Immune Adherence Research Articles

Autoimmune hemolytic anemia due to IgA class autoantibodies

Abstract Autoimmune hemolytic anemia clue to warm-reacting autoantibod-ies solely of the IgA class is very rare, and only live cases were identified among 5,177 patients referred during 13.5 years. All were females Cages 21-69 years) and all presented with idiopathic “Coombs negative1' autoimmune hemolytic anemia, a. diagnosis that was confirmed using monospecific anti-human IgA reagents. Red cell-bound IgG was reduced, but in two patients IgM was initially increased, an occurrence that was thought to reflect the developing autoimmune response. The autoantibodies had high affinity for red cells with very little free antibody detectable in the serum; in two instances Rh specificity was evident. Hemolysis was severe in four patients. Two of them had intravascular hemolysis, one of whom also had marked dyserythropoiesis and a transiently positive Ham′s test. Although IgA autoantibodies caused hemolysis predominantly through immune adherence, on occasions they also seemed to be able to induce complement activation, possibly via the alternative pathway. Prednisolone was the mainstay of treatment, and was occasionally augmented with azathioprine and intravenous immunoglobulin. Blood transfusion was required in two patients, both of whom eventually required splenectomy that resulted in full remission. The one patient with mild hemolysis recovered without treatment. Immunohematology 1996;12:14-19.

Detection of circulating immune complexes in dog sera by immune adherence hemagglutination method.

Circulating immune complexes (CIC) in the sera from 102 clinically healthy dogs and from 16 Dirofilaria immitis (D. immitis) infected dogs were measured by the immune adherence hemagglutination (IAHA) method. The level of CIC in the sera from healthy dogs was 28.2 +/- 29.1 micrograms/ml. There was no significant difference in the levels of CIC regarding sexes or ages. On the other hand, the level of CIC in the sera from D. immitis infected dogs was 230.3 +/- 117.1 micrograms/ml, which was significantly (p < 0.01) higher than that of healthy dogs. The results of this study confirm that IAHA method is one of the reliable means of detecting and/or diagnosing immune complex mediated diseases in dogs.

Beta 2-integrin (CD11b/CD18) is the primary adhesive glycoprotein complex involved in neutrophil-mediated immune injury to human sperm.

The pathogenesis of antisperm antibody (ASA)-mediated infertility is postulated to be related in part to complement (C)-dependent neutrophil-mediated injury to sperm in the female genital tract. We have reported that sperm-bound IgG activated human C and deposited C3 fragments on motile sperm. We also demonstrated that IgG and C3-bound motile sperm adhered to human neutrophils in vitro, and that this adhesion potentiated the localized release of oxygen radicals at the site of neutrophil/sperm membrane contact. The goal of the present study was to identify the neutrophil surface receptor(s) involved in neutrophil/sperm adhesion and to evaluate their relevance to the pathogenesis of neutrophil-mediated immune injury to sperm. Neutrophils were coincubated with motile sperm in the presence of C-fixing ASA+ sera or control sera. After defined incubation periods, the following neutrophil variables were evaluated: 1) surface expression of Fc (Fc gamma RII and Fc gamma RIII) and C receptors (CR1[CD35], CR3 [CD11b/CD18], and CR4 [CD11c/CD18]) by flow cytometry, 2) neutrophil aggregation by flow cytometry, 3) tyrosine phosphorylation of neutrophil proteins by flow cytometry, and 4) the immune adherence and ingestion of sperm by light and scanning electron microscopy (SEM). The functionality of adhesion receptors was studied by use of a panel of anti-leukocyte monoclonal antibodies (mAbs) for their ability to block neutrophil/sperm adhesion and neutrophil aggregation. Only the incubation of neutrophils and sperm in the presence of C-fixing ASA+ sera resulted in marked (> 70%) sperm binding to neutrophils. Consistent with this pattern, a significant (p < 0.0001) increase in surface expression of neutrophil CD11b and neutrophil aggregation was evident. According to SEM, most of the sperm were linked to the neutrophil by the acrosomal region of sperm head. Maximum expression of CD11b antigen was obtained when neutrophils were coincubated with sperm in the presence of C-fixing ASA+ sera. CD11b up-regulation correlated with a significant (p < 0.05) increase in tyrosine phosphorylation of neutrophil proteins during sperm phagocytosis. Only the combination of mAbs directed to the beta 2-integrin, CD11b (D12/SHCL-3), or CD18 (MHM23) subunits maximally inhibited ASA- and C-mediated sperm adhesion to neutrophils (by 70%) or sperm phagocytosis (by 75%), as well as neutrophil aggregation (by 96%). These findings strongly implicate the CD11b/CD18 glycoprotein complex (CR3) in the adhesive events involved in ASA- and C-mediated immune destruction of motile sperm by neutrophils.

Characterization of rat complement receptors and regulatory proteins. CR2 and Crry are conserved, and the C3b receptor of neutrophils and platelets is distinct from CR1.

In the mouse, CR1 and CR2 on B lymphocytes are encoded by alternatively spliced Cr2 gene transcripts. Immune adherence receptors that bind C3b are present on mouse platelets and unstimulated neutrophils, but they are not CR1. In this study, rabbit anti-mouse CR1/CR2 Ab immunoprecipitated a 145- to 150-kDa CR2 protein from rat platelets, neutrophils, and splenocytes, but not a approximately 200-kDa CR1 protein. By Northern analysis, cDNA for mouse CR2 hybridized to mRNA of 3.7 and 5.2 kb from both mouse and rat splenocytes. The murine decay-accelerating factor and membrane cofactor protein analogue Crry was present in rat platelets, neutrophils, E, and splenocytes as two distinct proteins of 65 to 70 kDa and 75 to 85 kDa. Rat platelets, neutrophils, and splenocytes contained a novel 200-kDa cell membrane protein that specifically bound to a rat C3b-Sepharose column. We have named this protein C3bR-200. C3bR-200 was not identified by anti-mouse CR1/CR2 or anti-human CR1 Ab. Rat E lacked C3bR-200. Rat neutrophils and splenocytes also contained an 80-kDa C3b-binding protein that was distinct from Crry, which we have named C3bR-80. Therefore, CR2 and Crry are present in the rat, and have similar qualities to those from the mouse, except that CR2 is located on rat platelets and neutrophils, which is not the case in the mouse. Rat platelets, neutrophils, and splenocytes have a 200-kDa C3b-binding protein, C3bR-200, that is likely to be the rodent immune adherence receptor.

Complement-independent binding of microorganisms to primate erythrocytes in vitro by cross-linked monoclonal antibodies via complement receptor 1.

Under certain circumstances, soluble antigens, particulate antigens, and/or microorganisms have been shown to bind to primate erythrocytes via complement receptor 1 (CR1) in the presence of specific antibodies and complement. This immune adherence reaction, specific for CR1, can lead to neutralization of antigens in the circulation and their subsequent clearance from the blood. The present experiments utilized cross-linked monoclonal antibody complexes (heteropolymers) with specificity for both CR1 and either 35S-labeled herpes simplex virus capsid or Haemophilus influenzae as prototype viral and bacterial particulate antigens, respectively. In each case, the respective specific heteropolymers facilitated binding of the target antigens (> or = 70 to 90%) in vitro to erythrocytes in the absence of complement. Several experimental protocols were employed to demonstrate that heteropolymers mediate specific, rapid (> or = 30 s), and quantitative binding of prototypical particulate pathogens to human and monkey erythrocytes but not to sheep erythrocytes, which lack CR1. These results extend the potential use of the erythrocyte-heteropolymer system to the neutralization and clearance of particulate viral and bacterial pathogens from the blood.

Identification of complement receptor type 1-related proteins on primate erythrocytes.

The purpose of this study was to characterize the structure and function of the immune adherence receptor (CR1, CD35, C3b/C4b receptor) of primates. Western blotting, immunoprecipitation, ELISA, and affinity chromatography with homologous C3b and C4b were utilized. The major cross-reactive E membrane protein of ten species of primates tested was lower in m.w. than was human CR1 and fell into two size groups of 55 to 75 and 130 to 165 kDa. There was 10- to 100-fold more CR1 per primate E than human E. Five species also expressed lesser quantities of a protein similar in m.w. (approximately 200 kDa) to human CR1. In contrast to E, the major cross-reactive protein on PBMC was similar in size to human CR1. Four species also expressed lesser amounts of a lower m.w. protein on their PBMC of the same M(r) as that found on their E. Affinity chromatography demonstrated that the approximately 200-kDa form, if present, was recovered with a similar efficiency to that of human CR1. Three patterns of binding, however, were identified among the lower m.w. proteins: 1) C3b > or = C4b; 2) C4b > C3b; and C3b only or predominantly. The fact that these E proteins cross-react with Ab to human CR1, bind homologous C3b and, in most cases, C4b, and for some species represent the only such protein expressed on their E identifies them as immune adherence receptors. The 70-kDa CR1 of the chimpanzee E seems to arise by alternative splicing of the mRNA encoding the 200-kDa protein. These data raise interesting questions relative to the evolution of CR1 in primates and provide a basis for analysis of structure-function relationships among these size forms of CR1.

Erythrocyte Complement Receptors

Primate erythrocytes express complement receptors (E-CR), which can extrinsically bind C3b and Cb4. This interaction allows primate erythrocytes to bind complement opsonized particles and immune complexes, a phenomenon historically referred to as immune adherence. The binding of C3b and C4b by E-CR also leads to inhibition of complement activation. The human E-Cr is the complement activation. The human E-CR is the complement receptor type 1, or CR1, which is codominantly expressed as four polymorphic allotypes, ranging in size from 190,000 to 280,000 M(r). Non-human primate E-CR are similar to CR1 in function and antigenicity and are likely homologous to CR1 in structure; however, they are one third to one half the size of CR1. The physiological role of E-CR, determined from studies in monkeys and humans, is to allow erythrocytes to perform as inert shuttles for circulating immune complexes (IC), safely directing IC to the organs of the monocyte phagocytic system, thus preventing indiscriminate IC deposition in vulnerable tissue. In IC-mediated diseases, such as systemic lupus erythematosus (SLE), detectable erythrocyte CR1 levels are reduced, an abnormality that in part is acquired during disease activity. The loss of erythrocyte CR1 may be an important pathogenic factor in the development and severity of SLE.

Antilipid a monoclonal antibody HA-1A decreases the capacity of bacterial lipopolysaccharide to activate human vascular endothelial cells by an immune adherence mechanism

Human monoclonal IgM antibody HA-1A, which recognizes the lipid A component of bacterial lipopolysaccharide (LPS), has been shown to reduce mortality in Gram negative septicemia. The vascular endothelial lining of blood vessels, which controls leucocyte traffic and activation, as well as haemostatic balance, may be one of the primary targets of LPS action during sepsis. In earlier studies we have described HA-1A-induced immune adherence of LPS to complement receptors on erythrocytes, and showed that pre-incubation with HA-1A, in the presence of complement and red blood cells, markedly reduced LPS-induced cytokine production from peripheral blood mononuclear cells. In the present study, we measured the effect of immune adherence of LPS in the presence of HA-1A on the responses of cultured endothelial cells, and found that subsequent expression of adhesion molecules such as E-selectin, ICAM-1 and VCAM-1, and secretion of the cytokines interleukin-6 and granulocyte-macrophage colony stimulating factor were markedly reduced. Moreover, the ability of LPS to increase levels of tissue factor procoagulant activity on endothelial cells was markedly diminished by LPS immune adherence to HA-1A. This decrease in endothelial activation in response to LPS following immune adherence to HA-1A may play a significant role in the protective effect of HA-1A in vivo during the course of Gram negative sepsis.

Antilipid A monoclonal antibody HA-1A: Immune complex clearance of endotoxin reduces TNF-α, IL-1β and IL-6 production

HA-1A is a human monoclonal IgM antibody which recognizes the lipid A component of lipoposaccharide (LPS). This antibody has reduced mortality in the septic shock syndrome resulting from Gram negative bacteria, in which many of the manifestations are considered to be due to cellular activation and secretion of cytokines, most notably TNF-α. However HA-1A does not directly neutralize LPS effectively in vitro, and studies reported to date have not defined its mechanism of action. Here we demonstrate that HA-1A, which in the presence of complement promotes immune adherence, may inhibit LPS action by facilitating its sequestration on red blood cells and clearance to an extent that cytokine production is reduced. Incubation of LPS at clinically significant (pg/ml) doses with HA-1A at therapeutic levels (e.g. 10 μg/ml) and complement resulted in LPS association with erythrocyte CR1 receptors. This reduced the ability of the residual, free LPS by 50–70% to induce the secretion of TNF-α, IL-1β and IL-6 from normal blood mononuclear cells. This mechanism is likely to be operative in vivo, ad could account for the protective effect of HA-1A, and its reduction of TNF-α production in vivo.

Human Anti-Endotoxin Antibody HA-1A Mediates Complement-Dependent Binding of Escherichia coli J5 Lipopolysaccharide to Complement Receptor Type 1 of Human Erythrocytes and Neutrophils

HA-1A has been shown clinically to decrease mortality in septic patients with gram-negative bacteremia. In this study, the ability of HA-1A to augment the serum complement-dependent immune adherence of 125I-labeled Escherichia coli J5 lipopolysaccharide (LPS) to human erythrocytes (RBC) and polymorphonuclear leukocytes (PMNL) was evaluated. In vitro studies indicated three things: HA-1A mediates immune adherence of 125I-J5 LPS to human RBC and PMNL in a dose-dependent manner; under these conditions, high concentrations of LPS (400 ng/mL) could be specifically bound. Immune adherence occurs via the classical complement pathway as demonstrated by its calcium dependence; HA-1A-J5 LPS-C' immune complexes bound to CR1 on human RBC and PMNL. PMNL binding and internalization of immune complexes was demonstrated by trypsin stripping of externally bound immune complexes. These studies support the proposal that HA-1A can lower the bioavailability of endotoxin by mediating binding and potential clearance of LPS via human RBC through the reticuloendothelial system or via direct internalization by peripheral blood PMNL.

A human monoclonal antibody that detects HLA-A1, A23 and A24 antigens.

We report the production and characterization of a human monoclonal IgM (mu, kappa) antibody recognizing the HLA A1, A23 and A24 antigens. B lymphocytes obtained from a multiparous Japanese woman were transformed in vitro by Epstein-Barr virus, screened with an immune adherence assay, and fused with a murine myeloma cell line, P3-X63-Ag8.653. After subcloning by limiting dilution three times, a stable antibody-secreting hybridoma cell line, 4-35-7, was identified. The culture supernant had a titer of 1:32-64 against each of A1-, A23- and A24-positive lymphocyte panels, and showed complete correlation (r = 1.00) with the A1, A23 and A24 antigens on a lymphocyte panel of 287 unrelated, class I HLA-typed donors by the NIH cytotoxicity assay. Monoclonality of the antibody was ensured by Southern blot analysis of the human immunoglobulin heavy chain gene of 4-35-7. In view of the published data on HLA class I nucleotide sequences, the antibody may recognize an antigeneic determinant including two amino acid residues, Asp-166 and Gly-167, in the alpha 2 helix of the class I molecule that are specific for A1, A23 and A24 so far analyzed.

The relative roles of C4A and C4B in prevention of immune precipitation, solubilisation and immune adherence.

C4A and C4B levels were measured in serum from 246 normal individuals. Complement-mediated solubilisation, assayed using alkaline phosphatase anti-alkaline phosphatase immune complexes (IC), correlated with both C4A and C4B levels. However, C4A and C4B levels showed no correlation with solubilisation of bovine serum albumin (BSA) ICs, or with the prevention of immune precipitation of BSA or alkaline phosphatase ICs, nor with immune adherence assayed using thyroglobulin and BSA ICs.

Complement-mediated opsonisation and phagocytosis of Renibacterium salmoninarum

Renibacterium salmoninarum depletes complement-mediated haemolytic activity of human and chinook salmon ( Oncorhynchus tshawytscha ) serum, and the extent of depletion depends on bacterial density. The bacterium also consumes human, chinook salmon, channel catfish ( Ictalurus punctatus ), and rainbow trout ( O, mykiss ) complement in the presence of Mg-EGTA indicating the involvement of the alternative complement pathway. Bacteria incubated in human serum with or without Mg-EGTA cause human erythrocytes to agglutinate. This immune adherence is blocked by incubation of the opsonised bacteria in the presence of a goat anti-human C3 antibody or incubation of the human erythrocytes with a monoclonal antibody against the human complement receptor CR1. Both results indicate that the bacteria have been opsonised with C3b. Furthermore, incubation of R. salmoninarum in haemolytically-active rainbow trout serum is required for the adherence of the bacterium to rainbow trout kidney macrophages prior to their phagocytosis.

Antibody responses to ganglio-series gangliosides in different strains of inbred mice

We studied antibody responses after immunization with ganglio-series gangliosides against 10 strains of inbred mice, including Balb/c, C57BL/6, A/J, C3H/HeN, C3H/HeJ, CBA/N, AKR/N, NZB/N, DBA/2 and nu/nu Balb/c. Twelve gangliosides having NeuAc as their sialic acid moiety (GM4, GM3, GM2, GM1, GD3, O-Ac-GD3, GD2, GD1a, GD1b, GT1a, GT1b and GQ1b), four gangliosides having NeuGc (GM3, GM2, GM1 and GD3) and four asialo-gangliosides (GA4, GA3, GA2 and GA1) were injected intravenously adsorbed to Salmonella minnesota. The antibody titers of the mice sera were determined by an enzyme-linked immunosorbent assay and an immune adherence assay. Antibody responses were found to depend not only on the ganglioside used as an immunogen but also on the mouse strain. Gangliosides having a trisaccharide sequence (NeuAcα2 → 8NeuAcα2 → 3Gal-) such as GD3, GD2, GDlb, GT1a and GQ1b, in particular O-Ac-GD3, induced high-titer antibody responses, whereas those having a disaccharide sequence (NeuAcα2 → 3Gal-) such as GM4, GM3, GM2, GM1, GDla and GTl1 induced low-liter antibody responses. On the other hand, gangliosides with NeuGc developed minimum titers. In contrast, asialo-gangliosides induced much higher responses than the corresponding gangliosides. The differences in ceramide portions of these gangliosides did not appear to be involved in inducing antibody responses. Mice could be divided into three groups according to the magnitude of their antibody responses: Group 1, those that produce the highest antibody responses (C3H/HeN and A/J); Group 2, those that demonstrate moderate antibody titers (Balb/c, C57BL/6, DBA/2 and nu/nu Balb/c); and Group 3, those that make minimum responses (AKR/N, C3H/HeJ, CBA/N and NZB/N). The pattern of reactivity to the various gangliosides was similar in all the strains tested.

Abnormal immune adherence and elimination of hepatitis B surface antigen/antibody complexes in patients with AIDS.

C3b-coated immune complexes (IC) adhere to complement receptor 1 (CR1) on human E in the circulation. E from AIDS patients have an acquired low CR1 number. To study immune adherence and IC elimination in AIDS, radiolabeled hepatitis B surface Ag/antibody complexes were injected i.v. in six AIDS patients and in 14 healthy controls. The binding of IC to E was reduced in AIDS patients (mean binding 2 min after injection: 24.9 +/- 13.3%) compared with healthy individuals (63 +/- 3.7%) (p = 0.0005). The low binding correlated directly with the number of CR1/E and to the capacity of these E to bind IC in vitro. During the first 15 min disappearance of IC was faster in AIDS patients than in normal subjects and correlated with CR1 number. Thereafter, elimination was very slow in AIDS patients, which suggested that a fraction of IC might be released back into the circulation similarly to what has been observed for C3b-coated E. When the data were analyzed with a mathematical model allowing for such release to occur, five of six AIDS patients had a high release rate compared with little or no release in normal individuals (p less than 0.001). Thus, low CR1 on E is responsible for defective immune adherence, and might determine abnormal disappearance of IC from the circulation as well.

Does reduced erythrocyte C3b receptor (CR1) activity contribute to the pathogenesis of yersinia triggered reactive arthritis?

Erythrocyte C3b receptor (CR1) activity was measured in 27 patients with yersinia triggered reactive arthritis and in 151 control subjects, including 36 patients with uncomplicated yersiniosis and 115 healthy subjects. CR1 was measured by the immune adherence haemagglutination method. Patients with yersinia triggered reactive arthritis had reduced levels of CR1 compared with the controls. This difference was mainly due to the finding that five out of six HLA B27 negative patients with arthritis had decreased CR1 activity. Such a quantitative difference may contribute to the pathogenesis of reactive arthritis by affecting the clearance of immune complexes.

免疫粘着反応および蛍光抗体法による婦人科腫ようのＯｎｃｏｆｅｔａｌ Ａｎｔｉｇｅｎ‐Ｉに関する研究

免疫粘着反応および蛍光抗体法による婦人科腫ようのＯｎｃｏｆｅｔａｌ Ａｎｔｉｇｅｎ‐Ｉに関する研究

Anti-ganglioside antibodies in myasthenia gravis (MG)

We studied antibody responses after immunization with ganglio-series gangliosides against 10 strains of inbred mice, including Balb/c, C57BL/6, A/J, C3H/HeN, C3H/HeJ, CBA/N, AKR/N, NZB/N, DBA/2 and nu/nu Balb/c. Twelve gangliosides having NeuAc as their sialic acid moiety (GM4, GM3, GM2, GM1, GD3, O-Ac-GD3, GD2, GD1a, GD1b, GT1a, GT1b and GQ1b), four gangliosides having NeuGc (GM3, GM2, GM1 and GD3) and four asialo-gangliosides (GA4, GA3, GA2 and GA1) were injected intravenously adsorbed to Salmonella minnesota. The antibody titers of the mice sera were determined by an enzyme-linked immunosorbent assay and an immune adherence assay. Antibody responses were found to depend not only on the ganglioside used as an immunogen but also on the mouse strain. Gangliosides having a trisaccharide sequence (NeuAcα2 → 8NeuAcα2 → 3Gal-) such as GD3, GD2, GDlb, GT1a and GQ1b, in particular O-Ac-GD3, induced high-titer antibody responses, whereas those having a disaccharide sequence (NeuAcα2 → 3Gal-) such as GM4, GM3, GM2, GM1, GDla and GTl1 induced low-liter antibody responses. On the other hand, gangliosides with NeuGc developed minimum titers. In contrast, asialo-gangliosides induced much higher responses than the corresponding gangliosides. The differences in ceramide portions of these gangliosides did not appear to be involved in inducing antibody responses. Mice could be divided into three groups according to the magnitude of their antibody responses: Group 1, those that produce the highest antibody responses (C3H/HeN and A/J); Group 2, those that demonstrate moderate antibody titers (Balb/c, C57BL/6, DBA/2 and nu/nu Balb/c); and Group 3, those that make minimum responses (AKR/N, C3H/HeJ, CBA/N and NZB/N). The pattern of reactivity to the various gangliosides was similar in all the strains tested.

Complement-opsonized IgG antibody/dsDNA immune complexes bind to CR1 clusters on isolated human erythrocytes

We used fluorescence microscopy, quantitative FACS analyses, and radioimmunoassays to examine the distribution of complement (C3b)-opsonized antibody/dsDNA immune complexes (IC) bound to normal human erythrocytes (RBCs) via immune adherence (IA). IC were detected with fluorescent anti-human IgG, and the RBC IA-receptor (CR1) was revealed with monoclonal antibodies (Mabs) to CR1 and fluorescent anti-mouse IgG. Under saturating conditions RBCs exhibit a large heterogeneity in binding; a significant fraction binds no IC. The positions of bound IC coincide with CR1 clusters on RBCs, as predicted by several investigators. FACS experiments indicate an excellent correlation between CR1 number and IC binding within an RBC population. The number of CR1 clusters able to bind IC is proportional to the number of CR1 per RBC. However, IC (fluorescent spots) detected per RBC (on average less than 10) are less than the average number of IC bound (20-30). This suggests that each fluorescent spot represents a small number of aggregated IC bound to a CR1 cluster. These "patches" of aggregated complexes may facilitate transfer of RBC-bound IC to cells of the mononuclear phagocytic system. Finally, not all the CR1 clusters bind IC, suggesting that proper geometric alignment of multiple C3b per IC with several CR1 in the cluster is required for IA.

Procoagulant activity and tumor necrosis factor in rat hepatic allograft rejection.

We used a model of rat hepatic allograft rejection to evaluate levels of procoagulant activity and tumor necrosis factor during acute cellular rejection. ACI livers were transplanted into Lewis rats, and Lewis-to-Lewis isografts and unoperated animals served as controls. Animals were killed on days 1, 2, 3, 4, 5, 6, 7 and 9. Splenic mononuclear cells were obtained by Ficoll-Hypaque gradients. Collagenase perfusion, metrizamide gradients and centrifugal elutriation were used to isolate Kupffer cells. Procoagulant activity assay of the splenic and Kupffer cells was done using a one-step clotting assay. Tumor necrosis factor was assayed using an L929 cytotoxicity assay. Histological evidence of acute rejection began on the 4th postoperative day, and rats died by the 9th or 10th postoperative day. Splenic procoagulant activity was significantly elevated in rejecting rats on day 4 and remained elevated until death. In contrast, Kupffer-cell procoagulant activity was elevated over controls by day 3 and remained significantly elevated until death. The tumor necrosis factor levels were elevated from day 1 and remained so until death. The data indicate that procoagulant activity is synthesized both by peripheral monocytes and locally by Kupffer cells and that procoagulant activity and tumor necrosis factor levels rise during hepatic allograft rejection. Because procoagulant activity and tumor necrosis factor mediate immune functions that are postulated to be important in acute rejection (immune cell adherence, vascular thrombosis and delayed-type hypersensitivity), these elevations may contribute to the pathogenesis of acute rejection.