Interferon-a Overrides the Deficient Adhesion of
Transcription
Interferon-a Overrides the Deficient Adhesion of
From www.bloodjournal.org by guest on October 25, 2016. For personal use only. Interferon-a Overrides the Deficient Adhesion of Chronic Myeloid Leukemia Primitive Progenitor Cells to Bone Marrow Stromal Cells By Charles Dowding, Ai-Pu Guo, Jurgen Osterholz, Martin Siczkowski, John Goldman, and Myrtle Gordon Primitive blast colony-forming cells (BI-CFC) from chronic myeloid leukemia (CML) patients are defective in their attachment t o bone marrow-derived stromal cells compared with normal BI-CFC. We investigated the effect of recombinant interferon-% (IFN-a) on this interaction between hematopoietic progenitor cells and bone marrow-derived stromal cells by culturing normal stromal cells with IFN-a (50 t o 5,000 U/mL). At 50 U/mL we found that: (1)the capacity of stromal cells to bind two types of CML primitive progenitor cells (BI-CFC and long-term culture-initiating cells) was increased; and (2)the amount of sulfated glycosaminoglycans (GAGS)in the stromal layer was increased. However, sulfated GAGS were not directly involved in binding CML BI-CFC, unlike binding by normal BI-CFC, which is sulfated GAG-dependent. R ECOMBINANT interferon-a (IFN-a) has proved to be useful in the treatment of Philadelphia chromosome (Ph)-positive chronic myeloid leukemia (CML). About 70% of patients show long-lasting hematologic control.’ In some responding patients, IFN-a permits the re-emergence of a significant degree of Ph-negative hematopoiesis,‘.’ but the mechanism of its apparently selective action in CML is unknown. IFN-a has a direct antiproliferative effect on both normal and CML progenitor cells in vitro but no selective anti-leukemic activity on progenitor cell proliferation has been dem~nstrated.”~ In contrast to IFN-a, recombinant IFN-y has proved to be ineffective in treating or in combinapatients with CML either as a single agent5*6 tion with IFN-a.’ The bone marrow (BM) microenvironment is a possible alternative target for IFN-a. The binding of CML progenitor cells to stromal cells in vitro is deficient,* in vivo this deficiency could lead to release of progenitor cells into the circulation and hematopoiesis in extramedullary sites. We therefore speculated that IFN-a might act by modifying the interaction between CML and stromal cells in a way that allows normal microenvironmental control to be restored. Accordingly, we investigated the binding of CML progenitor cells to stromal cells cultured with IFN-a or IFN-y and the effect of IFN-a on the proliferation of Ph-positive and Ph-negative long-term culture-initiating cells. We also investigated the influence of IFN-a and IFN-y on extracellular matrix (ECM) synthesis by stromal cells. Heparan sulfate and chondroitin sulfate, both sulfated glycosaminoglycan (GAG) molecules, are highly negatively charged ECM components identified in cultures of BMderived stromal cells? Heparan sulfate has been implicated in normal hematopoietic cell-stromal cell interactions” and as a growth factor sequestrator within the ECM of stromal cells.” Another negatively charged molecule, neuraminic acid, is also involved in cell-to-cell interaction~.’~”~ Therefore, we investigated the effects of IFN-a and IFN-y on (1) the synthesis of sulfated GAG by stromal cells, and ( 2 ) the role of sulfated GAG and neuraminic acid molecules in the attachment of primitive myeloid progenitor cells to BMderived stromal cells. Blood, Vol78, No 2 (July 15). 1991: pp 499-505 Neuraminidase-treated control stromal cells bound an increased number of CML BI-CFC, reproducing the effect of IFN-a, whereas the binding t o IFN-ortreated stromal cells was unaffected by neuraminidase treatment. Thus, the enhanced attachment by primitive CML progenitor cells t o INE-ortreated stromal cells might be due t o qhanges in the neuraminic acid composition in the stromal cell layer. Our in vitro evidence may provide insights into the mechanism of action of IFN-a in vivo. Prolonged administration may alter the marrow microenvironment in some patients such that it can restrain the aberrant proliferation of Philadelphia chromosome (Ph)-positive stem cells while permitting Ph-negative stem cells t o function normally. o 1991by The American Society of Hematology. MATERIALS AND METHODS Samples. Peripheral blood cells from 13 patients with Ph chromosome-positive CML in chronic phase were collected before initiation of therapy with cytotoxic drugs. The median white blood cell (WBC) count for the patients investigated was 198 x lOPLand ranged from 124 to 287 x 109/L.Normal BM cells collected from iliac crests of donors of marrow for allogeneic transplantion were used as controls. Informed consent was given to use samples in experimental procedures. For most experiments mononuclear cell fractions (MNCs) from blood or marrow were prepared by centrifugation over Lymphoprep (density 1.077 gimL; Nycomed, Oslo, Norway) according to the method of Boyum.14 For the long-term culture experiments, MNCs were prepared by centrifugation (400g, 30 minutes) over iso-osmotic Percoll (density 1.065 g/mL; Pharmacia, Uppsala, Sweden). MNCs prepared by both techniques were washed twice in Hanks’ balanced salt solution (HBSS; GIBCO, Paisley, UK) before use. Stromal cell cultures with IFN-a or IFN-y. Stromal layers were grown by plating 5 x lo5 normal BM MNCs per milliliter in a medium (GIBCO) containing fetal calf serum (FCS; 15%, GIBCO), methylprednisolone (1.6 x mol& Upjohn, Crawley, UK), with (50 to 5,000 UImL) or without recombinant IFN-(y, (2 X 10’ Ulmg protein; Roferon, Hoffman La Roche, Basel, Switzerland) or with (25 to 100 UImL) or without recombinant IFN-y (2 to 4 X lo8 U/mg protein; Biogen S.A., Basel, Switzerland) in 35” petri dishes or 25-cm’ flasks (GIBCO). Cultures were fed weekly by total replacement of medium and additives. Stromal layers achieved confluence after 3 to 5 weeks and were used in the procedures described below. The effect of IFN-a (50 to 5,000 UImL) on From the MRCILRF Leukaemia Unit, Royal Postgraduate Medical School; and the Leukaemia Research Fund Centre, Chester Beatty Laboratories, London, UlC Submitted December 6, 1990; accepted March 21, 1991. Supported by the Medical Research Council and the Leukaemia Research Fund of Great Britain. Address repnnt requests to Charles Dowding, MD, BSc, MRCILRF Leukaemia Unit, Royal Postgraduate Medical School, Hammersmith Hospital, Du Cane Rd, London W12 ONN, UK. The publication costs of thls article were defrayed in part by page charge payment. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. section I734 solely to indicate this fact. 0 1991 by The American Society of Hematology. 0006-4971I9117802-0034$3.0010 499 From www.bloodjournal.org by guest on October 25, 2016. For personal use only. 500 stromal cell colony formation was assayed using a modification of a standard CFU-F assay.” Stromal cultures were initiated and colonies composed of 2 50 cells were scored on day 10. Colonies of cells with fibroblastoid morphology were identified by staining with hematoxylin for 2 minutes. In one representative experiment all colonies stained positively with alkaline phosphatase, confirming a fibroblastoid phenotype.’& Attachment of progenitor cells to stromal layers. Adherent cells were removed from normal BM or CML peripheral blood MNCs by incubating 4 x loh cells in a medium (8 mL) containing FCS (15%) for 2 hours in 25-cm’ tissue culture flasks. The nonadherent MNCs were assayed for blast colony-forming cells (BI-CFC) using the standard assay previously described.” In brief, the nonadherent equivalent of 5 x 10’ MNCs was added to each petri dish culture of stromal cells (three replicate dishes per experimental point) and were incubated for 2 hours. Cells not attached to the stromal cells were decanted and the stromal layers were washed vigorously three times with HBSS, overlaid with a medium (1 mL) containing FCS (15%) and agar (0.3%; Difco Laboratories, Detroit, MI), and incubated at 37°C in 4% CO, in humidified air. Colonies were counted 5 or 6 days later and dishes were desiccated and stained with May-GrunwaldiGiemsa to identify colonies composed of blast cells. In different experiments, stromal cells grown ? IFN-a, f IFN-y, or f enzyme treatment (see below) were used. The binding efficiency of CML BI-CFC was determined by measuring the recovery of BI-CFC following incubation of cell suspensions with stromal layers. Adherent cell-depleted CML MNCs were incubated (‘panned’) on stromal cells (primary) for 2 hours. Unattached cells were removed by sucking the medium up and down 10 times with a Pasteur pipette and were then transferred to a second dish of control stromal cells (secondary) and incubated for a further 2 hours. Blast cell colonies were scored in both primary and secondary cultures and the attachment of CML BI-CFC to primary stromal cells as a proportion of the primary + secondary total was calculated.18 To determine the recovery of normal or CML burst-forming unit-erythroid (BFU-E) and granulocyte-macrophage colonyforming cells (GM-CFC) from stromal cells cultured ? IFN-a (50 U/mL), the nonadherent equivalent of 5 x 16‘ normal or CML MNCs were assayed directly using standard clonogenic techn i q u e ~ ’or~ ‘panned’ for 2 hours on stromal cells before assay. Unattached cells were decanted from the stromal layers and the number of BFU-E and GM-CFC was calculated. Attachment to and culture of long-term culture (LTC)-initiating cells with stromal cells grown f IFN-oL In replicate experiments, MNCs (4 x lob, of density I1.065 g/mL) from four newly diagnosed CML patients were suspended in a medium (8 mL) containing FCS (15%) and were then added to flasks (25 cm2) of irradiated (30 Gy, I3’Cs source) normal allogeneic stromal cells grown f 50 U/mL IFN-a from culture initiation. After 2 hours the flasks were rocked ten times and the nonadherent cells (‘2h non-adh’) were removed, and the remaining cells attached to stromal cells were cultured in LTC medium (8 mL) consisting of a medium containing horse serum (12.5%), FCS (12.5%),methylprednisolone (2 x mol/L) and ? IFN-a (50 U/mL). The ‘2h non-adh’ cells were pelleted, resuspended in LTC medium, and added to identical (secondary) flasks of irradiated stromal cells matched with the first according to exposure to IFN-a. LTCs set up using IFN-a-treated stromal cells were maintained in the presence of IFN-a (50 U/mL), whereas LTCs set up with control stromal cells were maintained without IFN-CY. Cultures were incubated at 33°C in 4% CO, in humidified air. On days 7, 17, and 27 the cultures were demi-depopulated and re-fed with fresh medium plus additives. In addition, cultures were harvested by trypsinization (1:20 in EDTA, 15 to 30 minutes at 37°C; GIBCO) of the DOWDING ET AL adherent layer and the numbers of GM-CFC per flask calculated. The numbers of GM-CFC on day 27 were presumed to reflect the numbers of LTC-initiating cells seeded because at this time few of the original GM-CFC seeded remain.” Generation of Ph-positivehegative GM-CFC. Individual granulocyte-macrophage colonies derived from the 27-day-old LTCs from the same four CML patients studied above were analyzed cytogenetically using a standard technique with minor modifications.*’In brief, on days 7 to 9 of culture a mixture of recombinant granulocyte-macrophage colony-stimulating factor (GM-CSF; 1,000 UlmL; Genetics Institute, Cambridge, MA) and recombinant interleukin-3 (IL-3; 1,000 UlmL; Genetics Institute) in a 50 p L volume was added to a 35-mm dish containing the colonies. After 24 hours incubation at 37”C, colcemid (50 pL, 5 pg/mL; GIBCO) was added dropwise to the culture and was incubated for 2 to 4 hours at 37°C. Individual colonies were then plucked from the culture and resuspended in a droplet of KCI (0.075 mol/L) placed on a poly-L-lysine-treated slide. After 30 minutes the colony was fixed in methanol-acetic acid (3:1)2’and stained with Giemsa stain and the chromosomes were analyzed. For ‘direct’ cytogenetic analysis of cells from LTC, colcemid (40 pL, 50 pg/mL) was added to the culture, and after 4 hours the cells were pelleted and resuspended in KCI (0.085 mol/L). After 30 minutes of incubation at 37”C, the cells were pelleted, fixed in methanokacetic acid (3:1), and stained with Giemsa stain. Sulfated GAG in stromal layers. To estimate the amount of sulfated GAG in stromal layers grown ? IFN-a or ? IFN-y, confluent cultures (in 25-cmZflasks or 35-mm dishes) were incubated for 72 hours in fresh culture medium f IFN-a (50 to 500 UlmL) or IFN-y (25 to 100 UlmL) containing [”SI sulfate (50 pCi/mL. specific activity 20 to 40 Ciimmol; Amersham, Aylesbury, UK). The medium was then decanted and the cell layer was rinsed three times in a medium and incubated in fresh medium for a further 4 hours at 37°C to chase free [’’SI. The stromal cells were washed three times, air dried, and the sides were cut off the culture vessel. The surface to which the stromal cells were attached was then placed against an X-ray film (X-Omat; Kodak, Rochester, NY) within a cassette. The autoradiographs were developed after 4 to 12 hours. The intensity of the film blackening gave an indication of the [”SI sulfate incorporation by GAGs in the stromal layer because 90% of sulfate is used in sulfation of GAG molecules2* with only a small proportion incorporated into glycoprotein molecules. ’’Autoradiographs were scanned using a light densitometer (Chromoscan 3; Joyce-Loebl, Gateshead, UK). Quantitation of [”SI sulfate incorporation into the stromal cell layer was also measured using G25 gel filtration. Fractions were counted on an LKB 1217 Rackbeta liquid scintillation counter (Pharmacia, Uppsala, Sweden) . The amount of incorporated isotope was determined from the isotope in the first of the two peaks. Role of GAGs or neuraminic acid in binding CML BI-CFC. Stromal layers (grown f 50 UimL IFN-a) were incubated with neuraminidase (isolated from Clostridium perfringens, 0.1 U/mL; Sigma), heparanase ( 5 UlmL Sigma), or chondroitinase ABC (0.1 U/mL; Sigma, St Louis, MO) dissolved in a medium, for 30 minutes at 37°C. Control cultures were incubated with CY medium alone. The stromal cells were rinsed three times in a medium and used in the B1-CFC assay (above). Statistics. Dr R. Szydlo (MRC/LRF Leukaemia Unit) performed the statistical analyses. A Student’s t-test was used to compare differences between colony numbers under various culture conditions. A paired t-test was used to evaluate the effect of IFN-a on the binding of LTC-initiating cells to stromal cells. The IFN-a dose-response curve was evaluated using an analysis of variance with repeated measures. A xz test was used to compare the effect of IFN-a on generation of Ph-negative GM-CFC in LTC. From www.bloodjournal.org by guest on October 25, 2016. For personal use only. IFN-a ON STROMNPROGENITOR INTERACTIONS 501 RESULTS Effect of IFN-a and IFN-y on stromal cell capacity to bind progenitor cells. The formation of blast cell colonies on IFN-a-treated (50 U/mL) stromal cells by CML (169.5% ? 21.5% of the control, mean ? SEM, P < .05; Fig lA), but not by normal, Bl-CFC (104.7% ? 10.2%) was increased. At higher IFN-a concentrations, there was a non-significant increase in colony formation by both normal and CML Bl-CFC (data not shown). We sought to confirm that the peak in CML blast colony numbers detected on IFN-a (50 U/mL)-treated stromal cells resulted from increased numbers of Bl-CFC attaching rather than an increased ability of stromal cells to support colony formation. In 'panning' experiments comparing binding to IFN-atreated and -untreated stromal cells, the proportion of CML BI-CFC attaching to IFN-a-treated stromal cells in 2 hours was 81% compared with 40% attachment to untreated stromal cells (Fig 1B). In contrast, blast colony formation by CML Bl-CFC on IFN-y-treated (25 to 100 U/mL) stromal cells showed a dose-dependent decrease (Fig 1C). At 100 U/mL of IFN-y, blast colony numbers were reduced to 37.1% of the control (P < .05, n = 3) compared with the twofold increase found in the presence of IFN-a. The effects of IFN-a on progenitor cell binding were further tested in the LTC system. The presence of IFN-a "1 A i Normal CML - IFN-a + IFN-a B f C D 0 Days Fig 2. Numbers of GM-CFC/flask in four LTCs of CML peripheral blood cells of density less than 1.065 g/mL seeded onto irradiated normal BM stromal cells. See Materials and Methods for culture details. ( 0 )Control (no IFN-a). ( 0 ) IFN-a (50 U/mL). + (50 U/mL) during the period of LTC had no significant effect on the relative numbers of LTC-initiating cells detected for any of the four patients studied (Fig 2 and see below). Using cells from these four CML patients, the mean proportion of LTC-initiating cells (d I1.065 g/mL) that bound to IFN-a-treated stromal cells (50 U/mL) during 2 hours of incubation was 0.59 compared with 0.38 on control stromal cells (P < .OS, n = 4; Fig 1D). Stromal cells cultured with IFN-a (50 U/mL) slightly increased the attachment, though not significantly,of CML GM-CFC or BFU-E during a 2-hour 'panning' procedure (89.0% rt 12.1% and 85.8% f 10.3% of numbers recovered from control stromal cells, respectively; mean rt SEM, n = 5 for both). Effect of IFN-a (50 UImL) on proliferation of CML cells in LTC. The total numbers of GM-CFC, in the 2h-adherent and -nonadherent fractions combined, after 27 days of LTC of four CML samples are shown in Fig 2. There was little or no effect on the total numbers of GM-CFC detected after 27 days of LTC despite the altered distribution of LTCinitiating cells between the 2h-adherent and -nonadherent cultures. The generation of Ph-negative GM-CFC colonies in 27-day-old LTCs grown ? IFN-a (50 U/mL) is shown in Table 1. In three of four patients, only Ph-positive GMCFC were detected before LTC. In the fourth patient a small proportion of Ph-negative GM-CFC were detected 1 -IFN-u +IFNU Fig 1. Binding of primitive progenitor cells to stromal cells grown f IFN-a or 2 IFN-y during 2 hours of incubation. (A) Blast colony formation by normal BM MNCs ( 0 )(mean f SEM, range 22 to 49/1@ MNCs on control stromal cells, n = 7) or CML peripheral blood MNCs (0) (mean f SEM, range 44 to 121/106 MNCs on control stromal cells, n = 13) on stromal cells grown with IFN-a(50 U/mL). (B) Recovery of CML blast colony forming cells from stromal cells grown f IFN-a (mean f SEM, range 51 to 131/106 MNCs on control stromal cells, n = 3). (C) CML blast colony formation on stromal cells grown with IFN-y (mean f SEM, range 37 to 120/10* MNCs on control stromal cells, n = 3). (D) Relative number of LTC-initiating cells binding to stromal cells grown with IFN-a (50 U/mL). (*)P < .05 level of significance compared with control. Table 1. The Effect of IFN-a on the Generation of Ph-Negative GM-CFC in LTC of CML Peripheral Blood % Ph-Negative GM-CFC (coloniesanalyzed) Patient No. 1 2 3 4 PreCulture 0 4.2 0 0 (78) (24) (30) (46) Post-Culture* - IFN-a + IFN-a 4.1 (73) 6.3 (206)t 0 (47) 0 (52) 6.0 (50) 7.5 (187)t See Materials and Methods for culture details. "Day 27 of LTC. tOnly direct metaphases analyzed. 0 (64) 5.3 (75) From www.bloodjournal.org by guest on October 25, 2016. For personal use only. 502 DOWDING ET AL I II 111 Fig 3. Effect of IFN-a or IFN-y on stromal cells. (A) Colony formation by normal B M stromal cells grown r IFN-a (mean f SEM, range 20 to 42 colonies/ lo6 B M MNCs, n = 3). (6) [35S]sulfate incorporation by stromal cells grown f IFN-a in 25-cmZflasks. I, no IFN-a; II, 50 U/mL; 111, 100 U/mL; IV, 500 UlmL. Film exposed for 8 hours. (C) ['5S] sulfate incorporation by stromal cells grown f IFN-y in 35-mm petri dishes. 1, no IFN-y; II, 25 U/mL; 111.50 U/mL; IV, 100 U/mL. Film exposed for 4 hours. See Materials and Methods for culture and metabolic labeling details. IV IFN-N (UlmL) A B C before LTC. After LTC without IFN-a, a small proportion of Ph-negative GM-CFC (one patient) and Ph-negative metaphases in an experiment where colony analysis failed (one patient) were identified. In two patients, no Phnegative GM-CFC were found. In cultures seeded on IFN-a-treated stromal cells and grown in the presence of IFN-a, the proportion of Ph-negative GM-CFC or metaphases detected was increased over that in control cultures. In patient 4, whose preculture cells and postcontrol culture cells contained no Ph-negative GM-CFC, there were 5.3% Ph-negative GM-CFC fotlowing LTC with IFN-a. Overall, the differences did not achieve statistical significance, but in no case was the number of Ph-negative GM-CFC lower in cultures grown on IFN-a-treated stromal cells. Effect of IFN-a and IFN-y on stromal cell proliferation and GAG synthesis. Normal BM-derived stromal cell colony formation was unaffected by culture with IFN-a at concentrations I100 U/mL (Fig 3A). At concentrations of 2 100 U/mL there was a dose-dependent reduction in stromal cell colony numbers (P < .02). Representative autoradiographs showing ["SI sulfate incorporation by stromal layers cultured with IFN-a or IFN-y are shown in Fig 3B and C. The intensity of film blackening corresponded to the amount of sulfate incorporated into GAG. Figure 3B shows that at 50 to 100 U/mL, IFN-a increased the amount of sulfated GAG in the stromal layer compared with controls, whereas at higher concentrations the amount of sulfated GAG was reduced in parallel with inhibition of proliferation (Fig 3C). Densitometric scanning of the autoradiograph at 50 U/mL IFN-a showed a 39% increase in intensity compared with the control. Quantitation using G25 chromatography indicated a mean 25% increase in sulfated GAG incorporation into the stromal layer (Table 2). Incubation of stromal cells with IFN-y (25 to 100 U/mL) did not appear to alter the amount of sulfated GAG in the stromal layer (Fig 3C). Involvement of stromal cell-derived GAGs and neuraminic acid on attachment by CML Bl-CFC. Stromal layers grown 2 IFN-a (50 U/mL) were stripped of sulfated GAGs and neuraminic acid to see if there was any relationship between the effects of IFN-a on these matrix components and its effect on CML BI-CFC binding. The results of these experiments are shown in Fig 4. Neither heparanase nor chondroitinase ABC treatment of stromal cells had a significant effect on attachment of CML Bl-CFC to stromal cells grown 2 IFN-a, but neuraminidase treatment of control stromal cells (ie, without IFN-a) increased blast colony formation by CML cells (P < .05; Fig 4). Thus, the enhancing effect of IFN-a on CML binding to stromal cells was reproduced by treating control stromal cells with neuraminidase. Consequently, the results are consistent with IFN-a acting via modulation of the neuraminic acid content in the hematopoietic microenvironment. Table 2. p5S]Sulfate Incorporation Into Stromal Layers Grown f IFN-a Counts Per Minute Experiment No. Control (no IFN-N) IFN-N (50 UlmL) 1 13,910 2 77,967 3 54.91 7 4 15,884 18,079 (130.0%)' 80,871 (103.7%) 68,519 (124.8%) 22,492 (141.6%) Background counts less than 30. See Materials and Methods for labeling details. *Percentage of control (no IFN-a) counts. From www.bloodjournal.org by guest on October 25, 2016. For personal use only. IFN-CXON STROMNPROGENITOR INTERACTIONS * * T,* T * cho 503 *** neu Fig 4. Blast colony formation by CML peripheral blood MNCs on stromal cells grown 2 IFN-a (50 UlmL) and 2 treatment: hep, heparanase (n = 3); cho. chondroitinase (n = 2); neu, neuraminidase (n = 4). Mean values k 1 standard deviation. Mean and range in control colony numbers were 82 and 66 to 118/14 MNCs, respectively. ("I P < .05 level of significance compared with 'no IFN-a' control. See Materials and Methods for conditions for enzyme treatment. (Dl no IFN-a; (8)no IFN-a treatment;).( + IFN-a;) . ( IFN-a + treatment. + + DISCUSSION A selective anti-leukemic activity of IFN-a has been identified in a significant proportion of patients with CML',' but there is no evidence that this is due to a direct cytostatic/cytotoxiceffect on the leukemic clone. We considered, therefore, that in addition to exerting a nonspecific antiproliferative effect on normal and malignant cells, IFN-a might act by modifying the BM microenvironment. The microenvironment consists of a complex network of cells with their associated extracellular matrix and growth factors and is thought to play an active role in the regulation of normal hematopoiesis. However, the Ph-positive stem cell population in CML escapes this control.% We have previously shown that a population of primitive progenitor cells (Bl-CFC) found in the BM and blood of patients with CML is deficient in its ability to bind to BM-derived stromal cells. Moreover, CML Bl-CFC are also less discriminatory than normal Bl-CFC because normal cell binding is restricted to stromal cells grown with methylprednisolone, whereas CML cells bind, albeit inefficiently, to stromal cells grown with or without methylprednisolone.' This abnormal attachment of CML primitive progenitor cells could underlie some of the pathologic features such as: (1) lack of responsiveness to microenvironment-derived control factors,% (2) inappropriate release from the BM, and (3) extramedullary hematopoiesis. It follows that any maneuver that tends to normalize the adhesive interaction between CML progenitor cells and microenvironmental cells might reinstate a more physiologic regulation of leukemic hematopoiesis. In this study we showed that IFN-a, but not IFN-y, increased the attachment of two different types of CML primitive progenitor cells (Bl-CFC and LTC-initiating cells) to stromal cells. The increased efficiency in binding by CML BI-CFC approached the efficiency reported for normal Bl-CFC of which greater than 90% bind to stromal cells.' In this previous study designed to determine the rate of attachment of B1-CFC to stromal cells using serial 2-hour panning steps, approximately 90% of normal BM-derived BI-CFC attached to stromal cells during the first 2-hour incubation step. However, apparently only about 40% of the total CML B1-CFC adhered to stromal cells during the first 2-hour panning step because progressively decreasing numbers of B1-CFC could be detected by re-panning the nonadherent cell fraction on stromal cells for further 2-hour incubations. Blast colony-forming cells from CML patients did not have a different binding rate to stromal cells because maximal binding occurred after 2 Instead, once bound, a time-dependent increasing proportion of CML B1-CFC detached upon further incubation in liquid culture at 37"C, unlike normal Bl-CFC, which remained bound. This result suggests that a kinetic equilibrium between adhesion and detachment favors detachment for CML Bl-CFC and attachment for normal Bl-CFC. The enhanced binding was most marked at an IFN-a concentration of 50 U/mL, which would fall within the range of concentrations achieved in CML patients treated with IFN-a. This concentration had no inhibitory effect on fibroblast colony formation and little or no significant effect on the proliferation of CML LTC-initiating cells or other hematopoietic cells? In contrast to primitive progenitor cells, the attachment of CML GM-CFC and BFU-E to IFN-a-treated stromal cells during a similar 2-hour 'panning' step (previously shown to be <15% of normal progenitor cells") was not significantly increased. However, we have also recently found that IFN-a (100 U/mL) greatly increased the number of adherent GM-CFC in a 10-day liquid co-culture of CML BM cells with confluent normal stromal cells,26suggesting that under different conditions the attachment or release of lineage-committed progenitor cells might also be enhanced by IFN-a. In two of four patients, between 4% to 8% of the circulating LTC-initiating cells were Ph-negative as previously reportedz4 and in three of four patients IFN-a increased, though not significantly, the proportion of Phnegative GM-CFC detected on day 27. These IFN-amediated increases in Ph-negative GM-CFC detected could reflect differences in the responses of Ph-positive and Ph-negative LTC-initiating cells to the direct effect of IFN-a or could reflect an effect through modification of the interaction with stromal cells. If the modification of primitive progenitor cell-microenvironmental cell interactions by IFN-a plays a role in the anti-leukemic process, microenvironment-derived negative regulators of hematopoiesis, which might include stem cell inhibitory factor," transforming growth factor+,= or tumor necrosis factor-a,29might then annul the proliferative advantage usually expressed by leukemic cells. The cell adhesion molecules (CAMS) and their receptors involved in attachment of primitive progenitor cells to BM microenvironmental cells are poorly characterized. He- From www.bloodjournal.org by guest on October 25, 2016. For personal use only. 504 DOWDING ET AL m~nectin,~' fibronectin," thrombospondin;' and ad@, integrin33may have a role in binding primitive progenitor cells. Heparan sulfate, found in the ECM of stromal cell cultures," has been implicated in binding normal BI-CFC to stromal cells."' Chondroitin sulfate is the major GAG in r ~ d e n t ~and '.~~ human' stromal cell-derived ECM and has been identified in the membrane of murine factordependent cloned progenitor cells?' Roles for chondroitin sulfate in the regulation of proliferation38s39 and the process of detachment of progenitor cells from stromal cells" have been proposed. We therefore investigated whether (1) stromal cell-derived heparan sulfate or chondroitin sulfate was involved in the attachment by CML BI-CFC to stromal cells, and (2) whether the increased attachment of CML BI-CFC to IFN-a-treated stromal cells was due to an alteration in the synthesis or availability of sulfated GAG. We found that IFN-a increased the total amount of sulfated GAG (heparan sulfate + chondroitin sulfate) but enzymatic removal of either GAG from stromal cells did not significantly affect CML BI-CFC attachment, suggesting that sulfated GAGS do not have a primary role in the attachment of CML BI-CFC to stromal cells. Neuraminic acid (sialic acid) is a negatively charged terminal nonreducing sugar important in some hematopoietic cell-stromal cell intera~tions.''~'~ Stromal cells of LTCs can bind neuraminic acid: but its function within the stromal layer is unknown. A role for neuraminic acid in the IFN-a-induced changes in binding was suggested by results of enzymatic stripping of neuraminic acid from control stromal cells. This resulted in a significant increase in CML attachment but had no significant effect on the ability of IFN-a-treated stromal cells to bind CML BI-CFC. Whether IFN-a qualitatively modifies or reduces the neuraminic acid content in stromal cell ECM remains to be determined. The influence of neuraminic acid and the lack of involvement of heparan sulfate in the improved binding by CML cells to stromal cells indicates that IFN-a probably does not restore the normal cell adhesion mechanism because this would be heparan sulfate-dependent. Instead, IFN-a might facilitate the operation of a different mechanism of adhesion because cells express a variety of cell adhesion molecules with apparently overlapping functions."' Also, we have recently demonstrated that CML BI-CFC lack a phosphatidylinositolanchored CAM43that is an essential requisite for efficient and selective binding by normal BI-CFC. It is possible that expression of CAMS by hematopoietic progenitor cells controls their precise location in the marrow microenvironment so that they are properly regulated." The operation of a different cell adhesion mechanism by CML Bl-CFC, in the context of stromal cells exposed to IFN-a, raises the possibility that the immobilized leukemic cells could be placed at a proliferative disadvantage because they are subjected to inappropriate rather than normal control mechanisms. In summary, we have presented evidence for a functional difference in the effect of IFN-a on normal and CML progenitor cells that may prove to be associated with the finding of IFN-a-induced normalization of nuclear proteinDNA complexes observed in IFN-a-sensitive leukemic cells.@ To a great extent IFN-a overcame the deficient attachment of CML BI-CFC to stromal cells and increased the proportion of LTC-initiating cells that attached to stromal cells. This effect was associated with changes in the neuraminic acid content of the stromal layer. It is possible that the mechanism of action of IFN-a in CML patients might include modification of BM microenvironmental cells such that the interaction between hematopoietic cells and stromal cells is enhanced; this could lead to suppression of the leukemic clone and thus allow re-emergence of normal hematopoiesis. REFERENCES 1. Talpaz M, Kantarjian H, McCredie K, Trujillo J, Keating M, Gutterman J: Hematologic remission and cytogenetic improvement induced by recombinant human interferon alpha A in chronic myelogenous leukemia. N Engl J Med 314:1065,1986 2. Niederle N, Kloke 0, Osieka R, May D, Wandl UB, Becher R, Opalka B, Schmidt CG: Treatment of chronic and acute phase chronic myelogenous leukemia with interferon-% and interferon-y, in Huhn D, Hellriegel KP, Niederle N (eds): Chronic Myelocytic Leukemia and Interferon: Pathophysiological, Clinical and Therapeutic Aspects. New York, NY,Springer-Verlag, 1988, p 119 3. Galvani D, Cawley C: Mechanism of action of a interferon in chronic granulocytic leukaemia: Evidence for preferential inhibition of late progenitors. Br J Haematol73:475,1989 4. Dowding CR, Gordon MY, Goldman JM: The effect of recombinant alpha interferon on normal and chronic myeloid leukaemia (CML) progenitor cells. Exp Hematol 15:32a, 1987 (abstr) 5. Herrmann F, Lindemann A, Otto J, Mertelsmann R: In vivo effects of interferon-gamma in chronic myelogenous leukemia, in Huhn D, Hellriegel KP, Niederle N (eds): Chronic Myelocytic Leukemia and Interferon: Pathophysiological, Clinical and Therapeutic Aspects. New York, NY,Springer-Verlag, 1988, p 113 6. Russo D, Pecile V, Zuffa E, Gallizia C, Grazia-Michieli M, Damiani D, Testoni N, Pecile V, Visani G, Colombini R: Treatment of Ph+ chronic myeloid leukemia by gamma interferon. Blut 59:15,1989 7. Talpaz M, Kantarjian H, Kurzrock R, Gutterman J: Clinical studies with interferons in Philadelphia positive chronic myelogenous leukemia. Bone Marrow Transplant 4:63a, 1989 (abstr) 8. Gordon MY, Dowding C, Riley G, Goldman JM, Greaves M: Altered adhesive interactions with marrow stroma of haematopoietic progenitor cells in chronic myeloid leukaemia. Nature 328:342, 1987 9. Zuckerman KS: Composition and function of extracellular matrix in the stroma of long-term marrow cultures, in Wright DG, Greenberger JS (eds): Long-Term Bone Marrow Culture. New York, NY,Liss, 1984, p 157 10. Gordon MY, Riley GP, Clarke D: Heparan sulfate is necessary for adhesive interactions between human early hemopoietic cells and the extracellular matrix of the marrow microenvironment. Leukemia 2804,1988 11. Roberts R, Gallagher J, Spooncer E, Allen TD, Bloomfield F, Dexter TM: Heparan sulphate bound growth factors: A mechanism for stromal cell mediated haemopoiesis. Nature 332:378,1988 12. Tonelli Q, Meints R H Surface molecules of hematopoietic From www.bloodjournal.org by guest on October 25, 2016. For personal use only. IFN-a ON STROMNPROGENITOR INTERACTIONS cells: Requirement for sialic acid in spleen colony formation. Science 195:897,1977 13. Kapsenberg ML, Stiekema FEM, Kallan A, Bos JD, Roozemond RC: The restrictive role of sialic acid in antigen presentation to a subset of human peripheral CD4’ T lymphocytes that require antigen-presenting dendritic cells. Eur J Immunol 19:1829, 1989 14. Boyum A Isolation of mononuclear cells and granulocytes from human blood. Scand J Clin Lab Invest 21:7731, 1968 (suppl 97) 15. Castro-Malaspina H, Gay RE, Resnick G, Kapoor N, Meyers P, Chiarieri D, McKenzie S, Broxmeyer H, Moore MAS: Characterization of human bone marrow fibroblast colony-forming cells (CFU-F) and their progeny. Blood 56:289,1980 16. Liesveld JL, Abboud CN, Duerst RE, Ryan DH, Brennan JK, Lichtman MA: Characterization of human marrow stromal cells: Role in progenitor cell binding and granulopiesis. Blood 73:1794, 1989 17. Gordon MY, Hibbin J, Keamey L, Gordon-Smith E, Goldman JM: Colony formation by primitive haemopoietic progenitor cells in cocultures of bone marrow cells and stromal cells. Br J Haematol660:129,1985 18. Gordon MY, Hibbin JA, Dowding C, Gordon-Smith EC, Goldman JM: Separation of human blast progenitors from granulocytic, erythroid, megakaryocytic and mixed colony-forming cells by ‘panning’ on cultured marrow-derived stromal layers. Exp Hematol 13:937,1985 19. Dowding CR, Gordon MY, Goldman JM: Increased DR antigen expression by committed progenitor cells but not primitive progenitor cells in chronic myeloid leukemia. Exp Hematol 16:171, 1988 20. Andrews RG, Takahashi M, Segal GM, Powell JS, Bernstein ID, Singer JW: The LIF3 antigen is expressed by unipotent and multipotent colony-forming cells but not by there precursors. Blood 68:1030,1986 21. Amenomori T, Tomonaga M, Matsuo T, Yoshida Y, Kuriyama K, Sadamori N, Ichimaru M: A micromethod for chromosome preparation from individual hematopoietic colonies cultured in methyicellulose. Int J Cell Cloning 3:133,1985 22. Beeley JG: Glycoprotein and proteoglycan techniques, in Burdon RH, van Knippenberg PH (eds): Laboratory Techniques in Biochemistry and Molecular Biology. New York, NY, Elsevier, 1985, p 372 23. Hortin G, Green ED, Baenziger JU, Strauss A W Sulphation of proteins by a hepatoma-derived cell line. Biochem J 235:407,1986 24. Eaves AC, Cashman JD, Gaboury LA, Kalousek DK, Eaves CJ:Unregulated proliferation of primitive chronic myeloid leukemia progenitors in the presence of normal marrow adherent cells. Proc Natl Acad Sci USA 83:5306,1986 25. Gordon MY, Dowding CR, Riley GP, Goldman JM, Greaves MF: Adhesive defects in chronic myeloid leukaemia: Mechanisms of myeloid tumorigenesis. Curr Top Microbiol Immunol 149:151, 1989 26. Osterholz J, Dowding C, Guo A-P, Siczkowski M, Goldman JM: Alpha-interferon alters the distribution of CFU-GM between the adherent and non-adherent compartments in long-term culture of Chronic myeloid leukemia. Exp Hematol (in press) 27. Graham GJ, Wright EG, Hewick R, Wolpe SD, Wilkie NM, Donaldson D, Lorimore S , Pragnell IB: Identification and charac- 505 terization of an inhibitor of haemopoietic stem cell proliferation. Nature 344:442,1990 28. Eaves AC, Eaves CJ: Maintenance and proliferation control of primitive hemopoietic progenitors in long-term culture of human marrow cells. Blood Cells 14:355, 1988 29. Pelus LM, Ottmann OG, Nocka K Synergistic inhibition of human CFU-GM progenitor cells by prostaglandin E and recombinant interferon a, and an effect mediated by TNF. J Immunol 140:479,1988 30. Campbell AD, Long MW, Wicha MS: Haemonectin, a bone marrow adhesion protein specific for cells of granulocytic lineage. Nature 329:744,1987 31. Patel VP, Lodish HF: The fibronectin receptor on mammalian erythroid precursor cells: Characterization and developmental regulation. J Cell Biol102449,1986 32. Long MW, Dixit VM: Thrombospondin functions as a cytoadhesion molecule for human hematopoietic progenitor cells. Blood 75:2311,1990 33. Verfaillie C, McCarthy J, McGlave P: Differentiation of primitive human kin-, 34+, DR- hematopoietic progenitors is accompanied by transition of adhesion from the heparin-binding domain to the cell-binding domain of fibronectin (FN). Blood 76:491a, 1990 (abstr, suppl) 34. Gallagher JT, Lyon M, Steward WP: Structure and function of heparan sulphate proteoglycans. Biochem J 236:313,1986 35. Gallagher JT, Spooncer E, Dexter M: Role of the cellular marrow haemopoiesis. I. Synthesis of glycosaminoglycans in mouse bone marrow cell culture. J Cell Sci 63:155,1983 36. Oguri K, Okayama E, Caterson B, Okayama M: Isolation, characterization and localization of glycosaminoglycans in rabbit bone marrow. Blood 70:501,1987 37. Minguell JJ, Tavassoli M: Proteoglycan synthesis by hematopoietic progenitor cells. Blood 73:1821,1989 38. Hart GW: Biosynthesis of glycosaminoglycans by thymic lymphocytes. Effect of mitogenic activation. Biochemistry 21:6088, 1982 39. Spooncer E, Gallagher JT, Krizsa F, Dexter TM: Regulation of haemopoiesis in long-term bone marrow cultures. IV. Glycosaminoglycan synthesis and the stimulation of haemopoiesis by p-Dxyloside. J Cell Biol96:510,1983 40. Del Rosso M, Appelletti R, Dini G, Fibbi G, Vannucchi S, Chiarugi V, Guazzeli C: Involvement of glycosaminoglycans in detachment of early myeloid precursors from bone marrow stromal cells. Biochim Biophys Acta 676:129,1981 41. Gabius S, Gabius H-J: Sugar receptors of the stromal cell layer in human long-term bone marrow cultures: Their presence, modulatory responses to changes in the microenvironment and potential role in cellular adhesion. Blut 61:232, 1990 42. Hynes RO: A family of cell surface receptors. Cell 38:549, 1987 43. Gordon MY, Ford AM, Greaves MF: Cell interactions and gene expression in early hematopoiesis. Int J Cell Cloning 8:11, 1990 (suppl 1) 44. Howard OMZ, Talpaz M, Kantarjian H, Seong D, Wedrychowski A, Paslidis N, Hester J, Cork A, Turpin J, Lopez-Berestein G, Trujillo JJ, Gutterman J, Freireich E, Deisseroth A Interferon affects nuclear proteins in cells of clinically sensitive chronic myelogenous leukemia patients. Blood 76:1117,1990 From www.bloodjournal.org by guest on October 25, 2016. For personal use only. 1991 78: 499-505 Interferon-alpha overrides the deficient adhesion of chronic myeloid leukemia primitive progenitor cells to bone marrow stromal cells C Dowding, AP Guo, J Osterholz, M Siczkowski, J Goldman and M Gordon Updated information and services can be found at: http://www.bloodjournal.org/content/78/2/499.full.html Articles on similar topics can be found in the following Blood collections Information about reproducing this article in parts or in its entirety may be found online at: http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests Information about ordering reprints may be found online at: http://www.bloodjournal.org/site/misc/rights.xhtml#reprints Information about subscriptions and ASH membership may be found online at: http://www.bloodjournal.org/site/subscriptions/index.xhtml Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036. Copyright 2011 by The American Society of Hematology; all rights reserved.