Immunoglobulin heavy variable (IGHV) gene replacement is a RAG-mediated recombination event that modifies a rearranged IGHV-IGHD-IGHJ sequence by replacing the original IGHV with another IGHV. As the excision of the original IGHV occurs at a cryptic heptamer near the end of FR3 of the original IGHV, a trace (“footprint”) remains embedded in the V-D (“N1”) junction of the heavy complementarity-determining region 3 (HCDR3) sequence of the new rearrangement. Since IGHV-replaced antibody sequences are over-represented in several autoantibody-associated syndromes, it is assumed that the process is designed to eliminate unwanted antibody specificities, particularly those which are autoreactive. Because CLL antibodies/B cell receptors (BCRs) are often autoreactive, we sought to determine if IGHV replacement plays a significant role in the formation of these antibodies. To this end, we analyzed IGHV-IGHD-IGHJ rearrangements in a CLL sample set of 6,121 sequences and in another set of 3,326 sequences from normal B lymphocytes culled from the GenBank Database. Sequences without positional information for their IGHV match were eliminated, resulting in 6023 CLL and 3202 normal sequences. From among these, we looked for the presence of IGHV footprints in the N1 and N2 junctional regions. Sequences were processed by IMGT to determine the particular IGHV used as well as junctional N1, N2 sequences. The resulting junctional regions were analyzed using the Java-based VH Replacement Footprint Analyzer (VHRFA) program [Huang, L et al (2013) PLOS ONE] to identify footprints (≥5-mer) in the N1 and N2 regions. Various CLL subgroups based on ZAP-70 expression, CD38 expression, IGHV mutation status, BCR stereotypy, and genomic aberrations were also examined for differences in the presence of footprints.

Comparison of CLL with normal sequences in the N1 regions revealed a slight yet significant difference in footprint frequency (12.62% versus 11.06%, respectively; P = 0.03). Comparison of IGHV-mutated (M) versus unmutated (U) subgroups of CLL also disclosed a significant difference in frequencies (M: 13.58%, U: 11.51%, P = 0.02). Using footprints in the N2 regions as an internal control, we found the frequency of ≥5-mer replacement footprint motifs to be significantly higher in N1 over N2 (P = 1.1E-7), supporting the validity of this approach. Other comparisons led to a significant enrichment of footprint frequencies, such as 13.71% for the ZAP-70-negative group (P = 0.029), 13.36% for the CD38-negative group (P = 0.013), and 14.70% for the 13q-deletion group (P = 0.016) compared with the normal group. Note, an overall comparison of stereotyped versus non-stereotyped cases revealed highly statistically significant differences in N1 footprint frequencies (9.4% for the 2084 cases having defined subsets vs 14.32% for the 3,939 cases without defined subsets; P = 2.7E-8). Focusing on stereotyped cases that exhibited evidence for IGHV replacement, we found significantly different footprint frequencies in the N1 region among different stereotyped subsets, even when restricting the comparison to subsets utilizing the same IGHV and belonging to the same mutational category (U-CLL or M-CLL). Relevant examples include differential frequencies among U-CLL, IGHV1-69-expressing subsets #3, #5, #6, #7, and #59 (0 - 15.5%; P = 0.01) and M-CLL, IGHV4-34-expressing subsets #4, #16, #29, and #201 (0 - 40.9%; P = 0.005).

In summary, CLL had an elevated IGHV replacement rate compared to normal controls. In addition, M-CLL, ZAP-70-neg, CD38-neg and 13q del CLL subgroups were also elevated above normal controls. The most dramatic differences in IGHV replacement frequency were found between CLL non-stereotyped and stereotyped CLL cases, with an intriguing difference within stereotyped subgroups exhibiting the same IGHV where IGHV replacement varied by as much as 40-fold. Because higher levels of IGHV replacement are found in ZAP-70-neg, CD38-neg, 13q del CLL, and non-stereotyped cases that in general are less aggressive, the presence of IGHV replacement may pinpoint CLL clones whose precursor B cells had edited their BCRs away from autoreactivity and therefore antigen-drive. This conclusion suggests a differential role for autoantigen drive in certain stereotyped subsets: CLL subsets not exhibiting IGHV replacement appear to be more susceptible to antigen drive than those that do.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.