DIFFERENTIATION OF MURINE MARROW MEGAKARYOCYTE PROGENITORS (CFUM): HUMORAL CONTROL IN VITRO

Differentiation of mouse marrow megakaryocyte progenitors (CFUm) was studied in vitro by a colony assay using a plasma clot system. Erythropoietin (EPO) from sheep plasma (6 units/mg protein) in doses from 1 to 5 units/ml induced a linear increase in CFUm to a maximum of 20 colonies/10⁵ cells plated. Human urinary EPO also induced a dose-responsive increase in CFUm, but the maximum was 9 colonies/10⁵ with 2·0 units/ml of EPO and there was a decrease in colonies above that concentration. Thrombocytopoiesis-stimulating factor (TSF) derived from human embryonic kidney culture supernatant fluids induced a dose-responsive increase in CFUm in concentrations from 0·01 to 0·32 mg protein/ml in the absence of added EPO. TSF did not support the growth in vitro of erythroid colonies from mouse marrow (CFUe and BFUe) indicating an absence of EPO activity. In these studies sheep EPO appeared more effective in supporting CFUe growth than human EPO. TSF also had a stimulatory function in megakaryocyte differentiation at a precursor level. Multiple humoral factors play a role in megakaryocytopoiesis in vitro.     

Maturation and regulation of megakaryocytopoiesis.

The in vitro cloning technique for detecting megakaryocyte precursor cells was employed to compare stimuli known to influence megakaryocytopoiesis. Preparations of thrombopoietic stimulating factor (TSF) did not directly stimulate the growth of megakaryocyte colonies (CFU-m) but increased the frequency of CFU-m when TSF was added to the cultures with a constant amount of megakaryocyte colony stimulating factor. Platelets or platelet homogenates did not influence the frequency of CFU-m or the size of individual colonies. Analysis of cell surface properties of megakaryocytes obtained either by isolation from bone marrow or from in vitro colonies revealed species differences. The possibility that megakaryocytopoiesis and platelet release are regulated both within the marrow as well as by humoral factors is discussed.     

The effect of erythropoietin on the growth and development of spleen colony-forming cells

The progressive growth and development of spleen colonies was studied in heavily irradiated host mice in which erythropoiesis was modified by various procedures. Erythropoietic activity in non-polycythemic hosts bearing spleen colonies was not increased by injections of exogenous erythropoietin. Detectable levels of erythropoietin were found in the heavily irradiated host mice suggesting that the failure of exogenous erythropoietin to modify erythropoiesis was because the host mice were already maximally stimulated by the high endogenous erythropoietin levels. Spleen colonies do not become erythroid in polycythemic mice. The injection of exogenous erythropoietin into heavily irradiated polycythemic hosts did not decrease the total number of spleen colonies produced by a given bone marrow transplant, as would be expected if erythropoietin acted directly on the colony-forming cells. Comparison of growth curves for colony-forming cells in the spleens of polycythemic hosts either receiving or not receiving erythropoietin indicated that the overall doubling time of colony-forming cells during the first ten days after transplantation was not changed by the daily injection of erythropoietin. These experiments are consistent with the concept that erythropoietin is necessary for the development of erythroid colonies. Erythropoietin acts upon some progeny of the colony-forming cell rather than the colony-forming cell itself.     

Monoclonal antibodies to human urinary thrombopoietin

Monoclonal antibodies (MA) to a thrombocytopoiesis-stimulating factor (TSF or thrombopoietin) were obtained from hybridomas derived from the fusion of P3 X 63/Ag 8 cells and spleen cells from TSF-immunized BALB/c mice. The immunizing protein was a partially purified TSF-rich preparation from the urine of a thrombocytopenic patient, and was shown to stimulate platelet production in rebound-thrombocytotic mice; i.e., platelet counts of recipient mice were increased to 133% of control and the values for percentage 35S incorporation into platelets were elevated to 225% of control. Media from several hybrid cultures were tested in a microantibody detection technique that measured the binding of MA to a 125I-purified TSF preparation from human embryonic kidney (HEK) cells. The immune complex was precipitated by the addition of goat anti-mouse IgG serum and centrifugation. One clone gave 25% binding of 125I-TSF after a sevenfold dilution of the medium. This cell line was recloned and four of the subclones produced MA that gave even greater binding capacities. Hybridized cells were injected into "pristane-primed" mice and the antibodies produced in the ascites fluid were also shown to bind the 125I-TSF. Compared to the results of normal mouse serum, ascites fluid containing MA was shown to bind the unlabeled TSF from HEK cells. The TSF activity was significantly reduced in the supernatant fluid after precipitating the TSF-anti-TSF immune complex by a second antibody when tested in an immunothrombocythemic mouse assay. After SDS-PAGE, the precipitate from this TSF-MA conjugate showed that the antiserum bound a single 32,000 mol wt component, indicating the monospecificity of the MA. MA directed toward human TSF will allow studies that were not previously possible.     

Physical separation of hemopoietic stem cells from cells forming colonies in culture

Mouse bone marrow cells in suspension were separated into a number of fractions on the basis of cell density by equilibrium density gradient centrifugation, or on the basis of cell size by velocity sedimentation. After each type of separation, the cells from the various fractions were assayed for their ability to form macroscopic spleen colonies in irradiated recipient mice, and for their ability to form colonies in a cell culture system. The results from either separation technique demonstrate that cells in some fractions formed more colonies in vivo than in the culture system, while cells in other fractions formed more colonies in culture than in the spleen. The results of control experiments indicate that this separation of the two types of colony-forming cells was not an artifact of the separation procedures. From these experiments it was concluded that the population of cells which form colonies in culture under the conditions used is not identical to the population of cells detected by the spleen colony assay.     

Production of thymocyte-stimulating factor by murine spleen lymphocytes: cellular and biochemical mechanisms.

Cultures of murine spleen lymphocytes treated with Thy 1.2 antiserum plus complement do not produce thymocyte-stimulating factor (TSF). The population of thymocytes composed of immunocompetent, low-density cells produces only small amounts of TSF. Experiments with cyclophosphamide-injected mice and with spleen cells treated in vitro with antiserum to the murine B lymphocyte antigen plus complement and experiments using spleen cells stimulated in vitro with Sepharose-bound phytohemagglutinin indicate that B lymphocytes neither cooperate with T lymphocytes for the production of TSF nor produce TSF. Some lectins (pokeweed mitogen, Lens culinaris hemagglutinins A and B) have been found to induce the production of TSF by spleen cells. Other lectins (wheat germ agglutinin, Agaricus bisporus agglutinin) and sodium periodate do not. Spleen cells of mice immunized in vivo with keyhole limpet hemocyanin bound to bentonite particles or with BCG produce TSF when challenged in vitro with the specific antigen. Experiments using inhibitors of the macromolecular metabolism showed that DNA synthesis is not required for the production of TSF by spleen lymphocytes, whereas RNA and protein synthesis are required. Resolution of spleen lymphocytes on a discontinuous albumin gradient into six subpopulations showed that the TSF activity was rather uniformly distributed among the various subpopulations of cells.     

Effect of thrombopoietin on in vitro production of megakaryocytes from fetal mouse liver cells

Plasma clots containing fetal mouse liver cells (FMLC) were used to study the effects of a thrombocytopoiesis-stimulating factor (TSF) from kidney cell culture medium on the proliferation and maturation of megakarocytes. Cells in the megakaryocytic series were identified by the presence of acetylcholinesterase (AChE) and by their morphological and ultrastructural characteristics. For these experiments, 1 X 10(3) to 1 X 10(5) FMLC were cultured for 1-7 days with 0-5 micrograms of TSF; control cultures were treated with production medium (PMC) in which kidney cells had not been grown. The number of AChE+ cells that were observed depended upon the number of cells plated, i.e., after 6 days of culture with 5 micrograms of TSF, an average of 187 AChE+ cells was found after plating 1 X 10(4) cells and 1020 AChE+ cells were observed after plating 1 X 10(5) cells. In dose-response experiments, the number of AChE+ cells rose with increasing doses of TSF. Significantly elevated numbers of AChE+ cells were observed after the addition of 1-5 micrograms of TSF. The optimum time of culture, based upon the number of AChE+ cells found, was 3-5 days. Ultrastructural analysis of megakaryocytes in plasma clots showed evidence of platelet shedding on Day 5. After the culture of FMLC with TSF, a larger number of AChE+ cells was formed from a given number of cells plated than in previous studies that used adult bone marrow cells. Therefore, because of its greater sensitivity, FMLC may be useful for the assay of low levels of TSF, and may be a valuable tool for studying the effects of megakaryocytic regulatory factors on megakaryocytopoiesis.     

Mock retroviral infection alters the developmental potential of murine bone marrow stem cells.

Retroviral vectors were used to introduce an activated ras gene into murine pluripotent hemopoietic stem cells. We attempted to reconstitute the hemopoietic system of lethally irradiated mice with isolated spleen colonies obtained in vivo after injection of infected bone marrow cells. Spleen colonies derived from infected bone marrow were inefficient in promoting long-term survival of irradiated hosts. This loss of reconstitutive capacity of spleen colonies was not due to the retroviral infection per se but to the in vitro culture of spleen colony precursors. Incubation for 24 h in the presence of fetal calf serum and interleukin-3 without virus-producing cells was sufficient to abolish completely the reconstitutive capacity of spleen colonies while maintaining both self-renewal and pluripotential capacities of spleen colony precursors. These results show that the in vitro manipulation of stem cells that is included in current protocols for retroviral infection can modify the developmental potential of these cells. This finding clearly indicates that the use of retroviral vectors can introduce a bias in the analysis of hemopoiesis.     

Augmentation of erythroid burst formation by the addition of thymocytes and other myelo-lymphoid cells

Bone marrow from barrier-sustained specific pathogen-free (SPF) CBA and C57BL/6 mice gave relatively low numbers of BFU-E colonies in methylcellulose culture, as compared to conventional mice. Addition of thymocytes to the marrow cultures increased the yield of BFU-E colonies more than fourfold in SPF mice but only 1.5-fold in conventional mice. Colony size was also increased. Increased yield of BFU-E colonies was also obtained by co-culture of bone marrow with lymph node cells or with bone marrow or spleen cells from 900R whole-body-irradiated mice. The effect appeared to be cellular rather than humoral. It was not reproduced by conditioned medium from thymus or pokeweed mitogen stimulated spleen cells. The helper effect of thymus cells was eliminated or reduced by freezing and thawing, or by 48 hours of incubation after irradiation. Treatment of bone marrow cells in vitro with anti-theta serum and complement did not decrease the number of BFU-E colonies. The putative helper cells appear not to be T cells, were non-adherent to the plastic culture dish, and were cortisone resistant and radioresistant. The low BFU-E colony yield from SPF mouse marrow is presumed to be largely the result of deficiency of these non-T helper cells in SPF bone marrow, rather than of BFU-E progenitor cells.     

Some effects thymocyte-stimulating factor.

Thymocyte-stimulating factor (TSF) produced in supernatants of murine spleen cells stimulated with mitogens or with allogeneic cells confers to thymocytes the ability to respond to concanavalin A (Con A) with the dose-response characteristic of mature, immunocompetent T cells. No enhancement of the responsiveness of bone marrow cells to lipopolysaccharide (LPS) was found. Thymocytes from mice of different strains acquire, after treatment with TSF, a responsiveness to phytohemagglutinin (PHA) and Con A proportional to that shown by spleen or lymph node cells from mice of the same strain. It was also shown that murine thymocytes treated with TSF cause a graft-vs-host reaction when injected into an appropriate hybrid. All these activities are thermolabile and disappear from supernatants at the same rate, thus showing that they are, very probably, due to the same substance. Spleen cells of mice bearing tumors causing splenomegaly (C3HBA and H2712 adenocarcinomas) show a decreased production of TSF if judged on the basis of TSF produced per million spleen cells. Rat spleen cells produce a substance (rat TSF) which stimulates the PHA and Con A responsiveness of murine thymocytes. Rat TSF has a molecular weight similar or identical to that of murine TSF. However, on the basis of the different rates of thermodenaturation, it appears that rat TSF and murine TSF are two different molecules.     

Thrombopoietin from human embryonic kidney cells causes increased thrombocytopoiesis in sublethally irradiated mice.

Previous work showed that mice treated with platelet-specific antiserum prior to whole-body irradiation did not suffer the degree or duration of thrombocytopenia as did irradiated control mice. We now report that a partially purified preparation of a thrombocytopoiesis-stimulating factor (TSF or thrombopoietin) mimics the biological effects of platelet-specific antiserum treatment in hematopoietically suppressed mice. Male C3H mice were exposed to 3.0 or 4.5 Gy of 137Cs gamma radiation and injected with a total dose of 4 units (U) of TSF. Human serum albumin (HSA) and rabbit anti-mouse platelet serum-injected mice, along with unirradiated mice, served as controls. Packed cell volumes (PCV), RBC counts, WBC counts, platelet counts, and percentage 35S incorporation into platelets were measured in mice at various days (7-14) following treatment. The results showed that irradiated mice treated with TSF had increased 35S uptake into platelets and higher platelet counts than HSA-treated controls. Also, PCV, RBC counts, and WBC counts of irradiated mice treated with TSF were significantly higher than values for HSA-treated mice. Additional experiments using 40,000 U/mouse of Interleukin-6 (IL-6), 227 U/mouse of granulocyte macrophage-colony stimulating factor (GM-CSF), or a combination of GM-CSF and IL-6 did not show increased platelet counts or 35S incorporation into platelets on Days 10 and 14 when compared to other mice treated with control substances. These results suggest that the radioprotective effects of platelet antibodies reported previously may be due to the release and action of thrombopoietin. These studies also demonstrate that thrombopoietin therapy will modulate the severe thrombocytopenia that occurs in radiation-induced bone marrow suppression.     

Recovery of thrombopoietin during purification

A thrombocytopoiesis-stimulating factor (TSF or thrombopoietin) was previously purified by a six-step purification procedure. However, the exact quantity of TSF that was recovered, through the various purification procedures, was unknown because of the absence of a method for establishing a unit of measure of TSF. In the present work dose-response relationships on both the crude TSF preparations and on the more highly purified TSF were determined. TSF units were calculated from the dose-response curves. A unit of TSF is defined as the amount of material (mg) that is required to increase the percentages 35S incorporation into platelets of immunothrombocythemic mice by 50% above the baseline. The results of determining the TSF units on the crude TSF preparation indicated that 0.11 unit (U) of TSF/mg protein was present. Results showed that the specific activity of TSF can be increased to about 3.6 U/mg by a single purification procedure using Sephadex G-75 column chromatography. Increased specific activities were obtained by additional purification steps, i.e., DEAE-cellulose column chromatography, SE-HPLC, DEAE-HPLC, and SDS-PAGE. The purified product appears to have a specific activity of about 11,000 U/mg of protein with 0.00003% of the protein and 1.1% of the TSF recovered from the starting material. Establishing a unit of measure for TSF will allow calculations of its degree of purity, provide a method for quantitation of recoveries of activities after various purification procedures, and allow comparisons of results from different experiments and different laboratories.     

Effects of recombinant transforming growth factor-beta 1 on hematologic recovery after treatment of mice with 5-fluorouracil

Transforming growth factor beta 1 (TGF-beta 1) has been shown to inhibit bone marrow colony formation after in vitro treatment as well as after in vivo administration to normal mice. These data suggest that TGF-beta might either protect, or further depress, progenitor cell levels in mice exposed to a cell cycle-active drug such as 5-fluorouracil (5FU). rTGF-beta 1 was administered repeatedly by either the i.v. or i.p. routes to mice during the hyperproliferative state of the bone marrow that occurs 7 to 9 days after the i.v. administration of 150 mg/kg 5FU. The formation of both multilineage and the more differentiated (CFU-c) colonies was inhibited by 20 to 40%/culture, and 66 to 93%/mouse. When multiple doses of rTGF-beta 1 were administered systemically immediately before the injection of 5FU, the resulting rebound in the number of CFU-c and multilineage colonies containing granulocyte, erythroid, megakaryocyte, and macrophage lineage colonies per culture was markedly inhibited by 30 to 77%, whereas the total number of CFU per mouse was inhibited up to 93%. This effect was maximal when rTGF-beta 1 was administered at daily doses of greater than or equal to 5 micrograms/mouse for at least 3 days. This inhibition of the recovery of the bone marrow from 5FU treatment induced by rTGF-beta 1 was a delayed transient response because by day 16 the progenitor cell numbers and bone marrow cellularity were identical to the 5FU-treated marrow controls.     

Effect of cortisone-resistant thymocytes on endogenous colony formation in inbred mice]

Endogenous colonies in the spleen of sublethally irradiated (CBA X C57BL/6) E1 hybrid mice were recorded after the injection of thymocytes and the lymph node cells from the hydrocortisone-treated and intact CBA mice. Cortisone-resistant thymocytes failed to suppress the endogenous colony formation, whereas the lymph node cells produced a distinct suppressive effect on the endogenous colonies. After the injection of cortisone-resistant thymocytes the number of colonies in the spleen of individual recipients was double that in control irradiated hybrids.     

Proliferation and differentiation of Abelson virus-infected murine myeloid leukemia cell lines does not involve an autocrine mechanism

Culture in agar of cloned promonocytic leukemia cell lines derived from Abelson virus-infected mice produced colonies of both a compact and diffuse morphology. Diffuse colonies contained fewer cells capable of forming colonies when recultured in agar than did compact colonies. Serial subcloning of cells from diffuse, but not compact, colonies ultimately led to the complete loss of colony-forming cells, i.e., to clonal extinction. The production of both compact and diffuse agar colonies was independent of the cell density of either the static liquid culture from which cells were taken for culture in agar, or the number of cells per agar culture. Furthermore, bioassays of culture supernatants indicated the leukemia cells did not secrete hemopoietic growth factors active on normal hemopoietic cells, transforming growth factors active on adherent cell lines, or factors that influenced the growth of the leukemic cells themselves. Collectively, these data suggest neither growth-factor independent replication nor the spontaneous differentiation of Abelson virus-infected myeloid cells involves autocrine secretion of growth regulators.     

Detection of lymphoid leukemia colony-forming cells in abelson virus infected mice: Differences in inbred strains

BALB/c or DBA/2 mice were infected with Abelson murine leukemia virus (A-MuLV), pseudotype Molony murine leukemia virus (M-MuLV). Infection of these mice with 10⁴ focus-forming units of A-MuLV (M-MuLV) induced overt leukemia, detectable grossly or microscopically in 90% of the mice at 20–38 days. However, these methods did not detect leukemia at 17 days or before. Bone marrow cells from A-MuLV-infected leukemic or preleukemic mice were placed in tissue culture in a soft agarose gel. Cells from leukemic or preleukemic BALB/c mice grew to form colonies of 10³ cells or more, composed of lymphoblasts, whereas marrow cells from normal uninfected mice did not. Cells from these colonies grew to form ascitic tumors after intraperitoneal inoculation into pristane-primed BALB/c recipient. Colony-forming leukemia cells could be detected in the marrow of A-MuLV-infected mice as early as 8 days after virus incoluation. The number of colony-forming leukemia cells increased as a function of time after virus inoculation. Colony-forming leukemia cells require other cells in order to replicate in tissue culture. Normal bone marrow cells, untreated or after treatment with mitomycin-C, provide this “helper” function. Only in the presence of untreated or mitomycin-C treated helper cells was the number of colonies approximately proportional to the number of leukemia cells plated. Marrow cells from leukemic BALB/c mice form more colonies than those from leukemic DBA/2 mice. The number of colonies formed per 10³ microscopically identifiable leukemia cells plated was determined to be 2–3 for leukemic BALB/c mice and 0.3 for DBA/2 mice. Cocultivation of leukemic DBA/2 marrow cells with mitomycin-C treated normal BALB/c cells did not increase the number of colonies formed by the DBA/2 leukemic cells. Thus, the decreased ability of DBA/2 leukemia cells to form colonies appears to be a property of the leukemia cell population.     

In vitro activation of the in vivo colony-forming units of the mouse yolk sac.

Experiments were performed to investigate the presence of colony-forming units (CFU) in the mouse embryonic yolk sac during the developmental period in which the yolk sac is the sole hemopoietic organ. Injection of yolk sac cell suspensions from normal embryos into syngeneic, lethally irradiated adult recipients evoked a very low number of spleen colonies. However, prior cultivation of yolk sacs in vitro caused a dramatic increase in the spleen colony-forming capacity--as high as 84-fold--following 48 hours in culture. The yolk sac origin of the spleen colonies was confirmed by: (a) Chromosomal marker analysis; (b) dose-response analysis; (c) demonstrating that the above colonies were not of endogenous origin induced by the mere injection of grafted cells. We conclude that the yolk sac contains many precursors of colony-forming cells which though undetectable by immediate grafting apparently become activated in culture by an as yet unknown induction process.     

Proliferative kinetics and differentiation of murine blast cell colonies in culture: Evidence for variable G0 periods and constant doubling rates of early pluripotent hemopoietic progenitors

Several investigators have described hemopoietic colonies expressing multilineage differentiation in culture. We recently identified a class of murine hemopoletic progenitors which form blast cell colonies with very high replating efficiencies. In order to clarify further the relationship between progenitors for blast cell colonies and progenitors for the multilineage hemopoietic colonies in culture, we carried out analyses of kinetic and differentiation properties of murine blast cell colonies. Serial observations of the development of blast cell colonies into multilineage (and single lineage) colonies in cultures of spleen cells obtained from 5-fluorouracil (5-FU)-treated mice confirmed the transitional nature of the murine blast cell colonies. The data also suggested that the early pluripotent progenitors are in G₀ for variable periods, and that when triggered into cell cycle, they proliferate at relatively constant doubling rates during the early stages of differentiation. The notion that some of the pluripotent progenitors are in G₀ was also supported by long-term thymidine suicide studies in which spleen cells were exposed to ³H-thymidine with high specific activity for 5 days in culture, washed, and assayed for surviving progenitors. Comparison of replating abilities of day-7 and day-16 blast cell colonies from normal as well as 5-FU-treated mice indicated that some of the day-7 blast cell colonies are derived from maturer populations of progenitors which are sensitive to 5-FU. In contrast, progenitors for the day-16 blast cell colonies are dormant in cell cycle and were not affected by 5-FU treatment. Previously we reported that progenitors for day-16 blast cell colonies have a significant capacity for self-renewal. These observations suggest the hypothesis that the capability for self-renewal is accompanied by long periods of G₀, and that once commitment to differentiation takes place, then active cell division occurs.     

Recombinant murine granulocyte-macrophage (GM) colony-stimulating factor supports formation of GM and multipotential blast cell colonies in culture: comparison with the effects of interleukin-3

We studied the effects of murine recombinant granulocyte-macrophage colony-stimulating factor (GM-CSF) on murine hemopoiesis in methylcellulose culture. The GM-CSF was purified from cultures of Saccharomyces cerevisiae transfected with a cloned murine GM-CSF cDNA. In cultures of spleen cells from normal mice, only granulocyte-macrophage (GM) colonies were supported by GM-CSF. Blast cell colonies were the predominant type in cultures of spleen cells from 5-fluorouracil (5-FU)-treated mice. Dose-response studies revealed that maximal GM and blast cell colony formation is achieved with 100 U/ml GM-CSF. Blast cell colonies revealed variable but high replating efficiencies, and the secondary colonies included multilineage colonies. Serial replating of washed blast cell colonies in cultures with GM-CSF provided evidence for the direct effects of GM-CSF on the proliferation of multipotential blast cells. A combination of GM-CSF and interleukin-3 (IL-3) did not increase the number of blast cell colonies over the level supported by IL-3. This observation indicates that the progenitors for blast cell colonies that responded to GM-CSF are a subpopulation of multipotential progenitors that are supported by IL-3. Cytological studies of colonies derived from GM-CSF and/or IL-3 suggest that the eosinophilopoietic ability of murine GM-CSF is less than that of IL-3.     

In vivo stimulatory effect of macrophage colony-stimulating factor on the number of stroma-initiating cells

The in vivo effect of human macrophage colony-stimulating factor (M-CSF) on the number of cells that formed stromal colonies in an in vitro culture system (stroma-initiating cells; SICs) was investigated. We found that the number of SICs in the femurs of C57BL/6 mice was significantly increased by the treatment with M-CSF. We also found that the SICs were resistant to at least three different chemotherapeutic reagents, 5-fluorouracil (5-FU), cytarabine, and cyclophosphamide, because the femoral cells of mice treated with these reagents contained higher numbers of SICs than those of untreated mice. M-CSF treatment also increased the number of SICs of the reagent-pretreated mice. The SICs detected in our culture system were present only in Mac-1(-)CD45(-) cells, and the M-CSF treatment of 5-FU-pretreated mice actually increased the number of Mac-1(-)CD45(-) SICs. The Mac-1(-)CD45(-) SICs collected from mice that were pretreated with 5-FU and then treated with M-CSF formed stromal colonies under in vitro culture conditions that did not contain M-CSF but did contain a high concentration of fetal calf serum. This result suggested that SICs collected following the treatment procedure did not necessarily require the presence of M-CSF for their in vitro proliferation. Our study indicated that M-CSF has the ability to increase the number of progenitor or precursor cells for bone marrow stromal cells in vivo system.