Synthetic lethality in ATM-deficient RAD50-mutant tumors underlies outlier response to cancer therapy

Hikmat Al-Ahmadie, Gopa Iyer, Marcel Hohl, Saurabh Asthana, Akiko Inagaki, Nikolaus Schultz, Aphrothiti J Hanrahan, Sasinya N Scott, A Rose Brannon, Gregory C McDermott, Mono Pirun, Irina Ostrovnaya, Philip Kim, Nicholas D Socci, Agnes Viale, Gary K Schwartz, Victor Reuter, Bernard H Bochner, Jonathan E Rosenberg, Dean F Bajorin, Michael F Berger, John H J Petrini, David B Solit, Barry S Taylor, Hikmat Al-Ahmadie, Gopa Iyer, Marcel Hohl, Saurabh Asthana, Akiko Inagaki, Nikolaus Schultz, Aphrothiti J Hanrahan, Sasinya N Scott, A Rose Brannon, Gregory C McDermott, Mono Pirun, Irina Ostrovnaya, Philip Kim, Nicholas D Socci, Agnes Viale, Gary K Schwartz, Victor Reuter, Bernard H Bochner, Jonathan E Rosenberg, Dean F Bajorin, Michael F Berger, John H J Petrini, David B Solit, Barry S Taylor

Abstract

Metastatic solid tumors are almost invariably fatal. Patients with disseminated small-cell cancers have a particularly unfavorable prognosis, with most succumbing to their disease within two years. Here, we report on the genetic and functional analysis of an outlier curative response of a patient with metastatic small-cell cancer to combined checkpoint kinase 1 (CHK1) inhibition and DNA-damaging chemotherapy. Whole-genome sequencing revealed a clonal hemizygous mutation in the Mre11 complex gene RAD50 that attenuated ATM signaling which in the context of CHK1 inhibition contributed, via synthetic lethality, to extreme sensitivity to irinotecan. As Mre11 mutations occur in a diversity of human tumors, the results suggest a tumor-specific combination therapy strategy in which checkpoint inhibition in combination with DNA-damaging chemotherapy is synthetically lethal in tumor cells but not normal cells with somatic mutations that impair Mre11 complex function.

Significance: Strategies to effect deep and lasting responses to cancer therapy in patients with metastatic disease have remained difficult to attain, especially in early-phase clinical trials. Here, we present an in-depth genomic and functional genetic analysis identifying RAD50 hypomorphism as a contributing factor to a curative response to systemic combination therapy in a patient with recurrent, metastatic small-cell cancer.

Conflict of interest statement

Conflict of interest: No potential conflicts of interest to disclose.

©2014 American Association for Cancer Research.

Figures

Figure 1. The treatment history and genomic…
Figure 1. The treatment history and genomic landscape of a metastatic carcinoma with an extreme outlier response to combination therapy
(A) Schematic representation of the treatment history of the index responder, a patient with metastatic small cell carcinoma of the ureter (PR, partial response; CR, complete response; NED, no evidence of disease). (B) Computed tomography images of the index patient prior to surgery of the recurrent tumor, prior to combined AZD7762 and irinotecan therapy, and one month after combined treatment (left, middle, and right respectively). (C) Somatic abnormalities in the responder’s genome (from outside to inside) included a heavy burden of copy number alterations; mutations at ~10-Mb resolution; regulatory, synonymous, missense, nonsense, and frameshift insertions and deletions (gray, black, orange, red, and green); and intra- and interchromosomal rearrangements (light and dark blue). (D) The allelic fraction of mutations is shown in genes identified by WGS of the post-etoposide/cisplatin tumor and also covered by the IMPACT assay and re-sequenced in the treatment-naïve primary tumor (blue and gray respectively).
Figure 2. D-loop and adjacent mutations in…
Figure 2. D-loop and adjacent mutations in RAD50
(A) DNA copy number segmentation inferred from WGS of the index case indicates a focal heterozygous loss spanning the RAD50 locus on 5q31.1, deleting the wildtype allele (as indicated by the sequence logo representing the allelic frequency), retaining only RAD50 L1237F. (B) The RAD50 L1237F mutation (red) is present in the D-loop of the ATPase domain near the Walker B motif (top and bottom, schematic of RAD50 protein at multiple scales). Directly adjacent appear a cluster of mutations in diverse malignancies (black). (C) Conservation of the RAD50 D-loop motif and adjacent sequence is indicated across 9 organisms in which the mutated leucine and adjacent aspartate residues are highlighted in green and yellow respectively, along with the positions of other mutations. The mutation position for human and yeast (in brackets) are given and indicated by arrowheads (Hs, Homo sapiens; Mm, Mus musculus; Dm, Drosophila melanogaster; At, Arabidopsis thaliana; Sc, Saccharomyces cerevisiae; Sp, Schizosaccharomyces pombe; Ec, Escherichia coli; Pf, Pyrococcus furiosus; T4, Bacteriophage T4). (D) The three dimensional structure of the Rad50 dimer indicating the position of the affected subunit with mutations colored in red. Inset indicates the position of mutant residues within close proximity to bound ATP.
Figure 3. RAD50 hypomorphism attenuates ATM signaling,…
Figure 3. RAD50 hypomorphism attenuates ATM signaling, synergizing with checkpoint inhibition to confer chemotherapy sensitivity
(A) While the Rad50L1240F protein level is reduced, the Mre11 complex is intact in rad50L1240F cells as well as those harboring similar D-loop or adjacent mutations. The Mre11-Rad50 interaction was assessed by co-immunoprecipitation with Rad50 or Mre11 antibodies (Rad50-IP or Mre11-IP) and western blot (anti-Rad50 or anti-Mre11) from yeast extracts of the indicated genotypes. Preimmune antibodies (PI) were included as a negative control. Rad50L1240R abundance was too low to rigorously determine whether complex formation was disrupted. (B) Mec1 (yeast ortholog of human ATR) deficiency dramatically potentiates the DNA damage sensitivity of rad50L1240F-mutant cells, an affect that could not be rescued by Sae2 deletion. Mec1-proficient (MEC1 WT, top 8 strains) or Mec1-deficient (mec1Δ, which also contains the sml1Δ suppressor, bottom 14 strains) cells of the indicated genotypes were 1/5 serially diluted and spotted on plates with or without the indicated concentrations of camptothecin (CPT) and grown for 3 days. (C) While Rad53 is phosphorylated (P-Rad53) upon MMS treatment in Rad50-wildtype cells with (+) or without Mec1 or Sae2 (Δ), these levels are attenuated in rad50L1240F mec1Δ sae2Δ triple-mutant cells or those with adjacent mutations, indicating that Tel1 (ATM) is not activated in rad50-mutant cells. (D) Graphical summary of the impact of the indicated rad50 alleles on Rad50 levels (blue); the integrity of the Mre11 complex; on CPT sensitivity in checkpoint-proficient (MEC1) and checkpoint-compromised (mec1Δ) cells (green), the ability of the Tel1 checkpoint to rescue cell survival (gray/black); Rad53 phosphorylation (blue) upon MMS treatment; telomere length (red). (E) A model of sensitivity to DNA damaging agents such as irinotecan driven by the synthetic lethality between simultaneous genetic and pharmacologic perturbation of both axes (ATM and ATR) of the DDR by RAD50 hypomorphism and checkpoint inhibition respectively.

Source: PubMed

3
Tilaa