Purification of a membrane-derived human erythroid growth factor

Biology of Exfoliation of Plasma Membrane-Derived Vesicles and the Radiation Response: Historical Background, Applications in Biodosimetry and Cell-Free Therapeutics, and Quantal Mechanisms for Their Release and Function with Implications for Space Travel

This historical review of extracellular vesicles in the setting of exposure to ionizing radiation (IR) traces our understanding of how vesicles were initially examined and reported in the literature in the late 1970s (for secreted exosomes) and early 1980s (for plasma membrane-derived, exfoliated vesicles) to where we are now and where we may be headed in the next decade. An emphasis is placed on biophysical properties of extracellular vesicles, energy consumption and the role of vesiculation as an essential component of membrane turnover. The impact of intercellular signal trafficking by vesicle surface and intra-vesicular lipids, proteins, nucleic acids and metabolites is reviewed in the context of biomarkers for estimating individual radiation dose after exposure to radiation, pathogenesis of disease and development of cell-free therapeutics. Since vesicles express both growth stimulatory and inhibitory molecules, a hypothesis is proposed to consider superposition in a shared space and entanglement of molecules by energy sources that are external to human cells. Implications of this approach for travel in deep space are briefly discussed in the context of clinical disorders that have been observed after space travel.

Erythroid differentiation regulator (EDR), a novel, highly conserved factor

In serum-free WEHI-3 supernatants an activity was detected inducing haemoglobin synthesis in human and murine erythroleukaemia cell lines. The absolute numbers of benzidine-positive cells induced with either DMSO or WEHI-3-conditioned medium were comparable. Terminal differentiation was not observed. An expression library from WEHI-3 RNA aided by PCR cloning revealed an open reading frame corresponding to a 209 amino acid protein. This was 100% identical to a sequence from human stimulated peripheral blood mononuclear cells. In contrast to human RNA, mouse RNA exhibited multiple bands of pre-mRNA in Northern blots. The gene was provisionally termed erythroid differentiation regulator (edr). In mammalian cells EDR is mostly expressed as a 56 kDa dimer showing higher activity than the recombinant monomer. The activity profile is bell-shaped. Expression was observed in many normal mouse tissues, yet in haematopoiesis it was largely confined to CD34+ cells. It was enhanced by a series of stimuli such as phorbol ester, and transformed cells generally showed a higher level of EDR expression than normal ones. The protein is localized at the inner side of the cytoplasmic membrane and is released in part via vesicles. In view of the broad range of EDR-expressing tissues the function obviously exceeds haemoglobin synthesis induction. Involvement in cell survival and growth control has been observed and will be dealt with in detail elsewhere.

A functional erythropoietin receptor is necessary for the action of thrombopoietin on erythroid cells lacking c-mpl

OBJECTIVE

We hypothesized that thrombopoietin (TPO) exerts its mitogenic effects on erythroid cells, at least in part, via an interaction of TPO with the cells' erythropoietin receptor (EPO-R).

METHODS

We used BaF3 cells stably transfected with EPO-R to demonstrate that TPO alone is sufficient to support the long-term growth and proliferation of BaF3/EPO-R cells and to develop a TPO-dependent variant, BaF3/EPO-R(T), which is highly sensitive to and dependent on TPO for its proliferation. Northern analysis and RT-PCR were used to verify that both BaF3/EPO-R and BaF3/EPO-R(T) cells express EPO-R but lack c-mpl, the TPO receptor. To confirm that TPO responsiveness of BaF3/EPO-R(T) is due to TPO's interaction with EPO-R, EPO-R was downregulated by antisense mRNA.

RESULTS

Downregulation of EPO-R in BaF3/EPO-R(T) cells abolishes responsiveness to both EPO and TPO. Viability of EPO-treated transfectants decreased from 95% to 36%, while that of TPO-treated transfectants decreased from 95% to 9% by 48 hours. Nontransfected BaF3/EPO-R(T), and BaF3/EPO-R(T) transfected with vector alone, remained viable and grew in either EPO or TPO.

CONCLUSION

Our results suggest a functional EPO-R may be necessary and sufficient for TPO to exert its mitogenic effects on erythroid cells.

Modulation of intercellular communication mediated at the cell surface and on extracellular, plasma membrane-derived vesicles by ionizing radiation

The plasma membrane is a dynamic organelle whose function includes receptor-mediated signal transduction into the cell. Conversely, the plasma membrane is the origin of inter-cellular signaling. In addition to expressing and releasing growth factors in a soluble form(through exocytosis) and via proteolysis of cell surface components, membrane ligands may signal nearby cells through juxtacrine stimulation or by the exfoliation or shedding of plasma membrane-derived vesicles. Ionizing radiation (IR) has a profound effect on plasma membrane structure and function. IR-induced ultrastructural alterations are mediated via lipid interaction with water radiolysis products (e.g., hydroxyl radicals, hydrogen radicals, and hydrated electrons). Ionizing radicals act directly on lipid molecules to promote lipid hydro-peroxides and lipid hydroperoxide breakdown products (e.g., alpha, beta unsaturated aldehydes) that contribute to altered plasma membrane lipid composition. A change in lipid composition increases membrane lipid microviscosity and results in membrane fenestrations that enhance permeability to small molecules and ions. Reactive ionizing species also stimulate sphingomyelinase activity, leading to sphingomyelin hydrolysis and ceramide generation that further contributes to altered membrane lipid composition and cellular apoptosis. In addition, exposure to IR results in impaired rate of and cumulative shedding of plasma membrane-associated growth factors. Mechanisms of exfoliation are reviewed for normal cells and the impact of radiation on modulating signal transduction mediated by exfoliation is summarized.

Troponin I inhibits capillary endothelial cell proliferation by interaction with the cell's bFGF receptor

Troponin I (TnI) is a novel cartilage-derived angiogenesis inhibitor, first demonstrated by Moses et al. (1999, Proc. Natl. Acad. Sci. USA 2645-2650) to inhibit endothelial cell proliferation and angiogenesis, both in vivo and in vitro, and to inhibit metastasis of a wide variety of tumors in vivo. Despite convincing evidence of its efficacy, little is known about the mechanism of action of TnI as an anti-proliferative and anti-angiogenic agent. In the current article we demonstrate that TnI inhibits both bFGF-stimulated and basal levels of endothelial cell proliferation, and we hypothesize that this inhibition is occurring, at least in part, via an interaction of TnI with the cell-surface bFGF receptor on capillary endothelial cells. We further support this hypothesis by providing the first evidence that TnI can act on nonendothelial as well as endothelial cells and by demonstrating that this inhibitory action is specific for the bFGF receptor on the target cells. Preliminary data suggest that TnI may be competing with bFGF for interaction with the bFGF receptor on responsive cells.

Biologically Active Fas Antigen and Its Cognate Ligand Are Expressed on Plasma Membrane-Derived Extracellular Vesicles

Exfoliation of plasma membrane components is a directed process that consumes energy and requires active cell metabolism. Proteins involved in regulating the survival and proliferation of eukaryotic cells are released on exfoliated vesicles. We examine here whether the Fas receptor and its cognate ligand (FasL) are present on vesicles shed from high metastatic potential CX-1 cells and low metastatic potential MIP-101 cells and from HuT 78 cells, respectively. Rates of exfoliation at 2 hours and cumulative levels of extracellular vesicles in serum-free medium conditioned by CX-1 cells are increased by 1.8-fold and 1.6-fold, respectively, relative to that in medium conditioned by MIP-101 cells. Although vesicles shed from both cancer cell lines contain Fas antigen, the amount of Fas per vesicle and the percentage of vesicles containing Fas are increased for vesicles isolated from MIP-101 cells, relative to those from CX-1 cells, as determined by immunogold particle labeling and electron microscopy and by immunofluorescence microscopy and flow cytometry. Results of metabolic labeling with 35S-methionine indicate that Fas biosynthesis is reduced by up to 3.3-fold for CX-1 cells, relative to that of MIP-101 cells, consistent with the finding of decreased Fas on vesicles shed from the plasma membrane of CX-1 cells. Although mRNA for soluble Fas receptor is detectable in both cell lines, depletion of shed vesicles from serum-free medium by ultracentrifugation removes all detectable biological activity. FasL is detected on vesicles exfoliated from HuT 78 cells by immunoelectron microscopy and Western blot analysis. FasL-bearing vesicles induce apoptosis of Fas-expressing cancer cells at the same level as observed by treatment with monoclonal anti-Fas antibody. Furthermore, Fas-bearing extracellular vesicles from MIP-101 but not from CX-1 cells protect the CX-1 cell line from FasL-induced and anti-Fas–mediated apoptosis, indicating that Fas present on shed vesicles is biologically active. We conclude that the Fas antigen and its cognate ligand are exfoliated from the cell surface in a bioactive configuration. Exfoliation may provide a mechanism for long-range signal-directed apoptosis while maintaining Fas/FasL on a membrane surface.

Biologically Active Fas Antigen and Its Cognate Ligand Are Expressed on Plasma Membrane-Derived Extracellular Vesicles

Exfoliation of plasma membrane components is a directed process that consumes energy and requires active cell metabolism. Proteins involved in regulating the survival and proliferation of eukaryotic cells are released on exfoliated vesicles. We examine here whether the Fas receptor and its cognate ligand (FasL) are present on vesicles shed from high metastatic potential CX-1 cells and low metastatic potential MIP-101 cells and from HuT 78 cells, respectively. Rates of exfoliation at 2 hours and cumulative levels of extracellular vesicles in serum-free medium conditioned by CX-1 cells are increased by 1.8-fold and 1.6-fold, respectively, relative to that in medium conditioned by MIP-101 cells. Although vesicles shed from both cancer cell lines contain Fas antigen, the amount of Fas per vesicle and the percentage of vesicles containing Fas are increased for vesicles isolated from MIP-101 cells, relative to those from CX-1 cells, as determined by immunogold particle labeling and electron microscopy and by immunofluorescence microscopy and flow cytometry. Results of metabolic labeling with 35S-methionine indicate that Fas biosynthesis is reduced by up to 3.3-fold for CX-1 cells, relative to that of MIP-101 cells, consistent with the finding of decreased Fas on vesicles shed from the plasma membrane of CX-1 cells. Although mRNA for soluble Fas receptor is detectable in both cell lines, depletion of shed vesicles from serum-free medium by ultracentrifugation removes all detectable biological activity. FasL is detected on vesicles exfoliated from HuT 78 cells by immunoelectron microscopy and Western blot analysis. FasL-bearing vesicles induce apoptosis of Fas-expressing cancer cells at the same level as observed by treatment with monoclonal anti-Fas antibody. Furthermore, Fas-bearing extracellular vesicles from MIP-101 but not from CX-1 cells protect the CX-1 cell line from FasL-induced and anti-Fas–mediated apoptosis, indicating that Fas present on shed vesicles is biologically active. We conclude that the Fas antigen and its cognate ligand are exfoliated from the cell surface in a bioactive configuration. Exfoliation may provide a mechanism for long-range signal-directed apoptosis while maintaining Fas/FasL on a membrane surface.

Early plasma membrane events occurring in ultraviolet-B-induced apoptosis

Whereas nonsolar ultraviolet C radiation primarily affects nuclei (i.e., where it is absorbed by nucleic acids) of eukaryotic cells, ultraviolet radiation of long (320-380 nm) wavelengths (ultraviolet A) and intermediate (290-320 nm) wavelengths (ultraviolet B) primarily affects lipid membranes. We have previously demonstrated that ultraviolet B irradiation alters the surface architecture of human B cells and impairs expression of an erythroid growth factor on their surface and on extracellular vesicles. Here, we examined the effects of ultraviolet B irradiation on the capacity of Chinese hamster ovary cells to undergo the process of exfoliation, and on the capacity of Chinese hamster ovary cells transfected with flt3/flk2 cDNA to express the cytokine flt3/flk2. Our results indicate that the rate of release of shed vesicles from untransfected Chinese hamster ovary cells is decreased after one to two h, at a time when there is electron microscopic evidence for retention of vesicles at the cell surface. These changes at the cell surface precede all other apparent morphological changes (including DNA condensation in the nucleus, swelling of the mitochondria and appearance of apoptotic bodies). Furthermore, plasma membranes and shed extracellular vesicles from ultraviolet B irradiated Chinese hamster ovary cells that have been transfected with flt3/flk2 cDNA fail to express the protein.

Expression of a negative regulator of human erythropoiesis by fluidized lymphocyte plasma membranes

Regulation of hematopoiesis by paracrine molecules occurs in vitro. In some cases, hematopoietic paracrine factors have been localized to the plasma membrane of accessory cells. We have purified a unique integral membrane glycoprotein from normal human B cells that functions in vitro as a paracrine factor whose activity is directed toward erythroid progenitor cells. This factor is also spontaneously exfoliated from the cell surface as a component of extracellular vesicles. Analysis of the lipid and protein compositions and membrane lipid order of these extracellular vesicles reveals them to be biochemically distinct and more fluid than their parent membranes. Evidence in nonhematopoietic culture systems indicates that cell membrane function may be altered by modifying membrane fluidity. In an effort to accelerate growth factor release, plasma membranes of B cells were fluidized by incubation with an emulsion of Liposin II, phosphatidylcholine, and phosphatidylethanolamine. Fluidity assessed by steady-state fluorescence polarization was reduced in lipid-treated cells. Exfoliation was 3-4-fold higher from lipid-treated cells relative to untreated cells. Unexpectedly, a negative signal for burst-forming unit-erythroid (BFU-E) proliferation was expressed in membranes, in shed extracellular vesicles, and in supernatants of medium conditioned by the fluidized cells. Purification of the inhibitor is under way. The data are consistent with the hypothesis that accessory cell plasma membranes may positively or negatively regulate erythroid differentiation, depending upon the exchange of cholesterol and phospholipids between plasma membrane and ambient lipid pools.

Platelet-activating factor increases glutamine synthetase activity in early and late passage C-6 glioma cells

Previous studies from this laboratory have shown that C-6 rat glioma cells (2B clone) exhibit specific phenotypic characteristics depending on passage in culture and that these populations respond differentially to addition of various exogenous compounds to the medium. Early passage (less than 25) C-6 glial cells express low glutamine synthetase activity (a marker for astrocytes) and with increasing cell passage (greater than 70) C-6 glial cells express more astrocytic properties with respect to both glutamine synthetase (GS) and morphology. In this study, cells from both early (glioblastic) and late (astrocytic) passage were examined for their response to the phospholipid, platelet-activating factor (PAF). We found that PAF increased GS activity in early passage (glioblastic) cells and more importantly it increased GS activity in late passage cells, already committed to the astrocytic phenotype. Furthermore, cells from both passages failed to respond to addition of lyso-PAF, the non-biologically active analog of PAF, to the medium. By following the uptake of 3H-PAF into cells, we observed that greater than 90% of the phospholipid was taken into the cells within the first hour of incubation. We compared the PAF effects with that of dibutyryl cyclic AMP (dBcAMP) and RO20-1724, a phosphodiesterase inhibitor. Cells from the early passage responded to both dBcAMP and RO20-1724 treatments with a significant increase in GS activity whereas cells from the late passage showed no significant change, confirming earlier reports from this laboratory. These findings indicate that the response of C-6 glioma cells to PAF (at least in the late passage) is not mediated via cyclic AMP. We suggest that in early passage cells PAF promotes expression of the astrocytic phenotype and in late passage cells PAF mediates a gliosis-type response.

Intracellular regulation of the production and release of human erythroid-directed lymphokines

Erythroid burst-promoting activity (BPA) is released from B lymphocytes in soluble (sBPA) and membrane-bound (mBPA) forms. To study intracellular processes involved in production of these physically separable factors, we measured their time course release into serum-free medium from B cells that were pulse-exposed for 5-240 min to nonmitogenic base medium or inhibitors of energy-dependent metabolism (2,4-dinitrophenol, sodium azide, and 2-deoxy-D-glucose), transcription and translation (actinomycin D and cycloheximide), replicative DNA synthesis (cytosine arabinoside), or posttranslational processing (monensin). mBPA and sBPA were initially detectable after 1 and 2 h, respectively. Maximum cumulative levels of 8 +/- 0.6 and 9 +/- 1.0 U/ml, respectively, were reached after 8 h. In contrast, cumulative mBPA and sBPA levels in medium prepared from cells treated with metabolic inhibitors were reduced by up to 90%. Both surface exfoliation and mBPA expression by intact plasma membranes were diminished. Whereas pulse-exposure to cytosine arabinoside had no effect, treatment with actinomycin D or cycloheximide abolished BPA expression. Exposure to monensin reduced mBPA and sBPA levels to zero in a concentration-and time-dependent fashion. We conclude that production and release of BPA is an energy-dependent process, requiring mRNA synthesis and translation and posttranslational remodeling of the protein but not replicative DNA synthesis.

Growth factor-dependent initiation of DNA replication in nuclei isolated from an interleukin 3-dependent murine myeloid cell line

To study the proliferative response of hematopoietic cells to growth factors at the molecular level, we developed a cell-free system for growth factor-dependent initiation of genomic DNA replication. Nuclei were isolated from the IL-3-dependent cell line NFS/N1-H7 after a 10-h period of IL-3 deprivation. Cytosolic and membrane-containing subcellular fractions were prepared from proliferating NFS/N1-H7 cells. Nuclei from the nonproliferating cells (+/- IL-3) showed essentially no incorporation of [3H]thymidine during a 16-h incubation with a mixture of unlabeled GTP, ATP, UTP, CTP, dGTP, dATP, dCTP, and [3H]dTTP. When the combination of IL-3, a cytosolic fraction, and a membrane-containing fraction from proliferating cells was added to nuclei from nonproliferating cells, a burst of [3H]thymidine incorporation into DNA began after a 12-h lag period, attained a maximal rate at 16 h, and reached a level of 860 pmol thymidine/10(6) nuclei at 24 h (corresponding to replication of approximately 56% total mouse genomic DNA). This DNA synthesis was inhibited approximately 90% by the specific DNA polymerase alpha inhibitor aphidicolin. Deletion of a single cellular component or IL-3 from the system resulted in a marked reduction of DNA replication (-membrane, 80 +/- 4%; -cytosol, 90% +/- 4%; -IL-3, 74 +/- 7% inhibition). This model requires a growth factor (IL-3), a sedimentable cell fraction containing its receptor and possibly additional membrane-associated components, and a cytosolic fraction. It appears to recapitulate the molecular events required for progression from early G1 to S phase of the cell cycle induced by IL-3 binding to its receptor.

Characterization of biologically active, platelet-derived growth factor-like molecules produced by murine erythroid cells in vitro and in vivo

Platelet-derived growth factor (PDGF) is an important serum regulator of erythropoiesis in vitro. We have now obtained evidence suggesting that PDGF-like molecules may also modulate erythropoiesis in vivo. Western blot analysis of cytoplasmic extracts from Rauscher murine erythroleukemia cells and phenylhydrazine-treated mouse splenic erythroid cells revealed the presence of several PDGF-like proteins. The presence of PDGF-like proteins in the cytoplasm of these two erythroid cell types was confirmed by immunohistochemical staining. Using a serum-free biologic assay, PDGF-like biological activity was found in cell lysates and conditioned medium of both Rauscher cells and phenylhydrazine-treated mouse erythroid cells. Subcellular localization experiments revealed the biological activity to be concentrated in the cytosolic fraction. Using a series of antibodies to hematopoietic growth factors we demonstrated that PDGF-like biological activity was specifically immunoprecipitated by both monoclonal and polyclonal anti-human PDGF antibodies but not by antibodies to burst-promoting activity, granulocyte-macrophage colony-stimulating factor, IL-3, or erythropoietin. Taken together, the data are consistent with the hypothesis that PDGF-like molecules play a role in the regulation of mammalian erythropoiesis in vivo.

Selective and indirect modulation of human multipotential and erythroid hematopoietic progenitor cell proliferation by recombinant human activin and inhibin

Activin and inhibin are biomolecules that, respectively, enhance and suppress the release of follicle-stimulating hormone from pituitary cells in vitro. Purified recombinant human (rhu) activin A and inhibin A were assessed for their effects on colony formation in vitro by human multipotential (CFU-GEMM), erythroid (BFU-E), and granulocyte-macrophage (CFU-GM) progenitor cells. It was found that (i) rhu-activin A enhances colony formation by normal bone marrow erythroid and multipotential progenitor cells; (ii) purified rhu-inhibin A decreases activin, but not rhu-interleukin 3, rhu-granulocyte-macrophage colony-stimulating factor, or rhu-interleukin 4, enhancement of erythropoietin-stimulated colony formation by erythroid and multipotential progenitor cells; (iii) modulatory actions of rhu-activin and rhu-inhibin are mediated through monocytes and T lymphocytes within the marrow; (iv) actions are apparent in the absence or presence of serum; and (v) rhu-activin and rhu-inhibin have no effect on colony formation by granulocyte-macrophage progenitor cells. This defines an indirect mode of action and a specificity for activin and inhibin on multipotential and erythroid progenitor cells.

Effect of recombinant and purified human haematopoietic growth factors on in vitro colony formation by enriched populations of human megakaryocyte progenitor cells

Nonadherent low density T-lymphocyte depleted (NALT-) marrow cells from normal donors were sorted on a Coulter Epics 753 Dye Laser System using Texas Red labelled My10 and phycoerythrin conjugated anti HLA-DR monoclonal antibodies in order to obtain enriched populations of colony forming unit-megakaryocyte (CFU-MK). The CFU-MK cloning efficiency (CE) was 1.1 +/- 0.5% for cells expressing both high densities of My10 and low densities of HLA-DR (My10 DR+). This procedure resulted in an 18-fold increase in CE over NALT- cells. The effect of purified or recombinant human haematopoietic growth factors including erythropoietin (Epo), thrombocytopoiesis stimulating factor (TSF), interleukin 1 alpha (IL-1 alpha), granulocyte colony stimulating factor (G-CSF), granulocyte-macrophage colony stimulating factor (GM-CSF), macrophage colony stimulating factor (M-CSF or CSF-1) and interleukin MK colony formation by My10 DR+ cells was determined utilizing a serum depleted assay system. Neither Epo, TSF, CSF-1, IL-1 alpha nor G-CSF alone augmented MK colony formation above baseline (2.5 +/- 0.8/5 x 10(3) My10 DR+ cells plated). In contrast, the addition of GM-CSF and IL-3 each increased both CFU-MK colony formation and the size of colonies with maximal stimulation occurring following the addition of 200 units/ml of IL-3 and 25 units/ml of GM-CSF. At maximal concentration, IL-3 had a greater ability to promote megakaryocyte colony formation than GM-CSF. The stimulatory effects of GM-CSF and IL-3 were also additive in that the effects of a combination of the two factors approximated the sum of colony formation in the presence of each factor alone. The CFU-MK appears, therefore, to express HPCA-1 and HLA-DR antigens. These studies also indicate that GM-CSF and IL-3 are important in vitro regulators of megakaryocytopoiesis, and that these growth factors are not dependent on the presence of large numbers of macrophages or T cells for their activity since the My10 DR+ cells are largely devoid of these accessory cells.

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.