The common variable immunodeficiency IgM repertoire narrowly recognizes erythrocyte and platelet glycans

BACKGROUND

Autoantibody-mediated cytopenias (AICs) regularly occur in profoundly IgG-deficient common variable immunodeficiency (CVID) patients. The isotypes, antigenic targets, and origin(s) of their disease-causing autoantibodies are unclear.

OBJECTIVE

To determine reactivity, clonality and provenance of AIC-associated IgM autoantibodies in CVID patients.

METHODS

We utilized glycan arrays, patient erythrocytes, and platelets to determine targets of CVID IgM autoantibodies. Glycan binding profiles were used to identify auto-reactive clones across B cell subsets, specifically circulating marginal zone-like (MZ) B cells, for sorting and IGH sequencing. The locations, transcriptomes and responses of tonsillar MZ B cells to different T helper cell subsets were determined by confocal microscopy, RNA-sequencing, and co-cultures, respectively.

RESULTS

Autoreactive IgM coated erythrocytes and platelets from many CVID patients with AICs (CVID+AIC). On glycan arrays, CVID+AIC plasma IgM narrowly recognized erythrocytic i antigens and platelet i-related antigens and failed to bind hundreds of pathogen- and tumor-associated carbohydrates. Polyclonal i antigen-recognizing B-cell receptors were highly enriched among CVID+AIC circulating marginal zone (MZ) B cells. Within tonsillar tissues, MZ B cells secreted copious IgM when activated by the combination of IL-10 and IL-21 or when cultured with IL10/IL-21 secreting FOXP3-CD25hiTfh cells. In lymph nodes from immunocompetent controls, MZ B cells, plentiful FOXP3+ regulatory T cells, and rare FOXP3-CD25+ cells that represented likely CD25hiTfh cells, all localized outside of GCs. In CVID+AIC lymph nodes, cellular positions were similar but CD25hiTfh cells greatly outnumbered regulatory cells.

CONCLUSIONS

Our findings indicate glycan-reactive IgM autoantibodies produced outside of GCs may contribute to the autoimmune pathogenesis of CVID.

Rapid cytotoxicity of human B lymphocytes induced by VH4-34 (VH4.21) gene-encoded monoclonal antibodies, II

We have previously described complement-independent killing of human B lymphocytes by two IgM MoAbs derived from the VH4-34 (VH4.21) gene. Analysis of 17 independently derived VH4-34-encoded MoAbs shows that B cell toxicity is not limited to the two described MoAbs, but is a general property shared by a subset of MoAbs derived from the VH4-34 gene. As observed by two independent microscopy techniques, giant membrane pores were formed on target B cells within 10-15 min of exposure to cytotoxic VH4-34-derived MoAbs. Toxicity by individual MoAb correlated directly to its B cell binding intensity measured by FACS, i.e. stronger the binding greater the killing. Sequence analysis showed that V(H) region in germ-line or in near germ-line configuration was necessary but not sufficient for B cell binding. In addition, a particular sequence motif enriched in basic amino acids in the CDR3 may be required to supplement the reactivity mediated by the V(H) region of the MoAb molecule. VH4-34-encoded antibodies that fulfil the above sequence requirements have cold agglutinin activity towards the i antigen of cord erythrocytes. In vivo, such anti-i/anti-B cell antibodies are rarely detected in healthy adults, but serum levels are dramatically elevated in selective pathological conditions, such as systemic lupus erythematosus and infectious mononucleosis. This strict regulation may be related to the novel and rapid mechanism of human B cell toxicity demonstrated by antibodies encoded by a single human V(H) gene.

Rapid cytotoxicity of human B lymphocytes induced by VH4-34 (VH4.21) gene-encoded monoclonal antibodies

We have previously described two human cold agglutinin MoAbs 216 and A6(H4C5), that are derived from the VH4-34 (VH4.21) gene that bind specifically to a cell surface ligand on human B lymphocytes. In this study, we report that binding of 216 and A6(H4C5) leads to rapid killing of target B cells. This complement-independent cytotoxicity was measured by three independent assays, cell viability dye uptake on FACS, 3H-thymidine uptake, and the 3(4,5)-dimethylthiazol-2,5-diphenyl tetrazolium bromide (MTT) assay. Cytotoxicity was specific for CD20+ mononuclear cells in human spleen and peripheral blood. The MoAbs were also cytotoxic to human B cell lines Nalm-6, OCI-LY8, Arent and SUP-B8, but not to T cell lines HuT 78 and PEER. As observed by scanning electron microscopy, membrane pores were formed within 15 min of exposure to the MoAbs. Cytotoxic activity was dependent on MoAb concentration and temperature of exposure. Killing with greater at 4 degrees C than 37 degrees C. Sodium azide and EDTA did not block the cytotoxic activity. No DNA fragmentation typical of apoptosis was observed. This rapid cytotoxic activity, independent of physiologic cellular process and independent of complement, suggests a novel mechanism of all death via membrane perturbations.

Deficiency of beta 1-6 N-acetylglucosaminyltransferase involved in the biosynthesis of blood group I antigen in the liver of LEC rats

The activities of the beta 1-6 and beta 1-3 N-acetylglucosaminyltransferase, which synthesize blood group I and i antigens, respectively, were measured in various tissues of hepatitis- and hepatoma-predisposed rats (LEC rats). In LEC rats the beta 1-6 N-acetylglucosaminyltransferase activity was barely detectable in the liver, while substantial enzyme activity was found in other tissues. In the control LEA rats the enzyme was expressed in most tissues, including the liver. Immunochemical studies using a monoclonal antibody which recognizes I antigen indicated that the expression of I antigen was less prominent in hepatocytes of LEC rats than in hepatocytes of LEA rats. The level of beta 1-3 N-acetylglucosaminyltransferase activity was constant in most of the tissues during the development. These results indicate that the biosynthesis of I antigen does not occur in the livers of the LEC rats.