Abstract
Several recent next generation sequencing studies have uncovered mutations in key genes that contribute to CLL pathogenesis. Most CLLs with deletions of chromosome 13q (del(13q), incidence 50-60%) or normal cytogenetics (incidence 15-20%) do not carry these recurrent driver mutations. Clues to the genetic basis of disease in this more indolent group likely lie elsewhere in the genome.
We present preliminary data from whole-genome sequencing (WGS) of a cohort of CLL patients with del(13q) (n=13) and normal cytogenetics (n=10), sequenced at 30X and 60X coverage for tumor and normal (saliva) respectively. Of the 23 patients, 43% were male and the median age at diagnosis was 53 years. The median time from diagnosis to sampling was 1.9 years, and 22 of 23 patients were previously untreated. With a median follow-up of 4.8 years in the overall cohort, 8 patients (35%) had been treated at a median 2.5 years from diagnosis. Although indolent by cytogenetics and of low Rai stage (median 1), the cohort includes balanced numbers of ZAP70± and IGHV mutated/unmutated cases.
A total of 68,928 point mutations with an average of 2,995 mutations per genome were detected in our cohort using the MuTect algorithm, which is approximately 60% fewer mutations than in multiple myeloma. The pattern of mutations in protein coding regions was similar to previous exome sequencing studies, with an average of 18.7 ±7.4 non-synonymous mutations per case (0.57±0.2 per Mb). The rate of mutations was significantly lower in patients diagnosed at a younger age (p=0.016, 2 sided t test). As expected based on the cytogenetic makeup of our cohort, the incidence of mutations in the 20 known CLL driver genes did not reach genome wide significance. However, the incidence of mutations in these driver genes was significantly higher in patients with normal cytogenetics as compared to del(13q) (60% vs 18%, p=0.039).
Rearrangement events were identified using the dRanger algorithm followed by mapping of the precise breakpoint to single base pair resolution using the Breakpointer algorithm. A total of 78 rearrangements were identified, corresponding to a median of 3 (range 0-15) per CLL genome, significantly fewer than in most other cancers. Similar to other hematologic malignancies, the incidence of deletions (70%) was significantly higher than that of tandem duplications (9%) (Wilcoxon signed rank test, p=0.0001). In addition, inter-chromosomal translocations exhibited a strong positive correlation with long range deletions (Pearson’s r=0.72, p=0.0001) and a moderate positive correlation with inversions (Pearson’s r=0.64, p=0.0009), suggesting a common mechanism. The only clinical feature associated with the presence of structural rearrangements was ZAP70 positivity (p=0.046).
Of the 13 patients with del(13q), 4 were due to inter-chromosomal rearrangement that led to loss of the common minimal region of deletion. The other most recurrent rearrangement partners were chr.2 (n=12 cases, 52%), chr.14 (n=9 cases, 39%) and chr.1 (n=5 cases, 21%), while no rearrangements were seen in chromosomes 8,11,17,18 and 19. The remaining chromosomes were affected in <20% cases. Chr.2 events accounted for 32% of rearrangements (25/78) with breakpoints scattered throughout the chromosome and consisting of 32% inversions, 28% long-range deletions, 20% inter-chromosomal translocations, 16% deletions and 4% tandem-duplications. Chr. 2p16 gain, a recurrent chromosomal abnormality in CLL, was not detected in our cohort, and therefore not associated with these rearrangements.
Chr.14 accounted for 17% of rearrangements (13/78), with breakpoints concentrated in the 5’IgH region (chr.14q32.33), in the vicinity of the KIAA0125 gene. Rearrangements involving chr.14q32.33 (n=11 total, 7 deletions) were associated with a shorter time to first therapy (TTFT) (HR=2.88, p=0.013), despite no association with IGHV or ZAP70 status.
In summary, this cohort of CLLs with del(13q) and normal cytogenetics represents the largest WGS cohort reported to date and underscores the stability of the CLL genome. Nonetheless, structural rearrangements are common, and the number of non-13q rearrangements was associated with higher risk ZAP70 status. Furthermore 5’IgH rearrangements enriched in deletions were associated with shorter TTFT within this lower risk subgroup. Additional WGS data on novel somatic mutations and regulatory regions will be presented at the meeting.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.