CLL: promiscuity leads to risks

In this issue of Blood, Seiler and colleagues confirm that phage display is a feasible technology to identify specific ligands of antigen receptors expressed by CLL B cells. Importantly, they demonstrate that somatic hypermutation influences the range of specificities of the clonotypic BCRs. Furthermore, they raise the intriguing possibility that mimetic ligands to CLL BCRs might be of therapeutic use.

A strong argument supporting a role of antigen in chronic lymphocytic leukemia (CLL) was the finding that somatic mutations can be present in the immunoglobulin heavy variable (IGHV) genes expressed by the leukemic B cells in more than 50% of cases.¹  It is now established that the mutational status of the IGHV genes has important prognostic implications: CLL patients with somatically mutated IGHV genes (M-CLL) generally follow a more indolent course than those with unmutated IGHV genes (U-CLL). Initially, it was assumed that U-CLL derives from naive B cells; however, subsequent studies have challenged this view and convincingly demonstrated that all CLL cells, irrespective of IGHV gene mutational status, are antigen experienced.¹ 

A role for antigen in CLL development is additionally supported by the skewing of the IGHV gene repertoire in CLL²  and by the remarkable fact that almost 30% of CLL patients can be assigned to subsets with quasi-identical, “stereotyped” B-cell receptors (BCRs).³  BCR stereotypy in CLL strongly implies the recognition of structurally similar antigenic elements, likely selecting the leukemic clones. Preliminary evidence suggests that stereotypy may extend from similarities in the BCR to shared biological and clinical characteristics and, perhaps, outcome. Therefore, it becomes crucial to analyze the BCRs in terms of antigen specificity and affinity. However, up to this point, both the nature of the selecting antigens and the functional consequences of antigen recognition have remained largely unknown.

Previous attempts to isolate immunoglobulin from malignant CLL clones for further structural and functional studies have been hampered by several limitations. More recently, the Chiorazzi laboratory has systematically used recombinant DNA techniques to obtain monoclonal antibodies (mAbs) from CLL clones.^4,5  An alternative approach was followed by Rosen's group, who used cell lines derived from the neoplastic CLL clone by Epstein-Barr virus (EBV) transformation as well as primary ex vivo CLL cultures.⁶  Both approaches have yielded interesting results, highlighting restriction in terms of (auto)antigen specificity of the CLL mAbs and emphasizing the recognition of neo-antigens created during the apoptotic process (eg, by oxidation).^4-6 

Despite such progress, little is known about the chemical nature of important antigens in CLL. To address this issue and identify actual or mimetic ligands to CLL BCRs and also determine differences in epitope reactivity, Seiler et al adopted the phage display approach, which allows identification of mimetic ligands for both nonprotein structures as well as proteins.⁷  They probed phage-displayed peptide libraries with recombinant mAbs from 4 M-CLL and 4 U-CLL patients, 6 of which derived from cases belonging to major subsets with distinct stereotyped BCRs.

Mimetic ligands from M-CLL and U-CLL mAbs differed significantly. Whereas M-CLL mAbs recognized multiple peptides that shared well-defined amino acid motifs, U-CLL mAbs bound a series of peptides of quite dissimilar sequence. The specificity of the various mAb-phage interactions was confirmed by enzyme immunoassay. Interestingly, for M-CLL mAbs, even slight variations in key amino acid motifs of the isolated peptides were found to change or abolish binding, indicating that these peptides can be highly specific for individual BCRs. On these grounds, the authors propose that the similarity between peptides isolated by distinct M-CLL mAbs might be considered as an indication of a restriction in terms of the antigens that selected the corresponding CLL progenitors and even the malignant cells themselves.

Previous studies from the same group have indicated that all CLL cells likely derive from B cells producing polyreactive, natural antibodies encoded by germline immunoglobulin genes that either retain (U-CLL) or lose (M-CLL) polyreactivity due to somatic hypermutation.⁸  In accordance with this view, the range of mimetic epitopes selected by CLL mAbs generally correlated with the mutational status of the clonotypic immunoglobulin. From the mAb's side, this meant that U-CLL mAbs generally had a broader reactivity profile than M-CLL mAbs.

The final part of the study raises the intriguing possibility that mimetic ligands could alter leukemia cell survival and growth by inhibiting the binding of the clonotypic mAb to the “native” antigen, thus serving as a potential treatment for CLL. Alternatively, such ligands might also be used to target and eliminate CLL cells via the BCR, either by inducing apoptosis or by delivering cytotoxic agents.

The quest is on for antigens in CLL!