Purification and biochemical characterization of human pluripotent hematopoietic colony-stimulating factor.

Pluripotent hematopoietic colony-stimulating factor (pluripotent CSF), a protein that is constitutively produced by the human bladder carcinoma cell line 5637, has been purified from low serum (0.2% fetal calf serum)-containing conditioned medium. The purification involved sequential ammonium sulfate precipitation, ion-exchange chromatography, gel filtration, and reversed-phase high-performance liquid chromatography. The purified protein has a molecular weight of 18,000 in NaDodSO4/polyacrylamide gel electrophoresis, both by the silver staining technique and by elution of biological activity from a corresponding gel slice, and has an isoelectric point of 5.5. Pluripotent CSF supports the growth of human mixed colonies, granulocyte-macrophage colonies, and early erythroid colonies and induces differentiation of the human promyelocytic leukemic cell line HL-60 and the murine myelomonocytic leukemic cell line WEHI-3B (D+). The specific activity of the purified pluripotent CSF in the granulocyte-macrophage colony assay is 1.5 X 10(8) units/mg of protein.     

Cell culture studies and oncogene expression in juvenile chronic myelogenous leukemia

Juvenile chronic myelogenous leukemia (JCML) may be distinguished from adult CML based upon in vitro cell growth characteristics. We studied four untreated children with JCML and report additional unique findings. Peripheral blood (PB) and bone marrow (BM) cells were grown in soft agar. Without exogenous colony-stimulating activity (CSA) there was exuberant "spontaneous" colony formation in both PB and BM cultures. In the absence of exogenous stimulus, PB colony morphology was predominantly, but not exclusively, monocyte/macrophage. When PB was depleted of adherent cells, "spontaneous" colony formation was nearly completely abrogated. Cultures were also performed in the presence of various sources of CSA including giant cell tumor-conditioned medium (GCT-CM), a melanoma cell line-CM (LD1-CM), human placenta-CM (HPCM), and normal PB mononuclear cell (PBMC) feeder layers. Colony formation was typically increased with HPCM and PBMC, whereas in two patients GCT-CM and LD1-CM failed to stimulate additional colony growth when compared to cultures without exogenous CSA and, in fact, appeared to inhibit baseline "spontaneous" growth. The morphology of colonies in the presence of exogenous stimuli was highly variable. Because of the recent association between the c-fms protooncogene product and the receptor for the monocyte growth factor CSF-1, we analyzed the PB cells from two JCML patients for c-fms expression. Although expressed, c-fms levels were less than that in an adult with Ph1-positive CML in chronic phase. These studies indicate that in JCML, there are dramatic increases in both PB and BM colony-forming cells and that "spontaneous" growth is dependent on an accessory adherent cell fraction. Furthermore, patterns of responsiveness to various sources of CSA suggest that the colony-forming cells may not be a uniform population of malignant cells.     

Production of growth factors by malignant lymphoma cell lines

Fourteen Epstein-Barr virus (EBV)-negative cell lines were raised from bone marrow (BM), peripheral blood (PB), or lymph node samples of patients with intermediate- or high-grade malignant lymphoma. The cell lines were propagated in liquid suspension culture. They contain clonogenic progenitors capable of forming lymphoma colonies in semi- solid culture medium. Cells of these lines were used to examine the growth factor requirements of their clonogenic progenitors and to assess their ability to produce their own growth factors. Two of the cell lines (OCI-Ly9 and OCI-Ly13.1) required addition of exogenous factors for colony growth. These factors were routinely provided by media conditioned by phytohemagglutinin-stimulated leukocytes (PHA- LCM). Three lines formed some and nine lines gave rise to optimal numbers of colonies without addition of growth factors. Eight of these factor-independent lines were able to function as feeder cells and promoted colony formation by both factor-dependent lines. Cell lines that displayed feeder cell function released activities into supernatants able to replace their cellular source. Some of these endogenously produced growth-promoting activities could be replaced by known hematopoietic growth factors. Both factor-dependent cell lines were cultured with recombinant IL-1 alpha, IL-2, IL-3, IL-6, and GM colony-stimulating factor (CSF) and semipurified B-cell growth factor (BCGF) interleukin-4 (IL-4). A heterogeneous response pattern was observed. Both lines formed colonies with IL-4. The colonies were comparable in frequency and size with colonies observed with (PHA-LCM). OCI-Ly9 responded to IL-6 but showed no growth with IL-2. In contrast, the TAC-positive line OCI-Ly13.1 gave rise to colonies with IL-2 while remaining unresponsive to IL-6. A moderate number of colonies was observed when cells of this line were cultured with GM-CSF. Colony formation of both lines was uninfluenced by IL1 alpha or IL-3.     

Interaction between normal and leukaemic human cells in agar culture.

The influence of leukaemic cells from 12 patients with acute leukaemia on normal granulopoiesis in agar culture was investigated using leukeamic cell feeder layers. Leukaemic feeder cells from 7 of the 12 patients elicited no colony growth, while cells from the remaining 5 stimulated normal colony growth. In 3 of the 7 non-stimulatory patients release of inhibitory factors from the leukaemic cells seemed responsible for the effect on normal granulopoiesis, while inappropriate colony stimulating factor (CSF) production by the feeder cells could not be ruled out in the remaining 4 patients. When the leukaemic cells were cultured with, as well as without, conditioned medium, cells from 5 of the patients formed clusters. Growth in these cultures did not correlate to the effect found in the feeder layer experiments.     

Studies on the myeloid synergistic factor from 5637: comparison with interleukin-1 alpha

A synergistic factor that is produced by the human bladder carcinoma cell line 5637 (SF-1) stimulates primitive bone marrow progenitor cells, termed high proliferative-potential colony-forming cells (HPP- CFC), in the presence of an optimal dose of macrophage colony stimulating factor (CSF-1). Recent reports have demonstrated that interleukin-1 alpha (IL-1) is identical to hemopoietin 1 and have suggested that IL-1 is the synergistic factor present in 5637 conditioned medium (cm). We have compared the ability of recombinant human IL-1 alpha and partially purified preparations of SF-1 to synergize with optimal doses of CSF-1 to stimulate HPP-CFC. In all experiments performed the numbers of HPP-CFC colonies formed with IL-1 were significantly less than with SF-1. Replating experiments demonstrated that SF-1 plus CSF-1 generated HPP-CFC (responsive to IL-3 plus CSF-1); however, IL-1 plus CSF-1 resulted in no generation of HPP- CFC. Multiple factor combinations of IL-1 and SF-1 with G-CSF, GM-CSF, and CSF-1 also resulted in less HPP-CFC colony formation in cultures containing IL-1 compared with SF-1. Incubation of SF-1 with an antibody to IL-6 had no effect on HPP-CFC colony formation and IL-6 did not synergize with IL-1 plus CSF-1 or SF-1 plus CSF-1. These data suggest the presence of a factor in 5637 cm, which is distinct from G-CSF, GM- CSF, and IL-6, which synergizes with IL-1 to produce the SF-1 effect.     

Production of colony-stimulating activity by circulating leukocytes from patients with chronic myeloid leukemia at various phases of the disease

The production of colony-stimulating activity (CSA) by peripheral blood leukocytes (PBL) from 107 patients with chronic myeloid leukemia (CML), 58 at diagnosis, 34 during hematological remission, 33 in relapse from hematological control, 39 in accelerated phase, and 28 in blast phase, was measured in the double-layer agar culture system using normal nonadherent bone marrow cells as the source of CFU-GM. The median CSA levels in various stages decreased to 28-60% of control, whereas in hematological remission, there was a normalization of CSA. Five fractionation experiments using monocytes as feeder layers did not show CSA production in untreated CML. The leukemic PBL of 6 patients in blast crisis had no apparently inhibitory effect on colony formation stimulated by normal PBL. There was no correlation between the CSA of PBL and the number of CFU-GM in their bone marrow or peripheral blood at different disease states. Likewise, we failed to find a relationship between the PBL CSA levels and the total WBC counts or the differential counts in the peripheral blood as well as in the feeder layers. Our study indicated that circulating leukocytes from patients with CML at various phases except in hematological remission produce less CSA, which is not attributed to a low number of monocytes or the inhibitory effects of leukemic cells.     

Regulation of the colony-stimulating activity produced by a murine marrow-derived cell line (H-1).

The production of molecular species that stimulate growth of granulocyte or macrophage colonies (GM-CSF) by the fibroblastoid H-1 cell line is unaffected by either native or iron-saturated lactoferrin, although some inhibition is detected with 10 microM prostaglandin E1. The H-1 GM-CSF is able to support the formation of macrophage, neutrophil, and mixed colonies. Feeder layers of H-1 cells are also able to support the development of colony-forming units stimulated by GM-CSF (GM-CFUc) although the number of colonies produced when the optimal H-1 cell concentration is plated (2.5 x 10(3) cells) is only 30% of the number with conditioned medium alone. This inhibitory effect is observed irrespective of the presence of an additional agar layer between the feeder cells and plated bone marrow cells, implying that diffusable substances are involved. Addition of indomethacin (10 microM) to feeder layers derived from 2.5 x 10(3) H-1 cells increases the number of GM-CFUc detected to 50% of that seen with conditioned medium alone. This result suggests that released prostaglandin may be responsible for some, but not all, of the observed inhibition of colony formation. In the presence of the H-1 feeder layers, only macrophage colonies are detected and hence it appears that the H-1 cells produce, in addition to prostaglandin, a diffusible inhibitory substance that preferentially inhibits granulopoiesis.     

Characterization of hemopoietic activities in media conditioned by a murine marrow-derived adherent cell line, B.Ad

The hemopoietic activities present in medium conditioned by a murine bone marrow-derived adherent cell line (B.Ad) have been studied. B.Ad-conditioned medium stimulated neutrophil, neutrophil-macrophage, and macrophage colonies in agar cultures of bone marrow cells and 90% of this activity was neutralized by antimacrophage colony-stimulating factor (anti-M-CSF). The conditioned medium supported the generation and/or maintenance of spleen colony-forming units (CFU-S) in liquid cultures and synergized with multilineage colony-stimulating factor (Multi-CSF; IL-3) to stimulate colony formation by day-3 post-5-fluorouracil (FU)-treated bone marrow cells. When used as feeder layers, B.Ad cells stimulated erythroid colony-forming units (CFU-E) and markedly enhanced erythroid burst-forming units (BFU-E) stimulation more than did maximal Multi-CSF (IL-3) and Epo stimulation. No CFU-E- or BFU-E-stimulating activities were detected in medium conditioned by B.Ad cells. Similarly, B.Ad-conditioned medium was unable to stimulate Multi-CSF (IL-3) or granulocyte-macrophage (GM)-CSF-dependent cell lines. The data suggest that medium conditioned by this bone marrow-derived adherent cell line contains M-CSF and other factors not detectable as CSFs that either directly or by means of a synergistic mechanism are able to stimulate CFU-S and colony-forming cells (CFC).     

Acidic isoferritins and E-type prostaglandins in sources of colony stimulatory factors mask detection of cycling granulocyte-macrophage progenitor cells

The effect of granulocyte-macrophage colony stimulatory factors (GM- CSF), acidic isoferritins, and E-type prostaglandins on the detection of the cycle status of human granulocyte-macrophage progenitor cells (CFU-GM) was investigated. Bone marrow cells were pulse-treated with control medium or high specific activity tritiated thymidine [3HTdr) and subsequently plated over feeder layers containing mononuclear blood leukocytes prepared in the absence or presence of anti-acidic isoferritins and/or indomethacin, or plated in the presence of medium conditioned by placental cell or GCT-conditioned media free of acidic isoferritins and prostaglandin-E. The presence of anti-acidic isoferritins and/or indomethacin in the blood leukocyte feeder layers increased the detectable stimulatory capacity of these cells and permitted detection of a larger proportion of marrow CFU-GM in cycle than in control cultures. The cycle status was not influenced by GM-CSF in conditioned medium regardless of the dilution of conditioned medium used to stimulate colony formation. This suggests that GM-CSF, supplied to the cells after treatment with 3HTdr, does not itself influence the detection of CFU-GM in cycle, but using sources of GM-CSF that contain acidic isoferritins or prostaglandin-E will underestimate the actual number of CFU-GM in S-phase.     

Biologic properties in vitro of a recombinant human granulocyte- macrophage colony-stimulating factor

Recombinant human granulocyte-macrophage colony-stimulating factor (rH GM-CSF) was purified to homogeneity from medium conditioned by COS cells transfected with a cloned human GM-CSF cDNA and shown to be an effective proliferative stimulus in human marrow cultures for GM and eosinophil colony formation. The specific activity of purified rH GM- CSF in human marrow cultures was calculated to be at least 4 X 10(7) U/mg protein. Clone transfer experiments showed that this proliferation was due to direct stimulation of responding clonogenic cells. Acting alone, rH GM-CSF did not stimulate erythroid colony formation, but in combination with erythropoietin, increased erythroid and multipotential colony formation in cultures of peripheral blood cells. rH GM-CSF had no proliferative effects on adult or fetal murine hematopoietic cells, did not induce differentiation in murine myelomonocytic WEHI-3B cells, and was unable to stimulate the survival or proliferation of murine hematopoietic cell lines dependent on murine multi-CSF (IL 3). rH GM- CSF stimulated antibody-dependent cytolysis of tumor cells by both mature human neutrophils and eosinophils and increased eosinophil autofluorescence and phagocytosis by neutrophils. From a comparison of these effects with those of semipurified preparations of human CSF alpha and -beta, it was concluded that rH GM-CSF exhibited all the biologic activities previously noted for CSF alpha.     

RAS-blocking bisphosphonate zoledronic acid inhibits the abnormal proliferation and differentiation of juvenile myelomonocytic leukemia cells in vitro

Juvenile myelomonocytic leukemia (JMML) is a clonal myeloproliferative/myelodysplastic disorder of early childhood with a poor prognosis. JMML cells are characterized by hypersensitivity to granulocyte-macrophage colony-stimulating factor (GM-CSF) caused by a continuously activated GM-CSF receptor-retrovirus-associated sequence (RAS) signal transduction pathway through various molecular mechanisms, resulting in spontaneous GM colony formation in vitro. Bisphosphonate zoledronic acid (ZOL), a RAS-blocking compound, suppressed colony formation from bone marrow (BM) cells of 8 patients with JMML and 5 healthy control subjects without and with GM-CSF (10 ng/mL), respectively, in a dose-dependent manner in clonal culture. At 10 microM ZOL, however, spontaneous GM colony formation from JMML BM cells decreased to 3%, but the formation of G colonies containing granulocytes, but no macrophages, was enhanced, whereas 40% of GM colonies were retained and G colony formation was not affected in culture of normal BM cells with GM-CSF. In suspension culture, cytochemical and flow cytometric analyses showed that 10 microM ZOL also inhibited spontaneous proliferation and differentiation along monocyte/macrophage lineage of JMML BM cells but not the development of normal BM cells by GM-CSF. The inhibitory effect of ZOL on JMML cells was confirmed at a single-clone level and observed even at 3 microM. The current result offers a novel approach to therapy in JMML.     

Interleukin-4 has basophilic and eosinophilic cell growth-promoting activity on cord blood cells.

Effects of human recombinant interleukin-4 (IL-4) on cord blood cells depleted of T cells and monocytes were tested in colony assays and liquid cultures. IL-4 did not induce colony formation in semisolid medium, but enhanced generation of basophil colonies induced by conditioned medium (CM) of the bladder carcinoma cell line 5637. In liquid cultures, variable degrees of basophil growth were observed in the presence of IL-3, granulocyte-macrophage colony-stimulating factor (GM-CSF), G-CSF, and 5637 CM, or even with IL-4 alone, but the highest number of basophils were obtained when IL-4 was used in combination with IL-3 or 5637 CM. Progressive basophil growth was observed during 3 to 4 weeks of culturing, whereafter the numbers of basophils remained stationary for another 3 weeks. Interestingly, cord blood cell cultures performed with IL-3 contained variable percentages of eosinophils that were further enhanced in the presence of combinations of IL-3 and IL-4. These latter cultures contained approximately 50% eosinophils and 50% basophils. Kinetic studies indicated that basophils were present 7 days after onset of the cultures, whereas eosinophils did not appear before day 13. In contrast to the pronounced effects of IL-4 and 5637 CM on basophil development, relatively low numbers of eosinophils were observed under these culture conditions. Our results indicate that eosinophil and basophil development are regulated by different sets of factors, and that IL-4 has an enhancing effect of both cell lineages in association with the appropriate factors.     

Ph1-negative T lymphocytic colonies in agar cultures of peripheral blood in chronic myeloid leukemia

T lymphocytic colony formation by peripheral lymphocytes separated by discontinuous albumin gradient centrifugation was evaluated in 8 patients with Philadelphia (Ph1)-positive chronic myeloid leukemia (CML). Colonies were obtained using a liquid-on-agar culture system recently introduced (PHA overlayer-leukocyte feeder layer assay) which has been shown to be simple and reliable. The pattern of colony growth in CML and in normal controls was similar, the peak ranging from the 4th to the 6th day. Also the morphological aspects of colonies did not differ in the two groups. Cells recovered from CML lymphocytic colonies were shown to belong to T cell lineage, as they are able to form spontaneous E-rosettes and to respond to mitogenic stimulation in vitro. In contrast, cells recovered from all other cultured fractions failed to display these properties. Cytogenetic analysis showed that T colony cells were Ph1-negative whereas the chromosome anomaly was found in nonlymphoid colonies of the same patients, thus suggesting a nonclonal origin of T lymphocytes in CML.     

Interleukin-6 and granulocyte-macrophage colony-stimulating factor are candidate growth factors for chronic myelomonocytic leukemia cells

Previous studies suggest that malignant cells from some patients with myeloid leukemias produce colony-stimulating factors (CSFs) that can function as autocrine growth factors in vitro. We have examined the roles of interleukin-6 (IL-6) and granulocyte-macrophage CSF (GM-CSF) in the proliferation of myeloid leukemia cells. IL-6 activity was assessed in conditioned medium (CM) from myeloid leukemia cell cultures or cell lysates using IL-6-dependent KD83 and 7TD1 murine cell lines. Media conditioned by cells from patients with chronic myelomonocytic leukemia (CMMoL), but not by normal monocytes, chronic myelogenous leukemia (CML), or acute myelogenous leukemia (AML) cells, contained substantial levels (50 to 1,000 U/10(6) cells) of IL-6. The IL-6 content of CM correlated directly with donor peripheral blood WBC count. CM from two of five CMMoL samples also contained greater than 350 pg/mL GM-CSF. Moreover, CMMoL cells spontaneously formed colonies in semisolid medium. CMMoL colony formation could be partially inhibited by antibodies to IL-6 or GM-CSF, whereas combination of these antibodies gave additive, and nearly complete (greater than 93%), inhibition of spontaneous colony formation. Cell lysates from uncultured CMMoL cells from one patient contained abundant GM-CSF protein but no detectable IL-6. These data suggest that IL-6 and GM-CSF act in vitro as autocrine growth factors for CMMoL cells, and that CMMoL cells in vivo may represent a GM-CSF-dependent autocrine growth system.     

Adult hemoglobins are synthesized in murine fetal hepatic erythropoietic cells

The hemoglobins present in murine fetal hepatic erythroblasts on days 12-15 of gestation were studied by biochemical and immunocytologic techniques. In addition, fetal hepatic hemopoietic progenitor cells obtained from normal and mutant f/f mouse fetuses on days 11-13 of gestation were cultured in vitro with added erythropoietin and adult spleen cell conditioned medium to form large erythroid colonies. In all instances, adult hemoglobin synthesis was detected in the fetal hepatic erythroblasts and in the erythroid cell cultures in vitro. The tumor promoter, 12-O-tetradecanoylphorbol 13-acetate, enhanced the fetal hepatic erythroid colony growth in vitro, but did not alter the hemoglobin phenotypic expression.     

Spleen stromal cell lines selectively support erythroid colony formation

Mouse stromal cell lines (MSS lines) have been established from the spleens of newborn mice in culture at a low serum concentration. These MSS lines support the proliferation and differentiation of the erythroid progenitor cells from mouse fetal livers and bone marrow in a semisolid medium in the presence of erythropoietin. Larger colonies of over 1,000 benzidine-positive erythroid cells were developed from the fetal liver cells on the MSS cell layers after 6 days of incubation. These layers also support the maturation of the erythroid cells since the enucleation process of the latter was observed in large erythroid colonies. Metabolically active MSS cells are apparently required to support the proliferation and differentiation of the erythroid progenitor cells, because neither the MSS cells inactivated with fixation nor the conditioned media of MSS cells promoted the erythroid colony formation. These studies demonstrate that MSS lines specifically support the proliferation and differentiation of the erythroid progenitor cells in vitro and that stroma cells may have a critical function in blood formation in the mouse spleen.     

Cholera toxin enhances factor-dependent colony growth of murine mast cell progenitors

Mast cell colonies were observed when mouse spleen or bone marrow cells were cultured in the presence of medium conditioned by concanavalin-A-stimulated spleen cells, indicating that the medium contains the factor(s) necessary for the formation of these colonies. This factor-dependent colony growth of mast cell progenitors was enhanced by cholera toxin and prostaglandin E, which act on cellular growth mainly by elevating the intracellular cyclic-AMP level. The effect of the toxin was neutralized by preincubation of the toxin with GM1 ganglioside, the receptor substance for cholera toxin, suggesting that cholera toxin exerts its action through GM1 gangliosides present on mast cell progenitors. The toxin B subunit, which binds to GM1 ganglioside but does not elevate intracellular cyclic AMP level, did not affect the colony growth of mast cell progenitors. From these results, it is suggested that intracellular cyclic AMP levels may be involved in colony growth of mast cell progenitors.     

The FLT3 ligand is a direct and potent stimulator of the growth of primitive and committed human CD34+ bone marrow progenitor cells in vitro

The present studies investigated the effects of the recently cloned flt3 ligand (FL) on the in vitro growth and differentiation of primitive and committed subsets of human CD34+ bone marrow (BM) progenitor cells. FL alone was a weak growth stimulator of CD34+ BM cells, but synergistically and directly enhanced colony formation in combination with interleukin (IL) 3, granulocyte colony-stimulating factor (G-CSF), CSF-1, granulocyte macrophage (GM) CSF stem cell factor (SCF), and IL-6. FL and SCF were equally effective in stimulating colony formation in combination with IL-3. However, the tri-factor combination of FL + IL-3 + SCF stimulated 2.3-fold and 2.5-fold more colonies than FL + IL-3 and SCF + IL-3, respectively. These additional recruited progenitors appeared to be predominantly located in a primitive (CD71-) subset of the CD34+ progenitors, as 4.5-fold more colonies were formed by CD34+CD71- cells in response to FL + IL-3 + SCF than to FL + IL-3 or SCF + IL-3. Similar findings were observed in serum-containing and serum-deprived cultures. Whereas FL did not enhance burst-forming unit-erythroid (BFU-E) colony formation of CD34+ BM cells in the presence of serum, a low number of BFU-E colonies were formed in response to FL plus erythropoietin (Epo) under serum-deprived conditions. In addition, FL both in serum-containing and serum-deprived cultures stimulated colony formation of more committed myeloid progenitors in CD34+CD71+ BM cells. Thus, FL potently stimulates the growth of primitive and more committed human BM progenitor cells.     

Mechanism of synergy between granulocyte-macrophage colony-stimulating factor and granulocyte colony-stimulating factor in colony formation from human marrow cells in vitro

The synergy of human granulocyte-macrophage colony-stimulating factor (GM-CSF) and human granulocyte colony-stimulating factor (G-CSF) in the colony formation derived from human marrow cells was studied. The colony formation stimulated by GM-CSF plus G-CSF was dependent on the dose of each CSF, with the plateau for the number of GM colonies being higher than the sum of the individual plateaus by GM-CSF or G-CSF. Analysis of the colonies formed by GM-CSF plus G-CSF revealed efficient formation of neutrophil and monocyte colonies. To study the effect of GM-CSF and G-CSF on the maintenance of the progenitors that respond to the synergy of the CSFs, addition of each CSF to the medium of clonal cell culture was delayed. The progenitors that formed colonies on day 7 due to synergy of the CSFs were perfectly maintained by GM-CSF for at least 72 h and the progenitors that formed colonies on day 14 due to synergy of the CSFs were partly maintained by G-CSF or GM-CSF. The DNA synthetic rate of the progenitor cells that respond to GM-CSF plus G-CSF was significantly lower than those that respond to GM-CSF or G-CSF. According to light scatter analysis of phagocyte-depleted marrow mononuclear cells (PD-MMCs) using a flow cytometer, the peak population of progenitors that respond to GM-CSF plus G-CSF was in the smaller part of the PD-MMCs than those to GM-CSF or G-CSF. These results indicated that the progenitors to the synergy of GM-CSF and G-CSF are in a different proliferative state than those to each CSF. The synergy of GM-CSF and G-CSF depends on each CSF maintaining the viability of a different population of GM progenitors that can form GM colonies by both CSFs together.     

Tumor promoters enhance myeloid and erythroid colony formation by normal mouse hemopoietic cells.

The diterpene tumor promoters enhance the proliferation in culture of myeloid and erythroid precursor cells from normal mouse hemopoietic tissues. The effect is observed only with those diterpenes that are tumor promoters, including 12-O-tetradecanoylphorbol 13-acetate (TPA); diterpenes that are inactive as tumor promoters are ineffective as stimulators of colony formation. Tumor promoters act synergistically with suboptimal concentrations of conditioned medium used as a source of colony-stimulating factor (CSF) to increase both the number and size of myeloid colonies. Formation of myeloid colonies is stimulated by tumor promoters even without addition of CSF. Both pure and mixed granulocyte/macrophage colonies develop; high concentrations (> 100 micrograms/ml) of TPA are more favorable for macrophage colony formation. In erythropoietin-stimulated cultures, tumor promoters enhance the development of relatively early and intermediate of erythroid precursors, whereas later erythroid precursors are unaffected.