B-cell chronic lymphocytic leukemia (CLL) clones from different patients can express stereotyped BCR immunoglobulins (IGs) that share very similar IGHV-D-J rearrangements and HCDR3s, often with similar IGLV gene usage, suggesting that they were selected by common antigenic epitope(s). Patients bearing IGHV4-34/D5-18/JH6/VK2-30/JK2 rearrangements belong to one of such stereotyped subsets, subset 4. In addition, subset 4 BCRs are always IgG switched and exhibit IGHVsomatic mutations.

The germline IGHV4-34 gene is inherently autoreactive and is often found in antibodies that bind to the surfaces of erythrocytes and B lymphocytes through the linear i blood group antigen, poly-N-acetyllactosamine (NAL). To test the reactivity of subset 4 monoclonal antibodies (mAbs) to auto and foreign antigens, we expressed a panel of CLL BCR IGs carrying IGHV4-34, 1-69, 3-21, and 4-39 as recombinant human IgGs and tested the binding of these CLL mAbs to HEp-2 cells by intracellular immunostaining. Three subset 4 mAbs (183, 240 and 342) showed only weak binding to permeabilized HEp-2 cells compared to other CLL mAbs which often showed strong binding to cytoplasmic structures. Furthermore, because apoptotic cells are a source of autoantigens in autoimmune conditions, the reactivity of CLL mAbs to non-permeabilized, irradiated Ramos B and Jurkat T cell line cells was also tested by flow cytometry. Surprisingly, the average binding percentage to apoptotic Ramos B cells of subset 4 mAbs 183, 240, and 342 was only 6, 5, and 4%, respectively. Moreover, all three subset 4 mAbs exhibited minimum binding to live or apoptotic Jurkat T cells (1-10% and 4-6% to each population, respectively). We next tested the reactivity of subset 4 mAbs to i/I carbohydrate epitopes, the known targets of many non-stereotyped IGHV4-34 mAbs, using an i/I focused microarray containing 60 i/I related oligosaccharides. Notably, the subset 4 mAbs showed no binding at a concentration of 10 μg/mL. Together, the above results suggest that CLL subset 4 mAbs have weak or no interaction to the known autoantigens recognized by other non-stereotyped IGVH4-34 IGs. Hence we hypothesized that the restricted IGHV-D-Jrearrangement, somatic mutations, and IgG switch of subset 4 were acquired by selection for other auto or foreign antigens.

To test this possibility, we screened CLL mAbs against a library of over 8,000 full length human proteins, including an influenza A virus and some human viral gene encoded proteins. Each subset 4 mAb showed selective and strong binding to influenza A virus H3N2 Texas 1/77 at concentration of 50 μg/mL; this binding was 6-15 times higher than that of other non-subset 4 mAbs, which included two IGHV4-34, non-stereotyped CLL mAbs. Because the binding to virus is usually mediated via its glycoprotein hemagglutinin (HA), we tested reactivity of subset 4 mAbs to the HA of influenza A H3N2 Texas 1/77. Subset 4 mAbs had on average 2 fold (1.8-2.2) higher binding signals than the two non-stereotyped IGHV4-34mAbs. To explore if subset 4 mAbs recognize other influenza A virus strains, we tested the binding of a representative subset 4 mAb 183 to HAs from 11 strains that belong to 3 subtypes: H1 (4 strains), H2 (1 strain) and H3 (6 strains). Clear preferential binding to all 6 H3s and to some of H1s was found. mAb 183 did not bind to H2 (A/Japan/305/1957 H2N2). The average binding signal ratio to H3s, H1s and H2 was 2.4:1.6:1. Testing to another hemagglutinin subtype H7 (A/Anhui/1/13 H7N9) showed that mAb 183 did not interact with H7.

Intrigued by the observation that CLL subset 4 BCR IGs are always the IgG isotype, we tested if isotype switching contributed to antigen recognition by expressing CLL 183 IGHV-D-Jas divalent human IgM (mAb183M) and checked its reactivity to virus A H3N2 Texas 1/77. Switching isotype from IgG to IgM led to a 48% loss in reactivity to intact virus and 74% loss to hemagglutinin H3 (A/Philippines/2/1982 H3N2).

Taken together, our results indicate that unlike other IGHV4-34 bearing IGs, CLL subset 4 mAbs have only limited reactivity to autoantigens, including linear NAL epitopes. Instead, this subset has acquired specific recognition of subtypes of an infectious agent, influenza A (H3 and H1), apparently via its envelope protein hemagglutinin. The shift in antigen recognition appears to be dependent upon both the stereotyped IGVH-D-Jrearrangement and isotype switching to IgG.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution