Effect of filtration of platelet concentrates on the accumulation of cytokines and platelet release factors during storage

BACKGROUND

Platelet transfusions are frequently accompanied by febrile nonhemolytic transfusion reactions. These may be due, in part, to the release of cytokines interleukin 1 beta (IL-1 beta), interleukin 6 (IL-6), interleukin 8 (IL-8), and tumor necrosis factor alpha (TNF-alpha) by white cells (WBCs) into the plasma during storage of platelet concentrates (PCs). Acting as endogenous pyrogens, these agents may induce inflammatory responses.

STUDY DESIGN AND METHODS

This study proposed to determine if WBC reduction in PCs by filtration significantly reduced the levels of cytokines normally generated during storage of unfiltered PCs up to 5 days. Serotonin, platelet-derived growth factor (PDGF-AB), and von Willebrand factor levels were also assessed to establish whether or not filtration or storage elicited significant platelet activation and granule release.

RESULTS

Filtration significantly reduced total WBC counts by 99.1 percent before storage (p < 0.001) without affecting total platelet counts. Compared to unfiltered PCs, filtration prevented a rise in the levels of each cytokine by Day 3 for IL-1 beta (27.7 vs. 0.6 pg/mL; p < 0.05), IL-6 (114.2 vs. 0.4 pg/mL; p < 0.001), and IL-8 (4.2 vs. 0.02 ng/mL; p < 0.001). By Day 5, further increases in the levels of all cytokines were noted in unfiltered PCs, but Day 0 levels remained in filtered PCs (IL-1 beta: 105.4 vs. 0.4 pg/mL, p < 0.001; TNF-alpha: 42.2 vs. 7.5 pg/mL, p < 0.025; IL-6: 268.8 vs. 0.4 pg/mL, p < 0.001; and IL-8: 7.6 vs. 0.02 ng/mL, p < 0.001). From Day 0 to Day 5, there were significant increases in serotonin (21.3 vs. 6.3 ng/mL, p < 0.05), PDGF-AB (72.6 vs. 25.8 ng/mL, p < 0.001), and von Willebrand factor (4.7 vs. 2.7 IU/mL, p < 0.05) in unfiltered PCs, with similar increased levels being observed in filtered PCs during storage.

CONCLUSION

These data indicate that the accumulation of high levels of cytokines in stored PCs could be prevented by WBC-reduction filtration of PCs without the induction of significant platelet activation or granule release. As cytokines have the potential to induce febrile nonhemolytic transfusion reactions in patients, the transfusion of WBC-reduced PCs would be expected to reduce the frequency and severity of such reactions.

Adenosine nucleotides and serotonin stimulate von Willebrand factor release from cultured human endothelial cells

Endothelial cells (ECs) synthesize and release von Willebrand factor (vWf) either constitutively or from Weibel-Palade bodies by a regulated pathway. Although stimulated release of vWf from ECs occurs following exposure to thrombin, histamine, interleukin, tissue necrosis factor and fibrin in vitro, these agents are unlikely to be present in physiologically relevant concentrations during the initial stages of primary hemostasis. Alternatively, agents known to be released from the dense granules of activated platelets at the sites of vascular injury may provide the initial physiological stimuli for vWf release from ECs in vivo. We have examined the effects of the platelet secretagogues ADP, AMP, ATP and serotonin on the release of vWf from ECs and demonstrated enhanced release in all cases. The extent and time at which optimum vWf release was observed depended on the agonist and its concentration. At 3 nM, optimum release occurred after 4 hours with ADP (330%/ml) or 1 h with AMP (153%/ml) or ATP (450%/ml). At 30 nM, optimum release was seen after 1 hour with ADP (315%/ml) or AMP (595%/ml) and after 15 min with ATP (938%/ml). With serotonin, optimal release was seen by 30 min at 0.3 microM (1034%/ml) and after 1 h at 1 microM (745%/ml) although the response after 15 min was nearly equivalent (667%/ml). The doses giving 50% of maximal response (ED50) after 1 h were 6.5 nM (ADP), 15.2 nM (AMP) and 2.4 nM (ATP) and 20 nM for ATP or 75 nM for serotonin after 15 or 30 min respectively. ADP also enhanced PGI2 release from ECs in a dose- and time-dependent manner.(ABSTRACT TRUNCATED AT 250 WORDS)

Platelet-activating factor secreted by DDAVP-treated monocytes mediates von Willebrand factor release from endothelial cells

We have previously shown that although DDAVP (1-deamino-8-D-arginine vasopressin), a synthetic analogue of the natural hormone arginine vasopressin, does not directly promote release of vWf from human umbilical vein endothelial cells (ECs), enhanced release does occur when ECs were exposed to either monocytes or to supernatants recovered from DDAVP-treated monocytes. In the present study, we have found that exposure of monocytes to DDAVP did not increase secretion of interleukins (IL)-1 beta, IL-6, IL-8, tumor necrosis factor (TNF-alpha), growth factors G-CSF (granulocyte-), GM-CSF (granulocyte, monocyte-colony stimulating factor), prostaglandins (PG) E2, PGF2 alpha, or PGI2 or purine nucleotides such as ATP and ADP. However, increased levels of platelet-activating factor (PAF) were secreted by DDAVP-treated monocytes in a time- and dose-dependent manner that positively correlated with the enhancement in vWf release from ECs. Moreover, this effect could also be elicited when lipid extracts of these supernatants or purified PAF were added directly to ECs. This response could be inhibited with (+/-)-trans-2,5-Bis(3,4,5-trimethoxyphenyl)-1,3-dioxolane, a specific PAF receptor antagonist, when the ECs were exposed to supernatants from DDAVP-treated monocytes or to pure PAF. The present data indicate that enhanced secretion of PAF from monocytes is one mechanism whereby DDAVP can provoke release of vWf from ECs.

Isolation of a novel 45 kDa plasminogen receptor from human endothelial cells

We have previously identified an endothelial cell membrane protein of M(r) 45 kDa that binds plasminogen in a kringle-dependent, specific and reversible manner (Dudani et. al. (1991) Mol. Cell. Biochem. 108: 133-139). In this study, we have developed and optimized a protocol for the isolation of the 45 kDa plasminogen receptor from venous endothelial cells using a four step procedure consisting of lysis and detergent extraction followed by ligand affinity chromatography and preparative polyacrylamide gel electrophoresis. Control experiments were carried out using BSA-Sepharose instead of plasminogen-Sepharose as the affinity matrix. No plasminogen binding proteins were recovered from the former columns. However, a 45 kDa protein was recovered from lysine eluates of plasminogen-Sepharose. This material was then purified to homogeneity using preoperative electrophoresis. Analyses of proteins at various steps in the purification by SDS-PAGE showed enrichment of a band of 45 kDa which superimposed with the observed binding activity of plasminogen in ligand blots. The above binding could be inhibited by excess lysine. The 45 kDa protein could be distinguished from alpha-enolase which also binds plasminogen by: (i) significant differences in the profile of retention times of CNBr-degradation fragments on reversed phase HPLC; and (ii) partial peptide sequencing of one of the CNBr-degradation fragments of the 45 kDa protein. Moreover, the derived sequence did not show any significant homology to any protein in the Swiss Prot (release 20) database. We thus propose that the 45 kDa protein represents a novel plasminogen receptor on human venous endothelial cells.

Target antigens of anti-platelet antibodies in patients with steroid-resistant ITP of recent onset. Canadian Apheresis Study Group

Recent application of Western blotting procedures in the detection of anti-platelet antibodies has permitted more refined definition of reactivity. We report on the results of anti-platelet antibody assays in a series of 19 patients with recently diagnosed (less than 6 months duration) steroid-resistant, idiopathic thrombocytopenia purpura (ITP). At presentation, six of the 19 patients had a positive test for platelet-associated IgG (PAIgG) as measured by the direct radial-immunodiffusion (RID) assay, whereas three of the 19 were positive with the direct-platelet-suspension-immunofluorescence test (DPSIFT). The indirect-platelet-suspension-immunofluorescence test (IPSIFT) demonstrated antibodies in seven of the sera. Following Western blot (WB) analysis, the serum of 7/19 patients (only three of which were positive in the IPSIFT) could be shown to react with platelet antigens. Two patterns were seen: in four cases there were bands of apparent molecular weights of 60,000, 55,000 and 50,000; in the other three samples, a single band near 90,000 was demonstrated. Unlike the situation reported for chronic ITP, no reactivity was seen against higher molecular weight antigens other than the 'non-specific' binding at apparent molecular weight 200,000 which is also seen with normal sera. The data suggest that antibodies reacting against specific platelet antigens are present in the serum of some patients with recent onset ITP.

Expression of stem cell factor and c-kit mRNA in cultured endothelial cells, monocytes and cloned human bone marrow stromal cells (CFU-RF)

Previously, we have shown that conditioned medium from a subpopulation of human marrow stromal cells (CFU-RF) contain an activity able to stimulate the growth of macroscopic epo-dependent erythroid colonies. The ligand for the product of the c-kit proto-oncogene (also known as stem cell factor or SCF), among other activities, has been reported to have similar effects on erythroid colony growth. We have also presented data showing that SCF together with phytohemagglutinin-stimulated leukocyte conditioned medium can stimulate erythroid colony growth in the presence of antibodies to erythropoietin. Using the human SCF cDNA probe (K. Zsebo, Amgen Inc.) we now show that cells derived from CFU-RF colonies express SCF but not c-kit. Human umbilical vein endothelial cells were also found to express SCF and this expression was increased by addition of monocyte supernatant, IL-1 beta or thrombin. Cells of the human erythroleukemia cell line HEL were found to express c-kit but not SCF. Neither c-kit nor SCF mRNA were detected in phytohemagglutinin-stimulated lymphocytes. Together, these data support the view that the behaviour of proliferating erythroid stem cells in the marrow, which may express c-kit, could be regulated by membrane-bound SCF present on surrounding stromal cells.

Identification of an endothelial cell surface protein that binds plasminogen

To identify and characterize endothelial cell surface components that bind plasminogen, we used ligand-blotting to study binding of plasminogen to sodium dodecyl sulphate solubilized extracts of human umbilical vein endothelial cells. It was observed that glu-plasminogen bound predominantly to a 45 kDa endothelial cell polypeptide. The interaction of labelled glu-plasminogen with this polypeptide was reversible and specific as the binding could be inhibited by both excess cold lysine and unlabelled glu-plasminogen but not by unrelated proteins. Binding of glu-plasminogen to cell extracts prepared from endothelial cells that had been pretreated with proteinase K was significantly reduced indicating that the 45 kDa polypeptide is a cell-surface protein. The cell-surface localization of the 45 kDa polypeptide was also indicated by the positive interaction of glu-plasminogen with membrane fractions of endothelial cells. Lys-plasminogen also interacted with the 45 kDa polypeptide in a specific manner and reversibility experiments indicated that lys-plasminogen could also displace the bound glu-plasminogen. Since binding of plasminogen to the 45 kDa endothelial cell surface polypeptide was very similar to plasminogen binding to intact endothelial cells, we propose that the 45 kDa protein represents one of the major receptors for plasminogen on human endothelial cells.

Characterization of novel platelet and endothelial cell target antigens in a family with genetic susceptibility to autoimmunity

This report describes a French Canadian family whose members exhibit a high incidence of allo- and autoantibodies to antigens present on both platelets and endothelial cells. This is correlated with various HLA specificities known to be associated with autoimmunity, such as A1, B8, DR3, and, in some cases, with clinical disorders, including nephritis, hypertension, and thrombocytopenia. Immunoblot analysis using platelet and endothelial cell lysates showed serum antibodies to a 75 kDa endothelial cell surface polypeptide and to polypeptides with apparent mass of 115 kDa and 26 kDa found on both platelets and endothelial cells. This 115 kDa internal platelet protein was also found in a variety of other cell types, such as mononuclear cells, and increased following cell activation. Monoclonal antibody immunobilization assays were used to characterize the 26 kDa polypeptide; in three of the four patients tested, an antibody to leukocyte differentiation antigen CD9 was identified. The asymptomatic child of the propositus also exhibited an autoantibody against an 80 kDa platelet protein which was sensitive to thrombin digestion, suggesting that this polypeptide may be platelet glycoprotein V. In addition, P1A1 alloantibody was identified in one sister who had given birth to a severely thrombocytopenic boy and who herself had a severe vascular rejection to a cadaver kidney 2 years prior to this study. The propositus also developed hypertensive renal disease following a pregnancy and became dialysis dependent. Thus, members of this family have developed a variety of antibodies, particularly to platelet and endothelial cell antigens. Some subjects have remained asymptomatic in spite of having autoantibodies. However, others have been seriously ill, and their immune response to these antigens is believed to have played a role in the pathogenesis of their neonatal alloimmune thrombocytopenic purpura, hypertensive renal disease, renal graft rejection, and thrombocytopenia.

Definition of the affinity of binding between human von Willebrand factor and coagulation factor VIII

Factor VIII and von Willebrand factor are two plasma proteins essential for effective hemostasis. In vivo, they form a non-covalent complex whose association appears to be metal ion dependent. However, a precise definition of the nature of the molecular forces governing their association remains to be defined, as does their binding affinity. In this paper we have determined the dissociation constant and stoichiometry for Factor VIII binding to immobilized von Willebrand factor. The data demonstrate that these proteins interact saturably and with relatively high affinity. Computer assisted analyses of the Scatchard data favour a two site binding model. The higher affinity site was found to have a Kd of 62 (+/- 13) x 10(-12) M while that of the lower affinity site was 380 (+/- 92) x 10(-12) M. The density of Factor VIII binding sites (Bmax) present on von Willebrand factor was 31 (+/- 3) pM for the high affinity binding site and 46 (+/- 6) pM for the lower site, corresponding to a calculated Factor VIII: von Willebrand factor binding ratio of 1:33 and 1:23, respectively.

Characterization of plasminogen binding to human capillary and arterial endothelial cells

Phenotypic diversity of endothelial cells that line the various vascular spaces has been well established. However, it is not known if biochemical differences also exist, particularly in the numbers of receptors for plasma proteins. Equilibrium binding techniques were used to assess potential differences in the binding of 125I-labelled plasminogen to cultured human umbilical arterial endothelial cells and capillary endothelium, as compared with umbilical venous cells. The kinetic behaviour of plasminogen binding to all three types of cells was similar, with optimal binding occurring between 20 and 30 min of incubation. Binding of plasminogen to arterial, capillary, and venous cells was concentration dependent and reversible upon addition to excess unlabelled plasminogen. Scatchard analyses showed that artery, capillary, and venous endothelial cells all possess low affinity sites for plasminogen with Kd values of 0.30 +/- 0.07, 0.40 +/- 0.06, and 0.40 +/- 0.08 microM, respectively. Vein cells also possess an additional higher affinity binding site with a Kd of 0.07 +/- 0.01 microM, exhibiting a 6-fold greater affinity for plasminogen than the lower affinity sites on capillary and arterial endothelial cells. Assuming a stoichiometry of 1:1 for binding, the data indicate that arterial and capillary endothelial cells contain approximately 4.2 (+/- 0.9) x 10(6) and 4.1 (+/- 0.6) x 10(6) plasminogen receptors per cell. Venous cells contain both low and high density binding sites with 6.2 (+/- 0.8) x 10(6) and 12.4 (+/- 2.4) x 10(6) sites per endothelial cell. The presence of a higher affinity site on vein cells, but not on artery or capillary cells, may signal functional differences relating to fibrinolytic activity on the surface of these cells.(ABSTRACT TRUNCATED AT 250 WORDS)

DDAVP-induced release of von Willebrand factor from endothelial cells in vitro: the effect of plasma and blood cells

The vasopressin analogue 1-deamino-8-D-arginine vasopressin (DDAVP) causes an immediate, transient rise in plasma levels of von Willebrand factor (vWF) after its administration. Although it is recognized that vascular endothelial cells play an essential role in this process, the molecular basis of the response is not understood. We have investigated the phenomenon using human umbilical vein endothelial cells as an in vitro model. When normal individuals were stimulated with DDAVP, plasma from blood samples collected subsequently caused the release of vWF from cultured endothelial cells over a 24 h period (22-46% increase over baseline), compared to control plasma (5-17%). DDAVP added directly to the endothelial cells produced no increase in vWF release. When whole blood was treated in vitro with DDAVP, and the plasma subsequently added to endothelial cells, a significant increase in vWF secretion was found. Peripheral blood mononuclear cells were then tested. In the presence of DDAVP, an increased response occurred. Further fractionation of these cells showed that monocytes were largely responsible, causing an increased vWF release of 162% at 2 h. These observations were reinforced by finding that the supernatants of monocytes incubated with DDAVP were also effective in causing increased vWF release (118% compared to 58% for the control sample). Our studies suggest that DDAVP plays an indirect role in causing the release of vWF from endothelial cells, and that peripheral blood monocytes may act as intermediary target cells, which then produce factor(s) acting directly on endothelial cells.

An antibody to platelet membrane antigen associated with thrombotic events in essential thrombocythemia

A 71-year-old man with essential thrombocythemia presented with acute onset of thrombosis in his fingers. The platelet count at presentation was 752 x 10(9)/liter, and his PTT was within the normal range. Platelet aggregation was normal in response to ADP, decreased to epinephrine, and absent to collagen, with the patient receiving 1.2 g aspirin/day. Plasma vWF:Ag was 158 U/dl, VIII:C was 150 U/dl, and the vWF:Ag multimer pattern was normal. HLA antibodies were not detected in his serum, and the platelet-associated IgG was normal at 3.9 fg/platelet. Incubation of normal, unactivated platelets with the patient's plasma did not result in agglutination of the platelets. However, a serum antibody reacting against a platelet antigen with an apparent molecular weight of 45,000 could be demonstrated by Western blotting. This antibody also reacted with endothelial cells. The reactive pattern was similar to that of commercially obtained actin; the serum antibody could be removed by preincubation with actin. Platelet-plasmapheresis was carried out daily for 10 days, after which the serum still reacted weakly with the platelet antigen. By day 20, the band could still be identified on the electroblot, but 2 weeks later no band could be seen. Similar assay of sera from six other patients with myeloproliferative disorders and comparable thrombocytosis but without thrombotic manifestations did not show any such reactions. The data suggest a correlation between the presence of an antibody possibly directed against platelet and/or endothelial cell actin and the vascular events in this patient with essential thrombocythemia.

Anti-endothelial cell antibodies: detection and characterization using a cellular enzyme-linked immunosorbent assay

Anti-endothelial cell antibodies (AECA) have been detected in autoimmune diseases such as systemic lupus erythematosus (SLE) and scleroderma (PSS) but their role in pathogenesis is unknown. Immunofluorescence, immunohistochemistry, complement-dependent antibody lysis, and radioimmunoassay have been used in the past to detect AECA. We have developed a rapid, sensitive, and quantitative cellular enzyme-linked immunosorbent assay (ELISA) to detect and characterize AECA. Sera were obtained from 28 normal volunteers, 28 patients with SLE, and 14 patients with PSS. We also performed studies in 47 patients with various monoclonal gammopathies. Endothelial cells (EC) were obtained from human umbilical veins by standard methods and subcultured on 96-well tissue culture plates without fixation. EC were then sequentially incubated with sera, peroxidase-conjugated goat anti-human Ig (IgG, IgM, or IgA), and substrate. Optical density readings were converted to arbitrary units by developing a standard curve. Heavy-chain specific antibodies were used to determine the class of AECA binding to EC. IgG was purified by using protein A columns and digested with pepsin to obtain F(ab')2 fragments. The mean units of AECA from normals were 19.3 for IgG and 12.5 for IgM. SLE sera showed significant levels of IgM AECA (37 units, P less than 0.001) but not IgG (29 units, P less than 0.1). PSS sera showed significant levels of both IgM AECA (38 units, P = 0.001) and IgG AECA (42.7 units, P less than 0.005). IgA AECA were not detected in normal, SLE, or PSS sera. Blocking Fc receptors with rabbit IgG did not affect the titer of IgG or IgM AECA.(ABSTRACT TRUNCATED AT 250 WORDS)

HLA-D restriction of "naturally occurring" MLR suppressor cells in acquired common variable hypogammaglobulinemia

Cells with capacity to suppress the mixed lymphocyte response (MLR) were detected in two patients with acquired common variable hypogammaglobulinemia (ACVH). No specificity with respect to the stimulating HLA type was observed. In one case of ACVH available for extensive study, there was evidence for HLA-D restriction of the suppressor cells. The patient's lymphocytes specifically suppressed the MLR of subjects who carried the same HLA-D type. Family studies confirmed that the suppressor activity was restricted to HLA-D, not DR, and segregated with the appropriate HLA haplotype. These observations suggest that an immune suppressor gene mapping in the HLA region may be involved in the pathogenesis of common variable hypogammaglobulinemia.