Endothelial cell support of hematopoiesis is differentially ... - Nature

Leukemia (1998) 12, 1210–1220  1998 Stockton Press All rights reserved 0887-6924/98 $12.00 http://www.stockton-press.co.uk/leu

Endothelial cell support of hematopoiesis is differentially altered by IL-1 and glucocorticoids B Jazwiec1,2, A Solanilla1, C Grosset1, F-X Mahon1, M Dupouy1, V Pigeonnier-Lagarde1, F Belloc1, K Schweitzer3, J Reiffers1 and J Ripoche1 1

Laboratoire de greffe de moelle, UMR 5540, Universite´ de Bordeaux II, Bordeaux, France, and 3Free University Hospital, Department of Hematology, Amsterdam, The Netherlands

We investigated the ability of endothelial cells (EC) to support hematopoiesis in contact and non-contact cocultures with isolated CD34+ or CD34+/CD38low cells. In the absence of exogenous cytokines, umbilical vein EC (HUVEC) efficiently support proliferation of hematopoietic cells and generation of colonyforming cells (CFC). Cytokines (IL-6, LIF, G-CSF, GM-CSF, MCSF, but not IL-1, IL-3, IL-7) were detected in HUVEC coculture supernatants. Neutralization of these cytokines profoundly inhibited the ability of EC supernatants to support the differentiation of hematopoietic progenitors and led to an accumulation of immature cells. Contact cocultures were significantly more efficient than non-contact cocultures. The expanded cell population essentially belonged to the myeloid and monocytic lineages. Contact cocultures generated cells expressing the CD61 or CD41 antigens. Interleukin-1␣ (IL-1␣) augmented EC capacity to support hematopoiesis, this property resulting from the upregulation of cytokine expression. Glucocorticoids (GC) reduced this capacity by downregulating the biosynthesis of cytokines by EC and not by a direct effect on the progenitor cells. EC from the bone marrow microvasculature (BMEC) support the proliferation and the differentiation of hematopoietic progenitors. Synergistic increase in progenitor cells expansion and generation of CFC occurred when EC cocultures were added with exogenous cytokines. Supernatants of IL-1␣-stimulated EC potentiated the effects of an association of IL-1, IL-3, IL-6, LIF, SCF, Flt3-ligand, TPO, G-CSF, GM-CSF, M-CSF and IL11 on the proliferation of hematopoietic progenitors suggesting that EC may produce other soluble growth factors potentiating the action of the above set of cytokines. Keywords: hematopoiesis; endothelial cell; cytokine; interleukin1; glucocorticoids

Introduction The generation of hematopoietic cells requires the coordinate interplay of many regulatory molecules which are secreted and/or remain cell associated in the bone marrow. Importance for stromal cells in the production of these molecules has been deduced from in vitro experiments aiming at mimicking the bone marrow microenvironment such as long-term bone marrow cultures (LTBMC).1 The differentiation and proliferation of hematopoietic progenitor cells critically relies on a functional stromal layer.2–6 A direct contact between hematopoietic progenitors and the stromal layer is not an absolute condition for the proliferation and differentiation of hematopoietic progenitors. A non-contact culture system is also efficient.7,8 These experiments demonstrated that soluble factors secreted by the bone marrow

Correspondence: J Ripoche, Laboratoire de Greffe de Moelle, UMR 5540, Universite´ de Bordeaux II, 146, rue Le´o-Saignat, 33 076 Bordeaux, France; Fax: 33 05 56 93 88 83 2 Present adress: Laboratory of Clinical Immunology, Institute of Immunology and Experimental Therapy, Czerska 12, 53-114 Wroclaw, Poland The first two authors contributed equally to this work. Received 4 November 1996; accepted 31 March 1998

stroma were essential to hematopoiesis. Among these are cytokines which have been shown to support in vitro hematopoiesis when they are used in combination.9 The mRNA transcripts for most of these cytokines have been found in stromal cells. Secretion of bioactive molecule by bone marrow stroma cells was demonstrated for some of them.10–16 Other stromaderived growth factors are supposed to be necessary for optimal expansion of hematopoietic progenitors. The bone marrow stroma has a heterogeneous composition. It is made of fibroblasts, EC,17–23 and other cells present in smaller number but with potentially functional importance, such as adipocytes, macrophages and osteoclasts. Such complexity makes the task of identifying which cytokines are secreted by which cell and how their production are regulated difficult. Stroma-derived cell lines have proven to be useful as a simplified model for these studies. Clonal stromal cell lines from the hematopoietic environment, either of human or murine origin, have been shown to be able to support growth of primitive human hematopoietic progenitors.24–34 However, these cell lines may not accurately reflect their in vivo counterparts. One of the major problems associated with cell lines is the possibility of extensive genetic change that they may undergo in culture. This is probably one reason for the large variation observed in the ability of stromal cell lines to support hematopoiesis.35,36 A further limitation for murine cell lines may be the incomplete species cross-reactivity of hematopoietic growth factors. EC have been shown to produce many of the cytokines which are known to play a role in the proliferation and differentiation of hematopoietic progenitors.37–47 The functional importance of this production remains to be fully appreciated. As a direct relevant application, EC may for instance offer an improved alternative to unfractionated stroma cells as a source of feeder cells in bioreactors devised for ex vivo expansion of hematopoietic progenitors. In this report we have studied the capacity of EC to support human hematopoiesis in vitro and have tried to answer the question whether this support was directly and exclusively linked to the production of hematopoietic growth factors using comparative coculture experimental settings and antibody neutralizing experiments. We have analyzed the modulation of this capacity by IL-1 and GC. We show that addition of exogenous cytokines to EC/hematopoietic progenitor cocultures induced a synergistic increase in cell proliferation and generation of CFC resulting in the generation of a large number of clonogenic cells and that EC may produce hitherto uncharacterized soluble factors that potentiate a combination of cytokines including IL-1, IL6, IL-3, IL-11, G-CSF, GM-CSF, M-CSF, flt3-ligand, SCF and TPO. The potential usefulness of EC for protocols aiming at improving the ex-vivo expansion of hematopoietic progenitors is discussed.

IL-1 and GC alter EC capacity to support hematopoiesis B Jazwiec et al

Materials and methods

Cytokines, antibodies and other reagents IL-1␣ and ␤ were gifts from Dr PT Lomedico, Hoffman-La Roche, Nutley, NJ, USA; IL-3 was a gift from Dr Seiler and Dr Fritzsche, Behring, Marburg, Germany; IL-6 and GM-CSF were a gift from Dr AS Blanc, Sandoz, Basel, Switzerland. Stem cell factor (SCF) and G-CSF were gifts from Dr McNiece, Amgen, CA, USA; IL-11 was a gift from Dr Tabah, ScheringPlough, Levallois Perret, France. M-CSF was purchased from R&D Systems, Oxon, UK. LIF was a gift from Dr JF Moreau, Bordeaux, France. Flt3-ligand was gift from Dr SD Lyman, Immunex Corporation, Seattle, WA, USA. TPO was a gift from Dr Foster, Zymogenetics, Seattle, WA, USA. Dexamethasone (DXM) sodium phosphate salt (sterile, apyrogenic solution for human therapeutic use) was purchased from Merck, Sharp and Dohme-Chibret, Paris, France. Polyclonal neutralizing antibodies against cytokines were from R&D systems. Polyclonal neutralizing anti-LIF antibody was a gift from Pr V Praloran, CHU Dupuytren, Limoges, France.

Cord blood and bone marrow specimens Umbilical cord blood was obtained during normal full-term deliveries with the mothers’ informed consent. After placental delivery, the umbilical veins were canulated and aspirated. Blood was collected into physiologic saline containing preservative-free heparine and stored at 4°C. The samples were processed within 12 h of collection. Normal fresh BM samples were collected from individuals undergoing orthopedic surgery or donating BM for allogeneic BM transplantation. Informed consent was obtained from each donor.

Isolation of CD34+ cells and CD34+/CD38low cells Low density mononuclear cells were isolated by layering bone marrow or cord blood over Ficoll–Hypaque (density 1.077g/cm3; Seromed, Biochrom KG, Berlin, Germany). Hematopoietic progenitors expressing the CD34 antigen were obtained by an immunomagnetic isolation system (Isolex 50; Baxter, Maurepas, France), according to the manufacturer’s instructions. The purity of the isolated population was assessed by cytofluorimetry after staining with anti-CD34 monoclonal antibody HPCA-2-FITC (Becton Dickinson, Pontde-Claix, France). Only preparations giving 85% purity or more were used. Isolated cells were then used immediately for coculture with EC or for cell sorting. To isolate CD34+/CD38low cells, a two-color staining of immunomagnetically isolated CD34+ cells was performed with anti-CD34FITC monoclonal antibody (HPCA-2), anti-CD38-PE monoclonal antibody (Leu-7; Becton Dickinson). Cells were incubated for 30 min at +4°C and washed twice. Negative controls were labelled with appropriate isotypic controls. Cells were sorted on a PC 3000 cell sorter (Odam-Brucker, Wissenbourg, France).

EC from the umbilical cord HUVEC were obtained from freshly collected umbilical cords. 0.1% w/v collagenase (Boehringer Mannheim, Meylan,

France) was used for treatment of the umbilical vein instead of 0.2%, as originally described.48 They were cultured in IMDM (Iscove’s modified Dulbecco’s medium; Gibco, Cergy-Pontoise, France) supplemented with 15% heat-inactivated foetal calf serum (FCS), 90 ␮g/ml heparin (Sigma), penicillin (100 ␮g/ml), streptomycin (100 ␮g/ml), amphotericin B (2.5 ␮g/ml). This culture medium is referred to as standard medium. Cells were grown to confluence in 75 or 25 cm2 flasks coated with 50 ␮g/ml collagen I (Jacques Boy Institute, Paris, France). Cells were passaged after treatment with trypsin-EDTA and routinely used between the second and fourth passages. After the second passage, there was no visible contamination by monocytes. Cells were maintained in a humidified atmosphere at 37°C, 5% CO2, and were identified by their characteristic morphology and by the expression of Factor VIII antigen.

EC from the bone marrow microvasculature EC from human bone marrow (BM) microvasculature (BMEC) were isolated following a two-step procedure with BNH-9 or S-Endo-1-coated microspheres. BNH-9 is a mouse IgM monoclonal antibody recognizing EC-specific H and Y antigens (Immunotech, Marseille, France). S-Endo-1 is a monoclonal antibody reactive with a 110 kDa protein expressed at the membrane of HUVEC (Biocytex, Marseille, France). The whole procedure is essentially the same as described by Rafii et al49 and Schweitzer et al.50 Briefly, the BM sample (5–30 ml total) was diluted 1:1 in Buffer A (RPMI containing 5 mM EDTA) and placed on a 40 ␮m Cell Strainer Falcon disposable filter, washed five times with Buffer A while being gently triturated. Mononuclear cells (MNC) were separated from filtrate by standard Ficoll–Hypaque centrifugation at 400 g for 20 min at 24°C, and kept for further isolation as they also contain HBMEC. To avoid binding of the BM fat ladder spicules to the surfaces, plastic wares at this step were treated with 1% BSA in RPMI. The filter was cut out, placed in a 50 ml Falcon tube and digested in 10 ml of 0.1% collagenase for 15 min at 37°C. The digested material was gently passaged several times through a 1 ml tip. This material was washed twice with RPMI containing 2% FCS and then gently triturated to further dissociate the EC. HBMEC were further isolated by positive selection using BNH-9 or S-Endo1 monoclonal antibody. This protocol gives a highly enriched population of BMEC.50 EC were collected, washed once with complete medium and plated in collagencoated dishes or flasks as above, on collagen I coated 6-well Nunc plastic dishes in IMDM containing 20% FCS, 15 ␮g/ml ECGF, 90 ␮g/ml heparin, antibiotics and amphotericine B. Forskolin and IBMX (Sigma, St Quentin Fallavier, France) were added at the beginning of the cultures at final concentrations of 0.5 ␮M, to inhibit fibroblastic growth.51

Experimental procedures: coculture EC/hematopoietic progenitor cells Non-contact cocultures: Endothelial cells were subcultured in 6-well plates (Nunc, Polylabo, Strasbourg, France). Hematopoietic progenitors (2500 cells for CD34+ or 1000– 2000 cells for sorted CD34+/CD38low population) were placed over the EC monolayers in 24-mm Transwell inserts (0.45 ␮m pore size, Costar, Cambridge, MA, USA) in a total volume of 5 ml. The medium used was standard medium with or without

1211


1212

IL-1␣ at 100 IU/ml or DXM at 10−6m final concentrations. Cultures were fed twice a week by removing 2 ml of medium and replacing it with fresh medium with or without IL-1␣ or DXM as indicated. At time intervals, cells were recovered from the inserts by vigorous washing and resuspended in IMDM supplemented with 10% FCS. Cells were counted under a hemocytometer, tested for viability by Trypan blue exclusion, phenotyped and plated in methylcellulose cultures for clonogenic assays. In some experiments a combination of IL-1␣ (100 IU/ml), IL-6 (10 ng/ml), IL-3 (50 ng/ml), IL-11 (10 ng/ml), SCF (50 ng/ml), flt3-ligand (100 ng/ml), TPO (10 ng/ml), GCSF (50 ng/ml), GM-CSF (10 ng/ml), M-CSF (10 ng/ml) was added to the culture medium. For measurements of cytokine concentrations, the coculture medium was recovered at indicated time intervals, centrifuged for 30 min at 50 000 g, 4°C, decanted to a clean tube and frozen at −80°C until further use. Finally, in some experiments, EC were prestimulated for 36 h with IL-1␣ then washed five times to remove IL-1␣ and cocultivated with the hematopoietic progenitors as above. Contact cocultures: In contact cocultures hematopoietic progenitors were directly added on to the EC monolayer in a total volume of 5 ml in the presence or absence of IL-1␣ at a concentration of 100 IU/ml or DXM at a concentration of 10−6M. Cultures were maintained and fed as above. At the indicated time intervals, non-adherent cells were collected for numeration, methylcellulose colony-forming assay, phenotyping, or morphologic analysis. Adherent cells were detached together with the EC monolayer by trypsinization, quantified under a hemocytometer and analyzed for their clonogenic progenitor content by methylcellulose colony-forming assay. Long-term cocultures: For long-term cocultures, the EC together with the adherent population of progenitors were trypsinized and reseeded on a confluent fresh layer of EC. Routinely, cells were trypsinized every 10 days to 2 weeks. Hematopoietic progenitor proliferation in the presence of EC supernatants: Isolated CD34+ progenitors were grown in liquid culture in the presence of 25% EC supernatants in DMEM supplemented with 15% FCS and antibiotics. In some experiments, a combination of IL-1␣ (100 IU/ml), IL-6 (10 ng/ml), IL-3 (50 ng/ml), IL-11 (10 ng/ml), SCF (50 ng/ml), flt3-ligand (100 ng/ml), TPO (10 ng/ml), G-CSF (50 ng/ml), GM-CSF (10 ng/ml), M-CSF (10 ng/ml) was added to the culture medium. Cytokine neutralization experiments: For inhibition studies with neutralizing anti-cytokine antibodies a mixture of neutralizing polyclonal antibodies directed against IL-1␣, IL1␤, IL-6, IL-11, G-CSF, GM-CSF, M-CSF, SCF and LIF was added to the culture medium. Experiments were performed in 6-well plates. Final concentrations for each neutralizing antibody was 5 ␮g/ml. Antibodies were readded at the same final concentrations when refeeding the coculture. Neutralizing polyclonal antibodies directed against IL-1␣, IL-1␤, IL-6, IL11, G-CSF, GM-CSF, SCF and M-CSF from R&D were Cat No. AB-200-NA, AB-201-NA, AB-206-NA, AB-218-NA, AB-214NA, AB-215-NA, AB-255-NA and AB-216-NA, respectively. Anti-SCF antibody is directed against the soluble form of SCF.

This mixture of antibodies when used in the range of 5 ␮g/ml final for each antibody, gave efficient neutralization of a combination of recombinant human IL-1␣, IL-1␤, IL-6, IL-11, GCSF, GM-CSF, M-CSF, SCF and LIF at concentrations of 50 ng/ml each.

Preparation of LTBMC and other cells For experiments aimed at comparing EC and human bone marrow stroma for their ability to support proliferation and generation of CFC, human marrow adherent cells were derived from 4 to 6 week-old human LTBMC. LTBMC were established in 25 cm2 tissue culture flasks from light-density bone marrow cells. Culture medium was IMDM supplemented with 12.5% FCS, 12.5% horse serum, penicillin streptomycin (100 ␮g/ml), kanamycin (100 ␮g/ml), (100 ␮g/ml), amphotericin B (2.5 ␮g/ml) and 10−7 M hydrocortisone (HC) (Sigma). Semi-confluent stromal layers were trypsinized, replated in 6-well plates and grown until confluency. Cells were then washed five times to remove HC and used for non-contact coculture with hematopoietic progenitors in parallel with EC as described above. Normal human fibroblasts were isolated from foreskins using standard protocols.

In vitro methylcellulose colony-forming assay The frequency of CFU-GM progenitors was enumerated in semi-solid methylcellulose cultures containing IL-1␣ (500 IU/ml), SCF (10 ng/ml), IL-3 (10 ng/ml) and G-CSF (10 ng/ml). In brief, cells were plated in 35-mm Petri culture dishes (Falcon, Subra, Toulouse, France) containing IMDM supplemented with 10% FCS, 0.4% human serum albumin (Inter Transfusion, Bordeaux, France) and 0.88% methylcellulose (Sigma). Colonies (⬎40 cells) were scored on day 14. All cultures were performed in triplicate. To analyze the expansion of clonogenic progenitor cells by the coculture supernatants, mononuclear cells or CD34+ cells from normal human marrow were plated as above in methylcellulose containing 15% (v/v) of the coculture supernatants or the cocktail of cytokines described above and colonies were scored at day 14.

Flow cytometry Hematopoietic cells obtained as described above were incubated with saturating doses of conjugated monoclonal antibodies for 30 min at 4°C. Cells were washed two times with PBS/BSA 0.1% and analyzed on a EPICS XL (Coulter) flow cytometer. To eliminate possible interference with EC which might have detached, staining was gated on CD45+ cells. Antibodies used were Leu-M9 (anti-CD33; Becton Dickinson), Leu-M3 (anti-CD14, Becton Dickinson), Leu-12 (anti-CD19, Becton Dickinson), 69 (Plt-1) (anti-CD41, Coulter Immunotech (Marseille, France)), Anti-Hle-1 (anti-CD45, Becton Dickinson).

Morphological analysis Morphological analysis was performed by conventional microscopic analysis of May–Gru¨nwald Giemsa-stained cytospin preparations.


Cytokine measurements Cytokine concentrations in EC supernatants were measured by EIA. EIA kits were purchased from Immunotech for IL-1␤, IL6, GM-CSF quantifications, R&D systems for G-CSF and MCSF quantifications and BioEnviroTech (Allauch, France) for IL-7 and IL-3 quantifications. Measurements were done according to the manufacturer’s instructions.

Statistical analysis Results of experimental points obtained from multiple experiments were reported as the mean ± standard error of the mean (s.e.m.). Significant differences between values obtained in each assay were determined by unpaired Student’s t-test for samples with equal variance. Differences were considered as significant when P ⬍ 0.05. Results

EC secretory products efficiently support the expansion of nucleated cells and CFC CD34+ cells were resuspended in standard medium and placed in Transwell inserts above confluent EC monolayers. For each experiment the EC came from different donors. The coculture was maintained for a time period of up to 3 weeks. Regular visual inspection showed that no adherent EC layer formed in the Transwell insert within this period of time. Cell expansion was measured at 7, 14 and 21 day intervals. A progressive, sustained and important cell expansion of the nucleated cells was seen (Figure 1a). Cell expansion was routinely above 1000-fold after 3 weeks of coculture.

The progressive expansion of cell number contrasted with the very rapid expansion that could be observed when hematopoietic progenitors were added with a combination of cytokines known to directly induce proliferation of progenitors, in the absence of an EC monolayer (below). These results were consistent with the hypothesis that the EC layer promoted the progenitor expansion by producing cytokines, the concentration of which built up progressively with time and the degree of expansion being proportional to the concentration of these cytokines. Similar results were obtained with CD34+/CD38low cells (3.1 ± 0.6, 59 ± 22, 1005 ± 401 expansion of the nucleated cells at days 7, 14, 21, respectively, average ± s.e.m. of three independent experiments). To rule out a possible contact occurring through the pore of the Transwell by EC which might have detached from the monolayer, control experiments were performed with cells separated by two sets of Transwell filters; these experiments did not show any differences. Larger coculture periods, up to 8 weeks, generated a large number of hematopoietic cells. Kinetics of cell expansion showed that after the initial rapid increase in cell number until week 3 there was a progressive decline in the rate of expansion (fold cell expansion 136 ± 4 at week 4 and 12 ± 4 at week 8). Starting with 5000 CD34+ cells, the cumulated total number of cells was over 107. The EC monolayer tends to detach after 3 weeks of culture. This could be resolved by transferring the hematopoietic cells to a fresh layer of EC every second week. Another way of preventing detachment of the EC monolayer was to add DXM to the coculture medium. We observed that this addition at a final concentration of 10−6M had a detachment-preventive effect. However, a drawback of the use of glucocorticosteroids (GC) in these experiments is that they significantly reduced cell expansion and the generation of clonogenic cells (see below, Results section).

Figure 1 EC secretory products support the expansion of hematopoietic progenitors. CD34+ cells were seeded in Transwell inserts above an EC monolayer. At the indicated time intervals, cells were collected and counted and the frequency of CFU-GM measured using an in vitro colony-forming assay. (a) Expansion of the nucleated cells in non-contact cocultures with HUVEC. (b) Expansion of CFU-GM in non-contact cocultures with HUVEC. (c) Expansion of the nucleated cells in non-contact cocultures with BMEC. (d) Expansion of CFU-GM in non-contact cocultures with BMEC. Data represent the mean ± s.e.m. (a) n = 8, (b) n = 3, (c) n = 3, (d) n = 3, fold proliferation in cocultures initiated with 2500 CD34+ cells.

1213


1214

To analyze the amplification of clonogenic precursors, cells were recovered from inserts after 1, 2 and 3 weeks of coculture and an aliquot replated in methylcellulose for clonogenic progenitor cell assay. There was a progressive amplification of the number of clonogenic cells with time that paralleled the increase in total cell number. This increase in CFC could be sustained for 2–3 weeks (Figure 1b). After 3 weeks there was a progressive decrease in the number of clonogenic cells suggesting that EC secretory products promote rapid differentiation of the progenitors (fold CFU-GM expansion 11 ± 3 week 4, no CFU-GM at week 8). The EC were not able to maintain erythroid progenitors in such a non-contact coculture system as there was an immediate decrease in the number of BFU-E. Similar results were obtained with CD34+/CD38low cells (3.5 ± 0.44, 7.3 ± 2.1, 14.7 ± 2.1 expansion of the CFUGM at days 7, 14, 21, respectively, average ± s.e.m. of three independent experiments). BMEC also supported the proliferation of hematopoietic progenitors. However, expansions of both total nucleated cells and CFC were consistently lower with BMEC than with HUVEC (Figure 1c and d). In short-term 3-week-old cocultures, progenitor cell expansion was higher in HUVEC/progenitor cocultures than in bone marrow stroma/progenitor cocultures (6.5-fold difference at day 21, average of two independent experiments) or fibroblast/ progenitor cocultures (3-fold difference at day 21, average of two independent experiments).

Progenitor cell expansion and CFC generation in contact cocultures In the non-adherent population both cell proliferation and generation of clonogenic cells were increased for a period of up to 3 weeks (Figure 2a and b). As in non-contact cocultures there was a peak of expansion at week 3 followed by a progressive decline (fold cell expansion 326 ± 14 at week 4 and 26.4 ± 1.6 at week 8). Starting with 5000 CD34+ cells, the cumulated total number of cells was over 108. The generation of clonogenic cells reached a maximum at 2 weeks and then declined (fold CFU-GM expansion 11 ± 3 at week 4, no colonies at week 8). The adherent population represented a smaller number of cells that remained clonogeneic after 3 weeks (fold CFU-GM 12 ± 3 at 4 weeks followed by a decline with no CFU-GM at 8 weeks) suggesting that the more primitive progenitor cells tend to remain in the EC-adherent cell population, giving a committed progeny that detached from the EC monolayer during the coculture. Accordingly, when the non-adherent cell population was regularly removed every week, the adherent cell population gave rise to a new progeny of non-adherent cells. An adherent population of progenitors could be maintained in that way for a period of up to 8 weeks, by replating them after trypsinisation on a fresh EC monolayer (fold cell expansion of the adherent cell population 80 ± 4 at week 4; 2 ± 0.5 at week 8, average ± s.e.m. of three independent experiments). An important difference between the non-contact and contact cocultures resided in the significantly higher capacity of the contact coculture system to support the proliferation of the nucleated cells and of the CFC. This increased capacity was apparent at times of coculture as early as 5 days. At day 12 there was a significant higher cell proliferation in contact vs non-contact cocultures (average of 3-fold, five independent experiments, P = 0.002).

Figure 2 Expansion of hematopoietic progenitors in contact cocultures with EC. CD34+ cord blood cells were seeded in contact with the EC monolayer. At the indicated time intervals, cells were collected and counted and the frequency of CFU-GM measured using an in vitro colony-forming assay. (a) Expansion of the nucleated cells. (b) Expansion of CFU-GM. Data represent the mean ± s.e.m. (n = 7) fold proliferation in cocultures initiated with 2500 CD34+ cells.

Phenotypic analysis of the expanded cell population In non-contact cocultures, by day 10, more than 80 ± 5% of the expanded cells were myeloid expressing the CD33 antigen and 20 ± 5% were monocytic expressing the CD14 antigen. A very low number of cells stained with the megakaryocytic CD41 antibody. Virtually no staining was observed with lymphoid markers such as CD19. A significant number of mature myeloid cells (5–9% neutrophils) was produced, particularly in cocultures of 2 weeks or longer. At day 10, in contact cocultures, a more important population of cells (15 ± 2%) stained for CD41. Other cells belonged to the myeloid (70 ± 5%) and the monocytic (10 ± 5%) lineages.

IL-1␣ enhances EC competence to support hematopoiesis We evaluated cell proliferation and CFC generation in noncontact EC cocultures supplemented with IL-1␣ at a final concentration of 100 IU/ml. The addition of IL-1␣ consistently resulted in an average 1.3-, 2- and 3-fold increase in the nucleated cell expansion at days 7, 14 and 21 respectively


(significant at day 21, p = 0.02) (Figure 3a). CFC generation was also increased in the presence of IL-1␣, by an average 1.5- and 1.6-fold at days 14 and 21 respectively (Figure 3b). To analyze the direct responsibility of IL-1␣, EC were stimulated for 3 days with IL-1␣, washed free of IL-1␣, and then tested for their capacity to increase the number of hematopoietic cells in a 7-day coculture system. Priming of the EC monolayer with IL-1␣ during 72 h followed by a thorough wash to remove IL-1␣ led to an increased capacity of the EC to support the expansion of nucleated cells. At day 14, there was an approximately 2.6-fold higher cell proliferation in IL-1␣primed EC, vs non-primed EC (average of four experiments, P = 0.003). Finally, the number of seeded CD34+ cells was an important parameter. In experiments where increasing numbers (1000–12 000) of CD34+ were seeded in the Transwell inserts the effects of IL-1␣ were more marked in the presence of the higher number of cells (not shown). When analyzed in contact cocultures, the effects of IL-1␣ were again significant on the proliferation of the non-adherent cell population and were responsible for the generation of a very high number of cells (above 4000-fold expansion at day 21) (significant at day 21, P = 0.008) (Figure 4a). There was a moderate gain in the number of adherent cells at day 14 (not significant) and a loss at day 21 (P = 0.05) (Figure 4c). In both adherent and non-adherent cell populations IL-1␣ induced a loss of CFC at day 21, compared to day 14. Compared to day 7 there was still a gain in CFU-GM, however not significant (Figure 4b and d).

GC are potent suppressors of the EC capacity to support hematopoiesis Synthetic GC such as DXM added to the cocultures at a concentration of 10−6 M reduced cell expansion and generation of CFC in both non-contact and contact protocols. The proliferation of nucleated cells and CFC was reduced by an average 2.6-fold and 3.3-fold, respectively at day 21 (P = 0.0008, P = 0.05, respectively) in non-contact cocultures (Figure 3). In contact cocultures the effects of DXM were even more pronounced with an average 3.2-fold (P = 0.0005) and 6.2-fold (P = 0.01) reduction at day 21 of the proliferation of nucleated cells and CFC in the non-adherent population, respectively and a reduction of 2.7-fold (P = 0.008) and 8.8-fold (P = 0.01) in the adherent cell population, respectively (Figure 4). Eightweek-old contact or non-contact cocultures in the presence of DXM showed that cell expansion continues to increase until weeks 4–5 before decreasing whereas in control cells it started to decrease at week 3 (fold cell expansion in the presence or absence of DXM 150 ± 3 and 136 ± 4 at week 4 respectively, 3 ± 0.5 and 12.4 ± 4 at week 8 respectively in non-contact cocultures; 223 ± 7 and 326 ± 14 at week 4 respectively and 2.75 ± 0.5 and 26.4 ± 1.6 at week 8 respectively in contact cocultures). Fold CFU-GM expansion at week 4 was 10.5 ± 1 5 in non-contact cocultures in the presence of DXM (11 ± 3 in its absence), 14 ± 1 in contact cocultures in the presence of DXM (11 ± 3 in its absence). Other concentrations of DXM were not investigated. To analyze whether DXM could have a direct inhibitory action on the hematopoietic progenitors, CD34+ cells from normal cord blood or bone marrow were incubated in the presence of 10% 5637 cell line supernatants or a combination of SCF, IL-3 and IL-1␣ in the presence or absence of 10−6 m DXM for 7 days. Cells were enumerated and the generation of clonogenic cells was assessed by clonogenic assay in methylcellulose. No differences between control cells and DXM-treated cells could be found, ruling out a direct inhibitory effect of DXM on the progenitor cells (not shown). These results indicated that the suppressive effects of DXM in the EC/progenitor cell cocultures were related to the inhibition of the production of essential cytokines by the EC monolayer. Finally, as judged by the colony phenotype, most CFU-GM (size and day of appearance) generated in the presence of DXM were more primitive than those generated in the absence of DXM. CFU-GM colonies generated in the presence of DXM were large colonies (⬎200 cells), whose optimal size was reached only at day 21.

EC conditioned medium and coculture supernatants contain cytokines known to induce the proliferation and differentiation of hematopoietic progenitors: effects of IL-1␣ and GC

Figure 3 IL-1␣ and GC differentially alter EC support of hematopoiesis: non-contact cocultures. Non-contact cocultures were initiated in the presence of IL-1␣ at a final concentration of 100 IU/ml or DXM at a final concentration of 10−6M. Cell proliferation and the frequency of CFU-GM were determined at days 7, 14 and 21 of coculture. Data represent the mean ± s.e.m. (n = 5) fold cell expansion (a) or fold CFUGM expansion (b) in cocultures initiated with 2500 CD34+ cells.

Quantitation of cytokines was performed in 3-day-old conditioned medium or at day 14 of contact coculture and showed the presence of IL-6, G-CSF, GM-CSF, M-CSF and LIF (Ref. 46, not shown) in the supernatants. IL-1␤, IL-3 and IL-7 were not detected. In 3-day-old conditioned medium IL-1␣ stimulated the production of G-CSF (P = 0.002), GM-CSF (P = 0.05), IL-6 (P = 0.02) and M-CSF (P = 0.0006). In 14-dayold coculture supernatants IL-1␣ stimulated the production of G-CSF (P = 0.02), GM-CSF and IL-6. DXM reduced the production of G-CSF, GM-CSF and IL-6 (P = 0.01). Neither IL-1␤, IL-3 nor IL-7 were inducible by IL-1␣ or DXM (Table 1).

1215


1216

Figure 4 IL-1␣ and GC differentially alter EC support of hematopoiesis: contact cocultures. Non-adherent cells. Contact cocultures were initiated in the presence of IL-1␣ at a final concentration of 100 IU/ml or DXM at a final concentration of 10−6M. Cell proliferation and the frequency of CFU-GM were determined at days 7, 14 and 21 of coculture in the non-adherent cell population. Data represent the mean ± s.e.m. (n = 5) fold cell expansion (a) or fold CFU-GM expansion (b) in cocultures initiated with 2500 CD34+ cells. Adherent cells. Cell proliferation and the frequency of CFU-GM were determined at days 7, 14 and 21 of coculture in the adherent cell population. Data represent the mean ± s.e.m. (n = 5) fold cell expansion (c) or fold CFU-GM expansion (d) in cocultures initiated with 2500 CD34+ cells.

Supernatants of EC cultures support in vitro cell expansion and generation of clonogenic cells and this support is inhibited by cytokine neutralizing antibodies

Table 1

Cytokine production in HUVEC cultures

HUVEC + IL-1

HUVEC + DXM

A. 3-day-old cultures G-CSF UN M-CSF 638 ± 20 GM-CSF UN IL-6 970 ± 347 IL-1 UN IL-3 UN IL-7 UN

2441 ± 141 1373 ± 14 1427 ± 470 26940 ± 5572 UN UN UN

UN 871 ± 17 UN 1574 ± 212 UN UN UN

B. 14-day-old cultures G-CSF 741 ± 792 M-CSF 354 ± 293 GM-CSF 451 ± 230 IL-6 18980 ± 9550 IL-1 UN IL-3 UN IL-7 UN

2122 ± 324 409 ± 292 809 ± 393 28420 ± 3118 UN UN UN

145 ± 155 321 ± 180 42 ± 27 9435 ± 5376 UN UN UN

HUVEC

+

CD34 isolated progenitors were grown for 7 days in the presence of 25% final concentration of EC supernatants (control or IL-1␣-stimulated). Results showed that EC supernatants were by themselves able to promote the expansion of hematopoietic cells. IL-1␣-stimulated EC supernatants were on average four times more efficient than control supernatants (not shown). When a combination of neutralizing antibodies to IL1␣, IL-1␤, IL-6, IL-11, G-CSF, GM-CSF, M-CSF, SCF, LIF was added to EC supernatants a major reduction of the EC supernatant-induced proliferation of progenitors was observed. In non-contact cocultures run in the presence of the same combination of neutralizing antibodies, a profound inhibition of the expansion of nucleated cells could also be observed (Figure 5). The reduction in proliferation mainly concerned differentiated cells as there was 3.45 ± 0.5 (average of four experiments, P = 0.001) more CFU-GM in progenitors grown in the presence of the neutralizing antibodies than in their absence. Confirming these results, morphological analysis done at day 14 of the non-contact cocultures showed a striking accumulation of immature progenitors at a myeloblastic and monoblastic stage in the presence of the neutralizing antibodies (24 ± 2% undifferentiated cells, 48 ± 1% myeloblasts, 15 ± 1% promyelocytes, 2.5 ± 1% myelocytes, 11 ± 2% promonocytes in the presence of the antibodies; 4 ± 1% undiffer-

Medium was collected from (A) 3-day-old EC cultures, (B) 14-dayold cocultures performed in the presence or not of IL-1␣ (100 IU/ml) or DXM 10−6 M. Cytokine concentrations (pg/ml) were determined by EIA. UN, undetectable (lower level of detection: 20 pg/ml).


respectively) (Figure 6). IL-1␣-stimulated EC supernatants were even more efficient in promoting this potentiation indicating that the putative factors that are responsible for this potentiation were inducible by IL-1␣. Discussion

Figure 5 Cytokine neutralizing antibodies suppress EC capacity to support hematopoiesis. CD34+ cord blood cells were seeded in Transwell inserts and cocultivated with HUVEC for 14 days in the presence (HECnc + Ab) or absence (HECnc) of a combination of polyclonal neutralizing antibodies (Ab) IL-1␣, IL-1␤, IL-6, IL-11, G-CSF, GM-CSF, M-CSF, SCF, LIF at a final concentration of 5 ␮g/ml for each antibody. Cultures were maintained in a fully humidified atmosphere at 37°C and 5% CO2 and fed twice a week by removing 2 ml of medium from the bottom wells and replacing it with fresh medium with or without the combination of neutralizing antibodies. To control the efficiency of the neutralization, progenitors were grown in the presence of IL1␣, IL-1␤, IL-6, IL-11, G-CSF, GM-CSF, M-CSF, SCF, LIF (CK) with or without the combination of neutralizing antibodies (CK + Ab).

EC produce most of the cytokines known to play a role in hematopoiesis.37–47 The relationship between the production of these cytokines and the functional ability of EC to support hematopoiesis may help in understanding the role played by these cells in the ontogeny of the hematopoietic system as EC are intimately associated with embryonic hematopoiesis and may be involved in the development of embryonic progenitors.52,53 EC are also part of the bone marrow stromal microenvironment.17–23 We show that in the absence of added exogenous cytokines, EC support progenitor proliferation and generation of CFC both in a contact and a non-contact coculture system. The pattern of cytokine production is consistent with the phenotypic analysis of the expanded cell population showing essentially the presence of cells of the myeloid and the monocytic lineages. No IL-1, IL-3, IL-7 or EPO are synthetized by these cells. The lack of IL-7 production may explain

entiated cells, 22 ± 2% myeloblasts, 25 ± 2% promyelocytes, 27.5 ± 2% myelocytes, 21.5 ± 1% monocytes and promonocytes, some macrophages in the control (average of four experiments)). Finally, culture of bone marrow mononuclear or CD34+ cells in semi-solid medium supplemented with 15% of the coculture supernants also resulted in a significant increase in CFC. We also compared supernatants from EC, IL-1␣-stimulated EC and 5637 bladder carcinoma cell line. Supernatants from IL-1␣-stimulated EC were found to be the most efficient for cell proliferation and generation of CFC (not shown).

Major cell proliferation and generation of CFC is obtained in EC cocultures supplied with exogenous cytokines: potentiation effects of EC supernatants on cytokine-driven expansion We next sought to establish whether addition of exogenous cytokines to EC/CD34+ cocultures could improve the efficiency of hematopoietic progenitor expansion. IL-3, SCF and IL-1 were first chosen as they are not produced in detectable amounts by the EC monolayer. Results showed a major synergistic increase in cell proliferation and generation of CFC when IL-3, SCF and IL-1␣ were added to the non-contact cocultures compared to the control cocultures (not shown). Finally, as neutralization experiments with antibodies suggested the presence of additional uncharacterized growth factors in EC supernatants, we undertook experiments in which various combinations of cytokines were added to EC supernatants to test the effects of this addition on the proliferation of progenitors. Results show that addition of EC supernatant (6-day-old cultures) at a final concentration of 25% to a combination of IL-1␣, IL-3, IL-6, SCF, LIF, flt3-L, TPO, G-CSF, GMCSF, M-CSF and IL-11 significantly increased the proliferation of CD34+ progenitors (HECnc + CK vs HECnc, P = 0.0009 and P = 0.002; HECsup (IL-1) + CK vs HECsup (IL-1), P = 0.001 and P = 0.05 for fold cell expansion and fold CFU-GM expansion

Figure 6 Potentiation effects of EC supernatants on cytokinedriven expansion. CD34+ cord blood cells were grown for 12 days either in the presence of 25% HUVEC supernatant (HEC sup), 25% IL-1␣-stimulated supernatant (HEC sup (IL-1)), or in Transwell inserts above a HUVEC monolayer (HECnc) with the addition (CK) or not of a combination of IL-1, IL-3, IL-6, SCF, LIF, flt3-ligand, TPO, G-CSF, GM-CSF, M-CSF and IL-11. At the end of the culture period the expansion of nucleated cells and of CFC was evaluated. Results are expressed as the mean ± s.e.m. (n = 5) fold cell expansion or CFU-GM expansion. (a) Fold cell expansion; (b) fold CFU-GM expansion.

1217


1218

why these cells failed to support the growth of lymphoid progenitors, and the lack of EPO production may explain why they failed to support the growth of erythroid progenitors. EC seem to be able to support, although to a lesser extent, the proliferation of megakaryocytic progenitors as well (this work, Ref. 49). Neutralization experiments demonstrate that these EC-derived cytokines are essential and show a direct relationship between EC capacity to support hematopoiesis and the production of growth factors. These experiments also showed that this neutralization essentially impairs the differentiation of the progenitors but that the more immature cells are still able to proliferate. They suggest that other early acting growth factors are being produced by the EC monolayer. Other factors that may for instance be involved are the flt3-ligand which has been shown to stimulate the growth of primitive hematopoietic cells either alone or in combination.54,55 Experiments showing that EC supernatants are able to potentialize the effect of an association of IL-1, IL-6, IL-11, G-CSF, GM-CSF, M-CSF, LIF, SCF, TPO and flt3-ligand may also suggest that still unidentified molecules acting alone or in synergy to support the proliferation of hematopoietic cells are produced by the EC. Comparative analysis between contact and non-contact cocultures showed a significantly greater efficiency of contact cocultures for maximal production of committed progenitors indicating that additional contact-dependent factors are necessary for optimal expansion of the progenitors. Contactdependent parameters may be the optimized presentation of cytokines via their proteoglycan-mediated immobilization at the EC surface.56–58 Alternatively, adhesive interactions between progenitors and EC may promote hematopoietic progenitor proliferation. Adhesive interactions may by themselves promote cell proliferation as has been shown for VLA-4, VLA5 or VLA-6 when binding to their respective ligands.59 Adhesive interactions may also alter the way progenitors respond to growth factors. Compared to other sources of feeder cells, EC appear to be more efficient in expanding hematopoietic progenitors in short-term cultures. However, longterm EC cocultures show exhaustion of the primitive progenitors and suggest that classical long-term bone marrow cultures remain superior in the maintenance of this population. The lag time between the seeding of the progenitor cells and the initiation of cell proliferation indicates that expansion is dependent on an appropriate threshold concentration of these growth factors. Accordingly, IL-1␣ increased their production and reduced this lag time and IL-1␣-prestimulated EC were more efficient in supporting this expansion than unstimulated cells. IL-1 is known to stimulate the secretion by EC of cytokines, some of them having functional importance in hematopoiesis such as IL-6,60 colony-stimulating factors61,62 or more recently LIF.13,41 Although a direct effect of IL-1␣ on the progenitors cannot be ruled out,63 our results suggest that it is mainly acting through the modulation of cytokine production by the stromal layer. These observations could be of interest for protocols using endothelialized biomatrix as a source of feeder cells for ex vivo expansion of hematopoietic progenitors as the use of IL-1 may significantly increase the biomatrix competence. The inhibitory effects of GC on cell proliferation and generation of CFC help to understand the complex effects that these compounds have on hematopoiesis. GC have been shown to decrease the production of some hematopoiesis-promoting cytokines by EC. IL-664–65 and LIF47 for instance, are downregulated by GC. Our results show that the production of IL-6, G-CSF and GM-CSF are reduced by DXM in EC/hematopoietic

progenitor cocultures. This GC-induced reduction in cytokine production is consequent with the impairment of primitive progenitors to differentiate and proliferate in the presence of DXM. Accordingly, progenitors generated in the presence of DXM were more primitive. GC have been shown to decrease the frequency and the clonogenic progeny of LTC-ICs cultured on MS-5, a murine marrow-derived stroma cell line but not on LTBMC-derived human stroma. The authors discussed the possible inhibition by DXM of a synergistic factor produced by MS-5.66 This mechanism could also be operating for EC. In vivo, administration of GC to man results in a shift of the pattern of circulating leukocytes, mostly a decrease in the number of monocytes, basophils, eosinophils and lymphocytes. In contrast, neutrophil number is elevated.67 This effect, however, is probably secondary to a mobilization of the marginated pool of leukocytes, as GC are strong inhibitors of the expression of cell adhesion molecules at the surface of the endothelium.68 Results of this study also suggest that EC-conditioned medium might be used as a standardized source of growth factors. The ex vivo expansion of bone marrow progenitor cells has important applications for bone marrow transplantation where rapid recovery of blood cell counts is desirable. The transplantation of large numbers of CFU-GM has been shown to result in rapid hematopoietic recovery. The transplantation of large numbers of more mature cells is also desirable to quickly restore a sufficient number of granulocytes. EC-conditioned medium might therefore have a potential clinical interest for ex vivo expansion protocols of peripheral blood stem cells (PBSCs) aiming at increasing quantity and quality of PBSCs required for successful hematopoietic recovery after myeloablative therapy. Another potential interest is that these supernatants appear to contain still unidentified hematopoiesis-promoting growth factors that may be more generally produced by EC from different sources.69

Acknowledgements This work was supported by the Fondation pour la Recherche Me´dicale, the Fondation de France, produits Roche and Laboratoires Sandoz.

References 1 Dexter TM, Allen TD, Lajtha LG. Conditions controlling the proliferation of hematopoietic cells in vitro. J Cell Physiol 1977; 91: 335–344. 2 Gartner S, Kaplan HS. Long term culture of human bone marrow cells. Proc Natl Acad Sci USA 1980; 77: 4756–4759. 3 Hocking WG, Golde DW. Long term human bone marrow cultures. Blood 1980; 56: 117–124. 4 Toogood IRG, Dexter TM, Allen TD, Suda T, Lajtha LG. The development of a liquid culture system for the growth of human bone marrow. Leukemia Res 1980; 4: 449–461. 5 Dexter TM. Stromal cell associated hemopoiesis. J Cell Physiol 1982; 1: (Suppl.1) 87–94. 6 Allen TD, Dexter TM. The essential cells of the hemopoietic microenvironment. Exp Hematol 1984; 12: 517–521. 7 Verfaillie CM. Direct contact between human primitive hematopoietic progenitors and bone marrow stroma is not required for longterm hematopoiesis. Blood 1992; 79: 2821–2826. 8 Verfaillie CM. Soluble factor(s) produced by human bone marrow stroma increase cytokine induced proliferation and maturation of primitive hematopoietic progenitors while preventing their terminal differentiation. Blood 1993; 82: 2045–2053.


9 Metcalf D. Hematopoietic regulators: redundancy or subtlety. Blood 1993; 82: 3515–3523. 10 Migliaccio AR, Migliaccio G, Johnson G, Adamson JW, TorokStorb B. Comparative analysis of hematopoietic growth factors released by stromal cells from normal donors or transplanted patients. Blood 1990; 75: 305–312. 11 Caldwell J, Palsson BO, Locey B, Emerson SG. Culture perfusion schedule influence the metabolic activity and granulocyte–macrophage colony stimulating factor production rates of human bone marrow stromal cells. J Cell Physiol 1991; 147: 344–353. 12 Eaves CJ, Cashman JD, Kay RJ, Dougherty GJ, Otsuka T, Gaboury LA, Hogge DE, Landsdorp PM, Eaves AC, Humphries RK. Mechanisms that regulate the cell cycle status of very primitive hematopoietic cells in long-term human marrow cultures. II. Analysis of positive and negative regulators produced by stromal cells within the adherent layer. Blood 1991; 78: 110–117. 13 Wetzler M, Talpaz M, Lowe DG, Baiocchi G, Gutterman JU, Kurzrock R. Constitutive expression of leukemia inhibitory factor RNA by human bone marrow stromal cells and modulation by IL1, TNF-␣ and TGF-␤. Exp Hematol 1991; 19: 347–351. 14 Kittler ELW, McGrath H, Temeles D, Crittenden RB, Kister VK, Quesenberry PJ. Biological significance of constitutive and subliminal growth factor production by bone marrow stroma. Blood 1992; 79: 3168–3178. 15 Ryan DH, Nuccie BL, Ritterman I, Liesveld JL, Abbond CN. Cytokine regulation of early human lymphopoiesis. J Immunol 1994; 152: 5250–5258. 16 Wickenhauser C, Lorenzen J, Thiele J, Hillienhof A, Jungheim K, Schmitz B, Hansmann ML, Fischer R. Secretion of cytokines (interleukins-1␣, -3, and -6 and granulocyte–macrophage colonystimulating factor) by normal human bone marrow megakaryocytes. Blood 1995; 85: 685–691. 17 Siena S, Castro-Malaspina H, Gulati SC, Lu L, Colvin MO, Clarkson BD, O’Reilly RJ, Moore MA. Effects of in vitro purging with 4-hydroperoxycyclophosphamide on the hematopoietic and microenvironmental elements of human bone marrow. Blood 1985; 65: 655–662. 18 Slovick FT, Abboud CN, Brennan JK, Lichtman MA. Survival of granulocytic progenitors in the nonadherent and adherent compartments of human long-term marrow cultures. Exp Hematol 1984; 12: 327–338. 19 Keating A, Singer JW, Killen PD, Striker GE, Salo AC, Sanders J, Thomas ED, Thorning D, Fialkow PJ. Donor origin of the in vitro hematopoietic microenvironment after marrow transplantation in man. Nature 1982; 298: 280–283. 20 Toogood IR, Dexter TM, Allen TD, Suda T, Lajtha LG. The development of a liquid culture system for the growth of human bone marrow. Leukemia Res 1980; 4: 449–461. 21 Andreoni C, Moreau I, Rigal D. Long-term culture of human bone marrow. I. Characterization of adherent cells in flow cytometry. Exp Hematol 1990; 18: 431–437. 22 Strobel ES, Gay RE, Greenberg PL. Characterization of the in vitro stromal microenvironment of human bone marrow. Int J Cell Cloning 1986; 4: 341–356. 23 Berneman ZN, Chen ZZ, Ramael M, Van Poucke K, Korthout M, Van Bockstaele DR, Peetermans ME. A quantitative and dynamic study of endothelial cells and megakaryocytes in human long-term bone marrow cultures. Leukemia 1989; 3: 61–67. 24 Lemoine FM, Humphries RK, Abraham DM, Krystal G, Eaves CJ. Partial characterization of a novel stromal cell derived pre-B-cell growth factor active on normal and immortalized pre-B-cells. Exp Hematol 1988; 16: 718–726. 25 Tamir M, Harris N, Trainin N, Toledo J, Zipori D. Multilineage hemopoiesis induced by cloned stromal cells. Int J Cell Cloning 1989; 7: 373–384. 26 Kincade PW, Lee G, Pietrangeli CE, Hayashi SI, Gimble JM. Cells and molecules that regulate B lymphopoiesis in bone marrow. Annu Rev Immunol 1989; 7: 111–143. 27 Dorshkind K. Regulation of hemopoiesis by bone marrow stromal cells and their products. Annu Rev Immunol 1990; 8: 111–137. 28 Rios M, Williams DA. Systematic analysis of the ability of stromal cell lines from different murine adult tissues to support maintenance of hematopoietic stem cells in vitro. J Cell Physiol 1990; 145: 434–443. 29 Sutherland HJ, Eaves CJ, Landsdorp PM, Thacker JD, Hogge DE.

30 31

32 33

34

35 36 37

38 39 40 41 42

43

44

45

46

47

48

49

Differential regulation of primitive human hematopoietic cells in long-term cultures maintained on genetically engineered murine stromal cells. Blood 1991; 78: 666–672. Deryugina E, Mu¨ller-Sieburg CE. Stromal cells in long-term cultures: keys to the elucidation of hematopoietic development? Crit Rev Immunol 1993; 13: 115–130. Issaad C, Croisille L, Katz A, Vainchenker W, Coulombel L. A murine stromal cell line allows the proliferation of very primitive human CD34++/CD38− progenitor cells in long term cultures and semi-solid assays. Blood 1993; 81: 2916–2924. Burroughs J, Gupta P, Blazar BR, Verfaillie CM. Diffusible factors from the murine cell line MS-10B4 support human in vitro hematopoiesis. Exp Hematol 1994; 22: 1095–1101. Thalmeier K, Meissner P, Reisbach G, Falk M, Brechtel A, Do¨rmer P. Establishment of two permanent human bone marrow stromal cell lines with long-term post irradiation feeder capacity. Blood 1994; 83: 1799–1807. Yoder MC, Papaionnou VE, Breitfeld PP, Williams DA. Murine yolk sac endoderm- and mesoderm-derived cell lines support in vitro growth and differentiation of hematopoietic cells. Blood 1994; 83: 2436–2443. Peled A, Kalai KJ, Toledo J, Zipori D. Stroma-cell dependent hematopoiesis. Semin Hematol 1991; 28: 132–137. Kincade PW. Molecular interactions between stromal cells and B lymphocyte precursors. Semin Immunol 1991; 3: 379–390. Bagby GC, Dinarello CA, Wallace P, Wagner C, Hefeneider S, McCall E. Interleukin 1 stimulates granulocyte–macrophage colony-stimulating activity release by vascular endothelial cells. J Clin Invest 1986; 78: 1316–1323. Sieff CA, Tsai S, Faller DV. Interleukin 1 induces cultured human endothelial cell production of granulocyte–macrophage colonystimulating factor. J Clin Invest 1987; 79: 48–51. Knudtzon S, Mortenson BT. Growth stimulation of human bone marrow cells in agar culture by vascular cells. Blood 1975; 46: 937–943. Quesenberry PJ, Gimbrone MA. Vascular endothelium as a regulator of granulopoiesis. Production of colony-stimulating activity by cultured human endothelial cells. Blood 1980; 56: 1060–1067. Bagby GC, McCall E, Bergstrom KA, Burger D. A monokine regulates colony-stimulating activity production by vascular endothelial cells. Blood 1983; 62: 663–668. Zuckerman KS, Bagby GC, McCall E, Sparks B, Wells J, Patel V, Goodrum D. A monokine stimulates production of human erythroid burst-promoting activity by endothelial cells in vitro. J Clin Invest 1985; 75: 722–725. Broudy VC, Kaushansky K, Segal GM, Harlan JM, Adamson JW. Tumor necrosis factor type alpha stimulates human endothelial cells to produce granulocyte/macrophage colony-stimulating factor. Proc Natl Acad Sci USA 1986; 83: 7467–7471. Buzby JS, Knoppel EM, Cairo MS. Coordinate regulation of steel factor, its receptor (kit), and cytoadhesion molecule (ICAM-1 and ELAM-1) mRNA expression in human vascular endothelial cells of differing origins. Exp Hematol 1994; 22: 122–129. Koenig A, Yakisan E, Reuter M, Huang M, Sykora KW, Corbacioglu S, Welte K. Differential regulation of stem cell factor mRNA expression in human endothelial cells by bacterial pathogens: an in vitro model of inflammation. Blood 1994; 83: 2836–2843. Suen Y, Chang M, Lee SM, Buzby JS, Cairo MS. Regulation of interleukin-11 protein and mRNA expression in neonatal and adult fibroblasts and endothelial cells. Blood 1994; 84: 4125– 4134. Grosset C, Jazwiec B, Taupin JL, Lui H, Richard S, Mahon FX, Reiffers J, Moreau JF, Ripoche J. In vitro biosynthesis of leukemia inhibitory factor/human interleukin for DA cells by human endothelial cells: differential regulation by IL-1 and glucocorticoids. Blood 1995; 86: 3763–3770. Jaffe EA, Nachman RL, Becker CG, Minick CR. Culture of human endothelial cells derived from umbilical vein. Identification by morphologic and immunologic criteria. J Clin Invest 1973; 52: 2745–2756. Rafii S, Shapiro F, Pettengel R, Ferris B, Nachman RL, Weksler B, Moore MAS, Asch AS. Human bone marrow microvascular endothelial cells support long-term proliferation and differentiation of myeloid and megakaryocytic progenitors. Blood 1995; 86: 3353–3363.

1219


1220

50 Schweitzer CM, van der Schoot CE, Dra¨ger AM, van der Valk P, Zevenbergen A, Hooibrink B, Westra AH, Langenhuijsen MMAC. Isolation and culture of human bone marrow endothelial cells. Exp Hematol 1995; 23: 41–48. 51 Rutkowski JL, Cassandra JK, Lerner MA, Tennekoon GI. Purification and expansion of human Schwann cells in vitro. Nature Med 1995; 1: 80–83. 52 Kelemen E, Calvo W, Fliedner TM. Atlas of Human Hemopoietic Development. Springer-Verlag: Berlin, Heidelberg, New York, 1979. 53 Fennie C, Cheng J, Dowbenko D, Young P, Lasky LA. CD34+ endothelial cell lines derived from murine yolk sac induce the proliferation and differentiation of yolk sac CD34+ hematopoietic progenitors. Blood 1995; 86: 4454–4467. 54 Rusten LS, Lyman SD, Veiby OP, Jacobsen SEW. The FLT3 ligand is a direct and potent stimulator of the growth of primitive and committed human CD34+ bone marrow progenitor cells in vitro. Blood 1996; 87: 1317–1325. 55 Broxmeyer HE, Lu L, Cooper S, Ruggieri L, Li ZH, Lyman SD. FLT3 ligand stimulates/costimulates the growth of myeloid stem/progenitor cells. Exp Hematol 1995; 23: 1121–1129. 56 Gordon MY, Riley GP, Watt SM, Greaves MF. Compartmentalization of a hematopoietic growth factor (GM-CSF) by glycosaminoglycans in the bone marrow microenvironment. Nature 1987; 326: 403–405. 57 Roberts R, Gallagher J, Spooncer E, Allen TD, Bloomfield F, Dexter TM. Heparan sulfate-bound growth factors: a mechanism for stromal cell mediated hematopoiesis. Nature 1988; 332: 376–378. 58 Tanaka Y, Adams DH, Shaw S. Proteoglycans on endothelial cells present adhesion-inducing cytokines to leukocytes. Immunol Today 1993; 14: 111–115. 59 Shimizu Y, van Seventer GA, Horgan KJ, Shaw S. Costimulation of proliferative responses of resting CD4+ T cells by the interaction of VLA-4 and VLA-5 with fibronectin or VLA-6 with laminin. J Immunol 1990; 145: 59–67.

60 Sironi M, Breviario F, Proserpio P, Biondi A, Vecchi A, Van Damme J, Dejana E, Mantovani A. IL-1 stimulates IL-6 production in endothelial cells. J Immunol 1989; 142: 549–553. 61 Zucali JR, Dinarello CA, Oblon DJ, Gross MA, Anderson L, Weiner RS. Interleukin-1 stimulates fibroblasts to produce granulocyte– macrophage colony-stimulating activity and prostaglandin E2. J Clin Invest 1986; 77: 1857–1863. 62 Vogel SN, Douches SD, Kaufman EN, Neta R. Induction of colony stimulating factor in vivo by recombinant interleukin-1␣ and recombinant tumor necrosis factor ␣. J Immunol 1987; 138: 2143–2148. 63 Kobayashi M, Imamura M, Gotohda Y, Iwasaki H, Sakurada K, Kasai M, Hapel AJ, Miyazaki T. Synergistic effects of interleukin1␤ and interleukin-3 on the expansion of human hematopoietic progenitor cells in liquid culture. Blood 1991; 78: 1947–1953. 64 Waage A, Slupphaug G, Shalaby R. Glucocorticoids inhibit the production of IL-6 from monocytes, endothelial cells and fibroblasts. Eur J Immunol 1990; 20: 2439–2443. 65 Zanker B, Walz G, Wieder KJ, Strom TB. Evidence that glucocorticosteroids block expression of the human interleukin 6 gene by accessory cells. Transplantation 1990; 49: 183–185. 66 Croisille L, Auffray I, Katz A, Izac B, Vainchenker W, Coulombel L. Hydrocortisone differentially affects the ability of murine stromal cells and human marrow-derived adherent cells to promote the differentiation of CD34++/CD38− long-term culture-initiating cells. Blood 1994; 84: 4116–4124. 67 Cupps TR, Fauci AS. Corticosteroid-mediated immunoregulation in man. Immunol Rev 1982; 65: 133–155. 68 Aziz KE, Wakefield D. Modulation of endothelial cell expression of ICAM-1, E-selectin and VCAM-1 by ␤-estradiol, progesterone and dexamethasone. Cell Immunol 1996; 167: 79–85. 69 Davis TA, Robinson DH, Lee KP, Kessler SW. Porcine brain microvascular endothelial cells support the in vitro expansion of human primitive hematopoietic bone marrow progenitor cells with a high replating potential: requirement for cell to cell interactions and colony-stimulating factors. Blood 1995; 85: 1751–1761.