The production of colony stimulating activity by monocyte enriched fractions from murine continuous bone marrow culture adherent layers

The production of colony stimulating activity (GM-CSA) within murine continuous bone marrow cultures was investigated by subjecting either the nonadherent or the adherent layer cells to separation by velocity sedimentation. The presence of GM-CSA in conditioned medium was defined by the support of granulocyte/macrophage colony formation in soft agar culture. Cell free conditioned medium from weekly feedings of intact continuous marrow cultures and medium conditioned by each fraction of velocity sedimentation separated, nonadherent cells did not contain assayable GM-CSA. However, medium conditioned by fractions of adherent layer cells with a modal sedimentation velocity of 8.8 mm/h (range 6.2-10.6 mm/h) contained GM-CSA. Cytochemical studies with Wright's-Giemsa and non-specific esterase stains in addition to immunofluorescent studies with anti-collagen III, anti-collagen IV and monoclonal anti-Mac I antibodies to define fibroblasts, endothelial cells and monocytes, respectively, demonstrated that the cells within the GM-CSA producing fractions were enriched with monocytes/macrophages. Fibroblasts and a small proportion of endothelial cells were also present. GM-CSA is produced within the microenvironment (adherent layer) of murine continuous marrow cultures. Either adherent layer monocytes and/or a monocyte-endothelial cell interaction account for the GM-CSA production.     

PRODUCTION OF COLONY-STIMULATING FACTOR BY HUMAN MACROPHAGES

The monocyte is the cell present in human peripheral blood which produces a factor stimulating bone-marrow colony formation in agar. Monocytes give rise to macrophages, both in vivo and in vitro. Therefore, to define a possible physiological role for the tissue macrophage in human leucopoietic regulation, monocyte-derived macrophages and human pulmonary macrophages were tested for colony-stimulating activity. Macrophages and conditioned media prepared from macrophage cultures were highly active in stimulating human bone-marrow granulocyte and mononuclear cell proliferation in vitro. These observations suggest that tissue macrophages may be involved in the control of leucopoiesis and that these cells may function both as phagocytes and recruit new granulocytes and monocytes for host defence.     

Mode of action of human urinary colony-stimulating factor

Human urinary colony-stimulating factor (CSF-HU) has been highly purified using procedures containing DEAE cellulose, phenyl Sepharose CL-4B, Sephadex G-200, hydroxylapatite, and high performance liquid chromatography. The final preparation had a specific activity of 3.3 X 10(7) U/mg protein. Although the purified CSF-HU was not active on human monocyte-depleted bone marrow cells, it stimulated human peripheral blood monocytes obtained from five healthy volunteers to produce human active granulocytic colony-stimulating factor (G-CSF), which stimulated human monocyte-depleted bone marrow cells to form granulocytic colonies. The human G-CSF-producing activity of CSF-HU was not neutralized by polymyxin B, which is known to inhibit the effect of endotoxin. Newly produced G-CSF had an approximate molecular weight of 24,000 daltons as judged by chromatography on Sephadex G-150. These results indicate that CSF-HU stimulates human monocytes to produce human G-CSF in vitro.     

Tumor necrosis factor type alpha stimulates human endothelial cells to produce granulocyte/macrophage colony-stimulating factor.

Tumor necrosis factor type alpha (TNF-alpha) is produced by monocytes and has been purified, sequenced, and cloned from the HL-60 cell line. Soluble products of monocytes stimulate endothelial cells to release multilineage hematopoietic colony-stimulating activity. To determine whether TNF-alpha could stimulate endothelial cells to produce these activities, we added recombinant human TNF-alpha to cultured human umbilical vein endothelial cells. Untreated endothelial cell conditioned medium and TNF-alpha-stimulated endothelial cell conditioned medium were tested for hematopoietic colony stimulating activity in colony-forming assays in methylcellulose. TNF-alpha stimulated growth factor production by endothelial cells. Fifth-passage human endothelial cells and multiply-passaged bovine aortic endothelial cells responded similarly to first-passage endothelial cells, indicating that the action of TNF-alpha on endothelial cells is direct and not due to contaminating lymphocytes or monocytes present in the first-passage cultures. To investigate the molecular basis for these findings, polyadenylylated RNA was prepared from the TNF-alpha-stimulated endothelial cells and probed for granulocyte-macrophage colony-stimulating factor and granulocyte colony-stimulating factor mRNA. Granulocyte-macrophage colony-stimulating factor, but not granulocyte colony-stimulating factor, message was detected. This finding suggests that at least some of the hematopoietic colony-stimulating activity released by the TNF-alpha-stimulated endothelial cells is granulocyte-macrophage colony-stimulating factor. These results demonstrate that a purified monocyte product can stimulate endothelial cells to produce the multilineage growth factor granulocyte-macrophage colony-stimulating factor and extend the role of this immunoregulatory protein to the regulation of hematopoiesis in vitro.     

The marrow colony forming cell and serum colony stimulating factor of the chicken.

Marrow cells of the chicken produced colonies in semisolid media. Developing colonies consisted of granulocytes, macrophages or a mixture of these two cell types. The granulocyte-macrophage CFC was nonadherent. An adherent 'CFC' was also present and it differed in several ways from the nonadherent CFC: (a) clones contained only macrophages, (b) they contained a core of nonrefractile cells, (c) their appearance was delayed 1-2 weeks, (d) they were unaffected by the presence of erythrocytes and (e) the efficiency of cloning was increased but the percentage of clones able to produce 50 or more cells was markedly decreased, i.e., the cluster/colony ratio was increased. The growth of both colony types was strictly dependent on the presence of CSF. Data obtained from dose-response studies on unfractionated marrow indicated that clusters and colonies were derived from single cells. The CSF of chicken serum yielded sigmoid dose-response curves when tested on marrow cells. Calf serum could not support cluster or colony formation when tested alone but it did have an enhancing effect on the CSF of chicken serum. Levels of serum CSF were increased by injecting chickens with bacterial endotoxin. This phenomenon occurred with five chicken lines tested, but certain chickens of the Kimber line did not respond to endotoxin with elevated levels of CSF.     

Regulation of early human hematopoietic (BFU-E and CFU-GEMM) progenitor cells in vitro by interleukin 1-induced fibroblast-conditioned medium

Stimulators of human erythroid burst-forming units (BFU-E) and multipotential colony-forming cells (CFU-GEMM) can be produced by a number of different cell types. A product of human peripheral blood monocytes, interleukin 1 (IL-1), was evaluated for its ability to stimulate fibroblast cultures to produce stimulators of human bone marrow BFU-E and CFU-GEMM colony formation. BFU-E and CFU-GEMM colony formation was evaluated using low-density, nonadherent low-density, and T lymphocyte-depleted nonadherent low-density human bone marrow cells cultured in the presence of a source of pure human erythropoietin. Both human monocyte conditioned medium (MCM) and human recombinant IL-1 (hrIL-1) induced fibroblasts to produce stimulators of BFU-E and CFU- GEMM in a dose-dependent fashion with maximal colony formation occurring when fibroblasts were stimulated by 25% MCM or 140 ng/mL ROO6B hrIL-1, or 1.25 to 5 ng/mL ROO6T hrIL-1. Preincubation of MCM and hrIL-1 with an antibody to IL-1 inactivated the ability of MCM and hrIL- 1 to induce the release of erythroid and multipotential colony stimulating activity from fibroblasts. These results suggest that monocyte-derived IL-1 is involved in regulating the production of humoral stimulators of early human hematopoietic progenitors.     

Erythroid burst-promoting activity produced by interleukin-1-stimulated endothelial cells is granulocyte-macrophage colony-stimulating factor

Interleukin-1 (IL-1) induces cultured human umbilical vein endothelial cells to elaborate heterogeneous hematopoietic growth factors, including granulocyte-macrophage and granulocyte colony-stimulating factors (GM-CSF and G-CSF, respectively). Because erythroid burst- promoting activity (BPA) is also elaborated by endothelial cells exposed to IL-1, we sought to determine whether the BPA released by IL- 1-induced endothelial cells simply reflects the known erythropoietic activity of GM-CSF or whether other uncharacterized factors might be involved. Media conditioned by multiply passaged endothelial cells cultured for three days with recombinant IL-1 alpha (ECMIL-1) stimulated erythroid burst and GM colony formation in cultures of human nonadherent T-lymphocyte-depleted marrow mononuclear cells. Pretreatment with an anti-GM-CSF antiserum neutralized all the BPA and 56% of the GM colony-stimulating activity (GM-CSA) in ECMIL-1. The antiserum used in these studies did not inhibit IL-3 or G-CSF activity and did not inhibit ECMIL-1-induced murine GM colony growth (a measure of human G-CSF). To examine whether GM-CSF induces BPA release by accessory cells, media conditioned by marrow cells cultured for three days with GM-CSF were tested in the colony growth assays. Pretreatment with anti-GM-CSF antiserum completely neutralized the BPA and GM-CSA of the marrow cell-conditioned medium. We conclude that GM-CSF is the BPA elaborated by IL-1-induced endothelial cells. The in vitro erythropoietic activity of GM-CSF is not dependent on induced BPA release by accessory cells and therefore likely results from a direct effect of GM-CSF on progenitor cells.     

Production of colony-stimulating factor(s) for granulocyte-macrophage and multipotential (granulocyte/erythroid/megakaryocyte/macrophage) hematopoietic progenitor cells (CFU-GEMM) by clonal lines of human IL-2-dependent T-lymphocytes

Human T-lymphocyte lines that were selected for recognition of HLA-DR6 antigen and were dependent for growth in vitro on an added source of interleukin-2 (IL-2) were derived from the peripheral blood of normal individuals. Each was tested for production of a lymphokine(s) with properties of granulocyte-macrophage colony-stimulating factor (GM-CSF) using as target cells nonadherent cells from human long-term bone marrow cultures (LTBMC) or fresh marrow. Each of eight T-lymphocyte lines that were OKT3, OKT4, and HLA-DR positive produced GM-CSF that stimulated colony formation by both LTBMC cells and fresh marrow. Individually examined single-cell-derived bone marrow colonies growing in T-cell GM-CSF contained peroxidase-positive neutrophils, and macrophage-monocytes (GM-CFUc). Supernatant from a single-cell-derived T-cell clonal line designated F1 stimulated formation of granulocyte-macrophage colonies, megakaryocyte colonies, macroscopic erythroid bursts, and multipotential colonies containing erythroid cells, megakaryocytes, neutrophilic and eosinophilic granulocytes, and monocyte-macrophages (CFU-GEMM) in the presence of added erythropoietin. These data indicate that human IL-2-responsive T-lymphocytes produce lymphokine(s) that stimulate proliferation of primitive as well as committed hematopoietic stem cells, and implicate human T-lymphocytes in regulation of human multipotential hematopoietic stem cells in vivo.     

Monocyte-derived recruiting activity: kinetics of production and effects of endotoxin

Cultured monocytes release a factor, monocyte-derived recruiting activity (MRA), which stimulates fibroblasts, endothelial cells, and T lymphocytes to produce colony-stimulating activity (CSA). We studied the kinetics of MRA production using a technique in which MRA levels were measured in a two stage bioassay. We used umbilical vein endothelial cells as the MRA-responsive (CSA-producing) cells, and normal colony-forming unit granulocyte-macrophage (CFU-GM)-enriched bone marrow cells (T lymphocyte- and monocyte-depleted, low density bone marrow cells) as the CSA-responsive cells. MRA stimulated a 30- fold increase in CSA production by endothelial cells. MRA production was detected in supernatants from as few as 10(3) monocytes per milliliter, required the presence of fetal calf serum, and was inhibited by cycloheximide (10 to 100 micrograms/mL) and puromycin (10 to 50 micrograms/mL). Production was detectable after 24 hours of monocyte incubation, was maintained for three days, and fell to undetectable levels by seven days. With the addition of bacterial endotoxin (lipopolysaccharide [LPS]) (50 micrograms per 10(6) cells), MRA was detectable after only three hours of incubation, and levels peaked at 24 hours. Further, maximum MRA levels in the supernatants of LPS-stimulated monocytes were up to ten times greater than peak levels in the supernatants of unstimulated monocytes. Endotoxin augmented monocyte production of MRA to a greater extent than it did CSA production, indicating that the stimulation of CSA production by endotoxin may be at least partly indirect. The responsiveness of MRA production to endotoxin in vitro is consistent with the notion that MRA may be a biologically relevant regulator of CSA production by cells of the hematopoietic microenvironment.     

Enhancement of in vitro beta-thalassemic and normal hematopoiesis by a noncytotoxic monoclonal antibody, 9.1C3: evidence for negative regulation of hematopoiesis by monocytes and natural killer cells

The enhancement of in vitro human hematopoiesis by the addition of a noncytotoxic monoclonal antibody, 9.1C3, is described. Enhancement of all aspects of in vitro hematopoiesis was observed on addition of 9.1C3 antibody to cultures of mononuclear cells from normal bone marrow, cord blood, and peripheral blood from beta-thalassemia major patients. In cultures with no exogenous colony-stimulating factor (CSF), the addition of 9.1C3 resulted in a two- to eightfold increase in nonerythroid colony formation. Similarly, for cultures maximally stimulated with CSF, the addition of 9.1C3 antibody resulted in a one- to fourfold increase in colony formation. These effects were abrogated by the removal of either adherent, Leu-M3+ or Leu-7+ cells. Colony- forming cells were shown to be present among the 9.1C3-negative cells when mononuclear cells were sorted by flow cytometry. Media conditioned in the presence of 9.1C3 and mononuclear cells were able to enhance colony formation in vitro for normal nonadherent bone marrow cells beyond that achieved with supramaximal amounts of human placental- conditioned medium and erythropoietin. The data suggest that natural killer cells interact with monocytes to exert a negative regulatory control on in vitro granulopoiesis and erythropoiesis. Consequently, the number of progenitor and multipotential cells in cultures of unfractionated cell populations may be greatly underestimated.     

Human serum stimulates the production of G-CSF, IL-1, IL-6 and IL-8 by human peripheral blood leucocytes

Human serum induces human peripheral blood leucocytes (PBL) to release an activity stimulating neutrophil colony formation (G-CSA) from human bone marrow cells. By titrating individual growth factors and using specific neutralizing antibodies we showed that: human serum contains very low levels of G-CSF which are by themselves insufficient to stimulate myeloid colony formation in primary human bone marrow cultures and cannot account for the serum releaser activity; that although no detectable levels of IL-1, IL-2, IL-3, IL-4, IL-6 or IL-8 are found in the serum, anti IL-1 antibodies partially block the release of G-CSA when added early during PBL incubation; that PBL incubated in the absence of serum for 2 d produce small amounts of IL-1, IL-6, IL-8 and G-CSF and this is increased 6-16 fold in the presence of human serum; and that the neutrophil colony-stimulating activity released by PBL incubated with human serum is G-CSF.     

Monocytes stimulate fibroblastoid bone marrow stromal cells to produce multilineage hematopoietic growth factors

In previous studies we have found that monocytes produce soluble factors that stimulate human umbilical vein endothelial cells to produce granulocyte-macrophage colony-stimulating activity (CSA), burst- promoting activity (BPA), and megakaryocyte colony-stimulating activity (Meg-CSA) as well as factors that stimulate T lymphocytes and neonatal fibroblasts to produce CSA. To test the hypothesis that monocytes would similarly stimulate the production of hematopoietic growth factors by autologous bone marrow stromal cells, multiply-passaged adherent fibroblastoid cells derived from the bone marrow of normal volunteers were exposed to conditioned media prepared by incubating autologous peripheral blood monocytes in complete medium for three days. When conditioned media from stromal cells incubated in monocyte-conditioned medium were compared with those of stromal cells cultured in the absence of monocyte-conditioned medium, BPA was increased fourfold and CSA was increased more than 30-fold. We conclude that mononuclear phagocytes recruit stromal cells of the marrow to produce multilineage growth factors in vitro. We suggest that these monocyte-derived recruiting activities may play an important role in orchestration of hematopoietic growth factor production by cells of the marrow microenvironment.     

Characterization of colony-stimulating activity produced by human monocytes and phytohemagglutinin-stimulated lymphocytes

Human colony-stimulating activity (CSA) may support the proliferation of both human and murine granulocyte-macrophage progenitor cells (CFU- C) or, in the case of human urinary CSA, may only stimulate murine bone marrow CFU-C. CSA produced in the culture media of monocytes and macrophages and phytohemagglutinin-stimulated lymphocytes from human peripheral blood was characterized for both human and mouse marrow CFU- C stimulating activities. During the initial phase of a long-term cultures of monocytes, both human- and mouse-active CSA (MnCM-HM) were produced. In later phases of culture, however, only mouse-active CSA (MnCM-M) was produced. Fractionation on Sephadex G-150 revealed two functionally distinct species from MnCM-HM and lymphocytes conditioned medium, a high molecular weight factor (MW greater than 150,000) which stimulated mouse but not human colony formation, and a low molecular weight species (MW 25,000-35,000) which was active against both mouse and human target cells. However, MnCM-M revealed only one high molecular weight species (greater than 150,000), active only on mouse marrow. The possible biologic significance of such an activity is discussed.     

Effect of lipopolysaccharide on the production of colony-stimulating factors by the stromal cells in long-term bone marrow culture

The production of colony-stimulating factors (CSFs) by murine bone marrow stromal cells was studied with Dexter long-term bone marrow culture (LTBMC). For induction of CSF release, various concentrations (0.5-40.0 microgram/ml) of bacterial lipopolysaccharide (LPS) were added to nonrecharged 3-week-old LTBMCs consisting of an intact or macrophage-depleted adherent cell layer. The depletion of monocytes/macrophages from freshly prepared bone marrow cell suspension was performed by carbonyl-iron incorporation before establishment of LTBMC. The supernatants (Sup) of normal LTBMCs contained a low level of macrophage colony-stimulating factor (M-CSF) that was produced by the adherent cells but not by the nonadherent cell elements. No colony inhibitor was found in the Sup of LTBMCs, whereas a colony-promoting activity (CPA) was detected in medium conditioned by the adherent marrow cells (AC-CM). CPA could enhance the colony formation of myeloid progenitor cells when used in combination with recombinant murine granulocyte-macrophage colony-stimulating factor (GM-CSF). The production of CSFs peaked at about 24 h after refeeding, but it then declined to only half the optimal activity at the end of the week. Addition of LPS to the intact LTBMC invariably increased the production of a GM-CSF-like cytokine. The release of this cytokine was dose dependent and peaked at a dosage of 20 micrograms/ml of LPS at 24 h after treatment. In contrast, macrophage-depleted marrow-adherent cells failed to respond to LPS for CSF secretion. These results suggest that LPS can stimulate marrow macrophages to directly release CSF or to potentiate the production of CSF by other stromal cells.     

Human blood-derived macrophages: differentiation in vitro of a large quantity of cells in serum-free medium.

Large quantities of human blood-derived monocytes have been cultured in suspension in nonadherent cell culture bags and maintained for up to 3 weeks in a serum-free medium. This serum-free medium contained Iscove's modified Dulbecco's medium (IMDM) supplemented with human albumin, alpha-phosphatidylcholine, transferrin, and insulin. Morphology, cell surface antigens, and functional properties of these in vitro maturing macrophages were studied in comparison with macrophages cultured in a standard medium containing 10% fetal calf serum. In this report we demonstrate that this serum-free medium allows a better yield of cell survival than the standard medium; it also allows the differentiation of blood monocytes into fully functional macrophagic cells that express the different antigens found in mature macrophages. The results indicate that the use of serum-free defined medium offers good conditions in which to culture large numbers of human monocytes and allows an accurate analysis of the effect of supplementation with growth factors such as granulocyte-macrophage colony-stimulating factor (GM-CSF) and macrophage colony-stimulating factor (M-CSF) on the differentiation and survival of monocytes and macrophages. Serum-free cultures could also be helpful for the precise analysis of the cell secretion activity and for determining the factors that are responsible for monocyte maturation into macrophages.     

In vitro differentiation of endothelial cells from AC133-positive progenitor cells

Recent findings support the hypothesis that the CD34(+)-cell population in bone marrow and peripheral blood contains hematopoietic and endothelial progenitor and stem cells. In this study, we report that human AC133(+) cells from granulocyte colony-stimulating factor-mobilized peripheral blood have the capacity to differentiate into endothelial cells (ECs). When cultured in the presence of vascular endothelial growth factor (VEGF) and the novel cytokine stem cell growth factor (SCGF), AC133(+) progenitors generate both adherent and proliferating nonadherent cells. Phenotypic analysis of the cells within the adherent population reveals that the majority display endothelial features, including the expression of KDR, Tie-2, Ulex europaeus agglutinin-1, and von Willebrand factor. Electron microscopic studies of these cells show structures compatible with Weibel-Palade bodies that are found exclusively in vascular endothelium. AC133-derived nonadherent cells give rise to both hematopoietic and endothelial colonies in semisolid medium. On transfer to fresh liquid culture with VEGF and SCGF, nonadherent cells again produce an adherent and a nonadherent population. In mice with severe combined immunodeficiency, AC133-derived cells form new blood vessels in vivo when injected subcutaneously together with A549 lung cancer cells. These data indicate that the AC133(+)-cell population consists of progenitor and stem cells not only with hematopoietic potential but also with the capacity to differentiate into ECs. Whether these hematopoietic and endothelial progenitors develop from a common precursor, the hemangioblast will be studied at the single-cell level.     

Juvenile chronic myelogenous leukemia: characterization of the disease using cell cultures

To characterize juvenile chronic myelogenous leukemia (JCML), the proliferative properties of bone marrow (BM) and peripheral blood (PB) cells from nine patients were studied using assays for CFU-C and CFU- GEMM and liquid cultures. All specimens showed two reproducible abnormalities: impaired growth of normal hematopoietic progenitors and excessive proliferation of monocyte-macrophage colonies in the absence of exogenous colony-stimulating activity (CSA). Cytogenetic studies in one patient indicated that the CFU-C were malignant because BM cells at diagnosis and monocyte-macrophage colonies showed an abnormal karyotype, whereas PB lymphocytes did not. In contrast to JCML, PB from six adults with Philadelphia (Ph1) chromosome-positive chronic myelogenous leukemia (Ph1 + CML) yielded CSA-dependent CFU-C colonies which were composed of granulocytes, macrophages, or both, as well as exuberant growth of BFU-E colonies. Co-cultures of JCML BM adherent or nonadherent cells with normal BM resulted in suppression of normal hematopoietic colony formation. Supernatant from JCML adherent cells in liquid culture or plasma from newly diagnosed untreated JCML patients also suppressed control BM colony growth in a dose-dependent manner. These findings confirm that JCML is a malignant disorder of monocytic lineage and suggest that the mechanism of hematopoietic failure in JCML is mediated by an inhibitory monokine secreted by malignant JCML cells.     

Interaction of monocytes and T cells in the regulation of normal human megakaryocytopoiesis in vitro: role of IL-1 and IL-2

Autologous or allogeneic peripheral blood T cells can stimulate the human megakaryocyte progenitor cell (CFU-Meg)-derived colony formation in a dose-dependent fashion in agar cultures of nonadherent (NA), T cell-depleted (NT) bone marrow (BM) cells. Low concentrations of monocytes and T cells can collaborate in the stimulation of CFU-Meg colony formation or in the production of megakaryocyte colony stimulating factor (Meg-CSF) by T cells in the presence of mitogens or IL-2. Monocytes alone can produce only negligible Meg-CSF under any conditions. When monocyte conditioned medium (CM) was added to T cell-stimulated NA, NT BM cell cultures, CFU-Meg colony growth was appreciably increased compared with that stimulated by T cells alone. Dose-dependent increase in CFU-Meg colony growth was noted when varying concentrations of IL-1 were added to T cell-stimulated NA, NT cell cultures, although IL-1 itself could support no CFU-Meg colony growth in the absence of T cells. These data suggest that a synergistic interaction between T cells and monocytes during the production of Meg-CSF by T cells could be partly mediated by IL-1. IL-2 was found to stimulate Meg-CSF production by T cells in the presence or absence of mitogens. IL-2-stimulated Meg-CSF production by T cells was augmented by the addition of monocytes. Although IL-2 itself had no stimulatory effect on CFU-Meg colony growth, dramatic augmentation in the CFU-Meg colony number was noted when IL-2 was added to T cell-stimulated NA, NT cell cultures. High concentrations of monocytes and prostaglandin E (PGE) inhibited the CFU-Meg colony formation. These results suggest that IL-1 and IL-2 may play a stimulatory role on the normal human in vitro megakaryocytopoiesis, and may be involved in the development of reactive thrombocytosis and bone marrow megakaryocytic hyperplasia in various inflammatory diseases.     

A monokine regulates colony-stimulating activity production by vascular endothelial cells

Human umbilical vein endothelial cells were cultured in supernatants of peripheral blood monocytes that had been cultured for 3 days with and without lactoferrin. Colony-stimulating activity (CSA) was measured in supernatants of the endothelial cell cultures and appropriate control cultures using normal, T-lymphocyte-depleted, phagocyte-depleted, low- density bone marrow cells in colony growth (CFU-GM) assays. Monocyte- conditioned medium contained a nondialyzable, heat labile factor that enhanced 4-15--fold the production of CSA by endothelial cells. The addition of lactoferrin to monocyte cultures reduced the activity of this monokine by 69%. Lactoferrin did not inhibit CSA production by monokine-stimulated endothelial cells. Therefore, vascular endothelial cells are potent sources of CSA, the production of CSA by these cells is regulated by a stimulatory monokine, and the production and/or release of the monokine is inhibited by lactoferrin, a neutrophil- derived putative feedback inhibitor of granulopoiesis. Inasmuch as a similar monokine is known to stimulate CSA production by fibroblasts and T lymphocytes, we suggest that mononuclear phagocytes play a pivotal role in the regulation of granulopoiesis by recruiting a variety of cell types to produce CSA.