Pharmacologic strategies targeting the DNA of tumor cells have been in use for much of the past century for many different cancer types. Radiation has also been a long-employed strategy to cause DNA damage and subsequent tumor cell death. However, the class of agents designed to inhibit the enzyme poly-(ADP-ribose) polymerase (PARP) have taken this a step further—these agents do not damage DNA themselves, but rather, inhibit the repair of DNA via inhibition of the base excision single-strand repair pathway. PARP inhibitors have been shown preclinically and clinically to enhance the affects of chemotherapies known to damage DNA or interefere with DNA replication. However, the most exciting use of PARP inhibitors may be in exploiting the concept of synthetic lethality. In this setting, the concept is based on two factors: (1) BRCA1/2-positive malignancies cannot use one of the major pathways to repair double-strand DNA breaks (ie, homologous recombination), and (2) making the base excision repair pathway nonfunctional via inhibition of PARP leads to tumor cell death, as unrepaired single-strand breaks are converted into double-strand breaks.
PARP-Sensitive Cell Lines
Drs. Comen and Robson have provided an excellent review of the development and early clinical use of PARP inhibitors. As described in their review, the initial preclinical data showing efficacy of these agents in a BRCA mutation population was described by two groups in Nature in 2005.[1,2] The work by Bryant et al focused on BRCA2-deficient cell lines, and showed that low concentrations of PARP inhibitors produced cytoxicity, which was not seen in cell lines with intact homologous recombination function, but was seen in cell lines with defects in homologous recombination. When gene function was restored in these cell lines, the sensitivity to PARP inhibitors was reversed. When BRCA2 was depleted, other cell line models for breast cancer (MCF-7 and MDA-MB-231) were similarly sensitive to PARP inhibitors.
Farmer et al similarly showed that cells with deficiencies in BRCA1/2 had reduced survival when transfected with short interfering RNA (siRNA) targeting PARP1. Similarly, when BRCA1- or BRCA2-deficient cells were exposed to the PARP inhibitors KU0058684 and KU0059848, enhanced cell death was seen. There was greater sensitivity to the inhibition of PARP in BRCA-deficient cells than with other DNA-damaging chemotherapies. These effects were limited to homozygous BRCA-deficient cells.
The question remains whether the sensitivity of BRCA-deficient cells is due inherently to loss of BRCA or whether the effects are due primarily to the inability to repair DNA by homologous recombination. McCabe et al published data examining this issue.[3] Cells that were deficient in other proteins important for DNA repair by homologous recombination (RAD51, RAD54, DSS1, RPA1) were found to be sensitive to PARP inhibition, as were cells with ATM and CHK2 depletion and deficiency in Fanconi anemia proteins. This may provide additional rationale for the use of PARP inhibitors, not only in patients with germline BRCA mutations, but also for other cancers that may have acquired deficiencies in homologous recombination.
Subsequent Investigations
These initial preclinical data showing proof-of-concept for synthetic lethality in BRCA-deficient cell lines were corroborated clinically in the first human study of the PARP inhibitor olaparib.[4] In a phase I dose-escalation study, Fong and colleagues selected a population enriched in BRCA-associated cancers. In the overall population, there were no objective responses. In the group of 19 patients with a documented BRCA mutation, including breast, ovarian, and prostate malignancies, there was a 47% response rate and a 63% clinical benefit rate. Importantly, toxicity did not differ in the patients with a BRCA mutation compared to those without one. The subsequent phase II studies performed in both breast and ovarian cancer patients—presented at the 2009 annual meeting of the American Society of Clinical Oncology (ASCO)—revealed objective response rates of 30% to 40% and good tolerability.
As the single-agent studies revealed promising activity, a number of combination studies have been undertaken in a variety of malignancies with a number of different chemotherapeutics. In triple-negative breast cancer, a recent randomized phase II study compared addition of the PARP1 inhibitor BSI-201 to chemotherapy with gemcitabine(Drug information on gemcitabine) (Gemzar) and carboplatin(Drug information on carboplatin) vs chemotherapy alone.[5] Preliminary data presented at ASCO 2009 demonstrated that the addition of the PARP inhibitor was associated with improvements in response as well as progression-free and overall survival that were highly significant. Objective responses with the addition of the PARP inhibitor were improved from 16% to 48% P = .002), and median progression-free survival was improved from 3.3 to 6.9 months P < .0001). A randomized phase III study is currently underway to confirm these results with gemcitabine, carboplatin, and BSI-201.
Unanswered Questions
Despite the exciting clinical results obtained with PARP inhibitors, there remain a number of important unanswered questions regarding optimal use. The mechanisms by which triple-negative or basal breast cancers are sensitive to these agents and the potential mechanisms of resistance in both BRCA-intact and BRCA-mutant malignancies have yet to be described fully. It has been shown that basal breast cancers are “BRCA-like.” Gene microarray expression profiling studies show considerable similarity between BRCA-mutated tumors and basal tumors.[6] What is unknown is whether the basal-phenotype is the cause or the effect of the BRCA mutation/inactivation.[7]
In sporadic basal tumors, data show decreased BRCA1 mRNA expression in up to 30% of cases, and one of the potential mechanisms is epigenetic modification.[8-10]. However, to date, none of the studies including triple-negative breast cancers have published any correlative data describing the potential mechanisms underlying the sensitivity of this breast cancer subtype to PARP inhibitors, or the underlying basis that leads to PARP overexpression. As Drs. Comen and Robson point out and as described by others, PARP inhibitors could potentially possess single-agent activity if, in fact, there is inactivation of BRCA with subsequent loss of homologous recombination function. Other mechanisms may result in an overexpression of PARP or reliance on the base excision repair pathway beyond deficiencies in homologous recombination. Recently, cells with PTEN mutations have been shown to be similarly sensitive to PARP inhibitors, which may be an important additional application for these agents.[11]
Another important unanswered question involves the mechanisms of resistance to PARP inhibitors in patients with BRCA1/2 mutations. In the phase I/II trials of olaparib, response rates—while impressively near 50%—indicate that half of the patients will not respond. Evidence in platinum-refractory patients with BRCA mutations has suggested the existence of intragenic “reversion mutations,” whereby patients have an additional mutation that restores homologous recombination functionality.[12-14] Similarly, PARP inhibitor–resistant clones have been produced from CAPAN1 cell lines (which cannot repair DNA by homologous recombination) by continuous exposure to increasing PARP inhibitor treatment. In these resistant clones, the ability to perform homologous recombination was restored.[12] The examination of samples from patients on PARP inhibitor trials should help to discern additional potential mechanisms of resistance.
At this point, we do not know: (1) which patients are best suited for PARP inhibitor therapy, (2) which chemotherapy is the best to combine with these agents, and (3) whether they have any potential for single-agent activity in tumors with indications of “BRCA-like” deficiencies in homologous recombination. But what cannot be denied is that this is the beginning of an exciting chapter in the development of these agents. With properly designed studies and informing correlative data, these answers should be forthcoming.
Financial Disclosure: The authors have no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.
