Decreased phagocytosis of apoptotic cells in diseased SLE mice.

Antibodies against nucleosomes are a serological hallmark of systemic lupus erythematosus (SLE). Apoptotic cells are the unique source of nucleosomes, which are formed through cleavage of chromatin by nucleases. These nucleosomes and other autoantigens targeted in SLE are expressed in apoptotic blebs or at the surface of apoptotic cells. Therefore, it is conceivable that circulating antibodies can influence apoptotic cell clearance. Using an in vitro phagocytosis assay, we analysed the phagocytic efficacy for apoptotic cells of resident peritoneal macrophages from pre-morbid and diseased lupus mice. The assay was carried out in the presence of autologous serum, using autologous apoptotic thymocytes as targets. Under these conditions macrophages from diseased MRL/lpr and NZBxNZW(F1) lupus mice, and from age-matched NZB mice showed a decreased phagocytic efficacy (decrease 47%, 48% and 37%, respectively compared to measurements in pre-morbid mice). The cause of this decrease resides in the serum, and is not due to an acquired defect of macrophages. In conclusion, during disease progression in murine SLE, apoptotic cell clearance becomes impaired, which might amplify further disease progression.

No constitutive defect in phagocytosis of apoptotic cells by resident peritoneal macrophages from pre-morbid lupus mice.

Antibodies against nucleosomes are a hallmark of systemic lupus erythematosus (SLE). Nucleosomes are uniquely formed during apoptosis, through cleavage of chromatin by nucleases. Increased exposure of nucleosomes to the immune system could play a role in the induction of the autoimmune repertoire in SLE. To determine whether there exists a constitutive defect in the clearance of apoptotic cells, resident peritoneal macrophages from pre-morbid SLE-prone MRL and New Zealand (NZ) mice were analysed for their efficacy to phagocytose apoptotic cells in vitro. Although differences in phagocytic efficacy of up to 50% between different strains of mice were found, these were not related to SLE development. To evaluate whether macrophages from SLE-prone mice are more susceptible to phagocytic 'exhaustion', resident peritoneal macrophages were challenged by 20 h of additional culture in the presence of apoptotic cells. In both lupus and control strains this led to an increased capacity to phagocytose fresh apoptotic cells (increase between 15 and 92%). As a control, macrophages from all strains were also exposed to 20 h of additional culture without apoptotic cells. Under this condition resident peritoneal macrophages from all SLE-prone strains, and of the SLE-parental strain NZB, displayed a significant decrease in their efficacy to phagocytose apoptotic cells (decrease between 16 and 55%). Together, these findings do not support the hypothesis that a constitutive defect in the clearance of apoptotic cells, as evaluated by testing resident peritoneal macrophages, plays an important role in the induction of SLE.

A combination of anti-CD3 and anti-CD7 ricin A-immunotoxins for the in vivo treatment of acute graft versus host disease.

This study evaluated the anti-graft versus host disease (GVHD) potential of a combination of immunotoxins (IT), consisting of a murine CD3 (SPV-T3a) and CD7 (WT1) monoclonal antibody both conjugated to deglycosylated ricin A. In vitro efficacy data demonstrated that these IT act synergistically, resulting in an approximately 99% elimination of activated T cells at 10(-8 )mol/L (about 1.8 microg/mL). Because most natural killer (NK) cells are CD7(+), NK activity was inhibited as well. Apart from the killing mediated by ricin A, binding of SPV-T3a by itself impaired in vitro cytotoxic T-cell cytotoxicity. Flow cytometric analysis revealed that this was due to both modulation of the CD3/T-cell receptor complex and activation-induced cell death. These results warranted evaluation of the IT combination in patients with refractory acute GVHD in an ongoing pilot study. So far, 4 patients have been treated with 3 to 4 infusions of 2 or 4 mg/m(2) IT combination, administered intravenously at 48-hour intervals. The T(1/2) was 6.7 hours, and peak serum levels ranged from 258 to 3210 ng/mL. Drug-associated side effects were restricted to limited edema, fever, and a modest rise of creatine kinase levels. One patient developed low-titer antibodies against ricin A. Infusions were associated with an immediate drop of circulating T cells, followed by a more gradual but continuing elimination of T/NK cells. One patient mounted an extensive CD8 T-cell response directly after treatment, not accompanied with aggravating GVHD. Two patients showed nearly complete remission of GVHD, despite unresponsiveness to the extensive pretreatment. These findings justify further investigation of the IT combination for treatment of diseases mediated by T cells. (Blood. 2000;95:3693-3701)

Role of nucleosomes for induction and glomerular binding of autoantibodies in lupus nephritis.

The cardinal feature of systemic lupus erythematosus is the formation of anti-nuclear antibodies. In recent years, it has become clear that the nucleosome is a major autoantigen that drives this T cell-dependent autoimmune response, as exemplified by the presence of nucleosome-specific T helper cells and the high prevalence of nucleosome-specific autoantibodies. The only way to generate nucleosomes in vivo is by the process of apoptosis. There is growing evidence that in systemic lupus erythematosus apoptosis is disturbed, leading to the release of nucleosomes. Moreover, apoptosis-induced modifications of these autoantigens may render them more immunogenic, especially if the removal of apoptotic cells is insufficient. The first indications for the impaired clearance of apoptotic cells in systemic lupus erythematosus are emerging. Nucleosomes are also important for mediating tissue lesions, especially glomerulonephritis. In lupus nephritis nucleosomes, nucleosome-specific antibodies and nucleosome/IgG complexes have been identified in the glomerular immune deposits. Via their cationic histone part nucleosomes mediate the binding of anti-nuclear antibodies to intrinsic constituents of the glomerular basement membrane, such as the anionic heparan sulfate and collagen IV. Appreciation of this binding mechanism may lead to new treatment strategies, as shown for non-coagulant heparinoids.

An assay for the quantitative measurement of in vitro phagocytosis of early apoptotic thymocytes by murine resident peritoneal macrophages.

Research into the mechanisms by which apoptotic cells are phagocytosed has grown considerably over recent years, together with a growing appreciation of the importance of clearance of redundant cells for tissue homeostasis. However, studies addressing the efficacy of phagocytosis have been rare. The few studies reported to date were either attempts to determine apoptotic cell clearance from the circulation or were focused on clearance in inflammation. We now describe an in vitro assay which permits the quantitative measurement of phagocytosis of apoptotic cells by murine resident peritoneal macrophages. The apoptotic cells used in the assay were murine thymocytes incubated with dexamethasone for only 3 h. Most apoptotic thymocytes were annexin V positive and propidium iodide negative and therefore still in the earlier stages of apoptosis. The assay was completed 7 h after the isolation of both macrophages and thymocytes, while macrophage culture time was only 4 h. Because of this short-term culture it is likely that the resident peritoneal macrophages largely maintained their in vivo phenotype. Using BALB/c macrophages and thymocytes, the maximal in vitro phagocytosis exceeded five thymocytes per macrophage in 1 h and two of these thymocytes were taken up within 10 min. Therefore, in vitro phagocytosis by resident peritoneal macrophages was rapid and of high capacity, as it is postulated to be in vivo. Under selected conditions, the mean uptake was 4.45+/-0.70 (mean +/- SD, n = 31) thymocytes per macrophage in 1 h. The inter-assay coefficient of variation, also representing the biological variability, was found to be 15.7%. The average intra-assay coefficient of variation was 13.6%. This assay permits comparisons of phagocytic efficacy between different strains of mice in vitro. In addition, a method of preparation is described which allows long-term storage of experimental results. Finally, our data suggests that internalization, but not binding of apoptotic cells to short-term cultured resident peritoneal macrophages, is critically dependent on the presence of serum. This allows separate analysis of binding and internalization of apoptotic cells with the assay, without the necessity to use agents blocking internalization.

Different CD3/T cell receptor monoclonal antibodies have distinct capacities to induce adhesion of T lymphocytes to endothelium.

Murine CD3/T cell receptor (TCR) monoclonal antibodies (mAbs) induce immediate peripheral lymphocytopenias of different degree and duration. Lymphocytopenia is of short duration after the administration of immunoglobulin A CD3 mAb, but it persists much longer after the administration of immunoglobulin G2a CD3 mAb. Peripheral lymphocytopenia after the administration of WT31, a murine immunoglobulin G1 TCR mAb, appears to be dependent on the polymorphism of Fc(gamma)RIIa. In high responders, lymphocytopenia is comparable to that observed after immunoglobulin G2a CD3 mAb; in low responders, no lymphocytopenia occurs. In vitro, both immunoglobulin A and immunoglobulin G2a CD3 mAbs induce immediate activation of CD11a/CD18, with concomitant up-regulation of CD11b/CD18 on T cells, each of which is shown to be involved in the concurrent adhesion of T cells to endothelium. WT31 induces an immediate activation of CD11a/CD18 as well as T cell adhesion to endothelium in Fc(gamma)RIIa high responders only, interestingly without changes in the level of expression of CD11b/CD18. We conclude that the immediate occurrence of peripheral lymphocytopenia after the administration of CD3/TCR mAb is mediated by changes in the level of expression or avidity (or both) of adhesion molecules on T cells, whereas the persistence of this lymphocytopenia depends on the isotype of the CD3/TCR mAb and on the presence of suitable Fc receptors.

Role of polymorphic Fc receptor Fc gammaRIIa in cytokine release and adverse effects of murine IgG1 anti-CD3/T cell receptor antibody (WT31).

Anti-CD3 monoclonal antibody (mAb) OKT3 is immunosuppressive, but causes severe adverse effects during the first administration ("first-dose reaction"). These adverse effects are presumably caused by cytokine release that results from T-cell activation. In vitro, T-cell activation by anti-CD3 mAb requires interaction with monocyte Fc receptors. The Fc receptor for murine IgG1, Fc gammaRIIa, is polymorphic. In some individuals, murine IgG1 anti-CD3 mAb causes T-cell proliferation and cytokine release in vitro (high responders [HR]), whereas in individuals with the low-responder (LR) phenotype it does not. We have now investigated the role of this Fc gammaRIIa polymorphism in the release of cytokines in vivo and the occurrence of adverse effects after the administration of WT31, a murine IgG1 anti-CD3/T cell receptor mAb. WT31 caused an increase of plasma tumor necrosis factor-alpha in all four HR patients and none of the five LR patients. In all HR patients except one, plasma gamma-interferon and interleukin 6 also increased, and a first-dose response was observed, whereas no cytokine release or adverse effects occurred in any of the LR patients. WT31 caused lymphopenia in all HR and none of the LR patients. FACS analysis demonstrated that in HR patients, after the initial disappearance of CD3+ cells from peripheral blood, modulation of CD3 occurred, whereas in LR patients a high degree of coating of the lymphocytes was observed. Surprisingly, WT31 also induced a marked granulocytopenia, as well as a decrease of thrombocytes, in three of the four HR patients (and in none of the LR patients). These data provide direct clinical evidence that Fc receptor interaction determines the release of cytokines and the occurrence of adverse effects after administration of anti-CD3/T cell receptor mAb. Furthermore, these data suggest that tumor necrosis factor-alpha by itself is not sufficient to induce the first-dose reaction.

Analysis at the DNA level of human Fc gamma RII isoforms in relation to the polymorphic binding of murine IgG2b to human B cells.

Three classes of human leucocyte Fc gamma receptors (hFc gamma R) have been identified so far: hFc gamma RI, hFc gamma RII, and hFc gamma RIII. Previous studies have demonstrated that genetically determined differences between individuals exist both with respect to the binding of murine IgG1 (mIgG1) to hFc gamma receptors, and with respect to the binding of murine IgG2b (mIgG2b). The polymorphism in binding of mIgG1 could be ascribed to hFc gamma RIIA, an isoform of hFc gamma RII. The authors have now investigated whether one of the isoforms of hFc gamma RII is also responsible for the polymorphism in binding of mIgG2b. In these studies the authors used EBV-transformed human B cells that demonstrated either binding or no binding of mIgG2b in EA-rosetting assays. mRNA obtained from these cells was amplified by reverse transcriptase and polymerase chain reaction (RT-PCR). Hybridization experiments with the RT-PCR products revealed that the hFc gamma RIIB but not the hFc gamma RIIA isoform was present in these cells. DNA sequencing further demonstrated that the nucleotide sequence of both the extracellular part and the cytoplasmic moiety of hFc gamma RIIB was identical for all individuals tested, regardless of their ability to bind mIgG2b. These findings indicate that the polymorphic binding of mIgG2b cannot be ascribed to one of the isoforms of hFc gamma RII. Since hFc gamma RI and hFc gamma RIII are not present on the cell surface of these cells, the authors conclude that an Fc receptor different from the known hFc gamma receptors must be responsible for the polymorphic binding of mIgG2b. These data further expand the complexity of hFc gamma R.

Increasing cytotoxic activity and production of reactive oxygen and nitrogen intermediates by peritoneal macrophages during the development of multiple organ dysfunction syndrome in mice.

A major problem in the intensive care unit nowadays is the development of multiple organ dysfunction syndrome (MODS), a cumulative sequence of progressive deterioration of organ functions. While the pathogenic pathways of MODS remain to be elucidated, it is assumed that cells of the host defence system, especially the macrophages, are altered in their function. During the development of MODS it is assumed that macrophages are overactivated and that an exaggerated inflammatory response may contribute to its pathogenesis. In order to gain insight into the alterations of the functional status of the macrophage during the development of MODS, a series of macrophage functions was measured in the subsequent phases of zymosan induced generalized inflammation in mice. Male C57BL/6 mice received a single dose of zymosan intraperitoneally and groups of animals were killed after 2, 5, 8, and 12 days. Peritoneal macrophages were collected for in vitro assessment of the ADCC, the production of superoxide (O2-) and nitric oxide (NO), and complement mediated phagocytosis and intracellular killing of Staphylococcus aureus. A single intraperitoneal injection with zymosan resulted in a three-phase illness. During the third phase the animals developed MODS-like symptoms. Peritoneal cells from control animals produced very low to non-detectable amounts of O2- and NO, and the cytotoxic activity was also low. During the development of MODS, from day 7 onwards, the ability to produce O2- and NO2- became strongly elevated, as did the cytotoxic activity. These findings are in parallel with the development of MODS whereas the phagocytic and killing capacity remained essentially unaltered. The changes found could be detrimental for the organism, thus possibly contributing to the onset and development of MODS.

In vivo ANA is a fixation artifact: nucleosome-complexed antinucleosome autoantibodies bind to the cell surface and are internalized.

It has been suggested that binding of anti-double-standed DNA antibodies to cell surfaces, followed by internalization and nuclear binding (so called in vivo ANA) is of pathophysiological significance for tissue damage in systemic lupus erythematosus. We have shown before that pathogenic antinuclear antibodies complexed to nucleosomal antigens can bind to heparan sulfate in the glomerular basement membrane in vivo. Because nucleosomes are also reported to bind to the cell surface, we hypothesized that in vivo ANA is a property of antinuclear antibodies bound to nucleosomal antigens. Therefore, we studied three antinucleosome monoclonal antibodies (mAb) that exhibit in vivo ANA as seen by immunofluorescence in mice inoculated intraperitoneally with the hybridoma producing the mAb. The same mAb complexed to nucleosomal antigens after intravenous injection into mice induced in vivo ANA, in contrast to purified noncomplexed mAb. To study this in more detail, we incubated complexed mAb with various cell lines and found binding to the cell surface and subsequent internalization into cytoplasmic vesicles. However, no binding to the nucleus was observed by immunoelectron microscopy (IEM) and confocal laser microscopy. Noncomplexed mAb did not bind to the cell surface. Next, from mice bearing the hybridomas producing the mAb intraperitoneally, a small part of the kidney was snap frozen in liquid N2, fixed with acetone, and studied in immunofluorescence, whereas the remaining part of the kidney was fixed in vivo by renal perfusion with a mixture of 0.01 M sodium periodate, 0.075 M lysine HCl, 0.0375 M Na2HPO4, and 2% paraformaldehyde (PLP) and studied in both immunofluorescence and IEM. In the acetone-fixed kidney sections obtained without in vivo fixation we again observed in vivo ANA. However, after in vivo PLP perfusion fixation, no nuclear binding was found. In IEM, localization in cytoplasmic vesicles was seen. In conclusion, antinucleosome antibodies complexed to nucleosomal antigens can bind to the cell surface and are transported into the cytoplasm, but do not bind to the nucleus. The reported nuclear localization of antinuclear antibodies is caused by a fixation artifact.

Ling Zhi-8: studies of a new immunomodulating agent.

Ling Zhi-8 (LZ-8) is a protein derived from the fungus Ganoderma lucidum and has immunomodulatory capacities. It was shown to be mitogenic toward mouse splenocytes in vitro and immunosuppressive in vivo by reducing antigen-induced antibody formation and by preventing completely the incidence of autoimmune diabetes in nonobese diabetic mice. In this study, the mitogenic effects of LZ-8 on human mononuclear cells are reported. In accordance to its mitogenic effect on mouse splenocytes, LZ-8 proved to be mitogenic for human PBMC. This mitogenic effect of LZ-8 apparently required the presence of monocytes. We also demonstrated it to be immunosuppressive in vitro in a human MLC performed in the absence of monocytes, using purified T cells and EBV-transformed allogeneic B cells. Furthermore, we tested LZ-8 for its possible suppressive effects in 2 different models of allogeneic tissue transplantation. LZ-8 proved to have a significant effect on cellular immunity, since its administration in an allografted mouse skin model resulted in an increased survival time. In a model of transplanted allogeneic pancreatic rat islets, LZ-8 was effective in delaying the rejection process of allografted islets. More frequent or continuous administration resulted in a further prolongation of survival time. No serious side effects of LZ-8 could be discerned in these experiments.

The extent of peritubular CD14 staining in renal allografts as an independent immunohistological marker for acute rejection.

Previously, we demonstrated that in acute interstitial rejection, immunohistological staining of renal allograft biopsies with the CD14 mAb WT14, reacting with human monocytes/macrophages, shows a characteristic peritubular increase of positive cells. To test the diagnostic value of this CD14 positivity, we compared, in 154 unselected renal allograft biopsies, the extent of peritubular WT14 staining with (a) the original histological diagnosis, made with knowledge of clinical data, (b) the retrospectively and blindly scored histological diagnosis according to the criteria of the Banff classification, and (c) the eventual clinical diagnosis, which included evaluation of the response to therapy. The extent of peritubular WT14 positivity, blindly scored on cryostat sections of the frozen part of the biopsies, correlated positively with the probability of acute rejection (AR). When using a cutoff of 70% WT14 positivity for the diagnosis of AR, as extracted from a receiver operating characteristic curve, the WT14 diagnosis had a positive predictive value of 91% and a negative predictive value of 56%, compared with the original histological diagnosis. Compared with the Banff diagnosis of AR (grade I-III), these values were 95% and 47%, and compared with the clinical diagnosis, 84% and 63%, respectively. The WT14 diagnosis essentially corrected the original histological diagnosis in 7 cases, and was consistent with the eventual diagnosis in 5 equivocal cases. We conclude that the extent of peritubular CD14 positivity can be used as a marker for AR and can serve as a valuable additional criterion for AR in the histological examination of renal allograft biopsies.

Endothelial cells promote anti-CD3-induced T-cell proliferation via cell-cell contact mediated by LFA-1 and CD2.

T-cell activation requires not only T-cell receptor (TCR) engagement and subsequent TCR/CD3 cross-linking, but also one or more secondary activation signals generated by accessory cells (AC). We investigated the accessory function of endothelial cells (EC) in an in vitro model for T-cell activation where the first cross-linking signal was delivered to peripheral human T lymphocytes by either immobilized anti-CD3 monoclonal antibody (MoAb) or by PHA. In a previous report, we showed that EC provided a potent costimulatory signal in this model system. We have now analysed the nature of the EC-derived costimulatory signal by testing whether EC could be substituted by cytokines, by studying the effect of EC fixation and by testing the involvement of a number of adhesion molecules. Our findings indicate that EC accessory function is mediated mainly by membrane-bound factors. The nature of these membrane-bound factors was analysed by studying the inhibitory properties of a series of MoAbs directed against several adhesion molecules. Antibodies directed against CD44, E-selectin, CD31, CD26, B7/BB1, VLA-4 or VCAM-1 were not inhibitory. However, an inhibition, was clearly observed with antibodies against LFA-1 and CD2. Remarkably, this inhibition was not found with MoAbs to their respective counterstructures ICAM-1 and LFA-3. In summary, we postulate that both LFA-1/ICAM-1, and CD2/LFA-3 interactions are involved in EC accessory function, although the role of the EC-associated adhesion partners is not fully understood.

Proteolysis increases the Fc-mediated binding of murine IgG2b to human EBV-transformed B cells, but decreases the expression of Fc gamma RII and Fc epsilon RII.

We have previously described a polymorphic human Fc receptor for murine IgG2b (mIgG2b). This receptor was defined by the binding of (complexed) mIgG2b to monocytes and Epstein-Barr virus (EBV)-transformed B cells. Three per cent of normal individuals were high responders with respect to mIgG2b (mIgG2b-HR), whereas the other individuals were low responders for mIgG2b (mIgG2b-LR). In the present study we investigated the effect of proteolytic enzymes on the Fc-mediated binding of mIgG2b to EBV-B cells. Pronase, human leucocyte elastase and cathepsin G caused an increased binding (in an EA-rosetting assay) of mIgG2b to EBV-B cells from mIgG2b-HR, but not from mIgG2b-LR. Simultaneous immunofluorescence studies demonstrated that these proteolytic enzymes strongly reduced the expression of Fc gamma RII and Fc epsilon RII on these cells, whereas HLA class I or HLA class II molecules were not affected. These findings strongly suggest that binding of mIgG2b is not mediated by Fc gamma RII or Fc epsilon RII. We also studied the effect of proteolysis on mIgG2b-HR EBV-B cells from an HLA class II-negative individual. In this case EA-mIgG2b rosetting was decreased after proteolysis, suggesting that HLA class II molecules may have a role in protecting the binding site for mIgG2b against proteolytic destruction.