IFN-gamma regulates murine interferon-inducible T cell alpha chemokine (I-TAC) expression in dendritic cell lines and during experimental autoimmune encephalomyelitis (EAE).

Murine interferon-inducible T cell alpha chemokine (I-TAC) is a potent non-ELR Cys-X-Cys (CXC) chemokine that predominantly attracts activated T lymphocytes and binds to the receptor CXCR3. Using semiquantitative reverse transcription-polymerase chain reaction (RT-PCR) we analysed murine I-TAC expression in two different progenitor dendritic cell (DC) lines, MTHC-D2 and JAWS II which were exposed to various cytokines, and Con A-activated splenocytes from a panel of knockout mice. Analysis of the progenitor DC lines and Con A cultures demonstrated that murine I-TAC is primarily regulated by interferon (IFN)-gamma via interferon regulatory factor (IRF)-1. It has been proposed that I-TAC may have a role in autoimmune diseases such as multiple sclerosis (MS). Because I-TAC appears to be secreted from antigen-presenting cells (APCs) and attracts activated T cells, we examined the level of murine I-TAC mRNA in the central nervous system (CNS) of wild-type and IFN-gamma-receptor knockout (IFN-gammaR-/-) mice with myelin oligodendrocyte glycoprotein (MOG)35-55 peptide-induced experimental autoimmune encephalomyelitis (EAE). Peak I-TAC expression was detected in wild-type mice on day 14 when the mice begin to recover, whereas very low levels of I-TAC were detected in the CNS of IFN-gammaR-/- mice which develop severe EAE and die. The expression characteristics of murine I-TAC suggest an important mediator of immune cell communication that could augment vaccines and autoimmune therapies.

Macrophage-induced muscle pathology results in morbidity and mortality for Ross River virus-infected mice.

Ross River virus (RRV) is an Australian alphavirus that is often responsible for chronic epidemic polyarthritis and myalgia in humans. Past studies have shown severe disruption of striated muscle fibers to be prominent in RRV pathology in mice; in the present study, macrophages were directly implicated as the primary mediators of muscle damage. General immunosuppressive therapies had only minor effects on mortality and morbidity in RRV-infected mice, with no inhibition of muscle damage. Treatment of mice with macrophage-toxic agents (e.g., silica) prior to RRV infection completely abrogated disease symptoms without significantly affecting titers of virus in organs. Further studies found that clinical signs of infection and muscle damage correlated with a massive influx of macrophages into hind leg muscle, whereas no such infiltrate or damage was observed for silica-treated mice. These observations are significant for the human disease context, as monocytic cells have been detected in the synovial effusions of persons with epidemic polyarthritis.

Myb-transformed hematopoietic cells as a model for monocyte differentiation into dendritic cells and macrophages.

Immune induction is effected through the interaction of antigen-presenting cells with specific receptors on the surface of thymus-derived lymphocytes. Cells most able to ingest, process, and present antigen appear to be related to the mononuclear phagocyte/neutrophil series. For example dendritic cells (DC) can be found in colonies of GM-CSF-responsive bone marrow cells, and under experimental conditions are routinely expanded as a population in vitro from GM-CSF-responsive progenitor cells. To address the question of DC lineage and to determine what genes are involved in lineage commitment, we have generated a series of GM-CSF-responsive cell lines that can be induced to differentiate in a homogeneous manner in vitro. The cloned cell lines are derived from 12-day fetal liver and are transformed with a truncated form of c-myb, which lacks the normal autoregulatory sequences. As far as we know, these myb-transformed hemopoi-etic cells (MTHC) differ from normal only in the unregulated expression of myb, a gene whose expression is obligatory for proliferation of hemopoietic cells. MTHC in the presence of TNF-alpha and IL-4 will differentiate into cells that have many of the properties of macrophages. When the same MTHC lines are exposed to TNF-alpha in combination with IFN-gamma, the cells instead become DC. The differentiated DC are potent presenters of antigen in mixed lymphocyte reactions and of soluble antigen to specific T cell lines. Thus, cells with the properties of both macrophages and DC can be derived from a single type of GM-CSF-responsive progenitor cell. We have used this MTHC system to analyze differences in gene expression as the cells mature along the DC and macrophage pathways. A distinctive pattern of differentially expressed cDNAs is evident where macrophage-specific cDNAs are homologous to genes encoding cytoskeletal and cell-surface proteins, whereas the DC-specific cDNAs are homologous to signaling, chemokine, and IFN-gamma-inducible genes. We discuss the utility of MTHC in analyzing the relationships between DC and macrophages, and suggest that DC and macrophages represent extreme phenotypes in a spectrum of antigen handling cells that are somewhat interchangeable, depending on their immediate environment.

Immune mechanisms associated with the rejection of fetal murine proislet allografts and pig proislet xenografts: comparison of intragraft cytokine mRNA profiles.

BACKGROUND: Previous in vivo depletion studies of CD4 and CD8 T cells indicated that different rejection mechanisms operate for proislet allografts and xenografts. The cellular and molecular mechanisms of acute proislet allograft and xenograft rejection have therefore been characterized and directly compared. METHODS: The intragraft cytokine mRNA profile in rejecting BALB/c (H-2d) proislet allografts was analyzed in control, CD4 T cell-depleted, and CD8 T cell-depleted CBA/H (H-2k) recipient mice using semi-quantitative reverse transcriptase-assisted polymerase chain reaction (RT-PCR). The cytokine profiles for proislet allografts and pig proislet xenografts at 3-10 days posttransplant were directly compared and correlated with graft histopathology. RESULTS: Allograft rejection was protracted (2-3 weeks), characterized by infiltrating CD8 T cells and CD4 T cells (no eosinophils) and was associated with a Th1-type CD4 T cell response (IL-2, IFN-gamma, and IL-3 mRNA) and a CD8 T cell-dependent spectrum of cytokine gene expression (IL-2, IFN-gamma, IL-3, and IL-10 mRNA). Xenograft rejection was rapid (6-8 days), involved predominantly CD4 T cells and eosinophils, and in contrast to allografts, exhibited intragraft mRNA expression for the Th2 cytokines IL-4 and IL-5. CONCLUSIONS: Proislet allograft and xenograft rejection differ in the tempo of destruction, phenotype of the cellular response and intragraft profile of cytokine mRNA. The recruitment of eosinophils only to the site of xenorejection correlates with IL4 and IL-5 mRNA expression. These findings suggest that different anti-rejection strategies may need to be developed to optimally target the allograft and the xenograft response.

Intragraft expression of cytokine transcripts during pig proislet xenograft rejection and tolerance in mice.

The rejection of pig proislet (islet precursor) xenografts in CBA/H mice is a CD4+ T cell-dependent process. The molecular mechanisms of xenograft rejection, xenograft survival during anti-CD4 mAb therapy and xenograft tolerance post-withdrawal of anti-CD4 mAb administration, were examined by using a semiquantitative PCR method. Temporal analysis of intragraft cytokine mRNA demonstrated a Th0-like pattern of expression (IL-2, IFN-gamma, IL-3, IL-4, IL-5, and IL-10) on day 4 of the acute xenograft rejection process. From day 5, however, only Th2-associated transcripts (IL-3, IL-4, IL-5, and IL-10) were enhanced in xenografts compared with isograft controls. Immunohistochemistry showed that the principal participants in the rejection infiltrate were CD4+ T cells and eosinophils, with smaller numbers of CD8+ T cells. In vivo depletion of CD4+ T cells prevented xenograft rejection but had minimal effect on the peak levels of IL-2, IFN-gamma, and IL-10 mRNA; in contrast, the enhanced expression of IL-3, IL-4, and IL-5 transcripts seen in rejecting xenografts was abrogated. This established a positive correlation between acute xenograft rejection, presence of CD4+ T cells, and enhanced intragraft expression of mRNA for the Th2-type cytokines IL-3, IL-4, and IL-5. In tolerant hosts, long-term proislet xenograft survival and function (> 190 days) was accompanied by intragraft expression of IL-2 and IL-10 mRNA; IFN-gamma, IL-3, IL-4, and IL-5 mRNA were either undetected or not enhanced. The induced rejection of long-term functioning xenografts (> 170 days) in nontolerant hosts resulted in selective enhancement of IL-4 transcript expression. This study suggests that Th2-like CD4+ T cells are differentially activated in response to xenoantigen and that xenograft tolerance is associated with lack of expression of the Th2 cytokine, IL-4.

Augmentation of tumor necrosis factor-alpha-induced monocytic differentiation of a myelomonocytic leukemia (WEHI-3B JCS) by pertussis toxin.

We have shown that pertussis toxin (PTx) modulates the effect of tumor necrosis factor-alpha (TNF-alpha) in inducing monocytic differentiation of WEHI-3B (JCS) myeloid leukemic cells in vitro. PTx (0.1-2 ng/ml) alone was not cytotoxic and did not induce morphological changes in JCS cells. In the presence of a suboptimal concentration of TNF-alpha (25 U/ml), however, PTx (1 ng/ml) acted synergistically in inhibiting proliferation and in inducing monocytic differentiation of the JCS cells. Expression of the macrophage differentiation marker (Mac-1) on JCS cells was increased by the combination of PTx and TNF-alpha, and phagocytic activity of the cells was also enhanced. Moreover, JCS cells treated with PTx and TNF-alpha had reduced tumorigenic capacity in vivo. The data suggest that a PTx-sensitive G protein may be involved in regulating the TNF-alpha-induced monocytic differentiation of the myeloid leukemic JCS cells and that combination of PTx and TNF-alpha may be useful in the treatment of some forms of myelomonocytic leukemia.

Synergistic effect of IL-4 and TNF-alpha in the induction of monocytic differentiation of a mouse myeloid leukaemic cell line (WEHI-3B JCS).

We have previously shown that non-cytotoxic concentrations (600-1200 U/ml) of recombinant mouse tumour necrosis factor-alpha (TNF-alpha) can induce differentiation of a subclone (JCS) of the WEHI-3B myelomonocytic leukaemia cell line into mature cells with the characteristics of macrophages. In the present study, the effects of recombinant mouse interleukin-4 (IL-4), either alone or in combination with mouse TNF-alpha, on the growth and differentiation of JCS cells were examined. IL-4 alone (20-5000 U/ml) inhibited the growth of JCS cells in a dose-dependent manner but did not induce cell differentiation. However, combinations of IL-4 and TNF-alpha acted in synergy to inhibit cell proliferation and induce monocytic differentiation of JCS cells, as shown by increased expression of the macrophage differentiation antigens (F4/80, Mac-1), stimulation of phagocytic activity, induction of non-specific esterase and NBT-reducing activities, increased plastic adherence and morphological criteria. Similar synergistic interactions were also shown by human TNF-alpha and mouse IL-4, indicating that TNF-alpha might exert its effects through the low-affinity (p55) TNF receptors. Moreover, the clonogenicity of JCS cells in vitro and their tumorigenicity in vivo were significantly reduced by combined TNF-alpha and IL-4 treatment. Our results indicate that TNF-alpha can act as a differential signal for JCS cells and that its effects are modulated by IL-4. Therefore, the combination of TNF-alpha and IL-4 may be useful in the treatment of some forms of myelomonocytic leukaemia.

Differentiation state and responses to hematopoietic growth factors of murine myeloid cells transformed by myb.

Murine hematopoietic cells can be transformed in vitro by recombinant retroviruses that express the myb oncogene, and hematopoietic growth factor (HGF)-dependent myeloid cell lines can be derived from these transformed primary cells. In this study, the differentiation state and responses of myb-transformed hematopoietic cells (MTHCs) have been investigated. We find that MTHCs exhibit properties of early myeloid progenitors including synergistic responses to combinations of HGFs and expression of certain surface markers. As reported previously, MTHCs respond well to granulocyte-macrophage colony-stimulating factor (GM-CSF) but can also respond to interleukin-3 (IL-3); the response to the latter factor depends on the mouse strain from which the cells are derived. Although these single factors stimulate MTHCs, combinations of these factors with colony-stimulating factor-1 (CSF-1 or M-CSF) or Steel factor (SLF or SCF) act synergistically to promote colony formation. The surface markers expressed by MTHCs include both granulocyte-macrophage lineage specific antigens Gr-1, 7/4, F4/80, and Mac-1, as well as two antigens found on early progenitors and stem cells--Thy-1 and Sca-1 (Ly6E). Expression of the latter markers is often heterogeneous and can be modulated by the growth factors to which the cells are exposed. Finally, we show that monocytic differentiation of MTHCs can be induced by exposure to tumor necrosis factor (TNF alpha). Taken together, these results suggest that MTHCs will be a useful model for studying HGF/cytokine responses in both proliferation and differentiation.

Monocytic differentiation of a myelomonocytic leukemic cell (WEHI 3B JCS) is induced by tumour necrosis factor-alpha (TNF-alpha).

We have isolated a subclone (JCS) of the WEHI 3B myelomonocytic leukemia, which acquires the characteristics of mature macrophage lineage cells in the presence of PMA or noncytotoxic concentrations of TNF-alpha (600-1200 U/ml). JCS cells were compared with D+ and D- subclones of WEHI 3B. Unlike D+ cells, JCS cells did not produce differentiated granulocyte-macrophage colonies in the presence of postendotoxin serum or recombinant G-CSF. Stimulation with PMA or TNF-alpha reduced proliferation of JCS cells. TNF-alpha decreased the level of cell surface J11D antigen with concurrent increased expression of Mac-1 and FcR antigens and phagocytic activity. These TNF-alpha-mediated effects were enhanced by addition of IFN-gamma to the cultures. Furthermore, differentiation-inducing activity of PMA could be prevented using neutralizing anti-TNF-alpha antibodies. The results indicate that exogenous TNF-alpha can act as a differentiative agent for JCS cells and that endogenous TNF-alpha is the active substance when PMA is used to stimulate macrophage differentiation of JCS cells.

Localization of the GTP-binding protein Gi alpha in myelomonocytic progenitor cells is regulated by proliferation (GM-CSF, IL-3) and differentiation (TNF) signals.

We have examined the role of Gi alpha in haemopoietic cells using the myelomonocytic progenitor cell lines FDC-P1 and WEHI-3B (JCS). During growth factor-dependent proliferation of FDC-P1 cells Gi alpha was found predominantly in the nucleus and associated with the plasma membrane. Following removal of growth factor, Gi alpha accumulated in the cytoplasm and at the plasma membrane. Treatment of FDC-P1 cells with pertussis toxin (PT) completely inhibited translocation of Gi alpha to the nucleus and reduced the sensitivity of FDC-P1 cells to the proliferative effects of growth factors, indicating that translocation of Gi alpha plays a regulatory role in, but may not be essential for, cell division. Gi alpha initially associated with DNA during S/G2 of the FDC-P1 cell cycle but separated from condensing chromosomes during mitosis. In contrast to FDC-P1 cells, WEHI-3B (JCS) cells proliferate in the absence of added growth factors but can be induced to differentiate by TNF-alpha. In proliferating JCS cells Gi alpha was again associated with the nucleus but when proliferation was inhibited by TNF-alpha, Gi alpha accumulated in the cytoplasm with none detected in the nucleus. Thus the cytokine regulated accumulation of Gi alpha at different intracellular sites correlated with the ability of the cell to progress through the proliferative cycle. When the tyrosine kinase inhibitor genistein was added to FDC-P1 cells prior to stimulation with IL-3 or GM-CSF, proliferation was almost completely inhibited but translocation of Gi alpha was not affected, suggesting that tyrosine phosphorylation was not involved in G protein translocation but was essential for cytokine induced cell division. Cholera toxin (CT) also inhibited proliferation of FDC-P1 cells but had no effect on translocation of Gi alpha to the nucleus. The near complete inhibition of cell division by genistein and CT without a corresponding effect on Gi alpha movement indicates that Gi alpha can be regulated independently of tyrosine kinase and adenylyl cyclase activities, respectively.

Bone marrow cells from A/J mice do not proliferate in interleukin-3 but express normal numbers of interleukin-3 receptors.

Haemopoietic cells from A/J mice do not form colonies (proliferate) in response to interleukin-3 (multi-CSF, IL-3). We have examined different populations of cells from A/J mice and shown that, despite their failure to proliferate in response to IL-3, cells from bone marrow, spleen and the peritoneum all bound 125I-labelled IL-3. A wide variety of cell types bound IL-3 as determined by autoradiography, including promyelocytes, myelocytes, metamyelocytes, polymorphs, promonocytes, monocytes, eosinophils and lymphocytes, but not nucleated erythroid cells, and the proportion of each cell type binding label was similar when cells from A/J mice were compared with those of C57B1/6 and Balb/c mice. Bone marrow cells from A/J mice internalized interleukin-3 with normal kinetics and mRNA extracted from these cells contains the same species of IL-3 receptor and IL-3 receptor-like mRNAs as are found in the other strains. Collectively the data suggest that the failure of haemopoietic cells from A/J mice to proliferate in response to IL-3 is related to a selective defect in signalling to proliferation specific genes. This defect is apparently not related to internalization or processing of the IL-3/IL-3-receptor complex, but may be due to failure to activate appropriate accessory molecules in the cell.

Cytokine immunoreactivity in plasma does not change after moderate endurance exercise.

We investigated whether increased concentrations of circulating cytokines may be responsible for exercise-induced priming of blood neutrophils (J. A. Smith et al. Int. J. Sports Med. 11: 179-187, 1990). The plasma concentrations of tumor necrosis factor-alpha, interleukin- (IL) 1 beta, IL-6, granulocyte-macrophage colony-stimulating factor, and neopterin in trained and untrained human subjects were measured by immunoassay before and after 1 h of cycling at 60% of maximal oxygen uptake. C-reactive protein and creatine kinase (CK) were also measured before and 24 h after exercise as markers of the "acute-phase response" and muscle damage (C. Taylor et al. J. Appl. Physiol. 62: 464-469, 1987), respectively. The small changes in the plasma concentrations of cytokines or neopterin observed after exercise in both trained and untrained subjects were not significantly different to those found in a control group of nonexercised subjects. However, untrained subjects did exhibit an acute-phase response (P = 0.04) 24 h after exercise without additional release of CK into plasma. Baseline training differences were confined to a twofold elevation in CK activity (P = 0.04). The results show that circulating cytokines are unlikely to be responsible for the priming of neutrophil microbicidal activity observed after moderate endurance exercise (J. A. Smith et al. Int. J. Sports Med. 11: 179-187, 1990).

Distinguishing between mouse IL-3 and IL-3 receptor-like (IL-5/GM-CSF receptor converter) mRNAs using the polymerase chain reaction method.

A set of primers (MF43, MF44 and MF45) were designed and used in the polymerase chain reaction to distinguish between the expression of mouse IL-3 receptor and mouse IL-3 receptor-like mRNAs. Primers MF43 and MF45 were specific for IL-3 receptor mRNA while the primers MF44 and MF45 were specific for IL-3 receptor-like mRNA. Primers MF44 and MF45 could not amplify IL-3 receptor cDNA even at an annealing temperature of 46 degrees C which is 20 degrees C below the melting temperature of the primers, or at high template concentrations (up to 100 ng cDNA). The optimal range of Mg2+ concentrations for the two pairs of primers MF43, MF45 and MF44, MF45, were essentially the same and this permits comparisons of the expression level of these two mRNAs under identical PCR conditions. Both the IL-3 receptor and IL-3 receptor-like mRNAs could be detected in normal bone marrow cells and IL-3-dependent cell lines (FDC-P1 and 32D cl-23), as well as in the IL-3 independent cell lines P388D1 and WEHI-3B, the latter being a constitutive producer of IL-3. In contrast, neither species of mRNA was detected in the T lymphoma cell line (EL-4). The ratio of IL-3 receptor-like mRNA to IL-3 receptor mRNA was usually greater than 1, except in 32D cl-23 cells where it was 0.66.

Decreased expression of J11d antigen during monocytic differentiation of M1 myeloid leukemic cells.

Four myeloid cell lines (M1, WEHI-3B D+, FDC-P1, and 32D) were screened for the presence of J11d antigen. One of these cell lines, the myeloid leukemia M1, was found to express a high level of J11d antigen on the cell surface. Recombinant mouse leukemic inhibitory factor (rm-LIF), recombinant human LIF (rh-LIF), and steroids (hydrocortisone and dexamethasone) could induce M1 cells to undergo monocytic differentiation. The level of J11d antigen was greatly reduced after treatment of the cells with LIF or steroids. Western blotting revealed that the apparent molecular weight of the J11d antigen on M1 cells was 45-48 kDa. Furthermore, the level of J11d mRNA was also reduced during LIF-induced differentiation of M1 cells.

Enhanced mucosal cytokine production in inflammatory bowel disease.

Proliferation, maturation, chemotaxis, and activation of neutrophils and monocytes are mediated largely by cytokines, including colony-stimulating factors and lymphokines. Cytokines produced in the intestinal mucosa contribute to the increased migration of neutrophils and monocytes into the lesion of inflammatory bowel disease and to the activation of these inflammatory cells. Lamina propria mononuclear cells isolated from colon tissue from 14 patients with inflammatory bowel disease (IBD) and from histologically normal controls were studied. Cells from IBD-affected tissue produced significantly more colony-stimulating factor activity (1402 +/- 252 U) per 2 x 10(6) cells than those from normal mucosa (362 +/- 85 U), mainly because of the increased production of granulocyte colony-stimulating factor and interleukin 1. This was accompanied by increases in the amount of specific messenger RNA for these two cytokines in lamina propria mononuclear cells from mucosa of patients with Crohn's disease (CD) compared with normal controls. By contrast, there was a substantial reduction in interleukin 3 production in CD and in ulcerative colitis lamina propria mononuclear cells, and this was reflected in significantly reduced expression of interleukin 3 messenger RNA in CD cells. Of the agents used in the therapy of IBD, hydrocortisone and 5-aminosalicylic acid, but not cyclosporin A, markedly suppressed in vitro production of cytokines by lamina propria mononuclear cells, suggesting that their therapeutic efficacy in vivo may be due in part to down-regulation of cytokine production in the inflamed mucosa.

Synergistic effects of interleukin-1 beta and interleukin-3 on the expansion of human hematopoietic progenitor cells in liquid cultures.

In the present study, we show that recombinant human interleukin-1 beta (rhIL-1 beta), which has no effect on the proliferation of human progenitor cells, has synergistic effects on the expansion of human progenitor cells induced by rhIL-3 in liquid cultures. The synergistic effects of rhIL-1 beta with rhIL-3 were observed in liquid cultures using not only fresh bone marrow mononuclear cells, but also selected populations of nonadherent cells, non-T nonadherent cells, and CD34-positive cells. Anti-granulocyte-macrophage colony-stimulating factor (anti-GM-CSF) antibody partially blocked the synergistic effects of rhIL-1 beta on the proliferation of colony-forming unit (CFU)-GM burst-forming unit-erythroid (BFU-E), and CFU-Mix in liquid cultures in the presence of rhIL-1 beta plus rhIL-3, suggesting that the synergistic effects of rhIL-1 beta plus rhIL-3 are explained in part by the secondary production of GM-CSF. Limiting dilution assays and liquid culture assays using CD34-positive cells indicate that rhIL-1 beta directly increases the numbers of colony-forming cells in liquid cultures. These results suggest that rhIL-1 beta has unique direct and indirect effects on the expansion of hematopoietic progenitor cells in liquid cultures.

Intestinal lymphokine-activated killer cells in inflammatory bowel disease.

The role of non-specific cytotoxicity in the pathogenesis of inflammatory bowel disease (IBD) was investigated by assaying the natural killer (NK) and lymphokine-activated killer (LAK) cell activity of lamina propria mononuclear cells (LPMC) from 22 specimens of intestinal mucosa affected by IBD. Only minimal levels of NK activity were detected against K562 cells, as well as colon carcinoma cells, adenoma cells and fibroblasts freshly isolated from the intestinal mucosa. Culture of LPMC from IBD in the presence of interleukin-2 (IL-2) generated LAK cells that mediated high levels of activity against K562 cells and against neoplastic epithelial cells and fibroblasts derived from the intestinal mucosa. A group of 20 histologically normal specimens of intestinal mucosa showed similar levels of LAK activity against the K562 and intestinal cell targets. The minimal mucosal NK activity in IBD suggests that the cytotoxic properties of NK cells are not important in the pathogenesis of IBD. The presence of LAK precursor cells in the inflamed mucosa of IBD and their ability to lyse biologically relevant targets in vitro suggests that LAK cells have the potential to contribute to intestinal mucosal injury in IBD.

Gliotoxin treatment selectively spares M-CSF- plus IL-3-responsive multipotent haemopoietic progenitor cells in bone marrow.

Gliotoxin, an epipolythiodioxopiperazine, is a fungal metabolite that causes genomic DNA degradation preferentially in certain blood cell types including T lymphocytes and macrophages. Gliotoxin has previously been used to treat murine allogeneic bone marrow prior to transplantation into irradiated recipients, and in this situation the drug prevents development of graft-versus-host disease, and permits the establishment of allogeneic bone marrow chimeras. We have examined the nature of the cells that survive gliotoxin treatment and report here that gliotoxin selectively spares a unique class of haemopoietic stem cell that forms large (HPP) colonies in the presence of mixtures of M-CSF and IL-3. We confirm that the cells which survive gliotoxin treatment are capable of reconstituting the haemopoietic system in allogeneic lethally irradiated mice.