SF3B1 Mutations Frequently Occur With Both ATM Mutations and TP53 Mutations In CLL Patients

Abnormalities of TP53 and ATM genes are well established adverse prognostic markers in CLL. Mutations in splicing factor SF3B1 have recently been described as recurrent and predominantly subclonal aberration. Previous studies inconsistently suggested mutual exclusivity or partial overlap with TP53 mutations. Concerning ATM defects, the association between SF3B1 mutations and del(11q) was reported, but relation to ATM mutations remains unclear.

The aims were: (a) to assess association between SF3B1 mutations and the most adverse classic genetic lesions represented by TP53 mutations and del(11q); (b) to investigate association between SF3B1 mutations and ATM mutations in a subset of ATM-characterized patients, and (c) to delineate SF3B1 mutation profile and proportion. We used Sanger sequencing of SF3B1 hot-spot exons 14-16, FASAY analysis of TP53 exons 4-10 and resequencing microarray for ATM mutation detection (all 62 coding exons).

We analyzed unfavorable cohort of 338 patients characterized by prevalence of unmutated IGHV (72%). At the time of analysis, 82.5% of the patients were previously untreated. We observed SF3B1 mutation in 17.5% (59/338), TP53 mutation in 20% (68/338), and del(11q) in 27.5% (93/338) of cases. All these genetic defects were significantly more frequent in treated patients (SF3B1: P=0.008, TP53: P<0.0001, del(11q): P=0.0295). Interestingly, we observed quite frequent co-occurrence of SF3B1 and TP53 mutations; 28% (19/68) of p53-affected patients in comparison with 15% (40/270) of p53-wt patients harbored SF3B1 mutation (P=0.019). This co-occurrence was apparent (albeit without statistical significance, P=0.166) also in patients investigated at diagnosis, when 20.6% (7/34) of p53-affected patients but only 11.5% (19/165) of p53-wt patients exhibited SF3B1 mutation. The previously reported increased SF3B1 mutation frequency in patients with del(11q) was also apparent but not significant (P=0.078) in our cohort since mutation frequencies were 24% (22/93) and 15% (37/244) in groups with and without del(11q). To analyze the relation of SF3B1 and ATM mutations we used samples with characterized ATM mutation status (n=112). The p53-defective samples and samples with sole del(11q) were omitted since these more frequently harbored SF3B1 mutation. In the remaining 37 patients we observed that SF3B1 and ATM mutations frequently co-occur: 8/21 ATM-mutated patients (38%) but only 2/16 ATM-TP53-wt patients (12.5%) exhibited SF3B1 mutation (P=0.137). This association should be, however, verified on larger cohort. Concerning the SF3B1 mutation profile, we observed previously reported hot-spot missense mutations with the most abundant mutation K700E (21/59, 36%). In addition, we found 2 short in-frame deletions. Altogether it shows that rather than SF3B1 loss-of-function only partial impairment or gain-of-function is possible. The vast majority of samples had mutation in heterozygous state that correlates with presumably preserved second allele. Interestingly, we found 4 mutations in a proportion around 90% that could be explained by loss of heterozygosity in the locus 2q33.1 by any cause.

Our study indicates that SF3B1 mutations frequently overlap with mutations in the DNA damage response pathway genes at least in prognostically unfavorable cohort. It seems that SF3B1 mutations do not lead to complete protein loss indicating rather active involvement of mutated SF3B1 in splicing processes in CLL cells. The work was supported by grants NT13519-4, NT11218-6, MSM0021622430, MUNI/A/0723/2012.