ISRT Research Review 2010

Transcription

The International Spinal Research Trust
Annual Research Review
2010
The International Spinal Research Trust
Annual Research Review
2010
Research Division
Bramley Business Centre,
Station Road, Bramley, Guildford, Surrey GU5 0AZ, UK
Telephone: +44 (0)1483 898786
Facsimile: +44 (0)1483 898763
E-mail: [email protected]
Website: http://www.spinal-research.org
Along with donations from private individuals, Spinal Research is pleased to acknowledge the generous support of
Charitable Trusts, Charitable Foundations and other organisations, including those listed below. Their generosity
and the significant part they are playing in helping us to beat paralysis is greatly appreciated.
Annett Trust
Astor Foundation
Bernard Piggott Trust
Douglas Turner Trust
Dr. Scholl Foundation
Freemasons’ Grand Charity
Charles & Elsie Sykes Trust
Anna Rosa Forster Charitable Trust
Andy Stewart Charitable Foundation
Arthur James Paterson Charitable Trust
Birmingham Hospital Saturday Fund Medical Charity & Welfare
David Saunders Family Charitable Trust
Constance Travis Charitable Trust
Ernest Kleinwort Charitable Trust
Charles Wolfson Charitable Trust
Bournville Charitable CAF Trust
G C Gibson Charitable Trust
Gerald Palmer Eling Trust
F.J. Wallis Charitable Trust
Eveson Charitable Trust
G J W Turner Trust
GlaxoSmithKline
Grimmitt Trust
Oddfellows
Inman Charity
John Avins Trust
Henry Smith Charity
Hasluck Charitable Trust
Murrayfield Centenary Fund
Miss J K Stirrup Charity Trust
Norman Family Charitable Trust
Ninth Duke of Newcastle’s 1986 Charitable Settl
Henry Lumley Charitable Trust L & R Gilley Charitable Trust
Sir Joseph Hotung Charitable Settlement
William A. Cadbury Charitable Trust
Sir Cliff Richard Charitable Trust
Robert Luff Foundation Limited
Simon Gibson Charitable Trust
Sir Samuel Scott of Yews Trust
Tallow Chandlers’ Company
Zochonis Charitable Trust
Souter Charitable Trust
Stour Charitable Trust
Swire Charitable Trust
PF Charitable Trust
Welton Foundation
Trust PA
Contents
Welcome from the Chairman of the Scientific Committee
4
Funding for Research
5
Research Network Meeting
Agenda
Meeting Report
List of Delegates
7
8
11
15
Strategy Grants
Recent Awards
17
19
Nathalie Rose Barr PhD Studentships
Recent Awards
20
22
Research Reports
23
Non-integrating lentiviral expression of GMCSF to promote spinal cord regeneration
Francia Carolina Acosta Saltos, G. Raivich, P. Anderson, A. Thrasher
24
Measuring central nervous system plasticity
Karen Bosch, J.W. Fawcett, S.B. McMahon
31
Do omega-3 fatty acids modify inflammatory changes following a spinal cord compression injury?
Jodie C.E. Hall, J. V. Priestley, V.H. Perry and A. Michael-Titus
42
AAV8shRNA-RhoA and AAV8nt-3 transfection of dorsal root ganglion neurons (DRGN) in vivo mediates
neuron survival and disinhibited regeneration of dorsal column (DC) axons
Steven J. Jacques, Ann Logan, Martin Berry and Zubair Ahmed
47
Promoting spinal cord repair by genetic modification of Schwann cells to over-express PSA
Juan Luo, Dr Yi Zhang and Dr Xuenong Bo
57
Spinal cord diffusion imaging: challenging characterization and prognostic
Torben Schneider, Claudia Wheeler-Kingshott, Daniel Alexander
60
Project Grants
70
Investigation into the conduction properties of surviving axons following chronic spinal cord contusion and whether
therapeutic intervention can restore normal function
Katalin Bartus, Elizabeth Bradbury and Stephen McMahon
71
Rewiring the central nervous system following spinal cord injury using neurotrophins and rehabilitative training
Karim Fouad and Wolfram Tetzlaff
77
Developing mTOR-based strategies to promote axon regeneration and functional recovery after spinal cord injury
Zhigang He
81
Optimising recovery by facilitating plasticity
Lyn B. Jakeman and D. Michele Basso
86
Comparative evaluation of surgical and pharmacological methods for removal of a mature scar in a chronic spinal
cord injury model and subsequent regeneration of stimulated sensory neurons through the treated wound
Daljeet Mahay, Ann Logan, Martin Berry, Zubair Ahmed & Ana Maria Ginzalez
99
Do experimental treatments for spinal cord injury induce functional plasticity in spared pathways?
John Riddell and Susan Barnett
102
Axonal Regeneration in the Chronically Injured Spinal Cord
Mark Tuszynski and Ken Kadoya
107
ISRT Scientific Committee Members
111
International Campaign for Cures of spinal cord injury Paralysis (ICCP)
113
ICCP Spinal Cord Injury Clinical Trials Guidelines
115
ICCP Participating Organisations
116
3
Welcome from the Chairman of the Scientific Committee
I would like to welcome you to the Annual Research Review
for 2010 in which we present progress reports for Nathalie
Rose Barr PhD studentships and project grants funded by
Spinal Research. I would like to start this issue by thanking
Prof John Priestley for his successful chairing of the
Scientific Committee over the last three years, and welcome
four new members to the Committee; two from the UK
(Prof Robin Franklin and Prof Karim Brohi) and two from
the USA (Prof Mark Tuszynski and Dr Lyn Jakeman).
This year Spinal Research was focused in the BBC
Lifeline Appeal presented by Richard Hammond. During
the appeal we heard from two long term spinal injuries
sufferers, Melanie Reid and Dan and his mum. They told
their moving stories and how they cope with day to day life.
Melanie Reid, a journalist was recently injured during a
horse riding accident and wrote daily about her accident in
the Times. From these moving appeals £127,843.58 was
raised to fund research. This is a fantastic sum of money
and will really help in funding excellent research projects.
Over this year three new project grant awards were made.
These went to: Dr Matthew Ramer at ICORD, University
of British Columbia for C$286,200 to study: “Peripheral
sympathetic and sensory plasticity in bladder/bowel circuitry
in chronic spinal cord injury” and Dr John Riddell,
University of Glasgow for £218,417 to study: “How does
function in long axonal tracts and local neuron circuits
change in the progression from acute to chronic stages of
spinal cord injury and how effective are cell transplants
performed at these stages? Lastly Dr Laurent Vinay from
CNRS Marseille €194,767 to study: “Modulation of
chloride homeostasis as a new target to treat spasticity and
chronic pain after SCI”. We also were able to fund two new
Nathalie Rose Barr PhD studentships which were awarded to
Dr Xuenong Bo at Queen Mary University of London,
“Promotion of neuroplasticity by modifying perineuronal
nets using polysialic acid” and Dr Lawrence Moon at King’s
College London “Overcoming spinal cord injury with
clinically-relevant sustained delivery of neurotrophin-3 to
muscles initiated after 24 hours or 4 months”
Also within these pages is a report from the 2nd joint
Spinal Cord Meeting of the Christopher and Dana
Reeve Foundation (CDRF/NACTN), Internationales
Forschungsinstitut fuer Paraplegiologie (IFP/EMSCI) and the
International Spinal Research Trust (ISRT) Network Meeting
“Spinal Cord Research on the Way to Translation” held in
Ittingen, Switzerland. In addition International Campaign
for Cures of Spinal Cord Injury Paralysis (ICCP) guidelines
and ICCP participating organisations are summarised.
We have an excellent grant holders meeting planned for
September 2011 and I hope we can forge new collaborations
and provide stimulating forums for the progress of research
into spinal cord injury.
Professor Susan Barnett
Chairman of the Scientific Committee
4
Funding for Research
The International Spinal Research Trust (Spinal Research)
had its inception at the Guildhall in the City of London in
1981, with the sole purpose to raise funds for research into
reparative treatments for paralysis caused by spinal cord
injury. In the intervening years the Trust has committed
nearly £19 million to a wide range of relevant research
projects, with significant advances made through these.
Considering the challenge faced when translating laboratory
results to clinical treatments, our funds must be considered
limited. Nevertheless the Trust has been able to support
essential research that has, in turn, enabled researchers to
secure major grants from government and other funding
bodies. This is one of the best ways to leverage our
supporters’ funds.
For a project to receive funding, the application must
pass through a thorough peer-review procedure. For this the
Spinal Research is privileged to be able to rely upon the
advice of a distinguished Scientific Committee as well as
external reviewers. Their comments, recommendations and
advice on each application are made available to the Trustees
of Spinal Research, who make the final decision on whether,
and to what extent, a project will receive funding.
The scientific referees take into account several factors
when making their recommendations. The feasibility of the
project, together with the ability of the applicant and their
team to undertake it, is assessed to ensure that all the work
funded is scientifically respectable and likely to reach a
tangible conclusion. In addition, the referees attach most
importance to the relevance of each project, both in relation
to the advertised topic and to advancing the field in the best
manner possible. Researchers may be expected, for example,
to transfer their experimental studies to the adult
mammalian spinal cord.
In accordance with the Trust’s Research Strategy review
document, (published in 2007), applications for funding
are normally accepted following an advertised call for
proposals on a specific area of interest. The frequency of
these rounds depends on the availability of funds.
Regrettably the Trust is not able to fund as much research as
we would like and competition for our grants is high; at
times, many excellent applications have had to be turned
away. Advertisements calling for proposals appear in the
scientific press (e.g. Nature), on our website and are sent via
e-mail to hundreds of researchers on our database. If you
are not already on this database and wish to be, please
contact our Research Department. After review of an initial
brief proposal, shortlisted applicants are invited to submit
a full proposal. Further information about all our research
programmes and the various projects that are supported
by Spinal Research can be found on our website at
http://www.spinal-research.org.
Financial support typically covers the costs of salaries for
young postdoctoral scientists based in laboratories that
already have some expertise in spinal cord injury research,
plus the necessary laboratory materials and consumables
to carry out the planned research. Assistance for items of
equipment, technical support and collaborative travel
essential to the project can be considered, but need strong
justification. Institutional overheads and administrative
costs are not covered in any award. Spinal Research aims
to be flexible and unsolicited proposals can be considered
at any time. However, any application that falls outside
our current phase of the Strategy would need to make
an exceedingly strong case, not only in research terms but
also in terms of its relevance to Spinal Research’s ultimate
goal. Applications are considered on a competitive basis
whenever possible.
The Trust has a number of funding streams that run
concurrently and are closely linked. These support projects
ranging from PhD studentships to our new Translational
Awards which are described briefly on the next page.
For further information on the ISRT Research Strategy, see:
Adams, M. et al., (2007) International Spinal Research Trust research. III: A discussion document. Spinal Cord 45, 2–14.
5
Initiated in 1998 and named in honour of a late benefactress,
these awards are aimed at encouraging the development of
talented, highly-motivated young scientists in the field of
spinal cord repair, in both clinical and basic science research
environments. Calls for project proposals are advertised as
detailed previously, with a straightforward, one-step, peer
review process. The successful project supervisors then recruit
suitable candidates. The PhD degree must be awarded from
a UK university and a high priority is given to collaborative
proposals between more than one laboratory or institution.
Support includes University fees, a stipend in line with that
offered by the Wellcome Trust, plus funds for consumables,
travel costs and IT equipment. Spinal Research has so far
provided support to 32 projects under this scheme. Typically,
the PhD student is recruited to a team that is already
established in the field of spinal cord injury research, where
they should receive an excellent quality of training and
support. As well as obligations within their own institution
all students are encouraged to attend and present their data
at research conferences and to attend our annual Research
Network Meeting.
Strategy Awards
These form the mainstay of our basic science programme
and are normally awarded following an internationally
advertised competitive call for proposals based on themes
identified by our Scientific Committee with reference to our
Research Strategy discussion document. Project grants are
generally for the support of postdoctoral researchers to
undertake the approved research plan over a period of up
to three years, plus necessary consumables, travel or
technical assistance. Support will also be considered for
equipment if essential to the project.
Translational Awards
Strategy awards are intended to predominantly cater for
basic science proposals, concentrating largely on hypothesisdriven research and discovery aimed at understanding the
pathology of spinal cord injury at the anatomical, cellular
and molecular level and to develop understanding of
therapeutic targets and treatment concepts. Ultimately, such
fundamental science must lead to benefits in the quality of
life for those with SCI through translation to clinical
application. We recognise that translational activities will
not flourish spontaneously and have therefore committed
to increasing activity on applied research in SCI through
new Translational Grant Awards.
As a member of the Association of Medical Research
Charities, the Trust follows their guidelines regarding best
practice, including peer review, monitoring, the use of
animals in medical research, and patient agreements.
Further information is available at www.amrc.org.uk. Once
a full application is approved for funding, the Trust
negotiates a legal contract with the Principal Investigator’s
Institution that details both parties’ responsibilities
regarding, for example, finance, reporting procedures and
intellectual property rights.
Guidelines for Applicants
Our guidelines for completing initial proposals or ‘letters of
intent’ are straightforward. We require a concise outline of
the proposed project on no more than two pages. This
should include:
•
•
•
•
•
•
•
•
•
a clear definition of the specific problem
a reasoned argument for how this problem will be
tackled
some background, with evidence from previous
published work and/or suitable preliminary data to
support the hypothesis
a brief plan of the proposed experiments
the predicted outcome(s)
potential pitfalls and how they will be overcome
how this work will move the field of spinal cord repair
forward
any potential direct clinical benefits
an outline of the proposed budget
Progress reports are of major importance to the Trust
because they are essential for both monitoring the project
and as part of our responsibility towards those who provide
the finances. For this reason fund holders are required to
submit an annual report of the work undertaken that
indicates the goals they expect to achieve in the forthcoming
year. Final reports on the work are also needed. In addition,
many of the major donors and charitable trusts who support
our work require specific progress updates.
If successful, the applicants are invited to complete a
full application form, where further details of these
considerations are required.
6
Network Meetings
added infusions of growth factors and blockers of inhibitory
molecules. Because of the necessity for direct intervention at
the lesion site it is essential that the first treatments are
delivered to a region of the cord where any collateral damage
from the surgery is unlikely to have significant adverse
effects on the patient. This makes it unlikely that the first
treatments should be delivered to patients with cervical cord
lesions, even though they are the group that would benefit
most from even minor regeneration. Spinal Research
considers that the most favourable group of patients for a
safe trial of the first treatments are those with functionally
complete lesions in the lower part of the thoracic cord.
After the first gatherings in 1999, The Trust’s annual
meetings have developed into a popular and important
event where we invite those involved with all of our current
grants to meet, discuss their research and learn more about
each other’s projects in a confidential environment. In
particular we are keen that all the postdoctoral researchers
and PhD students that we support attend, as well as
grantholders and project supervisors. For younger scientists
it is sometimes the first chance to present their work to their
international peer-group. In addition to the traditional oral
and poster presentations, open discussion sessions on
current themes and controversial areas of research can lead
to new avenues of investigation.
Therefore we began funding studies to develop
techniques for detecting functional, physiological and
structural changes over two or three spinal segments
following spinal cord repairs, and for high resolution
imaging of the progress of lesions and the behaviour of
implanted cells. Such assessment techniques with the
necessary resolution are not presently available in routine
clinical practice and it is our aim to have these in place in
advance of trials of interventions.
Details of the 2010 Network Meeting, including a
meeting report, can be found in the following pages. The next
meeting will be in London, 2nd and 3rd September 2011.
The first stage of this initiative began in 2000 with a
major collaborative project in the UK and the progress made
and techniques developed have been peer reviewed and
published.
The Clinical Initiative
Regardless of the present success in animal models of spinal
cord injuries, where axon regeneration has been induced for
up to 3 cm, there are considerable hurdles to be overcome
before any therapeutic strategy can be considered for testing
in human patients. Not the least of these is that
experimental treatments cannot be used safely until the
progress and effect of the treatments can be accurately
assessed. At present the techniques needed for this are not
sufficiently well developed. Therefore Spinal Research has
developed a unique initiative to develop techniques for the
clinical assessment of spinal injury treatments.
The second stage of the Clinical Initiative started in
2005. In this stage, researchers in spinal injury units in the
UK, Canada and Switzerland are testing and refining the
procedures developed in Stage 1. These studies involve
monitoring and assessing the effects of non-invasive
strategies such as weight-supported treadmill training,
repetitive transcranial magnetic stimulation (rTMS) and
functional electrical stimulation (FES) on patients with
spinal cord injuries. Funding for this second phase has now
finished but Spinal Research will continue to support the
further validation, development and clinical adoption of this
clinical toolkit.
Future therapy of spinal cord injury might involve
implantation of cells into the lesion site, to which could be
7
“Spinal Cord Research on the Way to Translation”
12th Research Network Meeting, 26–28 August 2010,
Switzerland
PROGRAMME – THURSDAY, 26TH AUGUST
Session I – Optimizing training approaches
Volker Dietz
Neuronal dysfunction in chronic spinal cord injury
Grégoire Courtine
Turning the balance of plasticity to your advantage
Session II – Assessment tools
Susan Harkema
Armin Curt
Huub van Hedel
Rüdiger Rupp
Neuromuscular recovery with activity dependent plasticity after neurologic injury
Advancing the appreciation of segmental changes in SCI
Does “no pain, no gain” apply to sensory-motor recovery after spinal cord injury?
From diagnostics to therapy – The possibilities of realtime gait analysis in the rehabilitation
of incomplete spinal cord injured subjects
Panel Discussion: Readouts for Clinical Trials
Poster Session (wine and cheese)
PROGRAMME – FRIDAY, 27TH AUGUST
Session III – Regeneration and Plasticity I
James Fawcett
Increasing the intrinsic regenerative ability of spinal cord axons
Martin Schwab
Spontaneous, training- and anti-Nogo-A antibody induced recovery after CNS injury
Zhigang He
PTEN deletion enhances the regenerative ability of adult corticospinal neurons
Session IV – Regeneration and Plasticity II
Joost Verhaagen
Molecular target discovery for neural repair in the functional genomics era
Mark Tuszynski
Combinatorial Approaches to SCI
Heike Vallery
Therabotics 2030
Session V – Stem Cell Treatments
Fred Gage
Modeling human spinal cord injury in vitro
Sam Pfaff
Preparation of clinical grade human astrocyte precursors from stem cells
Panel Discussion: Treatment Combinations and New Treatments in Development
Session VI – Clinical Trials
Michael Fehlings
Jane Lebkowski, Geron
Klaus Kucher, Novartis
Repair and regeneration of the injured spinal cord: from molecule to man
Development of human embryonic stem cells for therapeutic applications
Therapeutic anti-Nogo-A antibodies in acute spinal cord injury – Latest safety and
pharmacokinetic data from ongoing first-in-human trial
Panel Discussion: Problems of Clinical Trials in SCI
PROGRAMME – SATURDAY, 28TH AUGUST
Session VII – Physiology of the Injured Human Spinal Cord I
Steve McMahon
Cortical overexpression of neuronal calcium sensor 1 induces functional plasticity in spinal
cord following unilateral pyramidal tract injury in rat
Phil Waite
Studies on pain after dorsal root injury
John Riddell
Electrophysiological assessment of function in animal models of spinal cord injury
Session VIII – Imaging and Characterization of the Injured Human Spinal Cord
Patrick Stroman
Mapping of function in the injured human spinal cord by means of functional MRI
Spyros Kollias
Advanced techniques for imaging the spinal cord
Session IX – Physiology of the Injured Human Spinal Cord II
Karim Fouad
The challence with the balance: Wanted versus unwanted treatment effects
Lynn Jakeman
Glial bridges and endogenous cellular repair strategies in spinal cord injury
8
POSTER PRESENTATIONS
Timing of decompression in acute traumatic central cord syndrome associated with spinal stenosis
Bizhan Aarabi
The effect of non-integrating lentiviral expression of GM-CSF in the rodent central nervous system
Caroina Acosta-Saltos
Combined light stimulation of Channelrhodopsin-2 and Chondroitinase ABC treatment restores respiratory activity in
chronically C2 hemisected rats and reveals plasticity of spinal cord circuitry
Warren J. Alilain
Assessing transport of integrins in adult CNS axons in vivo
Melissa R. Andrews
CNS injury: development of a novel in vitro model
Sue C. Barnett
Characterisation of the stem cell-like population found within human olfactory mucosa biopsies
Sue C. Barnett
Characterising functional, anatomical and electrophysiological changes from acute to chronic stages of spinal contusion
injury
Katalin Bartus
Measuring CNS Plasticity
Karen Bosch
Neurotrophic factors restore locomotion in the untrained adult spinal rat
Vanessa S. Boyce
Experience-dependent plasticity and modulation of growth regulatory molecules at central synapses
Daniela Carulli
Significance of motor evoked potentials in the Abductor Digiti Minimi (ADM) muscle in the foot in incomplete Spinal
Cord Injury
Bernard A. Conway
Electrical perceptual threshold: reliability and validity of a test for cutaneous sensation in spinal cord injury
Peter H. Ellaway
Virtual reality for motor rehabilitation and functional pain treatment in incomplete SCI patients
Kynan Eng
Corticomotor representation to human arm muscle changes following cervical spinal cord injury
Patrick Freund
Investigating corticospinal tract integrity using diffusion tensor MRI following spinal cord injury
Patrick Freund
Disability, cortical reorganization and atrophy following spinal cord injury
Patrick Freund
Changes in trans-cranial MEPs and SSEPs in association with cellular injections into porcine spinal cord injury epicenters
after SCI
James Guest
The effects of eicosapentaenoic acid delivered as dietary treatment after spinal cord injury
Jodie C.E. Hall
9
Tropism of adeno-associated virus 8 for large diameter sensory neurons of dorsal root ganglia after direct injection or
intrathecal delivery
Steven J. Jacques
Multiple intrinsic and extrinsic factors restrict sensory axon regeneration in chronic spinal cord injury
Ken Kadoya
Development of a myelination assay of human neurons generated from HESCs
Bilal E. Kerman
Electrophysiological properties of bilateral VPL neurons after spinal cord hemisection injury
Li Liang
Delivery of Decorin to acute dorsal column lesion sites suppresses inflammation, scar formation and angiogenesis
Daljeet Mahay
In vitro assay to measure inflammatory response in human glial cells after injury
Maria Carolina Marchetto
Deciphering the regeneration-associated gene expression program: gene expression profiling of axotomized facial motor
neurons in conditional c-Jun knockout mice
Matthew R.J. Mason
Specific and synergistic functions of monoaminergic receptors in the control of spinal locomotion
Pavel Musienko
Local translation of MAP2K7 allows JNK-dependent neurite outgrowth
Olivier Pertz
Investigating the reactivation of structural plasticity after digestion of Chondroitin Sulfate Proteoglycans with
Chondroitinase ABC
Oliver Raineteau
A multi-step screening approach successfully uncovers novel genes involved in the regeneration-promoting properties of
olfactory ensheathing glia cells
Kasper C.D. Roet
Optimized diffusion MRI protocols for estimating axon diameter with known fibre orientation
Torben Schneider
Microstructural spinal changes detected by diffusion tensor imaging in chronic spinal cord injury
Torben Schneider
Ketogenic diet improves gross forelimb function and fine-motor skills after incomplete cervical SCI
Wolfram Tetzlaff
Transplantation of skin-derived precursors differentiated into Schwann cells at eight weeks after spinal cord contusion
Wolfram Tetzlaff
Neurochemical biomarkers concentrations in the CSF of patients with traumatic spinal cord injury
Henk van de Meent
The effect on migration of primary Schwann cells and SCTM41 cells expressing a modified chondroitinase ABC enzyme
Philippa M. Warren
Promoting the survival, migration and integration of Schwann cells after transplantation into spinal cord by overexpression
of polysialic acid
Yi Zhang
10
Joint Network Meeting
Ittingen, Switzerland
26–28th August 2010
an example of someone that moves from an assistive device to
no device often, and unsurprisingly, performing less well on
parameters relating to speed. Harkema described a new 4 stage
classification of function developed by the Network that allows
an assessment of function without assistive devices which
helped to remove much of the variability seen.
Mark Bacon, Director of Research
This year’s Annual Research Network Meeting was the
second to be jointly-hosted by Spinal Research, the
Christopher & Dana Reeve Foundation and Internationales
Forschungsinstitut fuer Paraplegiologie. Held once again in
the wonderful surroundings of the monastery at Ittingen near
Zurich, this 3 day meeting entitled “Spinal Cord Research on
the Way to Translation” welcomed over 100 delegates to hear
and discuss the latest developments in basic and translational
research in spinal cord injury. Topics presented and discussed
ranged from understanding and optimising functional
training, assessment tools, regeneration and plasticity as well
presentations on the physiology of the injured human spinal
cord and updates on ongoing clinical trials.
ARMIN CURT championed the need to appreciate the
importance of segmental changes in function if the field is to
evaluate the effectiveness of interventions. Shortcomings in
standard ASIA assessments are clear. The conversion rate from
ASIA A to ASIA ≥B in cervical patients is ∼30%. In paraplegic
patients this conversion is lower. Unfortunately, conversion
rates don’t relate to a patients outcome. Motor scores (upper
limb) also change to roughly the same extent regardless of the
level of injury in the cervical region. He argued that the
clinical value of many interventions will be best evaluated by
rather specific or detailed functional outcome measures (hand
function, postural stability, sensory feedback etc.) where the
level of lesion and segmental deficit is most relevant. Assessing
segmental function in this way is important for proper
stratification of patients and defining relevant outcome
measures for intervention with specific functional goals. Curt
went on to discuss the utility of changes in motor levels rather
than motor score, particularly for cervical injuries. Thoracic
injuries pose a greater problem; changes in motor function
are minimal in those with higher thoracic injuries; sensory
change are highly variable; and measures of independence are
less sensitive for this group. More reliable measures of sensory
function are therefore needed, he said, and suggested
dematomal SSEPs and contact heat evoked potentials
(CHEPs) as an adjunct measures to determine sensory and
pain function. Such techniques may also help distinguish
between changes in white matter and grey matter pathways.
Following opening remarks from Martin Schwab, VOLKER
DIETZ began the meeting by discussing neuroplasticity in SCI
and the changes that occur as the injury progresses from the
acute to chronic state. Whilst clinical experience shows that
after injury functional improvements in locomotion occur
with training, he highlighted the importance of feedback cues
to the cord during training to prevent neuronal dysfunction
that becomes established and maintained as time post injury
increases. He maintained that neuronal dysfunction was
independent of injury level.
Continuing the theme, GREGOIRE COURTINE examined
the substantial anatomical and functional remodelling of
spinal circuitries that occur spontaneously after severe spinal
cord damage which in turn lead to a progressive degradation
of functional capacities in the chronic stages of the injury. He
asked whether the neuroplasticity could be harnessed for the
good. To do this, Courtine employed a combination strategy
of neurorehabilitation enabled by electrical and
pharmacological stimulations. Using this system, remodelling
of the lumbosacral circuitries and spared intraspinal systems
around the lesion site allowed paralyzed rats to voluntarily
control the pharmaco-electrically activated spinal circuitry
and to regain the impressive capacity to not only walk but
initiate locomotion, walk freely over ground, cross obstacles
and climb stairs.
The ability to predict pathological complications such as
neuropathic pain is important to patient treatment and
management. HUUB VAN HEDEL discussed an emerging
sensory test called the electrical perceptual threshold (EPT)
test. This quantitative test is being evaluated as an adjunct to
standard tests in a number of labs around the world and is
considered a promising candidate to determine changes in
segmental level of lesion in experimental trials. Can this test
predict neuropathic pain? Preliminary data from a small
patient cohort suggest that higher levels of electrical perceptual
threshold may exist in patients at 1 month who have
neuropathic pain at 6 months. Van Hedel went on to discuss
the predictive power of a structured interview in predicting
the development of neuropathic pain. The two techniques
were similar in terms of sensitivity and specificity but the
questionnaire was reasonably sensitive and specific in
determining who would be pain free after 6 months.
Interestingly, in this study, pain did not appear to influence
functional outcome, although this was data collected over the
1st 6 months and not longer term. He finished with
preliminary data that suggested that pain was reduced
following active training.
SUSAN HARKEMA presented data from NeuroRecovery
Network database. The network aims to provide consistent
intensive locomotor rehabilitation across participating centres
and collect functional outcome measurements for analysis. Of
note was the significant variability in baseline clinical function
taken at time of enrolment with significant overlap in function
between ASIA C and ASIA D groups. Following enrolment to
the training programme there was significant functional
recovery within ASIA groups which was evident in individuals
even years after injury. Some deficits in outcome measures
were identified exemplified by individual case studies where
“non-responders” by standard measures had nevertheless
functional improvements leading to improved quality of life.
The type and use of assistive devices was also found to be a
significant contributor to variability in functional tests, citing
11
RÜDIGER RUPP identified ceiling effects for many clinical
gait assessment tools. He presented an instrumented gait
analysis system developed in Heidelberg with the potential
for real-time visual (kinematic) feedback to aid rehab strategy
and volitionally-driven rehabilitation. The application of this
system in patients clearly demonstrated there were benefits
to this system, particularly in those with mainly sensory
dysfunction. The next challenge is to develop compact
motion sensing devices that allows feedback guidance during
at-home rehabilitation.
neuronal intrinsic PTEN/mTOR activity represents a
potential therapeutic strategy for promoting axon regeneration
and functional repair after adult spinal cord injury.
JOOST VERHAAGEN discussed the utility of functional
genomics and systems biology techniques in providing a
mechanistic framework for microarray gene expression
profiling following injury and repair. Applying this approach
to the study of olfactory ensheathing cells (OECs) his group
have identified 819 genes that are regulated after olfactory
epithelium lesions. Cluster analysis showed obvious
coordination in regulated gene expression following the
lesion. Through this analysis he identified phagocytosis as an
integral function of OECs following injury. Phagocyctosis is
coordinate in a multi-step process that involves recognition of
cell (axon) debris in an opsonin-dependant pathway,
engulfment and phagosome formation requiring activation
of cytoskeleton and finally “waste management” via lysosomal
pathways. Interestingly, phagocytosis and waste management
are poor after SCI and the suggestion that OECs may be
professional phagocytes could be one of the neural tissue
repair-promoting properties of OECs. Verhaagen described
other targets identified using similar approaches which were
screened for axon growth promoting activity. Further work
will take these hits into in vivo studies.
JAMES FAWCETT kicked off the first session of day 2 by
returning to the basic biology of axon regeneration. While a
great deal of effort has focused on overcoming the inhibitory
environments of the CNS and injury site, robust regeneration
is still “an awfully long way away from where we need it to
be”. Can the poor regenerative drive of central axons be
improved upon? Describing his group’s recent work on
integrin biology, he suggested these membrane-bound
molecules may be important in equipping axons with the
cellular machinery to interact with the extracellular matrix
through which they are expected to grow. Citing the massive
upregulation of Tenacin-C (TC) following injury, he pointed
out that central axons nevertheless lack the corresponding TC
binding integrin. In vitro studies revealed extraordinary robust
axonal growth from DRGs transfected with TC-binding
integrin when plated on TC. Regeneration in in vivo studies
was more modest but this could be improved when neurons
were engineered to constitutively express Kindlin-1 and tallin,
both activators of integrin. These studies, however, identified
problems with axonal transport of integrin from cell body to
dendritic processes. Ways to unblock the transports problems
will help improve this strategy and reveal important new
biology in regenerative medicine.
MARK TUSZYNSKI summarised the multiple mechanisms
that inhibit and stimulate axonal growth and therefore the
requirement for combinatorial approaches to enhance repair in
SCI. Early work showed that combination of a permissive
cellular graft with growth factors only elicited growth into the
graft, but not beyond. This led to ever more elaborate
combinations, including pre-conditioning lesioning and NT3 sinks beyond the graft to enhance regenerative drive and
entice exit from the graft, respectively. More recent work
confirmed that such combinatorial approaches not only work
in acute lesions but also chronic (1yr post injury) models.
Adding growth factor gradients to the combinatorial schema
resulted in guidance of axons and synapse formation to
appropriate target regions distal to the injury, although this
was not accompanied by evidence of electrophysiological
activity. Despite this relative success, regeneration in motor
systems, in particular the CST, still remains extremely difficult.
Local actions of Nogo such as triggering the arrest of
growth cones and inhibition of neuritic outgrowth are well
established. Strategies to overcome inhibitory molecules, such
has Nogo-A, has yielded numerous antibody and decoy-like
interference concepts. In his talk, MARTIN SCHWAB presented
data suggesting Nogo may have alternate, non-local effects,
which act via specific signalosomes that are retrogradely
transported to the cell body to regulate the cell growth
programme. He presented the idea that Nogo acts as a tonic
growth suppressor in the adult CNS in a corollary to growth
factors, signalling the end of development and stabilising
neuritic wiring.
HEIKE VALLERY discussed the development and clinical
application of robotic technologies in the fields of
neurosurgery and rehabilitation. A vision of the future sees
more and more use of robotics in all phases of patient
management within the hospital and home setting.
ZHINGANG HE’S talk discussed the effects of manipulating
the PTEN/mTOR pathway to increase regenerative drive.
Encouraging early work in the optic nerve model had now
been translated to models of cord injury. Upon the completion
of development, the regrowth potential of CST axons is lost
and this is accompanied by a down-regulation of mTOR
activity in corticospinal neurons. Axotomy triggers yet further
down regulation. Forced up-regulation of mTOR activity in
corticospinal neurons by conditional deletion of PTEN, a
negative regulator of mTOR, enhances compensatory
sprouting of uninjured CST axons and even more strikingly,
enables successful regeneration of a cohort of injured CST
axons past a spinal cord lesion. Furthermore, these
regenerating CST axons possess the ability to reform synapses
in spinal segments distal to the injury. Thus, modulating
The models used to study SCI are useful but there remain
questions about how well rodent-based studies (in vitro or in
vivo) model the human condition. FRED GAGE described
some of the first attempts to use human cells in in vitro assay
systems to model aspects of SCI from inflammation,
myelination, Glial response and basic physiology. The
ultimate aim of this work is to provide in vitro tools for
higher-throughput assessment of biological processes and
therapeutic targets for SCI. SAMUAL PFAFF continued the
theme of exploring the use of human-derived stem cell, this
time as therapies. After an introduction of current commercial
endeavours within the field, he outlined a translational
approach his group have adopted which focuses on ALS to
12
tease out the practicalities and feasibility of stem cell-based
strategies before application to other diseases. This decision
was based on the risk benefit analysis for experimental
treatments in ALS which is favourable in comparison to SCI
as the life expectancy from diagnosis in ALS is on average 2
years. Numerous regulatory hurdles exist including satisfying
the GMP requirements, appropriate scale-up and the
development of stereotactic systems etc. The therapeutic
concept is based around isolation of astrocyte precursors using
controlled defined culture protocols to achieve favourable
reparative astrocytic phenotypes.
6 repeated bolus injections of 45 mg antibody (total 270 mg).
Pharmacokinetics indicate reasonable attainment of target
lumbar CSF concentrations, although this in itself does not
inform on the tissue concentration. Non-human primate
preclinical studies provided further insight into antibody
tissue distribution. Novartis have extended their assessments
scheme to include bladder and hand function and a placebocontrolled Phase II study is planned.
Day 3 began with a return to discussions of basic cellular
biology. STEPHEN MCMAHON described the possible role of
the calcium sensing protein NCS-1 in regulating axon growth.
NSC-1 is a member of a large family of calcium binding
proteins that interact with numerous proteins including those
important in trafficking and secretion, mitochondrial
localisation and interestingly IP3R, which has implications
for control of neurite outgrowth. He reported on elegant in
vitro work by others that showed NCS-1 blockade resulted in
arrest of neurite outgrowth in DRG neurones. His own work
has shown increased sprouting in cortical neurons overexpressing NCS-1 which is associated with elevated levels of
pAKT. Blockage of the AKT-PI3K pathway inhibits this
sprouting. In vivo studies using lentiviral vector-mediated
NCS-1 expression in the cortex tested the hypothesis that
NCS-1 expression in cortical neurons would lead to increased
sprouting into denervated cord following unilateral
pyrmidotomy. Increased sprouting into denervated cord was
found, although modest. Behavioural and electrophysiological
measures improve with NCS-1 treatment. While the cellular
mechanisms are not yet known, NCS-1 may represent
another therapeutic target for SCI repair.
MIKE FEHLINGS provided a personal perspective and
overview of ongoing clinical trials sharing with the audience
some of the latest data from these. One of the key questions
still faced is what is the role of decompression and the
influence of timing of such procedures? Initial findings from
the Surgical Trial in Acute SCI Study (STASCIS) – a
prospective cohort study in cervical trauma patients with SCI
and cord compression – suggest early (<24 hour)
decompression is associated with better outcomes by the
ASIA Impairment Scale (AIS) at 1 year post injury. Safety and
feasibility would also look to have been established. Major
hurdles to achieving higher rates of early decompression are
logistical such as delays to referral. Further work is continuing
to validate these initial results.
JANE LEBKOWSKI updated the meeting on the Geron
oligodendrocyte progenitor cell trial for SCI. Geron’s product,
GRNOPC1, has three characterised activities; (i) production
of trophic factors promoting neuritic outgrowth, (ii) remyelination and, (iii) induction of neovascularisation.
Preclinical evaluation has included both thoracic and cervical
injuries with improvements reported on BBB scale.
GNROPC1 survives in the injured spinal cord long term
though mature very slowly. Animals that received
GRNOPC1 showed reduced cavity volume and persistent
myelination in the lesion site. Preclinical safety studies have
looked at biodistribtution, dosing and delivery regime,
toxicity/tumourigenicity and immune rejection. No product
was found in the brain or outside the CNS. The greatest
concentrations were detected around (<5 cm) the injury
epicentre. There was no evidence of clinical adverse effects
and importantly, no evidence of allodynia above that of
controls. A phase 1 safety study has been cleared to progress
to recruitment of patients with neurologically complete, subacute, SCI. For this, Geron have developed and gained
clearance for use of a syringe positioning device to make the
50 μl volume injection.
Pain is a common consequence of dorsal root injury. PHIL
WAITE compared 2 rat models of cervical root injuries, one
which eliminates all sensory input from the forepaw, the other
a partial forepaw denervation. Deficits in performing skilled
reaching and ladder walking tests were seen in both injury
groups, with the degree of impairment dependent on the
lesion severity. In the partial lesioned animals, persistent
mechanical allodynia and thermal hyeralgesia evolved in the
affected paw, whereas after the more severe injury, reduced
sensitivity occurred. Skilled motor tasks and forepaw
sensitivity were tested in the partial lesion model with and
without olfactory ensheathing cell (OEC) grafts. The delayed
OEC treatment had no effect on the performance of skilled
reaching or ladder walking. However, in OEC injected
animals, the extent of allodynia and hyperalgesia was reduced
from week 3 onwards compared to control animals.
Histological studies suggest that the antinociceptive effect of
OECs may be independent of changes in sprouting of spared
afferents from adjacent roots. Other potential mechanisms
suggested include modified dorsal horn excitability and
changes in the inflammatory responses within the dorsal horn.
The commercial perspective continued with a talk by
KLAUS KUCHER (Novartis). Novartis are clinically progressing
a recombinant monoclonal antibody therapy targeted against
Nogo-A. Anti-Nogo-A antibody works by interfering with
Nogo-A:receptor interactions and consequent downstream
interference of signalling pathways responsible for axon
regenerative arrest. The first-in-man study reported to be close
to completion with apparent safety and tolerability to the
procedure. Fifty one patients (thoracic and cervical ASIA A
injuries) have been treated at the time of reporting. The initial
method of delivery by continuous infusion was not deemed
optimal, with increased risk of infection and interference with
rehabilitation schedules. Thus, the latest cohort has received
Reliable and sensitive methods for assessing function in
animal models of spinal cord injury are essential to the process
of developing safe and effective therapies. Behavioural tests
provide a useful “global” indication of function but they do
not provide information on specific pathways or on the
mechanisms underlying functional changes. JOHN RIDDELL
described novel electrophysiological approaches that have been
developed to monitored changes in function of the
corticospinal system and sensory circuits within the cord. The
13
cord dorsum potentials (CDPs) technique records population
potentials on the surface of the spinal cord which represent
the strength of pathway connections within regions of the cord
below the electrode. Recordings can be made at, above and
below the lesion site to measure the strength of connections at
these sites during spontaneous, or treatment-induced recovery.
Spontaneous functional plasticity occurs following transection
and contusion injuries with plasticity in injured and spared
fibres differing in time-course. Using this technique Riddell
was led to conclude that there was little evidence that OEC
treatment enhanced plasticity suggesting instead it caused an
increase in white matter sparing and provided a general
neuroprotective effect. In contrast, anti-Nogo-A treatment did
enhance functional plasticity in both injured and spared fibres
of the corticospinal and sensory systems.
misplaced and greater efforts should be applied to cross
sectional imaging of the cord as opposed to segmental and
with this change in emphasis greater effort towards increasing
contrast between grey and white matter. High resolution of
the morphology of the cross-sectional cord can be achieved in
high field strengths scans. Even with 3-T scans reasonable
imaging can be achieved and a number of examples were
presented. Diffusion tension imaging (DTI) provides
additional information which might be useful to describe
integrity and geometric organisation of spinal cord tissue but
individual tracts look similar and morphological demarcation
is not currently possible. Nevertheless, DTI parameters were
shown to correlate to clinical (AIS) and electrophysiological
measures. He concluded by encouraging the integration of
advanced imaging techniques into, rather than isolated from,
clinical assessment of individual patients.
In a session devoted to imaging, PATRICK STROMAN began
by describing a new clinically-applicable method of functional
MRI of the human spinal cord (spinal fMRI). The potential
impact of non-invasive techniques such as this is enormous.
Stroman summarised some of the hurdles that must be
overcome when developing fMRI methods, such as the need
for quick data acquisition methods without loss of anatomical
information, and overcoming artefacts due to fixation devices.
He presented a novel method based on SEEP which, while
requiring longer data acquisition times, doesn’t require
specialised hardware and has advantages over faster alternate
methods such as BOLD fMRI which are susceptible to
artefact. This technique allowed him to map regions of
activity throughout the cord and brainstem simultaneously
during thermal stimulation to specific dermatomal regions to
demonstrate stimulation-dependant cord activity. Obvious
differences could be detected as a result of left or right
stimulation and between healthy and injured volunteers.
Global analysis of all regions during stimulation could reveal
areas of what he called “effective connectivity”. The 7 min
scan provided a rich data set which could provide a whole
cord signature of functional connectivity to provide
additional insight into changes after injury and therapy.
The meeting concluded with a session on the physiology
of the injured spinal cord. Plasticity is the substrate for much
of the spontaneous recovery that is witnessed after SCI. Many
putative treatments aim to enhance plasticity to augment the
recovery process. KARIM FOUAD took a reflective look at the
challenge of balancing wanted and unwanted effects of
treatments that enhance plasticity. Experimentally, researchers
often look at systems in isolation and the consequences of
treatments on other systems are secondary. The healthy cord
is exquisitely controlled: motor systems require a balance of
neuromodulation to set the excitatory levels for motor
neurons without which corticospinal activation doesn’t give
rise to motor activation. Despite this, staggered lesions, in
which descending neuromodulation is evidently lost, can
regardless result in recovery of walking. Fouad suggests this is
due to a phenotypic change in excitatory receptors from one
of ligand activation to constitutive activity. However, this
phenotypic change is associated with increased spasticity. He
showed evidence of enhanced serotonergic fibre infiltration
after combination treatment which resulted in increased grip
strength in rats that was clearly due to spastic closure of the
forepaw. Understanding the balance needed to achieve useful
function recovery without detrimental effects on other
function is clearly important and the consequences of driving
plasticity in one direction or over-stimulating plasticity must
be carefully evaluated.
SPYROS KOLLIAS continued the imaging theme. With its
inherent contrast sensitivity, the high spatial and temporal
resolution, the multiplanar sampling of anatomy, the reliable
differentiation between normal and pathologic tissue and the
lack of irradiation hazards, MR imaging has quickly emerged
as the study of choice for virtually all disorders of the spine.
Indeed many, if not all, the clinical trials included MRI as an
assessment within the clinical protocol. He asked whether this
was in efforts to understand changes that would relate to
therapeutic effect and mechanism of action or for purely safety
reason. His conclusion was the latter. For despite its potential,
MRI has disappointed in diagnostic ability. Numerous
motivations exist to continue to develop better techniques, not
least to gain better understanding of the biochemical,
physiological and functional consequences of injury. This, he
stated, would enable better diagnostic specificity, improved
treatment methods and a better understanding of when to
apply therapeutic interventions. MRI for the brain still leads
the way, leaving spinal cord imaging some distance to catch up.
Motion artefacts – particularly due to pulsatile flow of CSF –
remain major challenges but many gating techniques improve
this situation. However, he believes current emphasis on
increasing contrast between CSF and spinal cord tissue is
LYN JAKEMAN concluded the session and the meeting with
a discussion on astrocyte biology. Often seen as the problem
cell after SCI, new understanding of the role of glial
progenitors in regeneration now see therapeutic possibilities
for this maligned cell type. Jakeman presented data to support
the concept of stimulating endogenous astrocytes and
progenitors populations with TGF-α. Cell proliferation,
astrocyte migration and morphology were all modified after
intrathecal infusion of TGF-α which was accompanied by
increased axon growth into the lesion site. Similar results were
obtained following viral vector delivery of TGF-α. These data
demonstrate that localized administration of factors targeting
the glial response to injury may be a promising step toward
promoting endogenous repair after spinal cord injury.
The meeting also included a number of discussion
sessions which provided lively debate. The next Network
Meeting is planned for September 2011 in London.
14
Conference Delegates
Bizhan Aarabi
Carolina Acosta-Saltos
Albert Aguayo
Warren Alilain
David Allan
Aileen Anderson
Melissa Andrews
Mark Bacon
Sue Barnett
Katalin Bartus
Jesus Benito
Karen Bosch
Vanessa Boyce
Elizabeth Bradbury
Anita Buchli
Daniela Carulli
Daniel Chew
Bernie Conway
Gregoire Courtine
Armin Curt
Volker Dietz
Jean-Jacques Dreifuss
Isabelle Dusart
Peter Ellaway
Kynan Eng
Vieri Failli
James Fawcett
Michael Fehlings
Linard Filli
Karim Fouad
Ralph Frankowski
Patrick Freund
Fred Gage
Guillermo Garcia-Alias
Robert Grossman
James Guest
Jodie Hall
Susan Harkema
Jim Harrop
Zhigang He
John Hick
Susan Howley
Ronaldo Ichiyama
Steve Jacques
Lyn Jakeman
Ken Kadoya
Yorck-Bernhard Kalke
Naomi Kleitman
Spyros Kollias
Timea Konya
Jiri Kriz
Klaus Kucher
Jane Lebkowski
Roger Lemon
Juan Luo
Daljeet Mahay
Doris Maier
Didier Martin
University of Maryland
UCL
McGill University
Case Western Reserve University
SIU Glasgow
University of California
University of Cambridge
ISRT
University of Glasgow
King’s College London
Institute Guttmann
State University of New York
University of Zürich
University of Turin
University of Strathclyde
University Hospital Balgrist
University of Geneva
Universite Pierre at Marie Curie
Imperial College London
Wings for Life
Toronto Western Hospital
University of Alberta
University of Texas
UCL
Salk Institute for Biological Studies
The Methodist Hospital
University of Miami
Barts and The London School
University of Louisville
Thomas Jefferson University
Children’s Hospital Boston
ISRT
CDRF
University of Leeds
University of Birmingham
The Ohio State University
University of Ulm
NINDS
University Hospital of Zurich
ISRT
University Hospital Motol
N/Avartis Pharma AG
Geron
UCL
BG Trauma Hospital Murnau
University Hospital of LIEGE
15
Stephen McMahon
Lorne Mendell
Adina Michael-Titus
Pavel Musienko
Clara Orlando
Matt Pankratz
Olivier Pertz
Samuel Pfaff
Tiffany Poon
John Priestley
Olivier Raineteau
Geoff Raisman
Gennadij Raivich
Louis Reichardt
John Riddell
Frank Röhrich
Ferdinando Rossi
Rüdiger Rupp
Torben Schneider
Martin Schwab
Jan Schwab
Christopher Shaffrey
Prithvi Shah
Michelle Starkey
Andreas Steck
Patrick Stroman
Charles Tator
Wolfram Tetzlaff
Mark Tuszynski
Heike Vallery
Huub van Hedel
Joost Verhaagen
Phil Waite
Philippa Warren
Claudia Wheeler-Kingshott
Yi Zhang
Rongrong Zhao
Min Zhuo
Jens Zimmer
State University of New York and Stony Brook
University of Basel
UCL
UCL
Berufsgenossenschaftliche Kliniken Bergmannstrost
University of Turin
University Hospital Heidelberg
UCL
University of Berlin
University of Virginia
UCLA
University of Basel
Queen’s University
Toronto Western Hospital
ICORD
Netherlands Institute for Neuroscience
University of New South Wales
UCL
University of Toronto
University of Southern Denmark
16
Strategy Grants
Grant Holder
Location
Modulation of the glial response after spinal cord injury
Prof. A. Logan
Award
Grant Term
£140,971
2000, 3 years
Synthetic fibronectin conduits as guidance channels for directed regeneration within the spinal cord
Prof. J. Priestley
Queen Mary and Westfield College, London £135,292
2000, 3 years
Can olfactory epithelial ensheathing glia (OEG) successfully promote cross-species spinal cord tract regeneration?
Dr A.J. Roskams
University of British Columbia, Canada
C$245,565
2000, 3 years
Transplantation of fibroblasts genetically modified to secrete neurotrophic factors into spinal cord lesions in adult rats:
development of therapies for functional repair of acute and chronic spinal injuries
Prof. M. Murray
Medical College of Pennsylvania
$581,705
2000, 3 years
Hahnemann University
Generation and testing of rat and human ensheathing glia for spinal cord transplantation
Prof. P. Wood
The Miami Project, Florida, USA
US$278,783
2000, 3 years
& Prof M. Bunge
Development of the therapeutic potential of a combined glial cell/biodegradable substrate in functional tissue repair
following chronic injury of the adult rat spinal cord
Dr E. Joosten
Maastricht University, The Netherlands
€271,600
2001, 3 years
Restoration of spinal cord circuitry and function after nerve root injury in man
Mr T. Carlstedt
RNOH Stanmore, London
£188,989
& Mr R. Birche
2002, 3 years
Mechanisms of autonomic dysreflexia following spinal cord injury
Dr A.G. Rabchevsky
University of Kentucky, USA
& Dr G. Smith
US$213,694
2002, 3 years
A practical combination of ex vivo and in vivo gene therapy for spinal injury
Dr A. Blesch.
University of California, San Diego, USA
US$170,765
& Prof. M. Tuszynski
2002, 3 years
The role of the ubiquitin pathway in the preservation of functional axons after spinal cord injury
Prof. V.H. Perry
University of Southampton
£139,479
2002, 3 years
& Dr M. Coleman
Spinal axon sprouting: characterisation, manipulation and functional consequences
Dr M. Ramer
CAN$305,844
Genetic analysis of Nogo and Nogo receptor function in spinal cord regeneration
Prof. M. Tessier-Lavigne Stanford University, California, USA
US$486,398
(Grant finished because Prof. Tessier-Lavigne moved to Genentech)
2002, 3 years
2002, 3 years
Canine models of spinal cord injury: characterising and establishing the regeneration potential of canine olfactory
ensheathing cells
Dr N. Jeffery
£184,765
2002, 3 years
Prof. R. Franklin
Enhancing the role of propriospinal neurons in the recovery of motor function after spinal cord injury
Dr K. Fouad
University of Alberta, Canada
C$271,113
2003, 3 years
Synaptic connectivity in regenerating neurones of descending motor tracts
Dr P. Kirkwood
Institute of Neurology, London
£206,274
17
2003, 3 years
Early selective blockade of intraspinal inflammation is neuroprotective and leads to improved motor, sensory and
autonomic outcomes
Dr D. Marsh
The John P. Robarts Research Institute,
C$68,500
2003, 2 years
Canada
Standing and stepping with intraspinal microstimulation after spinal cord injury
Dr V. Mushahwar
C$135,000
2005, 3 years
Can human lamina propria olfactory ensheathing cells expand, migrate and stimulate rat SCI repair as well as mouse
OECs?
Dr A.J.I. Roskams
C$371,835
2005, 3 years
Neuroprotective strategies after spinal cord injury
Prof. W. Tetzlaff
C$194,928
2005, 2 years
Chemorepulsive axon guidance molecules in adult CNS axon regeneration failure: Class 3 semaphorins and
their receptors
Dr B. Zheng
University of California, San Diego, USA
US$286,955
2005, 3 years
Neuregulin growth factors in the repair of spinal cord axons by olfactory glia
Dr G. Raisman
NIMR Mill Hill, London
£147,734
2005, 3 years
Repair of adult rat corticospinal tract by transplants of olfactory ensheathing cells
Dr G. Raisman
NIMR Mill Hill, London
£1,460,901
Promoting axon regeneration in the injured spinal cord by RNAi-mediated knockdown of receptors for neurite
growth inhibitors
Prof. J. Verhaagen
Netherlands Inst. for Brain Research
€299,108
2004, 3 years
& Dr S. Niclou
Development of functional magnetic resonance imaging for assessing human spinal cord injuries
Dr P.W. Stroman
Queen’s University, Ontario, Canada
C$292,320
2005, 3 years
Role of microglia in spinal cord injury pain
Prof. S. McMahon
£197,777
2005, 3 years
Improving cardiovascular function after spinal cord injury
Prof. P. Waite
University of New South Wales, Australia
Aus$499,500
2006, 3 years
Dr John Riddell
£274,453
2008, 3 years
Dr Karim Fouad
University of Alberta
C$395,555
2008, 3 years
Optimizing recovery by facilitating plasticity
Dr Lyn Jakeman
Ohio State University
US$228,918
2008, 2 years
Investigation into the conduction properties of surviving axons following chronic spinal cord contusion and whether
Dr E. Bradbury
King’s College London, Guy’s Campus
£251,862
2009, 3 years
Prof. M. Tuszynski
US$ 253,200
2008, 3 years
Comparative evaluation of surgical and pharmacological methods for removal of a mature scar in a chronic spinal
cord injury model and subsequent regeneration of stimulated sensory neurons through the treated wound
Prof. A. Logan
£157,180
2009, 2.5 years
Locomotor training in chronic adult spinal cord injured rats: plasticity of interneurons and motoneurons
Dr R. Ichiyama
University of Leeds
£97,174
2009, 2 years
18
Autologous transplantation of Schwann cells and skin-derived Schwann cell precursors to repair the chronically
damaged primate corticospinal tract
Dr James Guest
University of Miami
US$319,561
3 years
Regeneration and plasticity of respiratory pathways in chronic spinal cord injured animals
Dr Warren Alilain
US$206,308
3 years
Dr Zhigang He
Children’s Hospital Boston
US$239,670
3 years
Recent Awards/Approvals
Grant Holder
Location
Award
Grant Term
Peripheral sympathetic and sensory plasticity in bladder/bowel circuitry in chronic spinal cord injury
Dr Matthew Ramer
ICORD, University of British Columbia
C$286,200
3 years
How does function in long axonal tracts and local neuron circuits change in the progression from acute to chronic
stages of spinal cord injury and how effective are cell transplants performed at these stages?
Dr John Riddell
£218,417
3 years
Modulation of chloride homeostasis as a new target to treat spasticity and chronic pain after SCI
Dr Laurent Vinay
CNRS Marseille
€194,767
3 years
Optimisation of Engineered Chondroitinase for Treatment of Spinal Cord Injury in Humans
Dr Elizabeth Muir
Cambridge Centre for Brain Repair
£232,805
1+1 year TBC
Development of a docosahexaenoic acid formulation for clinical studies on neuroprotection in spinal cord injury
Dr Adina Michael-Titus Barts and The London School of Medicine £190,141
1+1 year TBC
Repair of the avulsed brachial plexus by olfactory ensheathing cell transplantation: culture and manufacture of
autologous human olfactory ensheathing cells
Mr David Choi
UCL Institute of Neurology
£256,000
1+1 year TBC
Clinical Initiative
Grant Holder
Location
Award
Grant Term
Stage I
Development of procedures for assessment of functional and structural recovery following spinal cord injury in man
Prof. P. Ellaway et al.
Imperial College School of Medicine
£785,916
2000, 3 years
& NSIC Stoke Mandeville
Stage II
Functional and Neuronal Recovery in incomplete SCI: Interproject of the ISRT (clinical initiative) and the European
Multicenterproject (EM-SCI) for monitoring motor recovery in human SCI
Dr A. Curt
University Hospital Balgrist, Switzerland
SFr491,100
2004–2008
& Prof. V. Dietz
Outcome evaluation of FES-assisted exercise therapy for hand function in quadriplegic people
Prof. A. Prochazka
C$125,750
2004–2009
Treatments to aid recovery from spinal cord injury: testing improved clinical, physiological and functional assessments
Prof. P. Ellaway et al.
£419,626
2005–2009
& Royal National Orthopaedic Hospital,
London
19
Comprehensive evaluation of the physiological and functional adaptations induced by locomotor training in
incomplete spinal cord injured subjects
Prof. B. Conway
£246,732
2006–2009
Supervisor
Student
Location
Grant Term
Central neural regulation of autonomic and somatomotor function in human spinal cord injury
Prof. P. Ellaway,
Pietro Cariga
Imperial College School of
1999, 3 years
& Prof. C. Mathias
Medicine, London
Promotion of spinal cord regeneration by targeted neurotrophin gene transfer to ascending and descending neural
systems
Dr A. Logan
Elspeth Brown
Queen Elizabeth Medical
1998, 3 years
& Prof. M. Berry
Centre, Birmingham
Do OBECs have advantages over Schwann cells in their ability to mediate repair following transplantation into
astrocyte-containing areas of CNS damage?
Dr S. Barnett
Andras Lakatos
1998, 3 years
& Dr R. Franklin,
Comparison of the effectiveness of viral vectors and transfected cells for delivering neurotrophins to the injured
spinal cord
Prof. D.S. Latchman
Filitsa Groutsi
University College London
1998, 3 years
Promoting long-distance axonal regeneration and funtional reconnection using combined treatments for spinal
cord injury
Dr J. Riddell
Thomas Sardella
2005, 4 years
& Dr S. Barnett
Assessment of a novel chondroitinase-based strategy in promoting nerve regeneration and recovery of function after
spinal cord injury
Dr R. Keynes
Phillipa Warren
Cambridge Centre for Brain
2006, 4 years
& Prof. J.W. Fawcett
Repair
An investigation of the role of machinery intrinsic to the spinal cord in human movement
Dr J. Iles
Alima Ali
University of Oxford
1999, 4 years
Comparative characterisation of the signalling mechanisms activated by OECs and Schwann cells during growth and
migration into astrocyte-rich environments
Dr S. Barnett
Richard Fairless
2000, 4 years
& Dr M. Frame
The use of antisense connexin 43 as a neuroprotective agent following spinal cord injuries
Dr D. Becker
Michael Cronin
2001, 4 years
The response of adult NG2+ glial cells to spinal cord injury
Dr A. Butt
Paul Hubbard
2000, 4 years
Magnetic resonance imaging of transplanted glia
Dr R. Franklin
Mark Dunning
2001, 4 years
Mechanisms by which lens lesions promote axonal regeneration of central nervous system neurons
Dr D. Tonge
Aliza Panjwani
2001, 4 years
20
A neurophysiological study of residual supra-sacral sensory-motor pathways and their influence on sacral reflexes in
incomplete spinal cord injury
Prof. M. Craggs
Vernie Balasubramaniam
Royal National Orthopaedic
2002, 4 years
Hospital, University College
London
Inflammation and regeneration of dorsal column fibres
Prof. P. Richardson
Sadashiv Karanth
The Royal London Hospital
2002, 3 years
Spinal tract regrowth after block of ephrin signalling
Prof. S. Bolsover
Jez Fabes
2002, 4 years
Electrophysiological study on corticospinal and reflex organisation in incomplete spinal cord injury
Dr B. Conway
Isam Izeldin
2003, 3 years
Using small molecules to promote regeneration in the spinal cord
Prof. P. Doherty
Michelle Starkey
2002, 4 years
Role of RAS as an intracellular mediator of central axonal sprouting
Dr G. Raivich
Milan Makwana
2003, 4 years
Effects of chondroitinase treatment on axon and glial functions
Dr A. Butt
Maria Ovejero-Boglione
2003, 4 years
An investigation into the use of repetitive transcranial magnetic stimulation (rTMS) to improve functional recovery
after incomplete spinal cord injury in man
Dr N. Davey
Nick King
King Imperial College London 2004, 3 years
& Prof. P. Ellaway
Targeting Eph/Ephrin mediated inhibition at the damaged dorsal root entry zone (DREZ)
Dr I. Gavazzi
Philip Duffy
2004, 4 years
& Prof. J. Wood
siRNA knockdown of the p75/RhoA axon growth inhibitory pathway in DRG both in vitro and in vivo to promote
DRG neurite and dorsal column regeneration
Prof. A. Logan
Ruth Seabright
2004, 4 years
& Prof. M. Berry
Mechanisms involved in chronic pain after avulsion injury
Dr P. Shortland
Daniel Chew
Bart’s & The London School
& Prof. T. Carlstedt
of Medicine
Novel anti-inflammatory and neuroprotective strategies in spinal cord injury
Prof. J.V. Priestley
Jodie Hall
Queen Mary University
Dr A. Michael-Titus
of London
& Prof. V.H. Perry
2006, 4 years
2006, 4 years
The use of a YFP-expressing mouse in studies of spinal cord injury: mechanisms of chondroitinase-mediated repair
Dr E. Bradbury
Lucy Carter
2007, 3 years
& Prof. S. McMahon
Dr Y. Zhang
Lou Juan
Queen Mary University
2007, 3 years
& Dr X. Bo
of London
AAV8shRNA-RhoA and AAV8nt-3 transfection of dorsal root ganglion neurons (DRGN) in vivo mediates neuron
survival and disinhibited regeneration of dorsal column (DC) axons
Prof. A. Logan
Stephen Jacques
2007, 3 years
Spinal Cord Diffusion Imaging: Challenging Characterisation and Prognostic Value
Dr C.Wheeler-Kingshott
Torban Scneider
21
2008, 3 years
Prof. G. Raivich
Francia Acosta-Saltos
2009, 4 years
Promoting Neurological Recovery by Maximising Sensory-Motor Activation During Stepping and Walking:
development and assessment of robotics-assisted delivery platforms
Prof. B. Conway
2010, 3 years
Dr H. van Hedel
Mr D. Allan
Physiological changes accompanying plasticity
Prof. J. Fawcett
Karen Bosch
Prof. S. McMahon
Developing advanced MR imaging to assess spinal cord function and tract integrity
Dr C. Wheeler-Kingshott
Moreno Pasini
2009, 3 years
2010, 3 year
Recent Awards/Approvals
Supervisor
Location
Promotion of neuroplasticity by modifying perineuronal nets using polysialic acid
Dr Xuenong
Queen Mary University of London
Grant Term
2011, 3 years
Overcoming spinal cord injury with clinically-relevant sustained delivery of neurotrophin-3 to muscles initiated after
24 hours or 4 months
Dr Lawrence Moon
2011, 3 years
22
Reports by Nathalie Rose Barr PhD students
Francia Carolina Acosta Saltos, G. Raivich, P. Anderson, A. Thrasher
Karen Bosch, J. W. Fawcett, S. B. McMahon
Jodie C.E. Hall, J.V. Priestley, V.H. Perry and A. Michael-Titus
AAV8shRNA-RhoA and AAV8nt-3 transfection of dorsal root ganglion neurons (DRGN) in vivo mediates neuron survival and
disinhibited regeneration of dorsal column (DC) axons
Steven J. Jacques, Ann Logan, Martin Berry and Zubair Ahmed
Juan Luo, Yi Zhang and Xuenong Bo
Spinal cord diffusion imaging: challenging characterization and prognostic
Torben Schneider, Claudia Wheeler-Kingshott, Daniel Alexander
23
Non-integrating lentiviral expression of GMCSF to
promote spinal cord regeneration
Francia Carolina Acosta Saltos *, G. Raivich1, P. Anderson1, A. Thrasher2
1University College London, UK, [email protected], [email protected]
2 Institute of Child Health, UK, [email protected]
*PhD Student, [email protected]
INTRODUCTION
The difference in regenerative capacity between the central
(CNS) and peripheral nervous system (PNS) has been
primarily attributed to the non-supportive environment the
injured CNS presents to regenerating axons, compared to
the more permissive environment of the injured PNS
(Bounge, 2002; Filbin, 2003). However, the ability of
neuronal cell bodies to mount an appropriate response to
injury could also be a significant factor to create differences
in PNS and CNS (Anderson et al., 1998; Raivich et al.,
2004; Raivich and Makwana, 2007). In fact, there are
dramatic differences between the regenerative efforts
demonstrated by PNS and intrinsic CNS neurons, after
axonal injury. Neurons projecting axons to the PNS mount
a strong molecular response to axonal injury, upregulating
regeneration-related transcription factors, cytoskeletal
proteins and adhesion molecules. In contrast, neurons whose
axons are confined to the CNS usually mount a weak,
transient or incomplete response (Shokouhi et al., 2010).
mediated by activated macrophages. Direct injection of
isogenous macrophages into the DRG is enough to stimulate
central axonal regeneration (Lu and Richardson, 1991).
Additionally, the regenerative effect of intraocular injection
of zymosan has been attributed to the production of
oncomodulin by macrophages, a potent growth-promoting
signal that acts directly on the cell bodies of regenerating
RGC neurons (Leon et al., 2000). Like peripheral
macrophages, activated microglial cells – their CNS
counterparts – express a number of potentially cytotoxic
molecules such as TNF, IL-1beta, NO, oxygen radicals and
components of the complement cascade, which could impair
neuronal survival (Raivich et al., 1999). However, there is
accumulating evidence suggesting that microglia also
produce signals which change neuronal gene expression to
promote regeneration, as avid producers of many
neurotrophic cytokines and cell adhesion molecules such as
BDNF, IGF1, TGFb1, oncomodulin and osteopontin
(Bouhy et al., 2006; Schroeter et al., 2006; Yin et al., 2006;
Lalancette Hébert et al., 2007; Makwana et al., 2007) which
exert survival and regeneration-supporting effects
The neuronal cell body response to axonal injury is
accompanied by perineuronal inflammation around the cell
body of the axotomised neurons. Interestingly, while axonal
injury causes inflammation around PNS projecting
neurons, axonal injury to CNS neurons does not cause
perikaryal inflammation (Richardson and Lu, 1994). The
dorsal root ganglion (DRG) sensory neurons project axons
to the CNS, via the dorsal spinal root. When these centrally
projecting axons are crushed or cut, there is a very transient
and weak neuronal response to axotomy, with little or no
perikaryal neuroinflammation and slow or absent axonal
regeneration. However, when inflammation is induced
around the axotomised DRG neurons by stimulating
macrophages with Corynobacterium Parvum or by
peripheral conditioning, DRG neurons are able to extend
their axons into the CNS; past the dorsal root entry zone,
as far into as lamina II of the spinal cord (Lu and
Richardson, 1991). Similarly, retinal ganglion cells (RGCs)
only show very modest axonal regeneration after optic nerve
injury. Stimulating RGCs through macrophage activation,
via lens trauma or injections of yeast wall zymosam, causes
RGCs to extend large numbers of regenerating axons distal
to the site of optic nerve injury (Leon et al., 2000). The
improved axonal regeneration mediated by increased
perineuronal inflammation is associated with increased
upregulation of regeneration-related proteins, such as c-Jun
and GAP-43 by axotomised DRG and RGC neurons (Lu
and Richardson, 1991; Leon et al., 2000).
The aim of the current project has been therefore to
target local microglia using a non-integrating lentivirus
expressing granulocyte-macrophage colony stimulating
factor (GMCSF), to enhance and substantially prolong
microglial activation in order to augment the normally very
poor regenerative response of corticospinal motoneurons
following spinal cord injury. GMCSF is a potent microglial
mitogen (Raivich et al., 1991; 1993; Kloss et al., 1997),
responsible for the indirect proliferative effects of
proinflammatory cytokines IL1beta, TNFalpha and
lipopolysaccharide via enhanced GMCSF synthesis by
astrocytes (Malipiero et al., 1990; Kloss et al., 1997). It
induces the secretion of cytokines, and promotes
phagocytosis and functional antigenpresentation in cultured
brain-derived macrophages (Giulian and Ingleman, 1988)
as well as in vivo (McQualter et al., 2001; Mirski et al.,
2003; Ponomarev et al., 2007).
In addition to its intrinsic immunological actions,
intraperitoneal injection of recombinant GMCSF peptide
have been shown to be neuroprotective and to promote
functional recovery following spinal cord injury (Ha et al.,
2005; Bouhy et al., 2006; Huang et al., 2009). GMCSF has
been shown to strongly enhance microglial synthesis of
neurotrophins such as BDNF and promote neuronal survival,
increased expression of regeneration-related proteins such as
GAP-43 and increasing neurite outgrowth in vivo after spinal
cord injury (Ha et al., 2005; Bouhy et al., 2006; Huang et al.,
2009). However, previous studies demonstrating proregenerative effects of GMCSF concentrated on changes at
The positive effect of inflammation around neuronal
perikarya on CNS axonal regeneration appears to be
24
the site of spinal cord injury and focused on only particularly
short-term effects of GMCSF. The use of a non-integrating
lentivirus vector expressing the GMCSF gene permits the
analysis of the effects of prolonged exposure to GMCSF on
the CNS. In addition, appropriate pseudotyping of the
GMCSF virus and bicistronic expression eGFP permits
transfected corticospinal neurons and their axons to be
identified and analysed. This allows the effects of chronic
GMCSF administration on corticospinal axon regeneration
in the injured spinal cord to be studied.
packaging plasmid pCMVdR8.74D64V and the pVSVG
plasmid were used to transiently transfect 293T cells using
polyethyleneimine reagent. The produced eGFP and
GMCSF/eGFP NILV s were concentrated separately by
ultracentrifugation and their pellets were re-suspended in
OptiMEM 1 and frozen down at 80°C.
GMCSF/EGFP NILV bioreactivity assay
HEK 293T cells were treated with increasing titres of
GMCSF or eGFP-only NILV (Figure 2). Supernatant from
transfected HEK 293T cell, obtained 24 hours after
infection, was used to treat cells of BV2 microglial linage.
BV2 cells were maintained for 24 hours in: a) GMCSF
conditioned medium b) eGFP conditioned medium c)
medium with GMCSF recombinant peptide. Cell
proliferation levels were estimated using the 3(4,5dimethythiazol-2-yl)-25-diphenyltetrazolium bromide
(MTT) colorimetric assay, which correlates cell number to
formazan light absorbance measured with an optical reader.
METHODS
Plasmids and Subcloning
Non-integrating lentiviral vector (NILV) expressing eGFP
was created using the previously described integrase-deficient
second-generation plasmid pCMVdR8.74intD64V and the
viral genome plasmid pHR0SINcPPT-SEW (Yanez-Munoz
et al. 2006). pHR0SIN-cPPT-SEW contains the eGFP
expression cassette driven by the spleen focus-forming virus
(SFFV) promoter (Figure 1a). In addition, mouse GMCSF
and XIAP IRES genes were subcloned from their respective
pGL3-GMCSF and pMA-XIAP IRES plasmids into
pSL301, before being ultimately inserted into the lentiviral
genome plasmid pHR0SINcPPT-SEW. The resultant
lentiviral genome plasmid pHR0SINcPPT-SGXEW
contained the murine GMCSF gene driven by the SSFV
promoter, followed by the XIAP IRES element regulating
eGFP expression (Figure 1b). Efficient gene delivery to the
central nervous system (CNS) was achieved by pseudotyping
using the plasmid expressing the VSVG, pVSVG, as
previously described by Rahim et al. (2009).
Figure 2. NILV mediated production of GMCSF causes microglial
cell proliferation in vitro. Green and red bars demonstrate the effect
of increasing eGFP and GMCSF viral titres on microglial BV2 cell
density, respectively. Blue bars represent the effect of increasing GMCSF
recombinant peptide concentrations.
Figure 1. Schematic representation of the lentiviral plasmid
constructs used in the study. (A) The enhanced Green Flourescent
Protein (eGFP) lentiviral plasmid (pHR’SINcPPT-SEW). The
lentiviral plasmid carries Spleen Focus-Forming Virus promoter
(SFFV), driving the expression of eGFP. The promoter is preceded by
long terminal repeats (LTRs), a rev Response Element (RRE) and a
central PolyPurine Tract (cPPT) and followed by the Woodchuck
Hepatitis Post-transcriptional Regulatory (WPRE) element to stabilize
the eGFP mRNA. In this study this plasmid was used as a control to
the current GMCSF virus.
(B) The Granulocyte Macrophage Colony Stimulating Factor
(GMCSF) lentiviral plasmid was cloned by inserting the murine
Granulocyte Macrophage Colony Stimulating Factor (GMCSF) gene
(640 bp) and the xiap derived Interna Ribosome Entry Site (xIRES),
between the SFFV promoter and eGFP of the eGFP plasmid. SFFV
directs the expression of GMCSF and xIRES permits bicistronic
expression of eGFP.
Stereotactic injections to the CNS of adult rodents
Adult Sprague-Dawley rats and CD1 mice (both are
outbred strains) were anaesthasised with isoflourane and
placed in a stereotactic frame. The control eGFP-only virus
and GMCSF virus were delivered using a 10 μl Hamilton
syringe controlled by a micromanipulation pump. Rat
motor cortex and striatum were injected separately at a rate
of 200 μl/min. The following coordinates (relative to
bregma) were used for rat- motor cortex: 1 ml of virus
injected at 2 mm lateral, 1.5 mm ventral, 0.0, 0.5, 1.0 and
1.5 mm posterior to the bregma; striatum: 2 ml of virus
injected at 1 mm anterior, 3 mm lateral, 4 mm ventral.
Mouse motor cortex was injected at a rate of 50 μl/min. The
following coordinates were used for mice motor cortex:
0.5 ml of virus injected at 0.7 mm lateral, 0.5 mm ventral,
0.5, 0.7, 1.0 and 1.5 mm posterior to the bregma.
Cell culture and production of pseudotyped NILVs
Human embrionic kidney 293T cells were cultured in
Dulbecco’s modified Eagle’s medium supplemented with
10% fetal bovine serum and 5% penicillin/streptomycin.
NILVs were produced using a three-plasmid transient
transfection system. Plasmids encoding for the viral
genomes of the eGFP and GMCSF viruses, the gag-pol
Tissue processing and Immunofluorescence
All animals were sacrificed 14 days after viral injection using
sodium pentobarbitone. The rats were perfused
25
intracardially using phosphate-buffered saline (PBS),
followed by 2% paraformaldehyde and the brains were
removed and post-fixed for 1 hour in 2% paraformaldehyde
at 4°C. Then brains were cryoprotected in 30% sucrose
overnight, frozen on dry ice and then sectioned coronally
at the cryostat (30 μm).
revealed only a moderate number of eGFP+ cells and
comparatively large fluid filled spaces at the site of the
original needle track.
Immunofluorescence was used to detect microglia at
each of the rat and mouse injection sites. To visualize
microglia, sections were stained for IBA1 (1:1200) antibody.
Incubation with primary antibody was performed in PB
blocking solution bovine serum albumin. The sections were
thoroughly rinsed before applying CY3-conjugated
secondary antibody (1:100).
Rat tissue was double labeled for microglia and
granulocytes or microglia and DNA fragmentation. For
both stains the sections were left overnight with the primary
antibody against IBA1. The infiltrating neutrophil
granulocytes were detected by staining for endogenous
peroxidase. The enzyme was detected by covalently binding
of biotinylated tyramide (1% solution in PBS) for 10 min
at room temperature (RT) in the presence of 0.001%
hydrogen peroxide and by applying an Amca-Avidin
(1:100) for 1 hour. DNA fragmentation was analysed
using terminal Transferase mediated biotinylated d-UTP
nick end-labeling (TUNEL) and visualised with AMCAAvidin. For both stains after the application of
AMCA-Avidin, Cy3-secondary antibody was applied to
visualize IBA1 positive microglia.
Figure 3. GMCSF expressing non intergrating lentivirus causes high
levels of microglial activation in rat motor cortex. Top row – Coronal
section through a GMCSF NILV injected rat motor cortex. Bottom
row – eGFP-only NILV injection site. Left: microglial IBA1
immunostaining, Middle: eGFP fluorescence, Right: IBA1 (red) and
eGFP (green) merged. High magnification images emphasise the
difference in microglial morphology around eGFP positive cells
between the GMCSF NILV and the eGFP control injected animals.
Bar scale: 1.5 mm.
RESULTS
GMCSF producing viral vector enhances microglial
proliferation in-vitro
To test viral vector mediated production of GMCSF invitro; supernatant from eGFP-only and the GMCSF
non-integrating lentiviral vector (NILV) infected HEK
293T cells was applied onto BV2 microglial cells. As shown
in figure 2, the MTT colometric assay demonstrated no
effect of eGFP condition medium on the BV2 cell
proliferation. On the other hand, conditioned medium,
from HEK 293T cells transduced with GMCSF/eGFPONLY NILV elicited a 2.5-fold increase, with half-maximal
effect at 1:1000 dilution, equivalent to a viral titre of 100
plaque-forming units (PFUs). Addition of recombinant
GMCSF peptide to otherwise untreated BV2 cells served as
a positive control, and showed a 2.2-fold increase, and a
half-maximal effect at 1 ng/ml.
These data were matched by the microglial IBA1
immunostaining. Following GMCSF/eGFP NILV
injection, most local IBA1+ microglia/macrophages showed
rounded phagocytic morphology, with many densely packed
in clusters around the injection site. A large number of
microglial clusters were seen at the boundaries of the of fluid
filled spaces and tissue, but many spread as far as the white
matter of the internal capsule and corpus callosum, approx.
100 μm away from the eGFP+ cells. This spread and
high level of microglial activation was contrasted by the
more or less normal, ramified (resting) morphology for
microglia surrounding the needle track and the numerous
green-fluorescent cells following injection of the eGFP
only vector.
GMCSF induces microgliosis, granulocyte influx and
signs of local tissue damage in vivo
Similar findings were also observed following injection
into the striatum, shown in figure 4. Injection of eGFP-only
virus resulted in numerous eGFP+ neurons and minimal
microglia activation. GMCSF vector injected to the striatum
resulted in the appearance of fluid-filled spaces, low number
of eGFP positive cells and pronounced microglial activation,
extending 200 μm beyond the eGFP+ cells.
To test NILV mediated production of GMCSF in the
rodent CNS, Sprague-Dawley rats were injected with 1 μl
eGFP-only or the GMCSF/eGFP NILV (107 PFUs) in
the motor cortex. Fourteen days after injection, motor
cortices injected with eGFP-only NILV showed numerous
eGFP+ neurons up to 200 μm away from the needle track,
and normal consistency of injected tissue (Figure 3). In
contrast, rats injected with GMCSF/EGFP NILV,
26
recruitment did not extend beyond regions directly apposite
to the injection site, containing the eGFP+ cells.
Figure 4. Striatal effects 14 post-operatively. GMCSF NILV injected
rat striatum (top) displays strong microglial activation and few eGFP
positive cells, contrasted by a large number of eGFP cells and minimal
microglial activation with the eGFP-only virus (bottom). A&D:
Microglial IBA1, B&E: eGFP, C&F: IBA1 (red) and eGFP (green)
merged. Bar scale: 500 μm.
Figure 6. Granulocyte recruitment in GMCSF NILV injected rat
CNS. Triple immonoflourescence at the injection sites show eGFP in
green, the granulocyte endogenous peroxidase (EP) in red and IBA1+
microglia/macrophages in blue. A&E: Colocalisation of the three
stains at a section through the striatum (STR) after eGFP-only and
GMSF/eGFP NILV injection. B&F: eGFP fluorescence, C&G: EP
stain, D&H: IBA1 in b/w. Bar scale: 500 μm.
To explore the potentially toxic and pro-inflammatory,
local effects of GMCSF, we next stained the tissue for TUNEL
and granulocyte recruitment, at the 14 day time point. In the
eGFP-only NILV injected striatum, nuclear TUNEL staining
which identifies cell death associated with DNA damage,
revealed in section for section, very few or no TUNEL positive
cells, a large number of eGFP+ neurons and no sorrounding
phagocytic microglia. Striatum injected with GMCSF
revealed numerous TUNEL+ cells,large number of phagocytic
microglia and few remaining eGFP+ cells (Figure 5).
Figure 7. EP+ Granulocytes closely surround eGFP+ cells; phagocytic
microglia are spread further apart following injection of GMCSF NILV
(A-D) into the motor cortex. Both are barely present following injection
of eGFP-only NILV (E-H). A,E: all 3 fluorescences merged, as in figure
6. B&F: eGFP C&G: EP D&H: IBA1. Bar scale: 500 μm.
Since there is a wide choice of transgenic mutant mice
that could help us to study molecular mechanisms of
GMCSF on the CNS, we wanted to see if similar results
were also present in mice. As a starting point, we decided to
investigate the effects of eGFP only and GMCSF/eGFP
NILVs on the comparatively widely used CD1 mouse strain,
first using the previously chosen day 14 time point. To our
surprise, the results were more ambivalent. As seen in figure
8, even eGFP-only vector showed some microglial
activation, which was not evident in saline injected controls
(not shown). In mice injected with GMCSF/eGFP NILV,
there was a mixed result, a minority (– ¼) showing small
fluid filled area, clusters of activated microglia and a small
number of eGFP positive cells. In contrast, the majority of
mice had no tissue damage, a large number of eGFP positive
cells and little sign of exuberant microglial activation.
Similar, predominantly low level of activation was also
observed at other time points (day 1–7 and day 28).
Figure 5. GMCSF NILV is associated with local cell death in the
striatum (right). The eGFP fluorescence is shown in green, TUNEL
in red, and microglial IBA1 immunostaining in blue. A&E are
tricolour composites, B-D and F-H show individual stainings for
eGFP (B,F), TUNEL (C,G) and IBA1 (D,H), respectively. Bar scale:
500 μm.
Similar results were also observed for granulocyte
recruitment. Staining for granulocyte endogenous
peroxidase (EP) at eGFP-only NILV injections showed very
few granulocytes directly at the injection site, at the striatum
(Figure 6a) and motor cortex (Figure 7a). However,
GMCSF/eGFP NILV injections were associated with heavy
infiltration with EP+ granulocytes in the close proximity of
eGFP positive structures and the injection site. Unlike
microglial activation, GMCSF mediated granulocyte
27
CONCLUSION
As shown in the current study, the GMCSF non-integrating
lentiviral vector (GMCSF-NILV), combined with XIAPIRES element and eGFP expresses biologically active mouse
GMCSF both in cell culture and in the living animal. In
vitro, application of conditioned medium from GMCSFNILV transfected HEK293T cells on BV-2 microglial cell
line resulted in dose dependent proliferation, with half
maximal effect at 1:1000 dilution (100 PFU/ml). These
half-maximal effects were equivalent to the 1ng/ml dose of
recombinant mouse GMCSF. In vivo, lentiviral expression
of GMCSF in adult rat CNS was also associated with
extensive microgliosis in rat striatum and motor cortex, far
above that present in the CNS injected with the control,
eGFP-only NILV.
In addition to microgliosis, CNS injection of the
GMCSF/eGFP NILV also caused strong granulocyte
recruitment, on par with the role of GMCSF as a chemotactic
agent not only for macrophages, but also for granulocytes.
Interestingly, granulocyte recruitment confined to the directly
transfected areas demarcated by the XIAP-IRES eGFP
expression, suggesting that in case of granulocytes, the
cytokine works locally. In the case of microglia/macrophages
where the effects are more at a distance, it is possible that we
are observing secondary effects due to anterograde
deafferentation and/or retrograde response.
Unlike Sprague Dawley rats, injection of mouse
GMCSF-NILV into the CNS of CD1 mice reveal
considerable variability. Both rat and mouse strains are
outbred, but we only encountered the variability with the
mouse strain. The majority of the tested CD1 mice
demonstrated little microglial activation, and only a minority
the prominent microglial response comparable to rats. Mice
directly injected with recombinant GMCSF are known to
show strong local response; it is possible that the CD1 mouse
strain tested in this study does not produce or secrete
bioactive GMCSF protein but this remains to be confirmed
with in situ hybridisation and ELISA. It is also possible that
this is a strain specific effect; we are currently testing a wide
range of different mouse strains to resolve this problem.
Finally, microglia may also play a key role in the
neurotoxicity caused by the GMCSF-NILV. Compared
with eGFP-only NILV, the striata and motor cortices of
mice injected with GMCSF/eGFP NILV carrying XIAPIRES element and eGFP revealed only few eGFP+ cells,
high number of TUNEL positive cells, as well as numerous
ameboid macrophages spread up to a considerable distance
away from injection site. Microglia are well known to
mediate cellular damage in the presence of inflammation
(Raivich et al., 1999), and the GMCSF toxicity may be
mediated by activating microglia and inducing the release of
toxic substances such as TNF, IL-1beta, NO, oxygen
radicals and glutamate. Interestingly, gap junction
hemichannels are the main avenue of excessive glutamate
release from neurotoxic activated microglia (Takeuchi et al.,
2006), and as shown by the group from Neurol Dept,
U Nagoya, a pharmacologic blockage of these hemichannels
using blood-brain permeable small molecule derivatives of
Figure 8. Unlike the Sprague Dawley rats, the outbred CD1 mice
demonstrate a very variable microglial response to GMCSF virus, and
also respond to eGFP-only NILV. Top row, middle row and bottom
row correspond to high responders (HR), low responders (LR) and
eGFP-only controls, respectively. The monochrome images in the left
column show eGFP and in middle column IBA1. The right column
shows both fluorescences merged, with red for IBA1 and green for
eGFP. Bar scale: 500 μm.
28
the glycyretinic acid is associated with considerable
neuroprotection (Takeuchi et al., 2006 & 2008; Liang et
al., 2008; Yawata et al., 2008). We have just started a
collaboration with this group, to explore whether a
combination of these agents with GMCSF could prevent
toxicity, while producing strongly enhanced and nondestructive microgliosis.
inflammation and promotes survival in adult CNS.
J. Neurosci. 27:11201–13.
McQualter J.L., Darwiche R., Ewing C., Onuki M., Kay
T.W., Hamilton J.A., Reid H.H., Bernard C.C. (2001)
Granulocyte-macrophage colony-stimulating factor: a new
putative therapeutic target in Multiple Sclerosis. J. Exp.
Med. 194:873–881
Malipiero U.V., Frei K., Fontana A. (1990) Production of
hemopoietic colony-stimulating factors by astrocytes.
J Immunol. 144:3816–21
Mirski R., Reichert F., Klar A., Rotshenker S. (2003)
Granulocyte macrophage colony stimulating factor (GMCSF) activity is regulated by a GM-CSF binding molecule
in Wallerian degeneration following injury to peripheral
nerve axons. J. Neuroimmunol. 140:88–96
Ponomarev E.D., Shriver L.P., Maresz K., PedrasVasconcelos J., Verthelyi D., Dittel B.N. (2007) GM-CSF
production by autoreactive T cells is required for the
activation of microglial cells and the onset of experimental
autoimmune encephalomyelitis. J. Immunol. 178:39–48
Raivich, G., Bohatschek, M., Dacosta, C., Iwata, O.,
Galiano, M., Hristova, M., Wolfer, D.P., Lipp, H.P., Aguzzi,
A., Wagner, E.F., Behrens, A. (2004). The AP-1
transcription factor c-Jun is required for efficient axonal
regeneration. Neuron. 43, 57–67.
Raivich G., Bohatschek M., Kloss C.U., Werner A., Jones
L.L., Kreutzberg G.W. (1999) Neuroglial activation
repertoire in the injured brain: graded response, molecular
mechanisms and cues to physiological function. Brain Res.
Rev. 1999 Jul;30(1):77–105. Review
Raivich G., Makwana M. (2007). The making of successful
axonal regeneration: genes, molecules and signal
transduction pathways. Brain Res. Rev. 53:287–311.
Raivich G., Jones L.L., Werner A., Blüthmann H.,
Doetschmann T., Kreutzberg G.W., (1999). Molecular
signals for glial activation: pro- and anti-inflammatory
cytokines in the injured brain. Acta. Neurochir. Suppl.
73:21–30.
Richardson P.M., Lu X. (1994). Inflammation and axonal
regeneration. J. Neurol. 242: 57–60.
Schroeter M., Zickler P., Denhardt D.T., Hartung H.P.,
Jander S.(2006) Increased thalamic neurodegeneration
following ischaemic cortical stroke in osteopontin-deficient
mice. Brain. 129:1426–37
Shokouhi B.N., Wong B.Z., Siddiqui S., Lieberman A.R.,
Campbell G., Tohyama K., Anderson P.N. (2010).
Microglial responses around intrinsic CNS neurons are
correlated with axonal regeneration. BMC Neurosci.
5:11:13.
Takeuchi H., Jin S., Wang J., Zhang G., Kawanokuchi J.,
Kuno R., Sonobe Y., Mizuno T., Suzumura A. (2006)
Tumor necrosis factor-alpha induces neurotoxicity via
glutamate release from hemichannels of activated microglia
in an autocrine manner. J. Biol. Chem. 281:21362–8.
Takeuchi H., Jin S., Suzuki H., Doi Y., Liang J.,
Kawanokuchi J., Mizuno T., Sawada, Suzumura A. (2008).
Blockade of microglial glutamate release protects against
ischemic brain injury. Exp. Neurol. 214:144–6.
Yawata I., Takeuchi H., Doi Y., Liang J., Mizuno T.,
Suzumura A. (2008), Macrophage-induced neurotoxicity is
mediated by glutamate and attenuated by glutaminase
REFERENCES
Anderson, P.N., Campbell, G., Zhang, Y., Lieberman, A.R.
(1998). Cellular and molecular correlates of the
regeneration of adult mammmalian CNS axons into
peripheral nerve grafts. In Van leeuwen, F.W., Salehi, A.,
Giger, R.J., Holtmaat, A.J.G.D., Verhaagen, J. (Eds.).
Neuronal Degeneration and Regeneration: From Basic
Mechanisms to Prospects for Therapy (pp.211–233).
Amsterdam: Elsevier.
Bouhy D., Malgrange B., Multon S., Poirrier A.L., Scholtes
F., Schoenen J., Franzen R. (2006) Delayed GM-CSF
treatment stimulates axonal regeneration and functional
recovery in paraplegic rats via an increased BDNF expression
by endogenous macrophages. FASEB J 20:1239–41
Filbin M.T., 2003. Myelin-associated inhibitors of axonal
regeneration in the adult mammalian CNS. Nat. Rev.
Neurosci. 4:703–13.
Giulian D., Ingeman J.E. (1988) Colony-stimulating
factors as promoters of ameboid microglia. J. Neurosci.
8:4707–17
Ha Y., Kim Y.S., Cho J.M., Yoon S.H., Park S.R., Yoon
D.H., Kim E.Y., Park H.C., (2005). Role of granulocytemacrophage colony-stimulating factor in preventing
apoptosis and improving functional outcome in
experimental spinal cord contusion injury. J. Neurosurg.
Spine. 2:55–61.
Huang X., Kim J.M., Kong T.H., Park S.R., Ha Y., Kim
M.H., Park H., Yoon S.H., Park H.C., Park J.O., Min
B.H., Choi B.H. (2009) GM-CSF inhibits glial scar
formation and shows long-term protective effect after spinal
cord injury. J. Neurol. Sci. 277:87–97.
Kloss C.U., Kreutzberg G.W., Raivich G. (1997)
Proliferation of ramified microglia on an astrocyte
monolayer: characterization of stimulatory and inhibitory
cytokines. J. Neurosci. Res. 49:248–54
Lalancette-Hébert M., Gowing G., Simard A., Weng Y.C.,
Kriz J. Selective ablation of proliferating microglial cells
exacerbates ischemic injury in the brain. J. Neurosci.
27:2596–605
Leon S., Yin Y., Nguyen J., Irwin N., Benowitz L.I. (2000).
Lens injury stimulates axon regeneration in the mature rat
optic nerve. J. Neurosci. 20:4615–26.
Liang J.Y., Wang S.M., Chung T.H., Yang S.H., Wu J.C.
(2008) Effects of 18-glycyrrhetinic acid on serine 368
phosphorylation of connexin43 in rat neonatal
cardiomyocytes. Cell Biol. Int. 32:1371–9.
Lu X., Richardson P.M., (1991). Inflammation near the
nerve cell body enhances axonal regeneration. J. Neurosci.
11:972–8.
Makwana M., Jones L.L., Cuthill D., Heuer H., Bohatschek
M., Hristova M., Friedrichsen S., Ormsby I., Bueringer D.,
Koppius A., Bauer K., Doetschman T., Raivich G. (2007).
Endogenous transforming growth factor beta 1 suppresses
29
FUTURE PLANS
Future project work will centre on 4 specific themes:
A. Using titration experiments, we want to determine
GMCSF/eGFP NILV dose at which there is a minimal
cell death but microglial activation and up-regulation of
neuronal proteins associated with regeneration like cJun. These studies will also include experiments to
determine chemotactic potency of GM-CSF on
different blood borne leucocytes, including
granulocytes, blood-borne macrophages and T-cells
B. Stereotactic GMCSF/eGFP NILV injections into the
motor cortex combined with dorsal or dorsolateral
hemisection of the spinal cord to see the effect of
GMCSF on corticospinal tract regeneration and
functional recovery.
C. Screening of different inbred mouse strains, to
determine the presence or absence of GMCSF-NILV
sensitive strain lines, using combination of microglial
immunohistochemistry, in situ hybridisation for
GMCSF and ELISA/bioassays for GMCSF protein and
bioactivity.
D. Finally, the neurotoxic effect of GMCSF opens an
avenue of research on neuroinflammatory brain injury,
in the traumatically injured spinal cord, but also in
neonatal cerebral palsy, or in multiple sclerosis. Here,
inhibition of the GMCSF mediated damage using
pharmacological agents (e.g. hemichannel blockers)
could establish models where we can explore the effects
of changing destructive to non-destructive microgliosis
inhibitors and gap junction inhibitors. Life Sci. 2008
82:1111–6.
Yin Y., Henzl M.T., Lorber B., Nakazawa T., Thomas T.T.,
Jiang F., Langer R., Benowitz L.I. (2006) Oncomodulin is
a macrophage-derived signal for axon regeneration in retinal
ganglion cells. Nat. Neurosci. 9:843–52
PUBLICATIONS AND PRESENTATIONS
Poster presentation:
Acosta Saltos C., Gonitel R., Rahim A., Acosta Saltos A.,
Thrasher A., Anderson P., Raivich G. (2010). The effect of
non-integrating lentiviral expression of GM-CSF in the
rodent central nervous system. Poster presented at the Spinal
Research Network Meeting in Ittingen, Switzerland, 26th28th August 2010.
Gonitel R., Acosta-Saltos C., Mary Joy T., Anderson P.,
Raivich G., Thrasher A., (2010). Non-Integrating
Polycistronic Lentiviral Vectors for Use in the Central
Nervous System. Poster presentes at the 7th Annual
Conference of the British-Society-for-Gene-Therapy in
London, England, 29th-31st March 2010.
30
Karen Bosch*, J.W. Fawcett1, S.B. McMahon2
of Cambridge, UK, [email protected]
2King’s College London, UK, [email protected]
1University
INTRODUCTION
Central consequences of peripheral nerve injury
The peripheral and central components of the nervous
system are functionally integrated so it is unsurprising that
peripheral nerve injury often results in profound cortical and
subcortical reorganisation (Wall and Kaas, 1986; Chen et al.,
2002; Wall et al., 2002; Lundborg, 2003; Navarro et al.,
2007). For example, at the level of the spinal cord, incorrect
peripheral nerve regeneration leads to inappropriate
innervation of second order dorsal horn neurons and results
in an increase in receptive field size, altered efficiency of
central connections and a change in laminar projection
(Devor and Wall, 1978; Koerber et al., 1995). After
peripheral nerve regeneration the receptive field for a given
dorsal horn cell is discontinuous and increased in size due to
the disorganised regrowth of injured axons (Koerber and
Mirnics, 1996; Koerber et al., 2006). Receptive field size
decreases over time, indicating synaptic reorganization
(Devor and Wall, 1978, 1981; Koerber et al., 2006).
Peripheral nerve injury
Disruption of a peripheral nerve leads to a loss of the sensory,
motor and autonomic functions conveyed by that nerve.
This can lead to long-term debilitating consequences due to
the loss of motor and sensory function, as well as secondary
consequences such as neuropathic pain and psychological
suffering (Jaquet et al., 2001). Various factors influence the
success of surgical repair of a peripheral nerve: timing, type
of injury, type of repair, age of the patient and lesion location
(Hoke, 2006). Of these, the only factor that has made a
difference is refinement of surgical technique. However it is
widely accepted that these techniques have been optimally
refined (Lundborg, 2003). Despite these improvements the
functional recovery after nerve injury is often only partial;
indeed, in one study satisfactory return of function after
median or ulnar nerve injury in human patients was found
to be only 43% for sensory and 52% for motor function
(Ruijs et al., 2005). The main reason for suboptimal recovery
after peripheral nerve injury is likely to be due to misdirected
re-innervation, i.e. motor and sensory axons connecting with
inappropriate targets in the periphery despite early, optimal
surgical repair (Lundborg, 2000). Accordingly, it has been
found that when muscle efferent fibres innervate skin they
maintain their preinjury phenotype (Johnson et al., 1995)
and that synaptic efficiency can be rescued if motor neurons
re-innervate their native muscle but only partially rescued by
growing into skin (Mendell et al., 1995). Long after surgical
repair and successful peripheral nerve regeneration motor
behaviour remains uncoordinated; it has been postulated that
this is due to the failure of recovery of the muscle stretch
reflex (Alvarez et al., 2010).
As far as the cortex is concerned, early studies in nonhuman primates showed that partial hand denervation led
to cortical receptive field reorganisation within hours, with
takeover of the injured nerve’s receptive field by the
uninjured nerve fields (Merzenich et al., 1983b; Merzenich
et al., 1983a; Kolarik et al., 1994; Silva et al., 1996). Rapid
reorganisation in brainstem (Xu and Wall, 1997, 1999) and
spinal cord dorsal horn (Devor and Wall, 1978; Kohama et
al., 2000) has also been described after peripheral nerve
injury. In addition to these studies of acute central changes
after peripheral nerve injury, a vast literature exists that
describes the chronic changes at cortical, brainstem and
spinal cord dorsal horn level which occur extensively during
the first two months and then more gradually over a period
of 2 years, after which no further changes are detected
(Garraghty and Kaas, 1991b, a; Sengelaub et al., 1997;
Florence et al., 1998). The chronic changes in functional
connectivity at various CNS locations can be attributed to
axonal misdirection in the periphery (Lundborg, 2003).
Indeed, Nguyen et al. (2002) showed, using a transgenic
mouse with YFP-expressing motor axons, that reinnervation of target tissues is indeed erroneous after a nerve
transection and repair, but not after nerve crush injury,
when the guiding endoneural tube remains intact (Nguyen
et al., 2002). It has long been known that nerve crush injury
does not lead to a significant change in cortical (Wall et al.,
1983) or spinal cord (Devor and Wall, 1981) representation,
so it is due to inappropriate peripheral wiring of neurons
that CNS plasticity occurs.
Although clinical outcomes are far from perfect, the fact
remains that peripheral axons have the ability to regenerate,
grow considerable distances and re-innervate targets. This
is in stark contrast with the neurons of the central nervous
system (CNS), which do not show this remarkable ability to
regenerate. The difference is due to a variety of both
intrinsic and extrinsic factors. Richardson and colleagues
illustrated that the environment surrounding a regenerating
axon affects its ability to grow by demonstrating that CNS
axons would regenerate into peripheral nerve grafts
(Richardson et al., 1980). This finding corroborated
observations made by Ramon y Cajal many years earlier
(1928). The difference between the intrinsic responses of
the cell bodies of central versus peripheral neurons also
contributes to regenerative failure in the CNS, with
peripheral neurons exhibiting a shift from a state of
maintenance to one of growth by a change in a host of
regeneration-associated genes (RAGs) after transection
(Neumann and Woolf, 1999), whereas this upregulation in
genes is limited after CNS injury (Plunet et al., 2002).
The profound and inevitable axonal misdirection after
PNI repair and successful axonal regeneration leads to
inappropriate innervations of target organs (Koerber et al.,
1989; Guntinas-Lichius et al., 2005) and these inaccuracies
31
become permanently wired into the periphery. It becomes
clear that strategies for recovery of function must
concentrate on exploiting the ability of the CNS to
reconfigure neuronal connections following nerve repair.
Thus, modulating processes of CNS compensation and
adaptation by manipulating CSPGs may lead to
improvements in outcome after nerve injury. Employing
plastic changes to achieve an improvement in function in
this context may also show promise in other nervous system
injuries, for example stroke or spinal cord injury.
produce accurate forelimb flexion. The radial nerve is
antagonistic to the flexors as it is an extensor nerve,
supplying muscles that extend the forelimb. In this study
animals received brachial plexus injuries of graded severity,
always leaving the ulnar nerve intact, and functional
recovery was assessed through electrophysiological testing.
The reflexes studied are described below.
The simple stretch reflex
The monosynaptic stretch reflex causes contraction of both
homonymous and heteronymous muscles when spindles in
a muscle detect a change in length (Nelson and Mendell,
1978). At the same time, contraction of antagonist muscles
is inhibited via interneurons. The afferent volley is
conducted by large diameter, myelinated Ia fibres and the
contraction is effected by α-motoneurons. In these
experiments we have studied the heteronymous
monosynaptic stretch reflex produced in the ulnar nerve by
stimulation of an agonist, the median nerve, and compared
the elicited response with that produced upon antagonist
(radial) nerve stimulation. Although there is extensive
evidence for central reorganisation after peripheral nerve
regeneration (see above) this very stereotypical reflex has
been shown to persist after regeneration of a nerve to
innervate inappropriate targets (Eccles et al., 1960), even
though this means that the spinal cord is receiving incorrect
proprioceptive information. Here we investigate whether
inappropriate re-innervation with or without ChABC
treatment affects the characteristics of this reflex.
Central CSPG modification after peripheral nerve or dorsal
root injury
Perineuronal nets (PNNs) in the spinal cord require normal
activity in early life for their consolidation into net
structures around motor neurons in the ventral spinal cord
(Kalb and Hockfield, 1988, 1990; Takahashi-Iwanaga et al.,
1998) marking the end of the ‘critical period’ of plasticity
(Galtrey and Fawcett, 2007). Mice lacking tenascin-R, a
PNN component, have shown improved regeneration and
functional recovery following facial nerve injury and repair,
compared to wild-type littermates (Bruckner et al., 2000;
Guntinas-Lichius et al., 2005). Together with the evidence
that ChABC treatment of the cortex extends the critical
period of plasticity in rats (Pizzorusso et al., 2002), these
findings suggest that manipulation of PNNs may open a
window of opportunity for CNS plasticity to compensate
for peripheral nerve injury.
Galtrey and colleagues (2007) used peripheral nerve
injuries of varying severity to study the effect of central
ChABC treatment on functional recovery. The injuries used
were: crush; transection; transection with correct repair
(median-median, ulnar-ulnar); transection with incorrect
repair (median-ulnar, ulnar-median); and transection
without repair. These injuries provided degrees of
misguidance for the regenerating axons. The animals were
injured and left to regenerate axons for four weeks and then
a single intraspinal injection of ChABC was administered.
Behavioural testing showed improvements in skilled
forelimb function and grip strength. There was also an
increase in the number of newly-grown sprouts in the spinal
cord, as seen by microtubule-associated protein 1b
immunoreactivity. This suggests that ChABC treatment
caused local sprouting, which is then responsible for the
behavioural improvements (Galtrey et al., 2007). This study
provides evidence that reorganisation of spinal cord circuitry
due to increased permissiveness of the spinal cord
environment could compensate for inaccurate peripheral reinnervation. My project aims to build on this premise by
using electrophysiological measures to examine outcome
after misrouting of peripheral nerves. The measures I use
are outlined below.
Flexion withdrawal reflex
This reflex protects a limb by causing its rapid withdrawal
from a painful stimulus. A noxious cutaneous stimulus
causes polysynaptic activation of α-motoneurons
innervating multiple limb flexors, which contract in a
coordinated fashion to remove the limb from harm. The
flexion withdrawal reflex has been the subject of extensive
research, mostly aiming to elucidate pain mechanisms
(Wolpaw and Tennissen, 2001) and a commonly studied
phenomenon is wind up. Repetitive nociceptor activation at
a frequency of >0.3 Hz results in a progressive increase in
excitability of many spinal neurons following each stimulus
of a peripheral nerve (Mendell and Wall, 1965), leading to
an increase in ongoing activity, a lowered threshold and an
expansion of dorsal horn receptive fields (McMahon and
Wall, 1984; Cook et al., 1987). As this reflex is polysynaptic
it is conceivable that alterations of synaptic efficiency and
organisation could allow this reflex to adapt to the incorrect
information reaching the spinal cord from a misrouted
peripheral nerve. Here we investigate the effect of median
and radial nerve injuries of varying severity on wind up as
recorded in the intact ulnar nerve.
Spinal Reflexes
We have studied three nerves of the brachial plexus: the
median, ulnar and radial. The median and ulnar are both
flexor nerves; the median is responsible for hand flexion,
whereas the ulnar is more important in forearm flexion.
These two nerves are synergistic, working together to
Visualising synapses
For centuries there has been interest in the organisation of
the healthy and diseased nervous system. The discovery of
the axoplasmic transport system (Weiss and Hiscoe, 1948)
led to an explosion in the discovery of new and powerful
techniques to study axonal morphology. Some of the most
important tracers include: horse radish peroxidase
32
(Kristensson and Olsson, 1978), wheat germ agglutinin
(Schwab et al., 1978), cholera toxin B (Luppi et al., 1990)
and biotin dextran amine. This myriad of techniques for
visualizing axons has been widely used to reveal the
morphology if regenerating or sprouting axons after injury.
This report describes the development of a lentiviral vector
that will use the same mechanism of axonal transport to
express tagged synaptic vesicle proteins, synaptophluorin
(Miesenbock et al., 1998; Burrone et al., 2006) or
synaptophysin with a green fluorescent protein (GFP) tag.
This technique will specifically label synapses, providing a
quantifiable method of studying plasticity from an
anatomical point of view. The vectors I use have been
previously shown to be highly effective at transducing
neurons in vivo and in vitro (Yanez-Munoz et al., 2006; Yip
et al., 2010).
Aldrich). Percutaneous electrodes in the left and right
forelimbs recorded the electrocardiogram and body
temperature was maintained around 37°C using a
homeothermic blanket. Tracheotomy was performed and a
tracheal cannula inserted. The brachial plexus of the right
forelimb was exposed and the median, ulnar and radial
nerves were dissected free from surrounding connective
tissue and cut distally. Skin flaps from the incision formed
a pool which was filled with paraffin oil. The ulnar nerve
was mounted on silver wire hook electrodes for recording.
The median and radial nerves were electrically stimulated
in turn, while mounted on silver wire hook electrodes.
Fast reflexes. Electrical stimuli of increasing amplitude
from 50 to 500 μA (100 μs pulse at 0.5 Hz) were applied to
the median or radial nerves and ulnar nerve response at each
amplitude was recorded. Recordings were captured after each
of 5 pulses at increasing stimulus intensity (50 μA, 100 μA,
150 μA, 200 μA, 300 μA, 400 μA, 500 μA). A PC with
Scope software (ADInstruments) was used to capture
recordings. An average of 64 sweeps at 400 μA was calculated
online for each nerve and used to find the difference in
amplitudes of reflexes evoked by median and radial nerve
stimulation. This was achieved using software to calculate the
absolute integral of any response between 1.8 and 2.8 ms,
regardless of whether a response is observed qualitatively.
METHODS
Surgical Procedures
All experiments were undertaken in accordance with the
UK Animals (Scientific Procedures) Act 1986. Adult male
Wistar rats were used in this study. Rats were anaesthetized
with 60 mg/kg ketamine and 0.25 mg/kg medetomidine.
Body temperature was monitored rectally and used to
regulate a homeothermic blanket.
Peripheral nerve injuries. Incisions to the ventral skin and
pectoralis major muscle of the right forelimb were made,
exposing the brachial plexus near the axilla. The median and
radial nerves were identified and underwent one of three types
of lesion and repair: (i) both nerves were cut and tied off; (ii)
both nerves were cut and repaired correctly by self
anastomosis (median-median and radial-radial); (iii) both
nerves were cut and a crossover repair was performed
(median-radial and radial-median). Nerve transection was
performed using spring scissors and repair entailed one or two
stitches with 10–0 sutures (Ethicon, EthilonTM) to the
epineurium. Unrepaired nerves were ligated with a 4–0 suture
(Ethicon, EthilonTM). Overlying muscle and skin was sutured
in layers. 1 mg/kg atipamezole hydrochloride subcutaneously
was used to reverse the anaesthetic. Animals recovered in an
incubator and received 0.05 mg/kg buprenorphine
postoperatively. Animals were then left for up to 8 weeks in
order for axonal regeneration to occur before animals
underwent electrophysiological assessment. Animals
recovered uneventfully and did not exhibit autotomy.
Wind up. A train of 25 stimuli was delivered to the
median or radial nerves in turn at a stimulus intensity of
4 mA (1 ms pulse at 0.5 Hz). Recordings of ulnar nerve
activity for 1 second after each impulse were captured using
Chart5 software (ADInstruments). Ulnar nerve activity was
also recorded for 20 seconds prior to and 50 seconds after
the stimulation period. Three trials were carried out for each
nerve, with a 5 minute interval between trials to allow the
ulnar nerve to return to its resting level of activity. A multiunit recording of all spikes approximately 50% greater than
the noise was made during each second during the period of
stimulation and plotted as a graph. Area under the curve
analysis was performed and a repeated measures two way
ANOVA used to detect any statistical difference.
Generation of Lentiviral Vectors
Transfer plasmids. Two overexpression plasmids were
generated, in which the cytomegalovirus (CMV) promoter
reporter gene drives expression of synaptophlourin and
synaptophysin-GFP, respectively. The original lentiviral
transfer plasmid was assembled by Dr Ping Yip, and altered
by a commercial service (GeneScript) to include mCherry
(Invitrogen) as reporter in place of GFP. Dr Leon Lagnado
(Cambridge University) kindly provided synaptophluorin
and synaptophysin-GFP cDNA plasmids which were used
to clone the two genes into the transfer plasmid backbone.
The cDNA sequences were amplified by PCR with
restriction sites at the end of the primers (XhoI/XbaI for
synaptophluorin; XhoI/SpeI for synaptophysin-GFP). The
resulting products were then ligated into the backbone.
Successful insertion was ascertained by restriction enzyme
digestion. Sequencing was also carried out for further
confirmation (MWG Eurofins).
Viral vector delivery to the sensorimotor cortex. Rats
were anaesthetised as described above and fixed in a
stereotaxic frame. The skull was exposed and microinjections
were made using previously determined coordinates (Neafsy
et al., 1986). With reference to Bregma, these were: AP: −
1.5 mm, L: 2.5 mm; AP: −0.5 mm, L: 3.5 mm; AP:
+0.5 mm, L: 3.5 mm; AP: +1.0 mm, L: 1.5 mm; AP:
+1.5 mm, L: 2.5 mm; AP: +2.0 mm, L: 3.5 mm; all
injections were at a depth of 2 mm.
In vivo Electrophysiological Recordings
Preparation. Rats were terminally anaesthetised with an
intraperitoneal injection of 1.25 g/kg urethane (Sigma33
electrode along the length of the ulnar nerve. Several
responses of a longer latency were also observed after both the
median and radial nerve stimulation (fig 1); these were
generally of greater amplitude during median nerve
stimulation. To quantify the difference between the fast
responses provoked by median or radial nerve stimulation
half-wave rectification of an averaged trace was performed
and an integral calculated. Only the fast, approximately
2 ms latency wave was included in the analysis. The area
under the rectified curve produced by radial nerve stimulation
was found to be 26.3% (+7.9%, n=7; figure 1C) of the area
under the curve when the median nerve was stimulated. This
may seem surprising as figure 1B shows no fast reflex.
However this is due to difficulties in quantification of a subset
of the early data, where the baseline of the traces recorded
were ‘sloping’ and this did somewhat distort the results. This
is further discussed in the ‘Discussion’ section. However, the
trend is clear – over a 1ms duration between 1.8 and 2.8 ms
there is a much greater response in the ulnar after median
versus radial nerve stimulation.
Packaging. A third generation lentivirus packaging system
was used to package the newly generated vectors into virus
particles. This system has been previously described (Naldini
et al., 1996; Dull et al., 1998). Briefly, the transfer plasmid is
co-transfected with plasmids carrying essential viral genes
(pMDLg/pRRE, pRSV.REV) and the viral envelope gene
(VSV-G) into human embryonic kidney (HEK-293T) cells.
All plasmids were the generous gift of Dr Rafael Yanez-Munoz
(Royal Holloway, University of London). The transfection
was carried out using polyethylenimine (PEI) as a transfection
reagent. The reaction was allowed to proceed for 4 hours at
37°C before cells were washed and fed with complete
DMEM daily while packaging occurred. Virus particles were
harvested on days 2, 3 and 4 post-transfection via
ultracentrifugation at 50000g for 2 hours at 4°C. Particles
were resuspended and stored at −80°C.
Hippocampal Culture
Primary hippocampal neuron culture. Methods used to
prepare hippocampal neurons have been previously
described (Brewer et al., 1993). Briefly, embryonic day
18 foetuses were obtained from female Sprague Dawley rats.
The foestuses were decapitated and the hippocampi
dissected out and stored in Hanks’ balanced salt solution on
ice. The hippocampi were incubated for 15 minutes at 37°C
in 0.05% trypsin. Cells were dissociated by trituration and
seeded at a density of 45,000 cells per 11 mm poly-L-lysine
coverslip in Neurobasal medium containing 0.5 mM Lglutamine, 2% B27 and 1% Penicillin/Streptomycin. Half
the medium was replaced with fresh medium the following
day, and twice a week thereafter.
Culture transfection with lentiviral vectors. 1 μl of
synaptophysin-GFP or synaptophluorin lentiviral vector
was added to the neuronal media on day 4. Cultures were
fixed on day 16 with ice cold 4% paraformaldehyde,
blocked for 30minutes in normal goat serum and incubated
with rabbit anti-GFP (1:1000, Invitrogen) and mouse antiβ3 tubulin (1:1000, Promega) in PBS + 0.1% Triton +
0.01% azide overnight at room temperature. Neurons were
incubated with secondary antibodies (donkey anti-rabbit
Alexa 488 and goat anti-mouse 546, Invitrogen) for 2 hours
before being mounted onto glass slides.
RESULTS
A. Mono- and polysynaptic potentials recorded in
ulnar nerve
Intact animals
Upon median nerve stimulation a very stereotypical fast wave
was recorded from the ulnar nerve. This occurred above
100 μA stimulation and had a latency of around 2 ms (Fig.
1A). This very fast wave was not seen upon radial nerve
stimulation. The latency of this response is consistent with a
heteronymous connection of Ia spindle afferents in the
median nerve exciting α-motoneurons innervating muscles
in the ulnar nerve territory (Nelson and Mendell, 1978) and
I will thus refer to this wave as the monosynaptic reflex. The
slight variability in latency between animals (1.9–2.5 ms) is
likely to be due to differences in distance of the recording
Figure 1. Monosynaptic reflexes recorded in the ulnar nerve of control
animals. A: representative ulnar nerve recordings from an intact
animal. Stimulation of the median or radial nerve was at 400 μA for
100 μs. A fast reflex response at 2ms is observed after median but not
radial nerve stimulation. Longer latency, polysynaptic responses were
also evoked. B: Quantification of the reflex response after median and
radial nerve stimulation. Quantification was performed using half
wave rectification of an average of 64 sweeps at 400 μA.
Injured animals – monosynaptic
Rats underwent nerve axotomy with or without surgical
repair. They were then left for 8 weeks to allow the damaged
34
axons to regenerate and re-innervate targets. In order to
investigate the effect of these surgeries on low threshold
reflexes, electrophysiological testing was then performed. In
all cases median nerve stimulation resulted in a much larger
fast reflex than radial nerve stimulation (figure 2I).
Four rats underwent median and radial axotomy
without repair. Three out of the four of these rats showed a
fast reflex still present in the median nerve at –2 ms, an
example trace from one of these rats is shown in figure 2C.
The third rat in this group had a minimal fast reflex upon
median nerve stimulation and this explains the variability
seen in the data (figure 2I).
Eight rats underwent surgeries to cut and repair their
median and radial nerves. This configuration serves as
comparison to incorrect repairs, so that here the repaired
nerve will innervate the same territory as it did before injury.
As can be seen in figure 2E and 2F, the repaired nerves
exhibit exactly the same pattern of response to median nerve
stimulation as uninjured animals and again, as in all groups,
the response of the radial nerve is minimal. Figure 2I shows
the absolute size of the wave produced by stimulation and,
although the trend for median nerve fast wave to be larger
than the radial fast wave is consistently present, the absolute
size of the reflex appears to be bigger in injured animals than
that observed in uninjured animals. This is likely to be due
to human experimental factors (as mentioned later in the
‘Discussion’ section).
Figure 2. Monosynaptic reflexes recorded in the ulnar nerve for all
injury groups. A–H: representative traces of median and radial nerve
stimulation at 400 μA showing the presence of a fast reflex at
approximately 2 ms after median but not radial nerve stimulation. I:
quantification of fast reflex size calculated using half-wave
rectification of an average of 64 sweeps at 400 μA for the time period
of 1.8–2.8 ms. Across all injury groups the monosynaptic reflex is
larger after median nerve stimulation.
Five animals have undergone median and radial nerve
cross-anastomosis. All but one animal recorded a fast wave in
response to median but not radial nerve stimulation (figure
2G–I). Again, this also appeared to be even larger than that
elicited by median nerve stimulation in uninjured animals.
All animals here, regardless of injury severity, exhibited a
common trend: a consistent, stereotypical reflex always
present after median nerve stimulation, not after radial nerve
stimulation. Indeed, two way repeated measures ANOVA
analysis revealed no effect of injury on the magnitude of
monosynaptic reflexes, but did indicate a significant effect of
nerve (p<0.001). Thus I have pooled all results, calculated the
magnitude of the radial nerve fast reflex response as a
percentage of the size of the median nerve response for each
animal and found that, across all injury groups, animals have
an average radial nerve 22.3% (+ 5.3%, p<0.0001; n=21) the
size of that elicited by the median nerve (figure 3).
B. Wind up recorded in the ulnar nerve
Intact animals
Whole nerve recordings of the ulnar nerve response to either
flexor (median) or extensor (radial) nerve stimulation were
made at supramaximal C-fibre threshold (4 mA, 1 ms,
0.5 Hz). This means that both myelinated and unmyelinated
fibres would be activated. A train of 25 stimuli was delivered
at 0.5 Hz and the number of spikes evoked was recorded for
1 second immediately following each stimulus. Recordings
of nerve activity during the 20 seconds prior to, and the 50
seconds following, the period of stimulation were also kept.
Figure 3. Relative response to stimulation. Ulnar nerve response to
radial nerve stimulation is shown here as a percentage of response to
median nerve stimulation for each animal (p<0.0001; n=21).
During median nerve stimulation the number of spikes
recorded increased with each successive stimulus delivered
(figure 4A,B), reaching a plateau after a number of stimuli
(typically 15–20) and, in some cases, gradually declining.
After the end of the stimulation period a level of heightened
activity was maintained in most animals, before gradually
settling back to the resting levels of activity. This was not
35
Injured animals
Animals underwent median and radial nerve injury 8 weeks
before electrophysiological testing, as described above. Ulnar
nerve response to radial nerve stimulation was minimal in all
animal groups (figure 5). Wind up was abolished in those
rats that had undergone axotomy without repair, with
median and radial stimulation resulting in similar spike
frequencies (mean AUC: median 168.03, radial 193.03;
p=0.768, n=5). The group that underwent nerve injury and
direct repair responded, on average, like uninjured animals
with an average AUC of 563.9 from median and
283.7 from radial nerve stimulation. This difference was,
however, still only 44+11.8% of the wind up displayed by
the median nerves of control animals and also not
significantly more than radial repaired nerves (p=0.202).
Animals that underwent cross-anastomosis of their median
and radial nerves behaved in a remarkably similar way to
animals in the repair group. The radial (i.e. proximal radial
that has been sutured to the median nerve distally)
produced a very small AUC and median nerve stimulation
again wound up about half as much as a control median
nerve (45.7+7.9%).
generally the case during periods of radial nerve stimulation,
where the nerve did not exhibit this classic increase in
activity (figure 4B) or not to the same extent; rather the
number of spikes was mildly and consistently elevated
during the period of stimulation and stopped immediately
after the end of the stimulation period. An example trace
comparing the number of spikes per second is shown in
figure 4D, demonstrating these differences.
To quantify the difference in wind-up between the two
nerves we carried out area under the curve analysis on plots
of the number of spikes per second, exemplified in figure
4D, for each animal. The resulting graph demonstrates a
consistent difference between the two nerves (figure 4E;
n=8, p=0.039). This increasing response of one nerve with
each stimulation of a synergistic nerve is known as windup. It contrasts with the radial nerve response because
stimulation of the extensor does produce an increase in
ulnar nerve discharge but this does not further increase with
each successive stimulus, which is the essential characteristic
of wind-up. This robust difference between the flexor and
extensor capacity to produce wind-up in a synergistic flexor
nerve provides a convenient paradigm to study the effect of
cross-anastomosis of flexor and extensor nerves, as discussed
in the following section.
Figure 5. Effect of injury on wind up recorded in the ulnar nerve. A:
quantification of wind up using area under the curve analysis of spike
frequencies. There is a significant effect of nerve (p=0.004). Median
nerve stimulation leads to greater wind up than radial nerve
stimulation in control animals (p=0.001). Axotomy abolishes wind up
(p=0.012 axotomy vs. control, median nerve) and median nerves
repaired in any conformation show a trend towards recovery. Results
obtained using a two-way repeated measures ANOVA statistical test.
B: C fibre activation is required for the generation of wind up. The
median nerve of one uninjured rat was stimulated at 5 minute
intervals at various stimulus intensities. Only application of a stimulus
sufficient to excite C fibres caused wind up to occur.
Figure 4. Wind up recorded in the ulnar nerve. A: when a train of
high intensity, low frequency stimuli are applied to the median nerve
the number of spikes evoked in the ulnar increase with each successive
stimulus. B: this does not occur upon radial nerve stimulation. C:
representative recordings of 1 second duration following the first and
sixth stimulation of the median nerve of a control animal. D: number
of spikes per second before, during and after stimulation. Data from
one typical control rat.
36
From these results we see that wind up is affected by
nerve injury. It is abolished after axotomy and there is a trend
towards recovery of wind up in animals that have had their
nerve re-anastomosed, whether to their original distal stumps
or crossed over. Here we see a difference in result between
the high and low stimulation paradigms: after nerve injury,
stimulation at low intensity elicits an unchanged fast reflex,
whereas high intensity stimulation uncovered differences
between experimental groups of animals.
When wind up was first describes, it was soon
established that unmyelinated, afferent fibre activation was
required. This was confirmed in our experiments. At low
stimulus intensities, such as are known to activate large
diameter myelinated axons, repetitive stimulation at a
frequency of 0.5 Hz did not induce wind up (Woolf and
Wall, 1982). Higher stimulation intensity and duration
were required to produce a robust, consistent wind up.
Figure 5B shows response of the median nerve to
incremental stimulation intensities. Wind up is only present
at the highest intensity and widest pulse width used.
C. Lentiviral vectors as anatomical markers of
plasticity
We developed a technique for labelling presynaptic vesicles
by generating two non-integrating lentiviral vectors that
express the GFP-labelled proteins synaptophysin and
synaptophluorin. These vectors efficiently transfect HEK
cells and can be used to label synapses in rat embryonic
hippocampal neurons in vitro, as can be seen from the
punctate staining for GFP on dendrite stained for beta-3
tubulin (figure 6A, B). We have also shown that the vectors
can transduce cortical neurons in vivo (figure 6C, inset). In
addition to neurons, other cell types, presumably glia, are
also transduced by the viruses. These cells have a stellar
morphology (figure 6C) but they have not yet been
fully characterised.
Figure 6. Synaptophysin-GFP and synaptophluorin lentiviral vectors.
A, B: embryonic hippocampal neurons in culture, infected with
synaptophluorin synaptophysin-GFP. Green = anti-GFP staining. Red
= β-3 tubulin staining. Blue = Hoechst. C: grey-scale image of a rat
cortex that had been injected with synaptophuorin lentivector 4 weeks
previously. Anti-GFP staining shown. Inset: enlarged image of a single
infected neuron.
Low threshold, monosynaptic reflexes
Upon low intensity stimulation a stereotypical fast reflex
was observed upon median but not radial nerve stimulation.
This was confirmed by quantification of the amplitude of
any wave between 1.8 and 2.8ms. Negligible readings were
produced by radial nerve stimulation in this period, but the
fast reflex after median nerve stimulation was present and
large in every injury group. These results are consistent with
the classic studies which showed that monosynaptic Ia
afferent fibres of any particular muscle are restricted to the
motoneurons of its own and synergistic fibres and that the
monosynaptic reflex in kittens is unchanged by crossanastomosis, even when very young (Eccles et al., 1960). In
agreement with the literature, here we found that
monosynaptic connections persist after nerve injury despite,
in the case of crossed nerves, the inappropriate mature of
this type of reflex from a flexor to an extensor.
CONCLUSION
In these experiments we have aimed to develop new
measures of plasticity. Two electrophysiological
measurements of well-described spinal reflexes were
developed and used to investigate spinal cord plasticity: (i)
wind up of the flexion withdrawal reflex and (ii)
heteronymous monosynaptic connections of Ia afferent
fibres with α-motoneurons innervating synergist muscles.
Progress on the development of two lentiviral vectors that
will be used to specifically label synapses made by
corticospinal neurons was also described in this report.
Electrophysiology
In these experiments we have exploited the ability of a cut
and repaired nerve to regenerate down a surrogate nerve
trunk and successfully innervate new targets (Holmes and
Young, 1942). The median and radial nerves of rats
were operated on to make injury groups of increasing
severity: uninjured, self-anastomosis, cross-anastomosis and
axotomy. Recordings were always made from the intact
ulnar nerve.
From our data it appears that the monosynaptic reflex
evoked by median nerve stimulation is actually bigger in
injury groups than uninjured animals. This is unlikely
because, although most motoneurons survive after
peripheral nerve transaction, only approximately half of
these regenerate (Welin et al., 2008) and thus, if anything,
the monosynaptic reflex might be expected to be smaller
than in uninjured controls. An explanation for this apparent
37
increase in the reflex could be explained by a human factor:
experimenter experience. Most control experiments were
carried out before injured animals were characterized, thus
improvements in recording technique could be responsible
for this perceived increase in size. To negate technique as a
confounding factor, a number of control only experiments
have been carried out recently, in optimized recording
conditions, and more will follow.
in the injury groups studied here, more animals in all groups
will need to be characterised.
Wind up can differe between the different muscles that
are activated. It has been found, for example, that the
pattern of wind up development varies between recordings
from different muscle nerves in the hindlimb (Solano and
Herrero, 1999). It is therefore unsurprising that the pattern
of wind up described here may differ from that found by
others as we are recording from a whole, mixed nerve that
supplies a whole group of muscles in the forelimb.
Flexion withdrawal reflexes
These have long been used to study the phenomenon of
wind up and although most studies have involved recording
from dorsal horn neurons, the use of whole nerve recordings
has been described (Schouenborg and Sjolund, 1983; Woolf
and Wall, 1986). Stimulating and recording in the periphery
is advantageous because surgery is less complicated and the
outcome more physiological – the result of greater
integration and of supraspinal modulation (Herrero et al.,
2000). In these experiments this is particularly important
because the plasticity of interest may be mediated at
different levels and by various neuronal cell types. Indeed,
the final outcome of plastic changes, such as an adaptive
limb nmovement, is arguably more important than changes
at the level of individual synapses. The original wind up
experiments were carried out on decerebrate spinal cats,
recording from dorsal horn neurons (Mendell and Wall,
1965; Mendell, 1966) and the neurons that produced the
most robust wind up were found to be those with a wide
dynamic range (Schouenborg and Sjolund, 1983). In the
present study recordings were made from peripheral nerves.
Wind up in this type of preparation was first demonstrated
to share the characteristics with those described for dorsal
horn neurons during experiments in the cat (Price, 1972).
Schouenborg and Sjolund (1983) stimulated the sural nerve
and found that wind up of recordings from the common
peroneal nerve had similar characteristics to wind up of wide
dynamic range neurons. Our findings were that wind up of
the ulnar nerve after median nerve stimulation peaked
consistently after 15 stimuli and then decreased. The rate
of decrease varied between animals. In this respect our
findings differ from those of others who found that reflex
wind up in intact, anaesthetized peaked after 8–10 stimuli
(Schouenborg and Sjolund, 1983; Gozariu et al., 1997;
Solano and Herrero, 1999). In contrast many studies in
decerebrate animal preparations have described wind up
increasing up to and beyond 16 stimuli (Woolf and Wall,
1986; Cook et al., 1987; Gozariu et al., 1997).
The results for wind up in the injury paradigms show an
emerging pattern. Statistical analysis showed a significant
effect of nerve for animals of all groups, i.e. median nerve
stimulation consistently results in more evoked activity than
radial nerve stimulation. Specifically, control, repaired and
crossed nerves induce significantly more wind up of the
ulnar nerve than their radial nerve counterparts. It is clear
that injury reduces the development of wind up in all
injured groups. Repaired and crossed nerves are less
susceptible to this reduction than unrepaired nerves.
Electophysiology – summary
These preliminary results have provided us with an insight
into the characteristics of spinal reflexes in the nerves of the
brachial plexus and, importantly, show a clear difference
between two nerves – a flexor and an extensor. The next step
will be to investigate whether these reflexes change when
there is an environment of enhanced plasticity in the spinal
cord, using the bacterial enzyme Chondroitinase ABC.
Visualising synapses with lentiviral vectors
This report has described the development of two lentiviral
vectors that express GFP-tagged presynaptic proteins,
synaptophluorin and synaptophysin. The aim of this part
of the project is to open new avenues for the evaluation of
synaptic plasticity in the central nervous system. The viruses
infect neuronal cell bodies and then induce the production
of presynaptic proteins which are subsequently transported
to axonal terminals. We aim to transduce the corticospinal
tract, which projects to the spinal cord so virus is injected
into the brain and we aim to investigate the distribution of
the synapses made by this tract within the spinal cord. So far
during this project we have produced the lentiviruses and
shown that they can infect HEK cells and hippocampal
neurons in culture, with punctate staining for GFP observed
on the neuronal dendrites. Following this promising result
we aimed to transduce the corticospinal tract by performing
microinjections into the sensorimotor cortex of a number of
adult rats and showed that a number of cells could be
transfected. It remains to be investigated whether the
terminals of the cortical neurons in the spinal cord are
expressing the tagged proteins. These results will be
obtained shortly. Titration of the viruses in human
embryonic kidney cells (HEK) is also underway currently.
It is clear from observations that the integrity of the
spinal cord and the presence of anaesthesia affect the
characteristics of wimd up so this is likely to be due to
modulation of wind up by supra spinal areas (Herrero et al.,
2000). Indeed, electrical stimulation of the cord of a
spinalised animal reduced wind up (Hillman and Wall,
1969). The level of anaesthesia that we achieved was found
to be variable between preparations, despite the consistent
use of a dose of 1.25 g/kg of urethane. Variability was
observed in the weight of the animals, the length of time a
rat took to reach surgical anaesthesia, heart rate and
respiratory rate. This may explain some of the variability
observed in our data. In order to get a true idea of wind up
REFERENCES
Alvarez F.J., Bullinger K.L., Titus H.E., Nardelli P., Cope
T.C. (2010) Permanent reorganization of Ia afferent
38
synapses on motoneurons after peripheral nerve injuries.
Ann. N.Y. Acad. Sci. 1198:231–241.
Bruckner G., Grosche J., Schmidt S., Hartig W., Margolis
R.U., Delpech B., Seidenbecher C.I., Czaniera R.,
Schachner M. (2000) Postnatal development of
perineuronal nets in wild-type mice and in a mutant
deficient in tenascin-R. J. Comp. Neurol. 428:616–629.
Burrone J., Li Z., Murthy V.N. (2006) Studying vesicle
cycling in presynaptic terminals using the genetically
encoded probe synaptopHluorin. Nat. Protoc. 1:2970–
2978.
Chen R., Cohen L.G., Hallett M. (2002) Nervous system
reorganization following injury. Neuroscience. 111:761–773.
Cook A.J., Woolf C.J., Wall P.D., McMahon S.B. (1987)
Dynamic receptive field plasticity in rat spinal cord dorsal
horn following C-primary afferent input. Nature. 325:
151–153.
Devor M., Wall P.D. (1978) Reorganisation of spinal cord
sensory map after peripheral nerve injury. Nature. 276:
75–76.
Devor M., Wall P.D. (1981) Plasticity in the spinal cord
sensory map following peripheral nerve injury in rats.
J. Neurosci. 1:679–684.
Dull T., Zufferey R., Kelly M., Mandel R.J., Nguyen M.,
Trono D., Naldini L. (1998) A third-generation lentivirus
vector with a conditional packaging system. J. Virol.
72:8463–8471.
Eccles J.C., Eccles R.M., Magni F. (1960) Monosynaptic
excitatory action on motoneurones regenerated to
antagonistic muscles. J. Physiol. 154:68–88.
Florence S.L., Taub H.B., Kaas J.H. (1998) Large-scale
sprouting of cortical connections after peripheral injury in
adult macaque monkeys. Science. 282:1117–1121.
Galtrey C.M., Fawcett J.W. (2007) The role of chondroitin
sulfate proteoglycans in regeneration and plasticity in the
central nervous system. Brain Res. Rev. 54:1–18.
Galtrey C.M., Asher R.A., Nothias F., Fawcett J.W. (2007)
Promoting plasticity in the spinal cord with chondroitinase
improves functional recovery after peripheral nerve repair.
Brain. 130:926–939.
Garraghty P.E., Kaas J.H. (1991a) Large-scale functional
reorganization in adult monkey cortex after peripheral nerve
injury. Proc. Natl. Acad. Sci. U.S.A. 88:6976–6980.
Garraghty P.E., Kaas J.H. (1991b) Functional
reorganization in adult monkey thalamus after peripheral
nerve injury. Neuroreport. 2:747–750.
Gozariu M., Bragard D., Willer J.C., Le Bars D. (1997)
Temporal summation of C-fiber afferent inputs:
competition between facilitatory and inhibitory effects on
C-fiber reflex in the rat. J. Neurophysiol. 78:3165–3179.
Guntinas-Lichius O., Angelov D.N., Morellini F., Lenzen
M., Skouras E., Schachner M., Irintchev A. (2005)
Opposite impacts of tenascin-C and tenascin-R deficiency
in mice on the functional outcome of facial nerve repair.
Eur. J. Neurosci. 22:2171–2179.
Herrero J.F., Laird J.M., Lopez-Garcia J.A. (2000) Wind-up
of spinal cord neurones and pain sensation: much ado about
something? Prog. Neurobiol. 61:169–203.
Hillman P., Wall P.D. (1969) Inhibitory and excitatory
factors influencing the receptive fields of lamina 5 spinal
cord cells. Exp. Brain Res. 9:284–306.
Hoke A. (2006) Mechanisms of Disease: what factors limit
the success of peripheral nerve regeneration in humans? Nat.
Clin. Pract. Neurol. 2:448–454.
Holmes W., Young J.Z. (1942) Nerve regeneration after
immediate and delayed suture. J. Anat. 77:63–96 10.
Jaquet J.B., Luijsterburg A.J., Kalmijn S., Kuypers P.D.,
Hofman A., Hovius S.E. (2001) Median, ulnar, and
combined median-ulnar nerve injuries: functional outcome
and return to productivity. J. Trauma. 51:687–692.
Johnson R.D., Taylor J.S., Mendell L.M., Munson J.B.
(1995) Rescue of motoneuron and muscle afferent function
in cats by regeneration into skin. I. Properties of afferents.
J. Neurophysiol. 73:651–661.
Kalb R.G., Hockfield S. (1988) Molecular evidence for early
activity-dependent development of hamster motor neurons.
J. Neurosci. 8:2350–2360.
Kalb R.G., Hockfield S. (1990) Large diameter primary
afferent input is required for expression of the Cat-301
proteoglycan on the surface of motor neurons. Neuroscience.
34:391–401.
Koerber H.R., Mirnics K (1996) Plasticity of dorsal horn
cell receptive fields after peripheral nerve regeneration.
Koerber H.R., Seymour A.W., Mendell L.M. (1989)
Mismatches between peripheral receptor type and central
projections after peripheral nerve regeneration. Neurosci.
Lett. 99:67–72.
Koerber H.R., Mirnics K., Mendell L.M. (1995) Properties
of regenerated primary afferents and their functional
connections. J. Neurophysiol. 73:693–702.
Koerber H.R., Mirnics K., Lawson J.J. (2006) Synaptic
plasticity in the adult spinal dorsal horn: the appearance of
new functional connections following peripheral nerve
regeneration. Exp. Neurol. 200:468–479.
Kohama I., Ishikawa K., Kocsis J.D. (2000) Synaptic
reorganization in the substantia gelatinosa after peripheral
nerve neuroma formation: aberrant innervation of lamina II
neurons by Abeta afferents. J. Neurosci. 20:1538–1549.
Kolarik R.C., Rasey S.K., Wall J.T. (1994) The consistency,
extent, and locations of early-onset changes in cortical nerve
dominance aggregates following injury of nerves to primate
hands. J. Neurosci. 14:4269–4288.
Kristensson K., Olsson T. (1978) Uptake and retrograde
axonal transport of horseradish peroxidase in botulinumintoxicated mice. Brain Res. 155:118–123.
Lundborg G. (2000) A 25-year perspective of peripheral
nerve surgery: evolving neuroscientific concepts and clinical
significance. J. Hand. Surg. Am. 25:391–414.
Lundborg G. (2003) Richard P. Bunge memorial lecture.
Nerve injury and repair--a challenge to the plastic brain.
J. Peripher. Nerv. Syst. 8:209–226.
Luppi P.H., Fort P., Jouvet M. (1990) Iontophoretic
application of unconjugated cholera toxin B subunit (CTb)
combined with immunohistochemistry of neurochemical
substances: a method for transmitter identification of
retrogradely labeled neurons. Brain Res. 534:209–224.
McMahon S.B., Wall P.D. (1984) Receptive fields of rat
lamina 1 projection cells move to incorporate a nearby
region of injury. Pain. 19:235–247.
Mendell L.M. (1966) Physiological properties of
unmyelinated fiber projection to the spinal cord. Exp.
Neurol. 16:316–332.
39
Mendell L.M., Wall P.D. (1965) Responses of Single Dorsal
Cord Cells to Peripheral Cutaneous Unmyelinated Fibres.
Nature. 206:97–99.
Mendell L.M., Taylor J.S., Johnson R.D., Munson J.B.
(1995) Rescue of motoneuron and muscle afferent function
in cats by regeneration into skin. II. Ia-motoneuron synapse.
Merzenich M.M., Kaas J.H., Wall J.T., Sur M., Nelson R.J.,
Felleman D.J. (1983a) Progression of change following
median nerve section in the cortical representation of the
hand in areas 3b and 1 in adult owl and squirrel monkeys.
Neuroscience. 10:639–665.
Merzenich M.M., Kaas J.H., Wall J., Nelson R.J., Sur M.,
Felleman D. (1983b) Topographic reorganization of
somatosensory cortical areas 3b and 1 in adult monkeys
following restricted deafferentation. Neuroscience. 8:33–55.
Miesenbock G., De Angelis D.A., Rothman J.E. (1998)
Visualizing secretion and synaptic transmission with pHsensitive green fluorescent proteins. Nature. 394:192–195.
Naldini L., Blomer U., Gallay P., Ory D., Mulligan R.,
Gage F.H., Verma I.M., Trono D. (1996) In vivo gene
delivery and stable transduction of nondividing cells by a
lentiviral vector. Science. 272:263–267.
Navarro X., Vivo M., Valero-Cabre A. (2007) Neural
plasticity after peripheral nerve injury and regeneration.
Prog. Neurobiol. 82:163–201.
Neafsey E.J., Bold E.L., Haas G., Hurley-Gius K.M., Quirk
G., Sievert C.F., Terreberry R.R. (1986) The organisation
of the rat motor cortex: a microstimulation mapping study.
Brain Res. 396:77–96.
Nelson S.G., Mendell L.M. (1978) Projection of single knee
flexor Ia fibers to homonymous and heteronymous
motoneurons. J. Neurophysiol. 41:778–787.
Neumann S., Woolf C.J. (1999) Regeneration of dorsal
column fibers into and beyond the lesion site following
adult spinal cord injury. Neuron. 23:83–91.
Nguyen Q.T., Sanes J.R., Lichtman J.W. (2002) Preexisting pathways promote precise projection patterns. Nat.
Neurosci. 5:861–867.
Pizzorusso T., Medini P., Berardi N., Chierzi S., Fawcett
J.W., Maffei L. (2002) Reactivation of ocular dominance
plasticity in the adult visual cortex. Science. 298:1248–1251.
Plunet W., Kwon B.K., Tetzlaff W. (2002) Promoting axonal
regeneration in the central nervous system by enhancing the
cell body response to axotomy. J. Neurosci. Res. 68:1–6.
Price D.D. (1972) Characteristics of second pain and
flexion reflexes indicative of prolonged central summation.
Exp. Neurol. 37:371–387.
Richardson P.M., McGuinness U.M., Aguayo A.J. (1980)
Axons from CNS neurons regenerate into PNS grafts.
Nature 284:264–265.
Ruijs AC, Jaquet JB, Kalmijn S, Giele H, Hovius SE (2005)
Median and ulnar nerve injuries: a meta-analysis of
predictors of motor and sensory recovery after modern
microsurgical nerve repair. Plast. Reconstr. Surg. 116:484–
494; discussion 495–486.
Schouenborg J., Sjolund B.H. (1983) Activity evoked by Aand C-afferent fibers in rat dorsal horn neurons and its
relation to a flexion reflex. J. Neurophysiol. 50:1108–1121.
Schwab M.E., Javoy-Agid F., Agid Y. (1978) Labeled wheat
germ agglutinin (WGA) as a new, highly sensitive retrograde
tracer in the rat brain hippocampal system. Brain Res.
152:145–150.
Sengelaub D.R., Muja N., Mills A.C., Myers W.A.,
Churchill J.D., Garraghty P.E. (1997) Denervation-induced
sprouting of intact peripheral afferents into the cuneate
nucleus of adult rats. Brain Res. 769:256–262.
Silva A.C., Rasey S.K., Wu X., Wall J.T. (1996) Initial cortical
reactions to injury of the median and radial nerves to the
hands of adult primates. J. Comp. Neurol. 366:700–716.
Solano R., Herrero J.F. (1999) Response properties of hind
limb single motor units in normal rats and after carrageenaninduced inflammation. Neuroscience. 90:1393–1402.
Takahashi-Iwanaga H., Murakami T., Abe K. (1998) Threedimensional microanatomy of perineuronal proteoglycan
nets enveloping motor neurons in the rat spinal cord.
J. Neurocytol. 27:817–827.
Wall J., Felleman D., Kaas J. (1983) Recovery of normal
topography in the somatosensory cortex of monkeys after
nerve crush and regeneration. Science. 221:771–773.
Wall J.T., Kaas J.H. (1986) Long-term cortical
consequences of reinnervation errors after nerve
regeneration in monkeys. Brain Res. 372:400–404.
Wall J.T., Xu J., Wang X. (2002) Human brain plasticity: an
emerging view of the multiple substrates and mechanisms
that cause cortical changes and related sensory dysfunctions
after injuries of sensory inputs from the body. Brain Res.
Brain Res. Rev. 39:181–215.
Weiss P., Hiscoe H.B. (1948) Experiments on the
mechanism of nerve growth. J. Exp. Zool. 107:315–395.
Welin D., Novikova L.N., Wiberg M., Kellerth J.O.,
Novikov L.N. (2008) Survival and regeneration of
cutaneous and muscular afferent neurons after peripheral
nerve injury in adult rats. Exp. Brain Res. 186:315–323.
Wolpaw J.R., Tennissen A.M. (2001) Activity-dependent
spinal cord plasticity in health and disease. Annu. Rev.
Neurosci. 24:807–843.
Woolf C.J., Wall P.D. (1982) Chronic peripheral nerve
section diminshes the primary afferent A-fibre mediated
inhibition of rat dorsal horn neurones. Brain Res. 242:77–85.
Woolf C.J., Wall P.D. (1986) Relative effectiveness of C
primary afferent fibers of different origins in evoking a
prolonged facilitation of the flexor reflex in the rat.
J. Neurosci. 6:1433–1442.
Xu J., Wall J.T. (1997) Rapid changes in brainstem maps
of adult primates after peripheral injury. Brain Res.
774:211–215.
Xu J., Wall J.T. (1999) Evidence for brainstem and suprabrainstem contributions to rapid cortical plasticity in adult
monkeys. J. Neurosci. 19:7578–7590.
Yanez-Munoz R.J., Balaggan K.S., MacNeil A., Howe S.J.,
Schmidt M., Smith A.J., Buch P., MacLaren R.E., Anderson
P.N., Barker S.E., Duran Y., Bartholomae C., von Kalle C.,
Heckenlively J.R., Kinnon C., Ali R.R., Thrasher A.J.
(2006) Effective gene therapy with nonintegrating lentiviral
vectors. Nat. Med. 12:348–353.
Yip P.K., Wong L.F., Sears T.A., Yanez-Munoz R.J.,
McMahon SB (2010) Cortical overexpression of neuronal
calcium sensor-1 induces functional plasticity in spinal cord
following unilateral pyramidal tract injury in rat. PLoS Biol.
8:e1000399.
40
•
FUTURE PLANS
I intend to add to the results described in this report in the
following ways:
• Having characterised the flexor and extensor reflexes
before and after cross union injuries the question that
remains to be asked is whether these reflexes can
undergo plasticity in the presence of chondroitinase
ABC, an enzyme known to induce sprouting in the
CNS. Without treatment, when a flexor nerve is
connected to an extensor muscle the monosynaptic
reflex remains and wind up partially recovers. Should
chondroitinase induce any compensatory changes it
would be reasonable to deduce that stimulation of the
flexor regenerated to the extensor muscle should now
not produce a monosynaptic or any wind up in the
ulnar nerve as this is an inappropriate way for a flexor
to respond to extensor stimuli and not conducive to
functional recovery. More interestingly perhaps is the
prospect that an extensor nerve redirected to innervate
a flexor muscle should begin to behave like a flexor due
to central synaptic reorganisation. Thus the redirected
radial nerve may begin to provoke a wind up response
or monosynaptic reflex in the ulnar nerve where before
there was none.
41
The synaptophluorin and synaptophysin-GFP
lentivectors will have measures of their titres taken.
Whether expression at the level of the spinal cord has
been achieved also remains to be seen and will be tested
by immunostaing sections of spinal cord for GFP.
Synaptophluorin is a fusion protein of VAMP2 and a
pH sensitive GFP molecule. The molecule does not
fluoresce at an acidic pH, such as that found in the
lumen of vesicles, but does at a neutral pH. This means
that fluorescence is at a low basal level, increases during
exocytosis and returns to basal levels during
endocytosis. To address whether this function remains
after packaging and expression by the lentivirus we
transduced hipposampal neurons and, in collaboration
with others (Burrone lab), stimulated the cultures. We
observed that fluorescence increased during stimulation
and decreased after the stimulation period. The
latencies of the changes in fluorescence were consistent
with exocytosis and endocytosis. Thus we have tested
the principle that, not only can synapses be visualised,
but that they are functioning. I have not included these
results in the report as the data are still being analysed.
Should the vectors prove to label synapses in the spinal
cord it presents the possibility that we can test the
functionality of these synapses using a slice preparation.
Novel anti-inflammatory and neuroprotective strategies in
spinal cord injury
Jodie C.E. Hall1*, J.V. Priestley1, V.H. Perry2 and A. Michael-Titus1
Centre, ICMS, Barts and The London, Queen Mary University of London, London E1 2AT
[email protected]
2Southampton Neuroscience Group, School of Biological Sciences and School of Medicine, University of Southampton,
Southampton SO16 7PX
1Neuroscience
METHODS
INTRODUCTION
Following spinal cord injury (SCI), there is an inflammatory
response which principally involves the white blood cells,
including neutrophils, and microglia, the resident
macrophages of the CNS (Trivedi et al., 2006). These cells
release cytokines such as tumour necrosis factor-α (TNFα) and interleukin-1 (IL-1), which aid in recruiting
additional immune cells to the site of injury (Profyris et al.,
2004). Systemically, acute-phase proteins (APPs) such as Creactive protein (CRP) are released into the circulation.
After SCI, the inflammatory response in the spinal cord is
considered to drive further death and degeneration of nerve
fibres above and below the lesion site and this ultimately
leads to loss of function (Popovich et al., 1997; Dusart and
Schwab, 1994). Decreasing the response of white blood cells
and the systemic response has been shown to be
neuroprotective after SCI (Popovich et al., 1999; Bao et al.,
2005; Campbell et al., 2005; Fleming et al., 2008). This
means that there is a window of opportunity for
pharmacological intervention with the use of antiinflammatory treatments.
Experimental design
1. Acute DHA or EPA bolus delivery after compression
SCI
Adult female Sprague Dawley rats (220–250 g) underwent
50 g static compression for 5 min as described (Huang et
al. 2007a), at vertebral level T12. Animals received the
following treatments via a tail vein under brief isoflurane
anaesthesia (2%): an injection of (i) saline (vehicle) (n=46;
0.9% NaCl), (ii) DHA (n=29; 250 nmol/kg), or (iii) EPA
(n=26; 250 nmol/kg), 30 min after injury. The injection
volume was 5 ml/kg. Animals were perfused and tissue
processed as described below.
2. Dietary enrichment with EPA and compression SCI
Following compression SCI as described in experiment 1,
on the same day, animals received control (5KB3 Certified
EURodent; IPS Product Supplies Limited, UK) diet or
EPA-enriched diet (150–170 mg/kg/day; Incromega
DHA700E SR; Croda Healthcare, UK) for four weeks after
SCI (n=6 per group).
3. Acute DHA bolus delivery after contusion SCI
Adult female Sprague Dawley rats (200–220 g) received a
moderate spinal contusion injury at vertebral level T8 with
a preset force of 200 kDynes. Animals received the following
treatments via a tail vein whilst still under ketamine and
xylazine anaesthesia: an injection of (i) saline (vehicle)
(n=10; 0.9% NaCl), (ii) DHA (n=10; 250 nmol/kg), or (iii)
DHA (n=8; 500 nmol/kg) 30 min after injury. The
injection volume was 5 ml/kg. In all three studies, treatment
was allocated randomly and researchers were blinded to the
treatment groups.
Omega 3 polyunsaturated fatty acids (omega-3 PUFAs),
such as docosahexaenoic acid (DHA) and eicosapentaenoic
acid (EPA), are essential for the brain and appear to be
extremely effective treatments for several conditions
including CNS and inflammatory conditions such as
Alzheimer’s disease, multiple sclerosis and rheumatoid
arthritis. Work in our laboratory in the rat has demonstrated
that administering one intravenous (IV) bolus injection of
DHA or EPA within 3 hours of compression SCI results in
improved locomotion, decreased lipid and protein oxidation
and reduced levels of cyclooxygenase-2 (Lim et al., 2010;
King et al., 2006, Huang et al., 2007b). The effect of DHA
was enhanced when the bolus was supplemented with DHA
in the diet. The effects of an EPA-enriched diet after SCI
have not yet been explored.
Locomotor recovery
Locomotor function was assessed after compression or
contusion SCI using the BBB scoring system to measure the
use of hindlimbs following SCI in rats (Basso et al., 1996).
The use of omega-3 PUFAs as treatment after SCI
should be relatively easy to transfer to the clinic, due to their
safety and minimal toxicity, but such translation would
benefit from having more information on their mechanism
of action and optimum dose regimens. The aim of this
preclinical project was to investigate the effect of DHA and
EPA on the acute inflammatory response (experiment 1),
to determine the efficacy of dietary delivery of EPA in
compression injury (experiment 2), and of IV delivery of
DHA in contusion injury (experiment 3).
Bladder size
Bladder volume was measured using a high resolution
portable digital ultrasound system (Sonosite® MicroMaxx®;
BCF Innovative Imaging, Livingston, Scotland, UK) with a
SLA/13–6 MHz 26 mm linear array transducer as described
by Al-Izki et al. (2009). Post mortem measurements were
taken with a ruler at the widest point.
42
Cytokine measurement
All reagents were purchased from MesoScale Discovery
(MSD, Gaithersburg, USA) unless otherwise stated and the
assays were performed as described in the manufacturer’s
instructions. 25 μl duplicates of fresh spinal cord
homogenates and plasma samples were dispensed into the
bottom of each well of a 96-well MULTI-SPOT MSD plate
containing capture antibodies. Cytokine levels were
quantitated using a cytokine-specific Detection Antibody
labelled with SULFO-TAGTM reagent. A standard curve
was generated and raw readings were converted to pg/mg
total protein.
The number of neutrophils in the injury epicentre
increased three to twenty-fold in the saline group from 4 to
24 hours after SCI (Fig. 1a–b; p<0.05). Following DHA
treatment, there were significantly fewer neutrophils in the
dorsal columns and ventrolateral white matter (VLWM) of
the epicentre than the saline treated group 4 hours after SCI
(Fig. 1a, p<0.05). 24 hours after SCI (Fig. 1b), there were
significantly fewer neutrophils in the ventral horn of the
DHA treated group (p<0.05).
C-reactive protein measurement
The quantitative measurement of CRP in the rat plasma
was performed using a commercial rat CRP ELISA kit (BD
Biosciences, Oxford, UK). 100 μl duplicates of each plasma
sample (1:4000) were allowed to react with antibodies
coated on specially treated microplate wells. Enzyme-labeled
rabbit anti-rat CRP (conjugate) was then added and
washed, followed by addition of a urea peroxide substrate
with tetramethylbenzidine (TMB) as chromogen to initiate
colour development. Stop solution turned the blue positive
reactions to yellow and absorbance was read at 450 nm on
a spectrophotometer. A standard curve was generated and
raw readings were converted to pg/ml.
Figure 1. (A) JT1 immunostaining revealed that there was a
significant increase in neutrophils in the dorsal columns (DC), dorsal
horns (DH) and ventrolateral white matter (VLWM) of the vehicle
group compared to sham (#p<0.05). There were significantly fewer
neutrophils in the VLWM of the DHA treated group than the vehicle
group (*p<0.05). There was a trend towards a reduction in the DC
following treatment with DHA and EPA but this was not significant.
(B) 24 hours after SCI the number of JT1 immunoreactive
neutrophils in the epicentre increased three to twenty-fold in the vehicle
group compared to 4 hours and was significantly greater than sham
(#p<0.05). There were significantly fewer neutrophils in the ventral
horns of the DHA treated group than the vehicle treated group
(*p<0.05). There was a trend towards a reduction in the number of
neutrophils in most areas of the injury epicentre after treatment with
EPA but this did not reach significance. Results are mean ± SEM
number of animals in brackets.
Tissue processing
For histological analysis of neutrophils and
macrophages/microglia, rats were deeply anaesthetised with
pentobarbitone and transcardially perfused with saline
followed by paraformaldehyde (4% in 0.1 M PB). Spinal
cord tissue was dissected, post-fixed in 4%
paraformaldehyde, transferred to 20% sucrose (in 0.1M PB)
and blocked in OCT embedding compound for cryostat
sectioning.15 μm transverse spinal cord sections in the
injury epicentre or 5 mm rostral were incubated with
primary antibodies (ED1 for macrophages, 1:1000, Serotec,
UK; JT1 for neutrophils, 1:1000, gift), washed and followed
by the addition of Alexa Fluor 488 or 594 secondary
antibodies (1:1000, Invitrogen).
DHA, but not EPA reduced systemic CRP levels
Plasma levels of CRP were significantly increased (Fig. 1c,
p<0.05) after SCI compared to laminectomy (sham)
surgery. CRP levels returned to sham levels following
treatment with DHA, but not EPA. There was a significant
increase (p<0.001) in the levels of the cytokine IL-6 in the
injury epicentre 4 hours after SCI (Fig. 2a–b). Levels
returned to baseline levels at 24 hours. Treatment with
DHA or EPA did not reverse this increase (Fig. 2a–b). A
similar effect was found in the levels of the cytokines and
chemokines KC/GRO/CINC (the rat IL-8 counterpart),
TNF-α and IL-1β (data not shown).
Quantification of histological markers
The quantitative analysis of neutrophils in all groups and
time points in spinal cord tissue was conducted by counting
all labelled cells within the field of view in areas of the dorsal
horn (DH) and ventral horn (VH), dorsal columns (DC)
and ventrolateral white matter (VLWM) using a 40×
objective. For ED1, using Q-Win software, an outline was
drawn around an image of the whole section. A binary
image was created representing areas of immunoreactivity
and expressed as a % of the area. Regions from at least
3 sections per animal were quantitated and data expressed as
means ± S.E.M. All treatment groups were kept blind until
after the counts were made.
Experiment 2
EPA dietary treatment had a detrimental effect on recovery
of locomotion following SCI
Unexpectedly, we found that treatment with
supplementation of EPA in the diet, commencing
RESULTS
Experiment 1
DHA reduced neutrophil infiltration to the injury epicentre
43
immediately after SCI led to a worse functional outcome
than controls (Fig. 3a–b). Within days after SCI the BBB
score in the IV saline, EPA diet group was significantly
lower and remained lower until the end of the study (BBB
score of 5.0, day 28; compared to 10.8 in the control group,
p<0.05, 2 way RM-ANOVA). A significantly larger increase
was observed in bladder volume in the IV saline, EPA diet
group (2.18 ± 0.43 ml, day 3) compared to the control
group (0.74 ± 0.41 ml, p<0.05, Bonferroni post-hoc test,
Fig. 3b). On post mortem analysis, a significant permanent
increase in size of the bladder width (Fig. 3c) was found in
the IV saline, EPA diet group at 28 days compared to the
control group (12.8 ± 1.1 mm vs. 8.0 ± 2.0 mm; p<0.05).
There was no significant difference in the amount of
macrophages in the spinal cord tissue (p>0.05, Fig. 4a–c)
and there was no correlation with neuronal, axonal,
oligodendrocyte or microglial markers (data not shown).
Figure 4. (A–B) ED1 labelled macrophages approximately 5 mm
rostral to the injury site. (C) Quantification revealed that there was
no significant difference (p>0.05) between the control diet and IV
saline group and the IV saline, EPA diet group. Results represent mean
± SEM; n=6 per group. Scale bar = 50 μm.
Figure 2. (A) SCI led to a significant increase (***p<0.001) in the
levels of IL-6 in the epicentre 4 hours after SCI compared to naïve
and sham control. Levels returned to baseline levels at 24 hours.
Treatment with DHA, or (B) EPA did not affect the levels of IL-6
after SCI (C) ELISA revealed a significant increase in the CRP
plasma levels in the vehicle (saline) treated group (p<0.05) 4 hours
after SCI. DHA significantly reduced the CRP levels to control levels
(p<0.05), whereas EPA had no significant effect (p>0.05).
Experiment 3
DHA injection restored stepping ability after contusion SCI
Both groups recovered to a BBB score of 11 by 28 days (Fig.
5a) and there was no significant effect of treatment (p>0.05,
2 way RM ANOVA), although there was a trend towards an
improved score in the DHA 500 nmol/kg group.
Importantly, on further analysis, there was a significant
difference (Fig. 5b, p>0.05, Fischer’s exact test) between the
control and DHA 500 nmol/kg groups based on the
frequency of stepping at 28 days post-SCI. In the control
saline-injected group 30 % of the group were stepping
frequently/consistently compared to 60 % in the DHA
250 group and 88 % in the DHA 500 nmol/kg group.
There was no difference (p>0.05) in the amount of
macrophages (Fig. 6) in the spinal cord tissue between the
control and DHA 500 nmol/kg groups and there was no
correlation with neuronal, axonal, oligodendrocyte or
microglial markers (data not shown).
CONCLUSION
We have shown previously that a single bolus of DHA or
EPA (both 250 nmol/kg) confers significant improvement
in neuronal survival and functional outcome following
compression SCI (Huang, et al., 2007; King et al., 2006;
Lim et al., 2010). In this study we now show that DHA
(500 nmol/kg) improves functional outcome after
contusion SCI. We also show that some aspects of the acute
inflammatory response are modified by the administration
of DHA, but not EPA after SCI. However this was a modest
Figure 3. (A) Locomotor recovery in the IV saline, EPA diet group
was significantly worse than the control group (IV saline, control diet;
*p<0.05). (B) Calculation of bladder dimensions from ultrasound
readings revealed that there was a significantly larger increase in
bladder volume in the IV saline, EPA diet group compared to the
control group (*p<0.05) (C) Measurement of the bladder size at the
termination of the experiment revealed a chronic significant increase
in bladder width in the IV saline, EPA diet group compared to the
control group (*p<0.05). Error bars represent SEM, n=6 per group.
44
effect, and unlikely to be the major mechanism of action of
PUFAs in SCI. In addition, and unexpectedly, dietary EPA
after SCI was associated with an adverse outcome.
Interestingly, these results suggest that the neuroprotective
properties of DHA and EPA are most likely due to
mechanisms other than their acute anti-inflammatory
properties. Furthermore, differences were observed between
DHA and EPA, consistent with the report of others (Sierra
et al., 2008).
REFERENCES
Al-Izki S., Pryce G., Giovannoni G., Baker D. (2009)
Evaluating potential therapies for bladder dysfunction in a
mouse model of multiple sclerosis with high-resolution
ultrasonography. Mult. Scler. 15:795–801.
Bao F., Dekaban G., Weaver L. (2005) Anti-CD11d
antibody treatment reduces free radical formation and cell
death in the injured spinal cord of rats. J. Neurochem.
94:1361–1373.
Campbell S., Perry V., Pitossi F., Butchart A., Chertoff M.,
Waters S., Dempster R., Anthony D. (2005) Central
nervous system injury triggers hepatic CC and CXC
chemokine expression that is associated with leukocyte
mobilization and recruitment to both the central nervous
system and the liver. Am. J. Pathol. 166:1487–1497.
Dusart I., Schwab M. (1994) Secondary cell death and the
inflammatory reaction after dorsal hemisection of the rat
spinal cord. Eur. J. Neurosci. 6:712–724.
Fleming J., Bao F., Chen Y., Hamilton E., Relton J., Weaver
L. (2008) Alpha4beta1 integrin blockade after spinal cord
injury decreases damage and improves neurological
function. Exp. Neurol. 214:147–159.
King, V.R., Huang, W.L., Dyall, S.C., Curran, O.E.,
Priestley, J.V., Michael-Titus, A.T. (2006). Omega-3 fatty
acids improve recovery, whereas omega-6 fatty acids worsen
outcome, after spinal cord Injury in the adult rat.
J. Neurosci. 26 (17): 4672–4680.
Huang W.L., King, V.R., Curran, O.E., Dyall, S.C., Ward
R.E., Lal, N., Priestley, J.V., Michael-Titus, A.T. (2007a).
The characteristics of neuronal injury in a static
compression model of spinal cord injury in adult rats. Eur.
J. Neurosci. 25 (2):362–72.
Huang, W.L. et al. (2007b). A combination of intravenous
and dietary docosahexaenoic acid significantly improves
outcome after spinal cord injury. Brain. 130:3004–3019.
Lim S.-N., Huang W., Ward R., Hall J., Priestley J.,
Michael-Titus A. (2010) The acute administration of
eicosapentaenoic acid is neuroprotective after spinal cord
compression injury in rats. In press.
Popovich P., Wei P., Stokes B. (1997) Cellular inflammatory
response after spinal cord injury in Sprague-Dawley and
Lewis rats. J. Comp. Neurol. 377:443–464.
Popovich P., Guan Z., Wei P., Huitinga I., van Rooijen N.,
Stokes B. (1999) Depletion of hematogenous macrophages
promotes partial hindlimb recovery and neuroanatomical
repair after experimental spinal cord injury. Exp. Neurol.
158:351–365.
Profyris C., Cheema S., Zang D., Azari M., Boyle K.,
Petratos S. (2004) Degenerative and regenerative
mechanisms governing spinal cord injury. Neurobiol. Dis.
15:415–436.
Sierra, S., Lara-Villoslada, F., Comalada, M., Olivares, M.,
Xaus, J. (2008) Dietary eicosapentaenoic acid and
docosahexaenoic
acid
equally
incorporate
as
decosahexaenoic acid but differ in inflammatory effects.
Nutrition. 24(3):245–54.
Trivedi A., Olivas A.D, Noble-Haeusslein L.J. (2006)
Inflammation and Spinal Cord Injury: Infiltrating
Leukocytes as Determinants of Injury and Repair Processes.
Clin. Neurosci. Res. 6:283–292.
Figure 5. (A) No significant difference was found between the three
treatment groups in the BBB score of locomotion. (B) However,
significantly more rats in the DHA 500 nmol/kg treated group recovered
frequent or consistent stepping compared to the saline-treated group
(*p<0.05). Error bars represent SEM. n = 8–10 animals per group.
Figure 6. (A–B) ED1 labelled macrophages approximately 5 mm
rostral to the injury site, 28 days post-injury. Quantification (C)
revealed that there was no significant difference (p>0.05) between the
saline and DHA 500 nmol/kg treated groups. Results represent mean
± SEM; n = 7–8 animals per group. Scale bar = 200 μm.
45
Hall, J.C., Perry, V.H. and Priestley, J.V. (2007) Systemic
and local inflammatory changes following a spinal cord
compression injury.
– Zurich: Spinal Research Annual Meeting.
– Spinal Research Christmas supporter reception.
November 2007.
– William Harvey Research Day, Barts and The London.
2007.
Posters:
Hall J.C., Priestley J.V., and Michael-Titus, A. (2010) The
effects of eicosapentaenoic acid delivered as dietary
treatment after spinal cord injury.
– ISRT annual meeting, Zurich, 2010.
2010.
– PU16. San Diego: 2010 SFN annual meeting.
Hall, J.C., Priestley, J.V., Perry, V.H. and Michael-Titus, A.
(2009) Does acute treatment with docosahexaenoic or
eicosapentaenoic acid affect inflammatory markers
following compression spinal cord injury?
– P542.14/S11. Chicago: 2009 Soc. Neurosci. Abstract
Viewer/Itinerary Planner and P172.
– Santa Barbara: 2009. The Second Joint Symposium of the
International and National Neurotrauma Societies.
– ISRT Annual meeting, Glasgow, 2009.
Hall, J.C., Priestley, J.V., Perry, V.H. and Michael-Titus, A.
(2008) The effects of omega-3 fatty acids on early
inflammatory events after spinal cord injury in the rat.
– ISRT Annual meeting, London, 2008.
2008.
–Spring School Cambridge, 2009.
Hall, J.C., Michael-Titus, A., and (2008) The inflammatory
response and locomotor recovery following a spinal
compression injury.
– P79. Dublin: UCD International Neuroimmunology
Symposium.
Publications:
Hall, J.C.E., Perry, V.H., Priestley, J.V. and Michael-Titus
A.T. The acute inflammatory response after compression
spinal cord injury and the effects of omega-3 fatty acids.
In submission
Lim, S.-N., Huang W., Ward R., Hall J., Priestley J.,
Michael-Titus A. (2010) The acute administration of
eicosapentaenoic acid is neuroprotective after spinal cord
compression injury in rats. Prostaglandins, Leukotrienes &
Essential Fatty Acids In press.
FUTURE PLANS
The studentship has been completed. However, there are
several questions arising from our results that are worth
noting.
1. Is there a linear dose-response to omega-3 PUFA
treatment after SCI?
2. What are the mechanisms underlying the beneficial
effects of IV DHA and EPA after SCI?
3. What are the mechanisms underlying the detrimental
effects of the EPA diet after SCI?
46
AAV8shRNA-RhoA and AAV8nt-3 transfection of dorsal
root ganglion neurons (DRGN) in vivo mediates neuron
survival and disinhibited regeneration of dorsal column
(DC) axons
Steven J. Jacques*, Ann Logan, Martin Berry, Zubair Ahmed
University of Birmingham, Birmingham, UK [email protected]
INTRODUCTION
The inability of the central nervous system (CNS) to
regenerate axons following injury is a well-recognised
phenomenon, which may be explained by a number of
potentially maladaptive components of the CNS injury
response (e.g. reviewed in (Sandvig, Berry et al. 2004)). For
example, lack of trophic support, significant neuronal death
and the presence of axon growth inhibitory ligands (AGIL)
all contribute to the lack of axon regeneration seen.
(e.g. (Blits, Oudega et al. 2003)). It was decided to target
NT-3 responsive DRGN since delivery of this neurotrophin
has not been associated with hyperalgesia, presumably due
to sprouting of nociceptive DRGN or sympathetic
terminals within the DRG (Dyck, Peroutka et al. 1997).
Our group has previously produced and evaluated
shRNARhoA and demonstrated effective RNA and protein
knockdown in DRGN (Ahmed, Dent et al. 2005).
Adeno-associated viruses (AAV) have been used as gene
therapy vectors for more than twenty years in a large variety
of tissues including the CNS (Chamberlin, Du et al. 1998).
AAV vectors offer a number of advantages over other
methods of gene delivery including high level, sustained
transgene expression, low pathogenicity and a large variety
of serotypes facilitating modification of cellular tropism
(Goncalves 2005; Vandenberghe, Wilson et al. 2009). The
genomes of all AAV-based vectors are composed of ssDNA,
containing an expression cassette spanned by inverted
terminal repeats (ITRs), usually from AAV2. Capsid
proteins from serotypes other than AAV2 are able to
assemble around AAV2 ITRs, forming a so-called
transencapsidated vector. Such vectors are described by
giving the origin of the ITRs followed by the origin of the
capsid genes (e.g. AAV2 ITRs with AAV8 capsid is
designated AAV2/8). The capsid proteins expressed
determine the interaction of the vector with its host cells,
and hence the majority of its tropism. However, it is known
that other factors downstream of viral entry, can contribute
to differential expression of delivered transgenes (Duan, Yue
et al. 2000).
This project used a dorsal column (DC) injury model to
examine the effects of axonal regeneration-promoting
treatments. In brief, axons of the dorsal column are a subset
of the central projections of neurones lying in the dorsal root
ganglia (DRG). These cells, the dorsal root ganglion neurones
(DRGN) convey proprioceptive and certain cutaneous
modalities. When the central axons of DRGN are transected,
they fail to regenerate their axons and an appreciable number
of them apoptose and undergo atrophic changes (Chelyshev,
Raginov et al. 2005). A subset of DRGN, whose axons project
to the DC express the neurotrophin receptor TrkC, and are
therefore responsive to the trophic effects of neurotrophin-3
(NT-3) (Chen, Zhou et al. 1996).
Nogo, myelin-associated-glycoprotein (MAG) and
oligodendrocytes-myelin-glycoprotein (OMGp) all exert
their effects as AGIL via the non-signalling NgR receptor.
Alone, NgR is unable to transduce a signal, so it is found in
association with ‘leucine-rich repeat and Ig domain
containing’ (LINGO) along with either p75NTR or
‘TNFRSF expressed on the mouse embryo’ (TROY).
Successful activation of this signaling complex by AGIL
results in recruitment of an intracellular signaling cascade,
eventually leading to RhoA mediated collapse of the actin
cytoskeleton, arresting further axogenesis (reviewed in
(Sandvig, Berry et al. 2004).
It is known that vectors with the AAV2 capsid, when
delivered by a variety of routes, target sensory neurones
including those of the auditory and visual systems (Konishi,
Kawamoto et al. 2008). Despite the observation that AAV2
transduces embryonic dorsal root ganglion neurones (DRGN)
in vitro, our group (unpublished data) and others have shown
that this serotype cannot elicit transgene expression in DRGN
in vivo (Fleming, Ginn et al. 2001; Storek, Harder et al. 2006).
In contrast, AAV2/8 targets DRGN when delivered by direct
injection to dorsal root ganglia (DRG) and skeletal muscle,
as well as by intrathecal and intravenous injection (Xu, Gu et
al. 2003; Foust, Poirier et al. 2008; Storek, Reinhardt et al.
2008; Zheng, Qiao et al. 2009).
It is becoming increasingly recognised that
combinatorial strategies offer a powerful way to mobilise
growth of axons and overcome the barriers to their
regeneration (Logan, Ahmed et al. 2006). The main aim of
this project was to examine regeneration of DC axons in the
spinal cord after delivery of recombinant adeno-associated
viruses (AAV8) containing a construct encoding a
neurotrophic factor gene (NT-3) and an RNA-interference
construct designed to knock down RhoA (shRNARhoA).
DRGN can be classified in a number of ways, based
upon functional and morphological criteria which often
overlap. Despite a number of clear demonstrations that
Neurotrophin-3 (NT-3) supports the survival and
axonal growth of a subset of DRGN in explants and in vivo
47
AAV2/8 transduces DRGN, there is no published account
of whether functional subsets of DRGN are differentially
targeted. This is important if therapies using neurotrophic
factors (NTF) are being delivered, since inappropriate
trophic stimulation of, for instance, nociceptors may lead
to side effects such as neuropathic pain (Dyck, Peroutka et
al. 1997). Here, we present a detailed analysis of the cellular
tropism of AAV2/8 in the adult rat DRG, demonstrating
preferential targeting of large diameter DRGN in the
absence of neuronal death. We show that the central
projections of these large DRGN are strongly labelled with
eGFP several centimetres from the site of injection. These
findings suggest that AAV2/8 vectors display an even higher
degree of cellular tropism than was previously thought.
Furthermore, the presence of eGFP in long tracts holds
promise for convenient, accurate tracing of regenerating
axons. However, these findings must be taken in the context
of our discovery that delivery of AAV2/8gfp is associated
with a significant peripheral inflammatory and central glial
response, which may potentially lead to difficulties in
interpreting the results of regenerative studies employing
these vectors.
Transfection of COS-1 cells with Lipofectamine 2000
Having seeded the cells as described above, the medium was
replaced with DMEM alone. 500 μl of a mixture containing
10 μl Lipofectamine 2000 and 4 μg of DNA was added to
each well and left on, at 37°C in 5% CO2 for five hours.
After this, medium was replaced with complete DMEM
containing FBS and antibiotics. Cells were allowed to grow
for 6 days following transfection with addition of 1 ml
complete DMEM after 3 days.
SDS-PAGE and Western blotting
Samples were boiled in 1× loading buffer for four minutes
before being loaded onto a 12% Tris-glycine gel. The gel
was run at 125 V for 1 hour 50 minutes followed by either
Coomassie or silver staining or transfer onto a nitrocellulose
membrane over 2 hours at 25 V. After blocking, the
membranes were stained with anti-FLAG M2 antibody at a
dilution of 1:1000 and NT3 antibody at a dilution of
1:200. Bands were visualized by exposure onto Kodak
Biomax light film, using HRP-conjugated secondary
antibodies followed by application of ECL according to the
manufacturer’s instructions.
Hypothesis
AAV8shRNA-RhoA and AAV8NT-3 transfection of dorsal
root ganglion DRGN in vivo mediates neurone survival and
disinhibited regeneration of dorsal column (DC) axons.
L4/L5 DRG injection
Six adult male Sprague Dawley rats (150–250 g) were operated
upon in accordance with the regulations of the UK Animal
Act 1986. Anaesthesia was induced with 4% isoflurane and
maintained at 2% throughout the procedure. Buprenorphine,
at a dose of 0.03 mg/kg, was used for analgesia, given at the
start of the procedure and twice daily for a further 2d following
surgery as required. Aseptic conditions were used throughout.
The animal was placed on a heat pad in a custom-made
stereotactic apparatus, allowing the whole animal to be moved
through all planes, ensuring that the spine was kept straight at
all times. A 2 cm incision was made in the midline over the
lumbar region and held open with a retractor. The
ligamentous insertions of erector spinae were visualised
allowing the lumbar vertebrae to be identified and a small
mark made on the contralateral side at the level of L4. A 2 cm
paramedian incision was made, with L4 as its midpoint,
around 1 mm to the left of the spinous processes through the
erector spinae muscles, down to the articulating surfaces of the
intervertebral facet joints. Ligamentous attachments to the
articular surfaces were severed, followed by further blunt
dissection to reveal the lateral processes which were removed
to reveal the underlying dorsal roots and DRG. Haemorrhage
was stopped using Spongostan gel foam. A solution of 1010
AAV2/8GFP viral genomes in a volume of 10 μl (kindly
provided by Professor Ron Klein, Louisiana State University)
diluted in sterile PBS was injected into the L4 and L5 DRG
using a glass microelectrode attached to a 20 ml syringe
containing air. Injection was deemed successful if the DRG
was observed to swell. The dorsal incisions were closed using
catgut for the muscle layer and skin staples. Animals were
observed closely post-operatively during recovery, and checked
daily for any signs of autophagia.
Overall aims of the project
• Construct and evaluate AAV8 vectors encoding NT-3
and shRNARhoA.
• Evaluate the effects on growth and survival on DRGN
of these vectors using a dissociated culture system
in vitro.
• Inject these viruses into L4/L5 dorsal root ganglia in
vivo after DC transection and examine their effects on
regeneration of the central axons of DRGN. DC
injuries will be performed 28d after injection of virus,
and tissue harvested 28d after lesion.
Aims addressed this year
• Compare the tropism and toxicity of AAV8 vectors
when injected directly into the DRG or injected
intrathecally into the CSF.
• Examine the effect of direct injection to the DRG of
AAV8 vectors encoding NT-3 and shRNARhoA in vivo
after DC transection.
METHODS
Specific details of each experiment are detailed in the results
section. The following describes general techniques used.
COS-1 cell culture
COS-1 cells were maintained in T75 tissue culture flasks in
DMEM medium, supplemented with 10% FBS and 1%
penicillin/streptomycin. They were passaged every 2–4 days
using 0.05% trypsin. When plating for subsequent
transfection, COS-1 cells were trypsinised and seeded onto
6-well tissue culture plates at 500 000 cells/well in 2 ml
DMEM. They were left for 24 hours before transfection.
Intrathecal injection
Perioperative care was identical to the above. Rats were placed
in the prone position, and a 2 cm incision made between the
48
Adobe Photoshop CS3 (Adobe Systems Incorporated).
DAPI and FITC channels were converted to.TIF format
and examined using ImagePro image analysis software
(Media Cybernetics Inc., Maryland, USA). DRGN were
identified in the DAPI channel using the following criteria:
(i) a large, round cell body (visible as an empty area in the
DAPI channel); (ii) a large, round pale DAPI+ nucleus; (iii)
clearly defined encircling satellite cells with elongated
nuclei. DRGN were identified during 4 passes of each
composite image at 25% and 100% digital zoom levels. The
diameter and position of each DRGN was recorded and
saved as a mask which was then applied to the FITC
channel, allowing the recording of the frequency and
diameters of GFP+ DRGN.
L5 and S1 spinal segments. The L6/S1 interspinous ligament
was incised, allowing the L6 spinous process to be removed
and reflected rostrally, allowing direct visualization of the
ligamenta flava. A blunt 25 G needle was inserted between
the ligamenta flava at an angle of 60° to the horizontal. Access
to the intrathecal space was confirmed by CSF in the needle
cup, and the presence of a tail flick. CSF could also be
expressed by gentle tail traction. The injectate (vehicle or
1012 vg in 30 μl PBS) was pipetted into the needle cup, and
injected gently with air from a 5 ml syringe.
Tissue preparation and histology
Animals were sacrificed at 30d post injection by CO2
narcosis followed by perfusion with 4% formaldehyde
(TAAB laboratories, Aldermaston, Berkshire, UK) in
phosphate buffered saline (PBS). DRG, dorsal and ventral
roots, spinal cords (separated into segments containing
L4/L5 dorsal root entry zone (DREZ), rostral lumbar cord,
thoracic cord and cervical cord) and brainstem were
removed and post-fixed in 4% formaldehyde in PBS
overnight at 4°C. Tissues were cryopreserved by
equilibration in sucrose at 10, 20 and 30% w/v
concentration, embedded in Optimal Cutting Temperature
(OCT) mounting medium, and stored at −80°C. Frozen
tissue sections cut at 15 μm thick were collected from tissue
blocks using a Bright OTF cryostat. For determination of
GFP expression, thawed sections from the centre of each
DRG were washed three times in PBS and mounted using
Vectashield mounting medium with DAPI.
Spinal cord sections were examined by producing
composite images in Photomerge using fields of view taken
through a 10× objective lens. Grey matter was demarcated
using tissue autofluorescence (neuronal cell bodies in grey
matter autofluoresce in the FITC channel). Images were
contrast adjusted to demonstrate GFP+ axons only.
Statistical analysis
For each DRG, frequency histograms were created depicting
total DRGN and GFP+ DRGN with somatic diameters
ranging from 0–100 μm, using 10 μm bins. From these,
histograms were derived of frequencies as a percentage of
total DRGN. The combined frequency distributions for
total DRGN and GFP+ DRGN were presented as means
±S.E.M. and compared using the Mann-Whitney U test.
For immunohistochemical staining for βIII-tubulin,
thawed sections were washed 2×5 min in PBS, 2×5 min in
PBS + 0.1% triton X-100 (PBST) and then incubated in a
humidified chamber for 1 hour at rt in 3% bovine serum
albumin (BSA; Sigma, Poole, UK) in PBST. Sections were
incubated overnight at 4°C with a mouse monoclonal
antibody against βIII-tubulin (Sigma T8660; diluted
1:1000 in 3% BSA in PBST). After this, sections were
washed 3×5 min in PBST and then incubated with an
Alexa-594 conjugated secondary antibody (Molecular
Probes, Oregon, USA A11005; diluted 1:500 in 3% BSA in
PBST) for 1 hour at rt. Sections were finally washed
3×5 min in PBST before being mounted in Vectashield
mounting medium with DAPI (Vector laboratories Ltd.,
Peterborough, UK).
Protocol for restriction digestion of DNA
In a typical restriction digest, 1μg of plasmid DNA was
digested using 10 units of restriction enzyme in the presence
of the appropriate buffer, with or without BSA at 37°C for
2 hours. All enzymes used in cloning are listed in Table 2.2.
The products of each reaction were usually run on a 0.7%
agarose gel.
Protocol for dephosphorylation of restriction digest products
If a plasmid was to become the recipient for a sub-cloning
procedure, following its digestion, it was dephosphorylated
to ensure that its own sticky ends did not ligate together.
The resulting self-ligation results in a high level of vector
background, as well as decreasing the efficiency of any subcloning procedure. Dephosphorylation was achieved by
adding 5 units of antarctic phosphatase per microgram of
DNA, along with an appropriate amount of antarctic buffer
(both supplied by New England Biolabs, Ipswich, MA,
USA) and incubating at 37°C for 30min. The resulting
dephosphorylated fragments were then be run on an agarose
gel, purified and stored for future use.
Image capture and analysis
The brightest region of DRGN autofluorescence was selected
for viewing in DRG sections from uninjected (control)
animals (tissues processed as described above) and the camera
exposure time recorded where no DRGN were visible (mean
exposure time from 2 sections from each of 3 animals,
319 ms). This exposure time was then used throughout the
subsequent experimental analysis of tissues.
Protocol for agarose gel electrophoresis
Agarose was dissolved by heating in TAE-EtBr (2M Trisacetate, 100 mM Na2EDTA (Geneflow), 1μg/ml ethidium
bromide (Promega)) before pouring into a gel casting unit,
and inserting a comb. Samples were diluted in 6× gel
loading buffer, and up to 20 μl of sample loaded into each
well alongside an appropriate DNA ladder. The gel was run
DRG sections were photographed throughout via the
10× objective using a Zeiss Axioplan 2 microscope (Carl
Zeiss Ltd., Hertfordshire, UK). Approximately 35 fields of
view per DRG section were captured and merged to create
a single composite image using the Photomerge feature in
49
at a current of 50 mA for 60–90 minutes, or until the bands
were seen to have migrated a sufficient distance. Progress of
each run was monitored using the Multigenius gel
documentation system (Syngene, Cambridge, UK).
of large diameter DRGN (Figure 3A); a phenomenon
demonstrated by a statistically significant shift to the right
of the size distribution of eGFP+ DRGN compared with
the distribution of all DRGN (Figure 3B; Mann-Whitney
U test, p=6 × 10–36). Despite a lower overall transduction
rate, intrathecal injection yielded similar results, with 0.04%
(S.E. ± 0.04%) of small diameter DRGN transduced
compared with 1.83% (S.E. ±0.52%) of medium diameter
and 16.53% (S.E. ±10.24%) of large diameter DRGN
(Figure 3C, D; Mann-Whitney U, p=1,57 × 10–12).
RESULTS
Viral constructs are capable of synthesis of NT-3 and
knockdown of RhoA
Conditioned medium from COS-1 cells subjected to the
following treatments were collected and analysed by ELISA
for NT-3: untransfected, transfected with pAAV8gfp and
transfected with pAAV8gfp-NT3. Briefly, standard curves
were set up in a 96 well plate using recombinant NT-3 in
serial 1:2 dilutions. Alongside this, samples were added and
serial 1:2 dilutions made of them. An HRP-bound tertiary
antibody provided the colorimetric output which was read
on a plate reader at 450nm. The concentration of NT-3 in
the conditioned medium of each well was determined by
comparison with standards followed by correction for
dilution factors. It was found that no NT-3 was present in
supernatant from untransfected cells and cells transfected
with GFP alone. Hence, COS-1 cells do not produce NT3. Transfection with pAAV8gfp-nt3 resulted in production
of high levels of NT-3, around the 60ng/ml level. This
provided a biologically relevant concentration of NT-3.
A
Direct
% Transduction 12.1 5.4 13.7 9.4 14.0 16.3 10.3 23.1 5.9 8.1 2.5 15.5
Mean DRGN
33.5 28.4 37.8 35.2 53.5 57.4 36.0 37.6 31.8 36.2 34.3 33.7
cell body
diameter (μm)
Mean GFP+
DRGN cell body 46.9 34.0 50.2 49.8 65.3 74.2 47.7 47.7 43.5 56.4 46.9 42.8
diameter (μm)
B
IT
% Transduction 0.6
0
0
0.2 5.1 2.7 1.7 1.8 2.5 0.6 1.1
33
31
31 37.8 35
61
36
33
35
36
Mean GFP+
DRGN cell body 51 n/a n/a 26 54.8 52
diameter (μm)
53
51
54
54
59
Mean DRGN
cell body
diameter (μm)
The amount of RhoA knockdown mediated by a variety
of shRNA constructs was examined by Dr Michael Douglas.
Briefly, COS-7 cells were transfected with a plasmid
containing rat RhoA, resulting in overexpression of RhoA in
these cells. A subsequent transfection of the RhoA-transfected
cells with plasmids encoding each of the shRNA constructs
allowed knockdown to be evaluated by western blotting of
COS cell lysates using a RhoA antibody (Figure 1).
31
Table 1. Transduction rates and mean DRGN diameters (total
DRGN versus GFP+ DRGN) for individual DRG after (A) direct
and (B) IT injection.
Figure 1. Preliminary validation of shRNA constructs. Sequence A,
giving maximal knockdown, was designed by Dr M. Douglas.
Sequences B and C were based on published sequences.
DRGN are targeted by AAV2/8 independent of delivery route
In all ganglia examined, eGFP expression was restricted to
DRGN (Figure 2A, B). No non-neuronal cells were eGFP+.
The mean transduction rate of total DRGN after direct
injection was 11%, with a range of 2.5–23.1% versus 1.5%
(range 0–5.1%) after IT injection (Table 1). After direct
injection, a small number of eGFP+ DRGN were seen on
the uninjected side (data not shown).
Figure 2. DRGN are targeted by AAV8. A. Only DRGN (asterisks)
and axons (arrowheads) were GFP+ in sections of DRG. (scale bar
50 μm). B. DRG, longitudinal section showing GFP+ DRGN and
axons (scale bar 500 μm). SN spinal nerve, DR dorsal root.
AAV2/8 preferentially transduces large-diameter,
parvalbumin positive DRGN
After direct injection, only 2% (S.E. 0.66%) of small
diameter DRGN were eGFP+, compared with 15% (S.E.
1.98) of medium diameter DRGN and 32% (S.E. 10.55%)
Of medium and large diameter DRGN, 17.4% were
PV+, compared with 33.3% of eGFP+ DRGN (Figure 4,
one-tailed t-test, p=0.046).
50
the dorsal horn (asterisks; Figure 6A). Some eGFP+ axons
projected to the ventral horn (daggers). There were no
eGFP+ axons in the contralateral grey matter, although very
occasional eGFP+ axons were seen in the contralateral
DREZ.
eGFP+ axons projected in the ipsilateral gracile
fasciculus (arrow heads) of the spinal cord at lumbar and
thoracic levels, with decreasing levels of eGFP seen in higher
segments (Figure 6B, C). Very faint labelling was seen in the
gracile fasciculus of the cervical cord, and no axons were
seen in the medulla at the level of the ipsilateral gracile
nucleus (data not shown). No eGFP+ neuronal somata were
seen in the spinal cord or brainstem grey matter.
Figure 3. AAV8 targets large-diameter DRGN. A, C. Proportion of
GFP+ DRGN in each of three arbitrary size classes (small 0–29 μm,
medium 30–59 μm and large >60 μm) after IT and direct injection,
respectively. B, D. Histogram of total and GFP+ DRGN somata sizes
after IT and direct injection, respectively.
Figure 6. The central projections of DRGN in the left gracile
fasciculus were clearly labelled by GFP. A. L4 dorsal root entry zone
(NB artefact from tissue processing removed); B. Rostral lumbar cord;
C. Mid-thoracic cord (not to scale, composed of merged fields at 100×
magnification). D. Longitudinal section of the rostral lumbar cord
(dashed line represents dorsal median sulcus). a–c refer to insets
demonstrated by the red boxes in A, B and C, respectively. DREZ solid
arrow; gracile fasciculus arrow heads; GFP+ axons open arrows; dorsal
horn asterisks; ventral horn daggers. Scale bar 50 μm.
Figure 4. AAV8 targets PV+ DRGN preferentially. A. GFP and PV
co-localisation in the DRG (scale bar 50 μm). B. Proportion of total
DRGN and GFP+ DRGN which are PV+ after IT injection.
Intrathecal injection resulted in widespread labelling of
central projections of large diameter DRGN at all levels of
the cord (Figure 7). eGFP+ axons were observed in dorsal
roots and the DC at all cord levels. eGFP+ axons were seen
terminating in the deeper parts of the dorsal horn,
particularly at lumbar and thoracic levels, with Clarke’s
column clearly defined in the latter. The intensity of labelling
in the gracile fasciculus was seen to diminish in successively
rostral segments, with very little in the cervical cord.
Differential expression of 67kDa Laminin receptor does not
explain tropism of AAV2/8 for large diameter DRGN
All DRGN expressed the 67kDa laminin receptor (LMR),
but in varying levels. There was no qualitative relationship
between neuronal diameter and LMR expression, with both
large and small diameter populations expressing both high
and low levels of LMR (Figure 5).
Figure 7. The central projections of DRGN were clearly labelled with
GFP after IT injection. A. Rostral lumbar cord. B. Mid-thoracic cord.
C. Mid-cervical cord. Orientation as in figure 6. Clarke’s column solid
arrows; cuneate fasciculus open arrows. Scale bar 500 μm.
Figure 5. LMR expression does not correlate with DRGN
transduction (scale bar 50 μm).
The central projections of DRGN are clearly labelled.
After direct DRG injection, eGFP+ axons (open arrows)
were visible entering the spinal cord at the DREZ (solid
arrow), and joining the dorsal columns (DC) or entering
Macrophages infiltrated the DRG after delivery by either
route
The intact DRG contained sparse CD68+ macrophages, in
contrast to DRG from animals receiving IT PBS, where a few
51
macrophages could be seen (data not shown). Direct and IT
delivery of AAV2/8gfp resulted in high levels of CD68
immunoreactivity in the DRG, frequently seen in cells
encircling individual DRGN (Figure 8). Occasional
macrophages were seen in the dorsal root, their presence
ending abruptly at the PNS/CNS boundary (data not shown).
Figure 9. Microglia are activated in the deep dorsal horn after direct
injection of AAV2/8gfp. A. High magnification view of the left deep
dorsal horn in the L1 cord segment demonstrating GFP+ axons and
activated Cd11b+ microglia. Scale bar 50 μm. B. Quantification of
the level of CD11b immunoreactivity in different regions of the L1 cord
segment. LDC and RDC: left and right dorsal column, respectively.
LSDH and RSDH: left and right superficial dorsal horn, respectively.
LDDH and RDDH: left and right deep dorsal horn, respectively.
Figure 8. Delivery of AAV2/8gfp led to macrophage infiltration in the
DRG. CD68 immunohistochemistry on L4 DRG in an intact animal
and after direct and IT injection of AAV2/8gfp. Scale bar 50 μm.
Direct DRG injection led to moderate microglial and astrocyte
activation
The lumbar cord at level L1/L2 was examined for microglial
activation in each delivery paradigm. After direct injection
of AAV2/8gfp, CD11b immunoreactivity in the deep dorsal
horn was upregulated compared with PBS injected animals
(Figure 9; t-test, p=0.004). Direct injection of PBS itself did
not appreciably increase the level of CD11b
immunoreactivity (data not shown). Interestingly, there
appeared to be a trend towards a reciprocal decrease in
Cd11b staining in white matter. IT injection of PBS or
AAV2/8gfp led to minimal amounts of microglial
activation, qualitatively no different from the intact state
(data not shown).
Compared with PBS injected controls, there was clearly
a trend towards a quantitative increase in astrocyte
activation in the rostral lumbar cord in all regions examined
(Figure 10). However, only one region (the right dorsal
column) demonstrated a statistically significant increase in
GFAP immunoreactivity (t-test, p=0.014). PBS injection
itself did not qualitatively increase the level of astrocyte
activation.
Figure 10. There is a trend towards activation of astrocytes in all cord
regions. A. GFAP immunohistochemistry in the region of Clarke’s
column at the L1 cord level (scale bar 50 μm). B. Quantification of
astrocyte activation based on pixel counts above threshold.
There was no evidence of DRGN central axonal
degeneration after delivery of AAV2/8gfp.
CONCLUSION
AAV2/8gfp appeared to be non-toxic in the DRG
Oil Red O staining did not reveal any evidence of
myelin breakdown products within the DC after direct
injection of AAV2/8gfp. A positive control using optic nerve
(14d after a crush lesion) revealed clear staining in the distal
stump and lesion site (Figure 11).
Published data using neuronal profile counting methods
show size distributions of DRGN in adult rats (Natalie,
William et al. 2002; Gaudet, Williams et al. 2004; Lu,
Zhang et al. 2006). These distributions correlate closely with
the results from this study, having a positively skewed
distribution with a modal diameter of around 30 μm.
52
DRG or delivered intrathecally. We have also examined
some further characteristics of the transduced population of
DRGN. Our data point to the preferential transduction of
sensory neurones involved in proprioception by AAV2/8.
The characteristics of these neurones were as follows: large
diameter; 67 LMR positive; PV positive; central projections
in the dorsal root, and DC; projections predominantly to
deep dorsal horn (including Clarke’s column) and ventral
horn; no projections to superficial dorsal horn or
contralateral side. Furthermore, in support of preferential
transduction of proprioceptive neurones destined to
terminate on neurones of the dorsal spinocerebellar tract,
we saw a diminution of DC labelling in successively rostral
cord segments after direct DRG injection. This is consistent
with eGFP+ axons leaving the DC and terminating in the
deep dorsal horn, predominantly Clarke’s column in the
thoracic and rostral lumbar cord.
Indeed, similar distributions are seen in studies examining
cell profile areas after a simple square root transformation of
the data (Jamieson, Liu et al. 2005). This distribution is also
extremely reproducible, manifest in the majority of sections
from individual DRG that were examined. Although not
definitive, the correspondence between our data and
published accounts suggests that there was no death of
DRGN after either delivery method. No evidence of gross
damage was observed in the DRG after either delivery
method, an observation in agreement with a recent
publication comparing AAV gene delivery across serotypes
(Mason, Ehlert et al. 2010). No evidence of Wallerian
degeneration was observed in the DC, consistent with the
other evidence for lack of toxicity. However, it has been
shown that neuronal death occurs when eGFP is delivered
to the CNS under the control of the CMV promoter (Klein,
Dayton et al. 2006). In the same study, toxicity attributable
to AAV8 was excluded by showing no detrimental effect of
an empty AAV8 vector. Thus, it appears that eGFP may be
less toxic to neurones of the PNS compared with those
found in the CNS.
The fact that large diameter DRGN are predominantly
transduced is important for a number of reasons. AAV2/8
based vectors are useful to research groups who use the DC
injury model, since many of the axons comprising the DC
project from large diameter DRGN. Thus, delivery of
growth factors to the exact DRGN population of interest
will minimise potential effects on other neuronal
populations which may confound results. It is particularly
significant that small diameter DRGN are not targeted,
since previous work has demonstrated that nerve growth
factor (NGF) stimulation of these neurones results in
lowering of pain thresholds (Dyck, Peroutka et al. 1997).
Further, delivery of NGF to the DRG via the cerebrospinal
fluid (CSF) has been shown to result in sympathetic
sprouting within the ganglion, which could also contribute
to neuropathic pain (Nauta, Wehman et al. 1999). Finally,
we feel that this result is important with respect to
experimental design. Given the fact that almost half of the
transduced DRGN are PV+, terminating in Clarke’s
column, it would appear that a mid-thoracic DC lesion
would fail to transect these neurones after L4/L5 DRG
injection. As a consequence of these findings, we now lesion
at L1 cord level, such that we are able to transect as many
transduced DRGN as possible, and still detect regeneration
by injection of cholera toxin B into the sciatic nerve. We
feel this serves to illustrate the importance of assessing the
cellular tropism of viral vectors for gene therapy before using
them in injury models.
The mechanism by which AAV2/8 targets large
diameter DRGN is not known. Much more is known about
the mechanisms whereby the archetypal AAV serotype
(AAV2) enters cells, including the exploitation of heparan
sulphate proteoglycans, integrins and fibroblast growth
factor receptors (Goncalves 2005). However, it is known
that AAV2/8 binds LMR, enabling it to transduce mouse
hepatocytes in vivo (Akache, Grimm et al. 2006). The
current study failed to observe any relationship between the
expression pattern of LMR and the transduction efficiency
of AAV2/8. However, given this result it must be stressed
that LMR is still probably involved in AAV2/8 transduction
of DRGN. It is probable that the differential transduction
seen is due to the interaction of a number of cell surface
Figure 11. Oil-Red-O staining reveals no evidence of Wallerian
degeneration within the DC. A. Overview of horizontal section of
cord at ∼0.75 mm depth from dorsal surface. B. High power view of
left DC shown in A. C. Distal stump of optic nerve, crushed 14d
previously, demonstrating strong red staining. D, E. High power views
of distal stump (D) and lesion site (E) demonstrating the expected
appearance of Wallerian degeneration in the CNS including lipid
droplets at the lesion site (asterisk). Scale bar 50 μm.
AAV2/8 targets large diameter, predominantly proprioceptive,
DRGN
We have presented the first detailed comparison of the
cellular tropism of AAV2/8 when injected directly into the
53
receptors (and possibly intracellular molecules), whose role
with regard to AAV vectors is presently unknown.
there was a trend towards bilateral microglial activation in
the deep dorsal horn; there was a trend towards a reciprocal
decrease in the amount of CD11b immunoreactivity in the
dorsal column; there was a trend towards astrocyte
activation throughout the entire cord segment.
AAV2/8gfp can transduce DRGN via CSF
The results presented from our study indicate that AAV2/8
has a clear tropism for large diameter DRGN. There have
been a number of publications describing the results of
delivery of AAV2/8 to other regions of the nervous system.
From these, the following can be concluded: (i) direct
injection of AAV2/8 to various regions of the brain results
in neuronal transduction (Klein, Dayton et al. 2006); (ii)
delivery of AAV2/8 to the cerebral ventricles transduces
neurones in multiple brain regions including cortex and
striatum (Broekman, Comer et al. 2006); (iii) intrathecal
injection of AAV2/8 transduces DRGN alone (Storek,
Reinhardt et al. 2008); (iv) intramuscular injection of
AAV2/8 results in transduction of DRGN (Zheng, Qiao et
al. 2009) and (v) intravenous or intraperitoneal delivery of
AAV2/8 results in transduction of DRGN alone (Foust,
Poirier et al. 2008). The synthesis of these results with our
findings provides a model of AAV2/8 delivery to the
nervous system consistent with AAV2/8 being: (i) unable to
cross the blood brain barrier (BBB); (ii) able to cross the
ependymal cell layer; (iii) unable to cross the pial barrier
and (iv) able to cross the arachnoid membrane.
The literature documenting inflammatory responses in
the nervous system induced by AAV8 vectors or eGFP is
scant and sometimes contradictory. One of the most
illuminating of these reports was from Klein et al. who
showed, by comparing an AAV8 containing eGFP with an
empty vector, that it is likely to be eGFP which is the toxic
agent and not the viral vector (Klein, Dayton et al. 2006).
Furthermore, eGFP has been shown to form aggregates in
HEK-293 cells and also in neurones under certain
conditions (Krestel, Mihaljevic et al. 2004; Link, Fonte et al.
2006). We acknowledge that our study did not employ
an empty AAV8 particle as a control, but feel that the
evidence is strong implicating eGFP as the proinflammatory agent. Also, at least in the majority of
pre-clinical studies, eGFP is a commonly used reporter
system. Therefore, we argue that it is informative to consider
the vector particle (AAV8 capsid) and its reporter gene
product (eGFP protein) together.
Microglia demonstrate multiple phenotypes, correlating
with their state of activation, ranging from highly ramified
(‘inactive’) to cells resembling activated macrophages
(‘phagocytic’). Here, we will use a third term – ‘moderately
activated’ – to describe microglia that have upregulated cellsurface markers but have not assumed a macrophage-like
phenotype. Since there is no universally agreed consensus
on how best to objectively define their level of activation, it
was decided to use a quantitative method (Ransohoff and
Perry 2009). By quantitative assessment of the level of
CD11b immunoreactivity, it was seen that microglia were
activated in the deep dorsal horn. Upon closer inspection
these cells were found to be in the ‘moderately activated’
state, where they up-regulate CD11b expression, but not in
the ‘phagocytic’ state associated with maximal activation.
These observations have been made at a single time point,
so the kinetics of microglial activation in this setting cannot
be commented upon.
The route by which AAV2/8 gains access to the DRG
from the CSF remains unclear, although it seems most likely
that it diffuses directly across the arachnoid membrane into
the DRG itself. It has been established already that
horseradish peroxidise (HRP), when injected epidurally, is
able to enter the DRG in a relatively short period,
apparently by direct diffusion (Byrod, Rydevik et al. 2000).
Given the fact that intrathecally injected AAV2/8 would not
have to cross the dural membrane, this seems a reasonable
explanation. There are also some unique features of the fine
anatomy of the DRG and dorsal root which may be of
relevance. For example, the fact that the relatively
impermeable sheath of the dorsal (and ventral) roots ends as
an open ‘shirt sleeve’ at the dorsal root entry zone, making
the endoneurium of the dorsal root in continuity with the
CSF (Haller, Haller et al. 1972). Also, the little-documented
‘lateral recess’ of the DRG, an evagination of the arachnoid,
usually filled with macrophages, may provide a portal for
entry of viral vectors (Himango and Low 1971). Clearly,
much more detail is required concerning the routes gene
therapy vectors take from their sites of delivery to the
neurones which they transduce. Ultimately, such knowledge
may help to optimise delivery protocols and assist in the
rational design of vectors.
It was interesting to note that there appeared to be a
strong, but not statistically significant, activation of
microglia in the contralateral deep dorsal horn, but not in
the contralateral superficial dorsal horn. We hypothesise that
this is due to the release of soluble inflammatory mediators
from the terminals of transduced DRGN. The strong
microglial activation in the deep dorsal horn compared to
more superficial layers may be explained by the following:
(i) the deep laminae of the dorsal horn are in close proximity
to one another, and are not separated by white matter. The
superficial laminae are separated by approximately 2 mm of
white matter in the dorsal columns. White matter, being
rich in lipid is likely to impede the diffusion of soluble
inflammatory mediators; (ii) most of the transduced DRGN
project to the deep laminae of the dorsal horn, so any effect
on the superficial layers is likely to be mediated by
infrequent collateral sprouts.
AAV2/8gfp elicits CNS glial activation and a PNS
inflammatory response
Despite the lack of any overt toxicity to the transduced
population of neurones, there was clear evidence of
macrophage infiltration within the DRG and a moderate
level of microglial and astrocyte activation within the cord
after direct injection of AAV2/8gfp. The following
observations warrant further discussion: CD11b
immunoreactivity was seen predominantly in grey matter;
54
In conclusion, we have constructed AAV8 based vectors
containing constructs that are capable of producing NT-3
and knocking down RhoA. We have performed an
experiment to examine the effects of these vectors on
DRGN axon regeneration in a DC crush model of SCI. The
experiment is currently being analysed and the results will be
reported separately on completion.
allodynia and lowered heat-pain threshold in humans.”
Neurology 48(2): 501–5.
Fleming, J., Ginn, S.L. et al. (2001). “Adeno-associated
virus and lentivirus vectors mediate efficient and sustained
transduction of cultured mouse and human dorsal root
ganglia sensory neurons.” Human Gene Therapy 12(1):
77–86.
Foust, K.D., Poirier, A. et al. (2008). “Neonatal intraperitoneal
or intravenous injections of recombinant adeno-associated
virus type 8 transduce dorsal root ganglia and lower motor
neurons.” Human Gene Therapy 19(1): 61–69.
Gaudet, A.D., Williams, S.J. et al. (2004). “Regulation of
TRPV2 by axotomy in sympathetic, but not sensory
neurons. Brain Research 1017(1–2): 155–162.
Goncalves, M.A.F.V. (2005). “Adeno-associated virus: from
defective virus to effective vector.” Virology Journal 2: 43.
Haller, F. R., Haller, C. et al. (1972). “The fine structure of
cellular layers and connective tissue space at spinal nerve
root attachments in the rat.” Am. J. Anat. 133(1): 109–23.
Himango, W.A. and Low, F.N. (1971). “The fine structure
of a lateral recess of the subarachnoid space in the rat.” Anat.
Rec. 171(1): 1–19.
Jamieson, S.M.F., Liu, J. et al. (2005). “Oxaliplatin causes
selective atrophy of a subpopulation of dorsal root ganglion
neurons without inducing cell loss.” Cancer Chemotherapy
and Pharmacology 56(4): 391–399.
Klein, R.L., Dayton, R.D. et al. (2006). “Efficient neuronal
gene transfer with AAV8 leads to neurotoxic levels of tau or
green fluorescent proteins.” Mol. Ther. 13(3): 517–27.
Konishi, M., Kawamoto, K. et al. (2008). “Gene transfer
into guinea pig cochlea using adeno-associated virus
vectors.” J. Gene Med. 10(6): 610–8.
Krestel, H.E., Mihaljevic, A.L. et al. (2004). “Neuronal coexpression of EGFP and beta-galactosidase in mice causes
neuropathology and premature death.” Neurobiol. Dis.
17(2): 310–8.
Link, C.D., Fonte, V. et al. (2006). “Conversion of Green
Fluorescent Protein into a Toxic, Aggregation-prone Protein
by C-terminal Addition of a Short Peptide.” Journal of
Biological Chemistry 281(3): 1808–1816.
Logan, A., Ahmed, Z. et al. (2006). “Neurotrophic factor
synergy is required for neuronal survival and disinhibited
axon regeneration after CNS injury.” Brain 129(Pt 2):
490–502.
Lu, S.-G., Zhang, X. et al. (2006). “Intracellular calcium
regulation among subpopulations of rat dorsal root ganglion
neurons.” The Journal of Physiology 577(1): 169–190.
Mason, M.R., Ehlert, E.M. et al. (2010). “Comparison of
AAV serotypes for gene delivery to dorsal root ganglion
neurons.” Mol. Ther. 18(4): 715–24.
Natalie, J.G., William, B.J.C. et al. (2002). “Expression of
gp130 and leukaemia inhibitory factor receptor subunits in
adult rat sensory neurones: regulation by nerve injury.”
Journal of Neurochemistry 83(1): 100–109.
Nauta, H.J., Wehman, J.C. et al. (1999). “Intraventricular
infusion of nerve growth factor as the cause of sympathetic
fiber sprouting in sensory ganglia.” J. Neurosurg. 91(3):
447–53.
Ransohoff, R.M. and Perry, V.H. (2009). “Microglial
Physiology: Unique Stimuli, Specialized Responses.” Annual
Review of Immunology 27(1): 119–145.
In addition, we have demonstrated that AAV8 vectors
preferentially target large diameter, parvalbumin positive
DRGN. This result is extremely important, given that most
of these neurones project in our pathway of choice – the
dorsal column. However, we did detect some inflammatory
effects of these vectors in both the DRG and spinal cord.
These effects are likely to be due to the presence of eGFP, a
reporter gene that remains common in preclinical work. We
conclude from this that care should be taken in the selection
of deliver vectors and reporter genes, since inflammatory
effects could potentially confound experiments examining
regeneration.
REFERENCES
Ahmed, Z., Dent, R.G. et al. (2005). “Disinhibition of
neurotrophin-induced dorsal root ganglion cell neurite
outgrowth on CNS myelin by siRNA-mediated knockdown
of NgR, p75NTR and Rho-A.” Molecular & Cellular
Neurosciences 28(3): 509–23.
Akache, B., Grimm, D. et al. (2006). “The 37/67kilodalton laminin receptor is a receptor for
adeno-associated virus serotypes 8, 2, 3, and 9.” J. Virol.
80(19): 9831–6.
Blits, B., Oudega, M. et al. (2003). “Adeno-associated viral
vector-mediated neurotrophin gene transfer in the injured
adult rat spinal cord improves hind-limb function.”
Neuroscience 118(1): 271–281.
Broekman, M.L.D., Comer, L.A. et al. (2006). “Adenoassociated virus vectors serotyped with AAV8 capsid are
more efficient than AAV-1 or -2 serotypes for widespread
gene delivery to the neonatal mouse brain.” Neuroscience
138(2): 501–10.
Byrod, G., Rydevik, B. et al. (2000). “Transport of
epidurally applied horseradish peroxidase to the endoneurial
space of dorsal root ganglia: a light and electron microscopic
study.” J. Peripher. Nerv. Syst. 5(4): 218–26.
Chamberlin, N.L., Du, B. et al. (1998). “Recombinant
adeno-associated virus vector: use for transgene expression
and anterograde tract tracing in the CNS.” Brain Research
793(1–2): 169–175.
Chelyshev, Y.A., Raginov, I.S. et al. (2005). “Survival and
phenotypic characteristics of axotomized neurons in spinal
ganglia.” Neuroscience & Behavioral Physiology 35(5):
457–60.
Chen, C., Zhou, X.F. et al. (1996). “Neurotrophin-3 and
trkC-immunoreactive neurons in rat dorsal root ganglia
correlate by distribution and morphology.” Neurochemical
Research 21(7): 809–14.
Duan, D., Yue, Y. et al. (2000). “Endosomal processing
limits gene transfer to polarized airway epithelia by adenoassociated virus.” J. Clin. Invest. 105(11): 1573–87.
Dyck, P.J., Peroutka, S. et al. (1997). “Intradermal
recombinant human nerve growth factor induces pressure
55
PUBLICATIONS
Poster presentation, ISRT Network Meeting 2008, 2009,
2010.
Jacques, S.J., Douglas, M.R., Ahmed, Z., Berry, M., Logan,
A. Delivery of gfp by AAV2/8 demonstrates targeting of
large diameter dorsal root ganglion neurones and labeling of
their central projections. In preparation for submission to
Gene Therapy.
Ahmed, Z., Jacques, S.J., Berry, M., Logan, A. (2009)
Epidermal growth factor receptor inhibitors promote CNS
axon growth through off-target effects on glia. Neurobiol.
Dis. 36:142–150.
Douglas, M.R., Morrison, K.C., Jacques, S.J., Leadbeater,
W.E., Gonzalez, A.M., Berry, M., Logan, A., Ahmed, Z.
(2009) Off-target effects of epidermal growth factor
receptor antagonists mediate retinal ganglion cell
disinhibited axon growth. Brain. 132:3102–3121.
Sandvig, A., Berry, M. et al. (2004). “Myelin-, reactive glia, and scar-derived CNS axon growth inhibitors: expression,
receptor signaling, and correlation with axon regeneration.”
GLIA 46(3): 225–51.
Storek, B., Harder, N. et al. (2006). “Intrathecal long-term
gene expression by self-complementary adeno-associated
virus type 1 suitable for chronic pain studies in rats.”
Molecular Pain 2(1): 4.
Storek, B., Reinhardt, M. et al. (2008). “Sensory neuron
targeting by self-complementary AAV8 via lumbar puncture
for chronic pain.” Proc. Natl. Acad. Sci. USA 105(3):
1055–60.
Vandenberghe, L.H., Wilson, J.M.et al. (2009). “Tailoring
the AAV vector capsid for gene therapy.” Gene Ther. 16(3):
311–9.
Xu, Y., Gu, Y. et al. (2003). “Adeno-associated viral transfer
of opioid receptor gene to primary sensory neurons: A
strategy to increase opioid antinociception.” Proceedings of
the National Academy of Sciences of the United States of
America 100(10): 6204–6209.
Zheng, H., Qiao, C. et al. (2009). “Efficient Retrograde
Transport of AAV8 to Spinal Cord and Dorsal Root
Ganglion after Vector Delivery in Muscle.” Hum. Gene Ther.
FUTURE PLANS
Final analysis of final in vivo experiment, giving a definitive
answer on the regenerative effect of NT-3 and shRNARhoA
containing vectors in the injured DC.
Completion of my PhD thesis by Easter 2011.
Awaiting responses from grant applications for postdoctoral
positions.
56
Promoting spinal cord repair by genetic modification of
Schwann cells to over-express PSA
Juan Luo *, Dr Yi Zhang and Dr Xuenong Bo
Queen Mary University of London, UK
INTRODUCTION
Cell therapy has been explored extensively as an important
strategy for the repair of spinal cord injury (SCI). Schwann
cells (SCs) are regarded as one of the most promising cell
type for transplantation in SCI. However, the poor survival
and limited integration of transplanted Schwann cells
within the CNS environment are two major issues that still
need to be addressed before they can be used clinically. We
proposed to genetically engineer the Schwann cells to
express polysialic acid (PSA) by transducing the cells with
lentiviral vector (LV) expressing polysialyltransferase (PST).
In the first two years, we have shown that
polysialyltransferase transduced SCs (PST/SCs) survived
better than the GFP/SCs and the PSA expressing SCs
integrated well within the normal spinal cord. We also
found that PSA expression on SCs did not promote their
migration in normal spinal cord, but significantly enhanced
their migration into the lesion sites in a spinal cord injury
model. Over-expression of PSA in host tissues around the
lesion site of spinal cord can also increase the penetration
of transplanted SCs into the lesion site. Expression of PSA
on transplanted SCs in combination with expression of PSA
in host spinal cord can further enhance the migration of
transplanted cells. In the third year, we mainly concentrated
on the study of the interaction of SCs with astrocytes in
vitro using confrontation assay and co-culture of the two
types of cells to mimic their interactions in vivo. As PSA is
able to promote the survival of transplanted SCs in early
stages, we also studied the underlying mechanisms and
attempted to identify other factors that contribute to the
death of transplanted SCs.
of PST/SCs with astrocytes in comparison with GFP/SCs.
A strip of SCs was set up opposing a parallel strip of
astrocytes. Cultures were then maintained until they came
into contact with each other. For quantification, a 300 μm
line was drawn along the interface between astrocytes and
SCs. The numbers of SCs that crossed the cell boundary
were counted and averaged over five randomly chosen fields
on each coverslip (8 coverslips per group per experiment)
and experiments were repeated three times. To assess the
hypertrophy of astrocytes in contact with PST/SCs or
GFP/SCs, the sizes of astrocytes defined by GFAP
immunoreactivity was measured using ImageJ.
To further study the interaction of PST/SCs with
astrocytes in comparison with GFP/SCs, astrocytes and SCs
were mixed at a ratio of 3:1 and cultured for 10–14 days. To
assess the proliferation of astrocytes, BrdU (20 μM) was
added 16 hours before staining.
3. Assessment of inflammatory cells at Schwann cell
transplantation site in spinal cord
Adult female Wistar rats were deeply anaesthetized and a
laminectomy was performed at T8-T9 level to expose the
spinal cord. Equal numbers of PST/SCs or GFP/SCs were
transplanted into the T8 dorsal column of rat spinal cords.
Animals were killed at 1 and 7 days after transplantation
and spinal cord containing the transplanted Schwann cells
was sectioned. Recruitment of macrophage and infiltration
of neutrophils to the transplantation site were assessed with
immunohistochemistry.
4. Immunocytochemistry and immunohistochemistry
Routine immunocytochemistry and immunohistochemistry
were performed for staining the cells and spinal sections.
The following primary antibodies were used: polyclonal
anti-GFAP or monoclonal anti-GFAP, polyclonal antip75NTR, monoclonal anti-PSA (mab735), and polyclonal
anti-P2X7 receptor. For nucleus staining, 4′,6′diaminidino2-phenylindole (DAPI) was applied. For identification of
inflammatory cells at the transplantation sites, spinal
sections were immunostained with anti-CD68 antibody for
macrophage/microglia and a specific antibody for
neutrophils (a gift from Prof. Hugh Perry) respectively.
METHODS
1. Schwann cell and astrocyte culture and lentiviral
transduction.
SCs were isolated from sciatic nerves and brachial plexus of
neonatal Wistar rats. Cells were maintained in medium
containing a cocktail of growth factors. Cultured SCs were
transduced with either LV/GFP (referred as GFP/SCs) or
LV/PST-GFP. In some experiments SCs were co-transduced
with LV/PST-GFP and LV/GFP for easy identification of
transduced cells due to the low visibility of PST-EGFP
fusion protein in primary cells. The efficiency of cotransduction was 95 ± 2%. Astrocytes were obtained from
the cortex of neonatal (P2–3) rat brains using standard
protocol as described previously (Noble and Murray, 1984).
5. Apoptosis assay
SCs were dissociated from the culture dishes and treated
with various concentrations of ATP (0, 3, 4 and 5 mM) or
glutamate (0, 0.1, and 1 mM) in the CO2 incubator for
1 hour. Apoptosis was detected with Annexin V Apoptosis
Assay kit using flow cytometry. For blockade of ATP
induced cell death, cells were pretreated with 0.35 mM
oxidized ATP (oxATP, a P2X7 receptor antagonist) for
2. Confrontation and co-culture assays
As astrocytes are a major factor that hinders Schwann cell
migration and integration in the CNS, in vitro
confrontation assay was set up to investigate the interaction
57
2 hours before exposure to various concentration of ATP
for 1 hour.
2. PSA has no significant effect on the recruitment of
inflammatory cells at Schwann cell transplantation
sites in normal spinal cord in early stages
As shown previously, the majority of cell death after
transplantation occurs in the first week. In this study, the
presence of macrophage/microglia around cell transplants
was obvious at 1 day after transplantation and CD68+ cell
numbers decreased significantly at 7 days. In contrast to
CD68+ cells, very few neutrophils infiltrated to the
transplantation site at 1 day after transplantation, but more
neutrophils infiltrated to the site at 7 day (data not shown).
However, there is no difference between PST/SCs and
GFP/SCs group in either CD68+ cell or neutrophil
cell numbers.
RESULTS
1. PST/SCs do not induce boundary formation in
SCs/astrocytes confrontation assays and PST/SCs
cause less stress response in astrocytes in culture
In the confrontation assay, when GFP/SCs and astrocytes
came into contact with each other, a distinct boundary was
formed between these two types of cells (Fig. 1B). However,
PSA-expressing SCs did not form a clear boundary with
astrocytes and were able to migrate within astrocytic
territory (Fig. 1A, C). PST/SCs caused less stress response in
astrocytes than GFP/SCs as indicated by the sizes of
hypertrophic astrocytes that came into contact with SCs
(Fig. 1D).
3. Glutamate and serum withdrawal do not induce
significant Schwann cell death in vitro
For the cell death assay, SCs were examined either by direct
observation under a microscope or using a flow cytometer
after exposure to glutamate. It was shown that SCs were
insensitive to high concentration of glutamate (Fig. 3). We
tested serum withdrawal and found that no significant cell
death occurred after 3 hours, indicating SCs can withstand
serum-free condition for a few hours (data not shown).
Figure 1. Schwann cell and astrocyte confrotation assay. (A)
PST/SCs (green) stained with p75 (a marker for SCs) penetrate the
astrocyte (red) boundaries and populate the astrocytic domain (stained
with GFAP, red). (B) A distinctive boundary forms when GFP/SCs
are confronted with astrocytes. The yellow line illustrates a typical
300 μm line drawn to quantify the number of cells that have crossed
the boundary between the astrocytes and either PST/SCs or GFP/SCs.
(C) Graph shows the number of SCs crossed the line. (D) Graph shows
the sizes of astrocytes along the boundary. Experiments were repeated
three times in duplicates. ***p < 0.001. Scale Bar= 100 μm.
Figure 3. Glutamate on the survival of Schwann cells. Phase
contrast images show Schwann cells in culture before (A) and after
exposure to 0.1 mM (B) or 1 mM glutamate (C) for 1 hour at 37°C.
(D) Flow cytometry shows the proportions of live cells before and after
exposure to 0.1 or 1 mM glutamate for 1 hour at 37°C. Glut, glutamate.
4. ATP dose-dependently induces Schwann cell death
via P2X7R activation in vitro
High concentration of ATP can induce death of certain
types of cells. In this study, we found that exposure of SCs
to high concentrations (over 3 mM) of ATP led to
significant cell death in vitro (Fig. 4).
Another indicator of the stress response of astrocytes is
the increased proliferation. In the co-culture assay, it was
found that significantly more astrocytes proliferated in the
GFP/SCs/astrocytes co-culture than in the PST/SCs
/astrocytes co-culture (Fig. 2).
However, when SCs were pre-treated with oxATP, ATP
induced cell death was prevented, which indicates that it is
the P2X7R that mediates the cell death. We have identified
the presence of P2X7R on SCs using a specific P2X7R
antibody (Fig. 5).
Figure 2. Proliferation of astrocytes in co-culture with Schwann
cells. (A–E) GFP/SCs and astrocytes co-culture with BrdU staining.
(F–J) PST/SCs and astrocytes co-culture with BrdU staining. (K)
Percentage of BrdU positive nuclei of astrocytes in co-culture. Scale
bar=100 μm.
5. PSA can partially protect ATP induced Schwann cell
death in vitro
We also found that PSA expression on SCs could partially
protect cell death induced by ATP in vitro (Fig. 6). There was
58
significantly more live cells in PST/SCs group compared with
GFP/SCs group after being exposed to 3, 4, 5 mM ATP.
Whether such protective effect of PSA against ATP induced
cell death reflects the improved survival of PSA-expressing
SCs in spinal cord, further studies need to be carried out.
3. ATP can induce SCs death via P2X7R activation in
vitro, which may be an important factor that causes the
cell death in the early stage after transplantation. It
certainly merits further studies using in vivo models.
PSA can partially protect ATP induced SCs death in
vitro, which is also an interesting phenomenon that
needs to be explored.
REFERENCES
Noble, M. and Murray, K. (1984). “Purified astrocytes
promote the in vitro division of a bipotential glial progenitor
cell.” EMBO J. 3(10): 2243–7.
1. Luo, J., Wu, D., Yeh, J., Richardson, P.M., Bo, X.,
Zhang, Y., Promotion of survival, migration, and
integration of transplanted Schwann cells by overexpressing polysialic acid. (to be submitted).
2. The 11th annual meeting of International Spinal
Research Trust, Glasgow, U.K. Sept., 2009,
“Engineered expression of polysialic acid on Schwann
cells in combination with its expression on spinal cord
enhances Schwann cells migration and integration after
transplantation into the lesioned spinal cord”.
Figure 4. ATP induces Schwann cell death in vitro. Phase
contrast images show Schwann cells in culture before (A) and after
exposure to 0.1 mM ATP (B) or 5 mM ATP (C) for 30min. (D)
Flow cytometry shows the proportions of live cells after exposure to 3,
4, 5 mM ATP for 1 hour at 37ºC. (E) Graph shows the percentage
of live cells with the increasing concentrations of ATP. **p< 0.01.
FUTURE PLANS
We have shown that PSA modified SCs survive better and
integrate well within the normal spinal cord and they do
not cause significant stress to the host astrocytes. The
modified SCs have stronger motility toward the lesion site
in a spinal cord injury model and have the ability to
myelinate axons. In the fourth year we’ll investigate the
effects of transplantation of PST/SCs in combination with
engineered expression of PSA in spinal cells on axonal
regeneration and functional recovery after rat spinal
cord injury.
Figure 5. Expression of P2X7 receptors on cultured Schwann
cells stained with an anti-P2X7 receptor antibody. Schwann cells
were identified with a Schwann cell marker S100. Scale Bar=50 μM.
MILE STONES AND OBJECTIVES
This year we first carried out the study on the interaction of
PST/SCs with astrocytes in vitro to see whether PSA overexpression on SCs causes less stress response in astrocytes.
Using confrontation and co-culture assay, we confirmed the
phenomenon we previously observed in vivo.
Although PSA expressing SCs survived better after
transplantation, the extent of cell death was still significant,
which prompted us to investigate the factors involved in
early cell death. The most striking discovery is that ATP can
induce significant Schwann cell death via P2X7 receptor
activation in vitro. This finding may signal that the blockade
of P2X7R may be beneficial for the survival of Schwann
cells in vivo. PSA can partially protect Schwann cell death
induced by ATP, via an unknown mechanism.
Figure 6. PSA partially protect ATP induced Schwann cell
death in vitro. Flow cytometry apoptosis assay was performed on
PST/SCs and GFP/SCs. Graph shows the percentage of live cells when
Schwann cells exposure to 3, 4, 5 mM ATP 1 hour. The values
represent the mean and SE. Each experiment was repeated three times.
*p<0.05, **p<0.01.
CONCLUSION
1. PST/SCs induce much less stress response in astrocytes
and can integrate well with astroctytes in vitro, which
confirms the phenomenon observed in vivo;
2. The better survival of PST/SCs in the early stage after
transplantation into normal spinal cord may not be due
to that PSA protects the attack of SCs by inflammatory
cells;
59
Spinal cord diffusion imaging: challenging characterization
and prognostic
Torben Schneider1*, Claudia Wheeler-Kingshott1, Daniel Alexander2
of Neuroinflammation, institute of Neurology, UCL, Queen Square, London WC1N 3BG
[email protected]
2Department of Computer Science, UCL, Gower Street, London WC1E 6BT
1Department
INTRODUCTION
Magnetic Resonance Imaging (MRI) is a established tool in
imaging the spinal cord (SC) and has shown to be very
useful in evaluating several aspects of spinal cord injury
(SCI) (see e.g. (Kadoya et al., 1987; Kulkarni et al., 1988)
but conventional MRI is limited to providing only
anatomical information on a macroscopic scale. Diffusionweighted imaging (DWI) is able to provide information
about the microstructure of biological tissue by imaging the
diffusion of water along one diffusion-sensitized direction
(Le Bihan, 1991). Under the assumption of a Gaussian
distribution of displacements, the three-dimensional
diffusion properties inside the tissue can be expressed in
terms of a first order diffusion tensor (DT) (Basser et al.,
1994). Diffusion tensor imaging (DTI) allows to compute
a variety of imaging metrics such as the mean diffusivity
(MD), fractional anisotropy (FA), axial (AD) and radial
diffusivity (RD), which have been shown to correlate with
white matter pathologies in the spinal cord as demonstrated
e.g. by (Schwartz et al., 2005; Ciccarelli et al., 2007).
spinal cord tissue. In the remainder of this report, we will
describe four experiments, each investigating a different
aspect of this process:
• In experiment 1, we carefully optimise both the DTI
acquisition protocol and analysis to obtain imaging
markers that are sensitive to the presence of the
collateral fibres and investigate the position dependency
of these DTI parameters at different levels of the
cervical cord.
• In experiment 2 we propose a novel partial volume
correction method for DTI metrics and investigate the
effect on accuracy and inter-subject variability of these
measurements in healthy subjects. Partial volume
averaging (PVA) is a common problem in quantitative
spinal cord DW imaging due to the small size of the
cord. Therefore we designed the method to be
independent of the DTI acquisition parameters.
• In experiment 3, we turn towards the more
experimental q-space imaging (QSI). For the first time
we present QSI of the cervical cord in a larger cohort of
9 healthy subjects and show inter- and intra-subject
reproducibility over the whole cord area and in
individual tracts.
• In experiment 4, we present a method that provides
DW imaging protocols for directly estimating
microstructural properties like axon radius and density
in the spinal cord. We extend the existing framework of
(Alexander, 2008) that uses a combination of simple
geometric compartments to describe the diffusion signal
in white matter. We incorporate a-priori known
information about the fibre organization in the cord.
We compare the efficacy of our approach with the
original framework using computer simulations and test
our method in a fixated sample of monkey spinal cord.
DTI has great potential in the investigation of spinal
cord disease and is readily available on most standard MRI
scanners. However, the simplistic underlying displacement
probability model is often inaccurate. As a result, different
microstructural changes can have the same effect on DTI
metrics and therefore cannot be told apart by DTI alone.
Over the years, several alternative DW methods have
been developed to overcome the limitations of DTI by
either (a) trying to directly infer the underlying
displacement probability distribution from the DW
measurements (q-space imaging) (Cohen and Assaf, 2002;
Assaf et al., 2005; Farrell et al., 2008) or (b) using DW
signal models based on a-priori anatomical knowledge about
the underlying tissue architecture (Stanisz et al., 1997; Assaf
and Basser, 2005; Assaf et al., 2005).
METHODS
EXPERIMENT 1:
Current DTI studies of the cervical cord mainly concentrate
on the longitudinal fibres of the SC. Only little is known
about the value of DTI for the assessment of the connective
collateral fibres. These fibres rise at an angle with the whitematter longitudinal tracts and enter the spinal cord gray
matter. They interconnect with other areas of the spinal cord
through the central gray matter and form part of many
functional connections within the spinal cord (Carpenter,
1991). Recently it has been demonstrated that the second
eigenvector is corresponding to sprouting collateral fibres
(Mamata et al., 2006). In this study we focus on DTI of the
spinal cord with particular interest in the diffusivity changes
caused by the presence of the collateral fibres. We aim to
Although advanced DW methods offer imaging
biomarkers that can be more specific to individual
pathological processes that occur in spinal cord injury and
recovery, they also have their own technical challenges in
clinical use. Moreover these methods are often only applied
to brain white matter and have not been tested yet in the
spinal cord.
It becomes apparent that there is the need for a
dedicated effort to develop spinal cord diffusion imaging
with the aim of optimising the whole process from the
acquisition design to the analysis methods specific to the
60
investigate whether these DTI parameters are specific to
nerve roots anatomy and therefore have the potential to be
used in spinal cord injury to assess the integrity of the axonal
connections.
Data acquisition
Diffusion-weighted scans are acquired on a 1.5T Signa
scanner (General Electric Company, Milwaukee, WIS)
using a cardiac-gated single shot CO-ZOOM EPI sequence
(Dowell et al., 2009) with imaging parameters TR=5RRs,
TE=95.5 ms, voxel size = 1×1×5 mm3 and an image matrix
of 64×64 (FOV=13×13 mm2). We acquire 8 distributed
diffusion weighted directions (see Table 1) interleaved with
4 non-diffusion weighted directions. A b-factor of
1000 s/mm2 was chosen for optimal DT reconstruction as
recommended in (Jones et al., 1999). To increase signal-tonoise-ratio we initially repeat each scan on each subject
22 times to determine the optimal number of averages
needed. Subsequent scans on the same subject are repeated
15 times (see Data Analysis).
Positioning of DWI scans
We use two sagittal oblique MRI scans to accurately reveal
the location of the neuroforamen, similar to (Goodman et al.,
2006). Based on a standard axial scan of the cervical spinal
cord, we prescribed a sagittal scan that is approximately
parallel to the spinal nerve leaving the neuroforamen (see
Figure 1A). To visualize the spinal cord and spinal nerve root
a second sagittal oblique scan perpendicular to the first one is
acquired. This scan is aligned so that at least one slice is
parallel to the nerve root (see Figure 1B).
Figure 1. Illustration of the positions of the two sagittal oblique scans.
The red colored slices illustrate the ideal positioning of the oblique
scans. (A) The position of the first sagittal scan is intersecting the
neuroforamen (B) The second scan is aligned parallel to the spinal
nerve root.
gx
gy
gz
0
0
0
*
0
0
0
*
1
1
0
*
0
1
1
1
0
1
0
0
0
*
−1
−1
0
*
0
1
−1
1
−1
0
−1
0
1
−1
1
0
*
0
0
0
*
*
Table 1. Gradient direction for DTI acquisition. Lines marked with
* are used for D⊥ reconstruction.
The first oblique scan is then used to position axial scans
so that one slice intersects the spinal nerve. Figure 2 presents
two scans acquired with this positioning. In Figure 2A one
can clearly appreciate the neuroforamina between C4 and
C7. Furthermore, in Figure 2B the spinal nerve root leaving
the spinal cord can be seen. Based on these scans we are able
to accurately position the DW scans with respect to the roots
anatomy. We assume that the diffusion parameters differ
mostly between P1 and P2, i.e. the positions shown in Figure
2C, where P1 coincides with the level of the spinal nerve
root leaving the spinal cord and P2 with the vertebral body.
Data analysis
After acquisition, all magnitude images are linearly
interpolated to a 128×128 matrix on a slice-by-slice basis
resulting in an in-plane resolution of 0.5 × 0.5 mm2. DT
reconstruction is performed using the camino toolbox
(Cook et al., 2006) and maps of the FA and RD are
calculated from the diffusion tensor. In addition we use an
alternative method of measuring diffusivity in the axial
plane (D⊥) from only the 4 co-planar acquisitions with
diffusion gradients perpendicular to the spinal cord as
described in (Fasano et al., 2009). The used diffusion
directions are marked “*” in Table 1. All calculations are
implemented in MATLAB (Mathworks, Natick, MA).
It is well known that in the low SNR regime the diffusion
indices are very prone to estimation errors as shown in
(Basser and Pierpaoli, 1996; Landman et al., 2008). Thus,
for reproducible measurements we need to acquire a
sufficient number of averages in each scan. To determine the
optimal number of averages for each subject we repeat the
diffusion measurements at both slice positions 22 times each
(overall scan time was approx. 1 hour). We then calculate the
diffusion indices described above using a subset of the first N
repeated measurements with N increasing from 1 to 22. A
Figure 2. Example slices of the two sagittal oblique scans from one
example subject. (A) The first sagittal oblique scan showing the
neuroforamina of C4-C7 (white arrows). (B) The second oblique scan
visualizing the spinal nerve roots (white arrows) (C) Positioning of the
two DW axial scans based on the location of the spinal nerve roots.
61
plot of mean diffusion indices over the spinal cord against
the number of averages is presented in Figure 3 for one
representative subject. A significant bias can be observed in
all diffusion indices when less than 10 averages are used. In
none of our subjects significant changes can be seen after
15 repetitions, so we choose the number of averages to be
15 in all subsequent scans.
computed FA maps as in (Wheeler-Kingshott et al., 2002).
A 2D distance transformation is applied to the binary
segmentation masks, i.e., determining the distance d of each
masked voxel to the border of the mask. Assuming that only
voxels close to CSF are affected by PVA, the fuzzy partial
volume correction factor w is then computed as
w=d/max(d) if d<c and w=1 if d≥c where c is a cutoff
distance determined on the basis of the DTI parameter
values (see Figure 4).
Figure 3. Plot of diffusion indices FA, RD and DRD against number
of averages. (A) shows the averaged diffusion parameter at nerve-root
level (B) shows mean parameters at level of the body.
Figure 4. (A) 1D illustration of computed weighting factors. (B)
Isolines of weighting factors overlayed on FA map in one subject.
FA, RD and D⊥ are then quantified over the whole
spinal cord at each position. We semi-automatically segment
the cord area on the average b=0 image of each slice using
ImageJ and the YAWI2D segmentation plug-in.
This approach ensures that for larger spinal cord areas,
the border voxels are weighted less than in the case of small
cord areas. The weighted average using the weighting factors
w is computed for all DTI parameters over the whole
segmented spinal cord area. We determine the optimal
cutoff voxel distance c’ in our dataset so that for c≥c’ the
average DTI parameters over the cord area reach a stable
plateau, i.e., assuming that CSF contribution effects are
minimized. In our data, DTI parameters reach the desired
plateau for c≥2voxels (see Figure 5) and are in agreement
with previously reported values in the healthy cord in
(Wheeler-Kingshott et al., 2002; Ellingson et al., 2007).
Thus the cutoff value c=2 is chosen for further analysis. A
two-tailed paired t-test is performed to compare significance
of differences between uncorrected and corrected
measurements among all subjects.
Pilot study
A pilot study was carried out on 4 healthy female subjects.
For each subject, parameter maps of FA, RD and D⊥ were
calculated as described above. We also calculated colourcoded maps of the DT eigenvectors V1, V2, V3 for each
scan. To assess intra-subject scan-rescan reproducibility, the
scans were repeated with the same parameters after
5–7 days. Reproducibility of parameters was assessed by
computing the coefficient of variation (COV) that is
defined as the ratio of the standard deviation δ and the mean
μ: COV = μ_δ
EXPERIMENT 2:
Due to the small size of the cord and the limited spatial
resolution, a large proportion of voxels are affected by partial
volume averaging (PVA) from surrounding cerebro-spinal
fluid (CSF). Water molecules in CSF are less hindered than
in nervous tissue, resulting in increased diffusivity measures
and decreased anisotropy in PVA voxel (Alexander et al.,
2001; Pfefferbaum and Sullivan, 2003). This can lead to
biased average measurements over specific regions of interest
(ROIs) and over the whole cord volume and potentially
conceal subtle disease effects. We introduce a robust partial
volume correction method for average DTI parameters that
avoids the manual exclusion of PVA affected voxels, and
reduces their contribution depending on their distance to
the interface between spinal cord voxels and CSF. We
investigate the accuracy of our approach in healthy
volunteers and demonstrate that our method significantly
reduces PVA effects on DTI indices.
PVA method
We semi-automatically delineate the cervical cord between
levels C1/2 and C4/5 using the active surface segmentation
(Horsfield et al., 2010) available in Jim6, performed on the
Figure 5. Weighted average and standard deviation among all subjects
for DTI parameters computed with different cutoff values. Columns
corresponding to the chosen cutoff value of 2 are colored red.
62
Data acquisition and DTI analysis
We have acquired diffusion-weighted images of 14 healthy
volunteers (13 male, age=35±11). In each subject cardiac
gated DTI of the cervical cord was performed (acquisition
matrix=96 × 96, sinc interpolated in image space to
192×192, FOV=144×144 mm2, slice thickness = 5 mm,
20 slices, TE=88 ms, TR≈4000 ms) with a total of
100 b=1000 s/mm2 diffusion weighted volumes (20 unique
diffusion directions repeated 5 times) and 5 non-diffusion
weighted volumes. In each voxel the diffusion tensor was
fitted to the data using camino (Cook et al., 2006) and maps
of fractional anisotropy and mean diffusivity, axial
diffusivity and radial diffusivity were generated.
For comparison we also computed the apparent diffusion
coefficient (ADC) from the monoexponential part of the
decay curve (b≤1100s/mm2) as in (Farrell et al., 2008) for
both XY and Z directions.
ROI analysis
We semi-automatically delineate the whole cervical spinal
cord area (SCA) between levels C1 and C3 on the b=0
images using the active surface segmentation (Horsfield et
al., 2010) available in Jim6. On the segmentation mask we
perform a morphological erosion (2 iterations) to exclude
voxels with potential partial-volume average effect from
surrounding CSF. In addition, four regions of interest (ROI)
were manually placed in specific white matter tracts and one
ROI was positioned in the gray matter on all slices between
level C1 and C3. The four white matter regions comprised
the left and right tracts (L&r-LT) running in the lateral
columns and the anterior (AT) and posterior tracts (PT)
similar to (Hesseltine et al., 2006; Freund et al., 2010).
EXPERIMENT 3:
The aim of this study is to investigate accuracy and sensitivity
of tract-specific q-space imaging (QSI) metrics in healthy
controls. The principle of QSI is to exploit the inverse
Fourier relation between the signal S(q) and the
displacement density function (DPDF) p(r), with q being
the diffusion wave number and r being the average
displacement of water molecules (Cohen and Assaf, 2002).
The clinical potential of q-space metrics in the assessment
of white matter pathologies has been shown (Cohen and
Assaf, 2002; Assaf et al., 2008, 2000). However, most clinical
QSI studies only focused on a small number of patients and
failed to demonstrate the reliability of QSI. We test accuracy
and reproducibility in 9 healthy controls and also assess QSI
measures both in-plane (XY) and parallel to the main spinal
cord axis (Z), not presented before. We compare QSI
measures derived in gray matter and different ascending and
descending white tracts of the cervical spinal cord in healthy
subjects and investigate associations between QSI parameters
and conventional apparent diffusion coefficient (ADC)
measures, both in plane and along the cord.
Statistical processing
We report reproducibility as the intra-subject coefficient of
variation (COV=SD measurements/mean measurements)
for the three scan/rescan subjects and the inter-subject COV
among all nine subjects. Further, we compare significant
differences in the group mean values of the ADC parameters
(ADCxy, ADCz) and QSI metrics (P0xy, P0z, FWHMxy,
FWHMz) between tracts by performing the Hotellings-T2
test (confidence interval=99%). To investigate the relevance
of measurements in the Z direction, we compute the same
significance test of XY-only QSI parameters (P0xy,
FWHMxy). Finally, we investigate the relationship between
individual ADC and QSI measurements in XY and
Z directions for each tract using the Spearman’s rho
correlation coefficient.
Study design & Data acquisition
We recruited 9 right-handed male healthy subjects (mean
age 35±11yrs) to be scanned on a 3T Tim Trio (Siemens
Healthcare, Erlangen). Three subjects were recalled for a
second scan on a different day to assess intra-subject
reproducibility of QSI derived parameters. We performed
cardiac-gated high b-value axial DWI (matrix=96×96,
b-spline interpolated to 192×192 in image space,
FOV=144×144 mm2, slice thickness=5 mm, 20 slices,
TE=110 ms, TR≈4000 ms) with 32 b values between
0–3000 s/mm2 in b=50s/mm2 steps (gradient duration=
45 ms, diffusion time=55 ms, maximum gradient strength=
23 mT/m). Three different DWI directions were acquired:
two directions perpendicular (XY) and one parallel (Z) to
the main spinal cord axis. The two perpendicular diffusion
directions were averaged to increase the signal-to-noise ratio.
The measurements were linearly regridded to be equidistant
in q-space and the DPDF was computed using inverse fast
Fourier transformation. To increase the resolution of the
DPDF, the signal was extrapolated in q-space to a maximum
q=166 mm-1 by fitting a bi-exponential decay curve to the
DWI data as suggested by (Cohen and Assaf, 2002). Maps
of the full width at half maximum (FWHM) and zero
displacement probability (P0) were derived for XY and Z.
EXPERIMENT 4:
We develop a method that provides a diffusion weighted
MRI protocol for directly estimating microstructural
properties like axon diameter and density in white matter
with uni-directional distribution of fibres. Feasibility of
estimating axon diameter distribution (Assaf et al., 2008;
Barazany et al., 2009) or mean axon diameter (Alexander et
al., 2010) with diffusion MRI has previously been
demonstrated. A computational framework has been
developed in (Alexander, 2008) which optimizes a multishell HARDI acquisition for sensitivity to those parameters
without knowledge of fibre orientation. We adapt this
framework to provide an optimized set of diffusion
weightings and gradient directions for structures like the
spinal cord with known single fibre orientation (i.e. single
direction approach (SD)). We show that those protocols
improve efficacy of axon measurements particularly in the
presence of low signal-to-noise ratio (SNR). Using
computer simulations we test the ability of our method to
estimate axon information on low SNR data and compared
the results with orientation independent protocols. Finally
we demonstrate the feasibility of measuring axon diameter
and density in a fixed monkey spinal cord on a 4.7T
experimental scanner using the SD approach.
63
Protocol Optimization
We use the protocol optimization in (Alexander, 2008) that
is based on minimizing the Cramér-Rao-Lower-Bound
(CRLB) but modify it to assume the fibre direction to be
known a-priori. The SD protocol was optimized for a
maximum gradient strength of 300 mT/m achievable on an
experimental 4.7T Varian experimental scanner. The
protocol contains four sets of 30 diffusion-weighted
acquisitions, where each set is assigned one individual
diffusion weighting b. Furthermore the direction of the
gradient is optimized for each of the 120 acquisitions.
Although the gradients are allowed to be applied in any
arbitrary direction, all resulting gradients are either aligned
perpendicular or parallel to the assumed fibre orientation.
The final protocol is presented in Table 2. For reference we
also generate an orientation-independent (OI) protocol as
described in (Alexander et al., 2010) for the same
experimental setting.
set
b
[s/mm2]
Nⱍⱍ
N⊥
1
3082.36
0
30
2
2356.83
14
16
RESULTS
EXPERIMENT 1:
Scan/Rescan reproducibility
Table 3 shows the COV of all measured parameters in all
four subjects. It can be seen that our careful approach
towards positioning and analyzing the data allows good
reproducibility (CoV<10% in all but one case) in the
scan/rescan experiment among all subjects. Furthermore, it
can be noted that parameter variation seems to be slightly
elevated in the D⊥ parameter compared to RD but
differences are negligible.
FA:
Gradient Directions
3
13463.65
0
30
4
2366.71
1
29
P1
P2
Subject 1
0.3%
9.9%
Subject 2
11.9%
1.0%
Subject 3
3.8%
2.9%
Subject 4
4.7%
8.1%
P1
P2
Subject 1
3.6%
8.0%
Subject 2
6.6%
4.2%
Subject 3
2.9%
5.7%
Subject 4
1.2%
3.4%
P1
P2
Subject 1
5.2%
8.5%
Subject 2
7.7%
3.6%
Subject 3
3.8%
3.3%
Subject 4
2.4%
2.4%
RD:
Table 2. Optimized diffusion weightings and gradient directions. Nⱍⱍ
is the number of parallel gradient directions; N⊥ is the number of
directions perpendicular to the assumed fibre.
D⊥:
Synthetic Data
The performance of our modified protocol is tested on
synthetic data. Similar to (Alexander, 2008) the noise-free
signal are synthesized from the model and Rician noise is
added to the signal. Using a Markov-Chain-Monte-Carlo
(MCMC) approach, we estimate the posterior distribution
of the model parameters.
Table 3. COV of estimated parameters in all 4 subjects at both
positions calculated from scan/re-scan experiment.
Post-mortem monkey SC
A sample of perfusion fixated cervical spinal cord of
velvetian monkey is scanned on an 4.7 T Varian
experimental scanner using the FD-protocol (Table 2)
together with 12 additional b=0 acquisitions. Imaging
parameters were: FOV: 10×10 mm2, TE=59 ms, TR=2 s,
slice thickness=1.5 mm, 30 slices, 64×64 matrix, 2-D
interpolated to 128×128. We fit a model to the data that
simplifies the AxCaliber model of (Alexander et al., 2010)
which assumes a single radius instead of a radii distribution
as in (Assaf et al., 2008). To stabilize the fitting the
diffusivity in the parallel direction is assumed to be 0.45
μm/ mm2 and the volume fraction of the restricted
compartment f is constrained to be in the range of [0.5,
1.0]. The posterior distribution of the model parameters is
then estimated using the MCMC method in each voxel.
The mean axon-size R calculated by the mean of the
posterior distribution of the radius and the axonal density
a=f/pi/R2 are reported.
Single subject position dependency of measured
parameters
Figure 6 compares the measured diffusion parameters
between the two investigated positions in all subjects. FA
and RD/D⊥ are closely dependent, i.e., when FA is low RD
and D⊥ values are high and increasing FA corresponds to
lower RD and D⊥ in both positions. This implies that
parallel diffusivity in the nerve fibres is position
independent and therefore changes of FA between spinal
cord levels can be explained by different diffusivities
perpendicular to the SC axis. Furthermore, it can be seen
that the two methods of measuring diffusivity crosssectionally give similar values in all subjects apart from
minor differences in their standard deviation through the
entire section of the cord. This can be explained by the lower
number of only 4 diffusion measurements that are used to
reconstruct D⊥ compared to the 8 diffusion directions used
for full DT reconstruction. Since D⊥ requires no
measurements parallel to the fibre, the number of scans
needed for reliable measurements is significantly reduced
64
compared to a DTI acquisition, which can be extremely
beneficial for future studies of spinal cord injury patients.
Between subjects comparison of position dependency
of parameters
Although in individual subjects we can see differences
between position 1 and 2 with little variation in parameters
between scan and rescan, we find that these trends are not
consistent between subjects. In subject 1 and subject 3 we
observe lower RD/ D⊥ and higher FA at nerve root level
compared to the vertebral body (see Figure 6). Subject 2
shows an opposite trend with higher FA at spinal root level
and lower RD/ D⊥ respectively. In subject 4 there appear to
be no differences between the two positions. It is unclear
whether these differences between subjects can simply be
explained by normal variation due to physiological noise or
if they can be attributed to different fibre architecture in
each individual. However, these differences between subjects
also become apparent in the direction of the second
eigenvector. It has been shown by (Mamata et al., 2006) that
the second DT eigenvector is sensitive to the presence of
sprouting fibres in the spinal cord. Figure 7 presents the
color-coded maps of V2 overlaid on the FA map for two
subjects with differing trends in diffusion parameters. Figure
7A displays V2 of subject 2, Figure 7B presents the V2-map
of subject 4. In both cases, the first row shows the result
from the first scan while the second row shows maps derived
from the re-scan. The position of the slice in the second row
(i.e. for the rescan) was chosen to correspond anatomically
with the position of the slice in the first experiment
presented in the first row. This was achieved using our 45°
localization scanning method presented above and using a
printout of the first scan positioning as reference. It has to
be noted that a higher angular in-plane resolution of the
diffusion gradient scheme would be needed to allow
mapping real anatomical directions of the sprouting
peripheral nerves. This however, would increase the number
of acquisitions needed and therefore further increase the
scan time. Moreover, even with our low-resolution scheme,
distinct patterns emerge in each subject in the directions of
the second eigenvector and are consistent over the first and
second scan. Furthermore, in subject 2, where lower FA and
higher RD/D⊥ are present at spinal root level compared to
mid-vertebra level, we also observe different patterns in
position 1 and position 2. In subject 4, which shows no
difference in mean diffusion parameters in P1 and P2, the
V2 map is also similar in both positions. The same geometry
is apparent in the repeated scans for both subjects.
These preliminary findings suggest that the diffusion
measurements in the spinal cord depend indeed on the
presence of sprouting fibres. However, the organization of
those fibres seems to be varying between subjects and needs
to be addressed. The consistency of the patterns at different
slice positioning between scans within each subject is
encouraging because it suggests that DT parameters, and in
particular RD/D⊥, can be used in longitudinal studies to
assess structural changes due to degeneration or regeneration
of fibers.
Figure 6. Measurements of FA, RD and D⊥.Blue bars represent mean
of measurements at nerve root level, red represent the mean of
measurements at mid-vertebra level. Black error bars display the
standard deviation between scan and rescan.
65
for the variability in number of white matter voxels. This
allows more reliable measurements, particularly in patients
who might suffer from white matter atrophy.
Average change in
mean (%)
Average change in
std (%)
FA
+6.2* ± 0.7
+2.1
MD
−12.2* ± 1.3
−10.6
AD
−7.0* ± 1.0
−6.6
RD
−21.0* ± 2.1
−9.8
* Significance p<0.001 (confidence interval 99%)
Table 4. Averaged relative change of mean and standard deviation
between uncorrected and PVA corrected DTI measurements over all
subjects.
EXPERIMENT 3:
Reproducibility
In both intra-subject scan/rescan experiments and among
subjects we observe a consistently lower COV in QSI
metrics compared to ADC measurements (see Figure 8). In
particular, ADCxy shows the largest intra- and inter-subject
variation (>25%) in most tracts. In contrast, tract-specific
QSI measurements vary less, and the majority of observed
CoVs are between 5–10%.
Figure 7. Color coded second eigenvector from DT overlaid on FA
for two subjects. First row shows the first scan, second row the re-scan
of two subjects. First column shows values at spinal root level, second
column shows mid-vertebra position.
Figure 8. Intra- and inter-subject COV for QSI and ADC parameters.
EXPERIMENT 2:
Tract-specific differences
Figure 9 reports QSI and ADC values among all 9 subjects.
We find significant group differences in QSI and ADC
parameters between different white matter tracts. Figure 10
illustrates the tract-specific differences DPDFs in one
exemplary subject. In particular, there are significant
differences in the ADCxy and ADCz between the PT, AT
and lateral tracts (p<0.01) that are not observable with QSI
parameters. On the other hand, XY & Z QSI parameters
showed significant differences between left and right lateral
tracts (p<0.01), as well as differences between AT and l-LT
(p<0.05). However, perpendicular QSI metrics alone do not
show significant differences in any white matter tract. Both
ADC and QSI metrics are significantly different between
white matter tracts and gray matter (p<0.001).
Accuracy and inter-subject variability of PVA corrected
measurements
Table 4 shows lower standard deviation of diffusivity
parameters among subjects when using our PVA correction
method, suggesting lower inter-subject variability compared
to the uncorrected measurements. Furthermore, the largest
reduction of DTI values is observed in the RD (p<0.0001).
We also find moderate decrease in the AD and MD and
increase in FA (all p<0.0001). These results can be explained
by CSF contribution to average measurements in uncorrected
values and are in agreement with similar findings in
simulations (Alexander et al., 2001) and in the brain
(Pfefferbaum and Sullivan, 2003).
Summary
In conclusion, we propose a novel fuzzy partial volume
correction method that removes CSF contribution effects
in measurements of DTI parameters over the whole spinal
cord volume. We avoid fully excluding all potentially CSF
contaminated voxels, and introduce a weighting factor that
is dependent on the size of the cord and therefore accounts
QSI – ADC correlation
We further observe significant correlations between ADC
and QSI parameters in both XY and Z direction in all tracts.
In XY direction, the strongest associations between
FWHMxy and ADCxy are found in AT and PT (p<0.001,
rho≈0.8), although weaker correlations are also found in the
66
r-LT (p<0.05, rho=0.66). In Z direction, we find strong
positive correlations only in PT and l-LT between ADCz
and FWHMz (p<0.001, rho≈0.9) and negative correlation
with P0z (p<0.001,rho≈−0.8) respectively. Over the whole
SCA, we found correlation between all XY and Z
measurements: the strongest correlation is found between
ADCz and P0z (p<0.01, rho=-0.9) and a weak correlation
is found between ADCxy and FWHMxy(p<0.05,rho≈0.7)
and P0xy(p<0.05, rho≈-0.7).
maximal angular error of 10% between the sample and the
assumed direction and distribute the sample orientations
uniformly over that range. As expected, the sample variance
is increased in the SD protocol is similar to the previous
experiment in the OI protocol. However, the accuracy of
diameter estimation remains to be higher in the SD protocol
than in the OI protocol.
Figure 9. Mean and standard error of tract-specific QSI and ADC in
XY and Z direction over all subjects.
Figure 11. Histogram of MCMC samples for (A) orientation
independent and (B) single direction protocol for diameters 2,4 and
10 μm.
Figure 10. Exemplary FWHM maps and DPDFs in five voxels
placed in white matter tracts AT, PT, l&r-LT as well as inside GM for
XY and Z directions.
Figure 12. Histogram of estimated diameters (A) for orientation
independent and (B) single direction protocol with maximal angular
error of 10% in sample orientation.
Summary
QSI metrics obtained without sequence development, using
standard DWI protocol available on a 3T clinical scanner,
show a good reproducibility that is superior to simple ADC
analysis. We observe tract-specific correlations between
ADC and QSI parameters. However, especially in the lateral
tracts, associations are weaker than in the anterior and
posterior tracts, suggesting additional information in both
XY and Z from QSI analysis in these columns. We further
demonstrate that QSI parameters provides complementary
metrics that allow discrimination of white matter tracts in
healthy controls that cannot be distinguished with ADC
alone. Our findings also suggest that the Z direction
provides additional information to perpendicular
measurements.
Postmortem monkey SC
Figure 13 presents maps of axon diameter (A) and axonal
density (B) in the upper cervical spinal cord obtained from
the post-mortem monkey SC scan. We observe left-right
symmetry of axon diameter and density in all tracts, which
corresponds with the known structure of the SC. Parameters
are also consistent along the SC within the limits of
anatomical variation. Further we can discriminate axon
diameter and axonal density between anatomically different
white matter tracts. Dorsal and lateral sensory tracts show
small axons diameters between 1–4 μm and a density of
0.03–0.08 μm-2. The smallest axon calibers (<1.5 μm) are
observed in the dorsal columns (DC) while mean axon size
in the anterolateral column (ALC) is 1.5–2.5 μm. The
largest axons (3–4 μm) are found in the corticospinal tract
(CST) together with low density of 0.01–0.02 μm-2.
EXPERIMENT 4:
Summary
We present a diffusion imaging protocol that allows to
estimate axon radius and density that is optimized for an apriori known fibre direction and compared it to the
orientational independent protocol of (Alexander et al.,
2010). Using computer simulations we showed that our
protocol allows higher accuracy in radii estimations under
low SNR even if the real fibre orientation differs slightly
from the assumed one. A preliminary study on post-mortem
Synthetic data
Figure 11 shows histograms of axon diameters sampled from
synthetic data from of the OI protocol (A) and our modified
SD approach (B). All fibre samples are oriented parallel to
the direction assumed in the protocol. The SD protocol
clearly lowers the sample variance compared to the reference
OI protocol particularly for smaller diameters 2 μm and
4 μm. In Figure 12 the same experiment is repeated but we
introduce some uncertainty in fibre orientation. We allow a
67
monkey SC shows the feasibility of measuring axon radii
and density under low SNR and high resolution.
Furthermore using their axonal characteristics we were able
to distinguish different tracts of the SC.
http://www.tau.ac.il/chemistry/cohen/Documents/47.pdf.
Assaf, Y., Basser, P.J. (2005) Composite hindered and
restricted model of diffusion (CHARMED) MR imaging
of the human brain. Neuroimage. 27:48–58
Assaf, Y., Blumenfeld-Katzir, T., Yovel, Y., Basser, P.J. (2008)
AxCaliber: a method for measuring axon diameter
distribution from diffusion MRI. Magnetic resonance in
medicine: official journal of the Society of Magnetic
Resonance in Medicine / Society of Magnetic Resonance in
Medicine 59:1347–54
Assaf, Y., Chapman, J., Ben-Bashat, D., Hendler, T., Segev,
Y., Korczyn, A.D., Graif, M., Cohen, Y. (2005) White
matter changes in multiple sclerosis: correlation of q-space
diffusion MRI and 1H MRS. Magn. Reson. Imaging.
23:703–710
Barazany, D., Basser, P., Assaf, Y. (2009) In vivo
measurement of axon diameter distribution in the corpus
callosum of rat brain. Brain. Available at: http://brain.
oxfordjournals.org/cgi/content/abstract/132/5/1210.
Basser, P., Matiello, J., Le Bihan, D. (1994) MR diffusion
tensor spectroscopy and imaging. Biophysical Journal.
66:259–267
Basser, P.J., Pierpaoli, C. (1996) Microstructural and
physiological features of tissues elucidated by quantitativediffusion-tensor MRI. J. Magn. Reson. B. 111:209–219
Carpenter, M. (1991) Core text of neuroanatomy 4th ed.
Baltimore: Williams & Wilkins.
Ciccarelli, O., Wheeler-Kingshott, C.A., McLean, M.A.,
Cercignani, M., Wimpey, K., Miller, D.H., Thompson, A.J.
(2007) Spinal cord spectroscopy and diffusion-based
tractography to assess acute disability in multiple sclerosis.
Brain. 130:2220–2231
Cohen, Y., Assaf, Y. (2002) High b-value q-space analyzed
diffusion-weighted MRS and MRI in neuronal tissues – a
technical review. NMR in Biomedicine. 15:516–42
Cook, P.A., Bai, Y., Nedjati-Gilani, S., Seunarine, K.K.,
Hall, M.G., Parker, G.J., Alexander, D.C. (2006) Camino:
open-source diffusion-MRI reconstruction and processing.
14th Scientific Meeting of the International Society for
Magnetic Resonance in Medicine: 2759
Dowell, N.G., Jenkins, T.M., Ciccarelli, O., Miller, D.H.,
Wheeler-Kingshott, C.A.M. (2009) Contiguous-slice zonally
oblique multislice (CO-ZOOM) diffusion tensor imaging:
examples of in vivo spinal cord and optic nerve applications.
Journal of magnetic resonance imaging: JMRI. 29:454–60
Ellingson, B.M., Ulmer, J.L., Schmit, B.D. (2007) Optimal
diffusion tensor indices for imaging the human spinal cord.
Biomed. Sci. Instrum. 43:128–133
Farrell, J., Smith, S., Gordon-Lipkin, E., Reich, D. (2008)
High b-value q-space diffusion-weighted MRI of the human
cervical spinal cord in vivo: feasibility \ldots. Magnetic
Resonance in Medicine Available at: http://www3.
interscience.wiley.com/journal/118821795/abstract.
Fasano, F., Bozzali, M., Cercignani, M., Hagberg, G.E.
(2009) A highly sensitive radial diffusion measurement
method for white matter tract investigation. Magnetic
Resonance Imaging. 27:519–30
Freund, P., Wheeler-Kingshott, C., Jackson, J., Miller, D.,
Thompson, A., Ciccarelli, O. (2010) Recovery after spinal
cord relapse in multiple sclerosis is predicted by radial
diffusivity. Mult. Scler. 16:1193–1202
Figure 13. Axial slice of upper cervical cord showing (A) axon
diameter in μm and (B) axonal density in μm-2 in the corticospinal
tracts (CST), anterolateral column (ALC) and dorsal column (DC).
CONCLUSION
All experiments described above are working towards the
common aim of defining novel imaging markers that are
more specific to spinal cord pathologies. We have also assessed
reproducibility and correlations between different metrics.
On the one hand, we developed methods that improve
the quality of DTI metrics and can provide quantification
of collateral fibre organization in the cord. Despite the
limitations of DTI, it is still the most clinically established
DWI technique and is readily available on most clinical
scanners. Therefore our techniques can be easily adapted in
clinical studies.
On the other hand, we tested alternative acquisition
strategies such as QSI and direct model-based axon diameter
and density estimation. Our preliminary results are
promising but the acquisition can be challenging and
requires careful optimisation. However, those techniques
can be of great value in the assessment of degenerative white
matter diseases like multiple sclerosis and will also help
evaluation of severity of and recovery after SCI.
REFERENCES
Alexander, A.L., Hasan, K.M., Lazar, M., Tsuruda, J.S.,
Parker, D.L. (2001) Analysis of partial volume effects in
diffusion-tensor MRI. Magnetic resonance in medicine:
official journal of the Society of Magnetic Resonance in
Medicine / Society of Magnetic Resonance in Medicine
45:770–80
Alexander, D.C. (2008) A general framework for
experiment design in diffusion MRI and its application in
measuring direct tissue-microstructure features. Magn.
Reson. Med. 60:439–448
Alexander, D.C., Hubbard, P.L., Hall, M.G., Moore, E.A.,
Ptito, M., Parker, G.J.M., Dyrby, T.B. (2010)
Orientationally invariant indices of axon diameter and
density from diffusion MRI. Neuroimage. 52:1374–1389
Assaf, Y., Mayk, A., Cohen, Y. (2000) Displacement
imaging of spinal cord using q-space diffusion-weighted
MRI. Magnetic Resonance in Medicine Available at:
68
Goodman, B.S., Geffen, J.F., Mallempati, S., Noble, B.R.
(2006) MRI images at a 45-degree angle through the
cervical neural foramina: a technique for improved
visualization. Pain Physician. 9:327–32
Hesseltine, S.M., Law, M., Babb, J., Rad, M., Lopez, S., Ge,
Y., Johnson, G., Grossman, R.I. (2006) Diffusion tensor
imaging in multiple sclerosis: assessment of regional
differences in the axial plane within normal-appearing cervical
spinal cord. AJNR Am. J. Neuroradiol. 27:1189–1193
Horsfield, M.A., Sala, S., Neema, M., Absinta, M., Bakshi,
A., Sormani, M.P., Rocca, M.A., Bakshi, R., Filippi, M.
(2010) Rapid semi-automatic segmentation of the spinal
cord from magnetic resonance images: application in
multiple sclerosis. Neuroimage. 50:446–455
Jones, D.K., Horsfield, M.A., Simmons, A. (1999) Optimal
strategies for measuring diffusion in anisotropic systems by
magnetic resonance imaging. Magnetic Resonance in
Medicine. 42:515–25
Kadoya, S., Nakamura, T., Kobayashi, S., Yamamoto, I.
(1987) Magnetic resonance imaging of acute spinal cord
injury. Report of three cases. Neuroradiology. 29:252–255
Kulkarni, M.V., Bondurant, F.J., Rose, S.L., Narayana, P.A.
(1988) 1.5 tesla magnetic resonance imaging of acute spinal
trauma. Radiographics. 8:1059–1082
Landman, B., Farrell, J., Huang, H., Prince, J. (2008)
Diffusion tensor imaging at low SNR: nonmonotonic
behaviors of tensor contrasts. Magnetic Resonance Imaging
Available at: http://linkinghub.elsevier.com/retrieve/pii/
S0730725X08000763
Le Bihan, D. (1991) Molecular diffusion nuclear magnetic
resonance imaging. Magnetic Resonance Quarterly. 7:1–30
Mamata, H., Girolami, U.D., Hoge, W.S., Jolesz, F.A.,
Maier, S.E. (2006) Collateral nerve fibers in human spinal
cord: Visualization with magnetic resonance diffusion
tensor imaging. NeuroImage. 31:24–30
Pfefferbaum, A., Sullivan, E.V. (2003) Increased brain white
matter diffusivity in normal adult aging: relationship to
anisotropy and partial voluming. Magn. Reson. Med.
49:953–961
Schwartz, E.D., Duda, J., Shumsky, J.S., Cooper, E.T., Gee,
J. (2005) Spinal cord diffusion tensor imaging and fiber
tracking can identify white matter tract disruption and glial
scar orientation following lateral funiculotomy. Journal of
Neurotrauma. 22:1388–1398
Stanisz, G.J., Szafer, A., Wright, G.A., Henkelman, R.M.
(1997) An analytical model of restricted diffusion in bovine
optic nerve. Magn. Reson. Med. 37:103–111
Wheeler-Kingshott, C.A., Hickman, S.J., Parker, G.J.,
Ciccarelli, O., Symms, M.R., Miller, D.H., Barker, G.J.
(2002) Investigating cervical spinal cord structure using
axial diffusion tensor imaging. NeuroImage. 16:93–102
Freund, P.*, Schneider, T.*, Nagy, Z., Wheeler-Kingshott,
C.A.M., Thompson, A.J. Diffusion Tensor Imaging Detects
Axonal Degeneration and its Extent is Associated with
Disability in Chronic Spinal Cord Injury.
* equal contribution
CONFERENCE PAPERS
Schneider, T., Wheeler-Kingshott, C.A.M., Alexander, D.C.
(2010) In-vivo estimates of axonal characteristics using
optimized diffusion MRI protocols for single fibre
orientation. 13th International Conference on Medical
Image Computing and Computer-Assisted Intervention
(MICCAI2010).
POSTER PRESENTATIONS
Schneider, T., Alexander, D.C., Wheeler-Kingshott, C.A.M.
(2010) Optimized diffusion MRI protocols for estimating
axon diameter with known fibre orientation. 18th Scientific
Meeting of the International Society for Magnetic
Resonance in Medicine.
(2009) Preliminary Investigation of Position Dependency
of Radial Diffusivity in the Cervical Spinal Cord. 18th
British Chapter ISMRM Annual Symposium.
(2009) Preliminary Investigation of Position Dependency
of Radial Diffusivity in the Cervical Spinal Cord. 17th
Scientific Meeting of the International Society for Magnetic
Resonance in Medicine.
INVITED TALKS
“Optimising Diffusion Pulse Sequences for Investigation of
Spinal Cord Microstructure”
Departmental Seminar
Danish Research Centre for Magnetic Resonance,
Copenhagen, August 2009
“Spinal Cord Diffusion MRI”
CMIC Seminar
Centre for Medical Image Computing, UCL, London,
January 2010
“Imaging microstructure in the spinal cord with diffusion
MRI”
Imaging & Biophysics Unit Seminar Series
Institute of Child Health, UCL, London, November 2010
FUTURE PLANS
We will concentrate the development of acquisition
strategies to directly measure axon diameter and density in
the spinal cord. Our results from excised tissue showed great
potential but they still need to be confirmed in scans of live
tissue. Therefore, we are currently focusing on the
implementation of our protocols on our 3T Philips Achieva
scanner (Philips Healthcare, Eindhoven, NL).
JOURNAL PAPERS IN PREPARATION
Ciccarelli, O., Thomas, D.L., De Vita, E., WheelerKingshott, C.A.M., Schneider, T., Kachramanoglou, C.,
Toosy, A.T., Thompson, A.J. Spinal cord spectroscopy,
tractography and q-space MRI in a case of NMO spectrum
disorder.
Furthermore we will continue to improve spinal cord
DTI analysis and acquisition techniques on our 3T Philips
Achieva scanner (Philips Healthcare, Eindhoven, NL) to
support ongoing spinal cord research at the UCL Institute
of Neurology.
69
Project Grant Reports
Investigation into the conduction properties of surviving axons following chronic spinal cord contusion and whether therapeutic
intervention can restore normal function
Katalin Bartus, Elizabeth Bradbury and Stephen McMahon
Karim Fouad and Wolfram Tetzlaff
Zhigang He
Lyn B. Jakeman and D. Michele Basso
Comparative evaluation of surgical and pharmacological methods for removal of a mature scar in a chronic spinal cord injury model
and subsequent regeneration of stimulated sensory neurons through the treated wound
Daljeet Mahay, Ann Logan, Martin Berry, Zubair Ahmed & Ana Maria Ginzalez
John Riddell and Susan Barnett
Mark Tuszynski and Ken Kadoya
70
Investigation into the conduction properties of surviving
axons following chronic spinal cord contusion and whether
*Katalin Bartus, Elizabeth Bradbury and Stephen McMahon
King’s College London, London SE1 3QD, UK
*post doc
[email protected]
INTRODUCTION
Traumatic spinal cord injury (SCI) leads to severe deficits in
motor, sensory and autonomic function below the level of
the injury (Olson, 2002; Onifer et al., 2007). The initial
insult to the spinal cord can be classified as a contusion,
compression or laceration injury (Bunge et al., 1997;
Norenberg et al., 2004). Intensive research efforts are
focused on developing therapies to treat SCI, however few
of the most promising therapeutic options (such as
chondroitinase ABC [ChABC]) have been assessed in
translational models that mimic human SCI pathology,
having mainly been tested in discrete injury models where
specific tracts have been injured as opposed to the
heterogeneous complex pathological profile of a contusion
injury. Contusion injuries are the most common form of
SCI in humans, which are the result of a blunt trauma to the
spinal cord, leaving the dura mater intact but damaging the
underlying nervous tissue (Bunge et al., 1997). Rat models
of spinal contusion injury closely resemble the human
pathology, showing predominant grey matter degeneration
and cyst formation surrounded by a preserved rim of white
matter (Basso et al., 1996; Metz et al., 2000; Scheff et al.,
2003; Onifer et al., 2007). Therefore, this experimental
contusion model is a particularly clinically relevant tool to
investigate any potential therapeutic interventions, and
detailed characterization of this model will aid the
development of repair strategies following SCI.
Using a combination of electrophysiological,
behavioural and anatomical analysis, our study aims to carry
out a detailed characterization of this clinically relevant SCI
model, assessing functional changes at acute, sub-acute and
chronic post-injury time points, which will form the basis
for testing potential treatments. We aim to evaluate changes
in conduction across a time course spanning acute (1 day –
1 week), sub-acute (2–4 weeks) and chronic (12 weeks)
stages post-injury. We also aim to determine how changes in
conduction across a contusion injury site correlate with
behavioural and morphological outcome. Subsequently, our
goal is to test whether some promising therapies enhance
the function of surviving axons following contusion injury
and whether this can improve functional recovery. Our
focus will be on ChABC therapy, which has been shown to
have a number of beneficial effects in vivo, in models of SCI,
including axonal regeneration, compensatory plasticity of
spared systems, neuroprotection and recovery of motor
function after injury (e.g. Bradbury et al., 2002; Barritt et
al., 2006; Carter et al., 2008).
METHODS
Animals:
Adult female Sprague-Dawley rats (n=54, Harlan
laboratories) were used for all experiments in this study and
were all 200–220 g at the time of contusion surgery.
Animals were housed under a 12 hour light/dark cycle with
free access to food and water.
An important similarity between the human pathology
and that observed in rats is the spared tissue containing
axons that surround the cavity. However, the viability and
function of these surviving axons, and whether they may
present a potential target for therapy, are poorly understood.
Furthermore, there has been controversy regarding the state
of myelination of the axonal projections remaining after
contusion injury, which is an important aspect that may
correlate with surviving axons failing to conduct under
physiological conditions. While some evidence suggests that
there is chronic ongoing demyelination following SCI in
both humans and rats, with incomplete remyelination
taking place (Guest et al., 2005; Totoiu and Kierstead,
2005), other studies have found that despite initial
demyelination after injury, axons are fully remyelinated in
chronic injury stages (Lasiene et al., 2008). Human SCI is
commonly characterized by incomplete injury with some
axonal projections remaining intact (Bunge et al., 1993;
Quencer and Bunge, 1996), and the assessment of the
structural and functional status of surviving axons will be
important to evaluate.
Study Design:
Animals were pre-trained on a number of behavioural tasks
and then underwent a moderate severity contusion injury.
At various post-injury time points (spanning acute to
chronic stages of SCI) the animals were used for either
terminal electrophysiological assessments or for EM
ultrastructural analysis. Behavioural testing was carried out
on all remaining animals throughout the study.
Spinal cord contusion and post-operative care:
Animals were anaesthetised using a mixture of ketamine
(60 mg/kg) and medetomidine (0.25 mg/kg), administered
i.p. Single doses of 0.05 mg/kg buprenorphine and 5 mg/kg
carprofen were given subcutaneously at the time of
induction and the morning after surgery. Laminectomies
were performed at the vertebral level T10 to expose the
spinal cord, and rats received a 150 kD contusion injury
using the Infinite Horizon (IH) device (Precision Systems
Instrumentation, Lexington, KY) (Scheff et al., 2003).
Impact analysis, including actual force applied to the spinal
71
cord, tissue displacement and velocity of impactor were
recorded. Overlying muscle and skin were sutured,
anaesthesia was reversed using atipamezole hydrochloride
(1mg/kg administered i.p.), and animals were allowed to
recover overnight in cages placed on a heated blanket. Saline
(3–5 ml) and baytril (5 mg/kg) were given subcutaneously
twice daily for 3 and 7 days, respectively, post-injury.
Bladders were manually expressed twice daily until reflexive
emptying returned, typically 6–9 days post-injury.
below the lesion site (see Figure 4A for schematic). Firstly, the
number of single units present in each filament was counted
whilst stimulating caudal to the lesion (filaments were of such
a size that 5–10 single units were normally present). This was
then repeated whilst stimulating rostral to the lesion in order
to allow the calculation of the percentage of nerve fibres
capable of conducting through the lesion. The dorsal columns
were stimulated using 0.2 ms duration, square wave pulses at
a frequency of 1 Hz and an incrementally increasing intensity
(0–600 μA). At the end of each experiment measurements
were made of the distance from each stimulating electrode to
the recording electrode to allow for the calculation of
conduction velocities.
Behavioural assessment:
BBB locomotor scoring
Open field hindlimb locomotor function was assessed using
the Basso, Beattie and Bresnahan (BBB) locomotor rating
scale (Basso et al., 1995). Briefly, this involved placing the
animal in a circular open field (1 m diameter) with two
experimenters assessing both hindlimbs for individual joint
movements, plantar placements of the paw, weight support,
consistency of stepping and hindlimb/forelimb coordination,
level of toe clearance during stepping, and overall trunk
stability. Notes were made on each of these and a score was
calculated according to the 22 point (0–21) BBB scale.
Testing was carried out for 4 min on each animal and took
place on days 1, 3, 5 and 7 post-injury and weekly thereafter.
Histology:
In vivo electrophysiology:
Toludine blue staining of semi-thin sections and
electronmicroscopy:
Animals were terminally anaesthetised using sodium
pentobarbital (Euthatal, 200 mg/ml) and transcardially
perfused with 0.9% saline followed by 3% glutaraldehyde and
4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB).
Immediately after perfusion a section of spinal cord was
removed (ca. 20 mm) with the lesion epicentre located
centrally. Approximately 1–2 mm sections were taken from
the lesion epicentre and from the rostral and caudal lesion
borders and post-fixed in the same fixative for a minimum of
48–72 hours at 4°C. After washing in 0.1 M PB, sections were
osmicated using 1.5% Os in 0.2 M PB, dehydrated in a
graded ethanol series, and embedded in resin (TAAB
Embedding Materials). Semi-thin sections (1 μm) were cut on
a microtome and stained with 0.25% Toluidine blue solution
before being examined using a Zeiss Axioskop microscope
equipped with a Zeiss AxioCam MRm. Ultra-thin sections
were cut on an ultramicrotome and stained with lead uranyl
acetate by the Centre for Ultrastructural Imaging (King’s
College London). Sections were mounted on unsupported
100 mesh grids, and uninhibited sections were visualized on
a Hitachi H7600 transmission electron microscope.
Measuring percentage conduction through the lesion and
conduction velocity
Electrophysiological assessments were performed at a number
of post-injury time points from acute to chronic stages (1, 7,
14, 28, and 84 days post-injury). Animals were deeply
anaesthetised with urethane (1.25 g/kg), administered i.p.,
and depth of anaesthesia was regularly assessed by monitoring
pedal withdrawal reflexes and respiratory rate. Core
temperature was maintained close to 37°C using a selfregulating heated blanket connected to a rectal temperature
probe. Laminectomy was then performed to remove the
dorsal portions of vertebrae T7 – L5 to expose the underlying
spinal cord and dorsal roots. The dura mater was removed
from the spinal cord, exposed nervous tissue was covered with
mineral oil, and silver-wire stimulating electrodes were placed
over the midline approximately 5 mm rostral and caudal to
the T10 lesion site. Small filaments were then teased from
various dorsal roots (L3 – S2) and mounted on silver-wire
recording electrodes, allowing for the recording and
quantification of single units (activity of single nerve fibres)
from each of these filaments whilst stimulating either above or
Immunohistochemistry
Following electrophysiological experiments, animals were
transcardially perfused with 0.9% saline followed by 4% PFA
in 0.1 M PB. Immediately after perfusion a section of spinal
cord was removed (ca. 20 mm), with the lesion epicentre
located centrally, and post-fixed in 4% PFA (in 0.1 M PB)
for 2 hours before being cryoprotected in 20% sucrose (in
0.1 M PB) for 48 hours. The tissue was then freezeembedded in OCT. Transverse sections (30 μm) were cut
using a cryostat and mounted onto positively charged slides.
Sections were double stained for GFAP (to label astrocytes)
and NeuN (to label neurons), or protein zero (to label
peripheral myelin) and NF200 (to label axons), or
proteolipid protein (PLP, to label central myelin) and
NF200). Briefly, after blocking with 10% donkey serum in
phosphate-buffered saline (PBS) containing 0.2% Triton X100 and 0.1% sodium azide (PBST azide) for 30–60 min at
room temperature (RT), the sections were incubated in
PBST azide containing polyclonal rabbit anti-GFAP
(1:2000, DakoCytomation) and monoclonal mouse antiNeuN (1:500, Millipore), or chicken anti-protein zero
Horizontal Ladder
Animals were trained for 1 week prior to injury to run across
a 1 m long horizontal ladder with unevenly spaced rungs.
On the final day of training the animals were filmed and
the recordings were later analysed in slow motion, allowing
quantification of the total number of hindlimb footslips
during the course of three runs across the ladder, giving each
animal a baseline score. This testing and analysis procedure
was then repeated once a week post-injury, beginning at day
7 at which time animals had regained their stepping ability
as verified in the BBB testing.
72
(1:500, Pierce Biotechnology) and mouse anti-NF200
(1:400, Sigma Aldrich) overnight at RT. After 4 washes of 5
minutes in PBS, sections were incubated in PBST azide
containing the complementary secondary antibodies
conjugated with Alexa 488 (1:1000, Invitrogen) or Alexa
546 (1:1000, Invitrogen) for 4–5 hours at RT. After 4 washes
of 5 minutes in PBS, sections were then coverslipped with
Vectashield mounting medium (Vector laboratories).
Tyramide signal amplification (PerkinElmer) was employed
to stain for PLP and NF200, the primary antibodies being
mouse anti-PLP (1:2500, Millipore) and mouse antiNF200. Sections were examined using a Zeiss Imager. Z1
microscope equipped with a Zeiss AxioCam MRm.
remaining significantly different from the mean BBB score
pre-injury (12.9 ± 0.3 vs. 21 ± 0, p < 0.001, one-way
ANOVA-Tukey’s post-hoc test). Additionally, we assessed the
number of hindlimb footslips whilst the animal walked across
a horizontal ladder. This walking ability test also showed a
substantial deficit, with the largest deficit being observed at
the earliest post injury time points and partial recovery
reaching a plateau with no further improvement after the first
few weeks after injury (Figure 3B). The largest deficit, as in
the BBB testing, was observed at the earliest time point
assessed (41.2 ± 3.6 at 7 dpi, n = 26, Figure 3B), with recovery
leveling off by 35 dpi (17.9 ± 2.6, n = 18, Figure 3B) and no
further improvement recorded beyond this intermediate time
point (16.6 ± 2.5 at 84 dpi, n = 18, Figure 3B). This persistent
deficit was significantly different from the mean number of
foot slips recorded pre-injury (0.9 ± 0.2, n = 26, p < 0.001,
one-way ANOVA-Tukey’s post-hoc test).
RESULTS
The grant has been running for 12 months. In this time we
have established the surgical techniques for producing
reproducible spinal cord injuries with similar pathology to
a human spinal injury (Figures 1 and 2A), with typical glial
scarring and neuronal cell death (Figure 2B).
Figure 2. Establishing optimal lesion severity and reproducible spinal
cord injuries using the IH device. A: A moderate severity contusion
injury (150 kD) in rats gives a classic injury pathology typical to that
seen in human spinal injuries (Figure 1), with a cavity starting to
form at immediate (2 week) time points, being more pronounced at
chronic (12 week) time points after a contusion injury with a spared
tissue rim containing axons. B: Merged images of GFAP (red) and
NeuN (green) immunoreactivity at the lesion epicentre at 2 weeks postinjury and 12 weeks post-injury compared to uninjured (naïve)
animals. Exemplified are the typical increase in GFAP
immunoreactivity and decrease in NeuN immunoreactivity over time,
indicating reactive gliosis around the injury site and demonstrating
marked loss of neuronal cell bodies in the grey matter respectively.
Figure 1. Comparison of a typical human SCI with a SCI in rats
produced by an experimental contusion injury device. Cross sections of
human and rat spinal cord tissue show the normal uninjured spinal
cord (A and B) and the spinal cord following a contusion injury (C
and D). The gross pathology observed in rat spinal cord tissue is very
similar to the human spinal cord tissue, providing a good model for
experimental studies aimed at repairing injured spinal cord.
To determine the extent of the functional deficit and the
time course of any spontaneous functional recovery after a
moderate lesion severity with a force of 150 kD, behavioural
tests aimed at detecting hindlimb locomotor deficits, such as
BBB assessment (Basso et al., 1995), were carried out on
contused rats over a period of up to 12 weeks. Acutely after
injury, animals showed severely impaired walking ability
during BBB scoring, with an average score of 3.43 ± 0.6 at
one day post-injury (1 dpi; n = 34, Figure 3A), reflecting only
slight movement of two joints of each hindlimb. Scores
recovered to 10.2 ± 0.4 at 7 dpi (n = 33, Figure 3A),
associated with occasional weight-supported steps with no
hindlimb/forelimb coordination, and to 11.9 ± 0.3 at 14 dpi
(n = 28, Figure 3A), which represents weight support and
some stepping ability of the hind paws but no
hindlimb/forelimb coordination. Recovery plateaued at a
walking ability score of 12, being recorded up to 84 dpi
Figure 3. Behavioural assessment of locomotor abilities. Both BBB
(A) and horizontal ladder (B) testing show an initially severe deficit
followed by some spontaneous recovery over the first few weeks after
injury. Partial recovery observed in early stages after injury reaches a
steady-state with no further functional improvement in chronic stages
post-contusion, with animals remaining severely impaired.
73
The behavioural tests employed in this study mainly
assess the overall motor deficit and ensuing pattern of motor
recovery following spinal contusion injury. In order to assess
the effect of a contusion injury on axonal function, a novel
electrophysiological technique was developed that allows
recordings from a specific spinal pathway in vivo. During the
time that the grant has been running, we have established
and refined the techniques for measuring conduction across
a contusion injury in vivo, in the rat spinal cord and have
completed the time course study assessing conduction from
acute to chronic stages. We assessed the function of the
ascending sensory pathway which travels in the spinal dorsal
columns. A comparison was made between the intact
pathway (below the lesion) and the injured pathway (above
the lesion) by stimulating below and above the lesion and
recording from the dorsal root filaments (Figure 4A). The
number of single units (activity of a single nerve fibre)
present in a given dorsal root filament was compared when
stimulating either below or above the lesion (Figure 4B). In
uninjured animals all single units recorded when stimulating
below the spinal level of the lesion are also detected when
stimulating above the lesion (Figure 4C). Following a
contusion injury, axonal conduction is entirely abolished
(0% conduction through the lesion at 1 dpi, n = 4, figure
4C), followed by a gradual, partial recovery over a chronic
time course that reaches 21.28%± conduction through the
lesion by 84 dpi (Figure 4C). This is a significant recovery
compared with the initial deficit observed at acute time
points after injury (p < 0.001, one-way ANOVA-Tukey’s
post-hoc test), but remains a significant deficit when
compared to uninjured animals (p < 0.001, one-way
ANOVA-Tukey’s post-hoc test). We also examined the effect
of a contusion injury on the conduction velocity of viable
axons (Figure 4D). Significantly reduced conduction
velocities were recorded when stimulating above the lesion
when compared to below the lesion at each post-injury time
point assessed (p < 0.05, paired t-test, Figure 4D).
partially, by Schwann cells. Schwann cell-mediated
remyelination at chronic post-injury time points was
confirmed by immunohistochemical detection of protein
zero (Figure 6A), which was absent in uninjured spinal cords
(Figure 6B).
Figure 4. Electrophysiological assessment of the ascending sensory
pathway. A: Schematic diagram of the electrophysiological protocol.
Stimulating electrodes (S) are positioned rostral and caudal to the T10
contusion site. Small filaments are teased away from different roots
and positioned across a “hook-shaped” recording electrode (R). B:
Example traces illustrating a number of single units conducting in a
dorsal root filament when stimulating below the lesion site, and a
solitary single unit conducting when stimulating above the contusion
site. Arrows indicate examples of single units; in the presence of
multiple units temporal summation can occur (arrowhead), but
individual units can be identified due to their different thresholds for
activation. C: Changes in number of axons conducting after a
contusion injury. The function of the ascending sensory pathway is
initially abolished by 150 kD contusion (0% of axons capable of
conducting through the lesion), but partially recovers over time. D:
Changes in conduction velocity after a contusion injury. Injury
significantly decreases conduction velocity when stimulating above the
lesion at all time points when compared to uninjured animals.
Tissue has also been processed for ultrastructural
assessment of uninjured spinal cords as well as cords at acute
(1 week, n = 3), sub-acute (4 week, n = 4) and chronic
(12 week, n = 4) time points after contusion injury.
Toluidine blue-stained sections of the lesion epicenter at
progressive time points illustrated the large-scale
morphological changes over time that occur following
contusion, highlighting mass tissue necrosis at acute postinjury stages and cavity formation at later time points
surrounded by a rim of residual white matter (Figure 5A).
Ultra-thin sections were cut and processed for
electronmicroscopy from the boxed area shown in Figure
5A, containing the ascending dorsal column axons that were
assessed in the electrophysiology. All electronmicroscopy
data has been collected and is currently being processed for
quantification of the changes in myelin sheath thickness and
total number of ascending sensory dorsal column axons at
the different stages following injury compared to naïve
spinal cords (n = 3). Preliminary assessment of
electronmicrographs shows evidence of progressive loss of
myelin sheath at early post-injury time points, followed by
some restoration of myelin sheath in long-established
contusion injuries (Figure 5B), most likely mediated, at least
Figure 5. Gross pathology and changes in myelination of axons
following a contusion injury. A: Toluidine blue-stained semi-thin
transverse sections of the lesion epicenter at progressive time points
illustrating the large-scale morphological changes over time that occur
following contusion, highlighting mass tissue necrosis and cavity
formation with a residual rim of surviving axons. Scale bar, 500 μm.
B: Ultra-thin sections from the boxed area indicated in A, containing
ascending dorsal column axons that are assessed electrophysiologically.
Compared with axons in the uninjured spinal cord, contusion injury
causes loss of myelin with some remyelination taking place in chronic
post-injury stages.
74
Chondroitinase ABC promotes sprouting of intact and
injured spinal systems after spinal cord injury. J. Neurosci.
26:10856–10867.
Basso, D.M., Beattie, M.S., and Bresnehan, J.C. (1995). A
sensitive and reliable locomotor rating scale for open field
testing in rats. J. Neurotrauma. 12:1–20.
Basso, D.M., Beattie, M.S., Bresnahan, J.C. (1996) Graded
histological and locomotor outcomes after spinal cord
contusion using the NYU weight-drop device versus
transaction. Exp. Neurol. 139:244–256.
Bradbury, E.J., Moon, L.D., Popat, R.J., King, V.R.,
Bennet, G.S., Patel, P.N., Fawcett, J.W., and McMahon,
S.B. (2002) Chondroitinase ABC promotes functional
recovery after spinal cord injury. Nature. 416:636–640.
Bunge, R.P., Puckett, W.R., Becerra, J.L., Marcillo, A.,
Quencer, R.M. (1993) Observations on the pathology of
human spinal cord injury. A review and classification of 22
new cases with details from a case of chronic cord
compression with extensive focal demyelination. Adv.
Neurol. 59:75–89.
Bunge, R.P., Puckett, W.R., and Heister, E.D. (1997).
Observations on the pathology of several types of human
spinal cord injury, with emphasis on the astrocyte response
to penetrating injuries. Adv. Neurol. 72:305–315.
Carter, L.M., Starkley, M.L., Akrimi, S.F., Davies, M.,
McMahon, S.B., and Bradbury, E.J. (2008). The yellow
fluorescent protein (YFP-H) mouse reveals neuroprotection
as a novel mechanism underlying chondroitinase ABCmediated repair after spinal cord injury. J. Neurosci.
28:14107–14120.
Guest, J.D., Hiester, E.D., and Bunge, R.P. (2005).
Demyelination and Schwann cell responses adjacent to
injury epicenter cavities following chronic human spinal
cord injury. Exp. Neurol. 192:384–393.
Lasiene, J., Shupe, L., Perlmutter, S., and Horner, P. (2008).
No evidence for chronic demyelination in spared axons after
spinal cord injury in a mouse. J. Neurosci. 28:3887–3896.
Metz, G.A.S., Curt, A., Meent, H., Klusman, I., Schwab,
M.E., and Dietz, V. (2000). Validation of the weight-drop
contusion model in rats: a comparative study of human
spinal cord injury. J. Neurotrauma. 17:1–17.
Norenberg, M.D., Smith, J., Marcillo, A. (2004) The
pathology of human spinal cord injury: defining the
problems. J. Neurotrauma. 21:429–440.
Olson, L. (2002). Clearing a path for nerve growth. Nature.
416:589–590.
Onifer, S.M., Rabchevsky, A.G., and Scheff, S.W. (2007).
Rat models of traumatic spinal cord injury to assess motor
recovery. ILAR Journal. 48:385–395.
Scheff, S.W., Rabchevsky, A.G., Fugaccia, I., Main, J.A.,
and Lumpp, J.E. (2003). Experimental modeling of spinal
cord injury: characterization of a force-defined injury
device. J. Neurotrauma. 20:179–193.
Totoiu, M.O. and Kierstead, H.S. (2005). Spinal cord
injury is accompanied by chronic progressive
demyelination. J. Comp. Neurology. 486:373–383.
Quencer, R.M., Bunge, R.P. (1996) The injured spinal cord:
imaging, histopathologic clinical correlates, and basic
science approaches to enhancing neural function after spinal
cord injury. Spine. 21:2064–2066.
Figure 6. Myelin in the chronically contused spinal cord. A: 12 week
contused spinal cord epicentre immunolabeled using anti-NF200 (red)
and anti-protein zero (anti-P0; green) shows that many dorsal column
axons are associated with peripheral myelin rings. The boxed area is
shown at higher magnification. B: No association of ascending dorsal
column axons with peripheral myelin (P0), but central myelin
(proteolipid protein, PLP), occurs in the uninjured spinal cord. Note
the positive but negative immunoreactivity of the adjacent dorsal root
for P0 and PLP respectively.
CONCLUSION
Following initial absence of conduction across a contusion
injury site there is a slight increase in the percentage of
conducting axons as the injury progresses to chronic stages.
Accordingly, the behavioural assessments followed a similar
pattern showing an initial severe functional deficit with
some improvement over time. The partial, limited level of
recovery seen in both behavioural tests is a good level for
assessing whether treatment strategies can improve walking
ability to co-ordinated stepping movements and fine
locomotor control.
Anatomical assessment suggests the prevalence of
progressive demyelination and remyelination of axons at
progressive time points, likely to be mediated at least
partially by Schwann cells, which is consistent with findings
in human studies (Guest et al., 2005). Our data provides a
detailed, comprehensive characterisation of physiological
changes over time in a clinically relevant spinal injury
model, which will form the basis for testing the effects of
potential treatment strategies.
REFERENCES
Barritt, A.W., Davies, M., Marchand, F., Hartley, R., Grist,
J., Yip, P., McMahon, S.B., Bradbury, E.J. (2006)
75
treatment. We will continue collecting the nerve conduction
data for all of the time points from acute to chronic spinal
injury stages for the studies using chondroitinase therapy and
will compare the percentage of fibres conducting across the
injury between treated and non-treated animals in the
different injury stages. We will also continue our anatomical
assessments and monitor the animals on tests of walking
ability (BBB and horizontal ladder), ultimately determining
the effects of chondroitinase therapy.
Poster presented at the ISRT meeting in Ittingen, August
2010.
FUTURE PLANS
In the next 6–12 months of the project we will complete the
quantification of the electronmicroscopy data assessing
demyelination/remyelination by calculating g-rations (i.e.
ratio of fibre diameter to diameter of total axon diameter
inclusive myelin sheath) as well as the number of spared
dorsal column axons over a chronic time course post-injury.
Additionally, we will determine the total terminal field
innervations of the gracile nucleus by surviving ascending
dorsal column axons and also quantify white matter and grey
matter sparing in the spinal cord after a moderate 150 kD
contusion injury. We will then correlate the anatomical data
with the electrophysiological data and determine the
relationship between conduction failure and myelin levels.
We have confirmed Schwann cell-mediated remyelination
in contused spinal cord 12 weeks post-injury by means of
protein zero staining (Figure 5A) and we are currently
completing the immunohistochemistry assessing peripheral
and central myelin levels at the earlier post-injury time
points. We have also begun the studies using chondroitinase
MILESTONES AND OBJECTIVES
There have been no changes to aims and objectives as
outlined in the original application. We have completed the
time course study assessing changes in axonal conduction
over a chronic post-injury time period. We have completed
data collection for the ultrastructural analysis, which is
currently being processed for quantification of changes in
myelination, and will be correlated with the
electrophysiological data. We are expecting these results to
give us further insights concerning the importance of
demyelination/remyelination in the functional deficits after
traumatic spinal cord injury. We are also now in the process
of collecting data for the ChABC treatment study.
76
Rewiring the central nervous system following spinal cord
injury using neurotrophins and rehabilitative training
Karim Fouad & Wolfram Tetzlaff
[email protected]
INTRODUCTION
The lesioned corticospinal tract (CST) has been shown to
spontaneously send collaterals into the grey matter above a
spinal cord injury (SCI) site (Fouad et al., 2001, Bareyre et
al., 2004). This adaptive response to an injury potentially
results in the creation of new spinal circuits when lesioned
corticospinal axons become successfully rewired via spared
neurons. When the CST is ablated at the cervical level of
the spinal cord, potential targets for rewiring and eventually
promoting recovery of arm and hand motor control are
spared motor tracts such as the rubrospinal tract (RST) or
the reticulospinal tract (RtST), apart from interneurons. In
this project, we will focus on the spared RtST to mediate a
detour for axotomized corticospinal fibers. Although the
RtST is mainly responsible for the control of axial muscles
and locomotion, recent reports suggest that it has the
potential to take over motor control of the arm and hand
after CST injury (Ballermann and Fouad, 2006, Pettersson
et al., 2007). Our aim for this project is to promote rewiring
of the injured CST via neurons within the reticular
formation and thereby enhance motor performance of the
affected arm and paw. In a set of three individual
experiments, we will investigate the effectiveness of different
drug treatments and rehabilitative training to successfully
rewire the injured CST via the spared RtST in a rat model
of SCI. In the first experiment, BDNF is delivered to the
motor cortex to increase corticospinal sprouting in order to
enhance the probability of new functional connections at
the level of the brainstem (comparable to our earlier findings
i.e. Hiebert et al., 2002, Vavrek et al., 2006). For the second
experiment we planned that, NT-3 and/or BDNF are
expressed in the reticular formation with the help of viral
vectors. NT-3 is expected to act as a potent attractant to
guide axonal sprouts towards the reticular formation (Zhou
et al., 2003). BDNF is delivered to increase sprouting of the
RtST onto denervated targets downstream of the SCI (see
Fig. 1). In a third experiment, the most successful drug
treatment will be combined with different rehabilitative
training regimes in a forelimb task to maximize meaningful
plasticity and recovery (as described in Girgis et al., 2007).
3.
4.
5.
6.
METHODS
Experiment 1 was reported last year. Here we present only
the methods for Experiment 2:
1. Spinal cord injury: On the side of preferred paw use,
rats received a unilateral cervical dorsal quadrant lesion,
which ablates the dorsal portion of the CST and the
majority of the RST, but leaves most of the RtST intact.
2. Drug delivery: Adeno-associated viral vectors expressing
either BDNF, NT-3 or GFP were injected into the
motor cortex contralateral to spinal injury and into the
reticular formation ipsilateral to the spinal injury,
77
respectively. We used glass micro-electrode connected
to a picospritzer for injection.
Assessment of motor performance: Animals were trained
before lesion and tested thereafter in the single pellet
reaching task, which specifically allows assessment of
fine motor performance of the lower arm and paw. A
qualitative scoring regime will obtain information about
how well the shoulder, elbow, wrist and digit movements
match pre-lesion performance in grasping. Additionally,
error rate for the affected forelimb for crossing a
horizontal ladder was assessed weekly following injury.
Spontaneous use of the injured forelimb while exploring
the walls of a plexiglass cylinder was evaluated before
lesion and 6 weeks after lesion.
Tracer injections: Two different fluorescent tracers were
injected into the motor cortex and into the reticular
formation at the end of the experiment. Animals
survived for two weeks after injection to allow the
tracers to travel down into the spinal cord.
Perfusion and tissue collection: Animals were perfused
with saline followed by 4% paraformaldehyde and
brains as well as spinal cords were harvested, stored in
30% sucrose solution and frozen. Brains and
brainstems were cross sectioned in a cryostat at 40 μm
and 30 μm, respectively. Consecutive pieces of spinal
cord at and around the lesion site were cross sectioned
at 25 μm.
Histology:
a) Lesion size and location was analyzed to allow for
exclusion of animals with unacceptable lesion
extent. Due to sparing of the main CST, we
excluded 6 animals from further analysis.
b) The total number of traced CST fibers is being
assessed in the pyramids as well as in the spinal cord
immediately rostral to the lesion for normalization
purposes. Also, the amount of traced RtST fibers
will be calculated in the spinal cord.
c) Corticospinal fiber sprouting at the level of the
brainstem is currently examined by counting fibers
crossing the midline towards the contralateral
reticular formation. Also, the amount of sprouting
of lesioned CST fibers into the grey matter above
the lesion and sprouting of spared corticospinal
fibers (ventral portion of CST) below the lesion will
be quantified to detect any effects of treatment.
d) Reticulospinal fiber sprouting will be analyzed by
counting the number of traced RtST fibers that
cross from the white matter into the grey matter in
the spinal cord below the lesion. Densitometry of
traced fibers in different grey matter laminae will be
considered as well to take branching and
distribution of fibers into account.
has been confirmed (examples given in Fig. 2) and so far
only the number of the traced fibers has already been
analyzed. Therefore the current result section is limited to
the behavioral outcome (see Fig. 3). Over 6 weeks of
recovery the success rate in single pellet reaching was
quantified and the best result of the last 3 sessions for each
animal was grouped according to their treatment. One
group received GFP expressing vectors to the cortex and
reticular formation (white columns), one group received
GFP expressing vectors into the cortex (gray) and NT-3 into
the reticular formation, and one group received BDNF into
the cortex and NT-3 into the reticular formation (black).
Although the rats that received BDNF and NT-3 show a
lower success rate than the other two groups, this difference
was not significant (Fig. 2A). However, when we analyzed
the rate of errors when the rats were crossing a horizontal
ladder, the animals receiving NT-3 expressing vectors
showed insignificantly more errors than GFP only controls,
and those with both NT-3 and BDNF performed
significantly worse (B). This stands in direct contrast to the
results of the Cylinder test, where both neurotrophintreated groups performed better than GFP only controls.
Rats that received both vectors performed basically like
unlesioned animals in this test, which was significantly
better than the control/GFP treated rats.
RESULTS
An outstanding part of the analysis from Experiment 1 (see
last year’s report) was the quantification of sprouting of
lesioned CST fibers directly above the injury, and that of
spared fibers below. We can now report that against our
hypothesis and contradictory to former experiments (with
different lesions), no changes following BDNF application
were found. Negative results are always hard to interpret as
they can be based on various reasons. As we found clear
effects of BDNF in combination with NT-3 in the same
lesion model when given via viral vectors (Experiment 2)
we will include a group in the next experiment receiving
only BDNF expressing vectors to the cortex and no NT-3,
The goal of the second experiment was to
pharmacologically enhance sprouting of lesioned CST fibers
and attract their growth towards spared descending neurons
originating from the reticular formation (Fig. 1) using
targeted application of viral vectors to force the expression of
neurotrophic factors. Due to recent findings in an unrelated
set of experiments we slightly modified the experimental
design. We found indications that over-expression of BDNF
in the spinal cord possibly contributes to spasticity, likely by
the known neuro-excitatory effect of BDNF. Therefore, we
intended to limit the overall level of BDNF especially in the
brainstem, and decided to apply BDNF expressing vectors
only to the cortex. NT-3 expressing vectors however were
injected into the reticular formation as suggested in our
original application to attract CST sprouts.
Figure 2. Histology: So far only tracing of the CST (A) and the RtST
(B) has been confirmed and quantified. A: a cross section of the
brainstem shows the traced pyramidal tract and fibers sprouting across
the midline (black line) towards the contra lateral reticular formation.
The numbers of collaterals is currently being analyzed. The schematic
illustrates the location of the reticular formation (grey) in relation to
traced pyramid (red).B: tracing of the reticulo spinal tract shown on
a cross section. The schematic to the right indicates the viewing angle.
CONCLUSION
Bearing in mind that the histological analysis is not yet
completed, conclusions as to underlying mechanisms for the
observed behavioral results are extremely challenging and
speculative at this point. When considering the role of the
motor tracts that have been manipulated by our treatment
(i.e., CST which is involved in digit function, and RtST
which is involved in axial and proximal muscle control) and
the functional tests (horizontal ladder and grasping requiring
Figure 1. Lesion paradigm: Lesions of the cervical spinal cord targeted
the dorsolateral quadrant (A) ablating the main portion of the
corticospinal tract (CST, red) and only a small portion of the reticular
spinal tract (RtST, green). The main hypothesis of this study is to
promote the use of the RtST as detour for lesioned CST fibers (B).
The behavioral part of the experiment has been finalized
and the histological analysis is currently under way. Tracing
success of both, the CST and the reticulospinal tract (RtST)
78
REFERENCES
Ballermann, M., Fouad, K. (2006) Spontaneous locomotor
recovery in spinal cord injured rats is accompanied by
anatomical plasticity of reticulospinal fibers. Eur. J. Neurosci.
23:1988–1996.
Bareyre, F.M., Kerschensteiner, M., Raineteau, O.,
Mettenleiter, T.C., Weinmann, O., Schwab, M.E. (2004)
The injured spinal cord spontaneously forms a new
intraspinal circuit in adult rats. Nat. Neurosci. 7:269–277.
Fouad, K., Pedersen, V., Schwab, M.E., Brosamle, C.
(2001) Cervical sprouting of corticospinal fibers after
thoracic spinal cord injury accompanies shifts in evoked
motor responses. Curr. Biol. 11:1766–1770.
Girgis, J., Merrett, D., Kirkland, S., Metz, G.A., Verge, V.,
Fouad, K. (2007) Reaching training in rats with spinal cord
injury promotes plasticity and task specific recovery. Brain.
130:2993–3003.
Hiebert, G.W., Khodarahmi, K., McGraw, J., Steeves, J.D.,
Tetzlaff, W. (2002) Brain-derived neurotrophic factor
applied to the motor cortex promotes sprouting of
corticospinal fibers but not regeneration into a peripheral
nerve transplant. J. Neurosci. Res. 69:160–168.
Pettersson, L.G., Alstermark, B., Blagovechtchenski, E., Isa,
T,, Sasaski, S. (2007) Skilled digit movements in feline and
primate--recovery after selective spinal cord lesions. Acta.
Physiol. (Oxf ). 189:141–154.
Vavrek, R., Girgis, J., Tetzlaff, W., Hiebert, G.W., Fouad, K.
(2006) BDNF promotes connections of corticospinal
neurons onto spared descending interneurons in spinal cord
injured rats. Brain. 129:1534–1545.
Zhou, L., Baumgartner, B.J., Hill-Felberg, S.J., McGowen,
L.R., Shine, H.D. (2003) Neurotrophin-3 expressed in situ
induces axonal plasticity in the adult injured spinal cord.
J. Neurosci. 23:1424–1431.
digit and wrist function cylinder test not requiring digit
function) it could be suggested that our treatment promoted
the original function of the RtST but did not help to extend
its function to digit control. On the other hand the
neurotrophin treatment (presumably promoting increased
CST projection to the reticular formation) might have
negatively affected the original function and natural
occurring plasticity of spared CST fibers. This might explain
why we see deficits in tasks involving digit function. These
results raise various important questions that will be
addressed in additional video analysis of the grasping process
and in the next experiment. For example, did sprouting of
CST fibers onto the reticular formation occur and result in
enhanced recovery of function controlled by the RtST (i.e,
cylinder task). Can the manipulation of naturally occurring
plasticity/sprouting be detrimental? Is there a treatment
induced enhanced deterioration in digit function? And lastly,
can neurotrophin treatment combined with regular
activity/rehabilitative training translate plasticity into
functional meaningful connections?
FUTURE PLANS
We have ambitious plans for the next year, which can be
divided into 2 main goals:
1) We will finalize the histology for Experiment 2. The
focus lies on identifying whether corticospinal tract
sprouting towards the reticular formation was actually
increased and whether anatomical changes in the
reticulospinal tract of animals that received NT-3 can
be detected. These results will then be correlated to the
behavioral changes, which will provide important
information on the possible mechanism for the
treatment effect on the different functional tests. We
are also extending our behavioral analysis using high
speed video recordings taken during the grasping
sessions. This will allow us to detect positive and
negative effects on different components of the grasping
movement, which can then be related to the different
motor systems (i.e., CST versus RtST). We would
specifically like to find out whether the untreated rats
showed better digit function than treated ones.
2) Originally we planned to utilize the most promising
pharmacological treatment combination and add task
specific rehabilitative training. However, not surprising
considering the complexity of our experiments, the
results were not that clear. We found positive and
Figure 3. Functional outcome: Following 6 weeks of post-injury
recovery, the GFP treated rats performed only insignificantly better in
the reaching task then the treated animals (A). The performance in the
group receiving NT-3 to the reticular formation and BDNF to the
cortex was significantly worse than in the other groups (B). However,
in the cylinder test this group performed significantly better than the
controls (GFP treated). Data are shown as group mean ± SEM. The
asterisks indicates P>0.05.
79
RtST controlled shoulder and axial muscles following lesion
and treatment. Lastly, we will add a new electrophysiological
approach prior to perfusion of the animals. If our
speculation made in the conclusion section is correct, we
should be able to see a change in CST and RtST innervation
patterns. This can be approached by recording
electromyographic activity in proximal and distal muscles
as a response to stimulation of the cortex or the reticular
formation. My laboratory is experienced in all these
techniques and together with careful functional analysis this
experimental design will allow a better interpretation of our
current results.
negative effects on motor recovery, which raises
important questions that have to be resolved in our quest
to find a functional meaningful treatment. Thus, in the
next experiment we plan to use all 4 pharmacological
combinations (GFP to cortex and brainstem; GFP to
cortex and NT-3 brainstem; BDNF to cortex and GFP
to brainstem, and BDNF to cortex and NT-3 to
brainstem) and half of the animals in each group will
undergo rehabilitative training. This is a huge enterprise
and in order to reach sufficient animal numbers, we will
perform this project in two sets of identical experiments.
Each one will include all treatment groups in low
numbers. We are confident that no extra funds will
be required.
To answer some of the questions raised by our recent results
we will add another behavioral outcome measure (i.e.
grooming abilities) to evaluate the function of the mainly
The overall objectives remain the same as originally described,
however due to our recent results we adjusted the readout of
our experiments. This is described under “Future Plans”.
80
Developing mTOR-based strategies to promote axon
regeneration and functional recovery after spinal cord
injury
Zhigang He
Children’s Hospital Boston, USA
[email protected]
INTRODUCTION
Axons do not regenerate after injury in the adult
mammalian CNS, a phenomenon attributed to two
properties of the adult CNS, the inhibitory extrinsic
environment and a diminished intrinsic regenerative
capacity of mature CNS neurons. Neutralization of the
extracellular molecules identified as axon regrowth
inhibitors only allows a limited degree of axon regeneration
in vivo, pointing to the importance of boosting the intrinsic
ability of mature neurons for axon regeneration. Our recent
studies based on optic nerve injury models implicated the
mTOR activity as a critical determinant of intrinsic ability
of axon regeneration in retinal ganglion neurons (RGCs).
The objective of this proposed study is to extend these
findings and test whether manipulating mTOR activity
could promote axon regeneration and functional recovery
after spinal cord injury. Our specific aims include:
Aim 1. Examine the effect of genetic activation of the mTOR
pathway on the regrowth of lesioned descending axons.
Aim 2. Examine the effects of chemical agents that boost
the mTOR activity on promoting axon regeneration and
functional recovery.
the basilar artery. The wound was closed in layers with
6.0 sutures. The mice were placed on soft bedding on a
warming For pyramidotomy, animals were anesthetized with
ketamine/xylazine. The procedure is similar to what
described previously. Briefly, an incision was made at the left
side of the trachea. Blunt dissection was performed to expose
the skull base, and a craniotomy in the occipital bone
allowed for access to the medullary pyramids. The left or
right pyramid was cut with a fine scalpel medially up to the
basilar artery. The wound was closed in layers with
6.0 sutures. The mice were placed on soft bedding on a
warming blanket held at 37°C till fully awake. Two weeks
later the intact CST was traced with BDA.
The procedure for T8 dorsal hemisection is similar to
what was described previously. Briefly, a midline incision was
made over the thoracic vertebrae. A T8 laminectomy was
performed. To produce a dorsal hemisection injury, the
dorsal spinal cord was first cut with a pair of microscissors to
the depth of 0.8 mm and then a fine microknife was drawn
bilaterally across the dorsal aspect of the spinal cord. The
muscle layers were sutured and the skin was secured with
wound clips. The mice were placed on soft bedding on a
warming blanket held at 37°C till fully awake. Urine was
expressed by manual abdominal pressure twice daily till mice
regained reflex bladder function. Six weeks post-injury, BDA
was injected into the sensorimotor cortex to tract the CST.
METHODS
Animals and Surgeries
All experimental procedures were performed in compliance
with animal protocols approved by the IACUC at Children’s
Hospital, Boston. AAV preparation was described in Park
et al.18.
The procedure of T8 spinal cord crush is similar to what
was described previously with modifications. Briefly, a
midline incision was made over the thoracic vertebrae. A
T8 laminectomy was performed. The exposed cord was
crushed for 2 seconds with modified no. 5 jeweler’s forceps,
keeping the dura intact. The muscle layers were sutured and
the skin was secured with wound clips. The mice were
placed on soft bedding on a warming blanket held at 37°C
till fully awake. Urine was expressed by manual abdominal
pressure twice daily for the entire duration of the
experiment. At 10 weeks post-injury the CST was traced
with BDA.
For AAV injection, neonatal PTENf/f were
cryoanesthetized and injected with 2 μl of either AAV-Cre
or AAV-GFP into the right sensorimotor cortex by using a
nanolitter injector attached with a fine glass pipette. Mice
were then placed on a warming pad and returned to the
mothers after regaining normal colour and activity. For the
mice at the age of 4 weeks, a total l.5 μl of AAV-Cre or AAVGFP was injected into the hindlimb sensorimotor cortex at
three sites (coordinates from bregma in mm: AP/ML/DV
0.0/1.5/0.5, −0.5/1.5/0.5, −1.0/1.5/0.5). The mice were
placed on a warming blanket held at 37°C till fully awake
and received a spinal cord lesion 4 weeks later.
Immunofluorescence staining of the spinal cord,
cortex and medulla
Immunostaining was performed following standard
protocols. All antibodies were diluted in a solution
consisting of 10% normal goat serum (NGS) and 1% Triton
X-100 in phosphate-buffered saline (PBS). Antibodies used
were rabbit anti-p-S6 (Ser235/236) (1:200, Cell Signalling
Technology) and rabbit anti-GFAP (1:1,000, DAKO).
Sections were incubated with primary antibodies overnight
at 4°C and washed three times for 10 minutes with PBS.
For pyramidotomy, animals were anesthetized with
ketamine/xylazine. The procedure is similar to what
described previously19,20. Briefly, an incision was made at
the left side of the trachea. Blunt dissection was performed
to expose the skull base, and a craniotomy in the occipital
bone allowed for access to the medullary pyramids. The left
or right pyramid was cut with a fine scalpel medially up to
81
Secondary antibodies (goat-anti-rabbit Alexa488) were then
applied and incubated for 1 hour at room temperature. To
detect BDA labeled fibers, BDA staining was performed by
incubating the sections in PBS containing streptavidinHRP. The remaining staining procedure was performed
according to the protocol provided by TSA™ Cyanine 3
system (Perkin Elmer).
The number of fibers caudal to the lesion was analyzed
with a fluorescence microscope. The number of
intersections of BDA-labeled fibers with a dorsal-ventral line
positioned at a defined distance caudal from the lesion
center was counted under a 40× objective. Every other
section of the whole spinal cord was stained. Fibers were
counted on 3–4 sections with the main dorsal CST and
1–3 lateral sections with collaterals in the gray matter. The
number of counted fibers was normalized by the number of
labeled CST axons in the medulla and divided by the
number of evaluated sections. This resulted in the number
of CST fibers per labeled CST axons per section at different
distances (fiber number index).
Axonal counting and quantifications
For quantifying total labeled CST axon, BDA labeled CST
fibers were counted at the level of medulla oblongata 1 mm
proximal to the pyramidal decussation. Axons were counted
in 4 rectangular areas (9506 μm2) per section on two
adjacent sections. The number of labeled axons was
calculated by multiplying with the total area.
For the animals with T8 crush, the number of fibers
caudal to the lesion was analyzed with a fluorescence
microscope. The number of intersections of BDA-labeled
fibers with a dorsal-ventral line positioned at a defined
distance caudal from the lesion center was counted under a
40× objective. Every other section of the whole spinal cord
was stained. Fibers were counted on 3 sections with the
main dorsal CST. The number of counted fibers was
normalized by the number of labeled CST axons in the
medulla and divided by the number of evaluated sections.
This resulted in the number of CST fibers per labeled
CST axon per section at different distances (fiber
number index).
For the groups of pyramidotomy, digital images of C7
spinal cord transverse sections were collected by using a
Nikon fluorescence microscope under a 4× objective.
Densitometry measurement on each side of the gray matter
was taken by using Metamorph software, after being subthresholded to the background and normalized by area. The
outcome measure of the sprouting density index was the
ratio of contralateral and ipsilateral counts. At least 3 sections
were measured for each mouse.
To quantify the number of sprouting axons, a horizontal
line was firstly drawn through the central canal and across the
lateral rim of the gray matter. Three vertical lines (Mid, Z1,
and Z2) were then drawn to divide the horizontal line into
three equal parts, starting from the central canal to the lateral
rim. While Mid denotes midline crossing fibers, Z1 and Z2
are for sprouting fibers at different distance from the midline.
Only fibers crossing the three lines were counted on each
section. The results were presented after normalization with
the number of counted CST fibers at the medulla level. At
least 3 sections were counted for each mouse.
Statistical analysis
Two-tailed Student’s t-test was used for the single
comparison between two groups. The rest of the data were
analyzed using one-way or two-way ANOVA depending on
the appropriate design. Post hoc comparisons were carried
out only when a main effect showed statistical significance.
P-value of multiple comparisons was adjusted by using
Bonferroni’s correction. All analyses were conducted
through StatView. Data are presented as means + SEM and
the asterisks indicate statistical significance under an
appropriate test.
In the animals of T8 dorsal hemisection, digital images
were taken at the CST end by using a confocal microscope
(Zeiss, LSM510) under a 63× objective to quantify the
number of retraction bulbs,. The number of bulbs was
counted within a square by the area of 21389 μm2 and
normalized by the number of BDA labeled CST at the
medulla. At least 3 sections with the main CST per animal
were examined. The results were presented as the number of
retraction bulbs per 0.1 mm2 per labeled CST.
RESULTS
Correlation between mTOR down-regulation and
repression of CST sprouting
We first investigated whether mTOR activity regulates the
sprouting responses of CST axons after unilateral
pyramidotomy 19,20. In this model, the CST was severed
unilaterally at the left medullary pyramid above the
pyramidal decussation. The anterograde tracer biotinylated
dextran amine (BDA) was injected to the right sensorimotor
cortex to label uninjured CST axons. In intact mice, most
labeled axons were detected on the left side of the spinal
cord, with few axons appearing in the right side. Thus
increased numbers of labeled axons in the right side of the
spinal cord following a pyramidotomy would represent
trans-midline sprouting of intact CST axons into the
denervated side.
The density of sprouting fibers of the main CST rostral
to the lesion site was analyzed quantitatively using digital
images taken with a Nikon fluorescence microscope under a
4× objective. A series of rectangular segments by the width of
100 μm and the height covering the dorsal-ventral aspect of
the cord were superimposed onto the sagittal sections, starting
from 1.5 mm rostral up to the lesion center. After subtracting
the background (the most caudal part of the section), the
pixel value of each segment was normalized by dividing with
the first segment (1.5 mm rostral). The results were presented
as a ratio at different distances (fiber density index). Every
other section of the whole spinal cord was stained.
3–4 sections with main CST per animal were quantified.
By this procedure, we found a sharp age-dependent
decline in trans-midline sprouting responses: while
82
PTEN deletion prevents mTOR down-regulation and
increases CST sprouting
To assess whether PTEN deletion elevates neuronal mTOR
activity, we injected AAV-Cre into the sensorimotor cortex
of PTENf/f mice at P1 and examined the p-S6 signal in the
adult. Indeed, compared to AAV-GFP injected PTENf/f
controls, immunostaining for p-S6 was significantly higher
in adult PTEN-deleted cortical neurons. This postnatal
AAV-Cre-mediated PTEN deletion does not appear to alter
the projections of CST axons in the spinal cord, as the
numbers and termination patterns of labeled axons in the
pyramids and different levels of spinal cord are not
significantly different in the PTENf/f mice injected with
either AAV-Cre or AAV-GFP. This result is consistent with
the observation that the development of CST projections is
largely complete in the early postnatal stage24. We next
performed pyramidotomy in adult PTENf/f mice which
had a neonatal injection of AAV-Cre or AAV-GFP.
Compared to the limited sprouting in controls, PTEN
deletion elicited extensive trans-midline sprouting of adult
CST axons from the intact side into the denervated side.
Thus, PTEN deletion is sufficient for maintaining high
mTOR activity characteristic of young neurons in adult
cortical neurons, and for these neurons to launch a robust
sprouting response after injury.
allowing the mice to survive for two additional weeks.
Transverse sections 5 mm caudal to the lesion sites were first
examined. The presence of any labeled axons in the dorsal
main tract and the dorsolateral tract caudal to the injury
was taken as evidence of incomplete lesions and these
animals were excluded from further analysis.
In 9 control mice, not a single CST axon was seen
extending directly through the lesion site. In two of these
control mice, we found a few axons extending to the distal
spinal cord via the ventral column, consistent with previous
observations. Instead, characteristic dieback of CST axons
from the injury site was observed, and individual axons
displayed numerous retraction bulbs. By contrast, when
PTEN was deleted, the main CST bundle extended to the
very edge of the lesion margin and few retraction bulbs were
associated with the labeled CST axons eight weeks after
injury. This phenotype could be due to either a lack of axon
dieback of PTEN-deleted neurons, or resumed axon
regrowth after the initial injury-induced retraction. To
distinguish between these, we examined CST axons at
10 days post-injury. Apparent dieback and large numbers
of retraction bulbs were observed at this early time point in
both control and PTEN-deleted axons. Thus, instead of
affecting the acute post-injury axonal degeneration, PTENdeletion likely reversed the normal abortive regenerative
attempts typical for injured adult CNS axons31 and
enhanced their regrowth.
CST regeneration after T8 dorsal hemisection after
neonatal PTEN deletion
Although sprouting of uninjured neurons might partially
compensate for lost function, inducing severed axons to
regenerate beyond the lesion site and to re-connect the
axonal pathways would be needed for functional recovery
in more severe injuries. We thus asked whether PTEN
deletion would sustain a high level of mTOR activity in
injured adult corticospinal neurons and elicit robust axon
regeneration. By immunohistochemistry we found that
axotomy diminished p-S6 levels in adult corticospinal
neurons identified by retrograde labelling. With the stepwise
down-regulation of mTOR activity, firstly an age-dependent
decline and secondly an injury-triggered further reduction,
lesioned adult corticospinal neurons exhibited low
p-S6 signal, suggesting a major reduction of mTOR
activity. Importantly, AAV-Cre-mediated PTEN deletion
not only increased basal p-S6 levels but also efficiently
attenuated the injury-induced loss of mTOR activity in
corticospinal neurons.
More strikingly, significant numbers of labeled axons
regenerated past the lesion site in all 11 PTEN-deleted mice.
Examination of the entire collection of serial sections from
these animals revealed two distinct routes by which labeled
CST axons reached the caudal spinal cord: either directly
growing through the lesion or circumventing the injury site
via the spared ventral white matter. We estimated that
approximately two thirds of labeled axons seen in the distal
spinal cord grew through the lesion site and the rest
projected along the ventral white matter. Critically, CST
axons that regenerated past the lesion were not restricted to
one side of the distal spinal cord and instead projected
bilaterally. This is important because normal CST
projections are largely unilateral. The presence of significant
numbers of CST axons on the side contralateral to the main
tract is strong evidence of regenerative growth and cannot be
accounted for by spared axons.
Importantly, similar results were obtained in an
independent set of experiments where the lesions were
performed in a double-blinded manner by an independent
surgeon, who had carried out extensive analyses of possible
CST regeneration in Nogo knockout mice. In these
experiments with control and PTEN deleted mice, the
genotypes (AAV-Cre vs control) could be predicted by a
blinded observer with great accuracy (∼95%) based on BDA
labeling (regenerator vs non-regenerator), further supporting
the highly robust effect of PTEN deletion. Thus, the effect
of PTEN deletion was consistent and robust enough to
overcome the inter-investigator surgical variability typical for
experimental spinal cord injury models.
Having established an experimental paradigm to
maintain a relatively high level of mTOR activity in adult
corticospinal neurons even after injury, we set out to
determine whether PTEN deletion would enable
regenerative growth of adult CST axons in two different
spinal cord injury paradigms: a dorsal hemisection, which
transects all traced CST axons but spares the ventral spinal
cord25–28, and a complete crush model that transects all
passing axons and leaves no bridge of uninjured tissue29,30.
Dorsal hemisection injuries were created at T8, and the
CST from one hemisphere was traced 6 weeks post-injury
by injecting BDA into the right sensorimotor cortex,
83
Notably, the majority of PTEN-deleted regenerating
axons projecting into the lesion site were associated with
GFAP-positive tissue matrix. These GFAP-positive matrixes
often appeared in the superficial and medial locations of the
spinal cord which would be highly unlikely, if not
impossible, to be spared in a dorsal hemisection. GFAPpositive bridges were rarely seen at a shorter timeframe after
injury, suggesting that these matrixes develop over time
following dorsal hemisection, possibly as a consequence of
an interaction between GFAP-negative cells and GFAPpositive cells at the injury site. However, the identity of these
GFAP-positive cells/or matrixes remains unknown.
much cortical area, likely due to less efficient diffusion of
injected viral particles in the more mature cortex.
We then followed introduction of AAVs into the
sensorimotor cortex of PTENf/f mice at 4 weeks with a T8
complete spinal cord crush injury at the age of 8 weeks, and
analyzed CST regeneration after 3 months post-injury. We
still found significant CST regeneration in the spinal cord
caudal to the lesion sites. Thus, PTEN deletion at both
neonatal and young adult stages promoted robust CST axon
regeneration past a complete spinal cord crush lesion.
Regenerating CST axons re-form synaptic structures
We next investigated whether regenerating CST axons from
PTEN-deleted corticospinal neurons are able to form
synapses. For this, we analyzed the samples taken from the
gray matter of the spinal cord caudal to the lesion site in
PTENf/f mice with neonatal AAV-Cre injection and T8
crush injury at the age of 2 months. First, we assessed
whether BDA-labeled regenerating CST axons are costained with vGlut1, a presynaptic marker for excitatory
synapses. Some BDA-labeled bouton-like structures exhibit
vGlut1-positive patches at the tip of BDA-labeled CST
collaterals and along the axonal length and, suggesting the
accumulation of the molecular machinery characteristic of
a presynaptic terminal. We quantified these BDA and
vGlut1-costained bouton-like structures in similar spinal
cord locations in wild type intact mice and PTEN-deleted
mice with crush injury. The incidence of vGlut1-positive
patches in regenerating CST axons is approximately two
thirds of that of CST axons in un-injured mice.
CST regeneration after T8 complete crush injury after
neonatal PTEN deletion
A complete spinal cord crush destroys all neural tissues at
the injury site and is considered an extraordinary barrier for
regeneration. In this model, the dura mater is not damaged
so that the two ends of the spinal cord do not pull apart. In
mice, the lesion site is filled with a connective tissue matrix.
Initially, the matrix was largely GFAP negative, but evolved
so that GFAP positive fingers extended into the connective
tissue matrix at later post-injury stages. At 12 weeks postinjury, no CST axons extended into or beyond the lesion
site in any of the 8 control mice. In contrast, in all
8 PTENf/f mice with AAV-Cre, numerous axons extended
into the lesion sites and beyond the lesion for up to 3 mm.
Many regenerating axons follow ectopic trajectories. For
example, instead of projecting in one side of the spinal cord
like normal CST axons, regenerating axons extended
bilaterally with many of them showing tortuous projection
patterns, again, strongly against the possibility of being
spared axons.
We further assessed whether BDA-labeled regenerated
axons form synapses at the ultrastructural level. Sections
from the spinal cords from PTEN-deleted mice with crush
injuries and BDA injections were stained for BDA and
further processed for electron microscopic analysis. We
found many structures with characteristics of synapses,
based on the presence of a contact zone with presynaptic
vesicles (partially obscured by the reaction products in the
labeled terminal) and a post-synaptic density (psd). These
results establish that regenerating CST axons from PTENdeleted corticospinal neurons appear to possess the ability to
reform synapses in caudal segments. Whether these synapses
are functional and the identity of the neurons contacted by
the regenerated axons remain to be established.
The results described above were obtained from the
animals that had AAV-Cre injection at a neonatal age and
spinal cord crush injury at 2 months of age. A question is
whether such increased CST regeneration ability persists in
corticospinal neurons in older mice. To assess this, we
performed another set of experiments in which the same T8
spinal cord crush was performed in 5 month-old PTENf/f
mice with neonatal cortical injection of AAV-Cre or control.
We found significant CST regeneration at 3 months postinjury in these mice, to an extent similar to what seen in the
mice with injury at the age of 2 months.
Deleting PTEN after the neonatal period can also
induce CST regeneration
While no significant changes of CST numbers and
projections were found in the spinal cord in PTENf/f mice
with neonatal AAV-Cre cortical injections, it is still possible
that the up-regulation of mTOR activity associated with
PTEN deletion at this early stage could block developmental
events that turn off axon regeneration ability. To assess this,
we first optimized a stereotaxic injection method to
introduce AAVs to the sensorimotor cortex of mice at the
age of 4 weeks. AAV-Cre injections resulted in efficient Credependent PLAP expression in reporter mice. We estimated
that in comparison to that with neonatal AAV-Cre injection,
Cre-Dependent PLAP expression affected about 25% as
CONCLUSION
Together, our results indicate that PTEN deletion enables
injured adult corticospinal neurons to mount a robust
regenerative response never seen before in the mammalian
spinal cord. Both compensatory sprouting of intact CST
axons and regenerative growth of injured CST axons are
dramatically increased by PTEN deletion, suggesting that
these two forms of regrowth share similarunderlying
mechanisms. PTEN inactivation is known to activates
different downstream pathways such as Akt and mTOR
signalling and inhibit other signalling molecules such as
GSK-3 and PIP3. In cortical neurons, mTOR activity
undergoes a development-dependent down-regulation and
84
axotomy further diminishes mTOR activity. On the other
side, PTEN deletion in these neurons could increase mTOR
activity and promote their regrowth ability. Together with
our previous findings in retinal ganglion neurons, these
results support a critical role of mTOR activity in
determining the regrowth ability in CNS neurons. Because
mTOR is a central regulator of cap-dependent protein
translation, it is likely that neuronal growth competence is
critically dependent on the capability of new protein
synthesis, which provides building blocks for axonal
regrowth. Other PTEN deletion-induced effects, such as
increased axonal transport as the result of inactivation of
GSK-3, might also be involved.
Sun, F. and He, Z. Intrinsic brakes for axon regeneration.
Curr. Opin. Neurobiol. 20, 510–518, 2010. (review)
Liu, K., Lu, Y., Lee, J.K., Samara, R., Willenberg, R., SearsKraxberger, Tedeschi, A., Park, K.K., Connolly, L., Steward,
O., Zheng, B., and He, Z. PTEN deletion enhances the
regenerative ability of adult corticospinal neurons. Nature
Neurosci. 13, 1075–1081, 2010.
Tedeschi, A. and He, Z. Axon regeneration: electrical
silencing is a condition for regrowth. Current Biol. 20,
R713–714. 2010. (review).
He, Z. Intrinsic controls of axon regeneration. Ann. Rev.
Neurosci. (in press).
Our results also indicate that regenerating CST axons
from PTEN-deleted corticospinal neurons are able to
reform synapses in the spinal cord caudal to the lesion site.
As CST axons that regenerate after PTEN deletion extend
bilaterally in contrast to normal CST axons that extend
unilaterally, it is unknown to what extents these regenerating
axons could make synaptic connections with their original
targets. Interestingly, at least in some species such as the
larval lamprey and goldfish, axons that regenerate past a
spinal cord lesion fail to reach their original targets, yet make
synapses that allow functional recovery. Thus, our future
studies will be aimed to determine whether these
regenerating axons and synapses could mediate functional
recovery after spinal cord injury.
FUTURE PLANS
We will continue this line to test whether regenerating CST
axons could mediate function recovery after C5 contusion
lesion. We will also test the effects of PTEN inhibitors in
promoting axon regeneration and functional recovery after
injury.
We have completed the experiments of testing the effects of
PTEN dletion on CST regeneration. We will focus on
functional recovery aspects in next funding years. We have
been delayed in testing the PTEN inhibitors due to the
difficulty of delivering the compounds. But we will continue
to do this in next years.
85
Lyn B. Jakeman & D. Michele Basso
The Ohio State University, USA
[email protected]
INTRODUCTION
Spinal cord injury (SCI) causes permanent loss of sensation
and motor function below the level of initial damage. In
addition, depending on the level and type of injury, SCI can
lead to the concomitant dysregulation of autonomic and
sexual function and systemic complications including bone
and muscle loss, impairment of wound healing, and chronic
pain (Lin, 2003). Over the past several decades, animal and
clinical studies have been done with the goal of improving
function and reducing these complications after SCI. For
example, a number of pre-clinical interventions are directed
at increasing axonal regeneration, sprouting, or
neuroprotection at the injury site (reviewed in Fitch and
Silver (2007), Kwon et al., (2010b, 2010c) and Tetzlaff et al.,
(2010)). While many of these approaches have proven very
promising in the animal models, most of them have not been
ready to translate to the clinical setting despite increasing
emphasis on efforts to do so (Lammertse et al., 2007; Kwon
et al., 2010a). In contrast, rehabilitation strategies are moving
forward at a rapid rate in terms of both basic research and
clinical application, as reviewed in Marsh et al., (2010). In
particular, locomotor training with body weight support has
become used increasingly to improve locomotor function
and promote systemic cardiovascular, bone, and muscle
function (Wernig et al., 1995; Hicks et al., 2005; Dobkin et
al., 2006, 2007), as reviewed in Wessels et al., (2010). There
is considerable evidence that functional locomotor circuitry
is present below the level of a cervical or thoracic injury and
that these rehabilitation paradigms can take advantage of that
circuitry. To date, however, even these training paradigms
have provided very limited functional neurological
improvement for patients with complete or incomplete SCI.
To understand why, it is essential to recognize that there is
still very little known about the capacity for plasticity after
injury in the adult human spinal cord. Importantly, while
one may view the spared circuits as being relatively “intact”,
the segmental and short propriospinal systems both above
and below the level of damage are denervated as the damaged
axons undergo Wallerian degeneration, while long
propriospinal fibers may be spared, but may function in an
abnormal manner. Indeed, recent reports clearly indicate that
spinal cord neuronal function deteriorates over long periods
of time after injury (Dietz, 2010).
improve synaptic efficiency of established connections at the
expense of forming new ones. Recent studies have focused
on the effects of modifying the interactions of CSPGs
within the extracellular matrix using the bacterial enzyme,
chondroitinase ABC (chABC). ChABC specifically cleaves
the glucosaminoglycan side chains from sulfated CSPGs,
resulting in changes in their interactions with other matrix
proteins, release of sequestered growth factors or
modulators, and disassembly of the dense perineuronal nets
(PNNs) that are formed around the synapses of highly active
neurons (Pizzorusso et al., 2002). It is through a
combination of these effects that chABC enhances plasticity
and sprouting in the adult central nervous system and can
lead to improved functional recovery after injury (Bradbury
et al., 2002; Massey et al., 2006; Barritt et al., 2006; Galtrey
and Fawcett, 2007; Garcia-Alias et al., 2008; Tester and
Howland, 2008; Bradbury and Carter, 2010; Jakeman et
al., 2010). Other effects of chABC administration that may
contribute to recovery include neuroprotection (Carter et
al., 2008) and improved conduction through uninjured
tracts in the vicinity of the site of a partial spinal cord lesion
(Hunanyan et al., 2010).
Despite the complex actions of chABC on plasticity and
recovery after injury, the sites of action in different models
are still unclear, and the effects on recovery of function are
fairly modest. Many of the reports of successful recovery of
function have incorporated repeated intrathecal infusion of
chABC in a paradigm of 6 μl of 10 U/ml every other day for
10 days (Bradbury et al., 2002; Caggiano et al., 2005; Barritt
et al., 2006). However, chronic infusions present a clinical
challenge (Protopapas et al., 2007; Deer et al., 2007), and
are limited by tissue diffusion properties, and it is difficult to
define the site of action. Notably, discrete intraparenchymal
microinjections of higher doses of chABC have been shown
to be sufficient to disrupt perineuronal nets and re-activate
plasticity in response to physiological stimuli in the adult
visual (Pizzorusso et al., 2002) and somatosensory systems
(Massey et al., 2006). Similarly, a single intraparenchymal
injection of chABC to the cervical spinal cord enlargement
is sufficient to improve appropriate functional recovery after
crossed reinnervation of forelimb peripheral nerves in
pretrained rats (Galtrey et al., 2007). Given the important
point that the targeted site for improving locomotor function
may well be through modification of synaptic circuitry in
the lumbar spinal cord, we proposed that a clinically feasible
strategy for promoting recovery of function after SCI could
be developed using intraparenchymal microinjections of
chABC to the lumbar cord encompassing the denervated
locomotor circuitry. The purpose of this project was to
determine if chondroitinase ABC and locomotor training
could be combined to enhance synaptic plasticity in
segments distal to the site of an incomplete mid-thoracic
spinal cord contusion injury in rodent models.
The adult central nervous system exhibits limited
endogenous plasticity in comparison to that seen early in
development. One key factor is the maturation of the
extracellular matrix surrounding established synaptic
connections. In particular, the expression and sulfation
pattern of chondroitin sulfate proteoglycans (CSPGs) is
altered with development. While the role of CSPG
maturation is still poorly understood, the contribution of
this family of growth inhibitory molecules to matrix
structures surrounding highly active neurons likely serves to
86
experiments. All mice were acclimated to the laboratory and
behavioral apparatus for 1 week. After completing baseline
motor and sensory testing, the mice were randomly assigned
to receive a contusion injury or laminectomy surgery only.
Then, at 1, 3 and 7 days post-injury, the mice were tested
for locomotor function, and on day 7 the injured mice were
assigned to one of four treatment groups. One half of the
injured mice received a microinjection of ChABC in the
lumbar spinal cord, while the other half received an identical
injection of phosphate buffered saline (PBS) vehicle.
Beginning at 8 days after injury (1 day post-injection), half
of each injection group was then housed in cages with 24 hr
× 7 day/week access to a running wheel, while the remaining
mice were housed in identical cages with no wheel. The
treatment groups were as follows: Cs= chABC/sedentary (no
wheels); Cr= chABC/running wheels; Vs=Vehicle/sedentary;
Vw=Vehicle/wheels; and Lam=laminectomy. The
experiments were done in two parts. In the first study,
34 mice were enrolled; with n=6–7 per group. Mice in the
first study were perfused at 4 or 7 weeks post-injury due to
complications associated with sensory testing (see below).
A total of 24 mice were enrolled treatment groups in the
second study (n=6/group). Mice in the second study were
perfused at 10 weeks post-injury. The timeline of the mouse
studies is shown in Figure 1A.
The experiments proceeded in two stages. The first set of
studies tested the hypothesis that chABC applied directly to
the gray matter in the lumbar enlargement would facilitate
the benefits of voluntary wheel running after a moderate
contusion injury in mice. While there were hints of
improvement, the combination did not yield robust
functional benefit as predicted in this model. Therefore, the
second set of experiments represented a modest change in
direction and was designed to answer critical questions
around the initial hypothesis. We sought first to establish
whether injury-induced changes in CSPG composition
indeed extended throughout the spinal cord after a midthoracic contusion. These results provided additional direct
support for the rationale of targeting distant segments of the
spinal cord for the enzymatic adjunct therapy. Then, we used
the rat contusion model with an assist-as-needed locomotor
training paradigm mimicking the conditions used in a
number of rehabilitation centers (Dobkin et al., 2006). We
sought to determine if weight supported treadmill training
with assist as needed therapy would be effective alone or in
combination with either intraparenchymal chABC injections
or every other day bolus intrathecal chABC treatment to
improve recovery of overground locomotion in rats In
addition, we evaluated whether the combined therapies
would normalize spinal reflex and/or sensory changes after
contusion injury. Notably, the treatment approach for both
the mouse and rat studies began at 7 days after injury, which
was chosen because this is a time when endogenous sprouting
is occurring. Together, the results of these studies indicate that
CSPG expression is increased throughout the spinal cord by
7 days post injury. In addition, characteristics of spinal
plasticity and functional recovery in segments distal to the
site of injury can be altered through a combination of training
and chondroitinase. However, there is still work to be done to
understand and exploit the nature of the subtle and task
specific improvements that are in currently in reach for
treating the moderate and severe contusion injuries that
represent a large proportion of the clinical condition.
METHODS
General methods
All procedures on animals were carried out according to the
guidelines of the NIH and the National Research Council
Guide to the Care and Use of Laboratory Animals. All
procedures and protocols were approved by the Ohio State
University Institutional Lab Animal Care and Use
Committee and animals and the housing facility inspected
regularly by the University Lab Animal Research Department.
Injured mice and rats were cared for according to established
procedures as described elsewhere (Jakeman et al., 2000).
Care for SCI injured rodents includes daily inspection and
cleaning as needed, bladder expression 2–3× per day, urine
pH and weight monitoring and nutritional supplements.
Figure 1. Time course of behavioral studies. A. Mouse studies combined
wheel running with chABC microinjection placed into the lumbar
enlargement. Preinjury testing was carried out on all mice and
contusion injury or laminectomy performed at day 0 (0d). At 7 days,
the spinal cord was exposed and mice injected with 1 μl of 50 U/ml
chABC or PBS vehicle. The following day, half the mice from each
group were housed with 24 hours access to running wheels. Some
animals from Study 1 were perfused at 4w after injury due to
overgrooming. B. Rat studies compared the effect of treadmill training
alone, or treadmill training combined with intraparenchymal 2 × 1
μl microinjections of 100 U/ml chABC at 7 and 14 dpi or every other
day intrathecal bolus chABC infusions of 6 μl at 10 U/ml. Treatments
began at 7 dpi and treadmill training began at 8 dpi. VFH= von Frey
hair plantar sensory testing; BMS= Basso mouse scale for overground
locomotion; BBB=Basso,Beattie,Bresnahan scale for rat overground
locomotion.
Part I: Intraparenchymal chABC and voluntary wheel
running in mice
Animals and treatment groups:
Adult female mice (10 weeks of age; Jackson Laboratories)
were used and singly housed for the duration of the
87
Spinal cord injury and microinjection surgeries:
On day 0, the mice were anesthetized with ketamine
(80 mg/kg) and xylazine (10 mg/kg) and subject to a
moderate contusion injury to the spinal cord at the T9
vertebral level by a 0.5 mm displacement using the Ohio
State University Electromagnetic Spinal Cord Injury Device
(ESCID) as described previously (Jakeman et al., 2000).
Controls received an identical laminectomy only. At 7 dpi,
the injured mice received a partial laminectomy at the
T12/L1 vertebral junction. One half of the mice received a
single 1.0 μl intraparenchymal microinjection of Acorda
chABCI in PBS (Caggiano et al., 2005); (50 U/μl in PBS;
generously provided through an approved materials transfer
agreement with Acorda Pharmaceuticals) and one half
received an identical microinjection of PBS alone. The
microinjections were positioned 0.8 mm deep to the dorsal
surface of the spinal cord. The injection was produced with
multiple steps of ∼100 nl each, spaced over a 20 minute
period. After the last injection, the pipette was kept in place
for 2 minutes and then slowly withdrawn and the overlying
muscles and skin sutured.
with similar segments from each treatment group, and frozen
in a chamber with powdered dry ice. Each tissue block was
cut in the transverse plane on a cryostat at 10 μm thickness,
thaw mounted on Superfrost + slides, and stored at −20°C
until staining. One set of sections spanning each tissue block
and spaced 100 um apart was stained with Eriochrome
cyanine (EC) to establish the size and distribution of residual
white matter at the lesion site and to define the spinal
segments within the cervical and lumbar enlargements.
Additional series of adjacent sections through the lumbar
enlargement were stained with Wisteria floribunda
agglutinin lectin (WFA) to identify perineuronal nets
(Hartig et al., 1992; Seeger et al., 1994), or mouse anti-Di6S
antibody (Seikagaku Corporation, Cape Cod Associates) to
identify 6S stubs present following chABC digestion of
CSPG-GAG sidechains. WFA staining was performed by
bringing slides to 37°C for 1 hour, washing 3× in PBS, and
incubating overnight in 0.1 μl/ml biotinylated WFA (Vector
Labs B1355). The following day, the sections were rinsed,
incubated in avidin-biotin-complex reagent (Elite ABC;
Vector Labs), and developed using diaminobenzidine (DAB)
as a chromagen and NiCl enhancement. For Di6S
immunostaining, sections were incubated in primary
antibody at 1:1000 dilution overnight, rinsed and incubated
in biotinylated goat-anti-mouse IgG (Vector labs), amplified
with ABC reagent, and developed with 0.6% H2O2 using
VIP or SG as a chromagen (Vector labs).
Behavioral baseline and post-injury testing:
Prior to injury, mice were acclimated to in-cage running
wheels (Mini-Mitter, Respironics, Inc.) for 1–3 days until
they reached criterion (>15,000 rev/day or >25,000 rev/2
days). The wheels were modified with the addition of a
lightweight textured vinyl surface Locomotor function was
assessed before injury and at 1,3 and 7 days post injury (dpi)
and weekly thereafter using the Basso Mouse Scale (BMS).
Sensory tests were performed before injury and at 2, 3, 4
and 6 weeks post-injury, and included plantar heat
withdrawal using a Hargraeves apparatus (Hargreaves et al.,
1988), and plantar mechanical withdrawal thresholds using
Seimmes-Weinstein microfilaments (von Frey Hairs; (see
Hutchinson et al., 2004)). At the end of study 1, the mice
were tested for % withdrawal response to pinprick or
withdrawal threshold to von Frey hair mechanical stimuli
applied above and below the level of injury. Application of
these tests in injured C57BL/6 mice has been described in
detail elsewhere (Hoschouer et al., 2010a). Additional
outcome measures included the latency to withdrawal to
cold, using ice pops applied to the plantar surface of the
hindpaw (Lindsey et al., 2000) and the presence of a
proprioceptive placing response, assessed from videotape of
mice subjected to contact of the dorsal aspect of the
hindpaw on a table top, leading to dorsiflexion of the ankle
joint and a full (score =2); partial (score=1) or no (score =0)
resulting placement response to position the plantar surface
of the paw on the table top (Basso, 2004).
Statistics:
Behavioral outcome measures were compared using two way
ANOVA with repeated measures (BMS, plantar heat
latencies, von Frey Hair thresholds) or Mann Whitney U
(reflex measures) with time post-injury and treatment group
as the independent variables. All analyses and graphing was
done using Prism 5 (Graphpad, Inc.).
Part II: CSPG expression at and distal to the injury
site following contusion injury in rats:
Animals and injuries:
A total of 40 adult female Sprague Dawley rats (225–250 g;
Charles River Laboratories) was used to characterize the
patterns of CSPG expression in cervical, thoracic and lumbar
spinal cord before, and at 3,7,14 and 28 days after moderate
contusion injury (n=8/time point). The rats were anesthetized
with ketamine and xylazine and a T8 laminectomy
performed. The surrounding vertebral processes were secured
to a holding frame and the exposed dura subjected to a
250 kDyne contusion injury (n=32) using the Infinite
Horizons (IH) spinal cord impactor (Scheff et al., 2003). The
remaining rats received a sham laminectomy with no injury
(n=4) or served as naïve controls (n=4). At the designated
time post injury, all rats received an overdose of ketamine and
xylazine for tissue harvest. One half of the animals were
transcardially perfused with PBS and 4% paraformaldehyde
and the tissue blocks frozen for histological analysis as
described above. The remaining rats (n=4/group) were
euthanized after anesthesia and the entire spinal cord was
rapidly removed by dissection, blocked into cervical, thoracic,
and lumbar segments of 10 mm, 6 mm and 10 mm in length,
respectively, and frozen in liquid nitrogen.
Anatomical assessment:
At the designated time post injury, the mice were deeply
anesthetized and perfused with phosphate buffered saline
(PBS) followed by 4% paraformaldehyde in PBS. The entire
spinal cord was removed and marked at spinal level C5, the
injury site, and L5. Tissues were postfixed 2 hours, rinsed
overnight in 0.2M phosphate buffer (PB), cryoprotected
2–3 days in 30% sucrose in dH2O and then blocked,
embedded in OCT (optimal cutting temperature) medium
88
sections per spinal cord were selected across the mapped C4C6 and L4-L5 spinal segments. Sections were viewed on a
Zeiss Axiophot with fluorescent filters and images of
selected regions of cervical and lumbar spinal cords were
collected using a Sony 970 CCD camera and the
Microcomputer Imaging Device (MCID) from Imaging
Research, Inc. All image collection was done under identical
microscopy and lighting conditions. The regions used for
analysis were left and right dorsal horn (DH), intermediate
gray (IG), ventral horn (VH), lateral dorsal ascending white
matter tracts (fasciculus cuneatis, LDC), medial dorsal
ascending white matter (faciculus gracilis, MDC); dorsal
corticospinal tract (DCST), ventral white matter (VWM)
and lateral white matter (LWM). Within each region, a box
of 1000 μm2 was digitized for analysis. For GFAP measures,
the total area of positive staining (target area) was
determined and values expressed as proportional area
(PA=target area/measured area). For neurocan measures, the
fluorescence intensity was used to compare the amount of
diffuse staining between regions and specimens. Values were
compared across regions and times post injury using
repeated measures 2 way ANOVA and post-hoc analyses
with Bonferroni corrected t-tests using Prism 5.0.
Western blot procedures:
Tissue samples were prepared for Western blot according to
the methods of Massey et al., (2008). Briefly, the tissue
samples were thawed on ice and homogenized in 10 volumes
of 40 mM Tris–HCl, pH 7.6, containing 40 mM sodium
acetate and a protease inhibitor cocktail (Complete, Roche
Applied Science, Indianapolis, IN). Homogenization was
performed on ice and aliquots of the homogenates at a final
protein concentration of 2–3 mg/ml were treated with
0.3 U/ml of protease-free chABC (chABC, Sigma Chemical)
for 8 hours at 37°C. Chondroitinase activity was stopped by
boiling the samples in the presence of 1× gel-loading buffer.
Samples containing 10 μg total protein were electrophoresed
on reducing 3–10% SDS-polyacrylamide gels and
transferred to nitrocellulose membranes. The resulting
blots were incubated with primary and HRP-labeled
secondary antibodies and immunoblots developed by
chemiluminescence using ECL reagent (Pierce ECL).
Antibodies used for western blots were: anti-neurocan clone
5212 (Chemicon/Millipore, Inc.,) which detects full length
250 kD neurocan and the 150 kD C terminal cleavage
product; anti-brevican core protein (mouse anti-brevican;
Becton-Dickinson clone 5284), anti-brevican N-terminal
ADAMTS4/5 cleavage product (rabbit polyclonal antibody
B61, Viapiano et al., 2003), anti-aggrecan core protein (clone
Cat-301, Matthews et al., 2002; Chemicon/Millipore clone
5284), anti-NG2 (rabbit anti-NG2; Chemicon cat#5230)
and anti-beta-tubulin (Sigma Chemical). The immunoblots
were scanned and the integrated optical density (O.D.) of
each target protein was quantified using Gel-Pro Analyzer
software (v3.1, Media Cybernetics, Silver Spring MD). O.D.
ratios (O.D. target protein/O.D. beta-tubulin) for each
protein was normalized to control tissue values and
compared by one-way ANOVA followed by Tukey’s multiple
comparison test using Prism 5.0 (Graphpad, Inc.).
Part III: Intraparenchymal vs. intrathecal chABC and
manual assist treadmill training in rats
Animal groups, injuries, and behavioral outcome measures:
Due to the manpower required for daily manual assist
treadmill training, this study was also completed in two
parts. Part 1 compared injured rats with treadmill training
with untrained injured control rats housed in normal caging
and handled for daily care and behavioral testing only. Part
2 compared rats with treadmill training plus chABC given
either by repeated intrathecal bolus infusion or
intraparenchymal injection at 7 and 14 dpi with untrained
injured control rats. All subjects were Female Sprague
Dawley rats (225–250g starting weight) obtained from
Charles River, Inc. as described above and housed 2–3 per
cage for the duration of the study. All behavioral testing was
done by individuals without knowledge of drug treatment
or treadmill training groups. The time course of the rat
study is illustrated in Figure 1B.
CSPG immunohistochemistry:
Blocks containing the cervical, thoracic, and lumbar spinal
cord were cryoprotected and frozen in OCT. Serial
transverse sections were made through the entire 10 mm
cervical and lumbar spinal cord blocks. The lesion epicenter
(thoracic spinal cord) blocks were cut in the center and serial
longitudinal sections of the rostral and caudal ends were cut
separately to facilitate mounting. All sections of rat spinal
cord were cut at 20 μm thickness, and stored at −20°C until
staining. Series of equally spaced sections separated by
200 μm were stained with Eriochrome cyanine and cresyl
violet (EC/CV) to define myelin and Nissl substance,
respectively. Adjacent series were then stained with
antibodies raised against glial fibrillary acidic protein (rabbit
anti-GFAP; 1: 1000; Dako, Inc.), neurocan (mouse
monoclonal clone 5212; 1:200; Chemicon, Inc.), aggrecan
(Cat 301; 1:200; Chemicon, Inc), brevican (mouse antibrevican; 1:200; Becton-Dickinson clone 5284), and NG2
(rabbit anti-NG2, 1:500; US Biologicals). Fluorescently
tagged secondary antibodies were used for detection
(Alexafluor goat anti-rabbit 488 or 694; Alexafluor goat
anti-mouse488 or 694). In some sections, cell nuclei were
counterstained with Draq5. For quantitative analysis of
GFAP and neurocan staining, a total of 8 equally spaced
First, 45 rats were used to compare effects of home cage
housing (injury control; IC) with daily treadmill training
(TT) beginning at 8 days post-injury. Prior to surgery, the
rats were acclimated to the open field testing pool, treadmill
apparatus and von Frey Hair elevated grid apparatus, and
baseline scores obtained for these measures. Then, on day 0,
30 rats were anesthetized and a T9 laminectomy performed;
they were then immediately subjected to T9 moderate-severe
contusion (250 kDyne) using the IH Impactor as described
above. The remaining rats served as naïve controls. The injury
force for impact was chosen based on prior studies indicating
that after a 250 kDyne T8 injury, rats typically recover some
stepping ability, but rarely exhibit forelimb/hindlimb
coordination without intervention. One rat was removed
because of a poor impact profile and one rat died after injury.
All remaining rats were then tested for overground
locomotion using the BBB rating scale (Basso et al., 1995) at
89
1, 3, and 7 days post-injury and von Frey hair thresholds
measured at 6 dpi. The injured rats were then assigned to two
balanced groups (n=14 each) based on BBB and von Frey
scores. Beginning on day 8 post-injury (dpi), one half of the
injured rats were enrolled in daily treadmill training
(20 minutes per session; 5 sessions per week) while the
remaining rats were left in their home cages and handled daily
for care. Overground locomotion was rated by 2 observers on
14, 21, 28 and 35 dpi using the BBB locomotor rating scale.
Von Frey testing of the threshold to withdrawal from
mechanical force was tested on 29 and 30 dpi based on prior
studies demonstrating that hindlimb allodynia, which can
develop following moderate-severe contusion injury, can be
alleviated by daily TT (Huchinson et al., 2004).
and the IT bolus injections were repeated at 9,11,13,15,17
and 19 dpi. At 11 and 15 dpi, the rats in the injury control
group (IC) were anesthetized with isofluorane and the skin
cut and closed with wound clips to control for some of the
anesthesia and surgical manipulations. A total of 2 rats per
treatment group were perfused after BBB testing at 14 dpi to
evaluate the distribution of chABC enzymatic activity using
WFA and Di6S staining.
All rats were tested on the BBB locomotor rating scale
at 14, 21, 28 and 35 dpi by investigators who were blind to
the treatment group assignment. Rats in the first study were
tested for hindlimb sensory withdrawal threshold using von
Frey hair. For the second study, a more complete panel of
endpoint behavioral outcomes was applied between 28 and
25 dpi with the naïve rats serving as controls. The endpoint
activity box measures were taken at 33 dpi. These included
inclined plane testing, Hargraeves plantar heat testing for
latency to withdrawal, and recordings of overground
locomotion using the Catwalk runway (Hamers et al.,
2001). At 21 and 28 dpi, all rats were videotaped during
the first 5 minutes of the 20 minute treadmill training
session including 1 minute unassisted for later analysis of
ankle joint movements and independence in walking.
In the second part of the rat study, 48 rats were used to
compare the effects of TT combined with every other day
bolus intrathecal infusion of chABC or intraparenchymal
injection in the lumbar enlargement. Rats were obtained from
Charles River as above and acclimated to the BBB testing pool
and treadmill apparatus as described above. Von Frey hair
testing was not performed because none of the rats in the first
study showed any signs of reduced thresholds after injury. The
animals were also acclimated to activity boxes to permit
evaluation of total activity following injury and treatment.
At 35 dpi, all remaining rats (n=9–10 per group) were
divided for biochemistry (n=5/group) or histology (n=5–
6/group). Tissues were harvested and prepared as described
for previous studies above. The analysis of CSPG and GFAP
expression is still in progress.
After baseline testing was completed, a total of 40 rats were
subjected to 250 kD contusion injury at the T8 level using
the IH impactor; 12 rats were randomly assigned to the
intrathecal infusion (IT) group and were subjected to a second
laminectomy at the T13 vertebral level and fitted with an
intrathecal cannula (Alzet; Durect, Inc.). After placement in
the intrathecal space, the cannula was secured to overlying
muscles using 4–0 sutures and the open end of the cannula
attached to an Alzet 2002 minipump containing the vehicle
(0.9% saline with 0.1% protease free BSA). A total of
12 uninjured rats served as controls. BBB scores were obtained
at 1, 3 and 7 dpi as described, and the unassigned injured
animals were balanced and assigned to the microinjection
(MI) or injury control (IC) groups. One rat was removed due
to an abnormal impact force curve, three rats were removed
because they had only slight hindlimb movements (BBB<4) at
7 dpi, and one rat died following anesthesia for the intraspinal
microinjection; leaving 11–12 rats per group. After assignment
at 7 dpi, rats in the MI group were reanesthetized and a second
laminectomy performed at the T13/L1 vertebral junction. A
micropipette (o.d. 40 um) filled with 100 U/ml chABC
(Seikagaku, Inc.; Garcia-Alias et al., 2009) in 0.9% saline with
0.1% BSA was lowered through the intact dura to a depth of
0.8 mm to target the left side of the lumbar enlargement gray
matter. A total of 2 μl of chABC was injected over 10 minutes
using a picospritzer pneumatic ejector, which dispersed the
drug in 200–300 nl increments. The rats in the IT group were
reanesthetized with 2% isofluorane in O2 and the
subcutaneous site of the intrathecal cannula exposed and the
Alzet pump removed. Using a 10 ul Hamilton syringe, the rats
received 6 ul of freshly thawed chABC at 10 U/ml (0.6 U) in
vehicle, followed by 10 ul of vehicle to flush the cannula. The
skin was resutured and the rats allowed to recover. Treadmill
training for both the IT and MI groups began at 8 days after
injury. The intraspinal MI injection was repeated at 14 dpi,
RESULTS
Part I: Voluntary wheel running and lumbar
intraparenchymal chondroitinase ABC (mice)
Study 1 Summary. Effects of ChABC and running wheels
on overground locomotion and sensory measures were not
significantly different from controls. Extensive pre-injury
sensory testing and contusion injury in mice are associated
with complications of self-directed biting behavior.
Beginning at 3 days post-injury, we discovered a much
higher than anticipated proportion of the injured animals
enrolled in the study were showing signs of overgrooming or
biting, particularly along the lower thoracic dermatomes of
the dorsal trunk. This phenomena, which we termed
“overgrooming” (OG), began prior to the initiation of
treatment (7 dpi), and was therefore not a result of the
microinjections or multiple surgeries. Most of the sites were
small and manageable by wrapping the torso between the
forelimbs and hindlimbs using vet wrap adhesive. Thus, the
study continued as planned, but by 9 dpi, as many as ½ of
the total injured mice were being treated for this condition.
The wrapped mice were able to eat, drink, and roam their
cages. However, nearly all those with running wheels were
not able to obtain the wheel running exercise required for
the proposed induction of plasticity. Therefore, at 4 weeks
post-injury, we removed all overgroomers from the study
and left only the unaffected mice for an additional 3 weeks.
After the removal of all overgroomers, the final groups sizes
ranged from n=1 – n=6 were too small for full comparison.
90
The behavioral results of Study 1 are shown in Figure
2. Although the data were underpowered, there was an
initial trend suggesting that the combined chABC and
wheelrunning mice (Cr) and chABC sedentary mice (Cs)
recovered more quickly on the BMS scale than the other
groups. From the histological specimens, we confirmed that
chABC cleavage of CSPG-GAGs in the lumbar spinal cord
leaves 6S stubs that can be identified at both 4 and 7 weeks
after injury (not shown). Follow-up studies were also
performed to test the interaction between the pre-injury
sensory testing and self-directed biting. In these
experiments, we showed that mice that are subjected to preinjury sensory stimulation on the trunk prior to contusion
injury have a higher likelihood of self-directed biting after
injury than mice that are not subjected to such stimulation.
This follow-up study was recently accepted for publication
(Hoschouer et al., 2010b). We then confirmed that a delay
of 2 weeks between sensory testing and injury is sufficient
to prevent the high incidence of this behavior.
For the second study, we modified our experimental
design in two ways. First, we reduced the pre-injury sensory
testing to a single session and extended the time between the
pre-injury testing and the injury date to 2 weeks to minimize
any combination of afferent overstimulation and
inflammation. Secondly, in collaboration with Dr. Sharon
Flinn from the Department of Occupational Therapy at the
Ohio State University, we also developed a high function
neck collar that could be used for mice to prevent access to
the most common site of irritation but not impede the ability
to run on the running wheels (described in (Hoschouer et
al., 2010b). In Study 2, 24 mice were acclimated to running
wheels and then subjected to a contusion injury. Only 4/24
mice required collars, these were applied for a maximum of
4 days and removed with no notable effects on mobility,
body weight, or locomotor scores. At 1 week post-injury, the
mice were re-anesthetized and received microinjections of
ChABC or PBS into the parenchyma as planned at spinal
level L4–5. All 24 mice completed this study. In the final
week of the study, the mice were subjected to a wide range
of sensory and motor tests, including BMS testing both
during the light and dark cycle, a propriospinal placing reflex
test, rotorod test, and sensory tests including withdrawal to
threshold and suprathreshold mechanical and pinprick
stimuli on the trunk above and below the injury level, and
plantar hindpaw heat withdrawal latency (Hargraeve’s
method), mechanical withdrawal threshold (up-down
method), and withdrawal to cold (ice pop) stimulus. The key
results from the two mouse experiments are shown in Figure
3, confirming that the single or combined therapies do not
affect the rate or final extent of recovery after SCI.
Figure 2. Behavioral results from mouse Study 1. A., B. Overground
locomotor recovery plotted according to the BMS scale and a BMS
subscale (ss). The subscale is used to score paw position and trunk stability
for mice with frequent to consistent stepping and serves to separate
differences in the quality of stepping. Lam= laminectomy; Cr= chABC
with running wheels; Cs=chABC sedentary (no wheels); Vr=vehicle with
running wheels; Vs=vehicle sedentary. There was a non-significant trend
toward more rapid recovery in the chABC groups. The n’s are small due
to early withdrawal of a number of mice for overgrooming. C.–F. There
were no differences in the sensitivity of the four injury groups in response
to heat or mechanical stimulation of the plantar surface of the hindlimbs
or pinprink or mechanical stimulation of the trunk below the injury.
Figure 3. Behavioral results from mouse Study 2. All groups contained
n=6 mice for the duration of the study. There were no differences
between treatment groups on the behavioral outcome measures in this
study and no differences in BMS or plantar tests when the results of the
two mouse studies were combined.
Study 2 Summary. There is no significant improvement in
recovery or abnormal sensation with single or combined
ChABC injection and wheel running following moderate
spinal contusion injury in mice.
91
Part II. Changes in CSPG expression after incomplete
thoracic contusion injury (rats)
We next examined the distribution of changes in CSPG
expression after contusion injury to determine if distal
spinal cord segments represent a reasonable target for
chABC treatment after injury. For these studies, we have
collaborated with Dr. Mariano Viapiano, who had done
previous work demonstrating changes in CSPG core protein
expression distal to the site of a dorsal column lesion in
rats (Massey et al., 2008). This work was performed in
rats because many of the tools used to examine CSPG
protein expression do not easily recognize the homologous
mouse proteins.
CSPG expression at the lesion epicenter
The extent of the contusion injury is illustrated by EC/cresyl
violet staining and GFAP immunohistochemistry in Figure
4. Following a 250 kDyne impact, there is a small rim of
spared white matter surrounding the lesion site. Subsequent
tissue damage extends as far as 5 mm rostral and caudal to
the site of initial impact.
Figure 5. Expression of neurocan and aggrecan at the lesion epicenter
after contusion injury in rats. A. Western blot shows three isoforms or
cleavage products of neurocan. B.–D. Quantitative analysis of the
time course of expression of total neurocan and the 150 kD and
250 kD products; n=4/time point. *=p<0.05;**p<0.01;***p<0.001
vs control (laminectomy or naïve samples). E.–G. Photomicrographs
of neurocan immunoreactivity (red) and colocalization with astrocytes
(GFAP; green). Arrow in G. shows a cell expressing both GFAP and
neurocan. H. Western blot of epicenter samples stained with Cat-301
antibody to identify aggrecan species. 3,7,14,28 represent samples at
each respective dpi. C=control sample. I. Quantitative analysis of
aggrecan western blots; *** p<0.001 vs. control. J. Double staining
illustrating aggrecan (green) in perineuronal nets surrounding a
cholinergic neuron as stained with an antibody to choline acetyl
transferase (ChAT;red). K. Normal distribution of aggrecan is highest
in spinal cord gray matter ventral horn (VH). L. Aggrecan staining of
transverse section at the lesion epicenter at 14 dpi. M. Illustration of
the loss of aggrecan staining extending as far as 2 mm from the caudal
border of the lesion site (arrow) at 14 dpi. Surviving motoneurons
without aggrecan are in red.
Figure 4. Histological examples of the rat contusion injury site
following a 250kDyn injury with the infinite horizons impactor. A.
Transverse sections illustrate the evolution of the lesion epicenter over
time as seen with eriochrome cyanine and cresyl violet (EC/CV)
staining. Scale = 200 μm. B.,C. Longitudinal sections through the
lesion site after staining with EC/CV or immunocytchemical staining
with antibodies to GFAP. Scale = 500 μm.
Aggrecan expression was dramatically reduced within
the lesion epicenter. Staining of Western Blots with the Cat301 antibody following chABC treatment of tissue samples
reveals a group of bands ranging from 200–300 kD in size
(Figure 5G,H). All of these bands were depleted by 3 dpi
and did not recover after 28 dpi. Staining of epicenter tissue
sections with Cat-301 antibody revealed labeling
throughout normal gray matter, especially surrounding
neurons in the intermediate gray and ventral horn. After
injury, there was no staining in the lesion site. In addition,
Cat-301 staining was depleted for several mm past the
lesion, including areas of neuropil that contained healthy
neuronal profiles (Figure 5I–L)
CSPG expression patterns at the injury epicenter
extended findings reported by others following a dorsal
column injury (Tang et al., 2003) or contusion injury (Iaci
et al., 2007). Total neurocan expression was increased up to
4 fold by 7 dpi and remained high at 28 dpi (Figure 5A–D).
This increase included a 40 fold increase in expression of
the full length (250 kD) form of neurocan as well as modest
increases in both the normal 150 kD adult neurocan and
appearance of a 180 kD cleavage product. Neurocan
immunoreactivity at the epicenter was associated with
astrocytes surrounding the lesion site. Neurocan was not
found within the lesion itself (Figure 5E–G).
Expression of the 100 kD isoform of brevican that is
recognized by the B61 antibody was depleted at the lesion
epicenter by 3 dpi, but it partially recovered by 28 dpi
(Figure 6A,B). Immunostaining with this antibody revealed
widespread expression in intact spinal cord including both
gray and white matter. After injury, there was a loss of
92
staining in the region of the lesion only, but a slight increase
in staining intensity in both gray and white matter
surrounding the injury site (Figure 6C–E).
CSPG expression in cervical and lumbar spinal cord segments
The most striking changes in CSPG expression in distal
segments were seen for neurocan (Figure 7).
Figure 7. Neurocan is upregulated chronically in the cervical and
lumbar spinal cord after mid-thoracic contusion in rats. A., E. Western
blots and B., C., F., G. Quantitative analysis of neurocan expression
after injury *=p<0.05;**=p<0.01;***p<0.001 vs. uninjured control.
D. GFAP (green) and neurocan (red) immunostaining in the dorsal
columns of the cervical spinal cord at 14 dpi. Both markers are
upregulated in the medial portion of the dorsal columns (MDC)
corresponding to the location of the fasciculus gracilis. This region
contains degenerating sensory axons from the lumbar segments. H.
GFAP (green) and neurocan (red) immunostaining in the ventral
horn of the lumbar spinal cord enlargement at14 dpi. Both markers
are upregulated in the gray matter corresponding to the location of
terminal fields of descending projections.
Figure 6. Expression of brevican and NG2 at the lesion epicenter after
contusion injury in rats. A. Western blot of brevican showns two
isoforms. B. Quantative analysis of total brevican expression
(n=4/time point). C. Distribution of brevican in gray and white
matter regions of spinal cord. D. Brevican staining is increased in
spared rim of white matter surrounding the lesion. E. Brevical staining
is present up to the edge of the lesion, nuclei are counterstained with
Draq5 (blue). F., G. Western blot and quantitative analysis of NG2
expression shows no significant change in NG2 expression at the lesion
site. H. Colocalization of neurocan (green) and NG2 in the
extracellular neuropil distal to the site of injury. I.–J.
Photomicrographs of the lesion border at 7 and 14 dpi. K. NG2 is
found throughout the lesion center at 14 dpi.
By Western Blot analyses, total neurocan was increased in
both cervical and lumbar spinal cord by 7 dpi. The effect
was earlier (3 dpi) and prolonged (through 28 dpi) in the
lumbar spinal cord tissues. Contributing to this increase in
neurocan expression were significant increases in all
3 isoforms or cleavage products of neurocan, with the most
marked increases (10–15 fold over control) for the large
uncleaved 250kD species. Immunostaining with the same
antibody revealed increased intensity of immunoreactivity in
both gray and white matter regions of the distal spinal cord
segments. To examine if these increases corresponded directly
to regions of increased astrocyte activation, adjacent sections
were stained with antibodies to GFAP. Semi quantitative
analysis of GFAP reactivity revealed significant increases that
differed between cervical and lumbar spinal cord. In cervical
spinal cord, GFAP was increased from 14–28 dpi, but only
in the medial dorsal columns (degenerating fasciculus
gracilis). In contrast, GFAP immunoreactivity was increased
in the intermediate gray and ventral horn as well as lateral
and ventral white matter of the lumbar spinal cord. These
increases were somewhat variable and peaked at 14 dpi.
These results were used to focus the evaluation of neurocan
immunostaining intensity to the MDC, DCST, IG and
VGM regions. Analysis of neurocan expression in cervical
NG2 expression levels were unchanged at the injury
epicenter by Western Blot analyses (Figure 6F,G). However,
immunocytochemistry revealed an underlying shift in
expression patterns that may have masked measures of
absolute protein expression. Specifically, NG2 staining was
initially lost at the lesion site, while NG2 staining in the
tissue surrounding the lesion was markedly increased at
7 and 14 dpi. By 14–28 dpi, NG2 staining was found
throughout the lesion site and surrounding tissues as
described in detail previously (Figure 6H–K)(McTigue et
al., 2006).
93
spinal cord revealed increased staining in the medial dorsal
columns with no change in gray matter. In contrast, increases
in neurocan intensity were found in the dorsal corticospinal
tract and ventral horn, with a trend toward increase in
intermediate gray matter. All of these increases peaked at
14 dpi and corresponded with regional increases in
GFAP immunoreactivity.
of the lumbar spinal cord parenchyma combined with
locomotor exercise would enhance recovery of overground
locomotion. We moved away from mice and wheel running
in part because the voluntary exercise had proven to provide
an inconsistent amount of locomotor training. We instead
explored this hypothesis using rats, which are more practical
for weight supported forced treadmill training, and for
which we now had biochemical assays available. In addition,
we hoped to address the hypothesis that intraparenchymal
injections would be directed toward plasticity induced
recovery by comparing treatment by microinjection with
that obtained using intrathecal bolus injections of chABC.
The doses for intraparenchymal and intrathecal
administration were determined using effective doses from
the literature. Also based on prior reports, we expanded the
microinjection paradigm to include a total of 2 injections
spaced one week apart. Importantly, we chose to begin our
treatments in all cases at 7 dpi.
Treadmill training beginning at 7 dpi does not improve
overground locomotion after mid-thoracic contusion
The behavioral results of the first arm of the study are
illustrated in Figure 8A. A 250kD injury resulted in
complete paralysis followed by a period of recovery in BBB
scores that reached a plateau by 14 dpi. Rats that received
TT 5 days per week from 7 dpi to 28 dpi did not show
improved recovery of overground locomotion compared
with rats housed in pairs in their home cages. Surprisingly,
none of the rats showed a reduced threshold for withdrawal
to tactile stimuli after injury, which has been seen in some
contusion injury models and is frequently used as a measure
of hindlimb allodynia.
Figure 8. Behavioral results from rat studies. A. Study 1. There were
no differences in the time course of locomotor recovery after contusion
injury in rats housed in their cages (sedentary) and rats subjected to
5×/week treadmill training (TT). B. Study 2. The rats with
intrathecal infusions were delayed in their recovery of overground
locomotion, but there were no differences between the groups in their
final BBB scores. C. Study 2. All rats showed deficits in ability to
maintain body position on the inclined plane at 4 weeks post-injury.
**p<0.01; ***p<0.001 vs. Naïve rats. D. Study 2. Rats receiving
chABC by intraparenchymal microinjection showed no difference in
latency to withdraw from a noxious heat stimulus, while the injury
control and intrathecal rats both had reduced latency or hyperalgesia.
**p<0.01 vs. Naïve. (n=9–10/group).
Neither intrathecal nor intraparenchymal chABC
combined with treadmill training beginning at 7 dpi
improves gross overground locomotion scores after midthoracic contusion.
Western blot analyses of aggrecan and brevican
expression did not reveal significant changes in cervical or
lumbar spinal cord segments after mid-thoracic contusion.
However, there was a significant increase in NG2 expression
in lumbar, but not cervical tissues. Detailed regional analysis
of NG2 expression is still underway. Together these findings
reveal changes in neurocan expression and NG2 expression,
especially in the lumbar spinal cord which is strongly
denervated of descending input following mid-thoracic
contusion. These CSPGs are strongly expressed by glial cells
that respond to Wallerian degeneration within distal
segments. In contrast, brevican and aggrecan, which
contribute to cell migration and perineuronal net structures,
respectivity, are unchanged in the distal regions of the spinal
cord. These findings suggest that there is a restricted
capacity of the distal segments to undergo synaptic plasticity
after contusion.
The behavioral results of the second arm of the study
are illustrated in Figure 8B. The injury produced a
functional deficit that was slightly more severe than that
seen in the first arm. Rats were paralyzed followed by a
period of recovery by 14 dpi. Rats that had only slight
movements of the joints of the hindlimbs by 7 dpi were
removed from the study to ensure that treadmill training
would be possible. At the end of the study most rats were
able to step frequently or consistently in the open field, and
only a few showed occasional passes with forelimb-hindlimb
coordination. It is important to note that there was some
variation in outcomes within the three groups, but the
outcomes across groups were not different at the study
endpoint. A recent study by (Kuerzi et al., 2010) suggests
that in cage locomotor activity of rats with incomplete
contusion injuries may be sufficient to provide task specific
training for control and “treated” groups. Compared with
the 20 minutes of assisted treadmill training, the nightly
movement of the rats in their cages may support
performance in overground locomotion that represents a
“ceiling” effect that is maximal for the amount of preserved
anatomical substrate in moderate or severe injury. In this
event, it is possible that chABC simply could not extend the
Part III. Manual assisted treadmill training and
chondroitinase ABC (rats)
With the knowledge that neurocan and NG2 expression are
increased distal to the site of injury and maintained for up
to 28 dpi we then reexamined whether chABC treatment
94
degree of locomotor recovery any more than is observed
with normal cage locomotor behavior alone. Unexpectedly,
the rats that received intrathecal bolus chABC infusions
showed impaired recovery on overground locomotion
during the period of the infusions as compared to the MI or
IC groups. The reasons for this are unclear. The injury
control rats underwent procedural anesthesia with
isofluorane and skin incisions, so it is unlikely that this
component of the treatment was deleterious to overground
locomotion. All rats were allowed at least 18 hours to
recover from isofluorane prior to BBB testing. We have
done an number of intrathecal cannula studies in the past,
including chronic minipump infusions in mice and rats
(Ankeny et al., 2001;Mire et al., 2008;White et al., 2008)
and intrathecal bolus infusions in mice (unpublished) and
have not seen deficits induced by this procedure. Likewise,
a number of other labs have used a similar approach with
chABC and observed no ill effects (Bradbury et al.,
2002;Caggiano et al., 2005;Barritt et al., 2006). At this
time, we are hypothesizing that because the cannulas used
in this study were new and the surgeons doing the infusions
were new to this approach, it is possible that the rate of
injection with the Hamilton syringe may have been too
high, inducing a transient deficit in behavioral function in
the open field. Notably, the detrimental effect seen in this
group was completely eliminated immediately after the
intrathecal infusions ended (19 dpi; See Figure 1B).
incomplete, moderate to severe contusive injury. This series
of studies was performed to test the hypothesis that
intraparenchymal chABC treatment (or ‘jab’) would
enhance the very limited degree of plasticity in the adult
nervous system and facilitate the limited benefits of
locomotor exercise or training to improve overground
locomotion after incomplete contusive SCI.
The results of these studies are certainly discouraging, as
we failed to demonstrate a robust improvement in BBB scores
with the treatment combination in spite of a compelling
rationale and well informed approach in two species.
However, when examined in the context of the many other
studies in the literature, one can use this design and results to
elucidate more specific hypotheses regarding the potential
mechanisms of chABC and rehabilitation strategies for
clinical application. Clearly, chABC exerts multiple effects on
the injured nervous system due to the complex interactions of
CSPG-GAG sidechains with the actively signaling
extracellular matrix components. The three best supported
effects on recovery from SCI include first, an early and robust
neuroprotective component that follows administration of
chABC immediately after trauma (Bradbury et al., 2002;
Caggiano et al., 2005). In addition, there is an early effect on
local axonal sprouting that can be exploited if there is a
permissive terrain for axon growth at the site of damage (Yick
et al., 2003; Houle et al., 2006), and finally, a modest capacity
to contribute to facilitation of specific activity dependent
synaptic remodeling that requires high doses placed directly
into the parenchyma and strong intrinsic activation of specific
behavioral or physiological circuitry (Pizzorusso et al., 2002;
Galtrey et al., 2007). Notably, the improved function that has
been shown to occur when chABC is combined with a
forelimb dexterity task was observed after treating the injured
spinal cord with both intrathecal and intraparenchymal
chABC beginning immediately after injury. The former two
effects and the lack of improvement in our delayed intrathecal
or intraparenchymal treatment studies suggest that immediate
and sustained administration of chABC directly to the site of
acute injury is required to exert the most robust benefits in
terms of recovery, which would be feasible and effective
following mild or partial SCI. In addition, the present
findings provide further support for the rationale of a delayed
treatment later in recovery to permit reorganization in distal
segments. Furthermore, the benefits of specific afferent
derived input from locomotor training on reducing
hyperexcitability or spasticity may be further enhanced in
combination with intraparenchymal chABC. However, this
approach to facilitate distal neural plasticity will require
combination with a strong and prolonged physiological
stimulus and the required components are not yet defined.
Taken together with other recent findings, we suggest that
chABC is a beneficial adjunct therapy for delayed treatment
and the modification of distal activity after incomplete SCI
through the appropriate combination of electrical and
pharmacological tone could be optimized with the addition
of chABC to facilitate remodeling.
Additional outcome measures suggest that hindlimb reflex
hyperactivity may be reduced by the combined treatment of
intraparenchymal chABC and treadmill training
At the conclusion of the study, the rats were tested on
additional behavioral tasks to determine if either treatment
arm had an effect on more subtle aspects of recovery of
hindlimb function. The results are shown in Figure 8C,D.
Notably, there was no effect of treadmill training in
combination with either chABC treatment on the angle
sustained on the inclined plane or total activity or number
of rearing events over 30 minutes in an activity box.
However, the group that received microinjections of chABC
directly into the lumbar spinal cord gray matter had a
greater or more normalized latency to respond to a
nociceptive heat stimulus applied to the plantar surface of
the hindlimbs. In contrast, those animals with intrathecal
chABC did not differ in the hyperalgesic response or
reduced latency to withdraw from the heat than the injury
control group. At the time of this final report, we have not
completed analysis of the Catwalk stepping patterns or
videotaped segments of treadmill walking for ankle
kinematics. These detailed analyses are underway.
CONCLUSION
There are numerous previous studies that illustrate the
efficacy of chABC as an enhancer of neural sprouting after
injury and as a facilitator of neurotrophic treatments in
models of synaptic plasticity and recovery in the intact and
injured nervous system. However, it is still unclear if, or
how, this enzyme might be used in a broader clinical sense
as a stimulus to improve the profound loss of descending
input required for gross locomotor or sensory recovery after
Ankeny, D.A., McTigue, D.M, Guan, Z., Yan, Q, Kinstler,
O.B., Stokes, B.T., Jakeman, L.B. (2001) Pegylated Brain95
derived neurotrophic factor shows improved distribution
into the spinal cord and stimulates locomotor activity and
morphological changes after injury. Exp. Neurol. 170:
85–100.
J., Yip, P., McMahon, S.B., Bradbury, E.J. (2006)
26:10856–10867.
Basso, D.M. (2004) Behavioral testing after spinal cord
injury: congruities, complexities, and controversies.
J. Neurotrauma. 21:395–404.
Basso, D.M., Fisher, L.C., Anderson, A.J., Jakeman, L.B.,
McTigue, D.M., Popovich PG (2006) Basso Mouse Scale for
Locomotion Detects Differences in Recovery after Spinal
Cord Injury in Five Common Mouse Strains. J. Neurotrauma.
23:635–659.
Bradbury, E.J., Carter, L.M. (2010) Manipulating the glial
scar: Chondroitinase ABC as a therapy for spinal cord
injury. Brain Res. Bull.
Bennett, G.S., Patel, P.N., Fawcett, J.W., McMahon, S.B.
(2002) Chondroitinase ABC promotes functional recovery
after spinal cord injury. Nature. 416:636–640.
Caggiano, A.O., Zimber, M.P., Ganguly, A., Blight, A.R.,
Gruskin, E.A. (2005) Chondroitinase ABCI improves
locomotion and bladder function following contusion
injury of the rat spinal cord. J. Neurotrauma. 22:226–239.
Carter, L.M., Starkey, M.L., Akrimi, S.F., Davies, M.,
McMahon, S.B., Bradbury, E.J. (2008) The yellow
fluorescent protein (YFP-H) mouse reveals neuroprotection
as a novel mechanism underlying chondroitinase ABCmediated repair after spinal cord injury. J. Neurosci.
28:14107–14120.
Deer, T.R., Raso, L.J., Garten, T.G. (2007) Inflammatory
mass of an intrathecal catheter in patients receiving baclofen
as a sole agent: a report of two cases and a review of the
identification and treatment of the complication. Pain Med.
8:259–262.
Dietz, V. (2010) Behavior of spinal neurons deprived of
supraspinal input. Nat. Rev. Neurol. 6:167–174.
Dobkin, B., Apple, D., Barbeau, H., Basso, M., Behrman,
A., Deforge, D., Ditunno, J., Dudley, G., Elashoff, R.,
Fugate, L., Harkema, S., Saulino, M., Scott, M. (2006)
Weight-supported treadmill vs over-ground training for
walking after acute incomplete SCI. Neurology. 66:
484–493.
Dobkin, B., Barbeau, H., Deforge, D., Ditunno, J.,
Elashoff, R., Apple, D., Basso, M., Behrman, A., Harkema,
S., Saulino, M., Scott, M. (2007) The evolution of walkingrelated outcomes over the first 12 weeks of rehabilitation
for incomplete traumatic spinal cord injury: the multicenter
randomized Spinal Cord Injury Locomotor Trial.
Neurorehabil. Neural Repair 21:25–35.
Engesser-Cesar, C., Anderson, A.J., Basso, D.M., Edgerton,
V.R., Cotman, C.W (2005) Voluntary wheel running
improves recovery from a moderate spinal cord injury.
Engesser-Cesar, C., Ichiyama, RM., Nefas, A.L., Hill, M.A.,
Edgerton, V.R., Cotman, C.W., Anderson, A.J. (2007)
Wheel running following spinal cord injury improves
locomotor recovery and stimulates serotonergic fiber
growth. Eur. J. Neurosci. 25:1931–1939.
Fitch, M.T., Silver, J. (2007) CNS injury, glial scars, and
inflammation: Inhibitory extracellular matrices and
regeneration failure. Exp, Neurol.
Fouad, K., Schnell, L., Bunge, M.B., Schwab, M.E.,
Liebscher, T., Pearse, D.D. (2005) Combining Schwann cell
bridges and olfactory-ensheathing glia grafts with
chondroitinase promotes locomotor recovery after complete
transection of the spinal cord. J. Neurosci. 25:1169–1178.
Galtrey, C.M., Asher, R.A., Nothias, F., Fawcett, J.W.
(2007) Promoting plasticity in the spinal cord with
chondroitinase improves functional recovery after peripheral
nerve repair. Brain. 130:926–939.
Galtrey, C.M., Fawcett, J.W. (2007) The role of chondroitin
sulfate proteoglycans in regeneration and plasticity in the
central nervous system. Brain Res. Rev. 54:1–18.
Garcia-Alias, G., Barkhuysen, S., Buckle, M., Fawcett, J.W.
(2009) Chondroitinase ABC treatment opens a window of
opportunity for task-specific rehabilitation. Nat. Neurosci.
12:1145–1151.
Garcia-Alias, G., Lin, R., Akrimi, S.F., Story, D., Bradbury,
E.J., Fawcett, J.W. (2008) Therapeutic time window for the
application of chondroitinase ABC after spinal cord injury.
Exp. Neurol. 210:331–338.
Hamers, F.P., Lankhorst, A.J., van Laar, T.J., Veldhuis, W.B.,
Gispen, W.H. (2001) Automated quantitative gait analysis
during overground locomotion in the rat: its application to
spinal cord contusion and transection injuries.
Hargreaves, K., Dubner, R., Brown, F., Flores, C., Joris, J.
(1988) A new and sensitive method for measuring thermal
nociception in cutaneous hyperalgesia. Pain. 32:77–88.
Hicks, A.L., Adams, M.M., Martin, G.K., Giangregorio,
L., Latimer, A., Phillips, S.M., McCartney, N. (2005) Longterm body-weight-supported treadmill training and
subsequent follow-up in persons with chronic SCI: effects
on functional walking ability and measures of subjective
well-being. Spinal Cord. 43:291–298.
Hockfield, S., Kalb, R.G., Zaremba, S., Fryer, H. (1990)
Expression of neural proteoglycans correlates with the
acquisition of mature neuronal properties in the mammalian
brain. Cold Spring Harb. Symp. Quant. Biol. 55:505–514.
Hoschouer, E.L., Basso, D.M., Jakeman, L.B. (2010a)
Aberrant sensory responses are dependent on lesion severity
after spinal cord contusion injury in mice. Pain. 148:
328–342.
Hoschouer, E.L., Finseth, T., Flinn, S., Basso, D.M.,
Jakeman, L.B. (2010b) Sensory stimulation prior to spinal
cord injury induces post-injury dysesthesia in mice.
Houle, J.D., Tom, V.J., Mayes, D., Wagoner, G., Phillips,
N., Silver, J. (2006) Combining an autologous peripheral
nervous system “bridge” and matrix modification by
chondroitinase allows robust, functional regeneration
beyond a hemisection lesion of the adult rat spinal cord.
J. Neurosci. 26:7405–7415.
Hunanyan, A.S., Garcia-Alias, G., Alessi, V., Levine, J.M.,
Fawcett, J.W., Mendell, L.M., Arvanian, V.L. (2010) Role
of chondroitin sulfate proteoglycans in axonal conduction in
Mammalian spinal cord. J. Neurosci. 30:7761–7769.
96
Hutchinson, K.J., Gomez-Pinilla, F., Crowe, M.J., Ying, Z.,
Basso, D.M. (2004) Three exercise paradigms differentially
improve sensory recovery after spinal cord contusion in rats.
Brain. 127:1403–1414.
Iaci, J.F., Vecchione, A.M., Zimber, M.P., Caggiano, A.O.
(2007) Chondroitin sulfate proteoglycans in spinal cord
contusion injury and the effects of chondroitinase
treatment. J. Neurotrauma. 24:1743–1759.
Ikegami, T., Nakamura, M., Yamane, J., Katoh, H., Okada,
S., Iwanami, A., Watanabe, K., Ishii, K., Kato, F., Fujita,
H., Takahashi, T., Okano, H.J., Toyama, Y., Okano, H.
(2005) Chondroitinase ABC combined with neural
stem/progenitor cell transplantation enhances graft cell
migration and outgrowth of growth-associated protein-43positive fibers after rat spinal cord injury. Eur. J. Neurosci.
22:3036–3046.
Jakeman, L.B., Hoschouer, E.L., Basso, D.M. (2010)
Injured mice at the gym: Review, results and considerations
for combining chondroitinase and locomotor exercise to
enhance recovery after spinal cord injury. Brain Res. Bull.
Kuerzi, J., Brown, E.H., Shum-Siu, A., Siu, A., Burke, D.,
Morehouse, J., Smith, R.R., Magnuson, D.S. (2010) Taskspecificity vs. ceiling effect: step-training in shallow water
after spinal cord injury. Exp. Neurol. 224:178–187.
Kwon, B.K., Hillyer, J., Tetzlaff, W. (2010a) Translational
research in spinal cord injury: a survey of opinion from the
SCI community. J. Neurotrauma. 27:21–33.
Kwon, B.K., Okon, E.B., Hillyer, J., Mann, C., Baptiste,
D.C., Weaver, L., Fehlings, M., Tetzlaff, W. (2010b) A
Systematic Review of Non-Invasive Pharmacologic
Neuroprotective Treatments for Acute Spinal Cord Injury.
J. Neurotrauma.
Kwon, B.K., Okon, E.B., Plunet, W., Baptiste, D.C.,
Fouad, K., Hillyer, J., Weaver, L., Fehlings, M., Tetzlaff, W.
(2010c) A Systematic Review of Directly Applied Biologic
Therapies for Acute Spinal Cord Injury. J. Neurotrauma.
Lammertse, D., Tuszynski, M.H., Steeves, J.D., Curt, A.,
Fawcett, J.W., Rask, C., Ditunno, J.F., Fehlings, M.G.,
Guest, J.D., Ellaway, P.H., Kleitman, N., Blight, A.R.,
Dobkin, B.H., Grossman, R., Katoh, H., Privat, A.,
Kalichman, M. (2007) Guidelines for the conduct of
clinical trials for spinal cord injury as developed by the
ICCP panel: clinical trial design. Spinal Cord. 45:232–242.
Lin, V.W. (2003) Spinal Cord Medicine. Principles and
Practice. New York: Demos Medical Publishing.
Lindsey, A.E., LoVerso, R.L., Tovar, C.A., Hill, C.E.,
Beattie, M.S., Bresnahan, J.C. (2000) An analysis of changes
in sensory thresholds to mild tactile and cold stimuli after
experimental spinal cord injury in the rat. Neurorehabil.
Neural Repair 14:287–300.
Marsh, B.C., Astill, S.L., Utley, A., Ichiyama, R.M. (2010)
Movement rehabilitation after spinal cord injuries:
Emerging concepts and future directions. Brain Res. Bull.
Massey, J.M., Amps, J., Viapiano, M.S., Matthews, R.T.,
Wagoner, M.R., Whitaker, C.M., Alilain, W., Yonkof, A.L.,
Khalyfa, A., Cooper, N.G., Silver, J., Onifer, S.M. (2008)
Increased chondroitin sulfate proteoglycan expression in
denervated brainstem targets following spinal cord injury
creates a barrier to axonal regeneration overcome by
chondroitinase ABC and neurotrophin-3. Exp. Neurol.
209:426–445.
Massey, J.M., Hubscher, C.H., Wagoner, M.R., Decker, J.A.,
Amps, J., Silver, J., Onifer, S.M. (2006) Chondroitinase ABC
digestion of the perineuronal net promotes functional
collateral sprouting in the cuneate nucleus after cervical spinal
cord injury. J. Neurosci. 26:4406–4414.
McTigue, D.M., Tripathi, R., Wei, P. (2006) NG2
colocalizes with axons and is expressed by a mixed cell
population in spinal cord lesions. J. Neuropathol. Exp.
Neurol. 65:406–420.
Mire, E., Thomasset, N., Jakeman, L.B., Rougon, G.
(2008) Modulating Sema3A signal with a L1 mimetic
peptide is not sufficient to promote motor recovery and
axon regeneration after spinal cord injury. Mol. Cell.
Neurosci. 37:222–235.
Nakamura, M., Nakano, K., Morita, S., Nakashima, T.,
Oohira, A., Miyata, S. (2009) Expression of chondroitin
sulfate proteoglycans in barrel field of mouse and rat
somatosensory cortex. Brain Res. 1252:117–129.
Pizzorusso, T., Medini, P., Berardi., N,, Chierzi. S,, Fawcett,
J.W., Maffei, L. (2002) Reactivation of ocular dominance
plasticity in the adult visual cortex. Science. 298:1248–1251.
Protopapas, M.G., Bundock, E., Westmoreland, S., Nero,
C., Graham, W.A., Nesathurai, S. (2007) The complications
of scar formation associated with intrathecal pump
placement. Arch. Phys. Med. Rehabil. 88:389–390.
Scheff, S.W., Rabchevsky, A.G., Fugaccia, I., Main, J.A.,
Lumpp, J.E., Jr. (2003) Experimental modeling of spinal
cord injury: characterization of a force-defined injury
device. J. Neurotrauma. 20:179–193.
Tang, X., Davies, J.E., Davies, S.J. (2003) Changes in
distribution, cell associations, and protein expression levels
of NG2, neurocan, phosphacan, brevican, versican V2, and
tenascin-C during acute to chronic maturation of spinal
cord scar tissue. J. Neurosci. Res. 71:427–444.
Tester, N.J., Howland, D.R. (2008) Chondroitinase ABC
improves basic and skilled locomotion in spinal cord injured
cats. Exp. Neurol. 209:483–496.
Tetzlaff, W., Okon, E.B., Karimi-Abdolrezaee, S., Hill, C.E.,
Sparling, J.S., Plemel, J.R., Plunet, W., Tsai, E., Baptiste,
D.C., Smithson, L.J., Kawaja, M.D., Fehlings, M., Kwon,
B.K. (2010) A Systematic Review of Cellular Transplantation
Therapies for Spinal Cord Injury. J. Neurotrauma.
Wernig, A., Müller, S., Nanassy, A., Cagol, E. (1995)
Laufband Therapy Based on ‘Rules of Spinal Locomotion’
is Effective in Spinal Cord Injured Persons. Eur. J. Neurosci.
7:823–829.
Wessels, M., Lucas, C., Eriks, I., de, G.S. (2010) Body
weight-supported gait training for restoration of walking in
people with an incomplete spinal cord injury: a systematic
review. J. Rehabil. Med. 42:513–519.
White, R.E., Yin, F.Q., Jakeman, L.B. (2008) TGF-alpha
increases astrocyte invasion and promotes axonal growth into
the lesion following spinal cord injury in mice. Exp. Neurol.
Yick, L.W., Cheung, P.T., So, K.F., Wu, W. (2003) Axonal
regeneration of Clarke’s neurons beyond the spinal cord
injury scar after treatment with chondroitinase ABC. Exp.
Neurol. 182:160–168.
Yick, L.W., Wu, W., So, K.F., Yip, H.K., Shum, D.K.
(2000) Chondroitinase ABC promotes axonal regeneration
of Clarke’s neurons after spinal cord injury. Neuroreport.
11:1063–1067.
97
targeting CSPG proteins and their extracellular
interactions is a viable target for therapies that seek to
enhance plasticity and functional recovery.
3) Found that neither intrathecal nor intraparenchymal
chABC, when administered at 7 days post injury, is
sufficient to improve the limited effects of treadmill
training on gross locomotor function after contusive
SCI in rats.
Peer Reviewed Articles:
Jakeman, L.B., Hoschouer, E.L., Basso, D.M.. 2010.
Injured mice at the gym: Review, results and considerations
for combining chondroitinase and locomotor exercise to
enhance recovery after spinal cord injury. Brain Res. Bull.
2010 Jun 15. [Epub ahead of print]
Hoschouer, E.L., Finseth, T., Flinn, S., Basso, D.M.,
Jakeman, L.B.. 2010. Sensory stimulation prior to spinal
cord injury induces post-injury dysesthesia in mice.
J. Neurotrauma. 2010 May;27(5):777–87
Over the coming year, we will complete our histological and
biochemical analyses of the spinal cord tissue samples
collected from the completed behavioral studies. We have
preliminary data confirming that the chABC treatments
exposed 6S stubs in the regions that were targeted at the
lesion epicenter and the lumbar spinal cord. We will extend
these findings to determine if perineuronal nets are lost and
recover by 35 dpi and evaluate the distribution of CSPG core
proteins, GFAP and microglial activation, and sprouting of
afferent and 5-HT fibers in the affected regions. We are
currently completing biochemical analysis of the effects of
treadmill training alone and in combination with IT or MI
chABC on the expression of neurocan and GFAP in the
lumbar spinal cord. In addition, we will perform more
detailed analysis of the step patterns of the treated and
control rats obtained using the Catwalk system, and
kinematic analysis of the ankle joint movements of the rats
during their sessions on the treadmill. Based on the evidence
that intraparenchymal chABC does modify the reflex
response to nociceptive stimulation, we predict that this
treatment group will shown differences in the patterns of
locomotion that are not evident from the BBB scores alone.
Abstracts, Manuscript in preparation:
Andrews, E.M., Yin, Q.F., Viapiano, M.S. and Jakeman,
L.B. 2009. Alterations in chondroitin sulfate proteoglycan
expression both at and distal to the site of spinal cord injury
in rats [Abstract]. Neuroscience Meeting Planner. no. 563.10.
Chicago, IL. (November) (Published)
Richards, R.J., Andrews, E.A., Yin, F.Q., Viapiano, M.S.,
Jakeman, L.B. 2010. The effects of severe spinal cord
contusion injury on regional glial reactivity and CSPG
expression in distant segments of the spinal cord [Abstract].
San Diego, CA, USA: Society for Neuroscience. (October)
(Forthcoming)
Oral Presentations:
Jakeman, L.B., Hoschouer, E.L., Basso, D.M.. 2009.
Injured mice at the Gym: Running wheels and
chondroitinase trials. Presented at Spinal Research Network
Meeting. Glasgow, Scotland, UK. (September 5)
There were no changes in the overall objectives of this
proposal over the two year course of study. The shift from
mouse to rat at the beginning of the second year of funding
was specified in the first year progress report and approved.
In our opinion, this change allowed us to do a much more
thorough test of the initial hypothesis and will provide
much more useful data for future analyses and directions.
There were a few experiments proposed to examine the
histological effects of the combined treatment in the mouse
model that have not yet been completed in order to allow
sufficient time and effort to complete the rat experiments.
Tissues are available to us to allow a direct species
comparison of histological effects of chABC treatment in
the lumbar spinal cord.
FUTURE PLANS
At the completion of this funding period, we have
completed most of the proposed objectives of this proposal.
In summary, we have:
1) Demonstrated that a single injection of chABC
administered to the parenchyma of the lumbar spinal
cord of mice at 7 days after a contusion injury is not
sufficient to improve functional recovery in overground
locomotion when administered alone or combined
with voluntary wheel running.
2) Shown that the expression of the CSPG core proteins,
neurocan and NG2 are upregulated in both gray and
white matter of the spinal cord at segments distal to the
site of a spinal cord contusion injury. This confirms that
98
Comparative evaluation of surgical and pharmacological
methods for removal of a mature scar in a chronic spinal
cord injury model and subsequent regeneration of
stimulated sensory neurons through the treated wound
Daljeet Mahay*, Ann Logan, Martin Berry, Zubair Ahmed & Ana Maria Ginzalez
The University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK
*post doc
[email protected]
INTRODUCTION
Following spinal cord injury (SCI), the terminals of the long
tract axons become arrested in the walls of a dense glial scar
(Sandvig et al., 2004). Astrocytes around the SCI site and
conspire with meningeal fibroblasts to lay down a basal
lamina of a glial limiting membrane and an extracellular
matrix (ECM) comprising of laminin, collagen, fibronectin
and chondroitin sulphate proteoglycans (CSPG). The
CSPG family which includes NG2, brevican, neurocan,
versican, aggrecan and phosphacan are inhibitory to axon
growth and are found at sites of injury (Davies et al., 2004).
In order to achieve functional repair in the chronically
injured patient it will be necessary to remove the established
scar and promote axon regeneration through the CSPG-rich
inhibitory barrier of the wound site. We suggest that
formation of the CSPG-rich glial scar can be acutely
suppressed and chronically removed by targeting
transforming growth factor beta (TGF-β) pro-inflammatory
cytokine and the epidermal growth factor receptor (EGFR),
which have been shown to induce CSPG production,
astrocyte proliferation and ECM remodeling (Asher et al.,
2000; Rabchevsky et al., 1998). Our previous studies have
shown that injection of TGF-β1 into a cerebral cortex lesion
increases the formation of astroglial limitans around the
lesion (Logan et al. 1994) and the expression of TGF-β2 is
up-regulated after SCI (Lagord et al., 2002). Furthermore,
it is known that local delivery of the small leucine rich TGFβ antagonist, dermatan chondroitin sulphate proteoglycans
or Decorin, binds to and inactivates TGF-β ligands
extracellularly and upregulate expression of matrix
metalloproteinases (Akhurst, 2006). In our previous studies
we showed that the delivery of Decorin after a cerebral
cortex stab injury suppressed astrocytic GFAP, laminin and
fibronectin expression (Logan et al. 1999). Davies et al.
(2004) showed that the delivery of Decorin in an acute
spinal cord injury abrogates inflammation and CSPG-rich
scar formation across the lesion site.
neuritogenic activity of Decorin; evaluation of neurite
outgrowth in Decorin-treated primary sensory neurons, and
3) to evaluate the anti-scarring activity of Decorin in vivo;
delivery of recombinant Decorin in an acute dorsal funicilus
lesions (DFL) by injection of, and insertion of collagen
matrices pre-impregnated with recombinant Decorin (Berry
et al. 2001).
METHODS
Preparation of primary adult rat meningeal fibroblasts
Adult male rats (200–250 g) were killed and the spinal
column was removed and submerged in sterile PBS. Spinous
processes were removed and dura mater and arachnoid were
collected from the dorsal surface of the cord and
enzymatically treated with Collagenase XI and Dispase. The
enzyme reaction was stopped and tissue was gently
triturated and the meningeal fibroblasts were seeded.
In vitro – western blot for NG2 expression
TGF-β2 stimulated meningeal fibroblasts were cultured in
conditioned medium from COS cells expressing a gene
construct WT Decorin for 3 days. Protein lysates were
prepared and western blot was undertaken for the expression
of NG2.
In vitro – Dorsal root ganglion neuron (DRGN)
neurite outgrowth assay
DRGN were seeded on a layer of TGF-β2 stimulated
meningeal fibroblasts in the presence of recombinant
Decorin. The cultures were stained for β-tubulin (DRGN)
and NG2 (meningeal fibroblasts) and neurite outgrowth
was measured.
In vivo – DFL lesion
Adult male rats (n=5 per group) received dorsal column
(DC) bilaterally crushed to a depth of 1.5 mm at the T8
level of the spinal cord. Freeze-dried collagen matrices were
implanted at the lesion site and injected with saline
(control), recombinant Decorin (treatment) or lesion alone.
Rats were harvested at day 21 and the cords were sectioned
(15 μm) and immunostained for inflammatory molecules
(ED1 – macrophages) and glial scar deposition (lamimin).
We aim to promote functional repair after chronic spinal
cord injury by promoting axon regeneration and
removing/preventing the scar by administrating Decorin
with and without scar resurrection in the chronically injured
spinal cord. The initial part of this study evaluated the antifibrotic and axogenic activity of Decorin after acute SCI.
Our aims were: 1) to set up an in vitro model to evaluate the
anti-fibrogenic bioactivity of Decorin; by monitoring
expression of NG2 in TGF-β stimulated primary rat
meningeal fibroblasts, 2) to monitor in vitro the
RESULTS
1. The biological activity of the Decorin was monitored by
the detection of suppressed NG2 production by TGF-β2
stimulated meningeal fibroblasts. The meningeal
99
fibroblasts were treated with conditioned medium from
cells expressing wild type (WT) Decorin for 3 days and
immunoblotted for the CSPG, NG2 (Figure 1). The
results showed that Decorin suppressed NG2 expression
in TGF-β2 treated meningeal fibroblasts.
Figure 3. DFL after 21 days with collagen matrices pre-impregnated
with PBS (A) and recombinant Decorin (B) for ED1 (green), laminin
(red) and nuclei (blue) round the lesion site (indicated by * centre of
lesion site and – surrounding the scar) / Image ×100 magnification.
Figure 1. NG2 western blot of TGF-β2 stimulated meningeal
fibroblasts cultured in the presence conditioned medium from COS
cells expressing a gene construct WT Decorin.
2. In vitro data showed the administration of recombinant
Decorin in primary sensory neuron cultures of dorsal
root ganglion neurons (DRGN) cultured on meningeal
fibroblasts pre-treated with TGF-β2 resulted in
increased neurite outgrowth (Figure 2) and support the
reports of an axogenic activity of Decorin.
Figure 4. Quantification of laminin deposition for % coverage of
basal lamina at lesion perimeter (A) and thickness of basal lamina
(B) at the glia limitans accessory (surrounding lesion site) and externa.
CONCLUSION
Our results indicate in vitro Decorin is biological active as
an anti-fibrogenic agent when added either as a
recombinant protein or by gene delivery to primary
meningeal fibroblast cells treated with TGF-β2. Hence, we
have shown that in culture Decorin blocks the TGF-β2
stimulated production of proteogylycans (inhibitory
molecules within scar tissue) by meningeal cells and
supports previous findings (Davies et al., 2004).
Furthermore, our data supports an axogenic activity of
Decorin for primary sensory neurons. Our more recent
work has translated this observation into an acute model of
SCI. The in vivo data shows that delivery of recombinant
Decorin protein to the site of a SCI significantly reduces
acute inflammation and scarring after 21 days. The results
are encouraging as they show that Decorin has the predicted
biological effect and we are now well placed to investigate
further the anti-scarring and axogenic effects of Decorin in
a chronic model of SCI.
Figure 2. Disinhibited neurite outgrowth of DRGN stained with βtubulin (green) seeded on a layer of TGF-β2 stimulated meningeal
fibroblasts stained with NG2 (red) and nuclei (blue) in the presence
of recombinant Decorin. Image ×100 magnification.
3. The administration of Decorin injected in to a collagen
matrix implanted in an acute DFL model after 21 days
resulted in smaller lesion area with suppressed
inflammation and reduced laminin deposition
compared to the PBS control (Figure 3). The coverage
and thickness of laminin deposition was reduced in the
Decorin treated group (Figure 4).
100
REFERENCES
Akhurst, R.J. A sweet link between TGFb and vascular
disease? Nature Genetics 2006. 38: 400–1.
Asher, R.A., Morgenstern, D.A., Fidler, P.S., Adcock, K.H.,
Oohira, A., Braistead, J.E., Levine, J.M., Margolis, R.U.,
Rogers, J.H., Fawcett, J.W. Neurocan is upregulated in
injured brain and in cytokine-treated astrocytes. J. Neurosci.
2000. 20: 2427–38.
Berry, M., Gonzalez, A.M., Clarke, W., Greenlees, L.,
Barrett, L., Tsang, W., Seymour, L., Bonadio, J., Logan, A.,
Baird, A. Sustained effects of gene-activated matrices after
CNS injury. Mol. Cell. Neurosci. 2001. 17(4):706–16.
Davies, J.E., Tang, X., Denning, J.W., Archibald, S.J.,
Davies, S.J. Decorin suppresses neurocan, brevican,
phosphacan and NG2 expression and promotes axon
growth across adult rat spinal cord injuries. Eur. J. Neurosci.
2004. 19(5):1226–42.
Lagord, C., Berry, M., Logan, A. Expression of TGFbeta2
but not TGFbeta1 correlates with the deposition of scar
tissue in the lesioned spinal cord. Mol. Cell. Neurosci. 2002.
20: 69–92.
Logan, A., Baird, A., Berry, M. (1999) Decorin attenuates
gliotic scar formation in the rat cerebral hemisphere. Exp.
Neurol. 159: 504–10
Logan, A., Green, J., Hunter, A., Jackson, R,, Berry. M,
(1999) Inhibition of glial scarring in the injured rat brain by
a recombinant human monoclonal antibody to
transforming growth factor-beta2. Eur. J. Neurosci. 11:
2367–74
Rabchevsky, A.G., Weinitz, J.M., Coulpier, M., Fages, C.,
Tinel, M., Junier, M.P. A role for transforming growth factor
alpha as an inducer of astrogliosis. J. Neurosci. 1998. 18:
10541–52.
Sandvig, A., Berry, M., Barrett, L.B., Butt, A., Logan, A.
Myelin, reactive glia-, and scar-derived CNS axon growth
inhibitors: expression, receptor signaling, and correlation
with axon regeneration. Glia. 2004. 46(3):225–51.
Manuscript in preparation
‘Delivery of Decorin in a collagen matrix suppresses
inflammation and scar formation after an acute dorsal
column lesion’
FUTURE PLANS
Our future aims are to (i), degrade acute scars in DFL by
injection of, and insertion of collagen matrices preimpregnated with vesicular stomatitis virus glycoprotein
(VSV-G) pseudotyped lentiviral vector (LV) encoding the
decorin gene; and (ii), convert a chronic DFL into an acute
wound by surgical resection of the scar and prevent new
cavitation and scar deposition using a hydrogel implant
coupled with the decorin strategy described in (i) above.
Unfortunately, our planned collaborator, Dr Stephen
Davies, has been unable to work with us on this project.
However, we have sourced our own supply of recombinant
Decorin from Catalent Pharma Solutions and prepared our
own lentivirus construct. This has slightly impacted on the
timelines but these should be recovered during the course of
the project.
101
Do experimental treatments for spinal cord injury induce
functional plasticity in spared pathways?
John Riddell & Susan Barnett
University of Glasgow, UK
[email protected]
INTRODUCTION
Many spinal cord injuries are incomplete, variable numbers
of spared fibres passing the lesion level and supporting some
residual function below the injury. One approach to
improving function following injury is to develop therapies
that maximise the potential of these spared fibres. We have
recently shown that OECs transplanted into a spinal cord
lesion improve the function of spinal cord circuitry in the
region adjacent to the lesion and thereby maximise
transmission in spared ascending pathways projecting to the
sensorimotor cortex (Toft et al. 2007). The mechanism
underlying these effects is not clear but one possibility is
that soluble factors released by the transplanted cells may
induce collateral sprouting or that regenerating axons
growing into the transplant receive some contact mediated
signal that stimulates axonal branching in lengths of axon
outside the transplant. Some evidence that OECs might
promote the sprouting of spared fibres has been reported by
Chuah et al. (2004). Following a lesion of the main dorsal
column component of the corticospinal tract, the minor
component projecting through the ventral white matter was
found to produce more collateral branches near OECtransplanted lesions than near lesions without transplants.
However, the density of terminal boutons was not
determined in this study. Chondroitinase treatments induce
increased anatomical plasticity in the nervous system,
including of damaged and spared fibre systems in animal
models of spinal cord injury. In some cases, this enhanced
anatomical plasticity has been observed in parallel with
improved recovery of function assessed behaviourally
(Bradbury et al. 2002). Sprouting of the corticospinal tract
following a cervical dorsal column lesion and chondroitinase
treatment occurs both above (presumably from damaged
fibres) and below the lesion site from spared corticospinal
fibres travelling in non-dorsal column white matter (Barritt
et al. 2006). Although chondroitinase treatment has been
shown to increase the number of labelled axon collaterals it
remains to be shown that this results in additional synaptic
connections and most importantly, it remains to be
demonstrated that they produce functionally useful
strengthening of corticospinal actions within the spinal
cord. The primary purpose of this project is to determine
whether olfactory ensheathing cell (OEC) transplants and
chondroitinase treatments improve function after spinal
cord injury by promoting functional plasticity. An
electrophysiological approach will be used to determine
whether OEC and chondroitinase treatments induce
functional plasticity in descending motor and ascending
sensory spinal pathways following injury and the relevance
of this to improved sensorimotor function will then
be investigated.
Specific aims: The project is divided into 3 main stages:
Stage 1. To develop electrophysiological assays of functional
changes in sensory motor pathways and to use these to
investigate the degree of spontaneous plasticity that occurs
in an ascending and descending pathway following a dorsal
column injury
Stage 2. To investigation whether OEC transplants modify
the response of an ascending and descending pathway to a
dorsal column injury by promoting sprouting or other
mechanisms of plasticity that enhance their activity.
Stage 3. To investigate whether delivery of chondroitinase
modifies the response of an ascending and descending
pathway following spinal cord injury by promoting
sprouting or other mechanisms of plasticity that enhance
their activity.
METHODS
Cell culture. For Stage 2 of the project, completed last year,
OECs were prepared from the olfactory bulbs of P7 rats
using a FACS purification step and cultured for
approximately 3 weeks prior to transplantation (Toft et al.
2007). Cells were modified using lentiviral infection to
express GFP. Cultures prepared for transplantation consisted
of pure p75 expressing cells and 9–90% also expressed GFP.
For Stage 3 of the project we aim to transplant OECs that
have been engineered to secrete chondroitinase. The aim of
this is provide sustained delivery of the enzyme and to do so
at a site within the spinal cord so as to reduce the problems
associated with tissue penetration. These cells will therefore
be infected in culture with a ChASE-A,C lentivirus
(provided by Dr George Smith).
Lesions and injuries. The SCI experiments are
performed on adult rats. For Stages 1 and 2 of the project
(completed in year 1) an inbred strain (F344) was used to
allow syngeneic transplantation of OECs and obviate the
need for immunosuppression. For Stage 3 of the study
(ongoing) F344 animals will used where cell transplantation
is involved, otherwise an outbred strain is preferred. For
Stages 1 and 2 of the project, lesions of the spinal dorsal
columns were made at a mid-cervical level (C4–5) using a
wire knife. This interrupts the ascending collaterals of
primary afferent fibres entering below the lesion and
descending fibres forming the main component of the
corticospinal tract. Those animals in which the effects of
OEC transplants on plasticity were to be investigated
received transplants of OECs immediately after injury. These
were made by pressure injection through a fine pipette
introduced into the lesion site. The cells were transplanted so
that they were distributed above and below the lesion rather
than filling the injury site. For stage 3 of the project we will
use an alternative model. During the second year of the
102
project we obtained a new piece of apparatus (Infinite
Horizon Impactor) that allows consistent production of
contusion injuries of a chosen and controllable severity. Since
this type of injury better reflects clinical injuries to the spinal
cord, we have adopted this for Stage 3 of the project and
developed a C6 injury that is compatible with our
electrophysiological outcome measures and with grip
strength testing of forelimb function (see below).
surrounding the stimulated pyramids. In order to investigate
plasticity in the corticospinal projection we first had to
design an approach that would provide reliable
quantification of the synaptic actions of the corticospinal
projection. Maximal unilateral activation of a pyramid was
found to be achievable if a bipolar stimulating electrode was
used and provided the tip of this was accurately positioned
within the target pyramid. It proved possible to determine
correct positioning during the experiment by collecting a
stimulus-response curve using graded increases in stimulus
intensity. By subsequently matching this to the histologically
verified position of the electrode track, a curve characteristic
of correct placement at the centre of the pyramid was
identified as were curves indicative of placement too near the
midline. Fig. 1. illustrates the method we have developed.
Electrophysiology. To assess the function of sensory
pathways and descending pathways, the ascending dorsal
column pathway and the corticospinal tract are electrically
stimulated at maximal intensity and recordings of cord
dorsum potentials made from the surface of the spinal cord.
CDPs are postsynaptic potentials which provide a measure
of the strength of connections between sensory afferents and
spinal cord neurones. The radial nerve is exposed and
mounted on bipolar stimulating electrodes in a paraffin pool
formed by adjacent tissues. The nerve is stimulated at
maximal A-beta fibre strength and cord dorsum potentials
(CDPs) recorded from the surface of the spinal cord for 8
mm above and below the lesion. To assess a descending
system, the corticospinal tract is activated electrically via
electrodes placed stereotaxically within the pyramids. Cord
dorsum potentials evoked by corticospinal volleys are again
recorded for 8 mm above and below the lesion in order to
assess the strength of corticospinal connections in the region
of the injury. Similar electrophysiological methods have
been used in all three Stages of the project.
Forelimb behavioral tests. To obtain functional
information on forelimb function to complement the
electrophysiological assessment of neural function will
introduce behavioral tests for Stage 3 of the project. Initially
we will use the grip strength test which we have found to
demonstrate persistent deficits in a 175kdyne C6 contusion
injury. One of the aims in the final year of the project will
be to develop a more sophisticated method of assessing
forelimb function based on kinematic analysis of reaching
and grasp.
Figure 1. Electrical stimulation of the pyramids. A. histological
verification of the correct positioning of a stimulating electrode in the
right pyramid. Scale bar=0.7 mm; B. schematic diagram of the
outline of the section and stimulating electrode track reconstructed
using Camera lucida drawings. Scale bar=0.7 mm; C. CDPs evoked
by stimuli of graded intensity applied to the pyramids using the
electrode which made the track shown in A and B; D, stimulusresponse curve obtained using the records illustrated in D.
RESULTS
First year
Before investigating the plasticity inducing potential of
OEC transplants and chondroitinase treatment (Stages 2 and
3), we aimed first to determine the background level of
change that occurs in the absence of plasticity promoting
treatments. We therefore used electrophysiological methods
(recording of cord dorsum potentials) to investigate
spontaneous plasticity in corticospinal and sensory fibre
systems following a dorsal column lesion at the C3/4 level
(see Fig. 2.). The function of these pathways, above and
below the level of the lesion was compared in normal animals
and at 1 week and 3 months after a dorsal column lesion. A
modest enhancement of transmission in both corticospinal
and sensory systems occurred following the dorsal column
lesion. Plasticity in the corticospinal projection occurred
both above the lesion, at the intact connections formed by
fibres axotomised more distally and also below the lesion at
connections formed by the spared fibres of the minor nondorsal column components of the corticospinal tract.
Stage 1 – Development of electrophysiological tests and
assessment of spontaneous plasticity following a dorsal column
lesion
The main aim of the project is to assess plasticity in both the
corticospinal (descending) and sensory (ascending) pathways
of the spinal cord using electrophysiology to assess changes
in the function of these pathways. To make valid quantitative
comparisons of the amplitudes of CDPs evoked by spinal
cord pathways and recorded in different animals it is
necessary to be able to activate the appropriate pathway
maximally, while at the same time avoiding stimulus spread
to adjacent pathways. This is relatively easily achieved for the
radial nerve which can be isolated for stimulation. However,
it is more difficult to achieve for the corticospinal tract where
there is the risk of spread of current from stimuli applied in
the pyramid on one side to the pyramid on the other and
also to spinally projecting neurons in the reticular formation
103
Stage 2 – Plasticity following OEC transplants
Transplants of OECs into a lesion cavity are thought to have
a neuroprotective action which can be detected
electrophysiologically (Toft et al. 2007). In order to test
whether OECs are also able to induce plasticity in
sensorimotor pathways we therefore aimed to transplant
cells so that they were distributed within the spinal cord
above and below the lesion but not within the lesion site
itself. Fortuitously, we found that following injection into a
cervical dorsal column lesion OECs spread caudally for
more than 6 mm along the dorsal columns but did not
remain within the lesion cavity (Fig. 4. shows examples of
caudally distributed OECs.
Figure 2. Schematic diagram showing recording locations for
CDPs and examples of CDPs. Cord dorsum potentials (CDPs)
evoked by electrical stimulation of the radial nerve and/or corticospinal
tract were recorded to assess spinal cord function in the region of the
lesion. Recordings were made at 1 mm intervals for 8 mm above and
below the C4/5 reference level (the lesion site in the lesioned animal
group) (A). Typical radial nerve evoked and corticospinally-evoked
CDPs were shown in B.
The time-course of plasticity in these two systems
differed. Plasticity at the connections of the axotomised fibres
above the lesion was fully developed within 1 week while
plasticity in the spared fibres below the lesion was seen at
3 months but not at 1 week. Fig. 3. summarises these
observations on corticospinal tract function. A modest
enhancement of the strength of connections formed by large
diameter sensory fibres in the radial nerve was also seen below
the level of the dorsal column lesion. This had a similar time
course to the plasticity of corticospinal connections above the
lesion occurring within 1 week of the injury.
Figure 4. Distribution of OECs 3 months after injections into
the lesion site. A. the cells were transplanted into the lesion site
immediately after lesioning. They were observed caudally for more
than 6 mm along the dorsal columns but did not remain within the
lesion cavity. B. confocal images from three transplanted animals show
the distribution of OECs from 2 mm above the lesion to 6 mm below
the lesion. GFP positive percentage were labelled.
Animals transplanted with cells in this way were therefore
used to investigate the effect of OECs on the function of
sensory afferents terminating in this region. Since the cells
did not spread rostrally in significant numbers, to investigate
the effect of OECs on the function of corticospinal fibres
terminating in this region, OECs were injected into the
dorsal columns at a depth corresponding to the corticospinal
tract at between 4 and 5 sites between the lesion and 4 mm
more rostrally. A further group of animals was injected with
similar volumes of medium as a control. Electrophysiological
methods were then used as above to investigate whether
transmission in the corticospinal and sensory fibre systems
following a dorsal column lesion was improved in
transplanted animals compared to 3 month survival animals.
However, corticospinal actions rostral to the lesion were not
enhanced by OEC transplants above the lesion and sensory
transmission caudal to the lesion was not enhanced by cells
below the lesion (Fig. 5. shows the results obtained for
sensory transmission). OEC transplants are therefore
unlikely to support recovery by promotion of sprouting.
Figure 3. Plot showing the amplitudes of pyramidal-evoked
CDPs recorded in different animal groups. Each data point is the
mean CDP amplitude for all animals from this group (15 three
months survival lesioned animals). The error bars show +/− SEM.
The recordings were made over the cervical spinal cord. Recording
positions are shown relative to the C4/5 border (0 mm) where dorsal
column lesions were made in lesioned animals. In the three months
group, CDPs below the lesion were much smaller than in normal
animals, they showed a tendency to be larger at all recording locations
than those recorded in animals one week after a dorsal column lesion.
This difference was significant at all recording sites except the closest
to the lesion. Above the level of the lesion, CDPs were significantly
larger than in normal animals and virtually the same as those recorded
in animals one week after a dorsal column lesion.
104
chondroitinase and then transplant them into the injury site
so that they act continually as cell factories producing the
chemical within the spinal cord. We have experimented
with the use of different viruses designed to modify OECs
so that they are able to make and release chondroitinase.
During initial attempts we encountered some technical
difficulties. Although OECs were shown to produce
chondroitinase either the virus or the infection product was
found to stunt the proliferation and expansion of the cells
in culture. However, a modified lentivirus appears to have
overcome this problem so that OECs are consistently
transduced to secrete chondroitinase into the medium but
now also proliferate normally. Cells prepared using this virus
will therefore be used in the next set of experiments to see
whether, with this more effective delivery, there is evidence
of functional plasticity. In addition, we have begun to
develop a new method for quantitative assessment of
forelimb function in rats, based on analysis of movements
made during reaching (kinematic analysis). This will be used
to assess the functional consequences of any plasticity
revealed by the electrophysiology.
Figure 5. Plots showing the amplitude of radial nerve evoked
CDPs recorded in different animal groups. Plots show distributions
of the amplitudes of CDPs recorded over the cervical spinal cord in
response to radial nerve stimulation. Recordings were made from
normal animals (black), three months survival animals (blue), and
lesioned animal with OEC transplanted into the lesion sites (red). CDP
amplitudes are averaged for all animals in each group, each data point
showing mean +/− SEM. Recording positions are shown relative to the
C4/5 border (0 mm) where dorsal column lesions were made.
Comparison of results obtained from normal animals (black line) and
animals with a dorsal column lesion but no transplants (blue line)
shows that the lesion substantially reduces the radial nerve-evoked CDPs
above the level of the lesion as expected. Compared to normal animals,
CDPs below the lesion in three month survival animals were larger.
However, comparison of the results obtained from animals with OECs
injected below the lesion (red line) and those with a lesion and no
transplant (blue line) suggest that OECs do not enhance the plasticity
of radial-nerve fibre terminations, but at the same time do not have any
detrimental effect on sensory function.
CONCLUSION
The Results obtained in Stages 1 and 2 of the project show
that corticospinal fibres terminating above the level of the
injury (but most likely damage more distally by the dorsal
column lesion) show rapid spontaneous plasticity. Spared
corticospinal fibres extending below the lesion also show
plasticity but over a longer time course. This spontaneous
plasticity does not appear to be enhanced by transplants of
OECs. We have previously shown (Toft et al. 2007) that
OECs, when transplanted into a dorsal column lesion at the
lumbar level, fill the lesion cavity and result in enhanced
sensory transmission in the region of the injury compared to
non-transplanted controls. In this work, we were not able to
differentiate between a mechanism involving sprouting and
one involving neuroprotection. The current findings suggest
that the enhanced sensory transmission produced by OEC
transplantation involves a neuroprotective mechanism. The
neuroprotective action most likely involves preservation of
local axon collaterals from sensory fibres as a result of
minimizing the die back of dorsal column fibres that
normally occurs progressively over several weeks following
a wire knife injury. Initial Results obtained in Stage 3 of the
project suggest that a single acute delivery of chondroitinase
enzyme into the spinal cord above a contusion injury is
not sufficient to induce detectable changes in the function
of sensory and descending pathways terminating within
the treated region. We anticipate that by transplanting
cells engineered to secrete chondroitinase a more
sustained delivery will be achieved and this is more likely to
induce functional plasticity and significantly improved
functional outcome.
Second year
Stage 3 – investigation of functional plasticity induced by
chondroitinase treatment.
During the course of the first year we were successful in
obtaining funding for an Infinite Horizon impactor device
for making controlled and reproducible contusion injuries
of the spinal cord in rodents. We have now used this device
to developed a contusion model at the cervical level and we
have investigated the effect it has on the corticospinal and
sensory pathways assessed using our electrophysiological
approach. Since this type of injury represents a more
relevant model for investigations of spinal cord injury, we
have incorporated this into the work for Stage 3. As a first
approach, we have injected chondroitinase directly into the
spinal cord at the time of injury, at numerous sites above
and below the injury. Electrophysiology was then used to
assess the function of sensory and descending pathways and
the animals ability to grip with their forepaws (grip strength
test) was also measured for several weeks after the injury.
No difference was seen between animals receiving
chondroitinase treatment and non-treated animals This
suggests either that chondroitinase delivered by this method
is not sufficiently effective to result in detectable plasticity
or that the sprouting of nerve fibres seen after
chondroitinase treatment does not lead to functional
connections. The next series of experiments will address
these two possibilities. A more effective way of achieving
delivery of chondroitinase so that it has a sustained action
in the spinal cord would be to engineer OECs to express
REFERENCES
P., Yip, P., McMahon, S.B. & Bradbury, E.J. (2006)
26, 10856–10867.
105
FUTURE PLANS
The aims for the third year of the project are to:
1) Finish writing up the results of Stages 1 and 2 for which
experimental work is now complete.
2) Investigate the effect of using transplants of cells
engineered to secrete chondroitinase to provide a
sustained delivery, and to test the effects of this on
functional plasticity at the spinal connections of
ascending dorsal column and descending corticospinal
fibres.
3) Develop methodology based on 3D kinematic analysis
of reaching and grasping for assessing functional
improvements in forelimb use that can be used to
evaluate functional outcome in this and future studies.
Bennett, G.S., Patel, P.N., Fawcett, J.W. & McMahon, S.B.
(2002) Chondroitinase ABC promotes functional recovery
after spinal cord injury. Nature. 416: 636–640.
Bradbury, E.J. & McMahon, S.B. (2006) Spinal cord repair
strategies: why do they work? Nature. vol 7: 644–653
Chuah, M.I., Choi-Lundberg, D., Weston, S., Vincent,
A.J., Chung, R.S., Vickers, J.C. & West, A.K. (2004)
Olfactory ensheathing cells promote collateral axonal
branching in the injured adult rat spinal cord. Exp. Neurol.
185:15–25.
Toft, A., Scott, D.T,, Barnett, S.C. & Riddell, J.S. (2007)
Electrophysiological evidence that olfactory cell transplants
improve function after spinal cord injury. Brain. 130:
970–984.
During the second year of the project we acquired new
apparatus (Infinite Horizons Impactor) that allows
consistent production of contusion injuries of a chosen and
controllable severity. This type of injury better reflects
clinical injuries to the spinal cord. We have therefore
developed a cervical contusion model and tested the
suitability of our electrophysiological outcome measures for
use in this model. Having determined that a 175kilodyne
injury was compatible with the electrophysiological
assessments and with grip strength testing of forelimb
function, we have used this model for Stage 3 of the project
in which we are investigating functional plasticity induced
by chondroitinase treatment.
The Results of Stages 1 and 2 of the project were presented
at the joint meeting of Spinal Research, The Christopher
and Dana Reeve Foundation and The International
Institute for Research in Paraplegia, in Ittingen, Switzerland,
September 2010.
106
Axonal Regeneration in the Chronically Injured Spinal
Cord
Mark Tuszynski & Ken Kadoya
University of California, San Diego
[email protected]
INTRODUCTION
Despite advances in promoting axon regeneration after
acute spinal cord injury (SCI), elicitation of axon
regeneration in chronic SCI remains a formidable challenge.
The overall purpose of this project is to increase the number
and length of regenerating sensory axons when
combinatorial therapies are applied in models of chronic
SCI. In Aim 1, we tested 15 months delayed treatment with
combinatorial therapies consisting of bone marrow stromal
cell (MSC) grafts, peripheral nerve conditioning lesions
(CLs), and neurotrophic factors beyond the lesion site.
Successful bridging sensory axon regenerations were
achieved only when the combinatorial therapies were
applied (Kadoya et al., 2009). However, the extent of this
axon regeneration was still significantly reduced compared
to acutely treated subjects.
Second lesions: To make second injury of chronically
injured dorsal column sensory axons at proximal to cell
bodies, DCLs were made at C7 at 6 weeks, 4 months, or
15 months after C3DCL. CLs were applied 1 week before
C7DCLs and MSCs were grafted at the same time of DCL.
Four weeks after second lesions and treatments, subjects
were perfused.
Labeling of dorsal column sensory axons: Three days
before perfusion, dorsal column sensory axons were traced
transganglionically by injections of CTB into sciatic nerves
(2 μl, 1%).
Immunohistochemistry and quantification: Dissected
spinal cords were sectioned sagittally at 30 μm and subject to
CTB, GFAP, phosphacan, brevican, neurocan, NG2, ED1
immunohistochemistry. Quantification of regenerating
sensory axons in the graft were performed by counting CTB
labeled axons crossing the line placed in the middle of the
graft. Dissected L4 and L5 DRGs were sectioned at 10 μm
and subject to double fluorescent immunostaining of NF200
and MAP2. Cell body size of NF200 positive neurons and
percentage of NF200 positive neurons per MAP2 neurons
were quantified.
Therefore, in Aim 2, we attempted to Identify multiple
mechanisms underlying, and potentially limiting, axonal
growth at prolonged time points after SCI. We examined
effects of acute and delayed treatments consisting of CLs
and MSC grafts on sensory axon regeneration, and
investigated the time course of intrinsic and extrinsic factors
influencing axonal regeneration in chronically injured
neurons up to 15 months after SCI. We assessed glial and
extra-cellular scar formation, inflammatory cell infiltration,
axonal retraction, survival and atrophy of injured neurons,
and neurite outgrowth ability from neuronal cell bodies.
Additionally, we explored how second injuries placed
proximal to cell bodies affect regeneration of chronically
injured axons to determine the relative contribution of
established extrinsic factors in chronic lesion sites to reduced
regeneration capacity in chronic SCI.
Neurite outgrowth of chronically injured sensory
neurons: After dissecting L4 and L5 DRGs 15 months after
C3DCLs and 1 week after CLs, sensory neurons were
dissociated and cultured for 48 hours on myelin substrate.
Control groups were lumbar DRG neurons from intact
subjects, subjects received CLs 1 week prior dissection, or
subjects underwent C3DCLs 15 months prior to dissection.
Cultured neurons were stained with NF200 and longest
neurite outgrowth of each neuron was measured.
METHODS
Lesions: Adult Fischer 344 rats underwent C3 DCLs using
a tungsten wire knife (Kadoya et al., 2009). Subjects were
perfused with 4% PFA at 1 week, 6 weeks, 6 months and
15 months time points after C3 DCL.
RESULTS
Chronically injured sensory axons regenerate into
spinal cord lesion sites; fewer axons regenerate into
chronic than acute injuries.
When MSCs were grafted at the same time as SCI, few
dorsal column sensory axons regenerated into grafts in the
lesion site (Fig. 1B). In contrast, when CLs were applied
1 week before C3DCL and MSCs were grafted at the same
time as injury, robust axonal regeneration was observed into
the graft (Fig. 1C). However, when this treatment was
delayed 6 weeks after injury, axon regeneration into the graft
occurred but was reduced in extent compared to acutely
treated subjects (Fig. 1D and G). When treatments were
delayed to 4 months after injury or later, few axons
regenerated into grafts containing MSCs alone (Fig. 1E–G).
This result is consistent with the result of Aim 1, which
tested the combinatorial therapy consisting of CLs, MSC
Treatments: One week before C3DCL or 6 weeks,
4 months and 15 months after C3 DCL, subjects
underwent crush of bilateral sciatic nerve CLs. One week
later, C3DCLs were made or chronic lesion sites were reexposed and syngenic MSCs were grafted into the lesion
cavity (Fig. 1A). The reason why MSCs were grafted in
lesion cavity is that physical substrate is required to occupy
the lesion cavity and to induce axon regeneration. In
chronically injured subjects, the peri-lesion scar was not
resected. Four weeks after the MSC graft, animals were
perfused with 4%PFA.
107
grafts and NT-3 gradient in 15 months old chronic C3
DCL. Only full combinatorial treatment group revealed
bridging sensor axon regeneration. But, the group of CLs
and MSC did not demonstrate more axonal growth into the
graft than the MSC alone group.
chronic lesion sites 6 weeks after injury or later. Thus, ED1
cells infiltrate into acute lesion sites but do not remain in
chronic lesion sites.
Figure 2. GFAP and CSPG labeling persist at the lesion
boundary from 6 weeks to 15 months after injury. Sagittal
sections of C3 lesion site. Rostral left, caudal right. Immunolabeling for
GFAP (A–D and a–d), phosphacan (E–H and e–h), brevican (I–L
and i–l), and NG2 (M–P and m–p). Lower case letters indicate
high magnification of boxed areas of same capital letters.
Immunoreactivities of these markers up-regulated around lesion sites
1 week after injury, surrounded the lesion cavity 6 weeks after injury,
and persisted 15 months after injury. Scale bar A–P = 500 μm; a–p
= 100 μm.
Figure 1. Sensory axon regeneration into graft is reduced when
CLs and cell grafts are delayed. (A) Experimental design. (B–E)
Sagittal sections of C3 lesion/graft site. CTB-labeled sensory axons
approaching lesion/graft site. Rostral left, caudal right. g, graft; h, host;
dashed lines, graft/lesion border. Scale bar 100 μm. While graft alone
showed few axons in the graft (B), acute treatment demonstrated
robust axonal growth (C). Six weeks delayed treatment exhibited
moderate axonal growth (D), but 4 months delayed (E) and
15 months delayed treatment (F) revealed little axonal growth into the
graft. (G) Quantification of the number of axons in the graft.
(ANOVA p < 0.001; post hoc *p < 0.01 to all other groups,
**p < 0.01 to all other groups.)
Axons remaining adjacent to the lesion site
significantly decline as a function of time after SCI.
One week after C3DCL, most CTB-labeled sensory axons
are located close to the lesion site, whereas 6 months and
15 months after injury, a fraction of CTB-labeled axons
have undergone retraction from the C3 lesion site. In these
groups, some endbulbs form in the dorsal columns several
mm caudal to the original injury site. Quantification
demonstrated that the total number of CTB-labeled axons
within 200 μm of the lesion boundary was significantly
reduced in subjects 15 months after injury compared to
1 week after injury. Thus, sensory axons exhibit significant
retraction from a site of SCI over time.
GFAP and CSPG immunolabeling persist at the lesion
boundary from 6 weeks to 15 months after injury.
Next, we explored multiple factors limiting axonal growth
in chronic SCI using this lesion and treatment model.
Because scar formation consisting of reactive astrocytes and
inhibitory CSPGs are well known for inhibiting axonal
growth especially in chronic SCI (Houle and Tessler, 2003;
Busch and Silver, 2007), the time course of deposition of
these markers was investigated. Immunoreactivity for GFAP
and CSPGs (phosphacan, brevican, neurocan, and NG2)
densely surround lesion sites 6 weeks after injury, and persist
15 months after injury (Fig. 2). Thus, scar tissue consisting
of reactive astrocytes and inhibitory extra-cellular matrix is
established by 6 weeks post-injury and persists over
extended time periods.
Chronically injured sensory neuronal cell bodies
survive 15 months after injury.
We investigated whether chronically injured neurons
atrophy or die by quantify NF200 positive neurons in L4
and L5 DRGs, which send dorsal column axons in cervical
spinal cord. There was no difference of the area of NF200
positive neurons and the percentage of NF200 positive
neurons per MAP2 positive neurons between subjects at
15 months after C3 DCL and intact subjects (Fig. 3). These
results suggest that injured sensory neurons survive and do
not atrophy 15 months after C3 DCL.
There are few ED1 positive cells in chronic lesion sites.
Infiltration of inflammatory cells was investigated by
immunolabeling for ED1(activated macrophages). One
week after injury, frequent ED1 positive cells were observed
in he lesion site, whereas few ED1 cells were detected in
108
Second injury and treatments can partially restore
reduced axon regeneration in chronic SCI.
The above data suggest that the intrinsic growth ability of
chronically injured sensory neuronal somata is maintained,
and that established extrinsic inhibitory conditions
surrounding injured axons and axonal retraction persists in
chronic SCI. We then tested the hypothesis that chronically
injured sensory axons can regenerate more extensively when
a re-lesion of the spinal cord is placed that lacks a chronic
inhibitory scar. To test this hypothesis, second DCLs were
made at C7 at 6 weeks, 4 months or 15 months after initial
C3DCL. Treatments consisting of CLs and MSC grafts were
applied at the same time that C7DCLs were made (Fig. 5A).
In this second C7 lesion site, there was no established glial
scar and little axonal retraction. Control subjects had no prior
C3DCL and received the same treatment at the C7DCL site.
Figure 3. Chronically injured sensory neurons survive and do
not atrophy 15 months after injury. Lumber DRG sections stained
for NF200 and MAP2 from intact subject (A) and subject 15 months
after C3 DCL (B). (C) Quantification of percentage of NF200
neurons per MAP2 neurons. (D) Quantification of cell body size of
NF200 positive neurons. There is no statistical difference of these
quantifications between subjects 15 months after C3 DCL and intact.
Sensitivity of sensory neurons to CLs is maintained
15 months after SCI: Sensory neurons from subjects
15 months after C3DCL revealed the same extent of neurite
outgrowth compared to sensory neurons from intact subjects
(Fig. 4C and E). In addition, sensory neurons from subjects
that underwent CLs 15 months after C3 DCL also exhibited
enhanced neurite outgrowth compared to neurons from
intact rats that was equal in degree to animals that underwent
CLs alone (Fig. 4D and E). These results suggest that the
intrinsic growth ability and sensitivity to CLs of chronically
injured sensory neuronal cell bodies are maintained.
Figure 4. Neurite outgrowth capacity and sensitivity to CLs of
injured neurons are still maintained 15 months after injury. (AD) Lumbar DRG neurons were cultured on myelin for 48 hours and
labeled with NF200. Neurons from naïve controls extend shorter
neurites (A) than CLs 1 week prior to isolation (B). Dorsal column
lesions 15 months prior to isolation have no influence on neurite
extension (C), whereas CLs applied 15 months following C3 lesions
significantly enhance neurite extension (D). Scale bar A–D = 50 μm.
(E) Quantification of neurite outgrowth as mean percentage of control
(A) indicates that both CL and CL 15 months after C3 lesions
significantly enhance neurite outgrowth on myelin compared to intact
animals and animals that received only C3 lesions (ANOVA p<0.0001;
post-hoc Fisher’s * p<0.001 to all other groups).
Figure 5. Second injury and treatments can partially restore
reduced axon regeneration in chronic SCI. (A) Experimental
design. (B–E) Sagittal sections of C7 lesion/graft site. CTB-labeled
sensory axons approaching lesion/graft site. Rostral left, caudal right.
g, graft; h, host; dashed lines, graft/lesion border. Scale bar 100 μm.
Acute treatment (B) and 6 weeks delayed second lesion and treatment
(C) demonstrated robust axonal growth into the C7 lesion site. Four
months delayed (D) and 15 months delayed second lesion and
treatment (E) exhibited moderate axonal growth. (F) Quantification
of the number of axons in the graft per subject. (ANOVA p < 0.001;
post hoc *p < 0.05 to 4 m and 15 m delayed groups).
109
Notably, re-lesioning and treatment at C7 resulted in a
368% increase in the number of axons penetrating the graft
in the lesion site, compared to animals treated at the chronic
C3 site of injury (Fig. 6). These findings indicate that axonal
retraction, the chronic scar, or both, contribute to the
reduced growth capacity of chronically injured axons.
REFERENCES
Busch, S.A., Silver, J. (2007) The role of extracellular matrix
in CNS regeneration. Curr. Opin. Neurobiol. 17:120–127.
Houle, J.D., Tessler, A. (2003) Repair of chronic spinal cord
injury. Exp. Neurol. 182:247–260.
Kadoya, K., Tsukada, S., Lu, P., Coppola, G., Geschwind,
D., Filbin, M.T., Blesch, A., Tuszynski, M.H. (2009)
Combined intrinsic and extrinsic neuronal mechanisms
facilitate bridging axonal regeneration one year after spinal
cord injury. Neuron. 64:165–172.
Kadoya, K., Lu, P., Blesch, A. and Tuszynski, M.H. “Axons
retain an ability to regenerate beyond a lesion site when
treated one year after injury, but from a reduced total axonal
pool,” Society for Neuroscience 39th Annual Meeting,
Chicago, IL, October 2009.
Kadoya, K., Tsukada, S., Lu, P., Coppola, G., Geschwind,
D., Filbin, MT., Blesch, A., Tuszynski MH (2009)
Combined intrinsic and extrinsic neuronal mechanisms
facilitate bridging axonal regeneration one year after spinal
cord injury. Neuron 64:165–172.
Kadoya, K. and Tuszynski, M.H. “Multiple intrinsic and
extrinsic factors restrict sensory axon regeneration in chronic
spinal cord injury,” Society for Neuroscience 40th Annual
Meeting, San Diego, CA, November 2010.
Figure 6. Summary of 15 months delayed manipulations. This
quantification summarizes acute and 15 months delayed
manipulations to injured sensory axons. While acute treatments could
induce robust sensory axon regeneration in the graft (left column),
15 months delayed treatments had little effect on axonal growth
(middle column). In contrast, when second C7 injury and treatments
were applied 15 months after initial C3 injury, the number of axons
regenerating into the graft significantly increased compared to
15 months delayed treatments (right column). However, these
manipulations failed to restore growth to acute treatment levels.
FUTURE PLANS
We will continue to identify mechanisms underlying the
inhibitory nature of chronic SCI and attempt to determine
factors critical for axonal growth in chronic stages of injury.
One approach is investigating gene expression profiles in
both acutely and chronically injured sensory neurons to
determine whether intrinsic growth programs of injured
sensory neurons change at prolonged time points after
injury. Another approach is to test the hypothesis that
chronically injured axons sprout below injury and this
sprouting reduces intrinsic regenerative capacity. Further,
we will test an expanded set of combination therapies
including chondroitinase abc for sensory axon regeneration
in chronic SCI.
CONCLUSION
(1) Chronically injured sensory axons exhibit a persistent
ability to regenerate when subjected to combinatorial
treatments.
(2) Regeneration into a graft is reduced when CLs and cell
grafts are delayed, compared to acute SCI.
(3) Scar tissue consisting of reactive astrocytes and CSPGs
densely surrounds the lesion site 6 weeks after injury
and persists 15 months after SCI.
(4) The sensory axon pool available for growth recruitment
is diminished at greater time points after SCI.
(5) Chronically injured sensory neurons survive and do not
atrophy 15 months after SCI.
(6) Neurite outgrowth capacity and sensitivity to CLs of
sensory neuron cell bodies are maintained 15 months
after SCI.
(7) Second injury and treatments can partially restore
reduced axon regeneration in chronic SCI.
(8) The extent to which intrinsic mechanisms underlie
reduced regeneration in chronic SCI remains to be
determined.
There is no change to specific aims in this project.
Overall Hypothesis: Combinatorial therapy with
conditioning lesions, neurotrophic gradients, CHASE and
BMSC grafts will increase the number and length of
regenerating sensory axons when treatment is initiated one
year after rodent SCI.
Aim 1: Confirm preliminary findings that bridging axonal
regeneration is supported by combinatorial therapies after
chronic SCI.
Aim 2: Identify cellular and extracellular mechanisms
underlying, and potentially limiting, axonal growth
12 months after SCI.
Aim 3: Determine whether expansion of combination
therapies to include CHASE enhances axonal regeneration
3 months after SCI.
110
ISRT Scientific Committee
The Scientific Committee is an international body of non-remunerated eminent scientists and clinicians who advise the
Trustees on research matters. The membership of the Committee reflects a diverse range of expertise from molecular and
cell biology to neurosurgery and clinical practice. The Scientific Committee is called upon as part of the grant review
process and their input instrumental in developing our strategic approach to research funding. They also provide advice
on other matters, such as in response to specific enquiries from the press or official bodies.
Chairman
Professor S. Barnett PhD
Division of Clinical Neuroscience
Glasgow Biomedical Research Centre
Room 4B/17, 120 University Avenue
Glasgow G12 8TA
Tel: 0141 330 8409/4353
Email: [email protected]
Professor J. Fawcett PhD FRCP
Centre for Brain Repair
Cambridge University
Forvie Site, Robinson Way
Cambridge CB2 2PY
Tel: 01223 331 188
Fax: 01223 331 174
Professor R. Franklin BSc BVetMed PhD MRCVS
FRCPath
MRC Centre for Stem Cell Biology and Regenerative
Medicine, & Department of Veterinary Medicine,
Madingley Road
Cambridge CB3 0ES
Tel: +44 (0)1223 337642
Professor K. Brohi BSc MBBS FRCS FRCA
Centre for Neuroscience and Trauma
Barts and The London School of Medicine and Dentistry
Trauma Clinical Academic Unit
The Royal London Hospital
Whitechapel
London E1 1BB
United Kingdom
Professor M. Craggs PhD BSc CBiol MIBiol CSci MIPEM
Spinal Research Centre
Royal National Orthopaedic Hospital Trust
Brockley Hill
Stanmore
Middlesex HA7 4LP
Tel/Fax: 020 8909 5343
Dr J. Guest MD PhD FRSC
University of Miami
Department of Neurological Surgery
Lois Pope LIFE Center
Miami
Florida 33136
USA
Tel: 00 1 305 575 7059
Fax: 00 1 305 575 3337
Professor Dr V. Dietz MD FRCP
Spinal Cord Injury Centre
Forchstrasse 340
CH-8008 Zurich
Switzerland
Tel: 00 41 1 386 39 01
Fax: 00 41 1 386 39 09
Professor S. Hunt PhD FMedSci
Department of Anatomy & Developmental Biology
Medawar Building
Gower Street
London WC1E 6BT
Tel: 020 7679 1332
Fax: 020 7383 0929
Professor P. Ellaway PhD
Department of Sensorimotor Systems
Division of Neuroscience & Psychological Medicine
Charing Cross Campus
St Dunston’s Road
London W6 8RF
Tel: 020 8846 7293
Fax: 020 8846 7338
Dr L. Jakeman PhD
Department of Physiology and Cell Biology
The Ohio State University
403 Hamilton Hall
1645 Neil Ave
Columbus, OH 43210
Phone: (614) 688-4424
Fax: (614) 292-4888
111
Professor A. Logan PhD
Department of Clinical Chemistry
Wolfson Research Laboratories
Edgbaston
Birmingham B15 2TH
Tel: 0121 627 2268
Fax: 0121 472 0499
Dr G. Raivich MD DSc
Perinatal Brain Repair Group
Department of Obstetrics and Gynaecology
Department of Anatomy
86–96 Chenies Mews
London WC1E 6HX
Tel: 020 7679 6068
Fax: 020 7383 7429
Professor S. McMahon PhD FMedSci
Centre for Neuroscience Research
Hodgkin Building
Guy’s Campus
London SE1 1UL
Tel: 020 7848 6270
Fax: 020 7848 6165
Professor J. Silver PhD
Professor of Neurosciences
10900 Euclid Avenue
Cleveland
Ohio 44106–4975
USA
Tel: 00 1 216 368 6251
Fax: 00 1 216 368 4650
Professor J. Priestley PhD MA DPhil
Neuroscience Centre
Medical Sciences Building
St Bartholomew’s & the Royal London School of Medicine
& Dentistry
Queen Mary University of London
Mile End Road
London E1 4NS
Tel: 020 7882 6343
Fax: 020 7882 7726
Professor M. Tuszynski MD PhD
Center for Neural Repair
Professor of Neurosciences
University of California, San Diego
La Jolla, California 92093-0626
Phone: 00 1858 534-8857
Fax: 00 1858 534-5220
Professor J. Verhaagen PhD
Netherlands Institute for Brain Research
Meibergdreef 33
1105 Az Amsterdam
The Netherlands
Tel: 00 31 20 55665500
Fax: 00 31 20 6961006
Professor G. Raisman DM DPhil FRS
Chair of Neural Regeneration
Director, Spinal Repair Unit
Institute of Neurology, UCL
Queen Square
London WC1N 3BG
Tel: 020 7676 2172
Fax: 020 7676 2174
112
International Campaign for Cures of Spinal Cord Injury
Paralysis (ICCP)
An alliance between the Christopher and Dana Reeve
Foundation, Institut pour la Recherche sur la Moëlle Épinière
et l’Encéphale, Japan Spinal Cord Foundation, International
Spinal Research Trust, Miami Project to Cure Paralysis, Neil
Sachse Foundation, Paralyzed Veterans of America, Rick
Hansen Foundation, Spinal Cure Australia and Wings for Life.
The ICCP is a body of affiliated, not-for-profit
organisations that are working to fund research into cures
for paralysis caused by spinal cord injury. The mission of
the ICCP coalition is “to expedite the discovery of cures for
spinal cord injury paralysis”.
With this in mind organisations that promote spinal cord
research determined how their collaborative efforts might
further hasten progress. On the 12th of May 1998 in
Charlottesville Virginia they signed a ‘Statement of Intent’,
and formed the International Campaign for Cures of Spinal
Cord Injury Paralysis (ICCP).
New members are welcome and the Japan Spinal Cord
Foundation joined the ICCP in 2004. The organisations
meet regularly and produce a multipurpose, general
information package that outlines current research and
statistics on the worldwide prevalence of spinal cord injury.
ICCP member organisations are credited with funding
research that has realised many significant discoveries,
brought scientists new optimism, and significantly changed
the outlook for people who have a spinal cord injury. The
question today is not if effective treatments and cures will be
found, it is a question of when. One of the latest ICCP
initiatives has been the development of a series of guideline
papers on clinical trial design and implementation.
113
ICCP objectives
1. To attract the best scientists, researchers and clinicians
to the area of nerve regeneration and repair in the CNS,
particularly those who are newly graduating, and
encourage their career commitment to spinal cord
research.
2. To promote public support for the development of
effective treatments and cures by highlighting the
individual vulnerability to injury and the benefits of
cures to present and future generations.
3. To promote government financial support for spinal
cord research by highlighting the economic cost of
lifetime care following injury.
4. To consider conducting collaborative awareness and
fundraising campaigns to promote the global nature of
spinal cord injury and paralysis cure research.
5. To promote links and communication between
laboratories, scientists, clinicians and other relevant
organisations.
6. To promote heightened communication between
fundraising groups and encourage shared utilisation of
resources and expertise.
7. To encourage the development of internationally
accepted strategies and priorities for spinal cord injury
research.
8. To evaluate the progress and success of the Campaign
against concrete, measurable outcomes and report
progress.
Information on the global impact of spinal cord injury is included on the ICCP website
(http://www.campaignforcure.org), and a downloadable information pack is available.
Also on the website is a searchable database that provides details of the grants awarded by all member organisations, and
includes links to the websites of all member organisations and opportunities for applying for research grants.
More information on the background and aims of the ICCP is included in
Adams, M. and Cavanagh, J.F.R. (2004) International campaign for cures of spinal cord injury paralysis (ICCP):
another step forward for spinal cord injury research. Spinal Cord 42, 273–280.
114
ICCP Spinal Cord Injury Clinical Trials Guidelines
•
The number of potential cellular-based and pharmaceutical
drug treatments aimed at repairing a spinal cord injury has
increased dramatically over the past few years. Some clinical
trials have already started, many involving drastic invasive
surgery, and several more are planned for the near future.
As the numbers of potential treatments and trials continues
to increase, there is concern that currently there is no
international standard that ensures trials are carried out
consistently and as safely as possible.
•
•
•
It is crucial that there is an effective, international forum
where all aspects of clinical trial design can be discussed,
including topics such as the strength of preclinical data,
participant inclusion criteria, trial design, trial management,
trial duration, validity of outcome measures and
interpretation of results. Enhanced communication between
groups conducting trials and information sharing will
benefit all, including investigators who conduct clinical
trials and patients who participate in them.
The degree and level of injury, timing of clinical
intervention, and the statistical power needed to
achieve a valid outcome
Determining appropriate clinical outcome measures for
each type of intervention
Patient selection criteria (inclusion/exclusion) and
ethics
Controlling potential confounding variables
(standardisation of adjunct treatments) and the
organisation of multi-centre trials
The initial workshop brought together researchers and
clinicians from around the world, all of whom are currently
involved in clinical trials in spinal cord injury treatments.
One consistent point of concern was the standard of
examination before and after the treatment. There are many
methods of assessing patients, from measures of muscle
strength to quality-of-life questionnaires. Spinal Research’s
Clinical Initiative was praised as taking a lead in developing
new, accurate and reliable methods of measuring small
changes in either sensation or muscle function that would
be applicable to most clinical trials.
With this in mind, Spinal Research co-sponsored a unique,
important scientific workshop that took place in Vancouver,
Canada in February 2004 under the auspices of the ICCP.
Consequently, three additional study groups met to consider
the particular practical and ethical issues that are associated
with, or peculiar to, clinical trials in spinal cord injury. One
significant outcome from the Workshop was the initiation
of a series of panel meetings with specialists in the field over
a period of 18 months to discuss in detail several of the
issues that must be dealt with before taking potential
therapies to clinical trial. These issues include:
The deliberations and the subsequent conclusions drawn
from this series of meetings culminated in the publication
of five guideline papers, one of which is intended as a
summary document for patients thinking of participating
in clinical trials. The summary document is available for
download on our website.
Fawcett, J.W., Curt, A., Steeves, J.D. et al. (2006) Guidelines for the conduct of clinical trials for spinal cord injury as
developed by the ICCP panel: spontaneous recovery after spinal cord injury and statistical power needed for therapeutic
clinical trials. Spinal Cord 45, 190–205.
Lammertse, D., Tuszynski, M.H., Steeves, J.D. et al. (2006) Guidelines for the conduct of clinical trials for spinal cord injury
as developed by the ICCP panel: clinical trial design. Spinal Cord 45, 232–242.
Steeves, J.D., Lammertse ,D., Curt, A. et al. (2006) Guidelines for the conduct of clinical trials for spinal cord injury (SCI)
as developed by the ICCP panel: clinical trial outcome measures. Spinal Cord 45, 206–221.
Tuszynski, M.H., Steeves, J.D., Fawcett, J.W. et al. (2006) Guidelines for the conduct of clinical trials for spinal cord injury
as developed by the ICCP Panel: clinical trial inclusion/exclusion criteria and ethics. Spinal Cord 45, 222–231.
115
ICCP Participating Organisations
Further information and links to members’ websites are included on the ICCP website http://www.campaignforcure.org
Mr David Prast
Spinal Cure Australia
PO Box 131
Artarmon NSW 1570
Australia
Tel: +61 2 9660 1040
Fax: +61 2 9660 4494
http://www.spinalcure.org.au
Dr Mark Bacon
Director of Research
International Spinal Research Trust
Bramley Business Centre
Bramley, Guildford, GU5 0AZ
UK
Tel: +44 (0) 1483 898786
Fax: +44 (0) 1483 898763
http://www.spinal-research.org
Mr Neil Sachse
Neil Sachse Foundation
141 Ifould Street
Adelaide SA 5000
Australia
Tel: +61 8 8227 1777
Fax: +61 8 8232 4311
http://www.nsf.org.au
Ms Susan Howley
Executive VP & Director of Research
Christopher Reeve Paralysis Association
500 Morris Avenue
Springfield, NJ 07081
USA
Tel: +1 973 379 2690
Fax: +1 973 912 9433
http://www.christopherreeve.com
Rosi Lederer
Wings for Life
Fürstenallee 4
5020 Salzburg
Austria
Tel: +43 662 6582 4206
Fax: +43 662 6582 5062
http://www.wingsforlife.com
Dr Alain Privat
Institut pour la Recherche sur la Moëlle Épinière
Université de Montpellier II, CC106
Place Eugène Bataillon
34095 Montpellier Cedex 05
France
Tel: +33 4 67 14 33 86
Fax: +33 4 67 14 33 18
http://www.irme.org
Beth Goldsmith
Executive Director
The Craig H. Neilsen Foundation
16633 Ventura Blvd., Suite 1050
Encino, CA 91436
USA
Tel: +1 818 8177616
Fax: +1 818 9957099
http://chnfoundation.org
Makoto Ohama
Japan Spinal Cord Foundation
4–7–16 Sumiyosi-cho
Fuchu-city
Tokyo 183-0034
Japan
Tel: 00 81 42 366 5153
http://www.jscf.org
The IRP Foundation
54, avenue Dapples
Case postale 655
CH-1001 Lausanne
Switzerland
Tel: +41 (0) 21 614 7777
Fax: +41 (0) 21 614 7778
http://www.irp.ch
Paralyzed Veterans of America
801 18th St., NW
Washington, DC 20006
USA
Tel: +1 202 416 7668
Fax: +1 202 416 7641
http://www.pva.org
116
Meg Speirs
The CatWalk Trust
PO Box 555
409 Queen Street
Masterton 5840
New Zealand
Tel: + 64 6 377 5430
Fax: +64 6 377 5432
http://www.catwalk.org.nz
The Miami Project to Cure Paralysis
1095 NW 14th Terrace
Lois Pope LIFE Center
Miami, FL 33136
USA
Tel: + 1 305 243-6001
Fax: + 1 305 243-6017
http://www.miamiproject.miami.edu
Rick Hansen Foundation
300–3820 Cessna Drive
Richmond, BC
V7B 0A2
Canada
Tel: + 604 295 8149
Fax: + 604 295 8159
http://www.rickhansen.com
117

ISRT Research Review 2010

Transcription

Similar documents

Plisic, Ljiljana

Lapidus Bunionectomy

Degenerative Cervical Myelopathy

Spine and Spinal Cord Injuries

sistem saraf (nervous system)

Gonzalez - SEP Monitoring (H)

SAFETY GUIDELINES TRAMPOLINE PARTS LIST

CHAPTER 19 Lesions of the Spine and Spinal Cord

program brochure