Complexity in the Gastric Cancer Genome and a Biomarker-Driven Trial of Poly (ADP-Ribose) Polymerase Inhibition in Gastric Cancer.

Abstract

Gastric cancer remains a global health problem, with poor results from treatment in metastatic or recurrent disease and 723,000 deaths worldwide per year. Recent efforts at molecular classification of gastric cancer, including publication of the work of the Cancer Genome Atlas Network, have yielded insights into potential targets in the principal categories of Epstein-Barr virus–related, microsatellite unstable, genomically stable, and chromosomal unstable gastric cancer and confirm prior work on genetic and protein expression abnormalities in gastric cancer, among them alterations in the DNA damage response gene ataxia telangiectasia mutated (ATM). This expanded understanding of the importance of dysregulated DNA damage response has opened up the possibility of personalized application of synthetically lethal therapies that exploit these alterations in DNA damage response to foster apoptosis or mitotic catastrophe. In the article that accompanies this editorial, Bang et al have undertaken an admirable attempt at a biomarker-enriched trial to pursue this approach. The strengths and limitations of their trial are best understood in the context of the molecular subtypes of gastric cancer. ATM regulates cellular signaling response to DNA doublestrand breaks via p53, and loss of ATM has been demonstrated to be synthetically lethal with inhibition of poly (ADP-ribose) polymerase (PARP), a stress-sensing protein critical in base excision repair. Inhibition of PARP is active in homologous DNA damage repair– deficient cells, but PARP inhibitors may differ in their off-target effects. The PARP inhibitor olaparib has been shown to have off-target effects that result in cell cycle arrest in a p53dependent manner, likely augmenting the anticancer activity of this compound. The preclinical work of Kubota et al has confirmed that sensitivity to PARP inhibition differs by p53 mutational status in gastric cancer cell lines as well, with loss of p53 function sensitizing to PARP inhibition and lesser sensitivity in lines with low ATM expression that are p53 proficient. However, the picture is complicated for a number of reasons. First, specific p53 mutations may result in truncation or loss of DNA binding on the one hand and gain of function on the other; the functional impact of specific mutations can be predicted via a variety of algorithms or established via functional assays of waf1 or PUMA activation. Second, the two clinically important molecular abnormalities in gastric cancer of loss of function of ATM and of p53 may not always coexist in individual patients. ATM mutation and loss of ATM expression are found in more than 60% of gastric cancers with the microsatellite instability (MSI) phenotype, and p53 mutation with loss of function is predominantly found in the chromosome unstable class of gastric cancers. Thus, comprehensive characterization of the mutational profile and functional p53 status may be required to optimally pursue the therapeutic implications of altered DNA damage response in gastric cancer. In addition, the group of patients susceptible to this approach may be somewhat larger if gastric cancers with altered expression of other DNA response genes are amenable to the same synthetic lethal approach. With ATM, as with any abnormality that impacts DNA repair, the effect on mutational landscape and immunophenotype are now also of interest in identifying patients who may be suitable for novel immunotherapies, and recent data suggest that MSI tumors may be particularly good candidates for therapy with immune checkpoint inhibition. The trial presented by Bang et al is an interesting phase II trial of the PARP inhibitor olaparib in combination with paclitaxel compared with paclitaxel alone as second-line therapy of metastatic or recurrent gastric cancer. The trial incorporated an intentional enrichment for a population with low or undetectable ATM expression (ATMlow), putatively the phenotype more sensitive to PARP inhibition, reflected in the preclinical work of Kubota et al discussed earlier. The primary end point of the trial was progression-free survival (PFS), which may have been difficult to determine in some patients because the trial accepted patients with nonmeasurable disease and peritoneal spread, which is notably more difficult to evaluate for response or progression than other metastatic sites and possibly associated with inferior outcome. In the event, both arms included a substantial proportion of patients with peritoneal metastases, possibly compromising determination of progression. The duration of chemotherapy to be administered in the trial was poorly defined, because patients were projected to receive six to 10 cycles of paclitaxel with olaparib or placebo, before continuing on olaparib or placebo. In addition, the sample size was small, and the hypothesis of a hazard ratio of 0.55 in the ATM-enriched population was ambitious. The trial did not meet its primary end point; the addition of olaparib to paclitaxel did not improve PFS in either the overall JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 33 NUMBER 33 NOVEMBER 2