Human platelet-derived growth factor: radioimmunoassay and discovery of a specific plasma-binding protein

The platelet-derived growth factor (PDGF) is the principal mitogen in serum for cultured cells of mesenchymal origin. PDGF also is a potent chemotactic protein for inflammatory cells and for cells required for wound repair. Because activity levels of PDGF in biological fluids are difficult to measure, we attempted to develop a radioimmunoassay for PDGF. Rabbits were immunized with purified PDGF; the antiserum obtained was monospecific for PDGF in immunodiffusion analysis against concentrated platelet lysates, serum, and plasma. A radioimmunoassay for PDGF was developed with a sensitivity of congruent to 0.2 ng/ml. Levels of PDGF in plasma/serum were measured and compared with PDGF levels determined by a receptor-competition assay and by a standard biological assay measuring incorporation of [3H]thymidine into 3T3 cells. Radioimmunoassay showed apparent PDGF levels of 50 ng/ml in human plasma and 103 ng/ml in serum. The 50 ng/ml PDGF in plasma was unexpected because the plasma samples contained little or no platelet release products as determined by very low levels of platelet factor 4. We therefore sought an immunologically reactive PDGF molecule in human plasma. No immunologically reactive protein was detected by immunodiffusion analysis or when plasma was treated with an immunoaffinity gel. Subsequently, a 125I-PDGF-binding protein was identified; the 125I-PDGF-plasma-binding protein complex was not reactive with anti-PDGF immunoglobulin. Correction for 125I-PDGF bound by the plasma-binding protein established serum levels of PDGF of congruent to 50 ng/ml; congruent to 50 ng/ml PDGF was found in serum by radioreceptor-competition assays and by mitogenic assays as well. The plasma-binding protein may serve to clear PDGF released in the circulation, thereby limiting PDGF activity to its local interactions at the site of blood-vessel injury.     

Stimulation of DNA synthesis in primary cultures of adult rat hepatocytes by rat platelet-associated substance(s)

Summary Experiments in whole animals have shown that normally quiescent adult rat hepatocytes are induced to proliferate by blood borne substances, which we are now probing in primary monolayer cultures. Under our conditions, freshly isolated adult hepatocytes do not proliferate actively in a defined medium, but are stimulated to synthesize DNA — an essential first step — by either serum or an EGF-hormone combination. Stimulation of [³H]thymidine incorporation into hepatocyte DNA by addition of dialyzed mouse, human, horse, or bovine (fetal, newborn, or calf) serum, whose activities are all similar, is regularly surpassed by an EGF-insulin mixture without serum. This, in turn, is exceeded by dialyzed normal rat serum, which is several times more potent than the other sera tested. Removal of blood platelets reduces the activity of normal rat serum by over 50%. Heat inactivation (56° C) causes a similar loss, but heat treatment of platelet-poor serum fails to cause further reduction. The activity of mouse and human serum is not reduced by platelet removal. Serum from partially hepatectomized rats is not significantly more stimulatory than normal rat serum, and its activity is depressed in the same way by platelet deprivation and heat inactivation. Lack of enhancement by partial hepatectomy is not consonant with whole animal studies and requires further investigation. The heat-labile portion of the DNA synthesis-stimulating activity of rat serum appears to derive from platelets. This activity differs from the well-characterized heat-stable human PDGF. Its relation to other reported platelet-associated growth factors is still undetermined. This work was supported by USPHS Grants CA-02146 and AM-19435.     

Regulation of 2'',5''-oligoadenylate synthetase gene expression by interferons and platelet-derived growth factor.

In murine BALB/c 3T3 cell cultures, either beta interferon or platelet-derived growth factor (PDGF) enhanced expression of the 2',5'-oligoadenylate synthetase mRNA and protein. The time course of induction in response to beta interferon was similar to that in response to PDGF. Of several growth factors known to be present in clotted blood serum (i.e., epidermal growth factor, transforming growth factor beta, and PDGF), only PDGF enhanced expression of 2',5'-oligoadenylate synthetase. The linkage of an interferon response element-containing segment from the 5'-flanking region of a human or murine 2',-5'-oligoadenylate synthetase gene made a heterologous gene responsive to interferon. The expression of such a gene construct in transfected cells was also induced by PDGF. Induction by PDGF was inhibited by mono- or polyclonal antibodies to murine interferon, which suggested that induction by PDGF requires interferon. Both PDGF and interferon induced nuclear factors that bound to this interferon response element-containing segment in vitro.     

Platelet concentrate as an additive to bone allografts: a laboratory study using an uniaxial compression test

Chemical cleaning procedures of allografts are destroying viable bone cells and denaturing osteoconductive and osteoinductive proteins present in the graft. The aim of the study was to investigate the mechanical differences of chemical cleaned allografts by adding blood, clotted blood; platelet concentrate and platelet gel using a uniaxial compression test. The allografts were chemically cleaned, dried and standardized according to their grain size distribution. Uniaxial compression test was carried out for the four groups before and after compacting the allografts. No statistically significant difference was found between native allografts, allografts mixed with blood, clotted blood, platelet concentrate and platelet concentrate gel regarding their yield limit after compaction. The authors recommend to chemical clean allografts for large defects, optimize their grain size distribution and add platelet concentrate or platelet rich plasma for enhancing as well primary stability as well bone ingrowth.     

Immunoelectron-microscopic studies of human platelet thrombospondin, Von Willebrand factor, and fibrinogen redistribution during clot formation

Summary The relative distributions of the human platelet -granule proteins fibrinogen, thrombospondin, and von Willebrand factor were mapped by immunoelectron microscopy in thin cryosections of activated platelets, platelet aggregates, and clots during the first 24 h ofin vitro clot formation. In early activated platelets, the results suggest that the canalicular system constitutes a significant component of the external platelet surface, and may act as a compartment for biochemical reactions occurring during granule relase. Further, detection of coagulation proteins by various non-morphological procedures may reflect protein contained within canalicular elements. Later in the release process, von Willebrand factor was detected as a major antigen on the platelet canalicular and plasma membranes; thrombospondin, on the other hand, showed minimal binding to platelets and only limited binding to the extensive fibrin network. Comparison of radioimmunoassays of supernatants of thrombin-stimulated platelets in plasma, clotted whole blood, and Triton X-100 platelet releasates indicated that virtually all of the platelet thrombospondin appears in serum. These data confirm the immunocytochemical results indicating that very little platelet thrombospondin binds to the platelet surface, compared with von Willebrand factor, studied here under the same conditions, which binds extensively to the platelet membrane following release and clot formation.     

Human platelet lysate contains growth factor activities for established cell lines derived from various tissues of several species

Summary Factors have been studied from human platelets that promote the growth of a hormone-responsive rat mammary adenocarcinoma cell line MTW9/PL, the BALB/c 3T3 mouse embryo fibroblasts, and numerous other established cell lines. A wide variety of the commonly employed cell lines, including lines of human, mouse, monkey, chicken, rat, Chinese hamster, and Syrian hamster origin, were tested for their growth response to a standard concentration of 200 μg/ml human platelet lysate, and the lysate was found to contain mitogenic activity for 24 of the 29 different lines assayed. A comparison was made between the platelet growth activity for the MTW9/PL cells and the well characterized platelet mitogen for the BALB/c 3T3 cells, platelet-derived growth factor (PDGF). When the platelet lysate was subjected to digestion by highly purified trypsin, the mitogenic activity for the MTW9/PL cells was not affected whereas that for the BALB/c 3T3 cells was essentially destroyed. Crude PDGF was prepared by heating the human platelet lysates at 100°C for 2 min followed by clarification, dialysis, lyophilization, and reconstitution. This PDGF material had no apparent growth activity for MTW9/PL cells, although chromatography of this material on Biogel P-100 revealed a high molecular weight (approximately 40,000 daltons) activity for the BALB/c 3T3 cells (presumably PDGF) and two growth activities for the MTW9/PL cells, one high molecular weight activity and a second activity of molecular weight less than 10,000. These studies demonstrated a form of epithelial tumor cell growth activity separable from the 3T3 type PDGF in crude heated extracts. This work was supported by American Cancer Society Grant BC-255B and NIH Grant CA 26617.     

The molecular biology of platelet-derived growth factor

PDGF is a connective tissue mitogen that has been associated with clotted blood serum for at least 300 million years. It regulates the expression of cell cycle "early genes" in normal fibroblasts. Induction of early genes is preceded by stimulation of a tyrosine-specific kinase. The putative structural gene for PDGF has been acquired by an acutely transforming retrovirus and is expressed in many connective tissue tumors. Further work is needed to determine whether (i) production of PDGF by tumor cells confers a proliferative advantage on these cells, (ii) tyrosine-specific phosphorylations mediate the induction of cell cycle early genes by PDGF, and (iii) products of cell cycle early genes play any functional role in the 10-12 hr chain of events that culminates in replicative DNA synthesis and cell division. In the meantime, these very issues represent candidate functions for other viral oncogenes and their cellular homologs. Some of these genes could act at the onset of the mitogenic cascade by causing the production of automitogenic growth factors. Others may function in the interior of the cascade by promoting tyrosine-specific phosphorylations. Still others may be mutated or rearranged homologs of cell cycle early genes whose expression is normally modulated by extracellular growth factors.     

Antibodies in the sera of patients with bronchogenic carcinoma that react with antigen from a tumour cell line

Summary Antigenic material was isolated from a human squamous cell bronchogenic carcinoma tissue culture line A549 that was found to react with antibodies in the serum of patients with lung cancer in an enzyme-linked immunosorbent assay (ELISA), while it did not react with sera from normal individuals. The antigen was tested with a panel of sera from a variety of patient groups by means of the ELISA. Results showed significantly higher numbers of sera from patients with lung cancer, particularly those of squamous cell origin, reacting with the antigen than of sera from 173 normal individuals or patients with breast and gynaecological cancers or melanomas.     

Balamuthia mandrillaris, agent of amebic encephalitis: detection of serum antibodies and antigenic similarity of isolates by enzyme immunoassay

ABSTRACT. We report the development of an enzyme-linked immunosorbent assay (ELISA) for detecting antibodies to Balamuthia mandrillaris , a free-living ameba that is an etiologic agent of granulomatous amebic encephalitis (GAE). As part of the California Encephalitis Project (CEP), we have tested serum and cerebrospinal fluid (CSF) samples from a subgroup of 130 hospitalized encephalitis patients (out of ∼430 samples) over a 16-month period. Case criteria were based on clinical, laboratory, and occupational/recreational histories. All serum samples initially underwent screening by immunofluorescent antibody (IFA) staining with results ranging from no detectable ameba antibodies to titers of 1:256. In addition to the 130 samples tested prospectively, sera and/or CSF from 11 previously confirmed cases of balamuthiasis, six healthy individuals, and earlier CEP submissions with high IFA antibody titers were also tested retrospectively. Among the 130 samples, two cases of balamuthiasis were identified by ELISA and confirmed by the polymerase chain reaction (PCR). The availability of sera from human and animal cases and from varied geographic areas allowed comparisons of serologic similarities of the different Balamuthia strains and human sera. All sera, whether from human or other mammals, reacted with all strains of Balamuthia , as they did with Balamuthia amebae from different geographic areas. Enzyme-linked immunosorbent assay results were consistent with the IFA results. Differences between readings were likely due to cross-reactivity between Balamuthia antigens and unidentified antibodies in serum.     

The effects of aspirin-containing serum in the continuous culture of Plasmodium falciparum

In vitro culture of Plasmodium falciparum-infected human erythrocytes (RBC) has permitted systematic study of human host-parasite relations. In this study the effect of aspirin in the culture system was examined by using serum from blood of fasting, healthy male volunteers, before and after the ingestion of aspirin. The addition of aspirin-containing serum disturbed parasite growth and development: 0-1/2 dilutions of treated/control sera inhibited parasite development, with nuclear pyknosis, pyknotic extracellular parasites (trophozoites) in the media, decreased numbers and sizes of "rings" (early trophozoites), and an increased number of later trophozoites and schizonts. Paradoxically, while the incorporation of [3H]isoleucine into protein was not affected by the aspirin-containing sera, the incorporation [3H]hypoxanthine was significantly changed and did not correlate with morphological evidence of cytotoxicity. Thus, the so-called "incorporation" of a radioactive tracer is not a fully reliable index of parasite growth in the presence of certain compounds. The findings underscore the importance, in this culture system which employs human serum, of avoiding serum from donors who have recently ingested aspirin.     

Immunochemical localization of the epitope for a monoclonal antibody that neutralizes human platelet-derived growth factor mitogenic activity.

A monoclonal antibody (mAb), sis 1, generated against human c-sis-encoded platelet-derived growth factor (PDGF) BB, was shown by enzyme-linked immunosorbent assay and Western blot (immunoblot) analysis to recognize human PDGF BB and human platelet PDGF AB but not the human PDGF AA. This monoclonal antibody potently inhibited PDGF receptor-binding and mitogenic activities of both human PDGF BB and PDGF AB but had no effect on PDGF AA. Finally, we demonstrated that an immunoaffinity-purified anti-c-sis peptide antibody (anti-V4) which also blocked binding of PDGF BB to its cognate receptor and competed with mAb sis 1 for binding to PDGF BB. All of these results suggest that mAb sis 1 recognizes an epitope of the c-sis gene product, PDGF BB, that spatially overlaps the V4 surface domain of PDGF BB, immunochemically localizing a region of PDGF BB critical for PDGF receptor binding and activation.     

Autologous and homologous immunofluorescent antibody to established breast cancer cell lines

Summary Indirect immunofluorescence tests were performed on 14 established human breast cancer cell lines using sera from a variety of subjects. Autologous reactions were studied on 10 cell lines, with positive reactions demonstrable in 8. Tests using sera from a randomly selected population of breast cancer patients showed reactivity in 40 to 66% depending on the target cell line used. Reactivity to other nonbreast cancer cell lines was rare. Several control populations were tested, including normal blood bank donors, persons with benign breast disease, and persons with other forms of cancer; immunofluorescent antibody was detected much less frequently in sera from these populations than those from the breast cancer group. Positive reactions remained in spite of absorption of serum with heterophile antigens, normal human breast tissue, and AB+ red blood cells. Thus established cell lines of human breast cancer possess antigens commonly recognized by sera from breast cancer patients. This work was supported in part by Grant CA 16789 from the National Cancer Institute, NIH, Department of Health, Education and Welfare.     

Biological properties of a hepatocyte growth factor from rat platelets

In an accompanying communication we demonstrated that about half of the potency of rat serum to stimulate DNA synthesis in cultured adult rat hepatocytes resides in a polypeptidelike substance from the platelets. A lysate of rat platelets was able to restore the potency of platelet-poor rat serum, whereas a lysate of human platelets inhibited thymidine incorporation by the hepatocytes. Moreover, addition to these cultures of either highly purified human platelet-derived growth factor (PDGF) or human platelet factor 4 (PF-4) failed to influence DNA synthesis either alone or in the presence of rat or human platelet-poor serum, which is required for expression of PDGF activity. Unlike the human platelet factors, rat platelet lysate (RPL) was moderately active by itself and was augmented equally well by platelet-poor serum from either source. At concentrations below 5%, platelet-poor serum from hypophysectomized rats was as potent as that from normal rats in augmenting RPL activity. This suggests that, unlike PDGF, which is not activated by hypophysectomized rat serum, the hepatotrophic component of RPL does not require the presence of exogenous somatomedins for activity, but interacts instead with other plasma constituents or with somatomedins produced by the hepatocytes in vitro. Rat platelets do, however, appear to contain PDGF or its rat equivalent in addition to the hepatocyte growth factor, since if they are heated to 100 degrees C for 10 min, their ability to stimulate nuclear labeling in confluent BALB/c 3T3 cells is not impaired, while their ability to stimulate DNA synthesis in rat hepatocytes is destroyed. These studies indicate that the hepatocyte growth factor from rat platelets differs from PDGF in its biological as well as physical characteristics, but that rat platelets also contain PDGF or an equivalent substance.     

Control of cytolysis of BALB/c-3T3 cells by platelet-derived growth factor: a model system for analyzing cell death

In culture medium supplemented with 10% clotted blood serum, the saturation density of BALB/c-3T3 cells is determined jointly by cell replication and cell loss. By prelabelling cellular DNA with ³H-thymidine and also by time lapse photography, we studied cell loss independently of replication. Cell loss was accelerated when BALB/c-3T3 cells were transferred from serum-supplemented medium, which contains the platelet derived growth factor (PDGF), to medium supplemented with platelet-poor plasma which lacks it. Loss occurred via the disintegration of cell attached to the surface of the tissue culture dish. Cytolysis of individual cells occurred rapidly; less than 15 minutes transpired between the first indication of a perturbance (by phase contrast microscopy) and fragmentation of the cell cytoplasm. Kinetic analysis was consistent with random cell death rather than a fixed lifetime. The percentage of cells undergoing cytolysis was governed by the cell density; at high densities, such as are present in confluent cultures, a higher percentage of cell loss was noted than at low density. Cell death was antagonized by partially purified or electrophoretically homogenous preparations of-PDGF. Pure PDGF stimulated cell survivial at ng/ml in a concentration dependent fashion. The process of cell replication was not necessary for survival because PDGF prevented cytolysis in the presence of methotrexate, an inhibitor of DNA synthesis. A brief (4 hour) treatment with PDGF prevented cell death; such PDGF treated cells displayed increased survival after being taken up with trypsin and planted onto a fresh surface in plasma supplemented medium. Pituitary fibroblast growth factor, a functional analogue of PDGF for induc of DNA synthesis in BALB/c-3T3 cells, also functioned as an anticytolytic agent. By contrast, epidermal growth factor and insulin did not. Cytolysis of SV40-transformed cells occurred at a constitutively low rate and was insensitive to PDGF.     

Sera and conditioned media contain different isoforms of platelet-derived growth factor (PDGF) which bind to different classes of PDGF receptor

Platelet-derived growth factor (PDGF) is encoded by separate genes for two possible subunit chains (A-chain and B-chain) which can form three possible dimers (AA, AB, and BB). We have recently presented evidence that multiple forms of PDGF receptor exist which distinguish between these isoforms (Hart, C. H., Forstrom, J. W., Kelley, J. D., Smith, R. A., Ross, R., Murray, M. J., and Bowen-Pope, D. F. (1988) Science 240, 1529-1531). We used this specificity to determine the amount of PDGF from different sources which is able to bind to each class of receptor and found that each source had a characteristic isoform composition. Levels of total PDGF activity in sera from different species ranged more than 15-fold, from less than 1 ng/ml in dog, chicken, pig, and calf, to greater than 13 ng/ml in mouse and human. Despite these differences in PDGF content, the total mitogenic activities of the sera were comparable indicating that the relative importance of PDGF as a serum mitogen may vary considerably between species. Analysis of the total PDGF into the amounts of each isoform revealed great differences in composition. PDGF-BB constitutes only about 15% of the total binding activity in human PDGF purified by the method of Raines and Ross (Raines, E. W., and Ross, R. (1982) J. Biol. Chem. 257, 5154-5160) but is the predominant isoform in whole blood serum from all other species. In contrast to serum, medium conditioned by cultured PDGF-secreting cell types contained no detectable PDGF-BB except in two cases: medium conditioned by vascular endothelial cells and by cells transformed by simian sarcoma virus. The existence of isoform-specific PDGF receptors and the large variation in PDGF isoform composition dependent upon source may provide an important mechanism through which the effects of PDGF can be targeted to different cell types and/or toward eliciting different cell responses.     

Label-free and quantitative analysis of C-reactive protein in human sera by tagged-internal standard assay on antibody arrays

We have developed a new, high-throughput, competition-based tagged-internal standard (TIS) assay to measure the levels of blood proteins in human serum. In this assay, target proteins in the sample serum compete with tagged-internal standard proteins for binding to an antibody array. Antibody arrays are fabricated by immobilizing a target protein-specific antibody on the carboxylate-modified latex bead surface of well-type arrays. A solution of Alexa 546-conjugated target protein is added to a sample of human serum and applied to the well-type antibody array. The array is then analyzed with a fluorescence scanner and the level of unlabeled target protein in the human sera is inferred from the amount of tagged protein bound to the array. We successfully applied this assay to measure the level of C-reactive protein (CRP) in 92 unlabeled human sera. The TIS assay was found to be specific and reproducible for the quantitative analysis of CRP. The antibody array data from the TIS assay correlate well with clinical laboratory data obtained using the commercialized latex-enhanced turbidimetry immunoassay (n=3, r=0.967, CV=0.32%). Thus, the antibody array-based TIS assay system is high-throughput, quantitative, and label-free and may be useful in the rapid serodiagnosis of human disease.     

Neutralization sensitivity of cell culture-passaged simian immunodeficiency virus.

CEMx174- and C8166-45-based cell lines which contain a secreted alkaline phosphatase (SEAP) reporter gene under the control of a tat-responsive promoter derived from either SIVmac239 or HIV-1(NL4-3) were constructed. Basal levels of SEAP activity from these cell lines were low but were greatly stimulated upon transfection of tat expression plasmids. Infection of these cell lines with simian immunodeficiency virus (SIV) or human immunodeficiency virus type 1 (HIV-1) resulted in a dramatic increase in SEAP production within 48 to 72 h that directly correlated with the amount of infecting virus. When combined with chemiluminescent measurement of SEAP activity in the cell-free supernatant, these cells formed the basis of a rapid, sensitive, and quantitative assay for SIV and HIV infectivity and neutralization. Eight of eight primary isolates of HIV-1 that were tested induced readily measurable SEAP activity in this system. While serum neutralization of cloned SIVmac239 was difficult to detect with other assays, neutralization of SIVmac239 was readily detected at low titers with this new assay system. The neutralization sensitivities of two stocks of SIVmac251 with different cell culture passage histories were tested by using sera from SIV-infected monkeys. The primary stock of SIVmac251 had been passaged only twice through primary cultures of rhesus monkey peripheral blood mononuclear cells, while the laboratory-adapted stock had been extensively passaged through the MT4 immortalized T-cell line. The primary stock of SIVmac251 was much more resistant to neutralization by a battery of polyclonal sera from SIV-infected monkeys than was the laboratory-adapted virus. Thus, SIVmac appears to be similar to HIV-1 in that extensive laboratory passage through T-cell lines resulted in a virus that is much more sensitive to serum neutralization.     

Partial purification and characterization of porcine platelet-derived growth factor (PDGF)

Platelet derived growth factor (PDGF) has been partially purified from porcine platelets. Purification steps included heparin-agarose chromatography of the material released by thrombin-stimulated washed porcine platelets and Blue-Sepharose chromatography. Preparative isoelectric focusing showed that isoelectric point of porcine PDGF is at pH 10.0–11.0 and elution experiments from sodium dodecyl sulfate (SDS) polyacrymlam de gels indicated that its molecular weight is close to 30 kD. The immunoglobulin fraction prepared from anti-human PDGF serum inhibited the mitogenic activity of porcine PDGF. These experiments suggest a homology of porcine and human PDGF. Porcine platelet factor 4 and porcine platelet basic protein were devoid of significant mitogenic activity.     

A 125I-protein A-binding assay detecting antibodies to cell surface antigens

A 125I-protein A-binding assay detecting antibodies to cell surface antigens on human blood cells was developed and evaluated using sera from multitransfused nonleukemic patients sensitized against HLA antigens. The binding assay was found to be reproducible and more sensitive than conventional HLA testing. Seven patients with acute myelogenous leukemia and two patients with acute lymphoblastic leukemia successfully treated by chemotherapy were then investigated. Sera from seven of the patients studied in partial or complete remission demonstrated significant binding to autochthonous leukemic cells obtained from bone marrow or peripheral blood. In two cases sera taken during the leukemic stage demonstrated the most pronounced binding to the patients' own leukemic cells. Sera from four patients with demonstrable significant binding to autochthonous leukemic cells failed to bind to autochthonous remission cells when both types of target cells were tested in parallel. Differences in serum concentrations of IgG, IgA, and IgM were not the cause of the demonstrated increased binding of leukemic sera to autochthonous target cells. We propose that the 125I-protein A-binding assay presented in this paper detects antibodies reacting selectively with acute leukemia cells.     

Novel peripheral blood-derived human cell lines with properties of megakaryocytes

For 18 mo, we derived 18 cell lines from 11 donors with various clinical profiles ranging from normal to leukemic. Suspension cultures were initiated with 1 X 10(6) mononuclear blood cells/ml of nutrient medium containing 10% human serum and 10% lectin-stimulated human lymphocyte conditioned medium. The cultures were monitored weekly by morphological analyses of Wright-Giemsa-stained cell preparations. All successful cultures showed a significant decline in viability during the first 3-4 wk with rate "lymphoid" cells observed in mitosis. Within the next 2 wk, the proliferating cells gave rise to a rapidly expanding population of mononuclear cells. As the cultures expanded, cell morphology became heterogeneous with respect to cell size and nuclear ploidy, with an accumulation of giant multinuclear cells that were suggestive of megakarocytes. Even though the cells did not have the classical morphology of mature platelet-forming megakaryocytes, 90% of the cells within a cell line were positive by direct or indirect immunofluorescence for the platelet membrane glycoproteins IIb and IIIa; for surface markers HLA-Dr and B2-microglobulin; for intracellular platelet-derived growth factor and platelet factor IV; and for membrane affinity or binding with serum platelet-derived growth factor and platelet factor IV. These results suggest that a blood precursor cell, most likely a primitive megakaryoblast, was isolated from the peripheral blood and was provided with an optimal culture environment for sustained growth. These cells did not mature to a more differentiated stage, perhaps owing to regulatory factor deficiencies in this in vitro system. The remarkable frequency of obtaining cell lines with megakaryocyte properties from normal peripheral blood and the capacity of some normal donors to repeatedly yield these cell lines make this cell culture system indeed unique by being selective for putative megakaryocyte precursors.