Poly (ADP-ribose) Polymerase Inhibitors

THE USE OF PARP INHIBITORS IN SPORADIC CANCERS

Germline mutations in BRCA1 or BRCA2 are relatively common in hereditary breast and ovarian cancer. However, inactivation of BRCA genes by mutation in sporadic cancers is rare, at least in breast cancer, which may seem to limit the application of PARP inhibitors to a wider range of patients. However, many tumors display features in common with BRCA-deficient tumors, including similar defects in DNA repair due to either epigenetic mutation of BRCA1, such as promoter methylation, or mutation of other components of BRCA-associated pathways.²⁸ This BRCAness may make these tumors also susceptible to PARP inhibition.²⁸ For example, phosphatase and tensin homolog (PTEN) mutations, which occur with a frequency estimated at 50% to 80% in sporadic tumors,³⁸ may cause PARP inhibitor sensitivity in preclinical models, possibly because PTEN-null cells display BRCAness phenotypes, such as the inability to efficiently repair certain forms of DNA damage.³⁹

Traditional histopathologic methods and, more recently, gene expression profiling approaches have shown the phenotypic overlap between triple-negative breast cancers, basal-like breast cancers, and BRCA1 familial breast cancers.^40,41 In gene expression profiling studies, it has been observed that BRCA1 familial cancers strongly segregate with basal-like tumors and share features such as high-grade and pushing margins.^28,40,41 Although the overlap is not absolute, it leads to the hypothesis that there may be a subset of sporadic breast cancers that exhibits features of BRCAness, including deficiencies in HR and that may be susceptible to treatment with drugs such as PARP inhibitors.²⁶

There have been several studies of PARP inhibitors in sporadic ovarian cancer. A study by Lederman⁴² showed in a maintenance study following the response to platinum therapy a significant benefit in terms of progression-free survival (PFS) of olaparib compared to placebo. This was even more pronounced when the subgroup of BRCA mutation carriers were examined.⁴³ In both cases, the overall survival (OS) advantage was less than the PFS, but in the case of the BRCA mutation group, this reached statistical significance. Gelmon⁴⁴ also showed activity in sporadic ovarian cancer. In contrast, a study in sporadic triple-negative breast cancer failed to observe any benefit, although the study was small and the patients were heavily pretreated.⁴⁴

Iniparib (initially reported as a PARP inhibitor) showed an overall survival benefit in a Phase II trial of triple-negative breast cancer in combination with gemcitabine and carboplatin compared with chemotherapy alone.⁴⁵ However, a subsequent Phase III study showed no improvement in PFS.⁴⁵ The reasons for this are uncertain, but significant questions have been raised about whether iniparib is indeed a bona fide PARP inhibitor. Therefore, it is now generally conceded that studies of iniparib have no implications for PARP inhibitors as a drug class.⁴⁶

Which population of patients lacking a BRCA1 or BRCA2 mutation might benefit from PARP inhibitors remains unclear. This is likely to require the development of a clinical test to identify prospectively tumors with intrinsic sensitivity. Presently, most efforts are directed at developing assays of DNA repair deficiency.⁴⁷

MECHANISMS OF RESISTANCE TO PARP INHIBITORS

Resistance to targeted therapy frequently occurs, but it was unclear how resistance might arise to a synthetic lethal therapy.⁴⁸ Potential mechanisms of resistance to PARP inhibitors have, however, been elucidated both directly in vitro, in mouse models, and in the clinic.⁴⁸ An in vitro model for resistance was developed by producing cells from the highly PARP inhibitor–sensitive BRCA2-deficient cell line CAPAN1, which carries a c.6174delT BRCA2 frameshift mutation. CAPAN1 cells cannot form damage-induced RAD51 foci, are defective for HR, and are extremely sensitive to treatment with PARP inhibitors.⁴⁹ PARP inhibitor–resistant clones were highly resistant (over 1,000-fold) to the drug and were also cross-resistant to the DNA cross-linking agent cisplatin, but not to the microtubule-stabilizing drug docetaxel. PARP inhibitors and cisplatin both exert their effects on BRCA-deficient cells by increasing the frequency of misrepaired DSBs in the absence of effective HR. Therefore, this observation indicates that the resistance of PARP inhibitor–resistant clones to PARP inhibitors might be because of restored HR. This contention was supported by the acquisition in PARP inhibitor–resistant clone cells of the ability to form RAD51 foci after PARP inhibitor treatment or exposure to irradiation.

DNA sequencing of PARP inhibitor–resistant clones revealed the unexpected presence of novel BRCA2 alleles that resulted in the elimination of the c.6174delT mutation and restoration of an open reading frame.⁴⁹ Therefore, in this case, resistance arises because of gain of function mutations in the synthetic lethal partner (BRCA2) rather than the direct drug target (PARP). Alternative mechanisms of PARP inhibitor resistance have also been described.⁴⁸ A mouse model of BRCA1-associated mammary gland cancer demonstrated the efficacy of olaparib in vivo and was used to study mechanisms of resistance.⁵⁰ Resistance seemed to be caused by the upregulation of ABCB1a/b, which encode P-glycoprotein pumps; this effect could be reversed with the P-glycoprotein inhibitor tariquidar. In addition, other alterations in DNA repair pathways have been proposed to compensate for BRCA1 deficiency resulting in PARP inhibitor deficiency.⁴⁸

Studies of the mechanisms of resistance to PARP inhibitors in patient material are still at an early stage. Initial studies addressed the mechanism of resistance to platinum salts in BRCA mutation carriers. Cisplatin and carboplatin are part of the standard of care for the treatment of ovarian cancer, including individuals with BRCA1 or BRCA2 mutations. Platinum salts are thought to exert their BRCA-selective effects by a similar mechanism to PARP inhibitors.¹⁵ Clinical observations suggest that BRCA mutation carriers with ovarian cancer usually respond better to these agents than patients without BRCA mutations^51,52; however, resistance does eventually occur. To investigate this effect, BRCA1 and BRCA2 have been sequenced in tumor material from mutation carriers.^49,53 These studies revealed mutations in BRCA1 or BRCA2