Characterizing ATM Aberrations in Chronic Lymphocytic Leukemia (CLL): Prognostic Implications and Sensitivity to PARP Inhibition

Abstract

Introduction ATM mutations (muts) frequently occur in CLL, yet their influence on ATM protein function and DNA break repair is poorly elucidated. Our prior study demonstrated functional hypomorphism of the rare germline variant L2307F in ATM and a higher occurrence of 11q deletions (del(11q)) with ATM germline variants in CLL (Lampson et al, JCO 2023). However, the role of other ATM germline variants in association with del(11q) or somatic muts remains unknown. Here, we analyzed the characteristics of CLL patients (pts) with ATM aberrancies and evaluated ATM function in primary CLL cells. Methods We retrospectively analyzed 278 CLL pts with ATM aberrancy, evaluated by clinical targeted sequencing from 2014 to 2019, as per our recent study (Lampson et al, JCO 2023). ATM, KAP1, and H2AX phosphorylation levels were examined after etoposide or irradiation (IR) exposure, by western blot or flow cytometry, in primary CLL cells with or without ATM aberrancy. We assessed the inhibitory effect of talazoparib, a PARP inhibitor, on CLL proliferation over 14 days using flow cytometry and Cell Trace Violet dye. Results We identified 294 ATM variants, 146 unique (germline N=74, somatic N=72). In CLL pts, 58 had dual ATM allele loss (biallelic group), including del(11q) together with only germline ATM variant(s) (N=29), only somatic mut(s) (N=18), or both (N= 11). 168 pts had ATM muts or variants without del(11q), including only germline (N=137), somatic (N=19), or both (N=12) ( ATM group), and 52 pts had del(11q) alone (11q group). Complex karyotype increased with ATM allele deficiency, particularly with 5+ abnormalities (complex karyotype 5+, biallelic vs 11q vs ATM: 22.6% vs 10.6% vs 5.8%, P=0.002). Unmutated IGHV was significantly higher in the groups with del(11q) versus ATM alone (biallelic vs 11q vs ATM: 85.1% vs 87.8% vs 53.9%, p<0.001). Interestingly, when we compared IGHV status in the ATM alone group with or without somatic ATM muts, unmutated IGHV significantly increased in those with somatic muts ( ATM-somatic vs ATM-germline alone, 81.8% vs 43.6%, p<0.001), and karyotypic complexity tended to be higher in the somatic ATM mut group (P=0.077). This might suggest that pts with ATM somatic muts have clinical characteristics and pathogenicity more similar to del(11q) as compared to those with germline muts. SF3B1 muts were more common in those with del(11q) (biallelic vs 11q vs ATM: 20.7% vs 26.9% vs 12.5%, P=0.036). For functional evaluation, we compared the five groups of primary CLL samples: del(11q) with somatic ATM mut with/without germline (11q+ ATM-S), del(11q) alone with no ATM muts (11q), somatic ATM muts without del(11q) with/without germline ( ATM-S), germline ATM muts alone ( ATM-G), and no ATM aberrancies (wild type, WT). Western blot demonstrated significantly lower phosphorylated protein levels (pATM, pKAP1, γH2AX) in response to DNA damage with etoposide in ATM aberrancy groups compared to WT (WT vs 11q, vs ATM-S, vs 11q+ ATM-S, all P<0.001 except for WT vs 11q in γH2AX, P<0.05) (Figure 1A). Flow cytometry also showed significantly lower phosphorylation levels of ATM after IR in 11q, ATM-S and 11q+ ATM-S groups compared to WT, while ATM-G group didn't show significance. To assess ATM deficient cell sensitivity to talazoparib, we compared proliferation inhibition across groups cocultured with talazoparib for 14 days. The 11q+ ATM-S group exhibited significantly greater sensitivity to talazoparib at lower dosage compared to the WT (talazoparib 0.1 μM, P<0.05; 0.5 μM, P=0.098) (Figure 1B). The monoallelic ATM deficient groups (del11q and ATM-G) tended to have lower proliferation, but not significantly. Higher concentrations of talazoparib (2.5μM) strongly inhibited proliferation of all CLLs including WT. Conclusions Pts with ATM somatic aberrancy including del11q had worse associated prognostic factors including unmutated IGHV and complex karyotype. Somatic events showed greater functional impact than some germline variants. CLL cells with ATM aberrancies had ATM pathway dysfunction, and those with biallelic ATM aberrancies were more sensitive to the PARP inhibitor, talazoparib, than WT. This work could potentially lead to additional avenues for treatment based on differential sensitivities of CLLs with ATM aberrancy. Further studies are needed to characterize specific ATM germline variants with or without somatic ATM events.