Genomic Landscape and Clonal Heterogeneity Underlying Progression and Relapse In Chronic Lymphocytic Leukemia (CLL)

Despite recent advances and improvement in treatment options, chronic lymphocytic leukemia (CLL) is still an incurable disease. Recent advancement in parallel sequencing technology has provided better understanding of genetic landscape and clonal architecture underlying disease course. Furthermore, mutation status of TP53, SF3B1, NOTCH1, and BIRC3 has improved current CLL risk stratification. In order to better understand the genetic basis and clonal evolution underlying disease progression we analyzed a homogeneously treated cohort of CLL cases at multiple time points before and after therapy.

We performed WES in 30 sequential samples from 12 CLL cases. All 12 cases received PCR (pentostatin, cyclophosphamide and rituximab) based chemoimmunotherapy as initial treatment. Of these 12 cases, two had samples available both >6 months prior to enrollment and at treatment initiation, seven patients who experienced relapse had samples available at both treatment initiation and at and the time of relapse, and three patients had samples >6 months prior to enrollment, at treatment initiation, and at relapse. These cases were previously characterized by aCGH/FISH. R function Mclust and kernel density estimate (KDE) plots were used to identify subclonal heterogeneity.

An average of 102-fold depth coverage was obtained with an average of 80% of the targeted regions being covered by at least 30X. Overall, we detected a total of 219 nonsynonymous Single Nucleotide Variants (SNV) and indels (average 19, range 11-33). Del13q (42%) was identified as most prevalent copy number abnormality followed by trisomy 12 (33%), del11q and del17p (17% each). Recurrent mutations were identified in NOTCH1 (33%), DDX3X (25%), TP53, SF3B1 XPO1 and MED12 (17% each). Other tumor implicated genes such as CARD11, NOTCH2, NOTCH4, DIS3, TRAF2, NFKBID, CHK2 and RB1 were found mutated in individual cases.

In all 12 cases we identified at least one relevant cytogenetic abnormalities (del11q32, del13q14, del17p13, trisomy 12) and/or a mutated driver gene. Interestingly we found NOTCH1, DDX3X and XPO1 mutated at a significantly higher prevalence than previous studies (10%, 2% and 4%, respectively). Since all cases in our cohort correspond to individuals with progressive disease, a plausible explanation for the higher prevalence of mutations in our cohort is that these mutations are associated with disease progression.

By WES we identified multiple subclones in 58% of cases (7 of 12 cases), as compared to previously identified 33% (4 of 12 cases) by CGH/FISH. In 5 chemotherapy naïve cases, no significant increase in genomic complexity was detected with disease progression. In contrast, 5 out of 10 cases with samples analyzed before CIT and at the time of relapse had changes in clonal dynamics, with initially dominant subclones diminishing in response to therapy and the available empty niche being occupied by expansion of another “fitter” subclone at the time of relapse. No acquisition of tumor associated gene mutations was seen post therapy or associated with relapse in 9 of 10 cases. In the remaining case we identified an interesting example of convergent evolution. The pre-therapy sample was comprised of a unique clone characterized with del11q13, mutations in SF3B1 (G742D) and DDX3X (pE196fs). After therapy that clone was either eradicated or reduced to below detection level. Interestingly another clone emerged after relapse that harbored an independent del11q22 and mutations affecting different aminoacids of SF3B1 (K700E) and DDX3X (pL21fs). The strong selection of impairment in these genes during disease progression suggests a key role of these pathways/genes in the pathogenesis of this patient.

In conclusion, these studies provide additional insight into the genomic landscape and clonal architecture in a cohort of homogeneously treated CLL cases before and after therapy. Stable clonal architecture was seen in prior to treatment, whereas clonal evolution leading to increased tumor heterogeneity occurred after the under selection selective pressure of therapy in relapsed cases. Early identification of tumoral heterogeneity and underlying genetic abnormalities might provide better targets for therapy to decrease the likelihood of relapse.