Immunoglobulin mutational status detected through single-round amplification of partial V(H) region represents a good prognostic marker for clinical outcome in chronic lymphocytic leukemia

Abstract

The immunoglobulin (Ig) mutational status in B-cell chronic lymphocytic leukemia (CLL) distinguishes two subsets of patients with different prognosis. Ig status detection is commonly performed with a panel of V(H) family-specific primers. Although this method detects clonal VDJ rearrangement in virtually all cases, it is technically cumbersome and therefore not widely used clinically. Here, we describe a simple and rapid method to establish the mutational status of IgV(H) in CLL. The method is based on a consensus V(H) FR2 primer, used in both polymerase chain reaction (PCR) and sequencing reactions. Overall, monoclonal B-cell populations were detected in 163 of 189 CLL patients (86%). The prognostic value of IgV(H) mutational status was then evaluated by analyzing survival in 146 CLL cases using different V(H) homology cutoffs. CLL prognostic groups were best separated by the classical 98% cutoff: median survival was 127 and 206 months in unmutated and mutated CLL cases, respectively (P = 0.0023). V(H) FR2 consensus and V(H) family PCR were compared in 41 cases, correctly assigning all cases by both methods. Therefore, we suggest a sequential strategy to detect immunoglobulin mutational status in CLL patients by first using the approach described in this study followed by alternative V(H) family-specific PCRs for negative cases.

Figure 1

Comparison between nucleotide sequences obtained with upstream FR2 and downstream JHE primers in a single case. Sequence alignment was performed using Clustal tool of Sequence Navigator 1.0.1 software. Sense and complementary reverted anti-sense (*) nucleotide sequences were aligned together (above); correspondent sense and anti-sense electropherograms relative to CDR3 region were shown below.

Figure 2

Differential CDR3 length between three groups of B-CLL patients with different mutational rate in IgV_H region expressed by leukemic population. The horizontal line inside boxes indicates the median value. LM-CLL patients have a percentage of mutations ranging from 2 to 5%, HM-CLL patients have a percentage of mutations greater than or equal to 5%, and UM-CLL patients have mutations in a percentage inferior to 2%. A statistically significant difference is shown between M-CLL (LM-CLL + HM-CLL) and UM-CLL groups using a nonparametric Mann-Whitney test (P < 0.0001).

Figure 3

Kaplan-Meier survival curve comparing 146 CLL patients with mutated and unmutated V_H gene (homology level cutoff = 98%). Median survival for 63 unmutated CLL, 127 months; median survival for 83 mutated CLL, 206 months. The difference is significant at the P = 0.0023 level (log-rank test).

Figure 4

Log-rank statistics were performed for the V_H homology rate to test for a possible cutoff value separating two groups with different survival distributions. V_H germ line homology rate from 96 to 98.5% is shown on the x axes. Corresponding P values calculated by log-rank test for the distinction of two groups with different survival probability are shown on the y axes. Classical 98% homology rate is able to best separate prognostic groups between B-CLL patients.

Figure 5

Scatter diagram of IgV_H mutation percentage determined analyzing both complete IgV_H sequence obtained by V_H family-specific leader primers (x axes) and partial sequence obtained by FR2 degenerate primer (y axes). A: Comparisons were performed in 41 cases of our cohort; median gap in percentage of mutations was 0.3%, ranging from 0 to 3.8%. These two parameters were positively correlated with Spearman’s rank order correlation coefficient of 0.96 (P < 0.0001). B: Comparisons were also performed on 80 cases selected from literature (set 1, n = 40, circle; set 2, n = 40, triangle) considering only the portion of V_H sequence 3′ downstream to the annealing site of the FR2 primer. Cases with 100% homology to germ line of the entire V_H sequence were obviously excluded from the analysis. Median gap in percentage of mutations was 0.85% (range, 0 to 3.7%) in set 1 and 0.4% (range, 0.1 to 3.4%) in set 2. These two parameters were positively correlated with Spearman’s rank order correlation coefficient of 0.96 (P < 0.0001) and 0.95 (P < 0.0001) in sets 1 and 2, respectively. The vertical and horizontal lines indicate the mutation cutoff point of 2% commonly used to distinguish B-CLL patients into mutated and unmutated subsets. All cases were assigned to the same prognostic group (mutated and unmutated CLL) by both methods.