Mutations In XPO1, POT1, BIRC3 and FBXW7 collectively Predict Poor Outcome At Diagnosis In CLL and MBL Independent From The SF3B1 and NOTCH1 Status

With the application of high-throughput sequencing, our understanding of the somatic mutations that contribute to the pathogenesis of CLL has expanded significantly. Amongst these mutations, NOTCH1 and SF3B1 are the most prevalent and at diagnosis are associated with reduced survival independent of clinical and biological variables. However, we have limited understanding of the prognostic impact of other, less prevalent mutations. To address this question, we investigated the distribution, clinical significance and clonal acquisition of SF3B1, NOTCH1, XPO1, POT1, BIRC3, MYD88 and FBXW7 mutations in a single-centre diagnostic cohort of 140 patients with CLL, of whom 115 had Binet stage A disease and 91 patients with clinical MBL based on the 2008 IWCLL/NCI diagnostic criteria. Ninety patients (35 MBL and 55 CLL) also had DNA samples analyzed at a second time point (median 72 months, 24-145 months from diagnosis). We defined progressive disease as either presentation with stage B/C disease, evolution of MBL or stage A to stage B/C and/or the need for treatment. With high-resolution melt (HRM) analysis and subsequent Sanger sequencing, we identified mutations of SF3B1 in 9% (21/225), NOTCH1 in 5% (12/228), POT1 in 6 %(7/123), XPO1 in 2 %(4/196), MYD88 in 1% (3/223) and BIRC3 mutations in 0.5 % (1/222) of patients. FBXW7 mutations were identified in 8% (3/36) of the trisomy 12 patients from our cohort. In the patients studied sequentially we confirmed acquired mutations of SF3B1 (n=5) and XPO1 (n=1) in 3 patients with MBL and in 3 with CLL. Whilst we confirmed the impact of NOTCH1 mutations on treatment free survival (76 vs. 107 months, p=0.002) and SF3B1 mutations on overall survival (96 vs. 117 months, p=0.004), the prevalence of mutations in the other genes was too low to assess individually. We therefore assessed the collective burden of mutations in our cohort (excluding the MYD88 cases which had long term stable disease ) and demonstrated the following: 1) In both the MBL and CLL cohorts, the frequency of a gene mutation was higher in patients with progressive disease than in those with long-term, clinically stable disease (MBL: 36 vs. 12% p=0.02, CLL: 41 vs. 10%, p<0.001),and also independent of SF3B1 and NOTCH1 mutations and deletions of 17p (MBL and CLL: 11 vs. 3%; p=0.031). 2) All patients and those in the MBL cohort, with mutations targeting genes other than SF3B1 and NOTCH1 were more likely to require treatment (Entire cohort: OR 3.1, p=0.008, MBL cohort: OR 20, p= 0.006). This resulted in a reduced treatment-free survival (Entire cohort: 40 vs. 102 months, HR 10, 95%CI 3-30, p<0.001) and overall survival. (Entire cohort: 97 vs. 120 months HR 2.7, 95% CI 1.2-5.9, p=0.013) which was independent of the presence of a 17p deletion (p=0.008) and of IGHV status in bivariate analysis (p=0.02). 4/6 patients who acquired a mutation at 38-208 months (median 127 months) after initial analysis required treatment. Taken together, our study suggests that identifying mutations in genes other than NOTCH1 and SF3B1 may have utility in the clinical management of MBL and CLL as they predict for progressive disease, treatment requirement and reduced survival. Much larger studies with sequential data post therapy will be required to elucidate whether specific rare gene mutations predict a poor outcome or whether they either result in, or are a marker of genomic instability and subsequent genomic complexity.