Katzmann, J A

Abstract Myeloma is a neoplasia characterized by the accumulation of malignant plasma cells in the bone marrow. In these studies, we have demonstrated that CD40 is expressed in human myeloma cells and have used a recently established IL-6-dependent myeloma cell line, ANBL-6, to examine the potential function of CD40 expression in myeloma cells. In addition to its expression on the ANBL-6 cells, we show that CD40 is expressed on freshly isolated myeloma cells from seven of seven patients tested. To address the role of CD40 expression in myeloma cells, we have examined the responsiveness of the ANBL-6 cell line to a CD40-specific mAb, G28-5. This cell line has previously been shown to proliferate only in response to IL-6. Of interest in this study, G28-5 also induced proliferation of the ANBL-6 cells. This proliferation was substantially inhibited by an IL-6-neutralizing mAb. Analysis of ANBL-6 cell culture supernatants by ELISA demonstrated that G28-5-stimulated cells secreted significant levels of IL-6, whereas unstimulated cell culture supernatants contained undetectable levels of IL-6. Furthermore, CV-1/EBNA cells expressing the human CD40 ligand also induced the proliferation of the ANBL-6 cell line, an effect that was inhibited by the anti-IL-6 mAb. Lastly, RNA blot analysis demonstrated an increase in IL-6 message in G28-5-stimulated ANBL-6 cells over unstimulated cells. These results indicate that the primary mechanism of anti-CD40-stimulated proliferation of the ANBL-6 cells is the induction of autocrine IL-6 production. Moreover, these data suggest that the expression of CD40 in malignant plasma cells may play a role in tumor cell expansion, possibly by stimulating the induction of autocrine IL-6 secretion.

For further study of the porcine hemostatic mechanism, we have prepared murine monoclonal antibodies, and F(ab')2 and Fab fragments, specific for porcine platelet membrane glycoproteins GP lb and GP Ilb/IIIa. To avoid production of antibodies to von Willebrand factor (vWF), mice were immunized with platelets obtained from pigs with severe von Willebrand,s disease. One monoclonal antibody (PP3-4C), of IgG1 subclass, caused 85% inhibition of Ristocetin-induced platelet binding of 125I-vWF (porcine) at ≥12 µg IgG/ml. PP3-4C did not affect ADP or collagen-induced platelet aggregation nor inhibit 125I-fibrinogen (porcine) binding. Pepsin and papain digestion, respectively, were used to prepare PP3-4C F(ab')2 and Fab fragments. PP3-4C F(ab')2 at concentrations ≥12 µg/ml caused 80% inhibition of washed platelet agglutination in the presence of vWF and Ristocetin, whereas Fab fragments at concentrations ≥10 µg/ml caused 60% inhibition. Another monoclonal antibody (PP3-3A), of IgG1 subclass, completely inhibited ADP or collagen-induced platelet aggregation at an IgG concentration of 6 µg/ml. At 10 µg IgG/ml PP3-3A completely inhibited binding either of 125I-fibrinogen or of 125I-vWF to ADP-stimulated porcine platelets. PP3-3A did not affect vWF-dependent Ristocetin-induced platelet agglutination, nor 125I-vWF binding to platelets in the presence of Ristocetin. PP3-3A did not bind to platelets which were treated with 10 mM EDTA at 37°C for 60 min. F(ab')2 and Fab fragments were isolated from PP3-3A pepsin or papain digests. Both types of PP3-3A fragments caused 100% inhibition of ADP-induced platelet aggregation, at concentrations ≥6 yg/ml. Immunoprecipitation of surface-radiolabeled porcine platelets and subsequent SDS-PAGE demonstrated that PP3-4C recognized a glycoprotein with molecular weight of 140,000 (under reducing conditions), and 165,000 (non-reduced). PP3-3A recognized glycoproteins with molecular weights of 115,000 and 100,000 (reduced), and 130,000 and 80,000 (non-reduced). Neither monoclonal antibody bound to human platelets. These monoclonal antibodies to porcine platelet membrane glycoproteins which are analogues of human GP lb and GP Ilb/IIIa will be useful for in vitro and in vivo studies to further understanding of mammalian hemostatic mechanisms.

Highly purified human Factor V was used for the development of a competitive double antibody radioimmunoassay (RIA) using 125I-human Factor V, burro anti-human Factor V antisera as the primary antibody and goat anti-burro antisera as the precipitating antibody. The standard curve allows the detection of as little as 20 ng of Factor V per ml of plasma. With this specific RIA for human Factor V, we have measured the level of Factor V in the plasma and platelets of normal individuals. The normal level of Factor V in plasma ranges from 4 to 14 μg per ml, with the average value equal to 7.0 ± 2.0 μg/ml (n = 64; 33 females; 31 males; 22 to 61 years of age). There appeared to be no correlation between antigen levels and age or sex. Factor V clotting assays were consistent with the RIA data for any given plasma preparation providing freshly drawn plasma was used in the bioassay. The bioassay data were quantitated based upon the specific activity of purified plasma Factor V; 1.7 units of Factor V equals 1 μg of protein. Plasma Factor V antigen levels were not affected by lyophilization of the plasma, prolonged storage of the plasma at -20°C or intentional conversion of the plasma to serum. The levels of Factor V present in washed human platelets were also determined using the RIA. Assay of washed platelets lysed in 0.2% Triton X-100 indicated that 0.6 to 0.85 μg of Factor V was present per 2.5 × 108 platelets (4400 to 6200 molecules of Factor V per platelet). This result is in marked contrast to our observations for the bovine system, where we found that bovine platelets possess approximately 400 to 800 molecules of Factor V per platelet, and plasma Factor V levels range from 30 to 50 μg per ml. In the bovine system, the platelets possess approximately 1% of the total Factor V present, while in human blood, the platelets possess as much as 10 to 15% of the total Factor V present.

Murine monoclonal antibodies to porcine Willebrand factor were used to study the role of Willebrand factor in hemostatic plug formation in the in vivo and in vitro skin bleeding times. The in vivo assay requires the intravenous injection of antibody-containing mouse ascitic fluids into normal animals with subsequent ear bleeding time tests performed over a 24-hour period. The in vitro assay was carried out with heparinized normal porcine blood flowing through a 0.5 cm incision in an excised piece of porcine skin. The heat inactivated ascitic fluids, containing the antibodies, were used ;Ln these assays at final dilutions of 2 × 102 through 2 × 105 The 7 antibodies used comprised at least 5 groups with differing reactivities based on assays other than bleeding time tests. The antibody titers in a poreineg Willebrand factor-binding radioimmunoassay ranged from 108 through 1012 . Doses of 10HL of ascitic fluid/kg of body weight (approximately 2 × 103 final dilution) were able to transiently prolong the in vivo bleeding time without alteration of other VIII complex values (VIII:C, ristocetin cofactor activity and Vlll-related antigen). Higher doses resulted in bleeding times similar to von Willebrand pigs (>15 min) immediately following infusion, with a decay of the effect over the next 2 hours. At dilutions of 2 × 104 selected monoclonal antibodies interfered with hemostasis in the in vitro model whereas other antibodies inhibited only at 2 × 102 dilution. With one exception, the potencies of the antibodies appeared to be similar in both assays. The titer of the antibodies in the radioimmunoassay does not appear to correlate with prolongation of the in vitro bleeding time, suggesting specific reactivity with functional sites involved in hemostatic plug formatio.

The peptides produced during the thrombin (IIa) catalyzed activation of human Factor V (HFV) were investigated using a murine monoclonal antibody. This antibody (HFV-1) binds HFV and activated HFV (HFVa) and releases them at high ionic strength. HFV-1 coupled to sepharose was used to purify the single chain, 330,000 apparent MW procofactor molecule. Gel electrophoresis in SDS indicated that IIa- mediated activation proceeded with the appearance of intermediates with apparent MW of 270K, 205K and 150K and resulted in at least 3 end products with apparent MW of 110K, 93K and 70K. Aliquots of partially and fully activated HFV were chromatographed on the HFV-1 affinity resin. The peptides not associated with HFVa were obtained by washing the column with 20 mM imidazole, 0.154 M NaCl, 5 mM CaCl2, pH 6.5; the peptides noncovalently bound to the epitope-containing peptides were obtained by elution in the above buffer made 5 mM EDTA in place of Ca++ and the epitope-containing peptides were eluted in buffer made 1.2 M in NaCl. The 150K intermediate was released in EDTA and the 205K and 270K peptides were eluted in high salt. Of the end products of activaion, the 110K peptide was not retained by the coliman, the 93K peptide was released in EDTA, and the 70K peptide was released only at high ionic strength. These data indicate that the 270K and 205K intermediates, and the 70K end product share the antigenic site involved in the interaction with the monoclonal antibody. In addition a Ca++- dependent interaction between the epitope-containing domain and the 150K intermediate and 93K end product is suggested. From this data and by analogy with the bovine protein, we suggest that HFVa is a Ca++-dependent multiple subunit protein consisting minimally of two peptides of apparent MW of 93,000 and 70,000.