Search

Mice defective in both granulocyte colony-stimulating factor (G-CSF) and granulocyte-macrophage colony-stimulating factor (GM-CSF) have severely impaired neutrophil production and function, yet these mice respond to acute pathogen challenge with a significant neutrophil response. We have recently reported the development of an in vitro system to detect granulopoietic cytokines secreted from cells isolated from G-CSF, GM-CSF double knockout mice. The conditioned media produced by these cells after stimulation with lipopolysaccharide or Candida albicans supports the production and differentiation of granulocytes (ie, the conditioned media contains neutrophil promoting activity [NPA]). We now show that the NPA in the G-CSF−/−/GM-CSF−/− conditioned media requires interleukin-6 (IL6), is abolished by soluble gp130, and can be specifically immunodepleted by an anti-IL6R antibody. NPA effects on bone marrow cells are also mimicked by Hyper-IL6, and the soluble IL6R is present in NPA. These results show that the IL6/sIL6R complex is the major effector of NPA. NPA production by mice defective for both G-CSF and GM-CSF uncovers an alternative pathway to granulocyte production, which is activated after exposure to pathogens.

GM-CSF, purified from mouse lung-conditioned medium, stimulated granulocyte-macrophage colony formation but was unable to stimulate the formation of any type of pure or mixed erythroid colony in agar cultures of 12-day CBA fetal liver cells. However, initial incubation of cells for 2 days with purified GM-CSF, followed by addition of mouse spleen-conditioned medium (SCM), allowed the formation of up to 50% of the erythroid or mixed hemopoietic colonies expected in cultures stimulated from the outset by SCM. These GM-CSF-responsive colony-forming cells were not distinguishable from the remainder of cells forming erythroid or mixed colonies by sedimentation velocity or cell cycle status. From 302 clones stimulated to proliferate by purified GM-CSF, 25 formed erythroid or mixed hemopoietic colonies on transfer to cultures containing SCM. Clones initiated by SCM were unable to form erythroid or mixed colonies when transferred to cultures containing purified GM-CSF. Single cells from 12-day CBA fetal peripheral blood also were stimulated to proliferate by purified GM-CSF and subsequently formed pure or mixed erythroid colonies when SCM was added. GM-CSF is thus able to directly stimulate up to 5 cell divisions of at least 50% of detectable multipotential and early erythroid precursor cells.

Human placental conditioned medium (HPCM) contans colony-stimulating factors (CSFs) required for the growth in vitro of neutrophilic granulocyte-macrophage (GM) and eosinophilic (EO) progenitor cells from human bone marrow. Fractionation of CSFs in HPCM was achieved by manipulation of the elution conditions on a column of phenyl-Sepharose. After equilibration of the phenyl-Sepharose column at high ionic strength (1 M ammonium sulfate), all of the CSF bound; one species of GM-CSF (alpha) and all of the elutable EO-CSF were eluted from the column simply by reducing the salt concentration, whereas the second species of GM-CSF (beta) was free of EO-CSF and was eluted only by increasing the concentration of tehylene glycol in the elution buffer. The two GM-CSFs were functionally distinct. GM-CSF alpha preferentially stimulated colony formation by day 14 of culture, and there was a decreased proportion of neutrophil colonies and increased proportion of macrophage colonies as the strength of the stimulus was decreased; GM- CSF beta, on the other hand, preferentially stimulated colony formation by day 7 of culture, and the proportion of neutrophil colonies was high (average 80%) and independent of the concentration of GM-CSF beta. GM- CSF alpha and GM-CSF beta were indistinguishable on the basis of apparent molecular size on tel filtration columns (molecular weight 30,000), charge properties on isoelectric focusing beds (isoelectric point, 4.9), and were not related to each other as a sialoglycoprotein is related to its asialo form. Adherent cell removal of the target bone marrow cells (to remove colony-stimulating cells) suggested that both GM-CSFs acted directly rather than by stimulating the production of GM- CSF. Mixing and titration experiments indicated that the differences in functional specificities of the two GM-CSFs (and the lack of EO-CSF associated with GM-CSF beta) were not due to the presence of specific inhibitory molecules or lower absolute levels of CSF in one fraction relative to the other. These two species of GM-CSF should be useful in separately enumerating subpopulations of different GM-progenitor cells inhuman hemopoietic disorders.

A sensitive and reliable sandwich enzyme-linked immunosorbent assay (ELISA) has been developed for recombinant human granulocyte-macrophage colony-stimulating factor (hGM-CSF). The assay is quantitative between 100 pg/mL and 2.5 ng/mL for bacterially synthesized hGM-CSF in human serum and is more sensitive and specific than the semisolid agar bioassay. As part of a phase I study, the pharmacokinetics of intravenous (IV) bolus injection and subcutaneous (SC) administration of hGM-CSF were studied. Following a single IV dose, an initial high blood level of hGM-CSF occurred, followed by a rapid decrease occurring in two apparent phases with a half-life (t1/2)alpha of less than five minutes and a t1/2 beta of 150 minutes. After an SC injection, detectable serum levels occurred within 15 to 30 minutes, and serum levels were sustained for a variable time depending on the dose. At the highest SC dose (10 micrograms/kg), a serum level of greater than 1 ng/mL (65 pmol/L) was maintained for greater than 12 hours after a single injection. This corresponds to the concentration of hGM-CSF supporting near-maximum proliferation in vitro.

It has been suggested that 12–0-tetradecanoylphorbol-13-acetate (TPA) may stimulate the proliferation of granulocyte-macrophage (GM) colony- forming cells (CFC) via the GM colony-stimulating factor (CSF) receptor. GM-CFC in unfractionated mouse bone marrow and light density fetal liver (LDFL) cells were induced by TPA to form colonies in the absence of exogenously added GM-CSF. The colonies induced by TPA (10(- 8)M) were smaller than normally seen with maximal concentrations of GM- CSF, and less than 30% of the GM-CFC formed colonies in the presence of TPA. The number of colonies stimulated by TPA in the absence of GM-CSF was dependent on the number of cells plated. When fewer than 10,000 bone marrow cells or 3000 LDFL cells were plated in the 1-ml semisolid agar cultures, no colonies were stimulated by the TPA. Similarly, GM- CFC purified from the LDFL cells stimulated with TPA did not form colonies. However, when the fetal liver accessory cells (macrophages) were recombined with cell-sorter-purified GM-CFC, colony formation was again observed in the presence of TPA (10(-7)-10(-8) M). The number of colonies formed from the CFC was dependent on the number of accessory cells present, suggesting that the macrophages were induced by TPA to produce CSF. Although the purified GM-CFC required CSF for proliferation, TPA (10(-8) M) increased (5–10-fold) the sensitivity of the GM-CFC to GM-CSF. These observations indicate that TPA does not stimulate GM-CFC proliferation directly, but rather by inducing GM-CSF production by accessory cells and by increasing the responsiveness of GM-CFC to GM-CSF.

Granulocyte-macrophage colony stimulating factor (GM-CSF) stimulates the in vitro proliferation and differentiation of granulocytic and macrophage cells. This regulator is now known to act at other levels of hemopoietic regulation. The heterogeneity of GM-CSFs is not only related to the tissue of origin and the in vitro production method, but also to functional subclasses of the molecule that have distinct biologic specificities. Most adult mouse organs produce GM-CSF (mol wt 23,000), but a macrophage (M)-CSF has been detected in fetal conditioned medium (CM) and isolated from L-cell CM. Murine endotoxin serum appears to contain a M-CSF, GM-CSF, and G-CSF, the last of which cofractionates with a differentiation factor active on leukemic cells. Human GM-CSFs, G-CSF, and EO-CSFs active on human cells have been detected in a variety of CM, but as yet none have been purified. Again, there are subclasses of progenitor cells that respond to particular forms of human active CSFs. GM-CSF isolated from mouse lung CM stimulates multipotential progenitor cells, the initial proliferatin of progenitors in the erythroid, eosinophil, and megakaryocyte series, as well as mature cells in the GM series. While GM-CSF is also able to stimulate the differentiation of myeloid leukemic cells, other factors appear to be more potent in this respect. Information on the regulation of GM-CSF production, on the modulators of its action on specific target cells, and on its role in vivo will be required before the physiologic function of this molecule can be properly assessed.

Human granulocyte-macrophage colony stimulating factor (GM-CSF) has been synthesized in high yield using a temperature inducible plasmid in Escherichia coli. The human GM-CSF is readily isolated from the bacterial proteins because of its differential solubility and chromatographic properties. The bacterially synthesized form of the human GM-CSF contains an extra methionine residue at position 1, but otherwise it is identical to the polypeptide predicted from the cDNA sequence. The specific activity of 2.9 X 10(7) units/mg of protein for purified bacterially synthesized human GM-CSF indicates that despite the lack of glycosylation, the molecule is substantially in its native conformation. This molecule stimulated the same number and type of both seven- and 14-day human bone marrow colonies as the CSF alpha preparation from human placental conditioned medium. Human GM-CSF had no activity on murine bone marrow or murine leukemic cells. There was no detectable, direct stimulation of adult human erythroid burst forming units (BFU-E) by the bacterially synthesized human GM-CSF. Although impure preparations containing native human GM-CSF (eg, human placental conditioned medium) stimulated the formation of mixed colonies, even in the presence of erythropoietin, the bacterially synthesized human GM-CSF failed to stimulate the formation of mixed colonies from adult human bone marrow cells. The bacterially synthesized human GM-CSF increased N-formyl-methionyl-leucyl- phenylalanine (FMLP)-induced superoxide production and lysozyme secretion. Antibody-dependent cytotoxicity and phagocytosis by human neutrophils was stimulated by the bacterially synthesized human GM-CSF and eosinophils were also activated in the antibody-dependent cytotoxicity assay.

Medium conditioned by human placental tissue was found to stimulate granulocytic and monocytic colony formation by human marrow cells in semisolid agar cultures. The colony-stimulating activity of unfractionated conditioned medium was equivalent to the activity of standard peripheral blood underlayers. Placentas were a reliable source of active material, and one placenta provided enough material to stimulate 5,000–10,000 cultures of normal or leukemic cells. The colony- stimulating factor in human placental conditioned medium (CSFHPCM) was concentrated and purified 1800-fold using ammonium sulfate precipitation, calcium phosphate gel absorption, DEAE-cellulose batch absorption, gel filtration on Sephadex G-150, and polyarcylamide gel gel electrophoresis. The active factor behaved on gel filtration as a macromolecule with an apparent molecular weight of 30,000 daltons. The active factor in placental conditioned medium was not dependent on the presence of adherent marrow cells with endogenous colony-stimulating activity.

A microculture assay for murine granulocyte-macrophage colony- stimulating factor (GM-CSF) has been developed using fetal liver GM colony-forming cells (CFC) isolated by fluorescence-activated cell sorting. These GM-CFC are free of mature hemopoietic cells, such as granulocytes and macrophages, which may interfere with direct assays for GM-CSF. The assay procedure allows the quantitation of GM-CSF within 48 hr by measuring the number of cells produced from 50 GM-CFC in microcultures (15 microliter). The assay is particularly simple to set up and score and yet, because of the reduced volumes, this assay is still capable of detecting 0.2 pg (i.e., 0.2 U) of GM-CSF within 48 hr, i.e., 100 times less GM-CSF than the conventional soft agar assay. By allowing the microcultures to develop for 7 days, the extra proliferation allows a further tenfold increase in the sensitivity of CSF detection. The time and cost of setting up hundreds of GM-CSF assays for fractions from chromatographic columns, e.g., reverse phase high performance liquid chromatography, is reduced by at least five- fold. Enough GM-CFC can be isolated and stored frozen in one afternoon to provide sufficient cells for the daily assay of 200 samples of GM- CSF for several months. Microassay results for several sources of GM- CSF at different stages of purification are compared to the results obtained from the soft agar assay.

Hemopoietic progenitor cells (colony- and cluster-forming cells in semisolid agar) were purified from light density CBA murine fetal liver cells using fluorescein-conjugated pokeweed mitogen (PWM) and a rhodamine-conjugated antineutrophil serum sandwich (alpha N) and three- parameter fluorescence-activated cell sorting. All clonable progenitor cells were highly enriched (36–50-fold) in PWM-positive (greater than channel 15), alpha N-negative (less than channel 30) fractions with relatively high intensity (greater than 100) low angle light scatter. No separation was achieved between different types of progenitor cells (granulocyte-macrophage and erythroid colony-forming cells). The enriched fraction was a pure population of large, basophilic, undifferentiated blast cells, and in agar cultures stimulated with colony-stimulating factors, up to 90% of the enriched cells were hemopoietic progenitor cells capable of varying levels of clonal proliferation. Further fractionation based on increasing fluorescence with PWM separated into discrete populations, nonproliferative morphologically recognizable erythroid cells, late erythroid progenitor cells (day 2 CFU-E), and cells forming pure or mixed erythroid burst colonies. In addition, the majority of pluripotential hemopoietic stem cells (CFU-SS) were clearly separated from progenitor cells forming colonies in vitro. The present techniques provide suitable numbers of enriched progenitor cells for a variety of biological and biochemical studies.

There has been recent interest in the synergistic interactions between the growth factors involved in the in vitro control of hematopoiesis and other cell lineages. As a convenient model system, such interactions governing the DNA synthesis in murine bone marrow-derived macrophages (BMMs) were studied. By themselves, murine colony- stimulating factor-1 (CSF-1) and recombinant murine granulocyte- macrophage CSF (GM-CSF) were stimulators of DNA synthesis in quiescent or noncycling BMMs, whereas recombinant murine interleukin-3 (IL-3) and the phorbol ester, 12-O-tetradecanoyl-phorbol-13-acetate (TPA), were weak mitogens. On the other hand, murine granulocyte CSF (G-CSF), concanavalin A (Con A), and lipopolysaccharide (LPS) were inactive on their own. When the quiescent BMMs were exposed to combinations of the CSFs, there were striking synergistic effects for both GM-CSF and IL-3 with suboptimal doses of CSF-1, with a smaller effect for GM-CSF with IL-3 and little or no effect for CSF-1 with G-CSF. CSF-1, GM-CSF, and IL-3 could also synergize with TPA; CSF-1 cooperated with 1-oleoyl-2- acetylglycerol (OAG), both sets of results pointing to an interaction with protein kinase C. LPS completely abolished the CSF-1-mediated stimulation of DNA synthesis. We propose that BMMs are suitable normal cells in which to examine in depth the various mechanistic possibilities for these interactions.