Post-Transformation IGHV-IGHD-IGHJ Mutations in Chronic Lymphocytic Leukemia B Cells: Implications for Mutational Mechanisms and Impact on Clinical Course

Abstract

Analyses of IGHV gene mutations in chronic lymphocytic leukemia (CLL) have had a major impact on the prognostication and treatment of this disease. A hallmark of IGHV-mutation status is that it very rarely changes clonally over time. Nevertheless, targeted and deep DNA sequencing of IGHV-IGHD-IGHJ regions has revealed intraclonal heterogeneity. We used a DNA sequencing approach that achieves considerable depth and minimizes artefacts and amplification bias to identify IGHV-IGHD-IGHJ subclones in patients with prolonged temporal follow-up. Our findings extend previous studies, revealing intraclonal IGHV-IGHD-IGHJ diversification in almost all CLL clones. Also, they indicate that some subclones with additional IGHV-IGHD-IGHJ mutations can become a large fraction of the leukemic burden, reaching numerical criteria for monoclonal B-cell lymphocytosis. Notably, the occurrence and complexity of post-transformation IGHV-IGHD-IGHJ heterogeneity and the expansion of diversified subclones are similar among U-CLL and M-CLL patients. The molecular characteristics of the mutations present in the parental, clinically dominant CLL clone (CDC) differed from those developing post-transformation (post-CDC). Post-CDC mutations exhibit significantly lower fractions of mutations bearing signatures of activation induced deaminase (AID) and of error-prone repair by Polη, and most of the mutations were not ascribable to those enzymes. Additionally, post-CDC mutations displayed a lower percentage of nucleotide transitions compared with transversions that was also not like the action of AID. Finally, the post-CDC mutations led to significantly lower ratios of replacement to silent mutations in VH CDRs and higher ratios in VH FRs, distributions different from mutations found in normal B-cell subsets undergoing an AID-mediated process. Based on these findings, we propose that post-transformation mutations in CLL cells either reflect a dysfunctional standard somatic mutational process or point to the action of another mutational process not previously associated with IG V gene loci. If the former option is the case, post-CDC mutations could lead to a lesser dependence on antigen dependent BCR signaling and potentially a greater influence of off-target, non-IG genomic mutations. Alternatively, the latter activity could add a new stimulatory survival/growth advantage mediated by the BCR through structurally altered FRs, such as that occurring by superantigen binding and stimulation.

Keywords: activation-induced deaminase; chronic lymphocytic leukemia; immunoglobulin genes; mutation mechanisms; somatic mutations.

Figure 1

Complexity of intraclonal IGHV-D-J variants. (A) Examples of the degree of intraclonal diversity in CLL samples based on NGS sequencing data. The phylogenetic trees were inferred with the R package Alakazam and plotted with igraph (see Methods). The size of the CDC (light blue circles at top of each example) is proportional to its abundance within the entire leukemic clonal family. IGHV-D-J complexity is defined and represented by the number of distinct sequences downstream of the CDC. Each downstream sequence has all the mutations of the upstream sequence plus at least one additional. Low complexity contains one to three mutation-defined sequences downstream of the initial branch from the CDC. High complexity represents four or more unique sequences downstream of each initial branch, some involving intricate branching. (B) Pie graphs indicating the representation of U-CLL and M-CLL cases in the Low and High complexity categories.

Figure 2

Intraclonal IGHV-D-J diversity is relatively common in CLL. (A) Fraction of the intraclonal IGHV-D-J variant sequences among the total number of mRNA molecules coding the CLL clonal family in the 61 patients examined. (B) Frequency of PSCs among the total number of mRNA molecules encoding the CLL clonal family. (C) Fraction of the subclonal IGHV-D-J variants that the PSCs represent after excluding the CDC sequences.

Figure 3

Subclonal variant cell counts. (A) PSCs are numerically expanded. Box plots indicate the percentage distribution of all post-CDC subclones. The solid colored dots represent the percentages of individual PSCs. The dotted horizontal line indicates the 0.008% sensitivity threshold used to define subclonal expansion (see Methods). (B) Absolute numbers of distinct IGHV-D-J subclones per ul of individual patient blood samples estimated from sequencing data, flow cytometry, and absolute blood lymphocyte counts. Box plots and solid colored dots as in (A). (C) Representative examples of the relative numbers of the post-CDC subclones (blue circles) and the PSCs (red circles) in relation to the CDC of five samples analyzed in this study.

Figure 4

Relative targeting of mutational mechanisms. (A–D) Percentage of samples exhibiting significant targeting to AID hot spots in subclones, based on IGHV-mutation status (A), IGHV-mutation status + level of IGHV-D-J complexity for all CLL samples (B) or for only M-CLL (C) or only U-CLL cases (D). (E–H) Percentage of samples exhibiting significant features of Polη repair in subclones, based on IGHV-mutation status (E), IGHV-mutation status + level of IGHV-D-J complexity for all CLL samples (F) or for only M-CLL (G) or only U-CLL cases (H). (I) Fraction of post-CDC mutations attributable to the actions of AID at hot spots (red), AID at cold spots (green), of Polη repair (blue), or other/unclear (purple). For each motif type (AID hotspot, AID coldspot, Polη hotspot, all others), we calculated the fraction of mutations targeting the motif. Within each sample, this was done separately for the CDC, post-CDC, and non-productive sequences, giving a pair of fractions per sample. After calculating pairs for each sample, then CDC, post-CDC and non-productive sequences were compared across all samples using a paired t-test. Table to the right indicates the mean fractional differences of mutations between the CDC and the post-CDC settings in regard to mutations targeting the three motifs and not attributable to the any of the three.

Figure 5

Transition to transversion ratios in the post-CDC mutations. Transition to transversion ratios (Ts : TV) analyzed for: (A) all post-CDC mutations, (B) all post-CDC mutations based on levels of complexity, (C) only those post-CDC mutations involving AID hotspots, (D) post-CDC mutations involving AID hotspots based on levels of complexity and (E) non-productive IGHV-IGHD-IGHJ rearrangements (**P ≤0.01, ***P ≤0.001, ****P ≤0.0001). NS, Not statistically significant.

Figure 6

Selection strength on mutations in VH CDRs and the VH FWRs. The BASELINe program was used to calculate Selection Strength (Σ; see Methods) starting from observed and expected replacement to silent mutation ratios (R:S). Selection strength of mutations found in CDC and post-CDC productive and non-productive rearrangements based on complexity (post-CDC^High and post-CDC^Low) as well as those of a series of normal human mature B cell subsets (IgM memory cells, mMem; isotype-class switched memory cells, sMem; marginal zone B cells (MZ); and IgD^-CD27⁻ double negative B cells (DN). Σ indicates selection strength favoring selection for (right of 0) or against (left of 0) replacement mutations. Table to the right indicates the statistical analyses for selected differences between comparisons in the VH CDRs and VH FRs involving mutations detected in CDC and post-CDC productive and non-productive sequences divided based on post-IGHV-IGHD-IGHJ diversity into High and Low as described in Figure 1 . All comparisons were significantly different except for the following: for CDR: CDC vs. MZ; mMem vs. sMem; and sMem vs. DN; for FR: post-CDC^High vs. post-CDC^Low; post-CDC^High vs. non-Productive; CDC vs. mMem; CDC vs. sMem; CDC vs. DN; mMem vs. sMem; mMem vs. DN; sMem vs. DN. A complete list of statistical comparisons can be found in Supplemental Table S2 .

Figure 7

IGHV-D-J complexity correlates with clinical course based on time to first treatment. (A) Time to first treatment (TTFT) in months of the patients falling into the IGHV-mutated (M-CLL; blue) and IGHV-unmutated CLL subsets (U-CLL; green). (B) TTFT in months of the patients falling into the post-CDC complexity categories, Low and High. (C) TTFT (months) of the patients based on a combination of IGHV-mutation status plus the post-CDC complexity categories, U-CLL^High, U-CLL^Low, M-CLL^High, and M-CLL^Low. Statistics are displayed only when P-value ≤0.05 (*P-value ≤0.05, **P-value ≤0.01). NS, Not statistically significant.