T cell and monocyte-derived burst-promoting activity directly act on erythroid progenitor cells

Signal transduction in primary cultured human erythroid cells

Development of erythrocytes is a complex process governed by multiple cytokines. Colony assays have revealed the physiologic importance of these cytokines, although biochemical studies of highly purified human colony-forming unit-erythroid (CFU-E) generated in vitro from CD34+ cells have only recently begun. Studies from our groups and others suggested that signal transduction in primary erythroid cells differs considerably from that in cell lines or primary cells from other species. In this review, we summarize results of these studies with emphasis on possible implications for hematotherapy.

T lymphocyte subsets in chronic uremic patients treated with maintenance hemodialysis

Blood T lymphocyte subsets have been studied using monoclonal antibodies in 10 chronic uremic patients treated with maintenance hemodialysis. Both total T lymphocytes identified by the antibody OKT3, and the helper-inducer T lymphocyte subset identified by the antibody OKT4 were found to be significantly lower than normal. The cytotoxic-suppressor T cell subset was only moderately, even if significantly reduced, so that the T4/T8 ratio in uremic patients was significantly lower than normal. These data provide an additional contribution to the interpretation of immunological and hematological deficiencies observed in chronic uremia.

The effects of gestation on circulating progenitor cells

The frequency of BFU-E in second-trimester fetal blood (484 +/- 104/10(5)) falls progressively during gestation to a value of 69 +/- 41/10(5) in cord bloods of 36 weeks gestation and beyond, but this is still significantly greater than adult blood values of 14 +/- 8 (P < 0.01). BFU-E obtained from unfractionated peripheral blood mononuclear cells from fetuses/neonates less than 36 weeks gestation were more sensitive to erythropoietin than adult BFU-E, but the sensitivity of highly purified BFU-E obtained from second-trimester fetal liver was similar to that in adult cells. Almost maximal growth of BFU-E from purified fetal progenitor cells could be achieved with erythropoietin alone, whereas adult cells required the presence of other factors with 'burst-promoting activity'.

Burst-promoting activity in the serum of patients with chronic renal failure undergoing long-term hemodialysis

Anemia in patients with chronic renal failure (CRF) is multifactorial, and while the majority will respond to the paternal administration of recombinant human erythropoietin (r-HuEPO), a role for coexistent plasma inhibitors and stimulators, such as burst-promoting activity (BPA), remains controversial. To evaluate the latter possibility, eight individuals with CRF on long-term hemodialysis, were studied before (mean hemoglobin 58.4 +/- 8.0 g/l and after 12 weeks of r-HuEPO therapy (mean hemoglobin 100.4 +/- 18.3 g/l). In vitro erythroid cultures using erythroid burst forming unit (BFU-E) and erythroid colony forming unit (CFU-E) assays were performed, plating 5 x 10(4) bone marrow mononuclear cells and comparing growth in heat-inactivated autologous serum with AB serum. Using Step III sheep erythropoietin (Connaught Laboratories, Willowdale, Ontario, Canada) (n = 4), mean BFU-E pre-therapy were 89.7 +/- 75.1, CFU-E were 418.5 +/- 150.6, whereas the corresponding figures in AB serum were 2.5 +/- 2.9 and 197 +/- 94.19, p = 0.1, p = 0.01, respectively. Similarly, with r-HuEPO (EPOCONN: Connaught Laboratories, Willowdale, Ontario, Canada) (n = 4), mean BFU-E were 145.25 +/- 103.3 in autologous serum and 31.0 +/- 56.75 in AB serum (p = 0.04). As controls, erythroid progenitors from two normal donors yielded 69 and 61 BFU-E colonies in autologous serum and 103 and 42 in AB serum; the corresponding CFU-E were 52 and 235 versus 136 and 137.(ABSTRACT TRUNCATED AT 250 WORDS)

Interactions between purified human cord blood haemopoietic progenitor cells and accessory cells

The role of accessory cells in haemopoiesis remains confused. This appears in large part to reflect the use of impure populations of accessory cells and progenitor cells in previous studies. In this study, cell sorter purified populations of both accessory cells and haemopoietic progenitor cells were used to examine interactions between these cell types. We used a double culture protocol in which purified CD34+ cells were cultured with purified NK, T or monocytic cells in the first liquid culture phase after which the cells were transferred to secondary agar cultures to determine the number of colony forming cells (CFC). NK cells co-cultured with CD34+ cells resulted in an increased number of erythroid progenitors with no effect on the number of nonerythroid progenitors. In contrast, there were increased numbers of erythroid and non-erythroid CFC when the CD34+ cells were co-cultured with either purified T cells or monocytes. CD34+ cells cultured with cell-conditioned media derived from NK cells, T cells or monocytes in transwells where the CD34+ cells and the accessory cells were separated by a 0.2 micron membrane, showed no enhancement in CFC. These results suggest that intimate cell-cell contact is required for these events.

Effect of cimetidine on burst-promoting activity of normal T lymphocytes

The effect of cimetidine, an inhibitor of suppressor T lymphocytes, on the burst-promoting activity (BPA) of normal T lymphocytes has been studied. Cimetidine has been shown to increase the BPA of normal T lymphocytes, both when added to the culture and when T lymphocytes were preincubated for 1 h with it. Cimetidine had no direct effect on the in vitro growth of burst-forming units (BFU-E). Our results show that CD8 suppressor T lymphocytes inhibit the in vitro growth of BFU-E in normal conditions, either directly or through inhibition of BPA of CD4 helper T lymphocytes. Cimetidine has proved to be a useful tool for investigating the hematological role of T-lymphocyte subsets in normal and pathological conditions.

"Pure" human hematopoietic progenitors: permissive action of basic fibroblast growth factor

Methodology has been developed that enables virtually complete purification and abundant recovery of early hematopoietic progenitors from normal human adult peripheral blood. A fraction of the pure progenitors is multipotent (generates mixed colonies) and exhibits self-renewal capacity (gives rise to blast cell colonies). This methodology provides a fundamental tool for basic and clinical studies on hematopoiesis. Optimal in vitro cloning of virtually pure progenitors requires not only the stimulatory effect of interleukin-3, granulocyte-macrophage colony-stimulating factor, and erythropoietin, but also the permissive action of basic fibroblast growth factor. These findings suggest a regulatory role for this growth factor in early hematopoiesis.

Autoreactive HLA-DR-specific autoreactive T-cell clones: possible regulatory function for B lymphocytes and hematopoietic precursors

The physiological significance of autoreactive T cells derived from normal individuals and activated in the absence of any identifiable foreign antigen by class II MHC-syngeneic molecules remains unexplained. Here we report that autoreactive T-cell clones (Tilkin et al. 1987) proliferate and are able to kill autologous or syngeneic EBV-cell lines but not autologous or syngeneic HLA-Class II-positive PHA-activated T-cell blasts. Furthermore, they are able to efficiently inhibit in vitro the differentiation of CFU-GM and BFU-E colonies, in agreement with the well-known observation that hematopoietic precursors express HLA-DR molecules (Cannistra et al. 1986). The reasons why the autoreactive clones do not recognize T-cell blasts, as well as their possible implications in regulatory mechanisms involving HLA-class II molecules are discussed.

Purification of human blood burst-forming units-erythroid and demonstration of the evolution of erythropoietin receptors

To facilitate the direct study of the molecular events that control the development of human burst-forming units-erythroid (BFU-E), we have developed a method to purify BFU-E from peripheral blood. Using density centrifugation, rosetting with a mixture of neuraminidase-treated and IgG-coated sheep erythrocytes, positive panning with anti-My10 monoclonal antibody, overnight adherence to plastic dishes, negative panning with monoclonal antibodies, and density centrifugation, human blood BFU-E were purified from 0.04% to 56.6%, a 1,400-fold purification with a 13% yield. More than 90% of purified BFU-E were recombinant interleukin-3 (rIL-3) dependent, which survived for 48 h with rIL-3 in the absence of recombinant erythropoietin (rEP), and 80% gave rise to erythroid bursts of more than 500 hemoglobinized cells. rEP dependency was not evident until after 72 h of incubation in vitro. The purified cells (day 1) were incubated with rIL-3 and rEP in liquid culture for 24 (day 2), 48 (day 3), and 72 (day 4) h and then were transferred into semisolid cultures and incubated until day 15. The size of the erythroid colonies observed in semisolid cultures decreased continuously in association with the incubation time of day 1 purified cells in liquid cultures. The first appearance of colony-forming units-erythroid (CFU-E) that gave rise to colonies of 8 to 49 cells was observed after 72 h of incubation of day 1 cells in the liquid culture. 125I-rEP was incubated for 5 h at 37 degrees C with purified cells (day 1) or with the cells that had been incubated in liquid culture for an additional 24-72 h, and the presence of erythropoietin (EP) receptors was investigated using autoradiography. Specific binding of 125I-rEP was detected in 19 +/- 7% of the initial day 1 BFU-E. The percentage of 125I-rEP-binding to erythroid progenitor cells and the amount of binding continuously increased as day 1 BFU-E matured. 125I-rEP specific binding was observed with all of the erythroid progenitor cells that had been incubated in liquid culture for 72 h. These data demonstrate that primitive BFU-E have a much lower number of EP receptors than CFU-E and develop an increased concentration of EP receptors in association with their maturation and loss of proliferative capacity.

Paroxysmal nocturnal hemoglobinuria: the biochemical defects and the clinical syndrome

Paroxysmal nocturnal hemoglobinuria is a disorder characterized by the lack of membrane proteins affixed to the membrane by an anchor dependent upon phosphatidyl inositol, suggesting that some acquired abnormality in the metabolism of this class of proteins is basic to the disease. Most of the clinical symptoms can be explained by the lack of these proteins. However, much work is needed to understand completely the relationship of the biochemical facts and the clinical syndrome.

Physiological concentrations of atrial natriuretic factor stimulate human erythroid progenitors in vitro

Human alpha-ANF[1-28] potentiated erythroid colony formation up to four-fold in cultures containing erythropoietin. Both early and late erythroid precursor cells responded to alpha-ANF[1-28] [0.032 to 1 nM] in a dose dependent fashion. Removal of T lymphocytes and macrophages which have been shown to modulate erythropoiesis did not abolish the stimulatory effect. All major circulatory forms of ANF (alpha-ANF[1-28], alpha-ANF[4-28] and alpha-ANF[5-28]) had potent erythropoietic activity. These results indicate that concentrations of ANF reached during hypoxia stimulate erythroid progenitor cells in the presence of erythropoietin.

Soluble factor(s) released by the PF-382 T-cell line enhances the stimulatory effect of monocytes on the BFU-E growth

PF-382 is a human T-cell line that has been shown to elaborate factors that modulate normal hemopoiesis in vitro. In the present study we report that this cell line constitutively releases in both serum-containing and serum-free supernatants a potent enhancer of BFU-E growth. The factor(s), partially purified by gel filtration, is a heat-stable molecule(s) degradable by trypsin and 2-mercaptoethanol treatments, equally active on bone marrow and peripheral blood erythroid progenitor cells, but not on CFU-GM. Unlike other sources of BPA, this stimulatory factor(s) exerts its effect in the presence of mononuclear adherent cells. In fact, the addition of conditioned medium obtained by 48 hr preincubation of isolated monocytes with 10% PF-382 supernatant (M-CM2) or the concomitant addition of supernatant from PF-382 cells (PF-382-CM) and from unstimulated monocytes (M-CM1) are capable of fully replacing the presence of monocytes in the BFU-E assay. Since the independent addition of PF-382-CM or of M-CM1 is devoid of stimulatory function, we suggest that the PF-382 derived BFU-E growth inducer, which differs from IL-1, IL-3, IL-4, GM and G-CSF, exerts its activity "via" a synergistic mechanism with a monokine.

Acetylated lipoproteins impair erythroid growth factor release from endothelial cells

Endothelial cells are a known source of hematopoietic growth-enhancing factors, including platelet-derived growth factor (PDGF). In addition, endothelium interacts directly with plasma lipoproteins which have been shown to modulate hematopoiesis. To determine the relationship of these properties, we measured the release of an erythroid growth-enhancing factor from bovine endothelial cells under lipid-loaded and control conditions. Human bone marrow cells cultured under serum-free conditions form more erythroid, granulocyte/macrophage, and mixed hematopoietic colonies when supplemented with endothelial cell-conditioned medium (ECCM) than do controls (P less than 0.05). The activity is expressed over a wide range of erythropoietin, lymphocyte-conditioned medium (LCM), recombinant human interleukin-3, and colony-stimulating factor (CSF) concentrations, and is related to ECCM dose. In contrast, enhancing activity in ECCM prepared with 0-400 micrograms/ml acetylated low density lipoproteins (AcLDL) or native LDL is diminished to 0% in a dose-dependent fashion (relative to ECCM from unexposed cells or from cells incubated with very low density lipoproteins, P less than 0.05). Upon dilution, medium prepared from cells incubated with LDL shows a rightward shift in the dose-response curve for erythroid colony formation, while that prepared from AcLDL loaded cells demonstrates a downward shift, indicating that the inhibitory activities are kinetically distinct. Delipidation of ECCM prior to addition to marrow culture removes the inhibitory action of native LDL (P less than 0.05) but not that of AcLDL (P greater than 0.10). Immunochemical analysis suggests that the erythropoietic activity in ECCM is unrelated to that of PDGF, recombinant human CSF, and erythroid burst-promoting activity (BPA) present in LCM. This conclusion is supported by Northern blot analysis of endothelial cells using a cDNA probe for the v-sis homologue of the PDGF beta chain and by immunoprecipitation of metabolically labeled PDGF. The relative amounts of c-sis transcripts and of secreted PDGF were similar in endothelial cells incubated with or without AcLDL. We conclude that AcLDL impair the synthesis or release of an erythropoietic growth-enhancing factor(s) which is biologically distinct from PDGF and BPA present in LCM.

Establishment of a hemopoietic stimulating factor producing murine leukemia cell lines: pathogenesis of granulocytosis in L8313 bearing mice

Thy 1.2+ cells of L8313 leukemia bearing mice were previously shown to produce granulocyte macrophage colony stimulating activity (GM-CSA) and interleukin-3 (IL-3). Cell lines with the Thy 1.2+ phenotype were established from spleen cells of the leukemic mice, and were designated STIL-3 C5 and STIL-3 D10. They produced and released GM-CSA and IL-3 activities in culture supernatant. C5 cells were phenotypically Thy 1.2+ and Lyt 2.1+ with rearrangement of the T-cell receptor beta chain gene also being identified. D10 was Thy 1.2+ and L3T4+ with T-cell receptor beta chain gene rearrangement not detectable. Since the same sized rearranged bands were observed between C5 and L8313 leukemic mouse spleen cells, it is indicated that C5 cells are derived from the leukemic cells. Inoculation of these cells into mice induced granulocytosis. These results indicated that L8313, which was thought to be a "granulocytic leukemia", is actually a T-cell leukemia, whose granulocytosis is a leukemoid reaction induced by normal hemopoietic cells responding to the hemopoietic stimulating factors, GM-CSA and IL-3, produced by the leukemic T cells. Thus, L8313 should be known as a T-cell leukemia associated with granulocytosis.

Early hemopoietic differentiation: the action of multi-CSF is complemented by lineage specific growth factors

Although mechanisms controlling differentiation of hemopoietic stem and early progenitor cells are still poorly understood, it is generally conceded that a pivotal role is played by hemopoietic growth factors (HGFs). However, in-vitro analysis of their action on early progenitors may be obscured by cell-cell interaction, as well as by the presence of fetal bovine serum (FBS). To overcome these limitations, we investigated the action of pure multipotent or lineage-specific HGFs on purified progenitors grown in FBS-free cultures. In the murine system, highly purified progenitors were cultured in the presence of multipotent colony-stimulating factor (multi-CSF, also termed interleukin-3), erythropoietin (Ep) and macrophagic-CSF (M-CSF). Each HGF was unable by itself to induce significant colony growth. However, combined addition of multi-CSF and either Ep or M-CSF gave rise only to pure erythroid or macrophagic colonies, respectively. Partly purified human progenitors were challenged by human granulomonocytic-CSF (GM-CSF), pluripotent CSF (PPO, also termed granulocytic-CSF, G-CSF) and Ep. Here again, each HGF was unable per se to promote colony growth, but combined addition of GM-CSF or PPO and Ep gave rise only to pure erythroid colonies. These results support a model of early hemopoietic differentiation according to which multi-lineage HGFs represent "competence" GFs, the action of which is complemented by lineage-specific "progression" HGFs.

Cloning of human erythroid progenitors (BFU-E) in the absence of fetal bovine serum

We describe culture conditions which enabled us to clone early human erythroid progenitors (BFU-E) in the absence of fetal bovine serum (FBS). Our medium, which is chemically fully defined, supports proliferation and differentiation of erythroid progenitors comparable to that in FBS-supplemented cultures. Furthermore, it allows cloning of BFU-E from all human haemopoietic tissues (adult marrow, embryonic liver and yolk sac) and adult or perinatal blood. This system should facilitate the investigation of human erythroid differentiation in vitro (e.g. cell-cell interaction, mechanism of action of haemopoietins and inhibitors, regulation of Hb synthesis) as well as the mechanisms underlying myeloproliferative disorders. As an example, we have found that recombinant Ep shows, in this culture system, the same dose response as Ep from conventional sources.

Stimulation of haematopoiesis in primates by continuous infusion of recombinant human GM-CSF

Certain proteins are known to play an important part in the proliferation, differentiation and functional activation of haematopoietic progenitor cells in vitro. These proteins include erythropoietin and various colony-stimulating factors (CSFs), one of which is granulocyte-macrophage colony-stimulating factor (GM-CSF). Recently, both murine and human GM-CSF have been purified to homogeneity and complementary DNAs encoding them have been cloned. Although the in vitro activity of recombinant human GM-CSF has been investigated intensively, little is known about the functional activity of this protein in vivo. There is strong evidence that colony-stimulating activities produced by various human and murine tumour tissues and cell lines can stimulate granulopoiesis in mice, as can human urinary extracts. A partially purified preparation of human urinary colony-stimulating factor, however, proved only marginally effective in stimulating granulopoiesis in humans. All these studies suffer from the lack of a homogeneous preparation of colony-stimulating factor. It has recently been shown that recombinant murine multi-CSF or interleukin-3 can stimulate haematopoiesis in mice in vivo. Large-scale production of recombinant human GM-CSF now permits us to examine its effects in vivo using a primate model. We find that the continuous infusion of GM-CSF in healthy monkeys rapidly elicits a dramatic leukocytosis and a substantial reticulocytosis. A similar effect has been observed in one pancytopenic, immunodeficient rhesus macaque. These results suggest that GM-CSF could prove useful in several clinical situations.

Isolation of a human stromal cell strain secreting hemopoietic growth factors

A diploid fibroblastoid cell strain, termed "ST-1," has been established from a long-term liquid culture of human fetal liver cells. ST-1 cells are nonphagocytic, nonspecific esterase negative and do not possess factor VIII-related antigen but stain with antibodies specific for fibronectin and type I collagen. The ST-1 cells produce nondialyzable hemopoietic growth factors capable of stimulating the development of erythroid bursts, mixed granulocyte-macrophage colonies, pure granulocyte colonies, and pure macrophage colonies. These factors are active on both human fetal liver and human adult bone marrow progenitors. When liquid cultures of human fetal liver hemopoietic progenitors are established with a preformed monolayer of ST-1 cells, the yields of nonadherent cells, erythroid progenitors, and myeloid progenitors are greatly increased. These studies demonstrate that the fibroblastoid ST-1 cells support hemopoiesis in vitro and may be a critical element in the stromal microenviroment in vivo.

Human recombinant granulocyte-macrophage colony-stimulating factor: a multilineage hematopoietin

Human recombinant granulocyte-macrophage colony-stimulating factor (GM-CSF) was tested for its ability to induce colony formation in human bone marrow that had been enriched for progenitor cells. In addition to its expected granulocyte-monocyte colony-stimulating activity, the recombinant GM-CSF had burst-promoting activity for erythroid burst-forming units and also stimulated colonies derived from multipotent (mixed) progenitors. In contrast, recombinant erythroid-potentiating activity did not stimulate erythroid progenitors. The experiments prove that human GM-CSF has multilineage colony-stimulating activity.

Purification of fetal hematopoietic progenitors and demonstration of recombinant multipotential colony-stimulating activity

To facilitate the direct study of progenitor cell biology, we have developed a simple and efficient procedure based upon negative selection by panning to purify large numbers of committed erythroid and myeloid progenitors from human fetal liver. The nonadherent, panned cells constitute a highly enriched population of progenitor cells, containing 30.4 +/- 13.1% erythrocyte burst forming units (BFU-E), 5.5 +/- 1.9% granulocyte-macrophage colony forming units (CFU-GM), and 1.4 +/- 0.7% granulocyte-erythroid-macrophage-megakaryocyte colony forming units (CFU-GEMM) as assayed in methylcellulose cultures. These cells are morphologically immature blasts with prominent Golgi. This preparative method recovers 60-100% of the committed progenitors detectable in unfractionated fetal liver and yields 2-30 X 10(6) progenitors from each fetal liver sample, and thus provides sufficient numbers of enriched progenitors to allow direct biochemical and immunologic manipulation. Using this technique, a purified recombinant protein previously thought to have only granulocyte-macrophage colony stimulating activity (GM-CSA) is shown to have both burst promoting activity and multipotential colony stimulating activity. Progenitor purification by panning thus appears to be a simple, efficient method that should facilitate the direct study of committed hematopoietic progenitors and their differentiation.

Production of human active BPA from TCGF independent T cell lines that do not excrete HTLV: proof of direct action of MLA-144 derived BPA using purified BFU-E

Burst promoting activity (BPA) and colony stimulating activity (CSA) were measured in the supernatants from six T-cell lines which are IL-2 independent and do not excrete human T-cell leukaemia virus. Two cell lines were shown to produce high levels of BPA, namely MLA-144 and the Jurkat cell line. MLA-144 is a gibbon T-cell line which can proliferate and produce BPA in serum free HB101 medium. The BPA from this line was shown to act directly on BFU-E by studies using highly purified peripheral blood derived BFU-E. The Jurkat CM had far less activity against purified progenitors. These cells lines will be of value in the study of erythropoietic regulation.