HOODIA GORDONII: QUALITY CONTROL AND
Transcription
HOODIA GORDONII: QUALITY CONTROL AND
HOODIA GORDONII: QUALITY CONTROL AND BIOPHARMACEUTICAL ASPECTS by ILZE VERMAAK Submitted in partial fulfillment of the requirements for the degree DOCTOR TECHNOLOGIAE in the Department of Pharmaceutical Sciences FACULTY OF SCIENCE TSHWANE UNIVERSITY OF TECHNOLOGY Supervisor: Prof AM VILJOEN Co-Supervisor: Prof JH HAMMAN April 2011 DECLARATION BY CANDIDATE “I hereby declare that the thesis submitted for the degree of DTech: Pharmaceutical Sciences, at Tshwane University of Technology, is my own original work and has not previously been submitted to any other institution of higher education. I further declare that all sources cited or quoted are indicated and acknowledged by means of a comprehensive list of references”. I Vermaak Copyright© Tshwane University of Technology 2011 ii ACKNOWLEDGEMENTS I would like to acknowledge the following people: § First and foremost my supervisor Prof Alvaro Viljoen for his excellent supervision and mentorship and for sharing with me his expertise and research insight. He is my role model of a successful researcher and the one who inspires me to become a researcher of note. § My co-supervisor Prof Sias Hamman for his thoughtful advice, endless support and encouragement as well as attention to detail. § Dr Gill Enslin (Head of Department) and my colleagues at the Department of Pharmaceutical Sciences for their support and encouragement. § My fellow students at the Department of Pharmaceutical Sciences: Thank you for always being willing to assist me when needed and for your support and encouragement. It certainly has been fun. § The people who provided Hoodia gordonii samples: § The Kalahari Hoodia Growers in Namibia. A special thank you for your kind hospitality during our visits. § Conrad Geldenhuys and Nico Laubscher from the Northern Cape Department of Environmental Affairs and Nature Conservation. § Dr Marietjie Stander, Kobus Engelbrecht and Adolf Joubert. § Dr Weiyang Chen (HPLC) for developing the quantification method for P57. § Dr Malgorzata Baranska (Raman spectroscopy) for teaching us about Raman spectroscopy and the productive collaboration that followed. Your hospitality during our visit to Poland was much appreciated. § Dr Jan Masthoff (CAMAG) for his assistance with HPTLC-densitometry training and hospitality during our visit to Switzerland. § Tshwane University of Technology (TUT) and National Research Foundation (NRF) for financial assistance. A special note of thanks to TUT for granting me sabbatical leave enabling me to complete the write-up of my thesis. § My special friends: your support and encouragement and especially your patience with me is greatly appreciated. § My family for their support and interest in my studies. Lastly and most importantly, I wish to thank my parents and sister for their encouragement, love and support, making it possible to complete this degree. iii ABSTRACT Hoodia gordonii is a popular weight loss product highly susceptible to adulteration. This highlights the need for rapid and simple quality control methods for authentication of raw material and products. High performance thin layer chromatography analysis was used to authenticate raw material. Quantitative high performance thin layer chromatography analysis as well as near infrared, midinfrared and Raman spectroscopy combined with chemometric techniques were used to develop alternative quantification methods for P57 (the perceived active ingredient of H. gordonii) in raw material and/or products. Liquid chromatography coupled to mass spectrometry was used to determine the concentration of P57 and used as reference data to develop calibration models based on the partial least squares projections to latent structures regression algorithm for the infrared spectra. The performance of each calibration model was evaluated according to the correlation coefficient (R2) and root mean square error of prediction. The high performance thin layer chromatography system produced good separation of compounds including that of the P57 band. Quantitative determination for P57 resulted in linear calibration curves with good correlation coefficient (R2) values of 0.9706-0.9993. For the near infrared spectroscopy data, the partial least squares projections to latent structures model with 2nd derivative pre-processing predicted P57 content with an R2 value of 0.9629 and a root mean square error of prediction of 0.03%. Pre-processing of the Raman data with orthogonal signal correction yielded a partial least squares projections to latent structures model with an R2 value of 0.9986 and a root mean square error of prediction of 0.004%. The qualitative and quantitative high performance thin layer chromatography analyses provided a chemical fingerprint technique for authentication and confirmation of the presence of P57 in H. gordonii raw material and products as well as for quantification of P57. The parameters of the calibration models demonstrated that both near infrared and Raman spectroscopy shows potential to rapidly quantify P57 in H. gordonii raw material. As H. gordonii is one of the most widely consumed anti-obesity products of natural origin, it is disturbing that key aspects such as pharmacokinetic and biopharmaceutical aspects have not been sufficiently investigated. The in vitro iv transport of pure P57 as well as P57 from crude plant extracts across excised porcine intestinal and buccal tissue was investigated using a Sweetana-Grass diffusion apparatus. In addition to using buffer as transport medium, in vivo conditions were mimicked with the use of saliva and simulated intestinal fluid. The apparent permeability coefficient and flux values were calculated and statistically analysed. Pure P57 was transported at a much lower rate and extent than P57 from the crude extract across porcine intestinal tissue. P57 was transported across the buccal mucosa when applied in the form of a crude plant extract but no transport could be detected when P57 was applied in the pure form. The transport of P57 across intestinal and buccal mucosal tissues is significantly affected by exposure to conditions simulating an in vivo environment. v TABLE OF CONTENTS PAGE ACKNOWLEDGEMENTS ....................................................................................... iii ABSTRACT ............................................................................................................ iv LIST OF FIGURES ................................................................................................. xi LIST OF TABLES ................................................................................................. xvi GLOSSARY ......................................................................................................... xvii CHAPTER 1 – Introduction 1.1 INTRODUCTION....................................................................................... 1 1.2 AIMS ......................................................................................................... 4 1.3 OBJECTIVES ............................................................................................ 4 1.4 THESIS LAYOUT ...................................................................................... 5 1.5 REFERENCES.......................................................................................... 6 CHAPTER 2 – Hoodia gordonii and other plants used for their anti-obesity property 2.1 INTRODUCTION....................................................................................... 7 2.2 HOODIA GORDONII ................................................................................. 7 2.2.1 Classification and botanical aspects ......................................................... 7 2.2.2 Ethnopharmacology ................................................................................ 10 2.2.3 Commercialisation................................................................................... 10 2.2.4 Intellectual property rights and benefit sharing issues ............................ 14 2.2.5 Phytochemistry........................................................................................ 15 2.2.6 Biological activity ..................................................................................... 16 2.2.6.1 Anti-obesity activity ................................................................................. 16 2.2.6.2 Other biological activities ........................................................................ 19 2.2.7 Toxicity .................................................................................................... 20 2.3 PLANTS USED FOR THEIR ANTI-OBESITY PROPERTY .................... 21 2.3.1 Introduction ............................................................................................. 21 2.3.2 Mechanisms of action of anti-obesity preparations ................................. 22 2.3.3 Other plants popularly used as anti-obesity agents ................................ 25 2.4 CONCLUSIONS ...................................................................................... 42 2.5 REFERENCES........................................................................................ 44 vi SECTION 1 – Quality control ............................................................................. 68 REFERENCES ..................................................................................................... 69 CHAPTER 3 – Chromatographic techniques 3.1 LIQUID CHROMATOGRAPHY ............................................................... 70 3.1.1 INTRODUCTION..................................................................................... 70 3.1.2 BACKGROUND AND LITERATURE REVIEW ....................................... 71 3.1.3 MATERIALS AND METHODS ................................................................ 73 3.1.3.1 Plant material, reagents and sample preparation ................................... 73 3.1.3.2 Liquid chromatography coupled to mass spectrometry (LC-MS) ............ 74 3.1.3.3 High performance liquid chromatography (HPLC) .................................. 74 3.1.4 RESULTS................................................................................................ 75 3.1.4.1 Liquid chromatography coupled to mass spectrometry (LC-MS) ............ 75 3.1.4.2 High performance liquid chromatography (HPLC) .................................. 77 3.1.5 CONCLUSIONS ...................................................................................... 77 3.2 QUALITATIVE HIGH PERFORMANCE THIN LAYER CHROMATOGRAPHY (HPTLC) ............................................................. 79 3.2.1 INTRODUCTION..................................................................................... 79 3.2.2 BACKGROUND AND LITERATURE REVIEW ....................................... 80 3.2.2.1 Analysis of plant material ........................................................................ 80 3.2.3 MATERIALS AND METHODS ................................................................ 81 3.2.3.1 Plant material and reagents .................................................................... 81 3.2.3.2 High performance thin layer chromatography (HPTLC) analysis ............ 81 3.2.3.3 Liquid chromatography coupled to mass spectrometry (LC-MS) analysis .................................................................................................... 83 3.2.4 RESULTS AND DISCUSSION................................................................ 84 3.2.4.1 High performance thin layer chromatography (HPTLC) .......................... 84 3.2.4.2 Liquid chromatography coupled to mass spectrometry (LC-MS) ............ 90 3.2.5 CONCLUSIONS ...................................................................................... 92 3.3 QUANTITATIVE HIGH PERFORMANCE THIN LAYER CHROMATOGRAPHY (HPTLC) ............................................................. 92 3.3.1 INTRODUCTION..................................................................................... 92 3.3.2 BACKGROUND AND LITERATURE REVIEW ....................................... 93 vii 3.3.3 MATERIALS AND METHODS ................................................................ 94 3.3.3.1 Plant material and reagents .................................................................... 94 3.3.3.2 Quantitative high performance thin layer chromatography analysis ....... 94 3.3.4 RESULTS AND DISCUSSION................................................................. 96 3.3.4.1 Quantitative high performance thin layer chromatography analysis ....... 96 3.3.5 CONCLUSIONS .................................................................................... 100 3.4 REFERENCES...................................................................................... 102 CHAPTER 4 – Vibrational spectroscopy 4.1 INTRODUCTION................................................................................... 104 4.2 BACKGROUND AND LITERATURE REVIEW ..................................... 105 4.2.1 Infrared spectroscopy (IR)..................................................................... 105 4.2.2 Chemometric data analysis ................................................................... 107 4.3 MATERIALS AND METHODS .............................................................. 108 4.3.1 Plant material ........................................................................................ 108 4.3.2 Near infrared spectroscopy (NIR) ......................................................... 109 4.3.3 Mid-infrared spectroscopy (MIR) ........................................................... 109 4.3.4 Raman spectroscopy ............................................................................ 110 4.3.5 Raman mapping .................................................................................... 111 4.3.6 Data analysis......................................................................................... 111 4.3.7 P57 quantification with LC-MS analysis ................................................ 112 4.4 RESULTS AND DISCUSSION.............................................................. 112 4.4.1 Chemometric data analysis methods and parameters .......................... 112 4.4.2 NIR spectroscopy .................................................................................. 113 4.4.3 MIR spectroscopy ................................................................................. 116 4.4.4 Raman spectroscopy ............................................................................ 117 4.4.5 Raman mapping .................................................................................... 118 4.5 CONCLUSIONS .................................................................................... 120 4.6 REFERENCES...................................................................................... 122 viii SECTION 2 – Biopharmaceutical aspects ...................................................... 125 CHAPTER 5 – Buccal and intestinal transport of P57 from Hoodia gordonii 5.1 INTRODUCTION................................................................................... 126 5.2 BACKGROUND AND LITERATURE REVIEW ..................................... 127 5.2.1 Introduction ........................................................................................... 127 5.2.2 In vitro models for intestinal permeation studies .................................. 130 5.2.2.1 Membrane-based models ..................................................................... 130 5.2.2.2 Cell culture models................................................................................ 131 5.2.2.3 Isolated tissue models........................................................................... 131 5.2.3 In vitro models for buccal permeation studies ....................................... 133 5.2.3.1 Isolated tissue models........................................................................... 133 5.2.3.2 Cell culture models................................................................................ 133 5.3 METHODS AND MATERIALS .............................................................. 134 5.3.1 In vitro intestinal and buccal permeation studies .................................. 134 5.3.1.1 Plant material and reagents .................................................................. 134 5.3.1.2 Tissue preparation ................................................................................ 134 5.3.1.3 Preparation of simulated intestinal fluid and artificial saliva .................. 136 5.3.1.4 Transport studies .................................................................................. 137 5.3.1.5 Quantification of P57 ............................................................................. 138 5.3.1.6 Data analysis and statistical evaluation ................................................ 138 5.4 RESULTS AND DISCUSSION.............................................................. 139 5.4.1 In vitro intestinal transport studies......................................................... 139 5.4.2 In vitro buccal transport studies ............................................................ 141 5.5 CONCLUSIONS .................................................................................... 143 5.6 REFERENCES...................................................................................... 144 CHAPTER 6 – General conclucions and recommendations 6.1 GENERAL CONCLUSIONS.................................................................. 147 6.2 RECOMMENDATIONS ......................................................................... 148 ANNEXURE 1: Chemical structures 1.1. Chemical structures of the steroidal glycosides isolated from H. gordonii with their corresponding aglycones indicated (*,#,$,@,^) ............................. 150 ix ANNEXURE 2: Publications emanating from this thesis 2.1. Publications in accredited journals ........................................................ 158 2.2. Publication in a non-accredited journal ................................................. 158 2.3. Publications submitted to accredited journals ....................................... 158 ANNEXURE 3: Conference contributions 3.1. Oral papers ........................................................................................... 159 3.2. Poster .................................................................................................... 159 x LIST OF FIGURES PAGE FIGURE 1.1: Examples of H. gordonii products available for purchase on 2 the internet FIGURE 1.2: Examples of H. gordonii products that are available in 3 various forms such as patches, oral strips, capsules, topical gels, fruit bars and chewing gum FIGURE 2.1: Hoodia gordonii in flower (Photograph by AM Viljoen) 9 FIGURE 2.2: Hoodia gordonii cross-sections of dried and freshly-cut 9 stems (Photographs by AM Viljoen) FIGURE 2.3: The natural distribution of H. gordonii in South Africa and 9 Namibia indicated in orange FIGURE 2.4: Timeline of the most important developments in the research 11 and commercial development of H. gordonii FIGURE 2.5: An article on H. gordonii published in the Sunday Times 13 newspaper on 20 September 2009 FIGURE 2.6: Photographs of commercially cultivated H. gordonii in 13 greenhouse tunnels in Namibia and South Africa and the processing of the raw material by slicing (Photographs by AM Viljoen) FIGURE 3.1: Schematic illustration of an HPLC apparatus coupled to an 73 ultraviolet detector and a mass spectrometer (LC-UV-MS) FIGURE 3.2: LC-MS chromatogram of the plant material used as the 75 H. gordonii control standard for the LC-MS analytical method FIGURE 3.3: An extracted mass (861.5 m/z) chromatogram of P57 (A) 76 and the H. gordonii control standard plant material (B) FIGURE 3.4: Calibration curve for the P57 reference standard as obtained 76 by LC-MS FIGURE 3.5: The percentage of P57 determined with LC-MS in 77 H. gordonii stem samples (n = 145) FIGURE 3.6: An HPLC chromatogram of the H. gordonii plant material from a commercial capsule xi 78 FIGURE 3.7: The calibration curve for the P57 reference standard with the 78 HPLC method FIGURE 3.8: A diagram of the CAMAG semi-automated HPTLC system including the developing automatic chamber TLC ADC2, Sampler 4, automatic chromatogram immersion 82 device, TLC plate heater III and documentation device Reprostar 3 FIGURE 3.9: The HPTLC chromatogram used to determine the limit of 84 detection for P57. P57 was spotted at 5 µg (track 1), 3 µg (track 2), 2 µg (track 3), 1.5 µg (track 4), 1 µg (track 5), 0.5 µg (track 6), 0.25 µg (track 7), 0.15 µg (track 8) and 0.1 µg (track 9) FIGURE 3.10: An HPTLC plate viewed under UV at 365 nm. Track 1: 85 H. gordonii standard raw material, track 2-8: H. gordonii commercial products (1-7) from different manufacturers and track 9: the P57 reference standard FIGURE 3.11: An HPTLC plate viewed under UV at 365 nm. Track 1: 86 H. gordonii standard raw material, track 2-8: H. gordonii commercial products (8-14) from different manufacturers and track 9: the P57 reference standard FIGURE 3.12: An HPTLC plate viewed under UV at 365 nm. Track 1: 86 H. gordonii standard raw material, track 2-8: H. gordonii commercial products (15-20) from different manufacturers and track 9: the P57 reference standard FIGURE 3.13: An HPTLC plate viewed under UV at 365 nm. Track 1: wild population sample of H. gordonii from South Africa, track 2: wild population sample of H. gordonii from Namibia, track 3: cultivated sample of H. gordonii from South Africa, track 4: cultivated sample of H. gordonii from Namibia, track 5: 2 year old actively growing plant sample, track 6: 2 year old dormant plant sample, track 7: a 3 year old actively growing cultivated plant, track 8: a 3 year old dormant cultivated plant and track 9: the P57 reference standard xii 88 FIGURE 3.14: An HPTLC plate viewed under UV at 365 nm. Track 1: 89 H. gordonii standard raw material, track 2-7: the same sample of H. gordonii standard raw material as for track 1 adulterated with O. ficus-indica in concentrations of 10%, 20%, 40%, 50 %, 60%, 80%, respectively, track 8: O. ficusindica raw material, track 9: the P57 reference standard FIGURE 3.15: An HPTLC plate viewed under UV at 365 nm. Track 1: 90 H. gordonii standard raw material, Tracks 2 and 3: Opuntia ficus-indica samples, Tracks 4 and 5: Aloe ferox samples, Track 6: Cereus jamacaru sample, Track 7: Agave americana sample and Track 8: the P57 reference standard FIGURE 3.16: LC-MS chromatograms of the P57 standard (A) and the 91 preparative TLC sample (B) FIGURE 3.17: Mass spectra of the P57 reference standard (A) and the 91 preparative TLC sample (B) FIGURE 3.18: A photograph of the CAMAG TLC scanner 3 used in the 94 quantification of P57 from commercial samples (Photograph by I Vermaak) FIGURE 3.19: An example of an HPTLC chromatogram that was used for 96 the quantification of P57 from H. gordonii commercial products. Track 1: 65.50 ng P57, Track 2-3: Product 1, Track 4: 75.00 ng P57, Track 5-6: Product 2, Track 7: 100.00 ng P57, Track 8-9: Product 3, Track 10: 125.00 ng P57, Track 11-12: Product 4 FIGURE 3.20: An example of the baseline chromatogram for one track (A) 97 and the corresponding integration of the P57 peak as indicated (B) FIGURE 3.21: The linear calibration curve for four standard levels of P57 97 (black) and the four H. gordonii products (green) in duplicate in terms of peak height FIGURE 3.22: The linear calibration curve for four standard levels of P57 (black) and the four H. gordonii products (green) in duplicate in terms of peak area xiii 98 FIGURE 4.1: Photograph of the Buchi NIRFlex N500 Fourier transform near infrared spectrophotometer and sample 109 tray (Photograph by I Vermaak) FIGURE 4.2: Photographs of the Bruker Alpha-P Fourier transform 110 infrared spectrophotometer (Photographs by I Vermaak) FIGURE 4.3: Photographs of the Bruker Multiram Fourier transform 110 Raman spectrometer (Photographs by I Vermaak) FIGURE 4.4: Typical FT-NIR spectra for H. gordonii powdered raw 114 material (A) and the spectrum for the P57 reference standard (B) FIGURE 4.5: Loadings line plot from the PLS calibration model using FT- 115 NIR data before (A) and after (B) removal of X variables with low loading weights (10000 – 7800 cm-1) FIGURE 4.6 PLS calibration model using the 2nd derivative pre- 116 processing technique calculated from FT-NIR data FIGURE 4.7 PLS calibration model with OSC pre-processing calculated 118 from FT-NIR-Raman data FIGURE 4.8 A microscopic image of a fresh H. gordonii slice with the 119 marked area where the Raman mapping measurement was performed (A), and chemical maps of the clusters (B) and the borders of the clusters (C) showing the distribution of various plant components (blue – cellulose; violet – carotenoids; green – lignin) FIGURE 4.9 A single Raman spectrum of H. gordonii extracted from the 119 three clusters coloured accordingly. At the bottom a Raman spectrum of pure P57 standard is presented FIGURE 5.1: Schematic illustration of the drug absorption pathways across intestinal epithelial cells. (A) paracellular diffusion, (B) paracellular diffusion through opened tight junctions, (C) transcellular passive diffusion with (C*) intracellular metabolism, (D) carrier-mediated transcellular transport, (E) efflux and (F) vesicular transcytosis (Hunter and Hirst, 1997:131) xiv 128 FIGURE 5.2: The preparation and mounting of porcine intestinal mucosa 135 in Sweetana-Grass diffusions chambers (Photographs by JH Hamman) FIGURE 5.3: A photograph of a diffusion chamber cell as well as a 136 schematic diagram of an assembled cell indicating the excised mucosal tissue mounted between two half-cells FIGURE 5.4: Transport of pure P57 as well as P57 from a crude plant 139 extract across porcine intestinal tissue in the absorptive direction ( ) and secretory direction ( ) expressed in terms of apparent permeability coefficient (Papp) values FIGURE 5.5: Transport of pure P57 as well as P57 from a crude plant extract across porcine buccal tissue in the absorptive direction expressed in terms of apparent permeability coefficient (Papp) values xv 142 LIST OF TABLES PAGE TABLE 2.1: Summary of studies on the in vivo anti-obesity activity of 17 H. gordonii and P57 TABLE 2.2: Summary of the plants investigated for their anti-obesity 26 properties in alphabetical order by species name TABLE 3.1: Comparison of the P57% LC-MS results and HPTLC results 87 TABLE 3.2: Peak height and area comparison for the calculation of intra- 99 day precision TABLE 3.3: Peak height and area comparison for the calculation of inter- 100 day precision TABLE 3.4: Accuracy of the P57 standards determined by recovery 101 analysis TABLE 4.1: Number of PLS factors, R2, Q2, RMSEP and RMSEE values 113 for determining P57 content with NIR, MIR and Raman spectroscopy utilising a PLS model with different preprocessing methods TABLE 4.2: P57 content of the outer layers and cortex samples of 120 H. gordonii determined by LC-MS analysis TABLE 5.1: Experimental groups used for the intestinal and buccal 138 transport studies TABLE 5.2: Flux (J) values (µg/cm2/h) for P57 across porcine intestinal 140 mucosa TABLE 5.3: Flux (J) values (µg/cm2/h) for P57 across porcine buccal mucosa xvi 142 GLOSSARY ACC Acetyl coenzyme A carboxylase ADC Automatic developing chamber AMPK 5ʹ Adenosine monophosphate-activated protein kinase ANOVA Repeated one-way analysis of variance APCI Atmospheric pressure chemical ionisation ATP Adenosine triphosphate cAMP 3ʹ-5ʹ-Cyclic adenosine monophospate CBD Convention on Biological Diversity CCAAT Cytidine-cytidine-adenosine-adenosine-thymidine CCK Cholesystokinin C/EBPs CCAAT-enhancer-binding proteins CF-FAB Continuous flow-fast atom bombardment CITES Convention on International Trade in Endangered Species CNS Central nervous system CoA Coenzyme A CSIR Council for Scientific and Industrial Research CYP Cytochrome P ESI Electrospray ionisation FAS Fatty acid synthase FDA Food and Drug Administration FT Fourier transform GDPH Glycerol-3-phosphate dehydrogenase HPLC High performance liquid chromatography HPTLC High performance thin layer chromatography 5-HTP 5-Hydroxytryptophan IAM Immobilised artificial membrane i.c.v. Intracerebroventricular IR Infrared LB Liebermann-Burchard LC Liquid chromatography LC-MS Liquid chromatography coupled to mass spectrometry LC-UV Liquid chromatography coupled to ultraviolet detection xvii LC-UV-MS Liquid chromatography coupled to ultraviolet detection and mass spectrometry LLC-PK1 Lewis lung carcinoma-porcine kidney 1 MDCK Madin-Darby canine kidney MIR Mid-infrared MS Mass spectrometry MSC Multiplicative signal correction NIR Near infrared OPLS Orthogonal partial least squares projections structures OSC Orthogonal signal correction PAMPA Parallel membrane permeability assays PCA Principal component analysis P-gp P-glycoprotein PLS Partial least squares projections to latent structures PPAR Peroxisome proliferator-activated receptors Q 2 Cumulative overall cross-validated R2X R2 Correlation coefficient RBP Retinoid binding protein Rf Retention front RMSEP Root mean square error of prediction ROS Reactive oxygen species SIF Simulated intestinal fluid SNV Standard normal variate SREBP Sterol-regulatory-element-binding protein TG Triglyceride TLC Thin layer chromatography TSP Thermospray UCP Uncoupling proteins UPLC Ultra performance liquid chromatography UV Ultraviolet xviii to latent CHAPTER 1 Introduction 1.1 INTRODUCTION Hoodia gordonii (Masson) Sweet ex. Decne. is a spiny succulent plant with fleshy, thorny stems, belonging to the Apocynaceae family of flowering plants. It is frequently erroneously referred to as the desert cactus (Cactaceae) in the media due to its appearance. According to traditional knowledge passed down from generation to generation, the San bushmen of Southern Africa consumed the flesh of this plant to suppress hunger and thirst while on long hunting trips or in times of low food supply (Van Heerden, 2008:435; Glasl, 2009:300). Not surprisingly, in a world intensely focused on obesity and its associated illnesses, the discovery of this plants’ promise as an appetite suppressant initiated substantial research and commercial interest. According to the World Health Organisation, 1.6 billion adults worldwide are overweight, a number expected to increase by 40% within the next decade (WHO, 2006:S.a.). This global fat-epidemic gives rise to a multimillion dollar market for weight loss products. Hoodia-containing products exploded onto the market being sold as the new miracle anti-obesity preparation. It is extremely popular, as evidenced by the more than 5 million links that appear when the word Hoodia is typed into the GoogleTM search engine, and has been widely advertised in magazines as well as on several television shows (Figure 1.1). The credible history attached to the use of H. gordonii, the aggressive advertising, the worldwide obesity problem and the general misconception by the public that all plant medicinal products are of natural origin and therefore safe makes H. gordonii the ‘perfect’ candidate to seize a large portion of the anti-obesity products market. Unfortunately, the shortage of plant material due to the slow growth rate of H. gordonii, the need for a collecting permit and the high market demand leads to the distribution of low quality products produced by unscrupulous companies. Many commercially sold “Hoodia” weight-loss preparations have been proven to be fake (Holt and Taylor, 2006:111). 1 FIGURE 1.1: Examples of H. gordonii products available for purchase on the internet The quality and authenticity of H. gordonii raw plant material purchased by manufacturers and H. gordonii products is determined by the quantity of P57AS3 (commonly referred to as P57) it contains. P57 is one of the steroidal glycosides isolated by the pioneering group of scientists who investigated Hoodia species for appetite suppression (Van Heerden et al., 2007:2546) and is the perceived active anti-obesity compound in raw plant material. Since then several other researchers have identified steroidal glycosides from H. gordonii which may have potential as appetite-suppressants but according to the herbal industry, P57 content is currently the only benchmark of quality. The preferred analytical method currently used to quantify P57 is high performance liquid chromatography (HPLC) coupled to mass spectrometry (LC-MS) although more recently Janssen et al. (2008:200) quantified steroidal glycosides by using HPLC coupled to an ultraviolet detector (LC-UV). These analytical methods are expensive, require extensive sample preparation, and needs to be performed by a well-trained technician. This highlights the need for simpler, faster, and less expensive quality control analytical methods such as vibrational spectroscopy including near infrared (NIR), midinfrared (MIR) and Raman spectroscopy. 2 High performance thin layer chromatography (HPTLC) can provide a simpler means of authentication of raw material, confirm the presence of P57, and in addition quantify its content by densitometry. In addition to the need for quality control, the most basic biopharmaceutical properties to validate the pharmacological effect of H. gordonii were not determined before it went into commercial production. It is only recently that some of these aspects such as the in vitro pharmacokinetic profile of P57 have been investigated. To complicate matters, H. gordonii products are available in many different forms including tablets, capsules, patches, oral strips, topical gels, fruit bars and even chewing gum (Figure 1.2). FIGURE 1.2: Examples of H. gordonii products that are available in various forms such as patches, oral strips, capsules, topical gels, fruit bars and chewing gum Many of these commercial products such as the Hoodia patches are sold regardless of the fact that there is no precedent proving benefit (Holt and Taylor, 2006:111). Currently, it has not even been concluded whether oral H. gordonii has the desired anti-obesity effect, let alone whether P57 and other H. gordonii phytoconstituents permeate through barriers such as the skin to the extent of eliciting a systemic effect. It is evident that much research is needed before the 3 extensive worldwide use of H. gordonii can be validated. In addition, the more immediate problem of a faster quality control method for H. gordonii products could increase the safety and reputation of these products. 1.2 AIMS The first aim of the research project was to develop and optimise rapid quality control methods for H. gordonii raw material and products. The second aim was to determine whether the perceived active component of H. gordonii (P57) is transported across porcine intestinal and buccal mucosa. 1.3 OBJECTIVES § To determine the quantity of P57 in various plant material samples using liquid chromatography coupled to mass spectrometry (LC-MS). § To develop a chemical fingerprint technique for H. gordonii plant material and products using high performance thin layer chromatography (HPTLC). § To develop a method to detect adulteration of H. gordonii raw material and products and to confirm the presence or absence of P57 in these samples using qualitative HPTLC. § To use quantitative HPTLC to quantify P57 in H. gordonii products. § To acquire vibrational spectroscopy data for powdered H. gordonii raw material using near infrared (NIR), mid infrared (MIR) and Raman spectroscopy and to determine the location of P57 within the plant stem using Raman mapping. § To develop calibration models using the LC-MS quantification data in conjunction with the NIR, MIR and Raman spectroscopy data using chemometric analysis. § To determine whether pure P57 and P57 in crude extract form is transported across porcine intestinal mucosa in Sweetana-Grass diffusion chambers and to quantify the amount with high performance liquid chromatography (HPLC). § To determine whether pure P57 and P57 in crude extract form is transported through porcine buccal mucosa in Sweetana-Grass diffusion chambers and to quantify the amount with HPLC. 4 1.4 THESIS LAYOUT The thesis is divided into the following chapters: Introduction (Chapter 1), Hoodia gordonii and other plants used for their anti-obesity property (Chapter 2), chromatographic techniques (Section 1, Chapter 3), Vibrational spectroscopy (Section 1, Chapter 4), Buccal and intestinal transport of P57 from H. gordonii (Section 2, Chapter 5) and General conclusions and recommendations (Chapter 6). The experimental part was divided into two sections dealing with the two main topics, i.e. quality control and biopharmaceutical aspects. 5 1.5 REFERENCES GLASL, S. 2009. Hoodia: A herb used in South African traditional medicine – A potential cure for overweight? Pharmacognostic review of history, composition, health-related claims, scientific evidence and intellectual property rights. Schweizerische Zeitschrift für Ganzheitsmedizin, 21(6):300-306. HOLT, S., TAYLOR, T.V. 2006. Hoodia gordonii: An overview of biological and botanical characteristics: Part 1. Townsend Letter for Doctors and Patients, 280:104-113. JANSSEN, H.-G., SWINDELLS, C., GUNNING, P., WANG, W., GRÜN, C., MAHABIR, K., MAHARAJ, V.J., APPS, P.J. 2008. Quantification of appetite suppressing steroid glycosides from Hoodia gordonii in dried plant material, purified extracts and food products using HPLC-UV and HPLC-MS methods. Analytica Chimica Acta, 617(1-2):200-207. VAN HEERDEN, F.R. 2008. Hoodia gordonii: A natural appetite suppressant. Journal of Ethnopharmacology, 119:434-437. VAN HEERDEN, F.R., HORAK, R.M., MAHARAJ, V.J., VLEGGAAR, R., SENABE, J.V., GUNNING, P.J. 2007. An appetite suppressant from Hoodia species. Phytochemistry, 68:2545-2553. WHO see WORLD HEALTH ORGANISATION. 2006. Fact sheet no 311. Available on: http://www.who.int/mediacentre/factsheets/fs311/en/print.html [Accessed: 2009/09/29]. Figure references (www) FIGURE 1.1 http://webforhealth.com/wp-content/uploads/2009/01/hoodiastar.gif [Accessed: 2010/12/13]. http://hoodiadietpill.cn/images/hoodia-Gordonii_header.jpg [Accessed: 2010/12/13]. http://free99.net/images/Hoodia%20diet%20patch.gif [Accessed: 2010/12/13]. http://www.dietsinreview.com/images/cache/75x35_hoodia-stack-logo.gif [Accessed: 2010/12/13]. http://www.fortmaywood.com/news/i/User%2520buy%2520hoodia.jpg [Accessed: 2010/12/13]. http://www.tvtopten.com/images/hoodia2.jpg[Accessed:2010/12/13]. http://i.inteletrack.com/images/purehoodia/update2/PureHoodia_Email.jpg [Accessed: 2010/12/13]. http://www.hoodiaworx.com/images/hoodiaplus430x600.gif [Accessed: 2010/12/13]. 6 CHAPTER 2 Hoodia gordonii and other plants used for their antiobesity property 2.1 INTRODUCTION Hoodia gordonii has enjoyed a great deal of popularity as the new miracle weightloss product and has established itself, from 2005 onwards, as one of the bestselling weight-loss supplements in the USA, Canada and Western Europe. The commercialisation of this plant has been highly controversial due to intellectual property rights and benefit sharing issues as well as the fact that several prominent pharmaceutical companies involved in its development have withdrawn from further commercial ventures. Despite the lack of convincing scientific studies on the quality and efficacy of H. gordonii products, it is widely consumed raising concerns about safety. The long history of use of fresh raw plant material sets a precedent for safety but it must be kept in mind that this cannot be extrapolated to the use of extracts or synthetic chemical analogues developed (Holt and Taylor, 2006a:111). The investigation into H. gordonii as an anti-obesity agent in this study led to further exploration of the numerous other plants that are used for the treatment of obesity. In most cases the scientific basis for worldwide consumption is clearly absent and information on toxicity is limited. This emphasises the need for natural products to be regulated more rigorously similar to the regulation of conventional pharmaceuticals. 2.2 HOODIA GORDONII 2.2.1 Classification and botanical aspects Hoodia gordonii is part of the genus classified as stapeliads within the tribe Ceropegieae of the subfamily Asclepiadoideae belonging to the Apocynaceae family. The previous classification of Trichocaulon is not recognised as all Trichocaulon species were reclassified as Hoodia and Hoodia is no longer considered part of the Ascepliadaceae. The complete botanical name Hoodia gordonii (Masson) Sweet ex. Decne. was derived based on the discoveries of 7 several people. Hoodia pilifera was the first Hoodia species to be discovered by Carl P. Thunberg and Francis Masson in 1774. Robert Gordon observed a Hoodia species in 1779 and made a drawing of the plant which was published by Masson as Stapelia gordonii Masson. Robert Sweet of England placed Stapelia gordonii into a new genus in 1830. The genus was Hoodia, named after Mr Hood, a well known succulent grower in Britain. Finally, Hoodia gordonii was first validly published by Joseph Decaisne in 1844 after the previous generic names were declared invalid (Bruyns, 2005:92; Van Heerden, 2008:434-435). H. gordonii is a spiny succulent that has rows of small thorns present along the fleshy, grey-green to grey-brown stems. Only one stem is produced in the early stages but as the plant matures as many as 50 individual branches are formed from the common base. It is usually about as broad as it is tall and can attain a height of up to 1 metre. The large flowers, borne on or near the terminal apex, are usually flesh-coloured but more intense purple-red flowers are found in some Namibian populations. Variation in flower size is common (50-100 mm) with flowers decreasing in size as the flowering period advances. The plant usually flowers in August/September and the large number of flowers almost completely obscure the stems. Pollination occurs as a result of the attraction of flies and blowflies to the flowers due to their unpleasant carrion-like smell. The seed capsules, produced in October and November, resemble antelope or goat horns and contain numerous flat, light brown seeds with silky seed hairs attached to one end (Van Wyk and Wink, 2004:171; Bruyns, 2005:115,117, Olivier, 2005:S.a.). Figure 2.1 shows H. gordonii in flower and Figure 2.2 the intact stem as well as cross-sections of dried and freshly cut stems. H. gordonii is widely distributed throughout South Africa and Namibia (Figure 2.3). It grows in parts of the Western Cape, the north and northwestern regions of the Northern Cape as far as Kimberley and parts of the southern Free State as well as in southwestern Namibia. It has a wide tolerance of growing habitats and can survive extreme heat (> 40 °C) as well as relatively low temperatures (-3 °C), but is susceptible to frost and mainly grows in summer rainfall areas. It is found in a variety of habitats such as dry sand, stony slopes or barren, flat areas (Bruyns, 2005:115; Olivier, 2005:S.a.). 8 FIGURE 2.1: Hoodia gordonii in flower (Photograph by AM Viljoen) FIGURE 2.2: Hoodia gordonii cross-sections of dried and freshly-cut stems (Photographs by AM Viljoen) FIGURE 2.3: The natural distribution of H. gordonii in South Africa and Namibia indicated in orange 9 2.2.2 Ethnopharmacology Hoodia species are reported to have been consumed by the San people of South Africa and Namibia as appetite and thirst suppressants during long hunting trips and when food supplies were low (Glasl, 2009:300). Usually, a small piece of the fresh stem is consumed after peeling to remove the thorns. Hoodia currorii, H. flava, H. gordonii and H. pilifera have all been consumed fresh as appetite and thirst suppressants and H. currori has been used to treat indigestion, hypertension, diabetes and stomach ache while Hoodia officinalis subsp. officinalis was used to treat pulmonary tuberculosis. According to records H. pilifera (‘ghaap’) was consumed more often as it has a cool and watery taste, while H. gordonii has a persistent bitter taste and was considered inferior, emphasised by the disparaging vernacular names such as ‘muishondghaap’ and ‘jakkalsghaap’ referring to it only being fit for consumption by animals. However, it is said that the bitterness is less prominent after good rains and the juicy stems would then be eaten raw or cooked. Adding to confusion, several species of stapeliads are referred to vernacularly as ‘ghaap’ (Van Wyk and Gericke, 2000:70; Van Wyk and Wink, 2004:171; Van Heerden, 2008:345; Glasl, 2009:302). In many papers on H. gordonii it is wrongly stated that it was also used as a cure for abdominal cramps, diabetes, indigestion, hypertension and so forth, but these applications are in fact attributed to other Hoodia species as mentioned previously. The source quoted refers to Hoodia species in general and these uses as well as the uses for H. gordonii are described in a paper by Van Heerden (2008:345). From this detailed ethnopharmacological account, H. gordonii was also used for the treatment of tuberculosis and the honey from the flowers could be used to treat cancer (Foden, 2005:S.a; Van Heerden, 2008:345). 2.2.3 Commercialisation The most important developments in the research and commercial development of H. gordonii is depicted in Figure 2.4. The use of Hoodia species, especially H. pilifera, by the San as a source of food and as a water substitute was recorded in 1932 and 1937. As a result of this the National Food Research Institute which is part of the Council for Scientific and Industrial Research (CSIR) in South Africa included Hoodia species in an investigation into edible wild plants in 1963. The 10 CSIR further investigated the appetite-suppressant activity of extracts of Hoodia species making a breakthrough with the structure elucidation of the compound called P57AS3 (commonly known as P57). The process to extract and synthesise P57 was patented in South Africa in 1995. In 1998, this was followed by a patent granted by the World Intellectual Property Organisation on pharmaceutical compositions with appetite suppressant activity. In the same year a licensing agreement to further develop P57 was signed between the CSIR and Phytopharm. Phytopharm, a British pharmaceutical company specialising in phytomedicines, in turn sub-licensed the pharmaceutical company Pfizer for further development and commercialisation. Due to a merger between Pfizer and Pharmacia in 2003, the natureceuticals branch responsible for the development of P57 was closed (Van Heerden, 2008:434; Glasl, 2009:301; Wynberg, Schroeder and Chennels, 2009:95). There were also unconfirmed reports that additional reasons for the discontinuation included difficulty in synthesising the P57 molecule and that the synthetic molecules were not as effective as the natural molecules. FIGURE 2.4: Timeline of the most important developments in the research and commercial development of H. gordonii 11 Despite wide media speculation on the safety of H. gordonii, no cases of adverse effects have been confirmed except for one study reporting two cases, one of acute hepatitis and another of anticholinergic syndrome for preparations containing Hoodia and concomitant drugs. In 2004, a patent was licensed to the company Unilever for the incorporation of H. gordonii extracts into food products to produce functional foods. Despite having invested vast sums of money in this project, including the development of an extraction facility, these plans were abandoned in December 2008 due to safety and efficacy concerns. Unilever also terminated all activities related to Hoodia in South Africa on 31 March 2009. Phytopharm remains optimistic about the Hoodia programme and insists that it will find other partners for further development of what should be a very lucrative industry (Wynberg, Schroeder and Chennels, 2009:96-97). Various pharmaceutical companies invested vast sums of money before withdrawing from this project. In addition, many farmers invested in H. gordonii crops during the “Hoodia boom” or uprise in the commercial interest in this plant. Cultivation of the H. gordonii plant is not easy and it takes a long time to mature (Van Heerden, 2008:435). Plants were initially grown from seed, very expensive at the time, greenhouses had to be built, and watering systems were needed. After all this the distribution of fake products led to a substantial decrease in the price of H. gordonii raw material as described in the newspaper article depicted in Figure 2.5. H. gordonii is cultivated for commercial purposes on large scale in South Africa and Namibia (Figure 2.6) as well as other parts of the world, though interest in continued cultivation has waned. 12 FIGURE 2.5: An article on H. gordonii published in the Sunday Times newspaper on 20 September 2009 FIGURE 2.6: Photographs of commercially cultivated H. gordonii in greenhouse tunnels in Namibia and South Africa and the processing of the raw material by slicing (Photographs by AM Viljoen) 13 2.2.4 Intellectual property rights and benefit sharing issues The San, more popularly known as ‘Bushmen’, are the oldest inhabitants of Africa and possess a wealth of knowledge about local biodiversity in Botswana, Namibia, South Africa and Angola. These hunter-gatherers used to live in small nomadic groups covering a vast expanse of land but oppression, genocide and dispossession has led to loss of land, culture and identity. Only a minority of the San people have secured rights to live on their own land, while others work as labourers on commercial game/cattle ranches or reside in government resettlement villages. The common factor among all San is the hopeless poverty they live in. The use of Hoodia by the San as a thirst and appetite suppressant is undisputed. Records of this knowledge resulted in subsequent research and patents on Hoodia and licensing agreements for its commercialisation were granted without the knowledge of the San who are the originators of this knowledge (Wynberg, Schroeder and Chennels, 2009:89-94). Until 2001, development and commercialisation continued without acknowledgement of the San leading them to launch a painstakingly long and hard challenge against the CSIR on their Hoodia patent. This only occurred after the foreign media were informed of this potential exploitation of the San people, depicted as poor and emaciated, in stark contrast with images of rich obese Westerners. This act portrayed as biopiracy by large pharmaceutical companies resulted in an eagerly followed high-profile law suit and forced negotiations concerning benefit-sharing (Wynberg, Schroeder and Chennels, 2009:100-102). This case raised issues on ownership of knowledge and resources. Previously, renewable resources such as plants, animals and microorganisms were regarded as common heritage of humankind. Any person could remove these resources from nature and use them to develop commercial products. However, the establishment of the Convention on Biological Diversity (CBD) in 1992 declared that biological resources fall under the sovereignty of the states. The convention encompasses the conservation of biological diversity and the sustainable use of its components as well as fair and equitable sharing of benefits. Traditional knowledge is considered to be removed from the common heritage of humankind and therefore is subject to formal prior informed consent and benefits must be shared equitably (Wynberg, Schroeder and Chennels, 2009:11). The San 14 opted to share in the royalties rather than continue with an expensive law suit to prevent use of their knowledge culminating in a mutual benefit agreement which was finally signed in 2003. It was important to ensure that the agreement was inclusive of all of the San people and would benefit all equally as the traditional knowledge is extended into several countries (Wynberg, Schroeder and Chennels, 2009:102). According to the agreement between the CSIR, its licensees and the South African San Council, the parties will share royalties from potential sales of drugs or other products derived from Hoodia. The money, 8% of all payments that the CSIR receives from licensees and 6% of all royalties after commercialisation, will be paid into a Trust controlled by the San to be used to the benefit of the San people (Glasl, 2009:302). However, the payment of these royalties is currently not applicable since no products have yet been developed out of this agreement. The Sunday Times newspaper of 20 September 2009 (Figure 2.6) describes how the San are still waiting for their royalties. It is clear that the commodification of traditional knowledge is complex and the laws concerning intellectual property rights are continuously evolving to adapt to issues relevant to each natural resource. 2.2.5 Phytochemistry The pioneering group of scientists who investigated Hoodia species, Van Heerden et al. (2007:2545), isolated two steroidal glycosides named Compound 1 and 2 from H. gordonii and H. pilifera extracts. Compound 1 is referred to as P57AS3, more commonly known as P57. Since then many new structural analogues have been isolated from Hoodia gordonii: eleven oxypregnane glycosides – hoodigoside A-K (Pawar et al., 2007:524); ten pregnane glycosides – hoodigoside L-U (Pawar, Shukla and Khan, 2007:882-884); ten steroidal glycosides – gordonoside A-I, L (Dall’Acqua and Innocenti, 2007:559); seven pregnane glycosides – hoodigoside V-Z and hoodistanaloside A-B (Shukla et al., 2009:675); and Formula 7-8 and 12 (Abrahamse, Povey and Rees, 2007:4-5). The backbone chemical structures of these compounds are the aglycones hoodigogenin A, calogenin, hoodistanal, dehydrohoodistanal and isoramanone formed as a result of acid and/or enzymatic hydrolysis. The structures of the compounds isolated from H. gordonii and the respective aglycone of each analogue is depicted in Annexure 1.1. The discovery of numerous new compounds provides the opportunity for research into their 15 biological activity and has also been used to develop new quality control methods for H. gordonii products (Avula et al., 2008:722). 2.2.6 Biological activity 2.2.6.1 Anti-obesity activity Scientific studies on the effect of H. gordonii and/or P57 on weight loss is scarce because much of the necessary research has not been done but also partly due the fact that some of the information is proprietary. The first study that investigated the mechanism of action of P57 was that of MacLean and Luo (2004:1-11) who determined the effect of P57 (P57AS3) injected directly into the third ventricle of the brain in rats. They found that intracerebroventricular (i.c.v) injection of P57 led to a reduced food intake of 50-60% over the 24-h period following injection. The duration of effect was about 24-48 h and dose-dependent. In addition, pair-fed rats showed increased adenosine triphosphate (ATP) in hypothalamic neurons compared to animals treated with the solvent only. As it has been reported that intrahypothalamic 5ʹ adenosine monophosphate-activated protein kinase (AMPK) may regulate food intake and that AMPK is regulated in part by ATP, this central nervous system mechanism of action of P57 is plausible. However, recently a paper was published which seems to refute these findings. Pharmacokinetic studies after oral administration of P57 revealed that P57 did not reach the brain indicating that its mechanism of action may not be due to increased ATP in the hypothalamus (Madgula et al., 2010a:1585). The first scientific reports of in vivo anti-obesity activity were by Tulp and Harbi (2001:A404; 2002:A654) where food intake in both lean (moderate reduction) and obese (marked reduction) rats were voluntarily reduced with a concomitant reduction in body weight. The effective dose of Hoodia species was determined to be 1.8-2.7 kg/kg of bodyweight (Tulp and Harbi, 2001:A404). In the follow-up study, 3% (w/v) aqueous homogenate of dehydrated Hoodia plant material was administered to lean and obese mice resulting in a decreased food intake within 48 h. The body weight of the obese rats, initially twice the weight of the lean rats, decreased to near normal weights after two to three weeks of Hoodia feeding and the body fat mass in lean and obese rats were decreased compared to rats fed a normal diet (Tulp and Harbi, 2002:A654). 16 However, these studies were only published as abstracts which severely limits the value in terms of interpretation of the results. In addition, the administration of Hoodia species is indicated and not H. gordonii specifically. In another study, P57 isolated from H. gordonii led to a decreased food consumption and body mass over an 8-day period after administration by oral gavage to rats in doses of 6.25-50 mg/kg. In comparison with the known appetite-suppressant compound fenfluramine at 15 mg/kg/day, P57 (30 mg/kg/day) caused a suppression in body weight while fenfluramine administration resulted in a small decrease in food intake and an increase in body weight, though less than in the control group (Van Heerden et al., 2007:2545; 2549-2553). H. gordonii has also been fed to broiler chickens to determine the effect on productivity and carcass characteristics. Daily supplementation with an H. gordonii meal (300 mg) had no effect of feed intake, digestibility and growth but fat pad weights were reduced by 40%. Therefore H. gordonii supplementation may have a beneficial effect on carcass quality (Mohlapo et al., 2009:1591), but the mechanisms need to be further investigated. Clinical studies in humans have apparently been done but remain unpublished partly due to the proprietary nature of drug development research. In human studies undertaken by Phytopharm, 20 obese free-feeding volunteers were treated with an extract of H. gordonii. Exercise was not required and it was not expected of volunteers to follow any eating plan. The subjects who received the capsule containing the extract reduced their daily calorie intake of their own accord, by approximately 1000 calories per day. The decreased food intake resulted in weight loss of about 2 kg and a reduction in blood glucose and triglycerides (Holt and Taylor, 2006a:108; Glasl, 2009:305). The Phytopharm press release reporting on this study has since been removed from their website. In another paper on H. gordonii, the authors report an open-label observational study performed by one of the authors. A specific product called Hoodia Supreme® said to contain 400 mg of pure H. gordonii per capsule was supplied to eight obese participants twice daily for a 4-week period. Data was gathered by direct questioning on eating habits as well as weight measurement. Weight loss (± 1-7 kg), reduction in appetite, a voluntary reduction in calorie intake (estimated 5001000 calories/day), a reduction in craving for carbohydrate-containing foods and a mild energising effect was reported. In the same paper there was reference to a 17 study by Goldfarb, but the original source as mentioned in the references could not be located on the internet. It stated that 7 participants receiving DEX-L10® containing 500 mg of H. gordonii for a 28-day period reported a median weight loss of ± 5 kg with no alteration in dietary or daily activity habits (Holt and Taylor 2006b:103). A summary of the in vivo studies performed on H. gordonii and P57 is shown in Table 2.1. TABLE 2.1: Summary of studies on the in vivo anti-obesity activity of H. gordonii and P57 In vitro/In vivo Tested in: Hoodia gordonii or P57 used and administration Effect(s) In vivo (animals) Rats Hoodia species crude aqueous homogenate/semipurified extracts (oral) Reduced food intake, body weight, gonadal fat pad weight and blood glucose Rats P57 (oral gavage) Reduction in food intake and body weight Rats P57 (i.c.v injection) Reduction in food intake Chickens H. gordonii meal (oral) Reduction of fat pad weights Obese males (20) Hoodia extract (oral) Reduction in food intake, body weight, blood glucose and triglycerides* Obese subjects (2 male, 6 female) Hoodia Supreme® capsules containing Hoodia gordonii raw material (oral) Reduction in food intake, appetite, body weight, craving for carbohydrates and a mild energising effect reported Overweight subjects (7) DEX-L10® capsules containing Hoodia gordonii raw material (oral) Reduction in body weight and calorie intake* In vivo (humans) * Could not be verified as the original source as cited could not be located The information related to clinical studies is scarce and for most of these studies the data are not currently accessible, raising even more questions about the use of H. gordonii as an anti-obesity preparation. It is evident that proper clinical studies in humans are needed to validate claimed anti-obesity effects. These intriguing and promising results may have triggered the entry of several drug development partners into agreements with Phytopharm, but despite the positive results the 18 question remains why two renowned pharmaceutical companies were unable to publish data and abandoned what seems to be a potential goldmine. 2.2.6.2 Other biological activities Hoodigosides A-K and P57 was evaluated for cytotoxic and antioxidant properties in vitro. These compounds were found not to be cytotoxic and did not show any inhibition of growth against a panel of cell lines including SK-MEL, KB, BT-549, SKOV-3, VERO and LLC-PK1 at the highest concentration of 25 µg/ml. No antioxidant activity was recorded in HL-60 cells and intracellular reactive oxygen species (ROS) generation was not inhibited (Pawar et al., 2007:533). In a US patent application the gastroprotective effect of P57 and the effects on gastric acid secretion and gastric motility were described. Gastric damage due to oral administration of aspirin was reduced by 91.5% for P57 and 93.2% for cimetidine (a pharmaceutically registered gastric acid secretion inhibitor). Oral administration of a spray dried extract at a dose of 50 mg/kg inhibited acid output by 50% and gastric emptying by 26%. Intraduodenal administration of purified P57 at a dose of 10 mg/kg decreased acid output by 88% while subcutaneous administration decreased acid output by 43% (Hakkinen, Horak and Maharaj, 2005:5-6). These effects were promising but Madgula et al. (2010b:65) found 100% degradation of P57 in simulated gastric fluid, which raised more questions about the actual active molecule. Antidiabetic activity has also been claimed, but ethnopharmacalogically it is H. currori that has been used as an antidiabetic (Van Heerden, 2008:435). In 2006, a patent was filed for antidiabetic properties from extracts of Hoodia. An acute (20 to 160 mg/kg) and chronic study (60 and 120 mg/kg) was conducted with P57 (Compound 1) and food and water intake, bodyweight, glucose, insulin and leptin levels in male diabetic ZDF rats were assessed. In the chronic study the doses were reduced to 60 and 30 mg/kg on day 7 due to concerns about the 50% reduction in food intake. The results showed that P57 had positive effects on bodyweight, normal glycaemia was maintained and water intake of diabetic rats were reduced within 4 days to that of the control lean rats (Rubin, Bindra and Cawthorne, 2006:11-14). 19 2.2.7 Toxicity Data pertaining to the toxicity/safety of H. gordonii is scarce. In a five year study in Italy, 233 spontaneous reports on adverse reactions to natural health products were collected and analysed. Anti-cholinergic syndrome was noted with one preparation containing 100 mg H. gordonii raw material (Menniti-Ippolito et al., 2008:629). However, it also contained 200 mg Ma huang (Ephedra sinica) raw material which most likely caused this effect (Homoud, 2009:658). Another product containing Hoodia was associated with a case of acute hepatitis, but six pharmaceutical preparations were taken concomitantly (Menniti-Ippolito et al., 2008:269). Acute toxicity studies were described in a patent application by Van Heerden et al. (2002:52-53). A plant extract administered orally to mice in doses of 100 mg/kg up to 3028.5 mg/kg, revealed no clinical signs of toxicity. Doserelated reversible histopathological liver changes in the form of moderate cloudy swelling and hydropic degeneration of hepatocytes was recorded for a dose of 200 mg/kg. The number of animals used in this study were small (n = 2 per dosage group) but the fact that all of the groups except one showed liver changes on histopathological examination warrants further investigation. During anti-obesity activity experiments in congenic lean and obese rats, high quantities of H. gordonii were orally administered and no adverse effects were noted (Tulp and Harbi, 2002:A654). No adverse effects were reported for the clinical study performed in humans but it should be kept in mind that not much was reported at all due to the proprietary nature and the occurrence of side effects is certainly not the type of information a company developing a product would attract attention to. Its traditional use by the San for decades bodes well for safety, but is only applicable to fresh H. gordonii and not extracts or molecules isolated from the plant as these would be in concentrated form. The scarcity of information in terms of scientific studies on the efficacy and safety of H. gordonii, which is such a widely-consumed anti-obesity product, stimulated an investigation into other plants used for obesity. These plants are also consumed by the general public despite the fact that safety and efficacy studies are in most cases lacking. In addition, the safety of some of these products has been called into question as side-effects and even deaths were reported highlighting the need for proper clinical studies on all natural products. 20 2.3 PLANTS USED FOR THEIR ANTI-OBESITY PROPERTY 2.3.1 Introduction Obesity is a serious health problem that has become one of the most common health concerns of modern times and should be considered a global epidemic. In 2005 it was estimated that 1.6 billion adults were overweight globally with at least 400 million classified as obese. These numbers are projected to increase to 2.3 billion overweight and 700 million obese adults by 2015. The number of obese teenagers and children in particular has increased dramatically in recent years leading to heightened awareness of this global problem. At least 20 million children under the age of five years were overweight in 2005. Obesity is associated with and can lead to many disease conditions including type-2 diabetes, cardiovascular disease, hypertension, sleep apnoea and cerebrovascular accidents amongst many others (WHO, 2006:S.a.). Obesity occurs as a result of a higher energy intake than energy expenditure and weight loss can only be achieved by a negative energy balance (Finer, 2002:718) such as through a combination of exercise and dieting. The body mass index (BMI = [weight in kg]/[height in m]2) is a clinical calculation based on the weight and height of the patient that can be used for diagnosis of obesity. The normal range for BMI is 19.8 – 26.0, while a range between 26.1 – 29.0 indicates an overweight state and a value above 30 would diagnose a patient as obese (Mycek, Harvey and Champe, 2000:474). The commercial market for anti-obesity preparations is considerable due to public awareness of the ill-effects of obesity and the general perception that being obese is less than attractive. This multibillion dollar industry expands on a daily basis as new ‘miracle cure’ products become available. Anti-obesity preparations of natural origin are particularly appealing to consumers as the general perception is that if it is natural it must be effective and safe. Many plant products are currently commercially available but unfortunately most of them have been poorly researched, if at all. Not surprisingly, associations between consumption of herbal products and instances of toxicity have been and are still being made. For example, severe neurologic and cardiovascular complications have been associated with the consumption of ephedra alkaloids (Ephedra sinica) in 21 hundreds of cases. Ephedra-containing products were banned by the Food and Drug Administration (FDA) in 2004 compelling manufacturers to develop “ephedrafree” products. Since then Citrus aurantium (bitter orange), which contains the sympathomimetic agent synephrine, has been frequently used as an ephedra replacement in supplements. Synephrine consumption has been associated with tachycardia, exercise-induced syncope and myocardial infarction (Nykamp, Fackih and Compton, 2004:812; Bouchard et al., 2005:541; Firenzuoli, Gori and Galapai, 2005:247). Conventional or allopathic medicines are subjected to rigorous testing procedures before they are allowed to enter the commercial market. In the natural products industry, the most basic scientific studies which should precede commercial development are in general lacking. If nothing else, toxicity studies should at least be conducted. In addition, medical practitioners should recognise the extensive use of natural products and include this information in patient history files as this would ensure that more information especially in terms of drug interactions is obtained. It is imperative that these plants and herbal products be extensively investigated in terms of quality, efficacy and especially safety in order to validate their widespread consumption. 2.3.2 Mechanisms of action of anti-obesity preparations Weight loss can only occur if energy consumption is less than energy expenditure, i.e. if a negative energy balance exists. The obvious approach of exercise combined with dieting has a relatively low success rate and therefore alternative forms of obesity treatment including surgery and medications have been developed (Goodman et al., 1992:217). A substance causing a reduction in food intake through satiety mechanisms increases energy expenditure and efficiency, resulting in dietary compensation. Sympathetic nervous system stimulation causes β 1-, β 2-, β 3- and α -adrenoreceptor activation resulting in decreased appetite. Compounds that selectively stimulate β 3-receptors, cause the thermogenic effects of sympathomimetic agents but do not exhibit the side effects associated with general stimulation such as increased blood pressure, increased heart rate (β1), increased smooth muscle contraction and tremors (β2) (Finer, 2002:734; Ohkoshi et al., 2007:1257; Arbo et al., 2008:2770; Arbo et al., 2009:114). Activation of 3ʹ-5ʹ-cyclic adenosine monophospate (cAMP) promotes the breakdown of stored fat, increases the basal metabolic rate, increases 22 utilisation of body fat and regulates the thermogenic response to food (WesterterpPlantenga, Lejeune and Kovacs, 2004:1196). Acetyl coenzyme A carboxylase (ACC) inhibitors moderately retard fat accumulation as it affects fatty acid biosynthesis (Watanebe, Kawabata and Kasai, 1999:489). Activation of the 5ʹ adenosine monophosphate-activated protein kinase (AMPK) pathway results in increased fatty acid oxidation, inhibition of cholesterol synthesis, lipogenesis and inhibition of adipocyte lipolysis and lipogenesis. Increased expression of uncoupling proteins UCP1, UCP2, and UCP3 results in increased thermogenesis and this process is induced by substances like caffeine which results in the dissipation of energy in the form of heat (Matsuda et al., 2003:2408; WesterterpPlantenga, Lejeune and Kovacs, 2004:1196-1197; Chen et al., 2009:785;792; Murase et al., 2009:78). Conventional pharmaceutical anti-obesity drugs are few and are classified either as centrally acting or peripherally acting. Centrally-acting products include sympathomimetics with central nervous system (CNS) stimulant properties such as phentermine (Duromine®), phendimetrazine (Obex-LA®, Obesan X®), amfepramone (Tenuate Dospan®), norpseudoephedrine (Nobese No.1®) and agents that specifically inhibit re-uptake of noradrenaline and serotonin in the CNS such as sibutramine (Reductil®). Peripherally acting products such as orlistat (Xenical®) inhibits lipase in the gastrointestinal tract leading to decreased dietary fat absorption (Gibbon, 2008:69). Not all of these preparations are registered in all countries, as the final decision lies with the regulatory authorities in each respective country. Amfepramone, phentermine, phendimetrazine and norpseudoephedrine (less potent) exhibits amphetamine-like effects including increased energy and wakefulness, the main reasons for its abuse potential, and decreased appetite. Due to extensive abuse, norpseudoephedrine containing preparations were recently converted from over-the-counter to prescription-only medicines in South Africa. Sibutramine inhibits the re-uptake of released serotonin and noradrenaline from hypothalamic neurons causing a predominant effect on satiety and causes thermogenesis. It has no effects on dopaminergic neurons. Adverse effects include nausea, insomnia, dry mouth, rhinitis and constipation and in some individuals increases heart rate and blood pressure. Patient deaths have been attributed to sibutramine (Finer, 2002:727) and 23 ephedrine use. Orlistat inhibits pancreatic and gastric lipases which functions to catalyse the hydrolysis of ingested fats, resulting in a reduction of dietary fat absorption and consequently high faecal fat excretion. This is bound to have a considerable effect as between 35 and 45% of energy intake in western diets is attributed to dietary fat. Unfortunately, malabsorption of fat leads to rather uncomfortable side effects such as flatus with discharge, oil spotting from the rectum, faecal incontinence, faecal urgency, loose or liquid stools and malabsorption of fat-soluble vitamins (Finer, 2002:730; Gibbon, 2008:69). However, only 1% of patients experience side-effects severe enough to withdraw from treatment, while 25% of patients may report such side-effects. It has also been suggested that these side-effects may aid in modification of the patient’s lifestyle by avoiding high-fat meals to avoid the side-effects and thereby it may contribute to healthier choices in food intake (Finer, 2002:730). Many new mechanisms to induce weight loss have been proposed as a result of ongoing research and new molecules that utilise ‘old’ pathways are still discovered. The focus of many research groups has been on discovering new effective pancreatic lipase inhibitors from plants with fewer side effects. This may be due to the fact that it is not a complicated study and large numbers of plants can easily be screened. In these studies, effects are compared to mainly Orlistat® (positive control) with few studies reporting on associated side-effects. What makes Orlistat® and other pancreatic lipase inhibitors desirable is that they are acting peripherally thereby decreasing chances of systemic side effects. Similar to pancreatic lipase, α-amylases function to digest carbohydrates into easily absorbable sugars. Inhibition of these enzymes therefore results in decreased absorption of sugars which have been shown to have a beneficial effect in diabetes and most probably obesity. Leptin is a protein hormone that is encoded by an obesity gene and regulates appetite and energy expenditure. Circulating levels of leptin are generally proportional to the amount of body fat (decreased body fat = decreased leptin). However, in leptin gene deficient individuals injections of leptin decreases food intake and increases thermogenesis and physical activity. Peroxisome proliferator-activated receptors (PPARs) and CCAAT (cytidine-cytidine-adenosine24 adenosine-thymidine) -enhancer-binding proteins (or C/EBPs) are important adipogenic transcription factors involved in the inhibition of differentiation into adipocytes. Inhibition of differentiation of preadipocytes into adipocytes is another common mechanism exhibited by many of the plants used as anti-obesity preparations. Hormone-sensitive lipase is a rate-limiting enzyme catalysing the lipolysis of triacylglycerol and diacylglycerol following the phosphorylations of the enzyme by protein kinase A. Peptide YY is released predominantly from the lower intestine following a meal and decreases appetite in both lean and obese subjects. Ghrelin is predominantly released from the stomach and its level increases in times of hunger and decrease immediately post-prandially. Administration of ghrelin to rats and humans increases appetite (Murray et al., 2008:747; Pang, Choi and Park, 2008:183-184; Shi et al., 2009:72;75; Smeets and WesterterpPlantenga, 2009:230). Adiponectin levels are inversely correlated to body fat percentage in adults (Kang et al., 2007:4389;4395). Cholesystokinin A (CCK A) acts as a physiological satiety factor and activation of its receptors peripherally, slows gastric emptying and induces satiety. Fatty acid synthase (FAS), an important enzyme in the fatty acid synthesis pathway, catalyses the reductive synthesis of long-chain fatty acid from acetyl-coenzyme A (CoA) and malonyl-CoA. FAS inhibition leads to inhibition of feeding and weight loss. Down-regulation of sterol-regulatory-element-binding protein (SREBP) decreases sterol biosynthesis (Watanabe, Kawabata and Kasai, 1999:489; Chen et al., 2009:790-792). It is clear that many targets for anti-obesity agents exist but the only effect that should qualify a product as an anti-obesity preparation is its positive effect on weight loss. The treatment can only be considered successful if it induces sufficient weight loss from the initial body weight (5-10%), prevents further weight gain and allows long term maintenance of the weight loss once it is achieved (Alarcon-Aguilar, 2007:69). 2.3.3 Other plants popularly used as anti-obesity agents Many plants are being consumed for their anti-obesity properties. Table 2.2 lists the botanical names, family, vernacular names, plant parts used, mechanisms of action, in vitro or in vivo studies, toxicity and anti-obesity compounds isolated for 63 plants that have been investigated scientifically for their anti-obesity properties. 25 TABLE 2.2: Summary of the plants investigated for their anti-obesity properties in alphabetical order by species name Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated References Acanthopanax senticosus (Rupr. et Maxim.) Harms. Siberian ginseng Fruit, cell cultured herb extract Pancreatic lipase inhibitor Yes Yes No Unknown, expect similar to A. sessiliflorus Silphioside F, copteroside B, hederagenin 3-O-β-Dglucuronopyranoside 6ʹO-methyl ester, gypsogenin 3-O-β-Dglucuronide Cha, Rhee and Heo, 2004:422-428; Li et al., 2007a:1087-1089 Wu jia Leaves Pancreatic lipase inhibitor Yes Yes No Not significant Chiisanoside, 11deoxyisochiisanoside, isochiisanoside, sessiloside Yoshizumi et al., 2006:335341; Yoshizumi et al., 2008:1126-1128 Japanese horsechestnut Seeds Pancreatic lipase inhibitor, suppresses blood triacylglycerol level, suppresses increase in body weight, increases triacylglycerol level in feaces Yes Yes No Unknown Escin Ia, Escin Ib, Escin IIa, Escin IIb Van Wyk and Wink, 2004:34; Kimura et al., 2006:1657-1665; Hu et al., 2008:12-15; Kimura et al., 2008:4783-4788 Lesser galangal Rhizomes Pancreatic lipase inhibitor Yes Yes No Unknown 3-methylethergalangin, 5-hydroxy-7-(4ʹ-hydroxy3ʹ-methoxyphenyl)-1phenyl-3-heptanone Shin, Han and Kim, 2003:854-856; Shin et al., 2004:138-140; Van Wyk and Wink, 2004:43 Peanut Nuts, shells Pancreatic lipase inhibitor (peanut shell), lipolysis of 3T3L1 adipocytes (peanut shell), increases satiety, increase in energy efficiency and expenditure, dietary compensation Yes Yes Yes None observed (peanut shell) and none known (peanuts) No Coelho et al., 2006:585591; Moreno et al., 2006a:2797-2802; Higgs, 2007:353-356; Mattes, KrisEtherton and Foster, 2008:1741S-1744S Araliaceae Acanthopanax sessiflorus (Rupr. et Maxim) Seem. Araliaceae 26 Aesculus turbinata Blume Hippocastanaceae Alpinia officinarum Hance Zingiberaceae Arachis hypogaea L. Fabaceae 26 Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated References Camellia sinensis (L.) Kuntze Tea Leaves Stimulates sympathetic nervous system, thermogenesis, inhibits gastric and pancreatic lipase, increase fat oxidation, suppresses leptin, increase triacylglycerol excretion, decrease adipocyte cell size, suppresses food intake, suppress adipocyte differentation, increase expression of genes involved in fatty acid synthesis and oxidation, decrease lipoprotein lipase, decrease hormonesensitive lipase, decrease UCP-2, activation of AMPK, increase adiponectin, increase ghrelin Yes Yes Yes Hepatotoxicity Epigallocatechin gallate Van Wyk and Wink, 2004:75; Han et al., 1999:98-104; Sayama et al., 2000:481-484; Han et al., 2001:1459-1464; Chantre and Lairon, 2002:3-8; Lin, Della-Fera and Baile, 2005:982-989; WesterterpPlantenga, Lejeune and Kovacs, 2004:1195-1203; Auvichayapat et al., 2008:486-491; Hsu et al., 2008:363-369; Sarma et al., 2008:469-482; Chen et al., 2009:784-792; Kapetanovic et al., 2009:28-36; Lambert et al., 2010:409-415 Caper Fruits Unknown No Yes No Unknown No Inocencio et al., 2000:70-73; Eddouks, Lemhadri and Michel, 2005:345-349; Lemhadri et al., 2007:106109 Red pepper Fruits Thermogenesis, increase energy expenditure, increase secretion of catecholamines, enhance expression of the adiponectin gene, activation of AMPK, reduce food intake, increase satiety, decrease ghrelin levels, down-regulates expression of UCP1 and UCP2, suppress fatty acid Yes Yes Yes Not significant Capsaicin, capsiate, dihydrocapsiate, 9oxooctadeca-10,12dienoic acids Watanabe, Kawabata and Kasai, 1999:489-492; Kanki et al., 2003:1337-1342; Masuda et al., 2003:24082415; Chanda et al., 2005: 66-75; Hwang et al., 2005:694-698; WesterterpPlantenga, Smeets and Lejeune, 2005:682-688; Kang et al., 2007:4389- Theaceae 27 Capparis spinosa L. Capparaceae Capsicum annuum L. Solanaceae 27 Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated biosynthesis (ACC inhibitor) Caralluma fimbriata Wall. References 4395; Smeets and Westerterp-Plantenga, 2009:229-234 Caralluma Stems Unknown No No Yes Unknown No Kuriyan et al., 2007:338-343 Nomame herba Aerial parts, fruits Pancreatic lipase inhibitor Yes Yes No Unknown Proanthocyanidin, flavan dimers ((2S)-3ʹ,4ʹ,7trihydroxyflavan-(4αà8)catechin) Hatano et al., 1997:893-899; Yamamoto et al., 2000:758763 Khat Leaves Sympathomimetic effect, decrease leptin levels, decrease blood triacylglycerol level Yes Yes Yes Liver, urinary, reproductive Cathinone, cathine Al-Habori et al., 2002:209216; Van Wyk and Wink, 2004:84; Al-Hebshi and Skaug, 2005:299-305; AlDubai, Al-Habori and AlGeiry, 2006:632-635; AlZubairi et al., 2008:298-302; Gibbon, 2008:401; Mahmood and Lindequist, 2008: 271-277; Murray et al., 2008:747-750 Roman chamomile Aerial parts Unknown No Yes No Unknown No Lemhadri et al., 2007:106109 Bitter orange Fruits Increased sympathomimetic activity, stimulates lipolysis, increased thermogenesis Yes Yes Yes Sympathomimetic effects Synephrine, octopamine Calapai et al., 1999:586592; Colker et al., 1999:145152; Fugh-Berman and Apocynaceae Cassia mimosoides L. var. nomame Makino Fabaceae 28 Catha edulis (Vahl) Endl. Celastraceae Chamaemelum nobile (L.) All. Asteraceae Citrus aurantium L. Rutaceae 28 Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated References Meyers, 2004:698-703; Marchei et al., 2006:14681472; Arbo et al., 2008:2770-2774; Arbo et al., 2009:114-117 Citrus unshiu Marcow. Satsuma mandarin orange Fruit (peel, segment wall extract, membranes) Inhibits lipase activity, stimulates lipolysis Yes Yes No Unknown (synephrine content suggest similar to C. aurantium) Hesperidin, pectin Kawaguchi et al., 1997:102103; Tsujita et al., 2003:340345; Tsujita and Takaku, 2007:547-551; Edashige, Murakami and Tsujita, 2008:409-414; Tsujita and Takaku, 2008:102 Coleus Roots Activates cAMP Yes Yes Yes Not significant Forskolin Lieberman, 2004:330-332; Godard, Johnson and Richmond, 2005:1335-1342; Han et al., 2005a:449; Henderson et al., 2005:5462; Canová et al., 2006:200209 Turmeric Rhizomes Inhibits adipokine-induced angiogenesis, suppresses 3T3-L1 differentiation, increases oxidation and fatty acid esterification, downregulates the expression of key transcription factors Yes Yes No High dose CNS stimulation, haematological changes Curcumin, demethoxycurcumin, bisdemethoxycurcu-min Qureshi, Shah and Ageel, 1992:124-126; Asai and Miyazawa, 2001:2932-2934; Van Wyk and Wink, 2004:118; Ejaz et al., 2009:919-924 Sweet tea tree Leaves Pancreatic lipase inhibitor Yes Yes No Unknown No Kurihara et al., 2003:383385; Zhang et al., 2010:1491-1496 Rutaceae Coleus forskohlii Briq. 29 Lamiaceae Curcuma longa L. Zingiberaceae Cyclocarya paliurus (Batal.) Iljinsk. Juglandaceae 29 Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated References Dioscorea nipponica Makino Wild yam Roots/ rhizomes Pancreatic lipase inhibitor, suppresses blood triacylglycerol level Yes Yes No Not significant except at high doses (anorexia, asthenia, weight loss, piloerection, ataxia, syncope) Dioscin, diosgenin, gracillin, prosapogening A, prosapogenin C Kwon et al., 2003:14511456; Son et al., 2007:30633069; Qin et al., 2009:543550; Ma Huang Aerial parts Increase release of norepinephrine and inhibition of re-uptake (stimulates α- and β-adrenergic receptors ) Yes Yes Yes Toxic effects known, sympathomimetic stimulation Ephedrine Fontanarosa, Rennie and DeAngelis, 2003:1568-1570; Shekelle et al., 2003:15371545; Van Wyk and Wink, 2004:134; Morton, 2005:242-247; Kim et al., 2008b:128-137 Bee bee tree, Evodia Fruits Increase lipolysis in BAT, enhances heat dissipation, reduced food intake, decrease expression of orexigenic peptides (NPY, AgRP), increase leptin levels, inhibit adipogenesis in 3T3-L1 cells, inhibit differentiation in human white preadipocytes Yes Yes Yes None reported Evodiamine, rutecarpine Kobayashi et al., 2001:628632; Kim et al., 2008b:128137; Yang, 2008:1317-1321; Hu et al., 2010:259-267; Kim et al., 2009b:437-442; Shi et al., 2009:71-77; Wang, Wang and Yamashita, 2009:3655-3659 Goat's rue Aerial parts Activates AMPK in H4IIE rat hepatoma, HEK293 human kidney cells, 3T3-L1 adipocytes and L6 myotubes, down-regulation of fatty acid synthase and sterolregulatory-element-binding protein (SREBP) Yes Yes No Liver, lung effects Galegine, 1-(4 chlorobenzyl) guanidine hemisulfate Palit, Furman and Gray, 1999:1313-1318; Van Wyk and Wink, 2004:150; Mooney et al., 2008:16691676; Rasekh et al., 2008:21-26; Coxon et al., 2009:3457-3463 Dioscoreaceae Ephedra sinica Stapf. Ephedraceae 30 Evodia rutaecarpa (Juss.) Benth. Rutaceae Galega officinalis L. Fabaceae 30 Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated References Garcinia cambogia Desr. Brindal berry Fruit rind Inhibit ATP-citrate lyase (suppress fatty acid synthesis), suppresses appetite (decreased food intake), decrease leptin levels Yes Yes Yes Testicular toxicity in rats (disputed) (-)-Hydroxycitric acid Mattes and Bormann, 2000:87-93; Van Loon et al., 2000:1445-1449; Preuss et al., 2004:171-179; Soni et al., 2004:1513-1527; Burdock, 2005:1683-1684; Saito et al., 2005:411-418 Gardenia Fruit Pancreatic lipase inhibitor, suppresses serum triglyceride and cholesterol level Yes Yes No Genotoxic (genipin, gardenia yellow), hepatotoxic (genipin) Crocin, crocetin Ozaki et al., 2002:16031609; Lee et al., 2005:21062110; Sheng et al., 2006:116-122 Liquorice Roots Blocks 11β-hydroxysteroid dehydrogenase type 1 (topical), down-regulation of genes for fatty acid synthesis (mice) Yes Yes Yes Hypermineralo corticoidism, pseudoaldosteronisms (due to glycyrrhizic acid), not relevant to LFO, blood abnormalities and anticoagulation (rats LFO) Glycyrrhetinic acid (topical) Armanini et al., 2005:538542; Isbrukcer and Burdock, 2006:167-188; Tominaga et al., 2006:672-682; Kamisoyama et al., 2008:3225-3231; Nakagawa et al., 2008:2349-2357; Tominaga et al., 2009:169177 Chinese liquorice Roots Inhibits pancreatic lipase Yes No No Unknown Licochalcone A Won et al., 2007:1046-1049 Clusiaceae Gardenia jasminoides Ellis Rubiaceae Glycyrrhiza glabra L. 31 Fabaceae Glycyrrhiza uralensis Fisch. Fabaceae 31 Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated References Griffonia simplicifolia (Vahl ex DC) Griffonia Seeds Increases serotonin and circulating leptin levels Yes Yes Yes Serotonin syndrome in rats at high doses 5-HTP Yamada, Sugimoto and Ujikawa, 1999:49-51; Amer et al., 2004:137-142; Turner, Loftis and Blackwell, 2006:325-335 Gurmar Leaves Decrease fat digestibility, decreased food and water intake (1 study) Yes Yes Yes (in combi nation) None reported Gymnemic acids, gymnemate Shigematsu et al., 2001:713716; Ogawa et al., 2004:8; Preuss et al., 2004:171-179; Kanetkar, Singhal and Kamat, 2007:77-81; Luo et al., 2007:93-97 Jiaogulan immortality herb Aerial parts Pancreatic lipase inhibitor Yes No No None observed, mild adverse effects Dammarane-type triterpenoids (20S, 23R)3β-,20βdihydroxydamma-24dien-21-oic acid 21,23lactone, (20S, 24S)20,24epoxydammarane3β,12β,25-triol Attawish et al., 2004:539550; Razmovski-Naumovski et al., 2005:197-214; Bai et al., 2010:306-309 Roselle Calyces, dried flowers Inhibition of expression of adipogenic transcription factors C/EBPα and PPARγ through PI3K and MAPK pathway - inhibition of differentiation into adipocytes Yes Yes No Testicular toxicity reduced sperm count and spermatogenesis No Akindahunsi and Olaleye, 2003:161-164; Kim et al., 2003:499-504; Orisakwe, Husaini and Afonne, 2004:295-297; Van Wyk and Wink, 2004:170; AlarconAguilar et al., 2007:66-70; Kim et al., 2007:260-266; Sowemimo et al., 2007:427432; Carvajal-Zarrabal et al., 2009:in press Fabaceae Gymnema sylvestre R.Br. Asclepiadaceae Gynostemma pentaphyllum (Thunb.) Makino 32 Cucurbitaceae Hibiscus sabdariffa L. Malvaceae 32 Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated References Hoodia gordonii (Masson) Sweet ex. Decne. Hoodia Stems Increase hypothalamic ATP Yes Yes Yes Not significant P57 MacLean and Luo, 2004: 111; Holt and Taylor 2006a:104-113; Van Heerden et al., 2007: 25452553; Glasl, 2009:300-305; Mohlapo et al., 2009:15911595 Yerba mate Leaves Thermogenic, decreases food intake, reduces size of adipocytes, reduces plasma leptin, up-regulate PPARγ2, SREBP1, FAS and HMR, increases UCP2 and UCP3 expression, increases AMPK phosporylation, inhibits pancreatic lipase Yes Yes Yes Genotoxic, mutagenic Caffeine Andersen and Fogh, 2001:243-249; Ruxton, 2004:25-28; Van Wyk and Wink, 2004:179; Heck and De Meija, 2007:138-139; Pang, Choi and Park, 2008:178-184; Martins et al., 2010:42-47 Manchurian walnut Fruits Pancreatic lipase inhibitor Yes Yes No Unknown 1,4,8trihydroxynaphthalene1-O-β-d-[6ʹ-O-(3ʹ,4ʹ,5ʹtrihydroxybenzoyl)]gluco pyranoside, αhydrojuglone-4glucoside Han et al., 2007:184-186; Li et al., 2007b:846-851 Chinese juniper Heartwood, berries Enhanced gene transcription, elevated AMPK protein expression, phosphorylation in visecral adipose tissue Yes Yes No Unknown (-)-cedrol Sánchez de Medina et al., 1994:197-200; Van Wyk and Wink 2004:184; Ju et al., 2008:110-114; Kim et al., 2008c:1415-1420; Sciencelab 2008:S.a.; Park, Jung and Kim, 2009:513 Banaba Leaves Inhibits differentiation in 3T3L1 adipocytes, inhibits Yes Yes No Kidney and liver effects, Tannic acid Glick and Joslyn, 1970:509514; Suzuki et al., 1999:791- Apocynaceae Ilex paraguariensis A. St.-Hil. Aquifoliaceae 33 Juglans mandshurica Maxim. Juglandaceae Juniperus chinensis L. Cuppressaceae Lagerstroemia 33 Species name Common name Plant parts used speciosa (L.) Pers. Proposed mechanism of action In vitro In vivo anim* In vivo hum# expression of PPAR-γ2 and PPAR- γ 795; Liu et al., 2001:22422246; Liu et al., 2005:165171; Sigma-Aldrich, 2006:S.a.; Klein et al., 2007:401-406 Bark, leaves Inhibits pancreatic and lipoprotein lipase, inhibits lipolysis of 3T3-L1 preadipocytes, increases feacal fat excretion, downregulates obesity-related genes in the liver and epididymal fat, inhibits αglucosidase, inhibits αamylase Yes Yes No Not significant Mangiferin Prasanth, Padmaja and Samiulla, 2001:179-181; Prasanth et al., 2001:686688; González et al., 2007:2526-2532; Yoshikawa, 2002:18191824; Moreno et al., 2006b:21-26; Birari and Bhutani, 2007:885; Rodeiro et al., 2007:2506-2511; Wauthoz et al., 2007:112117 Bitter melon Fruits (juice) Increases sympathomimetic activity (increase plasma epinephrine and norepinephrine), increases lipolysis, decreases blood and white adipose tissue leptin, decreases resistin mRNA expression, activates AMPK pathway Yes Yes Yes Abortefacient, reduced fertility, high dose toxicity Momordicoside S, karaviloside XI Chen, Chan and Li, 2003:1088-1093; Virdi et al., 2003:107-111; Grover and Yadav, 2004:123-129; Chen and Li, 2005:747-753; Shih, Lin and Lin, 2008:134-143; Tan et al., 2008:263-273; Kasbia, Arnason and Imbeault, 2009:127-132 Chinese bayberry Bark Inhibits lipase, suppresses increase in blood triglyceride levels Yes Yes No None reported Myricitrin, myricetin, gallic acid Kobayashi et al., 2008b:289292 34 Cucurbitaceae Myrica rubra Sieb. et Zucc. References Mango Anacardiaceae Momordica charantia L. Active anti-obesity compounds isolated mutagenic, carcinogenic (tannic acid) Lythraceae Mangifera indica L. Toxicity Myricaceae 34 Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated References Nelumbo nucifera Gaertn. Lotus Leaves Inhibit α-amylase and pancreatic lipase, increase lipolytic activity 3T3-L1 adipocytes, increase UCP3 (thermogenesis) Yes Yes Yes Yes (brine shrimp) Flavanoids (quercetin 3O-α-arabinopyranosyl(1α2)-βgalactopyranoside, (+)catechin, hyperoside, isoquercitrin, astragalin) Van Wyk and Wink, 2004:213; Huralikuppi, Christopher and Stephen, 2006:217-223; Ono et al., 2006:238-243; Rai et al., 2006; Ohkoshi et al., 2007:1255-1258 Black cumin Seeds Unknown No No Yes High dose thymoquinone No Zaoui et al., 2002:69-73; Gilani, Jabeen and Khan, 2004:441-447; Le et al., 2004:251-258; Van Wyk and Wink, 2004:216; Najmi et al., 2008:11-13; Khader, Bresgen and Eckl, 2009:129-132; Qidwai et al., 2009:639-644; Sultan, Butt and Anjum, 2009:2768-2775 Asian ginseng Roots, berries Inhibit pancreatic lipase, decreased hypothalamic expression of orexigenic peptide Y, anorexigenic cholecystokinin increased, reduced leptin levels, inhibit lipid accumulation in 3T3-L1 adipocytes, increased PPARγ, and lipoprotein lipase, activates AMPK Yes Yes No Safe Ginsenosides (Rg3, Rh2), protopanaxadiol saponins, protopanaxatriol saponins Carabin, Burdock and Chatzidakis, 2000:293-301; Attele et al., 2002:18511857; Coon and Ernst, 2002:323-344; Xie et al., 2002:254-258; Dey et al., 600-605; Van Wyk and Wink, 2004:224; Hwang et al., 2007:1002-1007; Karu, Reifen and Kerem, 2007:2824-2827; Hwang et al., 2008:262-266; Kim et al., 2008a:78-85; Mollah et al., 2008:220-225; Kim et al., 2009c:596-601 Nelumbonaceae Nigella sativa L. Ranunculaceae 35 Panax ginseng C.A. Mey. Araliaceae 35 Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated References Panax japonicus C.A. Mey. Japanese ginseng Rhizomes Pancreatic lipase inhibitor, decrease plasma triacylglycerol Yes Yes No Unknown Chikusetsusaponins II, IV, and V and 28deglucosylchikuketsusa ponins IV and V Han et al., 2005b:9-17; Gibbon, 2008:69 American ginseng Stems, leaves Pancreatic lipase inhibitor, decrease plasma triacylglycerol Yes Yes No Unknown Ginsenosides R[b.sub.1], R[b.sub.2], Rc, Rd Van Wyk and Wink, 2004:224; Liu et al., 2008:S.a. Guarana Seeds Unknown (thermogenesis) No Yes No Considered safe, mild adverse effects Methylxanthines Espinola et al., 1997:223229; Mattei et al., 1998:111116; Van Wyk and Wink, 2004:228; Lima et al., 2005:1019-1027; Ruxton and Gardner, 2005:111-125 Yohimbe Bark α2-adrenergic antagonist (enhance lipid mobilisation) No No Yes CNS overstimulation, irritability and anxiety attacks Yohimbine Tam, Worcel and Wyllie, 2001:215-239; Van Wyk and Wink, 2004:229; Ostojic, 2006:289-299 White kidney bean Beans Inhibits α-amylase, phytohemoagglutinin-mediated (lectin) release of cholecystokinin and glucagonlike peptides Yes Yes Yes Toxic in high doses Phytohemagglutinin Pusztai et al., 1998:213-220; Pittler and Ernst, 2004:529536; Van Wyk and Wink, 2004:237; Carai et al., 2009:145-153 Indian psyllium Seeds, husks Lower food intake, increased fullness, reduces ghrelin gene expression, lowers plasma leptin levels Yes Yes Yes Minor adverse effects No Turnbull and Thomas, 1995:338-342; Van Wyk and Wink, 2004:245; Galisteo et al., 2005:2399-2403; Singh, 2007:1-10; Wang et al., Araliaceae Panax quinquefolium L. Araliaceae Paullinia cupana Mart. var. sorbilis Mart. 36 Sapindaceae Pausinystalia yohimbe (K. Schum.) Beille Rubiaceae Phaseolus vulgaris L. Fabaceae Plantago ovata Forssk. Plantaginaceae 36 Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated References 2007:1635-1641; Gibbon, 2008:56; Rodríguez-Morán, Guerrero-Romero and Lazcano-Burciaga, 1998:273-277 Platycodon grandiflorum (Jacq.) A. DC Platycodon root Roots Inhibits pancreatic lipase, reduce plasma and liver triacylglycerol, inhibits 3T3-L1 differentiation, suppressed FABP expression, inhibit hepatic FAS, reduced food intake (saponin fractions) Yes Yes No Unknown Platycodin A, platycodin C, deapioplatycodin D, platycodin D (34.8%), deapioplatycodin D3, platycodin V, platycodin D2, SK1 Han et al., 2000:2760-2764; Han et al., 2002:2241-2244; Van Wyk and Wink, 2004:247; Xu et al., 2005:180-184; Zhao et al., 2005:983-990; Park, Yoon and Ahn, 2007:3493-3498; Jeong et al., 2008:245-251; Kim et al., 2009a:629-635; Shin et al., 2009:47-53 Japanese bitter orange Fruits Increases intestinal transit time Yes Yes No Not significant No Shim et al., 2009:294-299 Roseroot, Golden root Rhizomes Inhibits lipase, decreases lipid absorption Yes Yes No None reported Rhodionin, rhodiosin Kobayashi et al., 2008a:1716-1719; Blomkvist, Taube and Larhammar, 2009:11871189 Rosemary Leaves Inhibit adipocyte differentiation, increase intracellular GSH, induce phase2 enzymes (involved in GSH metabolism) Yes Yes No None known Carnosic acid, carnosol Van Wyk and Wink, 2004:276; Takahashi et al., 2009:549-554 Campanulaceae 37 Poncirus trifoliata (L.) Raf. Rutaceae Rhodiola rosea L. Crassulaceae Rosmarinus officinalis L. Lamiaceae 37 Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated References Salacia reticulata Wight Kotahla himbutu Roots, stems Inhibit pancreatic lipase, lipoprotein lipase, lipolytic effects, suppresses white adipose tissue accumulation, reduce size of the adipocytes, decrease triacylglycerol levels, decrease plasma adiponectin, leptin and free fatty acids Yes Yes No Low birth weight, greater postimplantation losses, low survival ratio and viability index (-)-Epigallocatechin, (-)epicatechin-(4βà8)-(-)4ʹ-Omethylepigallocatechin, (-)-epicatechin, 3β,22βdihydroxyolean-12-en29-oic acid, mangiferin, (-)-4ʹ-Omethylepigallocatechin, maytenfolic acid Shimoda et al., 2001:527; Yoshikawa et al., 2002:1819-1824; Ratnasooriya, Jayakody and Premakumara, 2003:931935; Jayawardena et al., 2005:215-217; Kishino et al., 2006:433-438; Im et al., 2008:645-651 Chinese willow Leaves Inhibit α-amylase, enhances norepinephrine-induced lipolysis in fat cells, inhibits palmitic acid uptake into small intestinal brush border membranes Yes Yes No Unknown Apigenin-7-O-dglucoside, luteolin-7-Od-glucoside, chrysoeriol7-O-d-glucoside Han et al., 2003a:11881194; Han et al., 2003b:1195-1198; Li et al., 2008:1530-1536 Sage Leaves Inhibit pancreatic lipase, inhibit triglyceride increase Yes Yes No Neurotoxicity of essential oil due to thujone Carnosic acid, carnosol, royleanonic acid, 7methoxyrosmanol, oleanolic acid Ninomiya et al., 2004:19431946; Van Wyk and Wink, 2004:283 Scabiosa Whole plant Pancreatic lipase inhibitor Yes No No Unknown Scabiosaponins E, F, G, I, hookerosides A, B, and prosapogenin 1b Zheng et al., 2004:604-613; Birari and Bhutani, 2007:881; Guo et al., 2009:1167-1174; Mahabala, country mallow Leaves Unknown No Yes Yes Unknown No Patel et al., 2009:233237;Thounaojam et al., 2009:413-429 Clove Dried unopened flower buds Pancreatic lipase inhibitor, decrease plasma triacylglycerol Yes Yes No Organ damage in rats Not characterized Van Wyk and Wink, 2004:315; Kim et al., 2005:84-87; Yu, Shirai and Celestraceae Salix matsudana Koidzumi Salicaceae 38 Salvia officinalis L. Lamiaceae Scabiosa tschiliensis Grun. Dipsacaceae Sida rhomboidea. Roxb Malvaceae Syzygium aromaticum (L.) Merr. et Perry 38 Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated Myrtaceae Taraxacum officinale Weber ex Wigg. Suzuki, 2007:155-161; Agbaje, Adeneye and Daramola, 2009:241-254 Dandelion Roots, leaves Pancreatic lipase inhibitor Yes Yes No Unknown No Rácz-kotilla, Rácz and Solomon, 1974:212-217; Van Wyk and Wink, 2004:318; Zhang et al., 2008:200-203; RodriguezFragoso et al., 2008:130 Triumphetta Leaves Unknown No Yes No Unknown No Ngondi et al., 2006:200 Fenugreek Seeds, fiber Inhibit lipid absorption, decreased expression of adipogenic factors, increased satiety (reduced energy intake) Yes Yes Yes CNS effects at high doses in mice, none observed during human studies 4-hydroxyisoleucine, galactomannan Petit et al., 1995:674-680; Muralidhara et al., 1999; 831-837; Abdel-Barry and Al-Hakiem, 2000:65-68; Van Wyk and Wink, 2004:325; Handa et al., 2005:11861187; Mathern et al., 2009:1543-1548; Vijayakumar et al., 2010:667-674 Grape vine Seeds Inhibits pancreatic lipase, inhibits lipoprotein lipase, inhibits lipolysis, increase adiponectin, reduce leptinaemia, reduce energy intake, reduce PPARγ and C/EBPα, inhibits adipocyte differentiation through cell Yes Yes Yes None observed Resveratrol, stilbesterol, ampelopsin A, vitisin B, 3,4ʹ,5-trimethoxy stilbene, vitisin A Bentivegna and Whitney, 2002:1731-1743; Yamakoshi et al., 2002:599-607; Moreno et al., 2003: 876879; Van Wyk and Wink, 2004:344;Vogels, Nijs and Westerterp-Plantenga, 2004:667-673; Décordé et Asteraceae Triumphetta cordifolia A. Rich References Tiliaceae 39 Trigonella foenumgraecum L. Fabaceae Vitis vinifera L. Vitaceae 39 Species name Common name Plant parts used Proposed mechanism of action In vitro In vivo anim* In vivo hum# Toxicity Active anti-obesity compounds isolated cycle arrest Wasabia japonica (Miq.) Matsum. Leaves, rhizomes and stalks Suppress differentiation of preadipocytes, suppress increase in glycerol-3phosphate dehydrogenase (GDPH) activity and triglyceride (TG) accumulation Yes No No Unknown No Hou et al., 2000:195-199; Morimitsu et al., 2000:125133; Ogawa et al., 2009:239-244 Japanese ginger Rhizomes Suppress differentiation of preadipocytes, suppress increase in glycerol-3phosphate dehydrogenase (GDPH) activity and triglyceride (TG) accumulation Yes Yes No Unknown No Iwashita, Yamaki and Tshushida, 2001:164-169; Miyoshi et al., 2003:113-118 Ginger Rhizomes Reduced expression of retinoid binding protein (RBP) and mRNA in the liver and visceral fat Yes Yes No None known No Weidner and Sigwart, 2000:513-519; Iwashita, Yamaki and Tshushida, 2001:164-169; Van Wyk and Wink, 2004:349; WesterterpPlantenga et al., 2006:8591; Matsuda et al., 2009:903-907 Zingiberaceae 40 Zingiber officinale Roscoe al., 2008:659-666; Kim et al., 2008d:108-113 Wasabi Brassicaceae Zingiber mioga Roscoe Zingeberaceae * in vivo animal studies # References in vivo human studies 40 This investigation showed that several plants including Cassia mimosoides, Dioscorea nipponica, Glycyrrhiza glabra and Panax ginseng possess promising anti-obesity activity with the potential to be developed into anti-obesity agents or blueprints for anti-obesity compounds, but only after additional research and clinical trials have been conducted. It was also revealed that the toxicity of an astonishing 35% of the plants included in the table has not been investigated while the toxicity of most of the other plants has not been investigated comprehensively. Consider the following example: Tea is the most consumed beverage worldwide and hence it is not surprising that tea products would automatically be considered safe. However, instances of hepatotoxicity have been linked to consumption of green tea extracts (Sarma et al., 2008:477). These extracts are prepared differently from ordinary tea and are therefore higher in tea catechins which may cause adverse effects. Marked mortality was observed when purified green tea extract was administered to Beagle dogs. The highest dose administered had to be decreased from 1000 mg/kg/day to 800 mg/kg/day whereafter unscheduled mortality occurred in 16 out of 24 dogs. The cause could not be determined due to extensive organ and system involvement and the severity of the toxicity. In a follow-up study at a lower dosage (200 mg/kg) the dogs presented with digestive disturbances, mild liver damage, moderate to progressive anaemia, mild disturbances in haematopoiesis and transient ocular symptoms (Kapetanovic et al., 2009:30-35). After a safety review by the United States Pharmacopoeia Dietary Supplement Information Expert Committee it was concluded that 27 out of 34 reports pertaining to liver damage were probably caused by the consumption of green tea products (Sarma et al., 2008:469-470). In addition to safety, the quality and efficacy of these plants were also neglected. Another concern is that much of the research conducted has been of an in vitro nature. Clinical trials in humans have not been done for an alarming 62% of these plants, yet they are freely sold and consumed daily. For most of the plants the molecule(s) responsible for the anti-obesity activity has not been isolated. Isolation of single compounds is important in basic science to determine the mechanism by which plants exert their effect. Effects may then be attributed to a single compound and efforts can then be made to substitute certain functional groups on a molecule either to increase its activity or to ameliorate side 41 effects. On the other hand, ease of interpretation should not detract from the way in which these plants or plant combinations were used traditionally as it is wellknown that different compounds in one plant extract act on different target sites resulting in synergistic interactions. One such example is a study on the antimicrobial activity of T. camphoratus, where the combination of the volatile and non-volatile fractions exhibited increased efficacy compared to the individual fractions (Van Vuuren and Viljoen, 2009:505). An example of a synergistic plant combination is the increased antimicrobial activity of extracts of Salvia chameleaegnea and Leonotis leonurus in combination against Gram-positive organisms compared to the extracts in isolation (Kamatou et al., 2006:634). A combination strategy seems particularly suited to obesity as body weight is controlled by many factors. However, this does not imply that many plants should be randomly combined into one anti-obesity formulation with the hope that this will have a synergistic anti-obesity effect as the combination may exhibited more side effects and interactions. 2.4 CONCLUSIONS The statistics on obesity prove that it should be considered a threat to the general health of the world population. This threat leads to exploitation of the public by manufacturing companies focused on profit with little consideration towards the overall effect that these products may have on the consumer. This is further compounded by the fact that there is a general misconception by the public that ‘green is good’. These products are deemed natural or derived from nature and are therefore considered harmless despite the fact that more and more evidence to the contrary is emerging. Manufacturers of herbal supplements often claim that their products are potent and effective with no or few adverse effects. If these supplements possess such potent biological activity, they should be considered to be conventional drugs and treated as such in terms of quality control and proof of effectivity (Fontanarosa, Rennie and DeAngelis, 2003:1569). At this stage, research into side effects and toxicity seems to take place only once serious harm or even death of patients has occurred. For the treatment of obesity and the treatment of any disease the risk-benefit ratio must be considered. Future studies should therefore focus on pursuing promising anti-obesity activity displayed by 42 many plants on in vivo models other than rats or mice to eventually progress to in vivo human studies. It should also be kept in mind that caution should be exercised when clinical results are interpreted. In some studies, dietary modification and exercise regimens were introduced at the same time as the administration of the plant or product to be tested. This means that weight loss cannot be ascribed to the action of the plant or product only as dietary modification and exercise will have a beneficial effect on weight loss. It is evident that the investigation into plants used for their anti-obesity activity raised numerous concerns and questions which by nature can only be answered through more research. H. gordonii is a prime example of most of the abovementioned statements. It is unscrupulously marketed as a miracle drug to cure obesity and is one of the bestsellers, but it is also frequently adulterated with other plant material and substances which may have deleterious effects on the consumer. In contrast to other plants, the correct route of research and development of a drug was followed to a large extent for H. gordonii. However, the aspiration of developing P57 into a pharmaceutical anti-obesity preparation has not been realised due to numerous factors and H. gordonii products are still freely sold. The in vivo studies that have been performed to prove anti-obesity activity and numerous other aspects that should be investigated prior to consumption are insufficient. Ideally, all plant products should be regulated like conventional or allopathic medicines as they should conform to the cardinal rule of ‘safety, efficacy and quality’. In addition, basic research to substantiate the worldwide use of medicinal plants must underpin and precede commercial development. 43 2.5 REFERENCES ABDEL-BARRY, J.A., AL-HAKIEM, M.H.H. 2000. Acute intraperitoneal and oral toxicity of the leaf glycosidic extract of Trigonella foenum-gracum in mice. Journal of Ethnopharmacology, 70:65-68. ABRAHAMSE, S.L., POVEY, K.J., REES, D.D. 2007. Appetite suppressant compositions. US Patent 2007/0207227. AGBAJE, E.O., ADENEYE, A.A., DARAMOLA, A.O. 2009. Biochemical and toxicological studies of aqueous extract of Syzigium aromaticum (L.) MERR. & PERRY (Myrtaceae) in rodents. The African Journal of Traditional Complementary and Alternative Medicines, 6(3):241-254. AKINDAHUNSI, A.A., OLALEYE, M.T. 2003. Toxicological investigation of aqueous-methanolic extract of the calyces of Hibiscus sabdariffa L. Journal of Ethnopharmacology, 89:161-164. ALARCON-AGUILAR, F.J., ZAMILPA, A., PEREZ-GARCIA, M.D., ALMANZAPEREZ, J.C., ROMERO-NUÑEZ, CAMPOS-SEPULVEDA, E.A., VAZQUEZCARRILLO, L.I., ROMAN-RAMOS, R. 2007. Effect of Hibiscus sabdariffa on obesity in MSG mice. Journal of Ethnopharmacology, 114:66-71. AL-DUBAI, W., AL-HABORI, M., AL-GEIRY, A. 2006. Human khat (Catha edulis) chewers have elevated plasma leptin and nonesterified fatty acids. Nutrition Research, 26:632-636. AL-HABORI, M., AL-AGHBARI, A., AL-MAMARY, M., BAKER, M. 2002. Toxicological evaluation of Catha edulis leaves: a long term feeding experiment in animals. Journal of Ethnopharmacology, 83:209-217. AL-HEBSHI, N.N., SKAUG, N. 2005. Khat (Catha edulis) – an updated review. Addiction Biology, 10:299-307. AL-ZUBAIRI, A., ISMAIL, P., PEI PEI, C., RAHMAT, A. 2008. Genotoxic effect of Catha edulis (khat) crude extract after sub-chronic administration in rats. Environmental Toxicology and Pharmacology, 25:298-303. AMER, A., BREU, J., MCDERMOTT, J., WURTMAN, R.J., MAHER, T.J. 2004. 5Hydroxy-L-tryptophan suppresses food intake in food-deprived and stressed rats. Pharmacology Biochemistry and Behaviour, 77:137-143. ANDERSEN, T., FOGH, J. 2001. Weight loss and delayed gastric emptying following a South American herbal preparation in overweight patients. Journal of Human Nutrition and Dietetics, 14(3):243-250. ARBO, M.D., LARTENTIS, E.R., LINCK, V.M., ABOY, A.L., PIMENTEL, A.L., HENRIQUES, A.T., DALLEGRAVE, E., GARCIA, S.C., LEAL, M.B., LIMBERGER, R.P. 2008. Concentrations of p-synephrine in fruits and leaves of Citrus species (Rutaceae) and the acute toxicity testing of Citrus aurantium extract and psynephrine. Food and Chemical Toxicology, 46:2770-2775. 44 ARBO, M.D., SCHMITT, G.C., LIMBERGER, M.F., CHARAO M.F., MORO, A.M., RIBEIRO, G.L., DALLEGRAVE, E., GARCIA, S.C., LEAL, M.B., LIMBERGER, R.P. 2009. Subchronic toxicity of Citrus aurantium L. (Rutaceae) extract and psynephrine in mice. Regulatory Toxicology and Pharmacology, 54:114-117. ARMANINI, D., NACAMULLI, D., FRANCINI-PESENTI, F., BATTAGIN, G., RAGAZZI, E., FIORE, C. 2005. Glycyrrhetinic acid, the active principle of licorice, can reduce the thickness of subcutaneous fat through topical application. Steroids, 70:538-542. ASAI, A., MIYAZAWA, T. 2001. Dietary curcuminoids prevent high-fat diet-induced lipid accumulation in rat liver and epididymal adipose tissue. The Journal of Nutrition, 131:2932-2935. ATTAWISH, A., CHIVAPAT, S., PHADUNGPAT, S., BANSIDDHI, J., TECHADAMRONGSIN, Y., MITRIJIT, O., CHAORAI, B., CHAVALITTUMRONG, P. 2004. Chronic toxicity of Gynostemma pentaphyllum. Fitoterapia, 75:539-551. ATTELE, A.S., ZHOU, Y.-P., XIE, J.-T., WU, J.A., ZHANG, L., DEY, L. PUGH, W., RUE,, P.A., POLONSKY, K.S., YUAN, C.-S. 2002. Antidiabetic effects of Panax ginseng berry extract and the identification of an effective component. Diabetes, 51:1851-1858. AUVICHAYAPAT, P., PRAPOCHANUNG, M., TUNKAMNERDTHAI, O., SRIPANIDKULCHAI, B.-O., AUVICHAYAPAT, N., THINKHAMROP, B., KUNHASARA, S., WONGPRATOOM, S., SINAWAT, S., HOGPRAPAS, P. 2008. Effectiveness of green tea on weight reduction in obese Thais: A randomized, controlled trial. Physiology and Behaviour, 93:486-491. AVULA, B., WANG, Y.-H., PAWAR, R.S., SHUKLA, Y.J., SMILLIE, T.J KHAN, I.A. 2008. A rapid method for chemical fingerprint analysis of Hoodia species, related genera, and dietary supplements using UPLC-UV-MS. Journal of Pharmaceutical and Biomedical Analysis, 48:722-731. BAI, M.-S., GAO, J.-M, FAN, C., YANG, S.-X., ZHANG, G., ZHENG, C.-D. 2010. Bioactive dammarane-type triterpenoids derived from the acid hydrosylate of Gynostemma pentaphyllum saponins. Food Chemistry, 119:306-310. BENTIVEGNA, S.S., WHITNEY, K.M. 2002. Subchronic 3-month oral toxicity study of grape seed and grape skin extracts. Food and Chemical Toxicology, 40:1731-1743. BIRARI R.B., BHUTANI, K.K. 2007. Pancreatic lipase inhibitors from natural sources: unexplored potential. Drug Discovery Today, 12(19/20):879-889. BLOMKVIST, J., TAUBE, A., LARHAMMAR, D. 2009. Perspective on roseroot (Rhodiola rosea) studies. Planta Medica, 75:1187-1190. BOUCHARD, N.C., HOWLAND, M.A., GRELLER, H.A., HOFFMAN, R.S., NELSON, L.S. 2005. Ischemic stroke associated with use of an ephedra-free 45 dietary supplement containing synephrine. Mayo Clinic Proceedings, 80(4):541545. BRUYNS, P.V. 2005. Stapeliads of Southern Africa and Madagascar volume 1. Pretoria: Umdaus Press. BURDOCK, G. 2005. Letter to the editor: Garcinia cambogia toxicity is misleading. Food and Chemical Toxicology, 43:1683-1684. CALAPAI, G., FIRENZUOLI, F., SAITTA, A., SQUADRITO F., ARLOTTA, M.R., COSTANTINO, G., INFERRERA, G. 1999. Antiobesity and cardiovascular toxic effects of Citrus aurantium extracts in the rat: a preliminary report. Fitoterapia, 70:586-592. CANOVÁ, N.K., LINCOVÁ, D., KMONÍČKOVÁ, E., KAMENÍKOVÁ, L, FARGHALI, H. 2006. Nitric oxide production from rat adipocytes is modulated by β 3-adrenergic receptor agonists and is involved in a cyclic AMP-dependent lipolysis in adipocytes. Nitric Oxide, 14:200-211. CARABIN, I.G., BURDOCK, G.A., CHATZIDAKIS, C. 2000. Safety assessment of Panax ginseng. International Journal of Toxicology, 19(4):293-301. CARAI, M.A.M., FANTINI, N., LOI, B., COLOMBO, G., RIVA, A., MORAZZONI, P. 2009. Potential efficacy of preparations derived from Phaseolus vulgaris in the control of appetite, energy intake, and carbohydrate metabolism. Diabetes Metabolic Syndrome and Obesity: Targets and Therapy, 2:145-153. CARVAJAL-ZARRABAL, O., HAYWARD-JONES, P.M., ORTA-FLORES, Z., NOLASCO-HIPÓLITO, C., BARRADAS-DERMITZ, D.M., AGUILAR-USCANGA, M.G., PEDROZA-HERNÁNDEZ, M.F. 2009. Effect of Hibiscus sabdariffa L., dried calyx ethanol extract on fat absorption-excretion, and body weight implication in rats. Journal of Biomedicine and Biotechnology, in press:doi: 10.1155/2009/394592. CHA, Y.-S., RHEE, S.-J., HEO, Y.-R. 2004. Acanthopanax senticosus extract prepared from cultured cells decreases adiposity and obesity indices in C57BL/6J mice fed a high fat diet. Journal of Medicinal Food, 7(4):422-429. CHANDA, S., MOULD, A., ESMAIL, A., BLEY, K. 2005. Toxicity studies with pure trans-capsaicin delivered to dogs via intravenous administration. Regulatory Toxicology and Pharmacology, 43:66-75. CHANTRE, P., LAIRON, D. 2002. Recent findings of green tea extract AR25 (Exolise) and its activity for the treatment of obesity. Phytomedicine, 9:3-8. CHEN, N., BEZZINA, R., HINCH, E., LEWANDOWSKI, P.A., CAMERON-SMITH, D., MATHAI, M.L., JOIS, M., SINCLAIR, A.J., BEGG, D.P., WARK, J.D., WEISINGER, H.S., WEISINGER, R.S. 2009. Green tea, black tea, and epigallocatechin modify body composition, improve glucose tolerance, and 46 differentially alter metabolic gene expression in rats fed a high-fat diet. Nutrition Research, 29:784-793. CHEN, Q., CHAN, L.L.Y., LI, E.T.S. 2003. Bitter melon (Momordica charantia) reduces adiposity, lowers serum insulin and normalizes glucose tolerance in rats fed a high fat diet. The Journal of Nutrition, 133:1088-1093. CHEN, Q., LI, E.T.S. 2005. Reduced adiposity in bitter melon (Momordica charantia) fed rats is associated with lower tissue triglyceride and higher plasma catecholamines. British Journal of Nutrition, 93:747-754. COELHO, S.B., LOPES DE SALES, R., IYER, S.S., BRESSAN, J., COSTA, N.M.B, LOKKO, P., MATTES, R. 2006. Effects of peanut oil load on energy expenditure, body composition, lipid profile, and appetite in lean and overweight adults. Nutrition, 22:585-592. COLKER, C.M., KALMAN, D.S., TORINA G.C., PERLIS, T., STREET, C. 1999. Effects of Citrus aurantium extract, caffeine, and St. John’s Wort on body fat loss, lipid levels, and mood states in overweight healthy adults. Current Therapeutic Research, 60(3):145-153. COON, J.T., ERNST, E. 2002. Panax ginseng: a systematic review of adverse effects and drug interactions. Drug Safety, 25(5):323-344. COXON, G.D., FURMAN, B.L., HARVEY, A.L., MCTAVISH, J., MOONEY, M.H., ARASTOO, M., KENNEDY, A.R., TETTEY, J.M., WAIGH, R.D. 2009. Benzylguanidines and other galegine analogues inducing weight loss in mice. Journal of Medicinal Chemistry, 52(11):3457-3463. DÉCORDÉ, K., TEISSÉDRE, P.-L., SUTRA, T., VENTURA, E., CRISTOL, J.-P., ROUANET, J.-M. 2008. Chardonnay grape seed procyanidin extract supplementation prevents high-fat diet-induced obesity in hamsters by improving adipokine imbalance and oxidative stress markers. Molecular Nutrition and Food Research, 53(5):659-666. DALL’ACQUA, S., INNOCENTI, G. 2007. Steroidal glycosides from Hoodia gordonii. Steroids, 72:559-568. DEY, L., XIE, J.T., WANG, A., WU, J., MALECKAR, S.A., YAN, C.-S. 2003. Antihyperglycemic effects of ginseng: comparison between root and berry. Phytomedicine, 10:600-605. EDASHIGE, Y., MURAKAMI, N., TSUJITA, T. 2008. Inhibitory effects of pectin from the segment membrane of citrus fruits on lipase activity. Journal of Nutritional Science and Vitaminology, 54:409-415. EDDOUKS, M., LEMHADRI, A., MICHEL, J.-B. Hypolipidemic activity of aqueous extract of Capparis spinosa L. in normal and diabetic rats. Journal of Ethnopharmacology, 98:345-350. 47 EJAZ, A., WU, D., KWAN, P., MEYDANI, M. 2009. Curcumin inhibits adipogenesis in 3T3-L1 adipocytes and angiogenesis and obesity in C57/BL mice. The Journal of Nutrition, 139(5):919-925. ESPINOLA, E.B., DIAS, R.F., MATTEI, R., CARLINI, E.A. 1997. Pharmacological activity of guarana (Paullinia cupana Mart.) in laboratory animals. Journal of Ethnopharmacology, 55:223-229. FINER, N. 2002. Pharmacotherapy of obesity. Best Practice and Research Clinical Endocrinology and Metabolism, 16(4):717-742. FIRENZUOLI, F., GORI, L., GALAPAI, C. 2005. Adverse reaction to an adrenergic herbal extract (Citrus aurantium). Phytomedicine, 12:247-248. FODEN, W. 2005. Information document on trade in Hoodia gordonii and other Hoodia species. Available from: http://www.plantzafrica.com/planthij/hoodia.htm [Accessed: 29/26/2010]. FONTANAROSA, P.B., RENNIE, D., DEANGELIS, C.D. 2003. The need for regulation of dietary supplements – lessons from ephedra. Journal of the American Medical Association, 289(12):1568-1570. FUGH-BERMAN, A., MYERS, A. 2004. Citrus aurantium, an ingredient of dietary supplements marketed for weight loss: current status of clinical and basic research. Experimental Biology and Medicine, 299:698-704. GALISTEO, M., SÁNCHEZ, M., VERA, R., GONZÁLEZ, M., ANGUERA, A., DUARTE, J., ZARZUELO, A. 2005. A diet supplemented with husks of Plantago ovata reduces the development of endothelial dysfunction, hypertension and obesity by affecting adiponectin and TNF-α in obese Zucker rats. The Journal of Nutrition, 135:2399-2404. GIBBON, C.J. (ed.). 2008. South African Medicines Formulary 8th ed. Cape Town: Health and Medical Publishing group. GILANI, A.H., JABEEN, Q. KHAN, M.A.U. 2004. A review of medicinal uses and pharmacological activities of Nigella sativa. Pakistan Journal of Biological Sciences, 7(4):441-451. GLASL, S. 2009. Hoodia: A herb used in South African traditional medicine – A potential cure for overweight? Pharmacognostic review of history, composition, health-related claims, scientific evidence and intellectual property rights. Schweizerische Zeitschrift für GanzheitsMedizin, 21(6):300-306. GLICK, Z., JOSLYN, M.A. 1970. Food intake depression and other metabolic effects of tannic acid in the rat. The Journal of Nutrition, 100:509-515. GODARD, M.P., JOHNSON, B.A., RICHMOND, S.R. 2005. Body composition and hormonal adaptations associated with forskolin consumption in overweight and obese men. Obesity Research, 13(8):1335-1343. 48 GOODMAN, A.G., RALL, T.W., NIES, A.S., TAYLOR, P. (eds). 1992. Goodman and Gilman’s: The pharmacological basis of therapeutics 8th ed vol 1. New York: McGraw-Hill. GONZÁLEZ, J.E., RODRÍGUEZ, M.D., RODEIRO, I., MORFFI, J., GUERRA, E., LEAL, F., GARCÍA, H., GOICOCHEA, E., GUERRERO, S., GARRIDO, G., DELGADO, R., NUÑEZ-SELLES, A.J. 2007. Lack of in vivo embryotoxic and genotoxic activities of orally administered stem bark aqueous extract of Mangifera indica L. (Vimang®). Food and Chemical Toxicology, 45:2526-2532. GROVER, J.K., YADAV, S.P. 2004. Pharmacological actions and potential uses of Momordica charantia: a review. Journal of Ethnopharmacology, 93:123-132. GUO, T., LIU, Q., WANG, P., ZHANG, L., ZHANG, W., LI, Y. 2009. Facile synthesis of three bidesmosidic oleanolic acid saponins with strong inhibitory activity on pancreatic lipase. Carbohydrate Research, 344:1167-1174. HAKKINEN, J., HORAK, R.M., MAHARAJ, V. 2005. Gastric acid secretion. US Patent 2005/0079233 A1. HAN, L.-K., TAKAKU, T., LI, J., KIMURA, Y., OKUDA, H. 1999. Anti-obesity action of oolong tea. International Journal of Obesity, 23:98-105. HAN, L-K., XU, B.-J., KIMURA, Y., ZHENG, Y.-N., OKUDA, H. 2000. Platycodi radix affects lipid metabolism in mice with high fat diet-induced obesity. The Journal of Nutrition, 130:2760-2764. HAN, L.-K., KIMURA, Y., KAWASHIMA, M., TAKAKU, T., TANIYAMA, T., HAYASHI, T., ZHENG, Y.N., OKUDA, H. 2001. Anti-obesity effects in rodents of dietary teasaponin, a lipase inhibitor. International Journal of Obesity, 25(10):1459-1464. HAN, L.-K., ZHENG, Y.-N., XU, B.-J., OKUDA, H., KIMURA, Y. 2002. Saponins from Platycodi radix ameliorate high fat diet-induced obesity in mice. The Journal of Nutrition, 132:2241-2245. HAN, L.-K., SUMIYOSHI, M., ZHANG, J., LIU, M.-X., ZHANG, X.-F., ZHENG, Y.N., OKUDA, H., KIMURA, Y. 2003a. Anti-obesity action of Salix matsudana leaves (Part 1). Anti-obesity action by polyphenols of Salix matsudana in high fat-diet treated rodent animals. Phytotherapy Research, 17(10):1188-1194. HAN, L.-K., SUMIYOSHI, M., ZHENG, Y.-N., OKUDA, H., KIMURA, Y. 2003b. Anti-obesity action of Salix matsudana leaves (Part 2). Isolation of anti-obesity effectors from polyphenol fractions of Salix matsudana. Phytotherapy Research, 17(10):1195-1198. HAN, L.-K., MORIMOTO, C., YU, R.H., OKUDA, H. 2005a. Effects of Coleus forskohlii on fat storage in ovariectomized rats. Yakugaku-Zasshi, 125(5):449-453. 49 HAN, L.-K., ZHENG, Y.-N, YOSHIKAWA, M., OKUDA, H., KIMURA, Y. 2005b. Anti-obesity effects of chikusetsusaponins isolated from Panax japonicus rhizomes. BMC complementary and Alternative Medicine, 5:9-18. HAN, L., LI, W., NARIMATSU, S., LIU, L., FU, H., OKUDA, H., KOIKE, K. 2007. Inhibitory effects of compounds isolated from fruit of Juglans mandshurica on pancreatic lipase. Journal of Natural Medicines, 61:184-186. HANDA, T., YAMAGUCHI, K., SONO, Y., YAZAWA, K. 2005. Effects of fenugreek seed extract in obese mice fed a high-fat diet. Bioscience Biotechnology and Biochemistry, 69(6):1186-1188. HATANO, T., YAMASHITA, A., HASHIMOTO, T., ITO, H., KUBO, N., YOSHIYAMA, M., SHIMURA, S., ITOH, Y., OKUDA, T., YOSHIDA, T. 1997. Flavan dimers with lipase inhibitory activity from Cassia nomame. Phytochemistry, 46(5):893-900. HECK, C.I., DE MEJIA, E.G. 2007. Yerba Mate Tea (Ilex paraguariensis): A comprehensive review on chemistry, health implications, and technological considerations. Journal of Food Science, 72(9):138-151. HENDERSON, S., MAGU, B., RASMUSSEN, C., LANCASTER, S., KERKSICK, C., SMITH, P., MELTON, C., COWAN, P., GREENWOOD, M., EARNEST, C., ALMADA, A., MILNOR, P., MAGRANS, T., BOWDEN, R., OUNPRASEUTH, S., THOMAS, A., KREIDER, R.B. 2005. Effects of Coleus forskohlii supplementation on body composition and hematological profiles in mildly overweight women. Journal of the International Society of Sports Nutrition, 2(2):54-62. HIGGS, J. 2007. The potential role of peanuts in the prevention of obesity. Nutrition and Food Science, 35(5):353-358. HOLT, S., TAYLOR, T.V. 2006a. Hoodia gordonii. An overview of biological and botanical characteristics: Part 1. Townsend Letter for Doctors and Patients, 280:104-113. HOLT, S., TAYLOR, T.V. 2006b. Hoodia gordonii. Separating science from speculation: Part 2. Townsend Letter for Doctors and Patients, 281:99-104. HOMOUD, M.K. 2009. The sale of weight-loss supplements on the Internet: A lurking health crisis waiting to strike. Heart Rhythm, 6(5):658-662. HOU, D., FUKUDA, M, FUJII M., FUKE, Y. 2000. Transcriptional regulation of nicotinamide adenine dinucleotide phosphate: quinine oxidoreductase in murine hepatoma cells by 6-(methylsufinyl)hexyl isothiocyanate, an active principle of wasabi (Eutrema wasabi Maxim). Cancer Letters, 161:195-200. HSU, C.-H., TSAI, T.-H., KAO, Y.-H., HWAN, K.-C., TSEN, T.-Y., CHOU, P. 2008. Effect of green tea extract on obese women: A randomized, double-blind, placebocontrolled clinical trial. Clinical Nutrition, 27:363-370. 50 HU, J.-N, ZHU, X.-M., HAN, L.-K., SAITO, M., SUN, Y.-S., YOSHIKAWA, M., KIMURA, Y., ZHENG, Y.-N. 2008. Anti-obesity effects of escins extracted from the seeds of Aesculus turbinata Blume (Hippocastanaceae). Chemical and Pharmaceutical Bulletin, 56(1):12-16. HU, Y., FAHMY, H., ZJAWIONY, J.K. DAVIES, G.E. 2010. Inhibitory effect and transcriptional impact of berberine and evodiamine on human white preadipocyte differentiation. Fitoterapia, 81(4):259-268. HURALIKUPPI, J.C., CHRISTOPHER, A.B., STEPHEN, P.M. 2006. Antidiabetic effect of Nelumbo nucifera Gaertn. extract: Part II. Phytotherapy Research, 5(5):217-223. HWANG, J.-T., PARK, I.-J., SHIN, J.-I., LEE, Y.K., LEE, S.K., BAIK, H.W., HA, J., PARK, O.J. 2005. Genistein, EGCG, and capsaicin inhibit adipocyte differentiation process via activating AMP-activated protein kinase. Biochemical and Biophysical Research Communications, 338:694-699. HWANG, J.-T., KIM, S.-H., LEE, M.-S., KIM, S.H., YANG, H.-J., KIM, M.-J., KIM, H.-S., HA, J., KIM, M.S., KWON, D.Y. 2007. Anti-obesity effects of ginsenoside Rh2 are associated with the activation of AMPK signaling pathway in 3T3-L1 adipocyte. Biochemical and Biophysical Research Communications, 364:10021008. HWANG, J.-T., LEE, M.-S., KIM, H.-J., SUNG, M.-J., KIM, H.Y., KIM, M.S., KWON, D.Y. 2008. Antiobesity effect of ginsenoside Rg3 involves the AMPK and PPAR-γ signal pathways. Phytotherapy Research, 23(2):262-266. IM, R., MANO, H., NAKATUNI, S., SHIMIZU, J., WADA, M. 2008. Aqueous extract of Kotahla himbutu (Salacia reticulata) stems promotes oxygen consumption and suppresses body fat accumulation in mice. Journal of Health Science, 54(6):645653. INOCENCIO C, RIVERA D, ALCARAZ F, TOMÁS-BARBERÁN FA. 2000. Flavonoid content of commercial capers (Capparis spinosa, C. sicula and C. orientalis) produced in Mediterranean countries. European Food Research and Technology, 212:70-74. ISBRUCKER, R.A., BURDOCK, G.A. 2006. Risk and safety assessment on the consumption of licorice root (Glycyrrhiza sp.), its extract and powder as a food ingredient, with emphasis on the pharmacology and toxicology of glycyrrhizin. Regulatory Toxicology and Pharmacology, 46:167-192. IWASHITA, K., YAMAKI, K., TSHUSHIDA, T. 2001. Mioga (Zingiber mioga Rosc.) extract prevents 3T3-L1 differentiation into adipocytes and obesity in mice. Food Science and Technology Research, 7(2):164-170. JAYAWARDENA, M.H.S., DE ALWIS, N.M.W., HETTIGODA, V., FERNANDO, D.J.S. 2005. A double-blind randomized placebo controlled cross over study of a 51 herbal preparation containing Salacia reticulata in the treatment of type 2 diabetes. Journal of Ethnopharmacology, 97:215-218. JEONG, S., CHAE, K., JUNG, Y.S., RHO, Y.H., LEE, J., HA, J., YOON, K.H., KIM, G.C., OH, K.S., SHIN, S.S., YOON, M. 2008. The Korean traditional medicine Gyeongshingangjeehwan inhibits obesity through regulation of leptin and PPARα action in OLETF rats. Journal of Ethnopharmacology, 119:245-251. JU, J.B., KIM, J.S., CHOI, C.W., LEE, H.K., OH, T.-K., KIM, S.C. 2008. Comparison between ethanolic and aqueous extracts from Chinese juniper berries for hypoglycaemic and hypolipidemic effects in alloxan-induced diabetic rats. Journal of Ethnopharmacology, 15:110-115. KAMATOU, G.P.P., VILJOEN, A.M., VAN VUUREN, S.F., VAN ZYL, R. 2006. In vitro evidence of antimicrobial synergy between Salvia chamelaeagnea and Leonotis leonurus. South African Journal of Botany, 72:634-636. KAMISOYAMA, H., HONDA, K., TOMINAGA, Y., YOKOT, S., HASEGAWA, S. 2008. Investigation of the anti-obesity action of licorice flavonoid oil in diet-induced obese rats. Bioscience Biotechnology and Biochemistry, 72(12):3225-3231. KANETKAR, P., SINGHAL, R., KAMAT, M. 2007. Gymnema sylvestre: A Memoir. Journal of Clinical Biochemistry and Nutrition, 41:77-81. KANG, J.-H., KIM, C.-S., HAN, I.-S., KAWADA, T., YU, R. 2007. Capsaicin, a spicy component of hot peppers, modulates adipokine gene expression and protein release from obese-mouse adipose tissues and isolated adipocytes, and suppresses the inflammatory responses of adipose tissue macrophages. FEBS Letters, 581:4389-4396. KANKI, K., NISHIKAWA, A., FURUKAWA, F., KITAMURA, Y., IMAZAWA, T., UMEMURA, T., HIROSE, M. 2003. A 13-week subchronic toxicity study of paprika color in F344 rats. Food and Chemical Toxicology 41:337-1343. KAPETANOVIC, I.M., CROWELL, J.A., KRISHNARAJ, R., ZAKHAROV, A., LINDEBLAD, M., LYUBIMOV, A. 2009. Exposure and toxicity of green tea polyphenols in fasted and non-fasted dogs. Toxicology, 260:28-36. KARU, N., REIFEN, R., KEREM, Z. 2007. Weight gain reduction in mice fed Panax ginseng saponin, a pancreatic lipase inhibitor. Journal of Agricultural and Food Chemistry, 55:2824-2828. KASBIA, G.S., ARNASON, J.R., IMBEAULT, P. 2009. No effect of acute, single dose oral administration of Momordica charantia L., on glycemia, energy expenditure and appetite: A pilot study in non-diabetic overweight men. Journal of Ethnopharmacology, 126:127-133. KAWAGUCHI, K., MIZUNO, T., AIDA, K., UCHINO, K. 1997. Hesperidin as an inhibitor of lipases from porcine pancreas and Pseudomonas. Bioscience Biotechnology and Biochemistry, 61(1):102-104. 52 KHADER, M., BRESGEN, N., ECKL, P.M. 2009. In vitro toxicological properties of thymoquinone. Food and Chemical Toxicology, 47:129-133. KIM, J.H., KANG, S.A., HAN, S.-M., SHIM, I. 2008a. Comparison of the antiobesity effects of protopanaxadiol-and protopanaxatriol-type saponins of red ginseng. Phytotherapy Research, 23(1):78-85. KIM, J.-Y., SO, H., YOUN, M.-J, KIM, H.-J., KIM, Y., PARK, C., KIM, S.-J., HA, Y.A, CHAI, K.-Y., KIM, S.-M., KIM, K.-Y., PARK, R. 2007. Hibiscus sabdariffa L. water extract inhibits the adipocyte differentiation through the PI3-K and MAPK pathway. Journal of Ethnopharmacology, 114:260-267. KIM, M.S., KIM, J.K, KIM, H.J, MOON, S.R., SHIN, B.C., PARK, K.W., YANG, H.O., KIM, S.M., PARK, R. 2003. Hibiscus extract inhibits the lipid droplet accumulation and adipogenic transcription factors expression of 3T3-L1 preadipocytes. Journal of Alternative and Complementary Medicine, 9(4):499-504. KIM, S.-K., KIM, Y.-M., HONG, M.-J., RHEE, H.-I. 2005. Studies on the inhibitory effects of Eugenia aromaticum extract on pancreatic lipase. Agricultural Chemistry and Biotechnology, 48(2):84-88. KIM, H.-J., PARK, J.-M., KIM, J.-A., KO, B.-P. 2008b. Effect of herbal Ephedra sinica and Evodia ruteacarpa on body composition and resting metabolic rate: A randomized, double-blind clinical trial in Korean premenopausal women. Journal of Acupuncture and Meridian Studies, 1(2):128-138. KIM, S.-J., JUNG, J.Y., KIM, H.W., PARK, T. 2008c. Anti-obesity effects of Juniperus chinensis extract are associated with increased AMP-activated protein kinase expression and phosphorylation in the visceral adipose tissue of rats. Biological and Pharmaceutical Bulletin, 31(7):1415-1421. KIM, S.-H., PARK, H.-S., LEE, M.-S., CHO, Y.-J., KIM, Y.-S., HWANG, J.-T., SUNG, M.J., KIM, M.S., KWON, D.Y. 2008d. Vitisin A inhibits adipocyte differentiation through cell cycle arrest in 3T3-L1 cells. Biochemical and Biophysical Research Communications, 372:108-113. KIM, J.-Y., MOON, K.-D., SEO, K.-I., PARK, K.-W., CHOI, M.-S., DO, G.-M., JEON, Y.-K., CHO, Y.-S., LEE, M.-K. 2009a. Supplementation of SK1 from Platycodi radix ameliorates obesity and glucose intolerance in mice fed a high-fat diet. Journal of Medicinal Food, 12(3):629-636. KIM, S.-J., LEE, S.-J., LEE, S., CHAE, S., HAN, M.-D., MAR, W. NAM, K.-W. 2009b. Rutecarpine ameliorates bodyweight gain through the inhibition of orexigenic neuropeptides NPY and AgRP in mice. Biochemical and Biophysical Research Communications, 389:437-442. KIM, S.N., LEE, J.H., SHIN, H., SON, S.H.,KIM, Y.S. 2009c. Effects on in vitrodigested ginsenosides on lipid accumulation in 3T3-L1 adipocytes. Planta Medica, 75:596-601. 53 KIMURA, H., OGAWA, S., JISAKA, M., KIMURA, Y., KATSUBE, T., YOKOTA, K. 2006. Identification of novel saponins from edible seeds of Japanese horse chestnut (Aesculus turbinata Blume) after treatment with wooden ashes and their nutraceutical activity. Journal of Pharmaceutical and Biomedical Analysis, 41:1657-1665. KIMURA, H., OGAWA, S., KATSUBE, T., JISAKA, M., YOKOTA, K. 2008. Antiobese effects of novel saponins from edible seeds of Japanese horse chestnut (Aesculus turbinata Blume) after treatment with wood ashes. Journal of Agricultural and Food Chemistry, 56(12):4783-4788. KISHINO, E., ITO, T., FUJITA, K., KIUCHI, Y. 2006. A mixture of the Salacia reticulata (Kotahala himbutu) aqueous extract and cyclodextrin reduces the accumulation of visceral fat mass in mice and rats with high-fat diet-induced obesity. The Journal of Nutrition, 136:433-439. KLEIN, G., KIM, J., HIMMELDIRK, K., CAO, Y., CHEN, X. 2007. Antidiabetes and anti-obesity activity of Lagerstroemia speciosa. Evidence-based Complementary and Alternative Medicine, 4(4):401–407. KOBAYASHI, Y., NAKANO, Y., KIZAKI, M., HOSHIKUMA, K., YOKOO, Y., KAMIYA, T. 2001. Capsaicin-like anti-obese activities of evodiamine from fruits of Evodia rutaecarpa, a vanilloid receptor agonist. Planta Medica, 67:628-633. KOBAYASHI, K., YAMADA, K., MURATA, T., HASEGAWA, T., TAKANO, F, KOGA, K., FUSHIYA, S., BATKHUU, J., YOSHIZAKI, F. 2008a. Constituents of Rhodiola rosea showing inhibitory effect on lipase activity in mouse plasma and alimentary canal. Planta Medica, 74:1716-1719. KOBAYASHI, K., IHARA, S., KOBATA, A., ITOH, K., KUSUNOLI, N., YOSHIZAKI, F. 2008b. Inhibitory effect of Myrica bark on lipase activity in mouse plasma and gastrointestinal tract. Journal of Medicinal Food, 11(2):289-293. KURIHARA, H., ASAMI, S., SHIBATA, H., FUKAMI, H., TANAKA, T. 2003. Hypolipemic effect of Cyclocarya paliurus (Batal) Iljinskaja in lipid-loaded mice. Biological and Pharmaceutical Bulletin, 26(3):383-385. KURIYAN, R., RAJ, T., SRINIVAS, S.K., VAZ, M., RAJENDRAN, R., KURPAD, A.V. 2007. Effect of Caralluma fimbriata extract on appetite, food intake and anthropometry in adult Indian men and women. Appetite, 48:38-344. KWON, C.-S., SOHN, H.Y., KIM, S.H., KIM, J.H., SON, K.H., LEE, J.S., LIM, J.K., KIM, J.-S. 2003. Anti-obesity effect of Dioscorea nipponica Makino with lipaseinhibitory activity in rodents. Bioscience Biotechnology and Biochemistry, 67(7):1451-1456. LAMBERT, J.D., KENNET, M.J., SANG, S., REUHL, K.R., JU, J., YANG, C.S. 2010. Hepatotoxicity of high oral dose (-)-epigallocatechi-3-gallate in mice. Food and Chemical Toxicology, 48:409-416. 54 LE, P.M., BENHADDOU-ANDALOUSSI, A., ELIMADI, A., SETTAF, A., CHERRAH, Y., HADDAD, P.S. 2004. The petroleum ether extract of Nigella sativa exerts lipid-lowering and insulin-sensitizing actions in the rat. Journal of Ethnopharmacology, 94:251-259. LEE, I.-A., LEE, J.H., BAEK, N.-I., KIM, D.-H. 2005. Antihyperlipidaemic effect of crocin isolated from the fructus of Gardenia jasminoides and its metabolite crocetin. Biological and Pharmaceutical Bulletin, 28(11):2106-2110. LEMHADRI, A., EDDOUKS, M., SULPICE, T., BURCELIN, R. 2007. Antihyperglycaemic and anti-obesity effects of Capparis spinosa and Chamaemelum nobile aqueous extracts in HFD mice. American Journal of Pharmacology and Toxicology, 2(3):106-110. LI, F., LI, W., FU, H., ZHANG, Q., KOIKE, K. 2007a. Pancreatic lipase-inhibiting triterpenoid saponins from fruits of Acanthopanax senticosus. Chemical and Pharmaceutical Bulletin, 55(7):1087-1089. LI, X., LIU, Z., ZHANG, X., WANG, L., ZHENG, Y., YUAN, C., SUN, G. 2008. Isolation and characterization of phenolic compounds from the leaves of Salix matsudana. Molecules, 13:1530-1537. LI, Z.-B., WAND, J.-Y., JIANG, B., ZHANG, X.-L., AN, L.-J., BAO, Y.-M. 2007b. Benzobijuglone, a novel cytotoxic compound from Juglans mandshurica, induced apoptosis in HeLa cervical cancer cells. Phytomedicine, 14:846-852. LIEBERMAN, S. 2004. A new potential weapon for fighting obesity. Forskolin – the active diterpene in Coleus. Alternative and Complementary Therapies, 10(6):330333. LIMA, W.P, CARNEVALI JR, L.C., EDER, R., COSTA ROSA, L.F.B.P., BACCHI, E.M., SEELAENDER, M.C.L. 2005. Lipid metabolism in trained rats: Effect of guarana (Paullinia cupana Mart.) supplementation. Clinical Nutrition, 24:10191028. LIN, J., DELLA-FERA, M.A., BAILE, C.A. 2005. Green tea polyphenol epigallocatechin gallate inhibits adipogenesis and induces apoptosis in 3T3-L1 adipocytes. Obesity Research, 13(6):982-990. LIU F., KIM, J,-K., LI, Y., LIU, X.-Q., LI, J., CHEN, X. 2001. An extract of Lagerstroemia speciosa L. has insulin-like glucose uptake-stimulatory and adipocyte differentiation-inhibitory activities in 3T3-L1 cells. The Journal of Nutrition, 131:2242-2247. LIU, W., ZHENG, Y., HAN, L., WANG, H., SAITO, M., LING, M., KIMURA, Y. 2008. Saponins (ginsenosides) from stems and leaves of Panax quinquefolium prevented high-fat diet-induced obesity in mice. International Journal of Phytotherapy and Phytopharmacology, available online at: http://www.thefreelibrary.com/Saponins+(ginsenosides)+from+stems+and+ leaves+of+Panax+quinquefolium...-a0191955988 [Accessed: 27/11/2009]. 55 LIU, X., KIM, J.-K., LI, Y., LI, J., LIU, F., CHEN, X. 2005. Tannic acid stimulates glucose transport and inhibits adipocyte differentiation in 3T3-L1 cells. The Journal of Nutrition, 135:165-171. LUO, H., KASHIWAGI, A., SHIBAHARA, T., YAMADA, K. 2007. Decreased bodyweight without rebound and regulated lipoprotein metabolism by gymnemate in genetic multifactor syndrome animal. Molecular and Cellular Biochemistry, 299(1-2):93-98. MACLEAN, D.B., LUO, L.-G. 2004. Increased ATP content/production in the hypothalamus may be a signal for energy-sensing of satiety: studies of the anorectic mechanism of a plant steroidal glycoside. Brain Research, 1020:1-11. MADGULA, V.L.M., ASHFAQ, M.K., WANG, Y.-H., AVULA, B., KHAN, I.A., WALKER, L.A., KHAN, S.I. 2010a. Bioavailability, pharmacokinetics, and tissue distribution of the oxypregnane steroidal glycoside P57AS3 (P57) from Hoodia gordonii in mouse model. Planta Medica, 76(14):1582-1586. MADGULA, V.L.M., AVULA, B., PAWAR, R.S., SHUKLA, Y.J., KHAN, I.A., WALKER, L.A., KHAN, S.I. 2010b. Characterization of in vitro pharmacokinetic properties of hoodigenin A from Hoodia gordonii. Planta Medica, 76(1):62-69. MAHMOOD, S.A., LINDEQUIST, U. 2008. A pilot study on the effect of Catha edulis Frosk., (Celastraceae) on metabolic syndrome in Wokw rats. African Journal of Traditional Complementary and Alternative Medicines, 5(3):271-277. MARCHEI, E., PICHINI, S., PACIFICI, R., PELLEGRINI, M., ZUCCARO, P. 2006. A rapid and simple procedure for the determination of synephrine in dietary supplements by gas chromatography-mass spectrometry. Journal of Pharmaceutical and Biomedical Analysis, 41:1468-1472. MARTINS, F., NOSO, T.M, PORTO, V.B., CURIEL, A., GAMBERO, A., BASTOS, D.H.M., RIBEIRO, M.L., CARVALHO, P., 2010. Maté tea inhibits in vitro pancreatic lipase activity and has hypolipidaemic effect on high-fat diet-induced obese mice. Obesity, 18(1):42-47. MASUDA, Y., HARAMIZU, S., OKI, K., OHNUKI, K., WATANABE, T., YAZAWA, S., KAWADA, T., HASHIZUME, S.-I., FUSHIKI, T. 2003. Upregulation of uncoupling proteins by oral administration of capsiate, a non-pungent capsaicin analog. Journal of Applied Physiology, 95:2408-2415. MATHERN, J.R., RAATZ, S.K., THOMAS, W., SLAVIN, J.L. 2009. Effect of fenugreek fiber on satiety, blood glucose and insulin response and energy intake in obese subjects. Phytotherapy Research, 23(11):1543-1548. MATSUDA, A., WANG, Z., TAKAHASHI,S., TOKUDA, T., MIURA, N., HASEGAWA, J. 2009. Upregulation of mRNA of retinoid binding protein and fatty acid binding protein by cholesterol enriched-diet and effect of ginger on lipid metabolism. Life Sciences, 84:903-907. 56 MATTEI, R., DIAS, R.F., ESPÍNOLA, E.B., CARLINI, E.A., BARROS, S.B.M. 1998. Guarana (Paullinia cupana): toxic behavioural effects in laboratory animals and antioxidant activity in vitro. Journal of Ethnopharmacology, 60:111-116. MATTES, R.D., BORMANN, L. 2000. Effects of (-)-hydroxycitric acid on appetitive variables. Physiology and Behaviour, 71:87-94. MATTES, R.D., KRIS-ETHERTON, P.M., FOSTER, G.D. 2008. Impact of peanuts and tree nuts on body weight and healthy weight loss in adults. The Journal of Nutrition, 138(9):1741S-1745S. MENNITI-IPPOLITO F., MAZZANTI, G., SANTUCCIO, C., MORO, PA., CALAPAI, G., FIRENZUOLI F., VALERI, A., RASCHETTI, R. 2008. Surveillance of suspected adverse reactions to natural health products in Italy. Pharmacoepidemiology and Drug Safety, 17:626-635. MIYOSHI, N., NAKAMURA, Y., UEDA, Y., ABE, M., OZAWA, Y., UCHIDA, K., OSAWA,T. 2003. Dietary ginger constituents, galanals A and B, are potent apoptosis inducers in Human T lymphoma Jurkat cells. Cancer Letters, 199:113119. MOHLAPO, T.D., NG’AMBI, J.W., NORRIS, D., MALATJE, M.M. 2009. Effect of Hoodia gordonii meal supplementation at finisher stage on productivity and carcass characteristics of Ross 308 broiler chickens. Tropical Animal Health and Production, 41:1591-1596. MOLLAH, M.L., KIM, G.-S., MOON, H.-K., CHUNG, S.-K., CHEON, Y.-P., KIM, J.K., KIM, K.-S. 2008 Antiobesity effects of wild ginseng (Panax ginseng C.A. Meyer) mediated by PPAR-γ, GLUT4 and LPL in ob/ob mice. Phytotherapy Research, 23(2):220-225. MOONEY, M.H., FOGARTY, S., STEVENSON, C., GALLAGHER, A.M., PALIT, P., HAWLEY, S.A., HARDIE, DG., COXON, G.D., WAIGH, R.D., TATE, R.J., HARVEY, A.L., FURMAN, B.L. 2008. Mechanisms underlying the metabolic actions of galegine that contribute to weight loss in mice. British Journal of Pharmacology, 153:1669-1677. MORENO, D.A., ILIC, N., POULEV, A., BRASAEMLE, D.L., FRIED, S.K., RASKIN, I. 2003. Inhibitory effects of grape seed extract on lipases. Nutrition, 19:876-879. MORENO, D.A., ILIC, N., POULEV, A., RASKIN, I. 2006a. Effects of Arachis hypogaea nutshell extract on lipid metabolic enzymes and obesity parameters. Life Sciences, 78:2797-2803. MORENO, D.A., RIPOLL, C., ILIC, N., POULEV, A., AUBIN, C., RASKIN, I. 2006b. Inhibition of lipid metabolic enzymes using Mangifera indica extracts. International Journal of Food Agriculture and Environment, 4(1):21-26. 57 MORIMITSU Y., HAYASHI, K., NAKAGAWA, Y., FUJII, H., HORIO, F., UCHIDA, K., OSAWA, T. 2000. Antiplatelet and anticancer isothiocyanates in Japanese domestic horseradish, Wasabi. Mechanisms of Ageing and Development, 116:125-134. MORTON, S.C. 2005. Ephedra. Statistical Science, 20(3):242-248. MURALIDHARA, NARASIMHAMURTHY, K., VISWANATHA, S., RAMESH, B.S. 1999. Acute and subchronic toxicity assessment of debitterized fenugreek powder in the mouse and rat. Food and Chemical Toxicology, 37:831-838. MURASE, T., MISAWA, K., HARAMIZU, S., HASE, T. 2009. Catechin-induced activation of the LKB1/AMP-activated protein kinase pathway. Biochemical Pharmacology, 78:78-84. MURRAY, C.D.R., LE ROUX, C.W., EMMANUEL, A.V., HALKET, J.M., PRZYBOROWSKA, A.M., KAMM, M.A., MURRAY-LYON, I.M. 2008. The effect of Khat (Catha edulis) as an appetite suppressant is independent of ghrelin and PYY secretion. Appetite, 51:747-750. MYCEK, M.J., HARVEY, R.A., CHAMPE, P.C. 2000. Lippincott’s illustrated reviews: Pharmacology 2nd ed. Philadelphia: Lippincott-Raven publishers. NAJMI, A., NASIRUDDIN, M., KHAN, R.A., HAQUE, S.F. 2008. Effect of Nigella sativa oil on various clinical and biochemical parameters of insulin resistance syndrome. International Journal of Diabetes in Developing Countries, 28(1):11-14. NAKAGAWA, K., KITANO, M., KISHIDA, H., HIDAKA, T., NABAE, K., KAWABE, M., HOSOE, K. 2008. 90-Day repeated-dose toxicity study of licorice flavonoid oil (LFO) in rats. Food and Chemical Toxicology, 46:2349-2357. NGONDI, J.L., MAKAMTO, S.C., ETAME, S.L., OBEN, J. 2006. Effect of Triumphetta cordifolia on body weight and blood lipids in normolipidemic guinea pigs. Drug Development Research, 66(3):200-203. NINOMIYA, K., MATSUDA, H., SHIMODA, H., NISHIDA, N., KASAJIMA, N., YOSHINO, T., MORIKAWA, T., YOSHIKAWA, M. 2004. Carnosic acid, a new class of lipid absorption inhibitor from sage. Bioorganic and Medicinal Chemistry Letters, 14:1943-1946. NYKAMP, D.L., FACKIH, M.N, COMPTON, A.L. 2004. Possible association of acute lateral-wall myocardial infarction and bitter orange supplement. The Annals of Pharmacotherapy, 38(5):812-816. OGAWA, T., TABATA, H., KATSUBE, T., OHTA, Y., YAMASAKI, Y., YAMASAKI, M., SHIWAKU, K. 2009. Suppresive effect of hot water extract of wasabi (Wasabia japonica Matsum.) leaves on the differentiation of 3T3-L1 preadipocytes. Food Chemistry, 118:239-244. 58 OGAWA, Y., SEKITA, K., UMEMURA, T., SAITO, M., ONO, A., KAWASAKI, Y., UCHIDA, O., MATSUSHIMA, Y., INOUE, T., KANNO, J. 2004. Gymnema sylvestre leaf extract: A 52-week dietary toxicity study in wistar rats. Shokuhin Eisegaku Zasshi, 45(1):8-18. OHKOSHI, E., MIYAZAKI, H., SHINDO, K., WATANABE, H., YOSHIDA, A., YAJIMA, H. 2007. Constituents from the leaves of Nelumbo nucifera stimulate lipolysis in the white adipose tissue of mice. Planta Medica, 73:1255-1259. OLIVIER, I. 2005. Hoodia gordonii (Masson) Sweet ex Decne. Available from http://www.plantzafrica.com/planthij/hoodgord.htm [Accessed: 05/02/2008]. ONO, Y., HATTORI, E., FUKAYA, Y., IMAI, S., OHIZUMI, Y. 2006. Anti-obesity effect of Nelumbo nucifera leaves extract in mice and rats. Journal of Ethnopharmacology, 106:238-244. ORISAKWE, O.E., HUSAINI, D.C., AFONNE, O.J. 2004. Testicular effects of subchronic administration of Hibiscus sabdariffa calyx aqueous extract in rats. Reproductive Toxicology, 18:295-298. OSTOJIC, S.M. 2006. Yohimbine: The effects on body composition and exercise performance in soccer players. Research in Sports Medicine, 14(4):289-299. OZAKI, A., KITANO, M., FURUSAWA, N., YAMAGUCHI, H., KURODA, K., ENDO, G. 2002. Genotoxicity of gardenia yellow and its components. Food and Chemical Toxicology, 40:1603-1610. PALIT, P., FURMAN, B.L., GRAY, A.I. 1999. Novel weight-reducing activity of Galega officinalis in mice. Journal of Pharmacy and Pharmacology, 51(11):13131319. PANG, J., CHOI, Y., PARK, T. 2008. Ilex paraguariensis extract ameliorates obesity induced by high-fat diet: Potential role of AMPK in the visceral adipose tissue. Archives of Biochemistry and Biophysics, 476:178-185. PARK, T., JUNG, H.J., KIM, H.W. 2009. (-)-Cedrol isolated from Juniperus chinensis reduces diet-induced obesity through increased uncoupling protein expression in the visceral adipose tissue of mice. New Biotechnology, 255:513. PARK, Y.S., YOON, Y., AHN, H.S. 2007. Platycodon grandiflorum extract represses up-regulated adipocyte fatty acid binding protein triggered by a high fat feeding in obese rats. World Journal of Gastroenterology, 13(25):3493-3499. PATEL, D.K., PATEL, K.A., PATEL, U.K., THOUNAOJAM, M.C., JADEJA, R.N., ANSARULLAH, PADATE, G.S., SALUNKE, S.P., DEVKAR, R.V., RAMACHANDRAN, A.V. 2009. Assessment of lipid lowering effect of Sida rhomboidea.Roxb methanolic extract in experimentally induced hyperlipidemia. Journal of Young Pharmacists, 1(3):233-238. 59 PAWAR, R.S., SHUKLA, Y.J., KHAN, S.I., AVULA, B., KHAN, I.A. 2007. New oxypregnane glycosides from appetite suppressant herbal supplement Hoodia gordonii. Steroids, 72:524-534, PAWAR, R.S., SHUKLA, Y.J., KHAN, I.A. 2007. New calogenin glycosides from Hoodia gordonii. Steroids, 72:881-891. PETIT, P.R., SAUVAIRE, Y.D., HILLAIRE-BUYS, D.M., LECONTE, O.M., BAISSAC, Y.G., PONSIN, G.R., RIBES, G.R. 1995. Steroid saponins from fenugreek seeds: Extraction, purification, and pharmacological investigation on feeding behavior and plasma cholesterol. Steroids, 60:674-680. PITTLER, M.H., ERNST, E. 2004. Dietary supplements for body-weight reduction: a systematic review. American Journal of Clinical Nutrition, 79:529-536. PRASANTH, D., PADMAJA, R., SAMIULLA, D.S. 2001. Effect of certain plant extracts on α-amylase activity. Fitoterapia, 72:179-181. PRASANTH D., AMIT, A., SAMIULLA, D.S., ASHA, M.K., PADMAJA, R. 2001. αglucosidase inhibitory activity of Mangifera indica bark. Fitoterapia, 72:686-688. PREUSS, H.G., BAGCHI, D., BAGCHI, M., RAO, C.V.S., DEY, D.K., SATYANARAYANA, S. 2004. Effects of a natural extract of (-)-hydroxycitric acid (HAS-SX) and a combination of HCA-SX plus niacin-bound chromium and Gymnema sylvestre extract on weight loss. Diabetes Obesity and Metabolism, 6:171-180. PUSZTAI, A., GRANT, G., BUCHAN, W.C., BARDOCZ, S., DE CARVALHO, A.F.F.U., EWEN, S.W.B 1998. Lipid accumulation in obese Zucker rats is reduced by inclusion of raw kidney bean (Phaseolus vulgaris) in the diet. British Journal of Nutrition, 79:213-221. QIDWAI, W., BIN HAMZA, H., QURESHI, R., GILANI, A. 2009. Effectiveness, safety and tolerability of powdered Nigella sativa (Kalonji) seed in capsules on serum lipid levels, blood sugar, blood pressure, and body weight in adults: results of a randomized, double-blind controlled trial. The Journal of Alternative and Complementary Medicine, 15(6):639-644. QIN, Y., WU, X., HUANG, W., GONG, G., LI, D., HE, Y., ZHAO, Y. 2009. Acute toxicity and sub-chronic toxicity of steroidal saponins from Dioscorea zingiberensis C.H. Wright in rodents. Journal of Ethnopharmacology, 126:543-550. QURESHI, S., SHAH, A.H., AGEEL, A.M. 1992. Toxicity studies on Alpinia galanga and Curcuma longa. Planta Medica, 58:124-127. RÁCZ-KOTILLA, E., RÁCZ, G., SOLOMON, A. 1974. The action of Taraxacum officinale extracts on the body weight and diuresis of laboratory animals. Planta Medica, 26:212-217. 60 RAI, S., WAHILE, A., MUKHERJEE, K., SAHA, B.P., MUKHERJEE, P.K. 2006. Antioxidant activity of Nelumbo nucifera (sacred lotus) seeds. Journal of Ethnopharmacology, 104:322-327. RASEKH, H.R., NAZARI, P., KAMLI-NEJAD, M.., HOSSEINZADEH, L. 2008. Acute and subchronic oral toxicity of Galega officinalis in rats. Journal of Ethnopharmacology, 116:21-26. RATNASOORIYA, W.D., JAYAKODY, J.R.A.C., PREMAKUMARA, G.A.S. 2003. Adverse pregnancy outcome in rats following exposure to a Salacia reticulata (Celestraceae) root extract. Brazilian Journal of Medical and Biological Research, 36(7):931-935. RAZMOVSKI-NAUMOVSKI, V., HUANG, T.H.-W., TRAN, V.H., LI, G.Q., DUKE, C.C., ROUFOGALIS, B.D. 2005. Chemistry and pharmacology of Gynostemma pentaphyllum. Phytochemistry Reviews, 4:197-219. RODEIRO, I., DONATO, M.T., JIMÉNEZ, N., GARRIDO, G., DELGADO, R., GÓMEZ-LECHÓN, M.J. 2007. Effects of Mangifera indica L. aqueous extract (Vimang) on primary culture of rat hepatocytes. Food and Chemical Toxicology, 45:2506-2512. RODRIGUEZ-FRAGOSO, L., REYES-ESPARZA, J., BURCHIEL, S.W., HERRERA-RUIZ, D., TORRES, E. 2008. Risks and benefits of commonly used herbal medicines in Mexico. Toxicology and Applied Pharmacology, 227:125-135. RODRÍGUEZ-MORÁN, M., GUERRERO-ROMERO, F., LAZCANO-BURCIAGA, G. 1998. Lipid-and glucose-lowering efficacy of Plantago psyllium in Type II diabetes. Journal of Diabetes and its Complications, 12(5):273-278. RUBIN, I.D., BINDRA, J.S., CAWTHORNE, M.A. 2006. Extracts, compounds and pharmaceutical compositions having anti-diabetic activity and their use. US Patent 2006/0159779 A1. RUXTON, C.H.S. 2004. Efficacy of Zotrim: a herbal weight loss preparation. Nutrition and Food Science, 34(1):25-28. RUXTON, C.H.S., GARDNER, E.J. 2005. A review of the efficacy and safety of key ingredients of over-the-counter products for weight management. British Food Journal, 107(2):111-125. SAITO, M., UENO, M., OGINO, S., KUBO, K., NAGATA. J., TAKEUCHI, M. 2005. High dose of Garcinia cambogia is effective in suppressing fat accumulation in developing male Zucker obese rats, but highly toxic to testis. Food and Chemical Toxicology, 43:411-419. SÁNCHEZ DE MEDINA, F., GÁMEZ M.J., JIMÉNEZ, I., JIMÉNEZ, J., OSUNA, J.I., ZARZUELO, A. 1994. Hypoglycemic activity of juniper “berries”. Planta Medica, 60:197-200. 61 SARMA, D.N., BARRETT, M.L., CHAVEZ, M.L., GARDINER, P., KO, R., MAHADY, G.B. MARLES, R.J., PELLICORE, L.S., GIANCASPRO, G.I., LOW DOG, T. 2008. Safety of green tea extracts: A systematic review by the US Pharmacopoeia. Drug Safety, 31(6):469-484. SAYAMA, K., LIN, S., ZHENG, G., OGUNI, I. 2000. Effects of green tea on growth, food utilization and lipid metabolism in mice. In Vivo, 14(4):481-484. SCIENCELAB. 2008. Material safety data sheet: Juniper oil. Available at: http://www.sciencelab.com/xMSDS-Juniper_oil-9924428 [Accessed: 07/12/2009]. SHEKELLE, P.G., HARDY, M.L., MORTON, S.C., MAGLIONE, M., MOJICA, W.A., SUTTORP, M.J., RHODES, S.L., JUNGVIG, L., GAGNÉ, J. 2003. Efficacy and safety of ephedra and ephedrine for weight loss and athletic performance: a meta-analysis. Journal of the American Medical Association, 289(12):1537-1545. SHENG, L., QIAN, Z., ZHENG, S., XI, L. 2006. Mechanisms of hypolipidaemic effect of crocin in rats: crocin inhibits pancreatic lipase. European Journal of Pharmacology, 543:116-122. SHI, J., YAN, J., LEI, Q., ZHAO, J., CHEN, K., YANG. D., ZHAO, X., ZHANG, Y. 2009. Intragastric administration of evodiamine suppresses NPY and AgRP gene expression in the hypothalamus and decreases food intake in rats. Brain Research, 1247:71-78. SHIH, C.-C., LIN, C.-H., LIN, W.-L. 2008. Effects of Momordica charantia on insulin resistance and visceral obesity in mice on high-fat diet. Diabetes Research and Clinical Practice 81:134-143. SHIGEMATSU, N., ASANO, R., SHIMOSAKA, M., OKAZAKI, M. 2001. Effect of administration with the extract of Gymnema sylvestre R. Br leaves on lipid metabolism in rats. Biological and Pharmaceutical Bulletin, 24(6):713-717. SHIM, W.-S., BACK, H., SEO, E.-K., LEE, H.-T., SHIM, C.-K. 2009. Long-term administration of an aqueous extract of dried, immature fruit of Poncirus trifoliata (L.) Raf. suppresses body weight gain in rats. Journal of Ethnopharmacology, 126:294-299. SHIMODA, H., FURUHASHI, T., KAZUYOSHI, N., TAKAHIKO, N., MASAAKI, O. 2001. Thirteen-week repeat dose oral toxicity study of Salacia reticulata extract in rats. Japanese Journal of medicine and Pharmaceutical Science, 46(4):527-540. SHIN, J.-E., HAN, M.J., KIM, D.-H. 2003. 3-Methylethergalangin isolated from Alpinia officinarum inhibits pancreatic lipase. Biological and Pharmaceutical Bulletin, 26(6):854-857. SHIN, J.-E., HAN, M.J., SONG, M.-C., SONG, M.-S., BAEK, N.-I., KIM, D.-H. 2004. 5-Hydroxy-7-(4ʹ-hydroxy-3ʹ-methoxyphenyl)-1-phenyl-3-heptanone: a pancreatic lipase inhibitor isolated from Alpinia officinarum. Biological and Pharmaceutical Bulletin, 27(1):138-140. 62 SHIN, S.S., JUNG, Y.S., YOON, K.H., CHOI, S., HONG, Y., PARK, D., LEE, H., SEO, B.I., LEE, H.Y., YOON, M. 2009. The Korean traditional medicine Gyeongshingangjeehwan inhibits adipocyte hypertrophy and visceral adipose tissue accumulation by activating PPARα action in rat white adipose tissue. Journal of Ethnopharmacology, 127:47-54. SHUKLA, Y.J., PAWAR, R.S., DING, Y., LI, X.-C., FERREIRA, D., KHAN, I.A. 2009. Pregnane glycosides from Hoodia gordonii. Phyochemistry, 70:675-683. SIGMA-ALDRICH. 2006. Material Safety Data Sheet: Tannic acid. Available from: http://people.ccmr.cornell.edu/~ralph/MSDS/tannic_acid.pdf [Accessed: 07/12/2009] SINGH, B. 2007. Psyllium as therapeutic and drug delivery agent. International Journal of Pharmaceutics, 334:1-14. SMEETS, A.J., WESTERTERP-PLANTENGA, M.S. 2009. The acute effects of a lunch containing capsaicin on energy and substrate utilisation, hormones and satiety. European Journal of Nutrition, 48:229-234. SON, I.S., KIM, J.H., SOHN, H.Y., SON, K.H., KIM, J.-S., KWON, C.-S. 2007. Antioxidative and hypolipidemic effects of diosgenin, a steroidal saponin of yam (Dioscorea spp.), on high-cholesterol fed rats. Bioscience Biotechnology and Biochemistry, 71(12):3063-3071. SONI, M.G., BURDOCK, G.A., PREUSS, H.G., STOHS, S.J. OHIA, S.E. BAGCHI, D. 2004. Safety assessment of (-)-hydroxycitric acid and Super CitriMax®, a novel calcium/potassium salt. Food and Chemical Toxicology, 42:1513-1529. SOWEMIMO, A.A., FAKOYA, F.A., AWOPETU, I., OMOBUWAJO, O.R., ADESANYA, S.A. 2007. Toxicity and mutagenic activity of some selected Nigeria plants. Journal of Ethnopharmacology, 113:427-432. SULTAN, M.T., BUTT, M.S., ANJUM, F.M. 2009. Safety assessment of black cumin fixed and essential oil in normal Sprague dawley rats: Serological and hematological indices. Food and Chemical Toxicology, 47:2768-2775. SUZUKI, Y., UNNO, T., USHITANI, M., HAYASHI, K., KAKUDA, T. 1999. Antiobesity activity of extracts from Lagerstroemia speciosa L. leaves on female KK-Ay mice. Journal of Nutritional Science and Vitaminology, 45(6):791-795. TAKAHASHI, T., TABUCHI, T., TAMAKI, Y., KOSAKA, K., TAKIKAWA, Y., SATOH, T. 2009. Carnosic acid and carnosol inhibit adipocyte differentiation in mouse 3T3-L1 cells through induction of phase2 enzymes and activation of glutathione metabolism. Biochemical and Biophysical Research Communications, 382:549-554. TAM, S.W., WORCEL, M., WYLLIE, M. 2001. Yohimbine: a clinical review. Pharmacology and Therapeutics, 91:215-243. 63 TAN, M.-J., YE, J.-M., TURNER, N., HOHNEN-BEHRENS, C., KE, C.-Q., TANG, C.-P., CHEN, T., WEISS, H.-C., GESING, E.-R., ROWLAND, A., JAMES, D.E., YE, Y. 2008. Antidiabetic activities of triterpenoids isolated from bitter melon associated with activation of the AMPK pathway. Chemistry and Biology, 15:263273. THOUNAOJAM, M., JADEJA, R., ANSARULLA, DEVKAR, R., RAMACHANDRAN, A.V. 2009. Dysregulation of lipid and cholesterol metabolism in high fat diet fed hyperlipidemic rats: protective effect of Sida rhomboidea. roxb leaf extract. Journal of Health Science, 55(3):413-420. TOMINAGA, Y., MAE, T., KITANO, M., SAKAMOTO, Y., IKEMATSU, H., NAKAGAWA, K. 2006. Licorice flavonoid oil effects body weight loss by reduction of body fat mass in overweight subjects. Journal of Health Science, 52(6):672-683. TOMINAGA, Y., NAKAGAWA, K., MAE, T., KITANO, M., YOKOTA, S., ARAI, T., IKEMATSU, H., INOUE, S. 2009. Licorice flavonoid oil reduces total body fat and visceral fat in overweight subjects: a randomized, double-blind, placebo-controlled study. Obesity Research and Clinical Practice, 3:169-178. TSUJITA, T., SUMIYOSHI, M., HAN, L.-K., FUJIWARA, T., TSUJITA, J., OKUDA, H. 2003. Inhibition of lipase activities by citrus pectin. Journal of Nutritional Science and Vitaminology, 49(5):340-345. TSUJITA, T., TAKAKU, T. 2007. Lipolysis induced by segment wall extract from Satsuma mandarin orange (Citrus unshu Mark). Journal of Nutritional Science and Vitaminology, 53:547-551. TSUJITA, T., TAKAKU, T. 2008. Lipolysis induced by various citrus fruits. Nippon Shokuhin Kagaku Kogaku Kaishi, 55(3):102-108. TULP, O.L., HARBI, N. 2001. Effect of Hoodia plant on food intake and body weight in lean and obese LA/Ntul//cp-rats. FASEB Journal, 15(4):A404. TULP, O.L., HARBI, N. 2002. Effect of Hoodia plant on weight loss in congenic obese LA/Ntul//-cp rats. FASEB Journal, 16(4):A654. TURNBULL, W.H., THOMAS, H.G. 1995. The effect of a Plantago ovata seed containing preparation on appetite variables, nutrient and energy intake. International Journal of Obesity, 19(5):338-342. TURNER, E.H., LOFTIS, J.M., BLACKWELL, A.D. 2006. Serotonin a la carte: supplementation with the serotonin precursor 5-hydroxytryptophan. Pharmacolopgy and Therapeutics, 109:325-338. VAN HEERDEN, F.R., VLEGGAAR, R., HORAK, R.M., LEARMONTH, R.A., MAHARAJ, V., WHITTAL, R.D. 2002. Pharmaceutical compositions having appetite suppressant activity. US Patent 6,376,657 B1. 64 VAN HEERDEN, F.R. 2008. Hoodia gordonii: A natural appetite suppressant. Journal of Ethnopharmacology, 119:434-437. VAN HEERDEN, F.R., HORAK, R.M., MAHARAJ, V.J., VLEGGAAR, R., SENABE, J.V., GUNNING, P.J. 2007. An appetite suppressant from Hoodia species. Phytochemistry, 68:2545-2553. VAN LOON, L.J.C., VAN ROOIJEN, J.J.M., NIESEN, B., VERHAGEN, H., SARIS, W.H.M., WAGENMAKERS, A.J.M. 2000. Effects of acute (-)-hydroxycitrate supplementation on substrate metabolism at rest and during exercise in humans. American Journal of Clinical Nutrition, 72:1445-1449. VAN VUUREN, S.F., VILJOEN, A.M. 2009. Interaction between the non-volatile and volatile fractions on the antimicrobial activity of Tarchonanthus camphoratus. South African Journal of Botany, 75:505-509. VAN WYK, B.-E, GERICKE, N. 2000. People’s plants. Pretoria: Briza publications. VAN WYK, B.-E., WINK, M. 2004. Medicinal plants of the world. Pretoria: Briza publications. VIJAYAKUMAR, M.V., PANDEY, V., MISHRA, G.C., BHAT, M.K. 2010. Hypolipidemic effect of fenugreek seeds is mediated through inhibition of fat accumulation and upregulation of LDL receptor. Obesity, 18(4):667-674. VIRDI, J., SIVAKAMI, S., SHAHANI, S., SUTHAR, A.C., BANAVALIKAR, M.M., BIYANI, M.K. 2003. Antihyperglycemic effects of three extracts from Momordica charantia. Journal of Ethnopharmacology, 88:107-111. VOGELS, N., NIJS, I.M.T., WESTERTERP-PLANTENGA, M.S. 2004. The effect of grape-seed extract on 24h energy intake in humans. European Journal of Clinical Nutrition, 58:667-673. WANG, Z.Q., ZUBERI, A.R., ZHANG, X.H., MACGOWAN, J., QIN, J., YE, X., SON, L., WU, Q., LIAN, K., CEFALU, W.T. 2007. Effects of dietary fibers on weight gain, carbohydrate metabolism, and gastric ghrelin gene expression in mice fed a high-fat diet. Metabolism, 56:1635-1642. WANG, T., WANG, Y., YAMASHITA, H. 2009. Evodiamine inhibits adipogenesis via the EGFR-PKCα-ERK signaling pathway. FEBS Letters, 583:3655-3659. WATANABE, J., KAWABATA, J., KASAI, T. 1999. 9-Oxooctadeca-10,12-dienoic acids as acetyl-CoA carboxylase inhibitors from red pepper (Capsicum annuum L.). Bioscience Biotechnology and Biochemistry, 63(3):489-493. WAUTHOZ, N., BALDE, A., BALDE, E.S., VAN DAMME, DUEZ, P. 2007. Ethnopharmacology of Mangifera indica L. bark and pharmacological studies of its main C-glucosylxanthone mangiferin. International Journal of Biomedical and Pharmaceutical Sciences, 1(2):112-119. 65 WEIDNER, M.S., SIGWART, K. 2000. The safety of a ginger extract in the rat. Journal of Ethnopharmacology, 73:513-520. WESTERTERP-PLANTENGA, M., DIEPVENS, K., JOOSEN, A.M.C.P, BÉRUBÉPARENT, S, TREMBLAY, A. 2006. Metabolic effects of spices, teas, and caffeine. Physiology and Behaviour, 89:85-91 WESTERTERP-PLANTENGA, M.S., LEJEUNE, M.P.G.M., KOVACS, E.M.R. 2004. Body weight loss and weight maintenance in relation to habitual caffeine intake and green tea supplementation. Obesity Research, 13(7):1195-1204. WESTERTERP-PLANTENGA, M.S., SMEETS, A., LEJEUNE, M.P.G. 2005. Sensory and gastrointestinal satiety effects of capsaicin on food intake. International Journal of Obesity, 29:682-688. WON S.-R., KIM, S.-K., KIM, Y.-M, LEE, P.-H., RYU, J.-H., KIM, J.-W., RHEE, H.I. 2007. Licochalcone A: A lipase inhibitor from the roots of Glycyrrhiza uralensis. Food Research International, 40:1046-1050. WHO see WORLD HEALTH ORGANISATION. 2006. Fact sheet no 311. Available on: http://www.who.int/mediacentre/factsheets/fs311/en/print.html [Accessed: 2009/09/29]. WYNBERG, R., SCHROEDER, D., CHENNELS, R. (eds.). 2009. Indigenous peoples, consent and benefit sharing. Lessons from the San-Hoodia case. London: Springer. XIE, J.T., ZHOU, Y.-P., DEY, L., ATTELE, A.S., WU, J.A., GU, M., POLONSKY, K.S., YUAN, C.-S. 2002. Ginseng berry reduced blood glucose and body weight in db/db mice. Phytomedicine, 9:254-258. XU, B.J., HAN, L.K., ZHENG, Y.N., LEE, J.H., SUNG, C.K. 2005. In vitro inhibitory effect of triterpenoidal saponins from Platycodi radix on pancreatic lipase. Archives of Pharmacal Research, 28(2):180-185. YAMADA, J., SUGIMOTO, Y., UJIKAWA, M. 1999. The serotonin precursor 5hydroxytryptophan elevates serum leptin levels in mice. European Journal of Pharmacology, 383:49-51. YAMAKOSHI, J., SAITO, M., KATAOKA, S., KIKUCHI, M. 2002. Safety evaluation of proanthocyanidin-rich extract from grape seeds. Food and Chemical Toxicology, 40:599-607. YAMAMOTO, M., SHIMURA, S., ITOH, Y., OHSAKA, T., EGAWA, E., INOUE, S. 2000. Anti-obesity effects of lipase inhibitor CT-II, an extract from edible herbs, Nomame Herba, on rats fed a high-fat diet. International Journal of Obesity, 24:758-764. 66 YANG, X.W. 2008. Toxicological assessment on safety of water and 70% ethanolic extracts of nearly ripe fruit of Evodia rutaecarpa. Zhongguo Zhong Yao Za Zhi, 33(11):1317-1321. YOSHIKAWA, M., SHIMODA, H., NISHIDA, N., TAKADA, M., MATSUDA, H. 2002. Salacia reticulata and its polyphenolic constituents with lipase inhibitory and lipolytic activities have mild antiobesity effects in rats. The Journal of Nutrition, 132:1819-1824. YOSHIZUMI, K., HIRANO, K., ANDO, H., HIRAI, Y., IDA, Y., TSUJI, T., TANAKA, T., SATOUCHI, K., TERAO, J. 2006. Lupane-type saponins from leaves of Acanthopanax sessiliflorus and their inhibitory activity on pancreatic lipase. Journal of Agricultural and Food Chemistry, 54:335-341. YOSHIZUMI, K., MUROTA, K., WATANABE, S., TOMI, H., TSUJI, T., TERAO, J. 2008. Chiisanoside is not absorbed but inhibits oil absorption in the small intestine of rodents. Bioscience Biotechnology and Biochemistry, 72(4):1126-1129. YU, L., SHIRAI, N., SUZUKI, H. 2007. Effects of some Chinese spices on body weights, plasma lipids, lipid peroxides, and glucose, and liver lipids in mice. Food Science and Technology Research, 13(2):155-161. ZAOUI, A., CHERRAH, Y., MAHASSINI, N., ALAOUI, K., AMAROUCH, H., HASSAR, M. 2002. Acute and chronic toxicity of Nigella sativa fixed oil. Phytomedicine, 9:69-74. ZHAO, H.L., SIM, J.-S., SHIM, S.H., HA, Y.W., KANG, S.S., KIM, Y.S. 2005. Antiobese and hypolipidemic effects of platycodin saponins in diet-induced obese rats: evidence for lipase inhibition and calorie intake restriction. International Journal of Obesity, 29:983-990. ZHANG, J., KANG, M.-J., KIM, M.-J., KIM, M.-E., SONG, J.-H., LEE, Y.-M., KIM, J.-I. 2008. Pancreatic lipase inhibitory activity of Taraxacum officinale in vitro and in vivo. Nutrition Research and Practice, 2(4):200-203. ZHANG, J., SHEN, Q., LU, J.-C., LI, J.-Y., LIU, W.-Y., YANG, J.-J., LI, J., XIAO, K. 2010. Phenolic compounds from the leaves of Cyclocarya paliurus (Batal.) Ijinskaja and their inhibitory activity against PTP1B. Food Chemistry, 119:14911496. ZHENG, Q., KOIKE, K., HAN, L.-K., OKUDA, H., NIKAIDO, T. 2004. New biologically active triterpenoid saponins from Scabiosa tschiliensis. Journal of Natural Products, 67(4):604-613. 67 SECTION 1 Quality control In this section of the study, the development of rapid and cost-effective quality control analytical techniques for H. gordonii raw material and products were investigated. Quality control of products from botanical origin is of great importance as they are widely consumed by the general public and it is well-known that these products often vary in quality and chemical constituent profiles due to environmental and other factors (Smillie and Khan, 2010:175). In a recent paper Smillie and Khan (2010:175-185) eloquently described a comprehensive approach towards identifying and authenticating botanical products. The available tools suggested for authentication include macroscopic and microscopic techniques, genetic fingerprinting techniques, analytical chemical fingerprinting techniques such as ultra or high performance thin layer chromatography (UPLC/HPLC), liquid chromatography coupled to mass spectrometry (LC/MS), and high performance thin layer chromatography (HPTLC), as well as chemometric/metabolomic techniques such as proton nuclear magnetic resonance 1H-NMR and infrared (IR) spectroscopy coupled with chemometric analysis (Smillie and Khan 2010:175185). If an objective would be to satisfy this suggested approach using H. gordonii as an example, the scope of this study fills some of the current voids. Several analytical techniques have been developed for H. gordonii and its constituents using UPLC/LC-MS and HPTLC. However, this is the first time quantitative determination of P57 was attempted using other suggested tools such as IR spectroscopy in combination with chemometric analysis and quantitative HPTLC analysis. The traditional as well as new methods are all comprehensively discussed in the following two chapters. Microscopic and macroscopic and genetic fingerprint techniques such as polymerase chain reaction (PCR) have been attempted by Joshi et al. (2009:253-264) to distinguish between different species of Hoodia, the common adulterant O. ficus-indica as well as other related species with limited success. Macroscopically and microscopically, Hoodia 68 species could be distinguished from each other as well from all the other species. However, in powder form, differentiation between the species of Hoodia was not possible due to the similar physical characteristics of the small particles. Genomic deoxyribonucleic acid (DNA) analysis by PCR was useful to distinguish between different species of Hoodia and O. ficus-indica. In addition, the PCR method could distinguish between H. gordonii, H. parviflora and H. ruschii from dried plant tissue. However, when tablets/capsules were analysed, this method was rendered useless possibly due to degradation of DNA during manufacturing processes (Joshi et al., 2009:253-264). REFERENCES JOSHI, V., TECHEN, N., SCHEFFLER, B.E., KHAN, I.A. 2009. Identification and differentiation between Hoodia gordonii (Masson) Sweet ex. Decne., Opuntia ficus indica (L.) P. Miller, and related Hoodia species using microscopy and PCR. Journal of Herbs Spices and Medicinal Plants, 15:253-264. SMILLIE, T.J., KHAN, I.A. 2010. A comprehensive approach to identifying and authenticating botanical products. Clinical Pharmacology and Therapeutics, 87(2):175-185. 69 CHAPTER 3 Chromatographic techniques 3.1 LIQUID CHROMATOGRAPHY 3.1.1 INTRODUCTION The most commonly used method to analyse raw plant material for P57 content is liquid chromatography (LC). As P57 is the only recognised anti-obesity compound in H. gordonii, the first analytical methods were focused on its quantification only. When liquid chromatography coupled to ultraviolet detection (LC-UV) is used, detection of P57 is performed at 220 nm. This absorbance is due to the presence of a tigloyl group, present in all H. gordonii steroidal glycosides (Jansen et al., 2008:202). Avula et al. (2006:606) developed a method for the quantification of P57 using liquid chromatography electrospray ionisation mass spectrometry (LCESI-MS) and liquid chromatography with ultraviolet detection (LC-UV). The LCMS method had a total run time of 45 min including a 5 min wash and 15 min calibration period while the total run time for the LC-UV method was 55 min. Product testing revealed that 4 of the 6 products tested did not contain P57 or was below the detection limit and no P57 was detected in the common adulterant Opuntia ficus-indica (prickly pear). The P57 results for LC-UV and LC-MS were very similar but the limit of detection of LC-MS (10 ng/ml) was lower than that of LC-UV (100 ng/ml) (Avula et al., 2006:607-611). The same research group discovered many novel compounds from H. gordonii and subsequently followed up these discoveries with a chemical fingerprinting technique for Hoodia species and 11 oxypregnane glycosides with LC-UV. Due to the complexity of these plant extracts and abundance of similar compounds the run time was increased to 80 minutes and various other adjustments were made to the chromatographic conditions. This method could distinguish between Hoodia species and 23 other closely related genera, but could not authenticate H. gordonii as all 11 compounds were present in all Hoodia species (Avula et al., 2007:1527-1530). In a follow-up study ultra performance liquid chromatography (UPLC) in combination with UV and MS was used for chemical fingerprinting of Hoodia species (12 hoodigosides) and related genera and the quantification of P57 in several different dosage forms 70 of dietary supplements claiming to contain Hoodia. The use of UPLC enabled a considerable reduction in analysis time from 80 min for HPLC to 15 min. The UPLC-UV method was used for quantification of P57 and all the plant extracts and products were analysed with LC-MS. P57 was not detected in 26 of the 35 commercially available products tested. The nine other products had low amounts of P57 and with UPLC-UV one product tested negative for P57 while its presence was confirmed with the more sensitive UPLC-MS method. This paper also provided extraction methods for the many different dosage forms in which products claiming to contain Hoodia is sold (Avula et al., 2008:722-730). Another HPLC-UV quantification method was developed by Janssen et al. (2008:200). This method can be used to quantify several steroidal glycosides from plant material and even complex food products such as yoghurt and maize oil. However, this paper is more focused on method development and quantification results as such are not shown. The development of these methods contributes greatly towards the quality control of H. gordonii plant material and products. However, liquid chromatography methods are expensive, time-consuming, and laborious and require skilled personnel. Thus, simpler and faster methods of quality control of H. gordonii raw materials and products are desirable and were therefore investigated in this study. For the purposes of this study, plant samples were sent to an independent analytical laboratory for P57 quantification by LC-MS analysis. These results were used as reference values in the study to develop alternative analytical methods for quantification of P57. Further on in the study, an in-house HPLC method with photodiode array (PDA) detection was developed for the quantification of P57 in the samples generated from the in vitro transport studies. The LC-MS and HPLC quantification methods are briefly described in this chapter. 3.1.2 BACKGROUND AND LITERATURE REVIEW In the field of phytochemistry, chromatography is the key to obtaining pure compounds for structure elucidation, pharmacological testing or for development into therapeutic entities. In addition, it is the main analytical technique used in the quality control and standardisation of phytochemical products. Crude plant extracts are very complex, contain many different molecules and metabolites and the chemical nature of these constituents varies considerably within a given 71 extract (Wolfender, Rodriguez and Hostettmann, 1998:299-300; Marston, 2007:2785). High performance liquid chromatography (HPLC) has the ability to separate, tentatively identify and quantify the compounds present in any sample that can be dissolved in a liquid. Coupled to ultraviolet (UV) photodiode array detection (PDA) it allows the running of a chromatographic separation with simultaneous detection at different wavelengths. The resultant UV spectra of natural products provide information on the type or class of constituents contained within a sample. However, the variability of the physicochemical and spectroscopic parameters of these compounds causes numerous detection problems. For example, a product that does not contain a chromophore cannot be detected by an UV PDA detector. HPLC chromatograms can be used as fingerprints to authenticate plant material for quality control purposes and different species can be detected by comparison of their chemical composition (Wolfender, Rodriguez and Hostettmann, 1998: 300; Marston, 2007:2787-2789). Since every molecule possesses a specific molecular weight, detection of compounds using mass spectrometry (MS) can be ideally considered as a universal detection method. LC-MS is one of the most sensitive analytical methods and can be used to obtain information on the molecular weight and structure of compounds. Different LC/MS interfaces have been developed due to basic incompatibilities between HPLC and MS. The compounds within plant extracts have to be ionised to allow detection with MS, and the complexity of these crude plant extracts leads to difficulties as not one interface can detect all of the compounds contained within such an extract. For the analysis of plant chemical constituents and their secondary metabolites, the main interfaces used include electrospray ionisation (ESI), atmospheric pressure chemical ionisation (APCI), thermospray (TSP) and continuous flow-fast atom bombardment (CF-FAB). Each of these interfaces possesses unique characteristics and ESI is used for detection of larger polar molecules such as saponins while TSP allows for detection of smaller moderately polar molecules (Wolfender, Rodriguez and Hostettmann, 1998:299-300; Marston, 2007:2789-2788). The use of LC-hyphenated techniques plays a significant role as a strategic tool in phytochemical investigations as it 72 provides a great deal of preliminary information about the nature and composition of crude plant extracts with only a small amount of plant material. Unfortunately the availability of general LC-MS libraries of spectra of crude plant extract constituents still presents challenges (Wolfender, Ndjoko and Hostettmann, 2003:453). A schematic diagram of an LC-UV-MS setup is shown in Figure 3.1. HPLC pump HPLC pump Injector 8-port valve Waste Cation exchange column RPLC column Mass spectrometer UV detector Waste or fraction collector FIGURE 3.1: Schematic illustration of an HPLC apparatus coupled to an ultraviolet detector and a mass spectrometer (LC-UV-MS) 3.1.3 MATERIALS AND METHODS 3.1.3.1 Plant material, reagents and sample preparation For the LC-MS method, the pure P57 reference standard (ChromaDexTM Inc California, USA), was dissolved in 50% acetonitrile, and diluted to concentrations of 0.005, 0.01, 0.025, 0.05 and 0.1 mg/ml. The in-house control sample of authentic unadulterated H. gordonii plant material and the samples (n=145) to be tested were prepared in the same way for analysis. Ground dry plant material (0.5 g) was weighed off on an analytical balance and extracted in 10 ml of 50% acetonitrile and 0.1% formic acid for 2 h in an ultrasonic bath (Sonorex digital 10p). The filtered extract was transferred to an HPLC vial for analysis. 73 For the HPLC method a crude plant extract was prepared by extraction of the powder (2.4 g) from a commercial H. gordonii capsule product with acetonitrile in an ultrasonic bath (Sonorex digital 10p) for 10 min at 25 °C. The process was repeated three times to ensure maximal extraction, the filtered extracts combined and air-dried. The dried extract was redissolved in methanol for analysis. The P57 content of the commercial H. gordonii product was determined by LC-MS analysis in an independent laboratory. Pure P57 (ChromaDexTM Inc, California, USA) as reference standard was dissolved in methanol in concentrations of 1, 2, 3, 5, 7.5, 10 and 20 µg/ml to develop a calibration curve. 3.1.3.2 Liquid chromatography coupled to mass spectrometry (LC-MS) P57 was quantified using a Waters (MA, USA) API Q-TOF Ultima system equipped with a BEH C18 column (1.7 µm, 2.1 x 50 mm). The MS detector was operated in electron impact mode with the capillary voltage at 3.5 kV and the cone voltage at 35 kV (positive switching - ES+). The flow rate of the HPLC was 0.4 ml/min, the cone gas flow rate was 50 l/h and the desolvation gas flow rate was 350 l/h. The source temperature was 100 °C and the desolvation temperature 350 °C. The accurate mass (861.5 m/z) was extracted in ES+ mode within a mass range of 200 to 1800 m/z. The injection volume was 2 µl. Gradient elution was used with 0.1% formic acid in water (Solvent A) and acetonitrile (Solvent B) as mobile phases. The gradient started with 100% solvent A, changed to 20% solvent B over 0.5 min, to 100% solvent B over 11.5 min, followed by a 2 min isocratic step and a return to initial conditions for 1 min for a total run time of 15 min. Masslynx® (version 4.2) software was used for data analysis. The extracted mass chromatogram of m/z 861.5 which corresponds to the [M-H2O+H]+ was used for quantification. 3.1.3.3 High performance liquid chromatography (HPLC) An HPLC system consisting of a Waters (MA, USA) 2690 separation module and a Waters 996 photodiode array detector was used and separation was achieved using a Gemini C18 column (250 x 4.6 mm; 5 µm particle size; Phenomenex) equipped with a guard column (4 mm x 3.0 mm, Phenomenex). The mobile phase consisted of water containing 0.05% acetic acid (A) and acetonitrile (B). Gradient elution was used as follows: 100% A at 0 min, adjusted over 40 min to 20% A and 74 80% B. Each run was followed by a 5 min wash with 100% B and an equilibration period of 8 min. The flow rate was 1.0 ml/min, the column operated at a temperature of 30 °C and the detection wavelength was 220 nm. The chromatographic data was collected and analysed with Empower® software. 3.1.4 RESULTS 3.1.4.1 Liquid chromatography coupled to mass spectrometry (LC-MS) H. gordonii is a complex plant containing an abundance of chemical constituents as illustrated in Figure 3.2, which is an example of the LC-MS chromatogram for the plant material used as the in-house control standard in this method. FIGURE 3.2: LC-MS chromatogram of the plant material used as the H. gordonii control standard for the LC-MS analytical method An example of an extracted mass (861.5 m/z) chromatogram of both P57 (A) and the H. gordonii control standard material (B) is shown in Figure 3.3. An example of a calibration curve calculated from the area under the peaks for the respective reference standards is shown in Figure 3.4. A good correlation coefficient (R2) of 0.9991 was obtained for the calibration curve. 75 FIGURE 3.3: An extracted mass (861.5 m/z) chromatogram of P57 (A) and the H. gordonii control standard plant material (B) FIGURE 3.4: Calibration curve for the P57 reference standard as obtained by LCMS The percentage content of P57 in several (n=145) H. gordonii plant material samples as determined by LC-MS is shown in Figure 3.5. It can be seen that 76 there is a high variation in percentage P57 content in stem samples of H. gordonii plant materials collected from different sites, which range from 0.000-0.430%. FIGURE 3.5: The percentage of P57 determined with LC-MS in H. gordonii stem samples (n = 145) 3.1.4.2 High performance liquid chromatography (HPLC) An example of the HPLC chromatogram for H. gordonii plant material from a commercial capsule product that was used in the transport study is shown in Figure 3.6. Figure 3.7 shows the calibration curve for the P57 reference standard for dilutions of 1, 2, 3, 5, 7.5, 10 and 20 µg/ml. A good correlation coefficient of 0.9967 was obtained for this calibration curve. 3.1.5 CONCLUSIONS The correlation coefficients for the calibration curves for the LC-MS method (R2=0.9991) as well as the HPLC method (R2 = 0.9967) indicated that good linearity was obtained for both these analytical methods over concentration ranges of P57 applicable to the samples that were analysed in this study. Thus, the LCMS method was acceptable to quantify P57 in H. gordonii raw materials and products while the HPLC method was an acceptable analytical method to quantify P57 concentrations in the transport studies. 77 0.40 0.35 0.30 AU 0.25 0.20 0.15 0.10 0.05 0.00 -0.05 -0.10 5.00 10.0 0 15.0 0 20.0 0 25.0 0 Time (min) 30.0 0 35.0 0 40.0 0 45.0 0 FIGURE 3.6: An HPLC chromatogram of the H. gordonii plant material from a commercial capsule FIGURE 3.7: The calibration curve for the P57 reference standard with the HPLC method 78 3.2 QUALITATIVE HIGH PERFORMANCE THIN LAYER CHROMATOGRAPHY (HPTLC) 3.2.1 INTRODUCTION The high consumer demand and limited geographical distribution of H. gordonii leads to gross adulteration of up to 75% of H. gordonii products. Allegedly some of these products contain small quantities of real H. gordonii material blended with adulterant plant materials like Opuntia ficus-indica (prickly pear), Agave americana, Aloe ferox and Cereus jamacaru (Adams, 2006:S.a.; Avula et al., 2006:606; Widmer, Reich and DeBatt, 2008:21). Most companies selling H. gordonii products are acutely aware of the high rate of adulteration of these products and tend to highlight positive analytical test results accompanied by an example of a Convention on International Trade in Endangered Species (CITES) certificate to demonstrate authenticity. Some companies also indicate that their products contain high percentages of P57. The triterpene glycoside (named P57) was first isolated by Van Heerden et al. (1998) who pioneered the investigation into Hoodia species, and is currently the only perceived active compound in H. gordonii. Therefore the quality of H. gordonii raw material and products is determined in part by the presence of P57. The preferred analytical method currently used for determination of P57 content is high performance liquid chromatography (HPLC) coupled to a mass spectrometer (LC-MS) or HPLC coupled to a photodiode array detector (LC-UV) (Widmer, Reich and DeBatt, 2008:21). However, LC-MS requires skilled personnel and in addition it is both laborious and costly. Thin layer chromatography (TLC) is an option that has distinct advantages over several other chromatographic methods such as its simplicity, low cost, high sample capacity, and rapid availability of results. It is unique due to the fact that it presents the results as an easy-to-interpret image or chemical fingerprint (Marston, 2007:2787). Several papers discussing the authentication of H. gordonii and to detect the presence of P57 have been published. Widmer, Reich and DeBatt (2008:21-25) first investigated the HPTLC method used by Alkemists Pharmaceuticals, wellknown for the authentication of H. gordonii raw material before embarking on the development of their own method. Extraction from raw material and development 79 conditions were optimised and enabled H. gordonii to be distinguished from H. currorii, H. parviflora and O. ficus-indica. They also concluded that β-sitosterol can be used to check for the presence of P57 as it has the same Rf value (Widmer, Reich and DeBatt, 2008:23-24). This was criticised in a paper by Rumalla et al. (2008: 3959) as they mention that using only P57 or a non-specific compound like β -sitosterol may not establish the complete chromatographic identity of H. gordonii which comprises of at least two chemically distinct steroidal aglycones. The elegant method developed by Rumalla et al. (2008:3961-3964) provided a characteristic fingerprint for Hoodia species, but not H. gordonii specifically, with Rf values for 11 pregnane glycosides used as standards. The chemical fingerprint of H. gordonii was compared with 24 closely related plant species and none was comparable to H. gordonii. This thesis provides an alternative rapid and simple HPTLC method to authenticate H. gordonii raw material and products with good band separation confirmed by LC-MS as an accurate and reliable method for the confirmation of the presence of P57 in raw material and product samples. The new mobile phase, the use of Liebermann-Burchard reagent and the specificity for P57 which is still the main measure of quality as determined by industry standards distinguishes this method from previously published methods. 3.2.2 BACKGROUND AND LITERATURE REVIEW 3.2.2.1 Analysis of plant material The quality of plant material is traditionally divided into two aspects. Qualitative analysis deals with the authentication of material while quantitative analysis is focused on the measurement of certain biomarkers or active substances (Reich and Schibli, 2006:1). The complexity of plant material presents difficulty in analysis as plants contain numerous constituents including active and inactive compounds as well as unknown compounds. In addition, there is an innate variability in the composition of plant material due to differences in maturity, geographical origin and environmental factors such as mineral nutrition and exposure to stress. Chemical composition differences from grower to grower as well as from crop to crop are common and postharvest handling and storage also has an effect (Lazarowych and Pekos, 1998:497; Reich and Schibli, 2006:3). 80 For herbal extracts, regulatory authorities often recommend fingerprint chromatography for proper identification purposes and the British Herbal Pharmacopoeia for example emphasises the use of TLC profiles for this purpose (Lazarowych and Pekos, 1998:497; Marston, 2007:2787). An authentic botanical sample which is extracted and chromatographed provides a unique chromatographic fingerprint and primary reference for comparison against test samples. TLC is the ideal screening method as it not only provides authentication or identification with determination of adulteration in some cases but also quantitative or semi-quantitative determination of biomarkers (Reich and Schibli, 2006:4; Marston, 2007:2787). HPTLC is an improvement on TLC in that the stationary phase particle size on the plate is much smaller, 5 µm as opposed to 12 µm, providing greater separation efficiency, reproducibility, improved detection, and faster separation over shorter developing distances (Marston, 2007:2787). 3.2.3 MATERIALS AND METHODS 3.2.3.1 Plant material and reagents A number of H. gordonii stem samples were collected at various locations throughout South Africa and Namibia from natural populations as well as from cultivated sites. The aerial parts of Opuntia ficus-indica, Aloe ferox, Cereus jamacaru and Agave americana from wild populations were collected in South Africa. The samples were hand-picked, sliced and air-dried. The dried plant material was ground with a Retsch® 400 MM ball mill (Haan, Germany) at a frequency of 30 Hz for 90 s. H. gordonii containing products (capsules, tablets, chips, spray, topical gel, patch, fruit bar, compote, instant drink mix, oral strips) were purchased from various sources including retail stores in South Africa as well as internet purchases from international companies. The P57 reference standard was purchased from ChromaDexTM Inc (California, USA). 3.2.3.2 High performance thin layer chromatography (HPTLC) analysis HPTLC analysis was performed on a CAMAG semi-automated HPTLC system (Muttenz, Switzerland) including the automatic TLC Sampler 4 (connected to a nitrogen tank), automatic developing chamber ADC2, chromatogram immersion device, TLC plate heater III, and documentation device Reprostar 3 (Figure 3.8) with winCATS version 1.4.4.6337 planar chromatography manager software. 81 FIGURE 3.8: A diagram of the CAMAG semi-automated HPTLC system including the automatic TLC Sampler 4, automatic developing chamber ADC2, chromatogram immersion device, TLC plate heater III and documentation device Reprostar 3 The powdered raw plant material and most products were extracted with acetonitrile in an ultrasonic bath (Sonorex digital 10p) at 100% power and 25 °C for 20 min and filtered. This process was repeated three times with fresh acetonitrile to maximise extraction efficiency and the filtrates were combined and air-dried. In a preliminary study, the combination of acetonitrile and sonication was determined to be a good extraction method for compounds from H. gordonii. The buccal spray product was placed in a vacuum oven (Vismara) at 40 °C to dry by evaporation. The compote, fruit bar and oral strips were extracted with a mixture of acetonitrile and water (50:50) to aid in dissolution and release of P57 from poorly soluble excipients used in these products. The dried extracts were made up to a concentration of 50 mg/ml and 8 µl were spray-applied with a 25 µl syringe together with 2 µl of a 1 mg/ml solution of the reference standard P57 as 10 mm bands on 20 x 10 cm silica gel plates (Silica gel 60 F254, Merck, Germany). No chamber saturation was performed prior to development. Temperature and humidity were not controlled via the HPTLC system but the experiments were carried out in a temperature controlled laboratory. The bands were sprayed 10 mm from the bottom of the plate, with the first position at 30 mm from the side 82 edge. After a pre-drying phase of 5 min, ascending development was carried out over a path of 85 mm with the mobile phase (15 ml) consisting of toluene, chloroform and ethanol (40:40:12.5 v/v/v). This mobile phase and chamber conditions resulted in plates with good resolution and well-defined spots. A 5 min drying phase followed development. To visualise the chromatogram, derivatisation was performed by immersion of the plate in Liebermann-Burchard (LB) reagent and heating at 100 °C for 5-10 min. This reagent is traditionally used for detection of triterpenes and steroids (saponins, bitter principles) and therefore it was expected that P57 (a triterpene glycoside) would be clearly visualised. This is the first time LB reagent has been used as derivatising agent in an HPTLC quality control method for H. gordonii raw material and products. LB reagent was prepared by adding 10 ml of acetic anhydride and 10 ml of concentrated sulphuric acid to 100 ml of absolute ethanol while cooling on ice (Wagner and Bladt, 1996:362). The images were captured under ultraviolet light at 366 nm. The limit of detection for P57 was determined by spotting dilutions containing 5 µg, 3 µg, 2 µg, 1.5 µg, 1 µg, 0.5 µg, 0.25 µg 0.15 µg and 0.1 µg per 5 µl (spot volume), respectively. The smallest amount of the analyte that provided a visual response was considered the limit of detection for P57. To establish that the co-elution of compounds (especially P57 analogues) does not occur in the area of the reference standard and to confirm that the bands perceived to be P57 were indeed P57, preparative TLC was done. Several plates were spotted and developed using a sample with a known high quantity of P57 as determined by LC-MS. The bands corresponding to the reference compound (P57) were scraped off from the plates, washed with methanol and analysed by LC-MS. 3.2.3.3 Liquid chromatography coupled to mass spectrometry (LC-MS) analysis P57 was quantified using a Waters (MA, USA) API Q-TOF Ultima system equipped with a BEH C18 column (1.7µm, 2.1 x 50 mm). The MS detector was operated in electron impact mode with the capillary voltage at 3.5 kV and the cone voltage at 35 kV (positive switching - ES+). The flow rate of the HPLC was 0.4 ml/min, the cone gas flow rate was 50 l/h and the desolvation gas flow rate was 350 l/h. The source temperature was 100 °C and the desolvation temperature 83 350 °C. The accurate mass (861.5 m/z) was extracted in ES+ mode within a mass range of 200 to 1800 m/z. The injection volume was 2 µl. Gradient elution was used with 0.1% formic acid in water (Solvent A) and acetonitrile (Solvent B) as mobile phases. The gradient started with 100% solvent A, changed to 20% solvent B over 0.5 min, to 100% solvent B over 11.5 min, followed by a 2 min isocratic step and a return to initial conditions for 1 min for a total run time of 15 min. The LC-MS data was analysed using Masslynx® (version 4.2) software. The specific mass of 861.5 (m/z) was used for quantification of P57. 3.2.4 RESULTS AND DISCUSSION 3.2.4.1 High performance thin layer chromatography (HPTLC) Figure 3.9 displays the HPTLC chromatogram for determining the limit of detection for P57. The smallest amount spotted that provided a visual response was considered the limit of detection. Although there is a band visible in track 9, it is so unclear that the limit of detection should be considered 0.15 µg (track 8) or even 0.25 µg (track 7). This indicated that the HPTLC method will be able to confirm the presence of P57 even in highly diluted samples. The P57 band eluted consistently at a retention front (Rf) value of 0.42 on consecutive days, indicating acceptable precision and reproducibility of the method. FIGURE 3.9: The HPTLC chromatogram used to determine the limit of detection for P57. P57 was spotted at 5 µg (track 1), 3 µg (track 2), 2 µg (track 3), 1.5 µg (track 4), 1 µg (track 5), 0.5 µg (track 6), 0.25 µg (track 7), 0.15 µg (track 8) and 0.1 µg (track 9) 84 Figures 3.10-3.12 display HPTLC chromatograms for several weight loss consumer products claiming to contain H. gordonii viewed under UV light at 365 nm with the first track being raw H. gordonii standard material and the last track being the P57 reference standard for each figure. FIGURE 3.10: An HPTLC plate viewed under UV at 365 nm. Track 1: H. gordonii standard raw material, track 2-8: H. gordonii commercial products (1-7) from different manufacturers and track 9: the P57 reference standard On comparison with H. gordonii raw material (track 1), it is evident that all the products presented in Figure 3.10 contain H. gordonii plant material and P57 where the lowest concentrations were obtained for Product 3 (track 4), Product 4 (track 5), Product 6 (track 7) and Product 7 (track 8) and the highest concentration for product 5 (track 6) as detected qualitatively through visual inspection of band intensity. In Figure 3.11 it is shown that products 13 and 14 (tracks 7 and 8) contain the lowest concentration of P57 and that all the products presented in this figure contain H. gordonii. 85 FIGURE 3.11: An HPTLC plate viewed under UV at 365 nm. Track 1: H. gordonii standard raw material, track 2-8: H. gordonii commercial products (8-14) from different manufacturers and track 9: the P57 reference standard It is evident that none of the products presented in Figure 3.12 contain H. gordonii nor P57 (or negligible amounts below the limit of detection) based on comparison with the standard H. gordonii raw material chromatogram (chemical fingerprint) in track 1, and these include Product 15 topical gel (0.006%), Product 16 transdermal patch (0.000%), Product 17 fruit bar (0.001%), Product 18 compote (0.003%), Product 19 instant drink mix (0.000%) and Product 20 oral strips (0.000%). These results were confirmed by the LC-MS results (Table 3.1) FIGURE 3.12: An HPTLC plate viewed under UV at 365 nm. Track 1: H. gordonii standard raw material, track 2-8: H. gordonii commercial products (15-20) from different manufacturers and track 9: the P57 reference standard 86 TABLE 3.1: Comparison of the P57% LC-MS results and HPTLC results Figure Track 3.10 1 3.11 3.12 3.13 3.14 Sample description LC-MS result %P57 HPTLC result + or – for P57 Namibian cultivated raw material 0.186 + 2 Product 1 (Capsules) 0.248 + 3 Product 2 (Capsules) 0.162 + 4 Product 3 (Capsules) 0.131 + 5 Product 4 (Capsules) 0.080 + 6 Product 5 (Capsules) 0.304 + 7 Product 6 (Capsules) 0.191 + 8 Product 7 (Capsules) 0.102 + 1 Namibian cultivated raw material 0.186 + 2 Product 8 (Capsules) 0.134 + 3 Product 9 (Capsules) 0.134 + 4 Product 10 (Tablets) 0.064 + 5 Product 11 (Capsules) 0.180 + 6 Product 12 (Capsules) 0.169 + 7 Product 13 (Chips – sliced H. gordonii) 0.058 + 8 Product 14 (Buccal spray) 0.048 + 1 Namibian cultivated raw material 0.186 + 2 Product 15 (Topical gel) 0.006 - 3 Product 16 (Transdermal patch) 0.000 - 4 Product 17 (Fruit bar) 0.001 - 5 Product 18 (Compote) 0.003 - 6 Product 19 (Instant drink mix) 0.000 - 7 Product 20 (Oral strips) 0.000 - 1 South African wild 0.127 + 2 Namibian wild 0.161 + 3 South African cultivated 0.115 + 4 Namibian cultivated 0.186 + 5 2 yr old actively growing cultivated 0.131 + 6 2 yr old dormant cultivated 0.288 + 7 3 yr old actively growing cultivated 0.048 + 8 3 yr old dormant cultivated 0.112 + 1 Namibian cultivated raw material 0.186 + 3.15 1 Namibian wild raw material 0.161 + + = visually detectable on the HPTLC plate; - = not visually detectable on the HPTLC plate 87 Figure 3.13 is an HPTLC plate of two wild population samples from different localities and samples of differing maturity and growing stages. There are some qualitative differences in terms of band intensity between the samples for example on comparison of track 2, a wild population sample and track 4, a cultivated site sample. This can be explained by the fact that certain phytoconstituents in cultivated plants can sometime be much lower compared to those in wild populations of the same plant species. Plants manufacture secondary metabolites, which are mainly responsible for the medicinal properties of the plant, under certain conditions of stress in their natural environment which may not be expressed when cultivated under more controlled conditions (Schippmann, Leaman and Cunningham, 2006:81). In addition, the two wild samples (tracks 1 and 2) showed a red band situated right below the P57 band which was absent in all the cultivated samples (tracks 3 to 8). It was visually evident from the intensity of the bands that both the 3 yr old plants (tracks 7 and 8) contained less P57 compared to the 2 yr old plants (tracks 5 and 6). This difference in the concentration of P57 was confirmed with LC-MS (Table 3.1). FIGURE 3.13: An HPTLC plate viewed under UV at 365 nm. Track 1: wild population sample of H. gordonii from South Africa, track 2: wild population sample of H. gordonii from Namibia, track 3: cultivated sample of H. gordonii from South Africa, track 4: cultivated sample of H. gordonii from Namibia, track 5: 2 year old actively growing plant sample, track 6: 2 year old dormant plant sample, track 7: a 3 year old actively growing cultivated plant, track 8: a 3 year old dormant cultivated plant and track 9: the P57 reference standard 88 Figure 3.14 is an HPTLC plate spotted with H. gordonii raw plant material and H. gordonii raw plant material with gradually increasing adulteration concentrations of O. ficus-indica plant material. FIGURE 3.14: An HPTLC plate viewed under UV at 365 nm. Track 1: H. gordonii standard raw material, track 2-7: the same sample of H. gordonii standard raw material as for track 1 adulterated with O. ficus-indica in concentrations of 10%, 20%, 40%, 50 %, 60%, 80%, respectively, track 8: O. ficus-indica raw material, track 9: the P57 reference standard It is evident that P57 was present in the samples spotted from track 1-7, but the intensity decreased as the percentage of adulteration increased up to a barely visible band with low intensity in track 7. Furthermore, O. ficus-indica (track 8) clearly does not contain all the compounds found in H. gordonii (including P57) and displays only eight bands for this mobile phase. Another visible difference between the composition of H. gordonii plant material and O. ficus-indica plant material is at an Rf value of 0.4 where H. gordonii shows a black band and O. ficus-indica yielded a red band. Figure 3.15 shows HPTLC chromatograms of plant materials allegedly used to adulterate H. gordonii products, including O. ficus-indica, A. ferox, C. jamacaru and A. americana. It is evident from these plates that it would be easy to distinguish A. ferox, C. jamacaru and A. americana from H. gordonii raw material 89 with HPTLC but that adulteration with O. ficus-indica would be more difficult to detect as shown in Figure 3.14. FIGURE 3.15: An HPTLC plate viewed under UV at 365 nm. Track 1: H. gordonii standard raw material, Tracks 2 and 3: Opuntia ficus-indica samples, Tracks 4 and 5: Aloe ferox samples, Track 6: Cereus jamacaru sample, Track 7: Agave americana sample and Track 8: the P57 reference standard 3.2.4.2 Liquid chromatography coupled to mass spectrometry (LC-MS) Figure 3.16 displays the LC-MS chromatograms used to establish that co-elution of compounds at the specific Rf value (0.42) of P57 does not occur. Chromatogram A in this figure shows the P57 reference standard and chromatogram B shows the band isolated with preparative TLC. It is evident from LC-MS analysis that the major peak obtained for the extracted compounds was P57 with an accurate mass of 861.5 m/z. 90 6.76 100 % A 6.94 0 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00 12.00 13.00 14.00 15.00 6.77 100 % B 6.95 6.66 0 1.00 2.00 3.00 4.00 5.00 6.00 7.24 7.00 14.57 8.00 9.00 10.00 11.00 12.00 13.00 14.00 Time 15.00 FIGURE 3.16: LC-MS chromatograms of the P57 standard (A) and the preparative TLC sample (B) The mass spectra for each of the chromatograms in Figure 3.16 are displayed in Figure 3.17. 313.190 100 A % 313.209 861.435 457.256 314.201 761.388 314.216 458.260 315.195 381.209 413.240 557.316 466.233 862.440 761.417 762.392 601.327 763.408 862.503 901.423 829.408 981.390 0 313.190 100 % B 861.435 457.256 314.201 761.388 458.260 314.216 315.195 0 300 350 381.209 413.240 400 557.316 466.228 450 500 550 862.440 762.392 601.327 600 862.503 763.408 650 700 750 800 829.402 850 901.423 900 950 m/z 1000 FIGURE 3.17: Mass spectra of the P57 reference standard (A) and the preparative TLC sample (B) 91 Co-elution of bands is a common problem with HPTLC analysis of complex samples such as plant material but clearly this method establishes that P57 elutes as a single compound at Rf 0.42. This was confirmed with preparative TLC and LC-MS analysis indicating specificity of this method. Thus, the HPTLC method has the ability to clearly confirm the presence of P57 with the specified mobile phase and derivatising agent used. To validate that P57 was indeed present in all the H. gordonii raw material and product samples that were spotted on the TLC plates, the concentration of P57 was determined with LC-MS and the results are collectively presented in Table 3.1. 3.2.5 CONCLUSIONS The quality of commercial H. gordonii-containing products is a cause for concern especially because these products are consumed on a daily basis by many people worldwide. The lack of quality in terms of P57 content in several commercially available products is highlighted in this study. Some of the products investigated, especially the topical products and some oral dosage forms, contain very low concentrations or no P57 at all. Since P57 is currently considered to be the only active ingredient in H. gordonii for weight-loss, this simple HPTLC technique provides a fast and cost-effective method for screening products for the presence of P57. Furthermore, developing new rapid and accurate methods for basic quality control could contribute to the distribution of only high quality raw material, and consequently, more safe high quality consumer products. This technique could be developed further to perform the quantitative determination of P57 in raw materials and products. 3.3 QUANTITATIVE HIGH PERFORMANCE THIN LAYER CHROMATOGRAPHY (HPTLC) 3.3.1 INTRODUCTION Several HPTLC-based quality control methods for H. gordonii raw material and products have been developed as discussed in the previous section. To the author’s knowledge quantitative HPTLC has not been used before to quantify the amount of P57 within H. gordonii products. It has been used however to quantify marker molecules found in several plant species. Eugenol and gallic acid from Syzygium aromaticum (L.) Merr. et Perry (Myrtaceae) (Pathak et al., 2004:241), 92 harpagoside from Harpagophytum procumbens DC (Pedaliaceae) (Günther and Schmidt 2005:817), luteolin from Thymus vulgaris L. (Lamiaceae) (Bazylko and Strzelecka, 2007:931) and steviol glycosides (steviolbioside, stevioside, rebaudioside-A) from Stevia rebaudiana Bertoni (Asteraceae) (Jaitak et al., 2008:790) amongst others have all been successfully quantified from powdered samples using HPTLC coupled to densitometry. The techniques traditionally used to quantify P57 (LC-MS, LC-UV) are laborious, time-consuming and expensive. HPLTC-densitometric quantification is a method of choice due to high sample throughput at lower operating costs and shorter analysis times. Quantification of the important biomarker, P57, from H. gordonii samples will contribute greatly to quality control of raw material and products. 3.3.2 BACKGROUND AND LITERATURE REVIEW Densitometry involves quantifying the bands on a chromatogram by measuring the amount of light reflected from the plate to a detector. It is particularly useful and accurate for the quantification of marker compounds and is sensitive because of its spectral selectivity. Light in the range of 190-800 nm from different light sources can be used to measure absorption or fluorescence with various light sources such as a deuterium lamp (190-450 nm), tungsten lamp (370-800 nm) and a mercury vapour lamp with high-energy output at several individual wavelengths (including 254 and 366 nm). A narrow beam of light of a specified wavelength and geometrical dimension is allowed to reach the plate and after detection the chromatogram data is integrated and mathematically evaluated to produce peak height and area for each band or substance zone. A calibration function can be calculated when a set of standards are chromatographed together with the samples to be quantified. The measurements of the unknown compound will then be relative to the standard. The functional relationship between the signal and the amount of substance is more complicated in TLC since light is reflected from both the surface of the plate as well as the surrounding area. With fluorescence measurement the signal is proportional to the intensity of the light used for excitation and a simple linear equation usually applies. For most absorption measurements the data is best fitted by polynomial functions (Reich and Schibli, 2006:61, 63, 116). 93 3.3.3 MATERIALS AND METHODS 3.3.3.1 Plant material and reagents H. gordonii containing products (capsules) were purchased from various sources including retail stores in South Africa as well as internet purchases from international companies. The P57 reference standard was purchased from ChromaDexTM Inc (California, USA). 3.3.3.2 Quantitative high performance thin layer chromatography analysis This quantitative method was based upon the previously discussed qualitative HPTLC method (3.2.3.2) used to detect the presence of P57 with minor modifications. HPTLC analysis was performed on a CAMAG semi-automated HPTLC system (CAMAG, Muttenz, Switzerland) including the automatic TLC Sampler 4 (connected to a nitrogen tank), automatic developing chamber ADC2, chromatogram immersion device, TLC plate heater III, documentation device Reprostar 3 and TLC Scanner 3 (Figure 3.20) with winCATS version 1.4.4.6337 planar chromatography manager software. FIGURE 3.18: A photograph of the CAMAG TLC scanner 3 used in the quantification of P57 from commercial samples (Photograph by I Vermaak) 94 The powdered raw plant material (1 g) from the commercial capsules were extracted with acetonitrile in an ultrasonic bath (Sonorex digital 10p) at 100% power and 25 °C for 20 min and filtered. This process was repeated three times with fresh acetonitrile to maximise extraction efficiency, the filtrates combined and air-dried. The dried extracts of four products were made up to a concentration of 10 mg/ml, sonicated for 1 min to facilitate dissolution and 3000 nl were sprayapplied in duplicate under a flow of nitrogen gas with a 25 µl syringe together with four concentrations (65.50, 75.00, 100.00, 125.00 ng) of the P57 reference standard (0.125 mg/ml). The 6 mm bands were applied 25 mm from the sides and 10 mm from the bottom on 20 x 10 cm silica gel plates (Silica gel 60 F254, Merck, Germany). Fresh samples were prepared daily and the plates were washed with methanol prior to use. The developing chamber was saturated with 25 ml of mobile phase and a saturation pad for 20 min ±30°C and conditioned to 47% relative humidity (RH) using potassium thiocyanate (KSCN). After a plate predrying phase of 5 min, ascending development was carried out over a path of 85 mm with the mobile phase (15 ml) consisting of toluene, chloroform and ethanol (40:40:12.5 v/v/v). The bands were visualised by immersion of the plate in Liebermann-Burchard (LB) reagent and heating at 100 °C for 5-10 min. LB reagent was prepared daily by adding 10 ml of acetic anhydride and 10 ml of concentrated sulphuric acid to 100 ml of absolute ethanol while cooling on ice (Wagner and Bladt, 1996:362). Image acquisition and archiving was done with the Reprostar 3 documentation device at a wavelength of 366 nm. Quantitative evaluation of the plate was performed with a mercury lamp (Hg) in remission-fluorescence mode with a K400 optical filter at a wavelength of 366 nm. The slit dimensions were 4.00 x 0.30 mm, scanning speed 20 mm/s and data resolution 100 µm/step. The results were compared in terms of height as well as area and the two values were averaged to evaluate the method. The method was validated with respect to intra-day precision (repeatability), inter-day precision (intermediate precision) and accuracy. Intra-day precision is the precision of the method over a short interval of time (day) under the same operating conditions while intermediate precision is the variation on different days. The analyses were done three times on one day to determine intra-day precision and on two consecutive days to determine inter-day precision. Accuracy of the standards was 95 expressed as a percentage by dividing the observed concentration by the theoretical concentration spotted multiplied by 100 (Jaitak et al., 2008:790). 3.3.4 RESULTS AND DISCUSSION 3.3.4.1 Quantitative high performance thin layer chromatography analysis The chamber conditions and mobile phase consisting of toluene, chloroform and ethanol (40:40:12.5 v/v/v) gave good resolution of P57. In addition, the Rf value of 0.49 was consistent on the same day as well as on consecutive days indicating precision and reproducibility of the method. Figure 3.19 is an example of a chromatogram that was used for the densitometric quantification of P57 from H. gordonii products. The test samples were spotted in duplicate next to each other and the four reference standard levels of P57 (65.50, 75.00, 100.00, 125.00 ng) were spread over the plate. FIGURE 3.19: An example of an HPTLC chromatogram that was used for the quantification of P57 from H. gordonii commercial products. Track 1: 65.50 ng P57, Track 2-3: Product 1, Track 4: 75.00 ng P57, Track 5-6: Product 2, Track 7: 100.00 ng P57, Track 8-9: Product 3, Track 10: 125.00 ng P57, Track 11-12: Product 4 Figure 3.20 is an example of the baseline graph for one track and the corresponding integrated peak where P57 is identified. 96 FIGURE 3.20: An example of the baseline chromatogram for one track (A) and the corresponding integration of the P57 peak as indicated (B) Figures 3.21 and 3.22 show the calibration curves for four standard levels of P57 and the four H. gordonii products in duplicate in terms of graph height and area, respectively. FIGURE 3.21: The linear calibration curve for four standard levels of P57 (black) and the four H. gordonii products (green) in duplicate in terms of peak height. 97 FIGURE 3.22: The linear calibration curve for four standard levels of P57 (black) and the four H. gordonii products (green) in duplicate in terms of peak area The calibration curve of the peak area at 366 nm (y) versus concentration (x) was found to be linear in the concentration range of 65.50-125.00 ng for P57 with good correlation coefficients of 0.9706-0.9992 based on peak height and 0.9720-0.9993 based on peak area. From Table 3.2 it can be seen that the amount of P57 detected for the same sample varied throughout the day (n=3) for product 1 from 57.19-67.33 ng, product 2 from 77.47-92.10 ng, product 3 from 62.86-74.72 ng and product 4 from 60.1470.94 ng. The results for the first test of the day were consistently higher than for the other two testing times. Table 3.3 displays the inter-day precision results for the duplicates of all the test samples. The amount of P57 detected for the same sample varied across the two days for product 1 from 63.19-67.37 ng, product 2 from 65.64-74.72 ng, product 3 from 79.27-92.10 ng and product 4 from 70.4493.56 ng. There was a consistent correlation between height and area for all the test samples on same day as well as on consecutive day testing. 98 TABLE 3.2: Peak height and area comparison for the calculation of intra-day precision Product Duplicate Height (ng) Area (ng) Average of height and area (ng) Average (total) (ng) Time 1 1 2 3 4 1 70.14 72.05 71.01 2 62.96 64.32 63.64 1 93.11 95.02 94.07 2 89.50 90.76 90.13 1 75.43 76.87 76.15 2 72.71 73.86 73.29 1 72.04 74.45 73.25 2 67.04 70.19 68.62 67.33 92.10 74.72 70.94 Time 2 1 2 3 4 1 57.17 57.66 57.42 2 56.98 56.93 56.96 1 80.14 78.52 79.33 2 76.87 74.94 75.91 1 64.77 64.01 64.39 2 61.36 61.30 61.33 1 62.47 64.34 63.41 2 55.84 57.87 56.86 57.19 77.62 62.86 60.14 Time 3 1 2 3 4 1 60.55 62.03 61.29 2 54.89 56.26 55.58 1 79.06 80.61 79.84 2 74.55 75.63 75.09 1 68.02 69.89 68.96 2 67.53 69.83 68.68 1 67.96 69.28 68.62 2 61.51 62.86 62.19 99 58.44 77.47 68.82 65.41 TABLE 3.3: Peak height and area comparison for the calculation of inter-day precision Product Duplicate Height (ng) Area (ng) Average of height and area (ng) Average (total) (ng) Day 1 1 2 3 4 1 67.17 70.72 68.95 2 56.32 58.51 57.42 1 64.55 66.86 65.71 2 64.59 66.52 65.56 1 76.45 82.11 79.28 2 74.14 78.38 76.26 1 95.56 104.95 100.26 2 83.68 90.03 86.86 63.19 65.64 79.27 93.56 Day 2 1 2 3 4 1 70.14 72.05 71.10 2 62.96 64.32 63.64 1 75.43 76.87 76.15 2 72.71 73.86 73.29 1 93.11 95.02 94.07 2 89.50 90.76 90.13 1 72.04 74.45 72.25 2 67.04 70.19 68.62 67.37 74.72 92.10 70.44 The average percentage recovery at four different levels of P57 was good at 95.32-107.52 % on day one and 98.29-101.99% on day two (Table 3.4). 3.3.5 CONCLUSIONS The proposed quantitative HPTLC method for the estimation of P57 from commercially available H. gordonii products was found to be accurate, precise, specific and reproducible with good correlation coefficients and a good correlation between height and area for the P57 estimations. A major advantage of quantification by HPTLC as opposed to LC-MS/HPLC is the ability to analyse several samples at one time with a small quantity of mobile phase thereby reducing the time and cost of analysis. 100 TABLE 3.4: Accuracy of the P57 standards determined by recovery analysis Standard Amount spotted (ng) Amount detected (ng) Average of height and area (ng) Recovery (%) Day 1 1 65.50 65.85 (height) 67.54 (area) 66.70 101.82 2 75.00 76.92 (height) 76.47 (area) 76.70 102.26 3 100.00 96.16 (height) 94.48 (area) 95.32 95.32 4 125.00 132.12 (height) 136.67 (area) 134.40 107.52 Day 2 1 65.50 62.73 (height) 66.02 (area) 64.38 98.29 2 75.00 77.13 (height) 75.84 (area) 76.49 101.99 3 100.00 100.12 (height) 97.78 (area) 98.95 98.95 4 125.00 123.83 (height) 129.04 (area) 126.44 101.15 101 3.4 REFERENCES ADAMS, M. 2006. Hoodia fraud: Counterfeit Hoodia gordonii weight loss pills dominate the market. Natural news.com. Available from: http://www.naturalnews.com/018677.html [Accessed: 03/06/2009]. AVULA, B., WANG, Y.H., PAWAR, R.S., SHUKLA, Y.J., SCHANEBERG, B., KHAN, I.A. 2006. Determination of the appetite suppressant P57 in Hoodia gordonii plant extracts and dietary supplements by liquid chromatography /electrospray ionization mass spectrometry (LC-MSD-TOF) and LC-UV methods. Journal of AOAC International, 89(3):606-611. AVULA, B., WANG, Y.H., PAWAR, R.S., SHUKLA, Y.J., KHAN, I.A. 2007. Chemical fingerprinting of Hoodia species and related genera: Chemical analysis of oxypregnane glycosides using High-performance liquid chromatography with UV detection in Hoodia gordonii. Journal of AOAC International, 90(6):1526-1531. AVULA, B., WANG, Y.-H., PAWAR, R.S., SHUKLA, Y.J., SMILLIE, T.J., KHAN, I.A. 2008. A rapid method for chemical fingerprint analysis of Hoodia species, related genera, and dietary supplements using UPLC-UV-MS. Journal of Pharmaceutical and Biomedical Analysis, 48:722-731. BAZYLKO, A., STRZELECKA, H. 2007. A HPTLC densitometric determination of luteolin in Thymus vulgaris and its extracts. Fitoterapia, 78:931-395. GÜNTHER, M., SCHMIDT, P.C. 2005. Comparison between HPLC and HPTLCdensitometry for the determination of harpagoside from Harpagophytum procumbens CO2-extracts. Journal of Pharmaceutical and Biomedical Analysis, 37:817-821. JAITAK, V., GUPTA, A.P., KAUL, V.K., AHUJA, P.S. 2008. Validated highperformance thin-layer chromatography method for steviol glycosides in Stevia rebaudiana. Journal of Pharmaceutical and Biomedical Analysis, 47:790-794. JANSSEN, H.-G., SWINDELLS, C., GUNNING, P., WANG, W., GRÜN, C., MAHABIR, K., MAHARAJ, V.J., APPS, P. J. 2008. Quantification of appetite suppressing steroid glycosides from Hoodia gordonii in dried plant material, purified extracts and food products using HPLC-UV and HPLC-MS methods. Analytica Chimica Acta, 617(1-2):200-207. LAZAROWYCH, N.J., PEKOS, P. 1998. Use of fingerprinting and marker compounds for identification and standardisation of botanical drugs: strategies for applying pharmaceutical HPLC analysis to herbal products. Drug Information Journal, 32:497-512. MARSTON, A. 2007. Role of advances in chromatographic techniques in phytochemistry. Phytochemistry, 68:2785-2797. PATHAK, S.B., NIRANJAN, K., PADH, H., RAJANI, M. 2004. TLC densitometric method for the quantification of eugenol and gallic acid in clove. Chromatographia, 60(3/4):241 102 SCHIPPMANN, U., LEAMAN, D., CUNNINGHAM, A.B. 2006. A comparison of cultivation and wild collection of medicinal and aromatic plants under sustainability aspects. In: BOGERS, R.J., CRAKER, L.E., LANGE, D. (eds.). Medicinal and aromatic plants. The Netherlands: Springer:75-79. REICH, E., SCHIBLI, A. 2006. High-performance thin-layer chromatography for the analysis of medicinal plants. New York: Thieme. RUMALLA, C.S., AVULA, B., SHUKLA, Y.J., WANG, Y.H., PAWAR, R.S., SMILLIE, R.J., KHAN, I.A. 2008. Chemical fingerprint of Hoodia species, dietary supplements, and related genera by using HPTLC. Journal of Separation Science, 31(22):3959-3964. VAN HEERDEN, F.R., VLEGGAAR, R., HORAK, R.M., LEARMONTH, R.A., MAHARAJ, V., WHITTAL, R.D. 1998. Pharmaceutical compositions having appetite suppressant activity. PCT/GB98/01100. WAGNER, H., BLADT, S. 1996. Plant Drug Analysis: A thin layer chromatography atlas 2nd ed. Heidelberg: Springer-Verlag. WIDMER, V., REICH, E., DEBATT, A. 2008. Validated HPTLC method for identification of Hoodia gordonii. JPC – Journal of Planar Chromatography – Modern TLC, 21:21-26. WOLFENDER, J.-L., RODRIGUEZ, S., HOSTETTMANN, K. 1998. Liquid chromatography coupled to mass spectrometry and nuclear magnetic resonance spectroscopy for the screening of plant constituents. Journal of Chromatography A, 794:299-316. WOLFENDER, J.-L., NDJOKO, K., HOSTETTMANN, K. 2003. Liquid chromatography with ultraviolet absorbance-mass spectrometric detection and with nuclear magnetic resonance spectroscopy: a powerful combination for the online structural investigation of plant metabolites. Journal of Chromatography A, 1000:437-455. 103 CHAPTER 4 Vibrational spectroscopy 4.1 INTRODUCTION Despite the discovery of new potentially appetite-suppressing compounds from H. gordonii, the quality and therefore the purchase of raw material by manufacturers from H. gordonii growers is still determined by quantifying P57. Traditionally, liquid-chromatography coupled to mass spectrometry (LC-MS) is used to determine the P57 content of raw material. LC-MS is cumbersome, requires considerable expertise, is expensive and requires extensive sample preparation. Vibrational spectroscopy techniques, including mid-infrared (MIR), near-infrared (NIR) and Raman, offer an attractive alternative to traditional analytical methods due to rapid analysis time, little or no pretreatment of samples, reduction in the use of chemicals and the possibility to evaluate several components simultaneously in different physical states (solids, liquids, gases) and perhaps most importantly, it is non-destructive in nature. The most commonly used technique in terms of number of publications has been NIR, despite its lower level of spectral information, followed by MIR and Raman spectroscopy (Moros, Garrigues and De La Guardia, 2010:578-579). NIR and MIR spectroscopy has been used with success in the food industry for quantifying various components such as peptides in cheeses (González-Martín et al., 2009:1564) adulterants in Mexican honeys (GallardoVelázquez et al., 2009:313), caffeine in green tea powder and granules (Sinija and Mishra, 2009:998), the acid value of peanut oil (Rao et al., 2009:249) and aflatoxin B1 in chilli powder (Tripathi and Mishra, 2009:840). It has also found application as an analytical technique in the agricultural (Camps and Christen, 2009:1125; Xie, Ying and Ying, 2009:34), petrochemical (Balabin and Safieva, 2008:2745; Zhan et al., 2008:165) and pharmaceutical (Qu et al., 2008:1146; Mantanus et al., 2009:186) industries amongst others. In combination with chemometric analysis, Raman spectroscopy has found application in various diverse industries such as; the medical industry to identify liver cirrhosis, differences between benign and malignant tumours and to detect teeth caries; the pharmaceutical industry for quantification of polymorphs in powder mixtures; in botany as a tool in taxonomy, 104 to optimise breeding and cultivation and for quality control of plant products like essential oils; in the food and textile industries for quality control; and to analyse works of art and gemstones for characterisation, restoration, conservation, dating and authentication (Schrader et al., 2000:35; Kachrimanis, Braun and Griesser, 2007:407; Braun et al., 2010:29; Moros, Garrigues and De La Guardia, 2010:580). Raman spectroscopy combined with micro-equipment can provide detailed molecular information with high spatial resolution at the cellular level. For example, mapping has been used to show the spatial distribution of certain components or active ingredients in plants and pharmaceutical products (Baranska et al., 2005:1; Zhang, Henson and Sekulick, 2005:262; Baranska and Proniewicz, 2008:153). Thus, it can be used to determine where a specific component is localised in a plant. Raman mapping as well has the advantages that it is non-destructive, does not require laborious sample preparation and provides extensive data about the sample. The large amount of information in the form of spectral data (multivariate data matrices) that is generated with vibrational spectroscopy techniques requires the use of chemometric analysis for interpretation. Regression methods like the partial least squares projections to latent structures (PLS) algorithm are used to link the spectra to quantifiable properties of samples, in this instance P57. Mathematical pre-processing techniques such as orthogonal signal correction (OSC) and multiplicative signal correction (MSC) that are applied to data decrease interferences thereby enhancing the interpretation of the data (Moros, Garrigues and De La Guardia, 2010:580). The aim of this study was to develop calibration models using NIR, MIR and Raman spectroscopy in combination with chemometric data processing which may potentially be used for simpler, less expensive and faster quantification of P57 in raw material. In addition, Raman mapping was investigated to determine where P57 is localised within the stem. 4.2 BACKGROUND AND LITERATURE REVIEW 4.2.1 Infrared spectroscopy (IR) Infrared spectroscopy techniques are advantageous because the cover a wide wavenumber interval and is customarily divided into three wavelength regions depending on the type of instrumentation used to acquire the spectral information. NIR covers wavelengths 750-2500 nm or 13333-400 cm-1, MIR 2500-25000 nm or 105 4000-400cm-1 and far infrared 25-1000 µm or 400-10 cm-1. However, as some overlap exists, a new spectral region known as short-wavelength near-infrared (SW-NIR) (600-1000 nm) has been introduced which includes part of the visible light region (700-750 nm) (Moros, Garrigues and De La Gaurdia, 2010:578; Lin et al., 2009:119-120). A typical NIR instrument consists of a radiation source, a wavelength selection device, a sample holder, a photoelectric detector that measures the intensity of detected light and converts it to electrical signals and a computer system for acquisition and processing of the spectral data (Lin et al., 2009:122). Radiation sources used in analytical spectroscopy are divided into continuum sources and line sources. Continuum sources emit light with a relatively stable and continuous intensity over a wide range of wavelengths and include tungsten filament lamps, deuterium tungsten light, high pressure mercury or xenon arc lamps and heated solid ceramics or heated wires. The tungsten filament lamp and deuterium tungsten light are used for the NIR region while heated solid ceramics or heater wires are used for the MIR region. NIR light provided by the radiation source can penetrate much further into samples than MIR radiation, 13 mm versus a couple of micrometers, therefore samples may be scanned within containers that are transparent to NIR light. MIR light on the other hand, does not penetrate through these materials but it provides better specificity (Lin et al., 2009:139-141). Line sources emit light at only a few discrete wavelengths and the intensity varies at different wavelengths. Examples include hollow cathode lamps, electrodeless discharge lamps, light-emitting diodes (LED), sodium or mercury vapour lamps and lasers. Lasers emit monochromatic radiation with high intensity and are used in Raman spectrometers (Lin et al., 2009:123-124). Several different radiation sources of NIR and MIR instruments are classified according to the mechanism of light separation into filter-based, monochromator-based, and Fourier transform (FT) instruments. FTIR has established itself as one of the most powerful techniques for chemical analysis. Most modern FT spectrometers have an optical device called a Michelson interferometer that allows the controlled generation of interference patters or interferograms. Fourier transformation (FT), a mathematical procedure, is applied to interferograms to obtain spectra of the 106 scanned samples (Lin et al., 2009:125-126; Subramanian and Rodriguez-Saona, 2009:146, 155). The interpretation of spectra obtained from IR spectrometers is based upon the overtones and vibrations of the atoms of a molecule when passing IR radiation through a sample. The energy at which any peak in an absorption spectrum appears corresponds to the frequency of a vibration of a part of the sample molecule. Therefore a range of functional groups can be identified according to absorption peaks at specific wavelengths or matched with characteristic vibrations (Lin et al., 2009:139). Overtone peaks in NIR spectra are mainly due to O-H, C-H, S-H, and N-H stretching modes and combination bands which are formed when two or more vibrations combine through addition and subtractions of energies to form a single band. NIR spectra are complex and overlapping peaks and large baseline variations may present difficulties in the interpretation highlighting the need for processing methods such as derivatisation to improve spectral characteristics (Subramanian and Rodriguez-Saona, 2009:161). With MIR spectroscopy, a chemical profile of the sample is obtained by monitoring the fundamental vibrational and rotational stretching of molecules. It provides information useful to analyse the composition of a sample as well as the structure of chemical molecules. Although MIR spectra are easier to interpret than NIR spectra, chemometric analysis may still be required to draw meaningful information from the spectral data (Subramanian and Rodriguez-Saona, 2009:164165). 4.2.2 Chemometric data analysis Chemometrics is used to extract useful information efficiently from spectra and is defined by the International Chemometrics Society (IC) as the science of relating measurements made on a chemical system to the state of the system via the application of mathematical or statistical methods (Moros, Garrigues and De La Guardia, 2010:580). The power of chemometrics is that it can be used to model systems that are both largely unknown and complex such as spectral data. Chemometrics has a holistic, exploratory approach to data and process modelling. No assumptions are made about the data prior to analysis because in chemometrics the process or data structures reveal the relations themselves 107 (Swanepoel and Esbensen, 2007:S.a.). Data from vibrational spectrometers combined with chemometrics provides calibration models and classification tools. Classification methods, also called discriminant analysis, places samples into different classes through pattern recognition while regression methods are used to construct calibration models as it links the spectra to quantifiable properties of the samples. Various pre-processing methods (mathematical pretreatments), such as derivatives and multiplicative corrections, are used to remove spectral variation unrelated to sample components from measured spectra to enhance spectral information. This is considered a critical step before calibration models are constructed as this can significantly enhance the prediction accuracy of the model. In addition, selection of spectral regions may enhance predictive ability as many wavenumbers in a spectrum are not relevant to the sample (Moros, Garrigues and De La Guardia, 2010:580-582). The combination of NIR, MIR and Raman with chemometric analysis provides fast tools for multi-component analysis of various samples as an attractive alternative to the tedious, time-consuming traditional analytical methods (Moros, Garrigues and De La Guardia, 2010:590). 4.3 MATERIALS AND METHODS 4.3.1 Plant material Natural population (wild) and cultivated site samples of H. gordonii stems (n = 145) were hand-picked throughout South Africa and Namibia, sliced and air-dried. Random sampling was performed and all possible aspects of variation were considered (e.g. young vs mature plants). To construct a good predictive calibration model, samples must be selected for their variability rather than for quantity (Moros, Garrigues and De La Guardia, 2010:582). H. gordonii plants were harvested at different levels of maturity according to market need, thus it was important to include this variation. The dried plant material was powdered with a Retsch® 400 MM ball mill at a frequency of 30.0 Hz for 90 s and sieved (500 µm) to control the particle size distribution. For the Raman mapping experiment an intact, fresh H. gordonii stem was sliced transversely to a thickness of 6 mm. For LC-MS confirmation, the outer layers and cortex of 5 stems from different plants were separated with a knife and processed as above. The P57 reference standard was purchased from ChromaDexTM Inc (California, USA). 108 4.3.2 Near infrared spectroscopy (NIR) The powdered H. gordonii samples were placed in Chromacol® glass vials. Near infrared spectra of the powdered samples were collected from 10 000-4000 cm-1 on a Büchi® (Flawil, Switzerland) NIRFlex N500 Fourier transform near infrared (FT-NIR) spectrophotometer (Figure 4.1) with NIRWare® (version 1.2.3000 advanced edition) software in reflectance mode. Spectra were collected at a spectral resolution of 4 cm-1 with 32 scans per sample. FIGURE 4.1: Photograph of the Buchi NIRFlex N500 Fourier transform near infrared spectrophotometer and sample tray (Photograph by I Vermaak) 4.3.3 Mid-infrared spectroscopy (MIR) FT-MIR spectra were generated with a Bruker® (Ettlingen, Germany) Alpha-P Fourier transform infrared spectrophotometer (Figure 4.2) in absorbance mode and recorded with OPUS® (version 6.5) software. Spectra were collected from 4000-375 cm-1 at intervals of 4 cm-1 and the data for each spectrum was the average spectrum of 32 scans. The spectra for each sample were collected three times and the mean of the three spectra was used in further analysis steps. A blank background measurement was recorded after every three samples scanned. 109 FIGURE 4.2: Photographs of the Bruker Alpha-P Fourier transform infrared spectrophotometer (Photographs by I Vermaak) 4.3.4 Raman spectroscopy A Bruker® (Ettlinger, Germany) Multiram Fourier transform Raman spectrometer (Figure 4.3) was used to generate Raman spectra of the powders in the range of 4000-100 cm-1, with a spectral resolution of 4 cm-1 set to 128 scans per sample. The spectrometer was equipped with a diode-pumped Nd:YAG laser set to 30 mW, emitting at 1064 nm, and a germanium detector cooled with liquid nitrogen. The Raman spectra were recorded and the instrument controlled with OPUS® (version 6.5) software. Spectral measurements performed on the powder were collected in duplicate and averaged. FIGURE 4.3: Photographs of the Bruker Multiram Fourier transform Raman spectrometer (Photographs by I Vermaak) 110 4.3.5 Raman mapping Two-dimensional mapping of fresh H. gordonii plant material was done with the same spectrometer as above and settings as in 4.3.4 using a 6 mm cross-section of the stem. The stem sample was mapped at an area of 18.5 mm x 12.5 mm with a spatial resolution of 570 µm. A focused laser beam at a power of 50 mW with a diameter of ca. 50 µm was used and 128 scans were collected at each point. 4.3.6 Data analysis The spectral data was exported to Microsoft® Excel (2007) to enable analysis using Simca-P® 11.0 (NIR, MIR) and Simca-P+® 12.0 (Raman) chemometric analysis software. Criteria such as the correlation coefficient (R2), the number of partial least squares projections to latent structures (PLS) factors and the root mean square error of prediction (RMSEP) were used to assess the ability of these models to accurately predict the chosen quality parameter (Mantanus et al., 2009:1750). The spectral data were analysed using partial least squares projections to latent structures (PLS) and/or orthogonal projections to latent structures (O-PLS) regression methods and various mathematical pre-processing methods were investigated to develop a mathematical calibration model (Thermo Scientific, 2009). extremely Partial least squares regression (PLS-R) represents an successful new extension of traditional statistical regression methodology, developed over the last 35 years within the discipline of chemometrics. In PLS the independent data (X) is related to the dependent data (Y) and PLS attempts to capture variance and achieve correlation between X and Y-data. O-PLS is based on PLS but includes an orthogonalisation step. Strong systematic variance present in X may be unrelated to Y and is therefore irrelevant for the prediction of Y. Removing the information in X which is orthogonal to Y, termed structured noise, will not influence the variation between X and Y. Therefore the interpretation of the model is improved while the correlation remains the same (Swanepoel and Esbensen, 2007:S.a.). A work set and training was randomly chosen by the software. The workset consisted of 30% of the Y-values (P57 concentration determined by LC-MS) while the other 70% was used as the training set. Due to this randomised approach there will be slight differences in the results but this gives an indication of the robustness of the model if the results are fairly consistent. Several repetitions were done for each pre-processing method to 111 determine whether results remained consistent and differences were not significant. The training and work sets remained constant for all the pre- processing methods investigated. 4.3.7 P57 quantification with LC-MS analysis Extracts of the H. gordonii stem samples (0.2 g) were prepared in an ultrasonic bath (Sonorex digital 10p) for 2 h with 4 ml of 50% acetonitrile and 0.1% formic acid. The detection and quantification of P57 in was performed using a Waters (MA, USA) API Q-TOF Ultima system equipped with a BEH C18 column (1.7µm, 2.1 x 50 mm). The MS detector was operated in electron impact mode with the capillary voltage at 3.5 kV and the cone voltage at 35 kV (positive switching - ES+). The flow rate of the HPLC was 0.4 ml/min, the cone gas flow rate was 50 l/h and the desolvation gas flow rate was 350 l/h. The source temperature was 100 °C and the desolvation temperature 350 °C. Accurate mass (861.5 m/z) was used to quantify P57 in ES+ mode within a mass range of 200 to 1800 m/z. The injection volume was 2 µl. Gradient elution was used with 0.1% formic acid in water (Solvent A) and acetonitrile (Solvent B) as mobile phases. The gradient started with 100% solvent A, changed to 20% solvent B over 0.5 min, to 100% solvent B over 11.5 min, followed by a 2 min isocratic step and a return to initial conditions for 1 min for a total run time of 15 min. The LC-MS data was analysed using Masslynx® (version 4.2) software. 4.4 RESULTS AND DISCUSSION 4.4.1 Chemometric data analysis methods and parameters Many different calibration models were generated with the sample set (n=145) according to the method of pre-processing used and are shown in Table 4.1 together with the most important parameters used to evaluate these models. It is important to avoid overfitting by choosing a model with a small number of PLS factors, which is an indication of the fit of the model. The correlation coefficient (R2) is a measure of how well the model fits the data. An R2 value close to one is a necessary condition for a good model, but even models with a good R2 can suffer from poor prediction ability. A poor R2 is found when there is poor reproducibility (a lot of noise) in the training data set, or (with PLS) when for any other reason X does not explain Y. RMSEP is the root mean square error of 112 prediction for observations in the prediction set. The Q2 value is an indication of the predictive ability of the model and is considered good when greater than 0.5. There are various causes for a poor Q2 such as noisy data, when the relationship of X to Y is poor (with PLS), or when the model is dominated by a few scattered outliers in the training set (SIMCA-P+ analysis advisor, 2008:S.a.). TABLE 4.1: Number of PLS factors, R2, Q2, RMSEP and RMSEE values for determining P57 content with NIR, MIR and Raman spectroscopy utilising a PLS model with different pre-processing methods Pretreatment method None 1st derivative 2nd derivative MSC SNV OSC NIR spectroscopy PLS factors 1 3 6 3 4 2 R2 0.2021 0.5716 0.9629 0.2497 0.3424 0.7074 RMSEP 0.10 0.09 0.03 0.13 0.11 0.06 MIR spectroscopy PLS factors 3 3 1 2 2 1 R2 0.2158 0.5918 0.2285 0.4997 0.4522 0.6630 RMSEP 0.07 0.06 0.10 0.07 0.09 0.04 Raman spectroscopy PLS factors 4 3 2 4 5 2 R2 0.3486 0.6851 0.7253 0.3856 0.5324 0.9986 RMSEP 0.054 0.052 0.056 0.062 0.062 0.004 MSC – mulitiplicative signal correction; SNV – standard normal variate; OSC – orthogonal signal correction; R2 – correlation coefficient; RMSEP – root mean square error of prediction. 4.4.2 NIR spectroscopy The general profile of the FT-NIR spectra of H. gordonii powdered raw material (A) and the P57 reference standard (B) is depicted in Figure 4.4. From visual inspection the wavenumbers 10000-7800 cm-1 does not contribute to the model as evident from the low absorbance values. Between 5000-4000 cm-1 strong peaks from combination bands are present in the P57 reference standard spectrum (B). These peaks were also present in the H. gordonii powdered raw material spectra (A) but overlapping of peaks occur due to “masking” by other components. CH2 113 and CH3 absorption peaks were observed between 7000-5000 cm-1. Many different calibration models can be generated according to the method of preprocessing used. The best calibration model was constructed using the 2nd derivative pre-processing method and this model was further developed. FIGURE 4.4: Typical FT-NIR spectra for H. gordonii powdered raw material (A) and the spectrum for the P57 reference standard (B) The T vs U plot, describing the intercorrelation between the X and Y illustrates the potential of a correlation that exists between these two variables. About 42% of the variance of X is explained by the PLS model after the exclusion of X-variables with low loading weights (wavenumbers 10000-7800 cm-1) as indicated on the loadings line plot before and after exclusion (Figure 4.5). This is a considerable improvement as before exclusion only 24% of the variance of X was explained. After construction of the PLS model six outliers were removed which improved the accuracy of prediction or the R2 value. 114 FIGURE 4.5: Loadings line plot from the PLS calibration model using FT-NIR data before (A) and after (B) removal of X variables with low loading weights (100007800 cm-1) The PLS model had an R2 value of 0.9629, RMSEP of 0.03% and 6 PLS factors (Figure 4.6). The cumulative overall cross-validated R2X (Q2) for this model was 75.6%, indicating a good prediction ability of this model. An OPLS model separates the PLS components into two groups, those related to Y (predictive) and those orthogonal to Y. The Y-orthogonal component may be part of the model, but is not useful for the prediction of Y. The main benefit of OPLS is model interpretation, while maintaining the predictive ability of the PLS method. As for the PLS model, the OPLS model had an R2 value of 0.9629, RMSEP of 0.03% and 6 PLS factors and in addition the T vs U plot indicated good correlation of 0.9848 between the observed values (obtained from LC-MS measurements) and predicted values for P57 (obtained from the spectra by FT-NIR measurements). The cumulative overall cross-validated R2X (Q2) for the O-PLS model was 56.5%. 115 This indicated that the prediction ability of the OPLS model is not as good as the prediction ability of the PLS model as some of the X-variables removed are in fact not orthogonal to the first component. However, the O-PLS model was only used for interpretation and the PLS model was used as the final calibration model. These results indicate that FT-NIR may be used for rapid determination of the concentration of P57 in H. gordonii raw plant materials. Considering the high cost of other analytical methods, the small range of error of prediction (RMSEP) using the FT-NIR method is negligible. FIGURE 4.6: PLS calibration model using the 2nd derivative pre-processing technique calculated from FT-NIR data 4.4.3 MIR spectroscopy Chemometric analysis of the FT-MIR data did not reveal promising results and even with the removal of several outliers the predictive ability of the model did not improve appreciably. Several pre-processing methods were applied to the FT-MIR data with little success as can be seen in Table 4.1 where the highest R2 value (i.e. the best correlation between LC-MS values and predicted FT-MIR values) was 0.6630. This implies that the accuracy of predicting P57 in H. gordonii plant material will be very low and this method was consequently not further investigated. 116 4.4.4 Raman spectroscopy The results for the model assessment criteria for several models after the application of various pre-processing techniques (Table 4.1) show that the best model was constructed after applying the OSC pre-processing method to the data. OSC is used to correct the X data matrix (spectral data) by removing information that is linearly unrelated or orthogonal to the response matrix Y of the calibration model. It is widely applied for the correction of excessive background noise of spectral data collected with several analytical techniques such as NIR and nuclear magnetic resonance (NMR) (Blanco et al., 2001:126; Qu, Ou and Cheng, 2005:838). It has been used successfully to develop calibration models for plant extracts and whole soybean protein with NIR spectra, and in a pharmaceutical setting for D-mannitol and paracetamol polymorphs with FT-Raman spectroscopy (Qu, Ou and Cheng., 2005:838; Kachrimanis, Braun and Griesser, 2007:407; Igne et al., 2009:57; Braun et al., 2010:29). Two principal components were used to describe the OSC calibration model with 66.2% of the variation in X explained by PLS factor 1 and 32.3% by PLS factor 2 (R2 X (cum) = 0.985). The predictive ability of the model is 99.9 % as indicated by the overall cross-validated R2X (Q2) value of 0.999. The calibration model shows good correlation between the observed values (obtained from LC-MS measurements) and predicted values for P57 (obtained from the spectra by FT-Raman measurements) with a regression of 0.9986 indicating high accuracy of this model (Figure 4.7). The difference between R2 and Q2 was less than 0.2 and the RMSEP value for this model was 0.004, indicators of a good model. Internal validation by response permutation (100) confirmed the validity of the model as Q2 intercepts the Y-axis below zero and R2 and Q2 are closely correlated. In addition, this indicated that noise is not being modeled and the model is not overfitted. No outliers were removed for any of the models and internal validation and overfitting of the model was assessed with 100 permutations. These results clearly demonstrate the feasibility of FT-Raman spectroscopy for rapidly determining the concentration of P57 in H. gordonii raw plant materials with a negligible error of prediction. 117 FIGURE 4.7: PLS calibration model with OSC pre-processing calculated from FTNIR-Raman data 4.4.5 Raman mapping Raman mapping was performed to determine where the perceived active molecule P57 is localised in the plant. Spectra collected from the plant show some diversity corresponding to the tissue and its various chemical compositions. Generally it is possible to obtain maps, showing the distribution of individual components, based on the integration of characteristic Raman bands. However, because of the low concentration of P57 in the plant material this procedure could not be applied in this case since in the measured spectra signals characteristic of P57 were hardly visible. For this reason chemometrics was used to analyse Raman mapping data. Application of cluster analysis in the spectral range from 1500 to 1800 cm-1 together with pre-processing of vector normalisation allowed grouping the spectra into three clusters (Figure 4.8 B,C). This wavenumber range was chosen since it covers the most specific bands of P57. A single spectrum extracted from the three clusters shown in Figure 4.9 demonstrates several bands which can be assigned to individual plant components. An outer part of the slice contains mainly carotenoids (violet ring) and additionally lignin (green spots), whereas the cortex is composed of cellulose (blue center). Spectra extracted from the inner part of the slice also shows bands with weak intensity which can be assigned to P57. 118 FIGURE 4.8: A microscopic image of a fresh H. gordonii slice with the marked area where the Raman mapping measurement was performed (A), and chemical maps of the clusters (B) and the borders of the clusters (C) showing the distribution of various plant components (blue – cellulose; violet – carotenoids; green – lignin) lignin Raman Intensity carotenoids cellulose H.gordonii P57 H.gordonii H.gordonii P57 standard 3500 3000 2500 2000 1500 Wavenumber / cm-1 1000 500 FIGURE 4.9: A single Raman spectrum of H. gordonii extracted from the three clusters coloured accordingly. At the bottom a Raman spectrum of pure P57 standard is presented The distribution is not homogenous and P57 is mainly localised in the inner fleshy circle of the plant. This was confirmed with LC-MS of several samples where the 119 outer layers and cortex of the stem were analysed separately (Table 4.2). In effect the inclusion of the outer layers with the thorns of the plant containing very small amounts of P57 causes dilution of the final P57 content of the raw material. TABLE 4.2. P57 content of the outer layers and cortex samples of H. gordonii determined by LC-MS analysis Sample Outer layers (% P57) Cortex (% P57) 1 0.000 0.031 2 0.017 0.074 3 0.023 0.165 4 0.051 0.173 5 0.021 0.084 4.5 CONCLUSIONS The quality of raw plant material is determined to a large extent by the content of P57. The results show that there was a high variation of P57 concentration ranging from 0.000-0.430% P57. The unwritten industry standard of good quality H. gordonii raw plant material is considered to be more than 0.300% P57 content. A mere 4% of the 145 plant samples collected and analysed contain more than this limit of P57 content, all from natural populations. If a 20% deviation is allowed, as per the industry standard of 0.300% P57 content, raw plant material containing more than 0.240% P57 would be acceptable and only 12% of the plant samples collected attained this standard. In addition, it is evident that this quality standard may be unrealistic in cultivated H. gordonii raw material as 9% of the 12% samples that comply with this standard were collected from natural populations in contrast to only 3% that was collected from cultivated site samples. In general the concentration of P57 is much lower in cultivated site samples compared to natural population samples. It is known that active ingredient levels can be higher in wild plant populations compared to cultivated stock, usually due to differences in growth rates. In addition, the conditions under which plants are cultivated are different compared to natural populations which may also account for these differences (Schippmann, Leaman and Cunningham, 2006:81). Raman mapping was used to observe the spatial distribution of P57 within the plant. This 120 information may be used to only include the part of the plant containing the highest concentration of P57 in the final raw material. By preventing dilution, the final P57 content of cultivated raw material may be increased. However, removing these parts may not be viable in the industry as this would considerably lengthen the processing procedures. Both FT-NIR and FT-Raman spectroscopy demonstrates potential as alternative methods to LC-MS to rapidly quantify P57 in H. gordonii raw material with high accuracy and a small error of prediction. 121 4.6 REFERENCES BALABIN, R.M., SAFIEVA, R.Z. 2008. Motor oil classification by base stock and viscosity based on near infrared (NIR) spectroscopy data. Fuel, 87(12):2745-2752. BARANSKA, M., PRONIEWICZ, L.M. 2008. Raman mapping of caffeine alkaloid. Vibrational Spectroscopy, 48:153-157. BARANSKA, M., SCHULZ, H., SIUDA, R., STREHLE, M.A., RÖSCH, P., POPP, J., JOUBERT, E., MANLEY, M. 2005. Quality control of Harpagophytum procumbens and its related phytopharmaceutical products by means of NIR-FTRaman spectroscopy. Biopolymers, 77:1-8. BLANCO, M., COELLO, J., MONTOLIU, I., ROMERO, M.A. 2001. Orthogonal signal correction in near infrared calibration. Analytica Chimica Acta, 434:125-132. BRAUN, D.E., MAAS, S.G., ZENCIRCI, N., LANGES, C., URBANETZ, N.A., GRIESSER, U.J. 2010. Simultaneous quantitative analysis of ternary mixtures of D-mannitol polymorphs by FT-Raman spectroscopy and multivariate calibration models. International Journal of Pharmaceutics, 385:29-36. CAMPS, C., CHRISTEN, D. 2009. Non-destructive assessment of apricot fruit quality by portable visible-near infrared spectroscopy. LWT – Food Science and Technology, 42(6):1125-1131. GALLARDO-VELÁZQUEZ, T., OSORIO-REVILLA, G., ZUÑIGA-DE LOA, M., RIVERA-ESPINOZA. 2009. Application of FTIR-HATR spectroscopy and multivariate analysis to the quantification of adulterants in Mexican honeys. Food Research International, 42:313-318. GONZÁLEZ-MARTÍN, I., HERNÁNDEZ-HIERRO, J.M., VIVAR-QUINTANA, A., REVILLA, I., GONZÁLEZ-PÉREZ, C. 2009. The application of near infrared spectroscopy technology and a remote reflectance fibre-optic probe for the determination of peptides in cheeses (cow’s, ewe’s and goat’s) with different ripening times. Food Chemistry, 114:1564-1569. IGNE, B., ROGER, J.-M., ROUSSEL, S., BELLON-MAUREL, V., HURBURGH, C.R. 2009. Improving the transfer of near infrared prediction models by orthogonal methods. Chemometrics and Intelligent Laboratory Systems, 99:57-65. KACHRIMANIS, K., BRAUN, D.E., GRIESSER, U.J. 2007. Quantitative analysis of paracetamol polymorphs in powder mixtures by FT-Raman spectroscopy and PLS regression. Journal of Pharmaceutical and Biomedical Analysis, 43:407-412. LIN, M., RASCO, B.A., CAVINATO, A.G, AL-HOLY, M. 2009. Infrared (IR) spectroscopy – near-infrared spectroscopy and mid-infrared spectroscopy. In: SUN, D. (ed.). Infrared spectroscopy for food quality analysis and control. Massachusetts: Academic Press:119-143. 122 MANTANUS, J., ZIÉMONS, LEBRUN, P., ROZET, E., KLINKENBERG, R., STREEL, B., EVRARD R., HUBERT, PH. 2009. Moisture content determination of pharmaceutical pellets by near infrared spectroscopy: Method development and validation. Analytica Chimica Acta, 642(1-2):186-192. MOROS, J., GARRIGUES, S., DE LA GUARDIA, M. 2010. Vibrational spectroscopy provides a green tool for multi-component analysis. Trends in Analytical Chemistry, 29(7):578-591. QU, H.-B., OU, D.-L., CHENG, Y.-Y. 2005. Background correction in near-infrared spectra of plant extracts by orthogonal signal correction. Journal of Zhejiang University SCIENCE B, 6(8):838-843. QU, N., MINGCHAO, Z., MI, H., DOU, Y., REN, Y. 2008. Nondestructive determination of compound amoxicillin powder by NIR spectroscopy with the aid of chemometrics. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 70(5):1146-1151. RAO, Y., XIANG, B., ZHOU, X., WANG, Z., XIE, S., XU, J. 2009. Quantitative and qualitative determination of acid value of peanut oil using near-infrared spectrometry. Journal of Food Engineering, 93:249-252. SCHIPPMANN, U., LEAMAN, D., CUNNINGHAM, A.B. 2006. A comparison of cultivation and wild collection of medicinal and aromatic plants under sustainability aspects. In: BOGERS, R.J., CRAKER, L.E., LANGE, D. (eds.). Medicinal and aromatic plants. The Netherlands: Springer:75-95. SCHRADER, B., SCHULZ, H., ANDREEV, G.N., KLUMP, H.H., SAWATZKI, J. 2000. Non-destructive NIR-FT-Raman spectroscopy of plant and animal tissues, of food and works of art. Talanta, 53:35-45. SIMCA-P+ analysis advisor. 2008. SIMCA-P+ version 12.0.0. Sweden: Umetrics AB. SINIJA, V.R., MISHRA, H.N. 2009. FT-NIR spectroscopy for caffeine estimation in instant green tea powder and granules. LWT – Food Science and Technology, 42:998-1002. SUBRAMANIAN, A., RODRIGUEZ-SAONA, L. 2009. Fourier transform infrared (FTIR) spectroscopy. In: SUN, D. (ed.). Infrared spectroscopy for food quality analysis and control. Massachusetts: Academic Press:145-178. SWANEPOEL, M., ESBENSEN, K H. 2007. Representative sampling at the grape intake: A critical issue - often forgotten. 35th Technical Issue - Australian Grapegrower and Winemaker. THERMO SCIENTIFIC. 2009. http://www.thermo.com/com/cda/resources/ resources_detail/1,2166,13465,00.html [Accessed: 13/05/2009]. 123 TRIPATHI, S., MISHRA, H.N. 2009. A rapid FT-NIR method for estimation of aflatoxin B1 in red chilli powder. Food Control, 20:840-846. XIE, L., YING, Y., YING, T. 2009. Classification of tomatoes with different genotypes by visible and short-wave near-infrared spectroscopy with leastsquares support vector machines and other chemometrics. Journal of Food Engineering, 94(1):34-39. ZHAN, J., YIN, W., SHANG, H., LIU, C. 2008. In situ FT-IR spectroscopy investigations of carbon nanotubes supported Co-Mo catalysts for selective hydrodesulfurization of FCC gasoline. Journal of Natural Gas Chemistry, 17(2):165-170. ZHANG, L., HENSON, M.J., SEKULICK, S.S. 2005. Multivariate data analysis for raman imaging of a model pharmaceutical tablet. Analytica Chimica Acta, 545:262-278. 124 SECTION 2 Biopharmaceutical aspects In Part A, the issue of quality of H. gordonii materials and products was comprehensively discussed. If the authenticity of H. gordonii raw materials and products has been established then it is necessary to move towards meeting efficacy requirements. One of the first steps in the development of an effective product is to determine biopharmaceutical aspects related to the active constituents of H. gordonii such as P57. The most basic scientific research would be to establish that the purported active ingredient (P57) of H. gordonii is in fact absorbed to provide acceptable systemic bioavailability. Information in this area has been lacking for H. gordonii though several recent studies have addressed some of these aspects as discussed in the following chapter. This project aimed to contribute to and build upon current pharmacokinetic and biopharmaceutical knowledge of P57. 125 CHAPTER 5 Buccal and intestinal transport of P57 from Hoodia gordonii 5.1 INTRODUCTION As mentioned before, Hoodia gordonii was traditionally used by the San people of South Africa and Namibia to suppress hunger and thirst while on long hunting trips or in times of famine. The fresh stems were peeled to remove the thorns and the flesh was chewed and/or retained in the buccal cavity before swallowing (Van Wyk and Gericke, 2000:70; Lee and Balick, 2007:404-405; Glasl, 2009:300). Research on biopharmaceutical aspects that should be investigated prior to commercialisation of products claiming biological activity such as the bioavailability of the active compound has been lacking, though recently some of the biopharmaceutical aspects have been the focus of in vitro and in vivo investigations in animals. A mouse model was used to determine the bioavailability, pharmacokinetic profile and tissue distribution of pure P57 (Madgula et al., 2010b:1582-1586). In addition, the in vitro metabolic stability of P57 and interactions with drug metabolising enzymes (Madgula et al., 2008:1269-1275), the pharmacokinetic profile of hoodigogenin A, the aglycone of P57, and the stability of P57 in simulated gastrointestinal fluids (Madgula et al., 2010a:62-69) have been studied. The intestinal transport of pure P57 using the Caco-2 cell model has been investigated before (Madgula et al., 2010a:65-67), but this study is the first to compare the intestinal transport of pure P57 with that of P57 in a crude extract across intestinal mucosa. In addition, no transport studies across buccal tissue have been conducted yet for P57, which may contribute to the overall bioavailability of P57 specifically when the traditional use of the plant is considered. It was therefore the aim of this study to contribute to the current biopharmaceutical knowledge of the active ingredient of H. gordonii. This was done by investigating the in vitro transport of P57 across porcine intestinal and 126 buccal tissues in a Sweetana-Grass diffusion apparatus. In addition, the transport obtained for pure P57 with the transport of P57 when applied in the form of a crude plant extract was compared. 5.2 BACKGROUND AND LITERATURE REVIEW 5.2.1 Introduction The oral route of medicine administration is the most frequently used and preferred administration route when a systemic effect is required. The oral route is preferred over other routes of administration due to several reasons that include the following aspects: it is safe, simple, convenient, patients find it acceptable, good patient compliance, the administration of medicines can be controlled by the patients themselves and does not require skilled medical intervention, a large surface area for systemic absorption is available, many drugs are well-absorbed from the gastrointestinal tract and it enables the use of both immediate release and sustained release dosage forms (Chan, Lowes and Hirst, 2004:26; Daugherty and Mrsny, 1999:144; Hamman, 2007:64). Disadvantages associated with the oral route include a relatively slow onset of action, the possibility of irregular absorption, exposure to the first-pass metabolism effect and destruction of certain medicines by the enzymes and secretions of the gastrointestinal tract (York, 2007:7). The gastrointestinal tract is a hollow muscular tube that is divided into the oesophagus, the stomach, the small intestine and the colon. The small intestine can be further sub-divided into the duodenum, jejunum and ileum. The wall of the gastrointestinal tract consists of four main layers namely the mucosa, submucosa, muscularis externa and the serosa (Hamman, 2007:90-91). The intestinal epithelium separates the lumen of the gastrointestinal tract from the systemic circulation and is the main physical barrier to the absorption of orally administered compounds from the gastrointestinal tract. This membrane is semi-permeable and allows rapid transport of some molecules while impeding the transport of others. Lipophilic molecules are transported across the bilayer of apical membrane, while water and small hydrophilic molecules are transported through its numerous aqueous pores. The two main mechanisms of transport by which molecules can traverse the intestinal epithelium in both directions is transcellular (across the 127 cells) and paracellular (between the cells). The transcellular pathway is further divided into passive diffusion, carrier-mediated transport (active and facilitated) and endocytosis (fluid-phase endocytosis or pinocytosis, receptor-mediated endocytosis, phagocytosis and transcytosis). Several absorption pathways are schematically illustrated in Figure 5.1 (Hunter and Hirst, 1997:131; Ashford, 2007a:279-283). FIGURE 5.1: Schematic illustration of the drug absorption pathways across intestinal epithelial cells. (A) paracellular diffusion, (B) paracellular diffusion through opened tight junctions, (C) transcellular passive diffusion with (C*) intracellular metabolism, (D) carrier-mediated transcellular transport, (E) efflux and (F) vesicular transcytosis (Hunter and Hirst, 1997:131) The rate of transport of a substance is determined by its physicochemical properties, the nature of the membrane and the concentration gradient of the substance across the membrane. A balance between the aqueous solubility of the molecule, which dictates the amount of drug available for absorption, and its lipophilicity, which is important for membrane permeability, will provide the best oral bioavailability (Ashford, 2007a:279). The apical intestinal epithelial cell membrane also contains countertransport efflux proteins that actively expels molecules back into the lumen of the gastrointestinal tract after absorption. One of the key countertransport proteins is P-glycoprotein (P-gp), which is expressed at high levels on the apical surface of columnar cells in the jejunum. Efflux via P-gp is recognised as a mechanism which may significantly decrease bioavailability of 128 orally administered substances, especially at low doses. The P-gp transporters may also be blocked when two substances are co-administered allowing increased transport of one substance while the other substance blocks the P-gp transporter. The P-gp transporters are also saturable meaning that administration of high doses of a substance may cause saturation of the transporters (Ashford, 2007a:283; Hamman, 2007:108). Several in vitro methods have been investigated to determine intestinal and buccal permeability as an indicator of drug absorption in vivo. Drug delivery through the buccal mucosa (the cheek lining of the oral cavity) has several advantages as an alternative route of administration compared to oral administration. These include: easy access to the oral cavity and patient acceptance, avoidance of first-pass metabolism in the liver and intestine, avoidance of pre-systemic elimination within the gastrointestinal tract, bypassing peroral absorption variability as a result of gastric emptying rate or the presence of food, its relatively high permeability and good vascular supply (Shojaei, 1998:15, 27; Nicolazzo and Finnin, 2008:90; Obradovic and Hidalgo, 2008:167). The buccal route of administration is more suited to sustained release formulations and delivery of less permeable molecules than the sublingual mucosa. The sublingual mucosa is more permeable and is used when a rapid onset of action is required. A major disadvantage of the buccal mucosa is low flux resulting in low bioavailability, therefore penetration enhancers are often included in buccal formulations (Shojaei, 1998:19). The structure of the buccal mucosa consists of the stratified squamous epithelium (outermost layer) and lamina propria separated by the basement membrane and the submucosa. The buccal mucosa thickness is about 500-800 µm and the epithelium consists of about 40-50 cell layers. The buccal and sublingual areas within the oral cavity are non-keratinised while the gingivae and hard palate are keratinised (Shojaei, 1998:16; Nicolazzo and Finnin, 2008:91). The major barrier to the transport of molecules across the buccal mucosa is onethird to one-quarter from the apical surface of the buccal epithelium. Other barriers include the salivary film and mucous layer, the basement membrane and a metabolic barrier. Permeation of molecules through the buccal mucosa typically 129 occurs via passive diffusion but carrier-mediated transport has also been demonstrated for certain nutrients such as D-glucose (Nicolazzo and Finnin, 2008:91, 94). Passive diffusion involves transcellular (through the cells) as well as paracellular (between the cells) routes and is influenced by the physicochemical properties of the molecule (lipophilicity or hydrophilicity) as well as the pH partition theory. Lipophilic molecules and also the unionised form of a substance, predominantly permeates via the transcellular route while hydrophilic molecules permeate mainly via the paracellular route (Nicolazzo and Finnin, 2008:94; Obradovic and Hidalgo, 2008:167). 5.2.2 In vitro models for intestinal permeation studies 5.2.2.1 Membrane-based models One of the first steps to assess oral absorption is to determine the partition coefficient of a molecule indicating its lipophilicity which can then be used to predict the ability of a molecule to permeate through a biological membrane (Ashford, 2007b:307). This in itself is not sufficient and membrane-based models can be used to screen the transport of substances across lipophilic membranes. Parallel artificial membrane permeability assays (PAMPA) involve the use of organic solutions of artificial lipids to form bilayer structures simulating the cell membrane. The bilayer separates the aqueous donor and acceptor chambers and passive transport of small molecules is mimicked. It provides complementary information to other models in that passive transport permeability is assessed but it lacks paracellular and active transport mechanisms. It is a popular screening method due to its rapid results, low cost, versatility and easy adaptability in terms of increasing permeability of the membrane to investigate specific molecules (Bermejo et al., 2004:430; Hamman, 2007:17). In immobilised artificial membrane (IAM) chromatography, artificial monolayers of synthetic phospholipid analogues are bound to the silica of a high performance liquid chromatography (HPLC) column which then simulates a biological membrane. It therefore provides a measure of how well the solute in the aqueous phase will partition into biological membranes through determination of its retention on the column. When partitioning dominates the interaction of the molecule with the membranes and when it is the rate limiting step of drug movement, it may be used to predict the permeation of molecules (Ashford, 2007b:308; Hamman, 2007:18). 130 5.2.2.2 Cell culture models Cell culture techniques are well-accepted models to predict absorption of molecules. The most widely used cell line, Caco-2, is a human epithelial colon adenocarcinoma cell line that was first proposed and characterised as a model for oral drug absorption in 1989. The Caco-2 cell model has many advantages including determining the mechanisms of absorption of a substance, it can be used as a rapid screening tool, requires only a small amount of the substance for transport studies and it can be used as an indicator of cellular toxicity. In culture, Caco-2 cells spontaneously differentiate to form a monolayer of polarised enterocytes with distinct apical and basolateral sides that are separated by tight junctions. These enterocytes resemble those in the small intestine morphologically and biochemically, in that they contain microvilli and many of the transport systems present in the small intestine, for example the P-gp efflux transporter. Disadvantages of the Caco-2 cells are that the paracellular permeability is more comparable to the colon than the small intestine, they lack a mucus layer and lipophilic compounds may be retained within the cell layer (Ashford, 2007b:308-309; Hamman, 2007:18-23). To determine absorption, the substance in question is added to the apical chamber which is separated from the basolateral chamber by the Caco-2 monolayer. The appearance of the substance in the basolateral chamber is determined over time. If the disappearance of the substance from the apical chamber does not match its appearance in the basolateral chamber, the substance may be susceptible to enzyme degradation as it is transported through the cells. The substance may also be susceptible to metabolism within the epithelial cells due to the presence of cytochrome P450 (CYP) enzymes (Ashford, 2007b:310). Other cell models used in drug permeability studies include the Madin-Darby canine kidney (MDCK), Lewis lung carcinoma-porcine kidney 1 (LLC-PK1) and HT-29-A18C1 (subclone of human intestinal adenocarcinoma) cell lines as well as a TC-7 clone of Caco-2 cells. However, these cell models are used infrequently due to various disadvantages (Ashford, 2007b:310; Hamman, 2007:24). 5.2.2.3 Isolated tissue models Isolated tissue models for use in absorption studies include mounting excised sheets of intestinal mucosa in a diffusion chamber, the use of everted intestinal 131 sacs or solutions containing brush border membrane vesicles (Ashford, 2007b:310, Hamman, 2007:24-26). The diffusion chamber method involves sourcing intestine and removing the musculature to isolate the intestinal mucosa. The intestinal tissue from various animal species including pigs and rats has been used to investigate drug permeation and it has been shown that permeability data obtained using porcine intestinal tissue correlated well with human data (Ashford, 2007b:310; Nejdfors et al., 2000:501). The intestinal mucosa is cut into small pieces and mounted between the donor and acceptor chambers of a diffusion cell. The substance is added to the donor chamber and samples are removed from the acceptor chamber at specific time intervals and analysed. In addition, the chambers can be analysed independently allowing fluxes from the apical to basolateral and basolateral to apical sides to be measured. Advantages of this method include careful control of the experimental conditions by maintenance of the system at 37 °C and oxygen is provided to the tissue mimicking the in vivo condition. In addition, the pH of the buffer solutions can easily be altered and permeability across different sections of the intestine can be assessed (Ashford, 2007b:310, Hamman, 2007:26). In the everted intestinal sac method segments of the whole intestine are used as opposed to excised sheets of intestinal mucosa. The intestines are excised from rats, everted, and cut into segments which are tied up at both ends to form sacs. These sacs are placed into a solution containing the test substance for a period of time to allow substance uptake and samples are withdrawn at various time intervals and analysed. The method can distinguish between passive and active transport and can also show the influence of P-gp. The disadvantage of this method is the rapid loss of viability of the tissue (Ashford, 2007b:311; Hamman, 2007:25). The brush border membrane vesicle method involves the formation of vesicles, produced from intestinal scrapings, which contain luminal surface proteins and phospholipids with most of the brush border enzymatic and carrier activity. After incubation of the substance in a buffer solution with the vesicles, the vesicles are removed and the amount of substance absorbed by the vesicles is calculated (Hamman, 2007:24). 132 5.2.3 In vitro models for buccal permeation studies 5.2.3.1 Isolated tissue models The most widely used in vitro method for buccal permeation assessment is via excised buccal tissue from various animal sources mounted in diffusion chambers (Franz-type, flow-through, modified Ussing chamber). Many animal tissues including rabbit, monkey, dog, hamster, rat and pig buccal mucosa have been investigated as permeation models due to the shortage of sufficient quantities of human buccal mucosa. It has been found that the use of most of these tissues are limited due to the small area of non-keratinised tissue in the case of rabbits, the epithelium of dogs and monkeys is much thinner and therefore more permeable, and the surfaces of rat and hamster tissues are keratinised. The most commonly utilised tissue is porcine buccal mucosa as it is similiar to human mucosa in terms of structure, morphology, composition, thickness and shows good correlation in terms of permeability. In addition, porcine buccal mucosa is easily acquired at low cost from slaughterhouses (Shojaei, 1998:21; Nicolazzo and Finnin, 2008:97-99; Obradovic and Hidalgo, 2008:168). However, the use of isolated tissue models suffers several disadvantages including the limited surface area, large variability in permeation among replicates, the presence of tissue damage from mastication, minimal potential for automisation of the assay and the laborious and timeconsuming process of tissue excision (Obradovic and Hidalgo, 2008:172). 5.2.3.2 Cell culture models In addition to isolated tissue models, cell cultures have been developed as an in vitro technique to assess the permeability of compounds across buccal epithelia, particularly to overcome some of the disadvantages of isolated tissue models (Obradovic and Hidalgo, 2008:172). Cultures must be harvested from an appropriate source and grown under controlled conditions (temperature and humidity) using suitable growth medium. It was found that cultures from hamster cheek pouch do not differentiate into keratinised cells and therefore may be an appropriate model to predict permeability as the human buccal mucosa is nonkeratinised. Human buccal mucosa has been cultured from human buccal carcinoma cells (TR 146 cell line) but it is more permeable rendering it a less suitable model (Nicolazzo and Finnin, 2008:102). The most suitable cell culture model derived from healthy human buccal mucosa showed good correspondence 133 to intact human buccal mucosa in terms of morphology, protein expression and lipid composition with good reproducibility. In addition, good correlation was found on comparison between the human buccal cell culture and isolated human buccal mucosa in terms of permeability but only a limited amount of substances have been investigated. These cultures may also be used to evaluate drug metabolism, previously difficult to investigate due to the lack of fresh human buccal mucosa (Nicolazzo and Finnin, 2008:102; Obradovic and Hidalgo, 2008:172-173). 5.3. METHODS AND MATERIALS 5.3.1 In vitro intestinal and buccal permeation studies 5.3.1.1 Plant material and reagents A crude plant extract was prepared by extraction of the powder from a commercial H. gordonii capsule product with acetonitrile in an ultrasonic bath (Sonorex digital 10p) for 10 min at 25 °C. The process was repeated three times to ensure maximal extraction, the filtered extracts combined and air-dried. The P57 content of the commercial H. gordonii product was determined by LC-MS analysis. The pure P57 was purchased from ChromaDexTM Inc (California, USA). Solvents used in high performance liquid chromatograpy (HPLC) for quantification of P57 were of analytical grade. 5.3.1.2 Tissue preparation The numbers in the brackets in this section refers to Figure 5.2, which depicts photographs of the different step during tissue preparation for transport studies in the Sweetana-Grass diffusion apparatus. Porcine buccal and intestinal tissues from the oral and peritoneal cavities respectively were obtained from freshlyslaughtered pigs at a local abattoir (R&R Abattoir, Pretoria, South Africa) on each study day and washed with as well as transported in cold Krebs-Ringer bicarbonate buffer (referred to henceforth as buffer) (1). The intestinal tissue was prepared by sliding the intestine onto a glass tube and removing the serosal layer from the intestinal mucosa with blunt dissection (2). Thereafter the mucosal layer was cut along the mesenteric border (3) (Legen, Salobir, and Kerč, 2005:185), rinsed with buffer, placed on filter paper and cut into smaller pieces (4). The intestinal mucosa was placed on one half of the clamp, the filter paper removed (5) 134 FIGURE 5.2: The preparation and mounting of porcine intestinal mucosa in Sweetana-Grass diffusions chambers (Photographs by JH Hamman) 135 and the clamp was closed (6). The clamps containing the intestinal mucosa with an exposed surface area of 1.13 cm2 (7-8) were mounted between the two halfcells or chambers of the Sweetana-Grass diffusion apparatus with the mucosal side facing the apical chamber (Grass and Sweetana, 1988:374). The buccal mucosa was prepared by carefully removing excess underlying connective tissue using fine-point forceps and a scalpel, rinsed with buffer, clamped and mounted in the Sweetana-Grass diffusion half-cells (Ceschel et al., 2002:S46; Van Eyk and Van der Bijl, 2004:388). The assembled cells were placed in the six-cell aluminum block heater of the Sweetana-Grass apparatus, which is connected to a water bath maintaining the temperature at 37 °C. The apical and basolateral chambers were filled with 5 ml of fresh warmed buffer (37 °C). The tissues were continuously oxygenated with medical oxygen and an equilibration period of 30 min was allowed before commencement of the transport studies (Figures 5.2 (7-8) and 5.3). FIGURE 5.3: A photograph of a diffusion chamber cell as well as a schematic diagram of an assembled cell indicating the excised mucosal tissue mounted between two half-cells 5.3.1.3 Preparation of simulated intestinal fluid and artificial saliva Simulated intestinal fluid (SIF) was prepared by dissolving 6.8 g monobasic potassium phosphate (KH2PO4) in 250 ml water. After mixing, 190 ml 0.2 M sodium hydroxide (NaOH), 400 ml water and 10.0 g pancreatin (Sigma Aldrich) were added. The pH was adjusted to 7.5 ± 0.1 with 0.2 M sodium hydroxide and distilled water was added to provide a volume of 1000 ml (USP, 1990:1789). The artificial saliva was prepared by dissolving 5.208 g NaHCO3, 1.369 g K2HPO4.3H2O, 0.877 g NaCl, 0.447 g KCl, 0.441 g CaCl2.2H2O, 2.160 g mucin 136 and 200,000 U of hog pancreas α -amylase (Sigma Aldrich) in distilled water and making up the resulting solution to one litre. The resulting artificial saliva solution was adjusted to a pH of 7.0 ± 0.1 (Boland et al., 2004:402). 5.3.1.4 Transport studies Bi-directional intestinal transport experiments (apical-to-basolateral and basolateral-to-apical) were conducted for pure P57 and for P57 in H. gordonii crude extract across porcine intestinal tissue. Buccal transport experiments in the apical-to-basolateral direction were conducted for pure P57 and P57 in H. gordonii crude extract across porcine buccal tissue. In each experiment, 5% methanol was used to dissolve the pure P57 or the crude extract to increase solubility before it was made up to volume with the specific transport medium. Solutions made up to volume with buffer was adjusted to pH 7.4 ± 0.1, while that of SIF was adjusted to pH 7.5 ± 0.1 (USP, 1990:1789) and for artificial saliva to pH 7.0 ± 0.1 (Boland et al., 2004:402). A concentration of 200 µM of P57 was added to the chambers as previously reported by Madgula et al. (2008:1272) in an in vitro transport study using a Caco2 cell model. Therefore pure P57 was added to each chamber in a concentration of 1.2 mg/5 ml. It was calculated that 2.4 g of plant material had to be extracted to provide approximately 1.2 mg of P57 per chamber based on the LC-MS result of 0.304% P57 determined for the commercial H. gordonii product. Samples (200 µl) drawn from each of the original solutions before they were applied to the tissues in the transport studies were analysed for P57 content. The transport studies were initiated by adding 5 ml of each solution as indicated in Table 5.1 to the donor chamber (the apical or basolateral chamber, respectively). Samples (200 µl) were drawn from the receiver chamber over a period of 2 h at 20 min intervals and the volume of each sample was replaced with warm (37 °C) transport medium. 137 TABLE 5.1: Experimental groups used for the intestinal and buccal transport studies Intestinal transport studies Absorptive direction Secretory direction Apical chamber Basolateral chamber Apical chamber Basolateral chamber Pure P57 + buffer Buffer Buffer Pure P57 + buffer Extract + buffer Buffer Buffer Extract + buffer Extract + SIF Buffer SIF Extract + buffer Buccal transport studies Absorptive direction Apical chamber Basolateral chamber Pure P57 + buffer Buffer Pure P57 + artificial saliva Buffer Extract + buffer Buffer Extract + artificial saliva Buffer 5.3.1.5 Quantification of P57 Quantification of P57 was performed with an HPLC system consisting of a Waters (MA, USA) 2690 separation module and a Waters 996 photodiode array detector and separation achieved using a Gemini C18 column (250 x 4.6 mm; 5 µm particle size; Phenomenex) equipped with a guard column (4 mm x 3.0 mm, Phenomenex). The mobile phase consisted of water containing 0.05% acetic acid (A) and acetonitrile (B). Gradient elution was used as follows: 100% A at 0 min, adjusted in the following 40 min to 20% A and 80% B. Each run was followed by a 5 min wash with 100% B and an equilibration period of 8 min. The flow rate was 1.0 ml/min, the column was temperature controlled at 30 °C and the detection wavelength was 220 nm. The chromatographic data was collected and analysed with Empower® software. 5.3.1.6 Data analysis and statistical evaluation The transport results were corrected for dilution and the cumulative P57 transport was plotted as a function of time. The apparent permeability coefficient (Papp) values (cm/s) were calculated according to the following equation (Hansen and Nilsen, 2009:88): 138 Papp = (dC/dt) (1/A.60.C0) (cm/s) where dC/dt is the transport rate (slope), A is the surface area of the tissue (1.13 cm2) and C0 is the initial concentration of P57 applied. Statistical analysis of the Papp values for both the buccal and intestinal transport experiments was done by means of a repeated one-way analysis of variance (ANOVA) to indicate whether differences are significant (p ≤ 0.05) or not. The P57 flux (J) values were calculated according to the following equation (Hansen and Nilsen, 2009:88): J = V(dC/dt)/A (µg/cm2/h) where V is the receiver chamber volume, dC is the receiver chamber P57 concentration (µg/ml), dt is the transport time (h) and A is the tissue surface area (cm2). The net flux of P57 (JNET) was calculated as follows (Hansen and Nilsen, 2009:88): JNET = JB-A – JA-B (µg/cm2/h) where JB-A is the flux of P57 in the B-A direction and JA-B is the flux of P57 in the AB direction. 5.4 RESULTS AND DISCUSSION 5.4.1 In vitro intestinal transport studies The Papp values for P57 transport across intestinal mucosa are shown in Figure 5.4, while the flux (J) values for P57 are shown in Table 5.2. FIGURE 5.4: Transport of pure P57 as well as P57 from a crude plant extract across porcine intestinal tissue in the absorptive direction ( direction ( ) and secretory ) expressed in terms of apparent permeability coefficient (Papp) values 139 In general, the in vitro transport of pure P57 in buffer across porcine intestinal tissue was very low, but it was significantly higher in the secretory direction (Papp = 0.059 x 10-6 cm/s) than in the absorptive direction (Papp = 0.022 x 10-6 cm/s) indicating that P57 is actively effluxed by intestinal membrane transporters. This is also shown by the positive JNET flux value (Table 5.2). TABLE 5.2: Flux (J) values (µg/cm2/h) for P57 across porcine intestinal mucosa Transport direction A-B (µg/cm2/h) B-A (µg/cm2/h) JNET flux (µg/cm2/h) Pure P57 + buffer 0.174 0.501 0.3265 Extract + buffer 10.523a 10.476a -0.0469 Extract + SIF 1.357b 1.356b -0.0009 A-B: Apical to basolateral; B-A: Basolateral to apical. Values with the same superscript letter (a,b) are not significantly different from each other. These results are congruent with a previous study by Madgula et al. (2008:1272) where the intestinal transport of pure P57 across Caco-2 cell monolayers revealed lower transport in the absorptive direction than in the secretory direction. In this previous Caco-2 cell study it was also shown that in the presence of inhibitors of two membrane transporters namely verapamil which inhibits P-gp and MK-571 which inhibits multidrug resistance-associated protein, the permeability of P57 increased in the absorptive direction indicating it is a substrate of these two active membrane transporters. In the absence of these efflux pump inhibitors, the P57 permeability was concentration-dependent indicating saturation of transporters (Madgula et al. 2008:1273). The intestinal transport of P57 applied as a H. gordonii crude extract in buffer had a significantly higher Papp value of 74.479 x 106 cm/s compared to the Papp value of pure P57 of 0.022 x 10-6 cm/s in the absorptive direction as well as in the secretory direction Papp of 73.466 x 10-6 for the crude extract and 0.059 x 10-6 cm/s for pure P57. Furthermore, P57 transport from the crude extract in buffer was slightly but not significantly higher in the absorptive direction (Papp = 74.479 x 10-6 cm/s) than in the secretory direction (Papp 73.466 x 10-6 cm/s) with a negative flux value of -0.0469 µg/cm2/h indicating that phytochemicals in the crude extract influences P57 transport by means of efflux 140 inhibition but other mechanisms such as changes to P57 solubility and membrane permeability cannot be excluded. The phenomenon that crude extracts are in some cases more biologically active than the pure compounds has been established before. For example, pure aspalathin was transported to a lower extent across Caco-2 cell monolayers compared to aspalathin in a crude Rooibos tea extract (Huang et al., 2008). The transport of P57 in the H. gordonii crude extract dissolved in SIF was significantly decreased in both the absorptive and secretory directions (Papp = 10.617 x 10-6; 14.608 x 10-6 cm/s) compared to the P57 transport from the crude extract dissolved in buffer (Papp = 74.479 x 10-6; 73.466 x 10-6 cm/s). The transport in SIF was significantly higher in the secretory direction compared to the absorptive direction. These lower transport results can be explained by the fact that the stability of P57 was significantly affected by SIF. This corresponds to a stability study by Madgula et al. (2010a:65) where it was determined that 8.6% of the initial concentration of pure P57 applied was degraded within 180 min when exposed to SIF. In addition, 100% of pure P57 was degraded in simulated gastric fluid (SGF) after 60 min. As the simulated intestinal fluid is used to mimic in vivo conditions, these results indicate that the bioavailability of orally administered P57 may be significantly affected by enzymatic degradation and/or hydrolysis. An in vivo study by Madgula et al. (2010b:1582) in CD1 female mice revealed a peak plasma level of P57 after 0.6 h and moderate bioavailability of 47.5% after oral administration of a methanolic extract of H. gordonii (equivalent to a dose of 25 mg/kg P57). The moderate oral bioavailability is likely due to degradation in gastric and intestinal fluids. Acid hydrolysis in the stomach may be circumvented by the use of enteric-coating which enables release of a substance from a dosage form only upon reaching the small intestines. In addition, the buccal route of administration may be considered as an alternative route of administration to avoid the harsh conditions in the gastrointestinal tract that degrades P57. 5.4.2 In vitro buccal transport studies The Papp values of the buccal transport of P57 from H. gordonii crude extract across porcine buccal mucosa are shown in Figure 5.5 while the flux values (J) are shown in Table 5.3. 141 FIGURE 5.5: Transport of pure P57 as well as P57 from a crude plant extract across porcine buccal tissue in the absorptive direction expressed in terms of apparent permeability coefficient (Papp) values TABLE 5.3: Flux (J) values (µg/cm2/h) for P57 across porcine buccal mucosa Transport direction A-B (µg/cm2/h) Pure P57 + buffer 0 Pure P57 + saliva 0 Extract + buffer 2.223a Extract + saliva 2.204a Values with the same superscript letter (a) are not significantly different from each other. In this study, no buccal transport of pure P57 in either buffer or artificial saliva was obtained at the concentration tested. However, when crude extract was dissolved in buffer, the Papp value was 19.149 x 10-6 cm/s (Figure 5.5), indicating that phytoconstituents present in the crude extract significantly enhanced the transport of P57 across the buccal tissue. Interestingly, the transport was significantly higher when the crude extract was dissolved in artificial saliva (Papp = 22.081 x 10-6 cm/s) as compared to the buffer indicating that α-amylase does not degrade P57. These results further indicate that significantly more P57 is transported across the buccal membrane when in vivo conditions are simulated probably by effects of the saliva such as solubility improvement. 142 5.5 CONCLUSIONS It was determined with in vitro transport experiments that pure P57 is not transported across excised porcine buccal mucosa, while it is transported across excised porcine intestinal mucosa. The transport of P57 from a crude extract was significantly higher than the transport obtained for pure P57 across both porcine intestinal and buccal mucosal tissues. This indicates that the crude extract contains phytoconstituents that influence the transport of P57 across these mucosal tissues. The results obtained reinforce the credibility of the traditional method of chewing and oral consumption of the plant itself as opposed to oral administration of the pure isolated compound, P57. It can further be concluded that buccal transport of P57 from plant material may contribute to the overall bioavailability and therefore the pharmacological effects of P57, the perceived active ingredient of H. gordonii. 143 5.6 REFERENCES ASHFORD, M. 2007a. Gastrointestinal tract – physiology and drug absorption. In: AULTON, M.E. (ed.). Aulton’s Pharmaceutics: the design and manufacture of medicines. Hungary: Elsevier Limited:270-285. ASHFORD, M. 2007b. Assessment of biopharmaceutical properties. In: AULTON, M.E. (ed.). Aulton’s Pharmaceutics: the design and manufacture of medicines. Hungary: Elsevier Limited:304-323. BERMEJO, M., AVDEEF, A., RUIZ, A., NALDA, R., RUELL, J.A., TSINMAN, O., GONZÁLEZ, I., FERNÁNDEZ, C., SÁNCHEZ, G., GARRIGUES, T.M., MERINO, V. 2004. PAMPA – a drug absorption in vitro model. 7. Comparing rat in situ, Caco-2, and PAMPA permeability of fluoroquinolones. European Journal of Pharmaceutical Sciences, 21:429-441. BOLAND, A.B., BUHR, K., GIANNOULI, P., VAN RUTH, S.M. 2004. Influence of gelatin, starch, pectin and artificial saliva on the release of 11 flavour compounds from model gel systems. Food Chemistry, 86:401-411. CESCHEL, G.C., MAFFEI, P., SFORZINI, A., LOMBARDI BORGIA, S., YASIN, A., RONCHI, C. 2002. In vitro permeation through porcine buccal mucosa of caffeic acid phenetyl ester (CAPE) from a topical mucoadhesive gel containing propolis. Fitoterapia, 73 (Suppl.1):S44-S52. CHAN, L.M.S., LOWES, S., HIRST, B.H. 2004. The ABCs of drug transport in intestine and liver: efflux proteins limiting drug absorption and bioavailability. European Journal of Pharmaceutical Sciences, 21:25-51. DAUGHERTY, A.L., MRSNY, R.J. 1999. Transcellular uptake mechanisms of the intestinal epithelial barrier Part one. Pharmaceutical Science and Technology Today, 2(4):144-151. GLASL, S. 2009. Hoodia: A herb used in South African traditional medicine – A potential cure for overweight? Pharmacognostic review of history, composition, health-related claims, scientific evidence and intellectual property rights. Schweizerische Zeitschrift für GanzheitsMedizin, 21(6):300-306. GRASS, G.M., SWEETANA, S.A. 1988. In vitro measurement of gastrointestinal tissue permeability using a new diffusion cell. Pharmaceutical Research, 5(6):372376. HAMMAN, J.H. 2007. Oral drug delivery: biopharmaceutical principles, evaluation and optimisation 2nd ed. Pretoria: Content solutions. HANSEN, T.S., NILSEN, O.G. 2009. Echinacea purpurea and P-glycoprotein drug transport in Caco-2 cells. Phytotherapy Research, 23:86-91. HUANG, M.J., DU PLESSIS, J., DU PREEZ, J., HAMMAN, J.H., VILJOEN, A.M. 2008. Transport of aspalathin, a Rooibos tea flavonoid, across the skin and intestinal epithelium. Phytotherapy Research, 22:699-704. 144 HUNTER, J. HIRST, B.H. 1997. Intestinal secretion of drugs. The role of Pglycoprotein and related drug efflux systems in limiting oral drug absorption. Advanced Drug Delivery Reviews, 25:129-157. LEE, R.A., BALICK, M.J. 2007. Indigenous use of Hoodia gordonii and appetite suppression. Explore, 3(4):404-406. LEGEN, I., SALOBIR, M., KERČ, J. 2005. Comparison of different intestinal epithelia as models for absorption enhancement studies. International Journal of Pharmaceutics, 291:183-188. MADGULA, V.L.M., AVULA, B., PAWAR, R.S., SHUKLA, Y.J., KHAN, I.A., WALKER, L.A., KHAN, S.I. 2008. In vitro metabolic stability and intestinal transport of P57AS3 (P57) from Hoodia gordonii and its interaction with drug metabolizing enzymes. Planta Medica, 74:1269-1275. MADGULA, V.L.M., AVULA, B., PAWAR, R.S., SHUKLA, Y.J., KHAN, I.A., WALKER, L.A., KHAN, S.I. 2010a. Characterization of in vitro pharmacokinetic properties of hoodigonenin A from Hoodia gordonii. Planta Medica, 75(1):62-69. MADGULA, V.L.M., ASHFAQ, M.K., WANG, Y.-H., AVULA, B., KHAN, I.A., WALKER, L.A., KHAN, S.I. 2010b. Bioavailability, pharmacokinetics, and tissue distribution of the oxypregnane steroidal glycoside P57AS3 (P57) from Hoodia gordonii in mouse model. Planta Medica, 76(14):1582-1586. NEJDFORS, P., EKELUND, M., JEPPSSON, B., WESTRÖM, B.R. 2000. Mucosal in vitro permeability in the intestinal tract of the pig, the rat, and man: species- and region-related differences. Scandinavian Journal of Gastroenterology, 35(5):501507. NICOLAZZO, J.A., FINNIN, B.C. 2008. In vivo and in vitro models for assessing drug absorption across the buccal mucosa. In: EHRHARDT, C., KIM, K.-J. (eds.). Drug Absorption Studies - In Situ, In Vitro, and In Silico Models. New York: Springer:89-111. OBRADOVIC, T., HIDALGO, I.J. 2008. In vitro models for investigations of buccal drug permeation and metabolism. In: EHRHARDT, C., KIM, K.-J. (eds.). Drug Absorption Studies - In Situ, In Vitro, and In Silico Models. New York: Springer:67181. SHOJAEI, A.H. 1998. Buccal mucosa as a route for systemic drug delivery: a review. Journal of Pharmacy and Pharmaceutical Sciences, 1(1):15-30. USP. 1990. United States Pharmacopoeia 22nd ed. Rockville: The United States Pharmacopoeial convention, Inc:1789. VAN EYK, A.D., VAN DER BIJL, P. 2004. Comparative permeability of various chemical markers through human vaginal and buccal mucosa as well as porcine buccal and mouth floor mucosa. Archives of Oral Biology, 49:387-392. 145 VAN WYK, B.-E., GERICKE, N. 2000. People’s plants. Pretoria: Briza publications. YORK, P. 2007. Design of dosage forms. In: Aulton, M.E. (ed.). Aulton’s Pharmaceutics: the design and manufacture of medicines. Hungary: Elsevier Limited:1-14. Figure references (www) FIGURE 5.3 (left): http://warneronline.com/img_lg/690993_p2300_gr.jpg [Accessed: 2010/12/12]. 146 CHAPTER 6 General conclusions and recommendations 6.1 GENERAL CONCLUSIONS The occasional use of H. gordonii and other Hoodia species as appetite and thirst suppressants by people indigenous to South Africa has been well-documented leading to the popular consumption of H. gordonii preparations as anti-obesity preparations. Use of products of natural origin is particularly appealing to consumers due to the general perception that if it is natural it must be effective and safe. All of the quality control studies performed to date showed evidence of fake products claiming to contain H. gordonii therefore quality control has been the main focus of scientific studies. Furthermore, the supply of H. gordonii plant material is limited due to its sparse geographical distribution, slow maturation rate, need for a permit to cultivate or export material as well as high public demand, contributing to adulteration of a large amount of products. Despite the isolation of numerous steroidal glycosides from H. gordonii, the main focus has been on the pregnane glycoside P57, considered to be the active ingredient and marker molecule to determine quality of H. gordonii raw material and products. The first aim of this thesis was to develop and optimise rapid quality control methods for H. gordonii raw material and products. In terms of quality control a chemical fingerprint for H. gordonii raw material was developed and the presence or absence of P57 in raw material and products was successfully detected using high performance thin layer chromatography (HPTLC). Several novel quantification methods for P57 were developed using near infrared (NIR) and Raman spectroscopy coupled with chemometric data analysis as well has quantitative HPTLC. These methods are advantageous in that they allow simpler and faster quantification methods as opposed to the traditional analytical methods such as liquid chromatography coupled to a photo diode-array detector (LC-UV) and liquid chromatography coupled to a mass spectrometer (LC-MS). The wellestablished traditional methods were used to develop reference sets needed to expand the current quality control methods. Adulteration of raw material could be 147 detected to a certain extent with HPTLC and the location of P57 within a stem sample was determined using Raman mapping. Publications based on scientific studies of key aspects such as in vivo biopharmaceutics, the biological activity of all chemical constituents, clinical efficacy and especially safety are insufficient or completely absent for H. gordonii This causes great concern as H. gordonii is one of the most widely consumed antiobesity products of natural origin. Therefore, the second aim was to contribute towards increasing the biopharmaceutical knowledge of P57. It was shown in this study for the first time that P57 is transported across porcine intestinal mucosa which is closely related to human mucosa to a significantly higher extent when applied in the form of a crude extract as compared to pure P57. Furthermore, P57 in its pure form was not transported across porcine buccal mucosa but it was transported in the crude extract form. The results therefore indicate the possibility that some of the more than 40 other compounds in H. gordonii crude extract may beneficially affect P57 transport across mucosal epithelia. The use of media such as artificial saliva and simulated intestinal fluid which closely represents in vivo conditions revealed that saliva may have a beneficial effect in the transport of P57 while exposure to intestinal fluid has a deleterious effect on P57. In addition to meeting the aims and objectives of the study, comprehensive reviews on the literature pertaining to H. gordonii as well as anti-obesity preparations of plant origin were conducted. The H. gordonii case encompasses all issues related to the development of commercial products from natural origin including intellectual property rights, benefit sharing and biodiversity issues, protection of natural resources, the complexity of plant extracts, the paramount importance of quality control methods and the need for proper research including clinical studies prior to commercialisation. 6.2 RECOMMENDATIONS As a result of this study, new questions have arisen which should be investigated. Many of the Hoodia species have been reported to be used for weight loss and contain P57 and other steroidal glycosides, but only H. gordonii has been scientifically investigated to some extent. All pharmacokinetic studies completed 148 up to this point indicated that the oral bioavailability of P57 is low suggesting that other molecules or metabolites may be responsible for the anti-obesity effect. The aglycone of P57, hoodigogenin A, is a prime candidate for further research. In addition, the investigation of other Hoodia species for anti-obesity activity should be considered. Many aspects that should be investigated prior to commercialisation of products claiming biological activity such as bioavailability of the active ingredient has been lacking. For example, H. gordonii is also marketed in the form of a topical gel, yet there is no evidence to support sufficient transdermal permeation to produce pharmacological effects. A future avenue of research should be to determine whether P57 and other compounds from H. gordonii are in fact transported across the skin. A void still exists in terms of quality control of commercial H. gordonii products as compounds unique to H. gordonii has not been found and none of the methods developed showed that it could distinguish between different species of Hoodia in powder form used to manufacture products. In addition, the HPTLC quantification method developed during this study should be optimised for possible implementation as a standard quality control method in practice. 149 ANNEXURE 1: Chemical structures 1.1 Chemical structures of the steroidal glycosides isolated from H. gordonii with their corresponding aglycones indicated (*,#,$,@,^) O O O HO O O H3CO OH O O OH O O OCH3 OCH3 P57/Compound 1/Formula 6* O O O HO H3CO O O O OH O OH OCH3 Hoodigoside A* O O O HO H3CO O O O OH H CO 3 O O OH O OH OCH3 Hoodigoside B* O O O HO O H3CO OH O O O O OCH3 O O OCH3 OCH3 Hoodigoside C/Gordonoside C* 150 OH O O O O O HO H3CO O O O O OH H3CO O O OH OH O OCH3 Hoodigoside D * OCH3 O O O OH O HO HO OH O O H3CO OH O O O O OH O OCH3 OCH3 Hoodigoside E/Gordonoside H* O OH O HO HO O O O O O O OH H3CO O O H3CO O O OH OH O OCH3 Hoodigoside F OCH3 * O OH O HO HO O O O O O O OH O O H3CO OCH3 O O OH OH O OCH3 Hoodigoside G OCH3 * O OH HO HO O O O O O O O OH OCH3 O O OCH3 O O OCH3 Hoodigoside H 151 * OH O OCH3 O OH O HO O O HO O O O O OH H3CO O O OCH3 OH O O O OCH3 Hoodigoside I OCH3 * O OH O HO O O HO O O O O OH H3CO O O O O OH OH O OCH3 Hoodigoside J OCH3 * O OH O O O HO HO O OH HO O O O HO OH O OH OCH3 Hoodigoside K* H HO O O OH OH HO O O O O H3CO O O OH H3CO OH OH OH HO O O O OH OH O HO Hoodigoside L# H HO O OH OH HO O HO H3CO O O OH H3CO O O OH O OH OH HO O O OH OH O Hoodigoside M# 152 HO H OH O HO H3CO O O H OH 3CO OH O Hoodigoside N# H HO O OH OH HO O O O O O H3CO O O H3CO OH OH O HO Hoodigoside O # H HO O O OH OH HO O O O O H3CO O O OH OH OH HO O O O OCH3 HO # H HO O O OH OH HO O O O O O O O OCH3 OH O OCH3 HO # H HO O O OH OH HO O O O O O O OCH3 OH OH OH HO O O O OCH3 HO H HO O O OH OH HO O O O O OCH3 O O O O O OCH3 OH OH O Hoodigoside R# O H3CO OH OH O Hoodigoside Q O OH OH HO O O OH O H3CO OH OH O Hoodigoside P O H3CO OH OH O OH O OCH3 OH OH HO O O OH OH O Hoodigoside S# 153 HO H HO O O OH OH HO O O O H3CO O O O O O O O OCH3 OH O OCH3 OH OH O HO OCH3 Hoodigoside T# H HO O O OH OH HO O O O O O OCH3 OCH3 O O O O O OCH3 OH OH OH HO O O O OCH3 OH OH O HO Hoodigoside U# HO O O H O OH OH OH O O H3CO O O H OH 3CO OH O Hoodigoside V/Compound 2# O O O O HO H3CO O O H3CO O O O O OH OH O OCH3 Hoodigoside W OCH3 * O OH O HO H3CO O O O O OH OH O OCH3 OCH3 Hoodigoside X$ HO O H O OH OH OH HO H3CO O O OH H3CO O OH O Hoodigoside Y# 154 HO O H OH OH O O OH O O H3CO O O O O OH O OH OCH3 OCH3 Hoodigoside Z# O O O HO O H3CO OH O O O O O O H3CO OH O OCH3 Gordonoside B * OCH3 O O O O HO O O O O O O H3CO HO OH O OCH3 Gordonoside D * OCH3 O O O HO O O O H3CO O O OH O O H3CO O OCH3 Gordonoside E * OCH3 O O O HO O O O O O H3CO OCH3 O O OCH3 OCH3 Gordonoside F/Formula 9* 155 OH O O O O O HO O O O O H3CO O O OCH3 OH O OCH3 OCH3 Gordonoside G/Formula 10* O O O O HO O O H3CO HO O O O O O O H3CO OH O OCH3 Gordonoside I OCH3 * O O O O OH O O H3CO O O H3CO O O OCH3 O O OH O OCH3 Gordonoside L/Formula 11 * OCH3 HO O H O O H3CO O OH H O H O O O OH OH OH OH OH CHO OCH3 Hoodistanaloside A@ HO O H O O H3CO H O H O O O OH OCH3 Hoodistanaloside B^ 156 OH CHO O OH OH OH O O O O HO O O O O H3CO OCH3 O O OH OH O OCH3 Formula 7 OCH3 * O O O O HO O O O O H3CO O O OH OH O OCH3 Formula 8 OCH3 * O O O O HO O O O O H3CO O O OCH3 OH OH O OCH3 Formula 12 * OCH3 Chemical structures of aglycones O O O OH OH H OH OH HO OH HO Hoodigogenin A* /Gordonoside A HO Calogenin# Isoramanone$ OH OH H H H H HO OH H H OH CHO Hoodistanal O HO @ OH CHO Dehydrohoodistanal^ 157 ANNEXURE 2: Publications emanating from this thesis 2.1. Publications in accredited journals § Vermaak, I., Hamman, J.H., Viljoen A.M. 2010. High performance thin layer chromatography as a method to authenticate Hoodia gordonii raw material and products. South African Journal of Botany, 76:119-124. § Vermaak, I., Hamman, J.H., Viljoen A.M. 2010. A rapid spectroscopic method for quantification of P57 in Hoodia gordonii raw material. Food Chemistry, 120:940-944. § Vermaak, I., Viljoen A.M., Hamman, J.H., Baranska, M. 2010. The potential application of FT-Raman spectroscopy for the quantification and mapping of the steroidal glycoside P57 in Hoodia gordonii. Phytochemistry Letters, 3:156-160. § Vermaak, I., Hamman, J.H., Viljoen, A.M. 2010. Hoodia gordonii: An up-todate review of a commercially important anti-obesity plant. Planta Medica, in press. § Vermaak, I., Viljoen, A.M., Chen, W., Hamman, J.H. In vitro transport of the steroidal glycoside P57 from Hoodia gordonii across excised porcine intestinal and buccal tissue. Phytomedicine, in press. 2.2. Publication in a non-accredited journal § Vermaak, I., Viljoen, A.M. 2008. Indigenous South African Medicinal Plants. Part 9: Hoodia gordonii. South African Pharmaceutical Journal, 75(4):37. 2.3. Publication in preparation for submission to an accredited journal § Vermaak, I., Hamman, J.H., Viljoen, A.M. The quantification of P57 from H. gordonii using HPTLC coupled to densitometry. 158 ANNEXURE 3: Conference contributions 3.1. Oral papers § Vermaak, I., Viljoen, A.M., Manley, M. Towards a rapid quality control method for Hoodia gordonii. Indigenous Plant Use Forum (IPUF) 2008, 7-11 July, Graaff-Reinet, South Africa. § Vermaak, I., Viljoen, A.M., Manley, M. Towards a rapid quality control method for Hoodia gordonii raw materials and products. World Conference on Medicinal and Aromatic Plants IV 2008, 9 – 14 November, Cape Town, South Africa. § Vermaak, I., Viljoen, A.M., Hamman, J.H. Quality control of Hoodia gordonii raw material and products with HPTLC and FT-NIR spectroscopy. 5th International Conference on Pharmaceutical and Pharmacological Sciences (ICPPS) 2009, 23 – 26 September, Potchefstroom, South Africa. § Vermaak, I., Viljoen, A.M., Hamman, J.H. Quality control of Hoodia gordonii raw material and products. 5th University of Johannesburg Botany Symposium 2009, 27 October, Johannesburg, South Africa. § Vermaak, I., Viljoen, A.M., Hamman, J.H. Quality control of Hoodia gordonii raw material and products. TUT Faculty of Science Research Day 2010, 4 August, Pretoria, South Africa. § Vermaak, I., Viljoen, A.M., Hamman, J.H. In vitro permeability of the steroidal glycoside P57 from Hoodia gordonii through porcine buccal and intestinal mucosa. 31st Annual Conference of the Academy of Pharmaceutical Sciences South Africa (APSSA) 2010, 14 – 17 September, Polokwane, South Africa. § Vermaak, I., Viljoen A.M., Hamman, J.H. In vitro permeability of the steroidal glycoside P57 from Hoodia gordonii through porcine buccal and intestinal mucosa. Department of Pharmaceutical Sciences (TUT) Research showcase 2010, 22 October, Pretoria, South Africa. 3.2. Poster § Vermaak, I., Viljoen, A.M., Hamman, J.H. Quality control of Hoodia gordonii raw material and products. 58th International Congress and Annual Meeting of the Society for Medicinal Plant and Natural Product Research (GA) 2010, 29 August – 2 September, Berlin, Germany. 159