A series of setbacks almost caused the demise of PARP inhibitors, but four companies are now beginning pivotal trials of these agents in breast and ovarian cancer.

Nature Reviews Drug Discovery, October 2013

by Ken Garber

Inhibitors of poly ADP-ribose polymerase (PARP) are staging a comeback. These anticancer drugs, given up for dead by many observers in 2011, are now poised for possible regulatory approval within three years. At least six Phase III trials, in ovarian and breast cancer, have either begun or will launch by the end of the year. “I think we’ve made that turnaround,” says Susan Domchek, an oncologist and breast cancer researcher at the University of Pennsylvania. “A lot has been going on to get us excited again in this field.”

Sanofi, once the field’s leader, has dropped out because of Phase III failures, but has acknowledged that its compound, iniparib, was not a functional PARP inhibitor. And AstraZeneca is now scrambling to recover a future for olaparib, after it prematurely terminated its ovarian cancer programme. But the field is now much more tightly bunched, with four other companies also launching pivotal trials soon. Subtle differences in clinical trial design may spell the difference between winners and losers.

A case of whiplash

Excitement around PARP inhibitors, past and present, stemmed from the prospect that inhibition of the enzyme would exploit a pre-existing weakness in cancer cells, triggering their self-destruction while avoiding collateral damage. PARP-1 and PARP-2 facilitate DNA repair. They recognize and bind single-strand breaks in DNA and catalyze the formation of branched chains of poly-ADP ribose (PAR), which in turn recruit DNA-repair enzymes to the damage sites.

Beginning in the 1990s companies began developing PARP inhibitors as chemopotentiators, because many chemotherapy drugs cause DNA damage that is repaired via PARP activity, leading to drug resistance. But although PARP inhibitors showed anticancer activity, they proved highly toxic when used in combination with chemotherapy. Then, in 2005, two breakthrough Nature papers showed that PARP inhibitors given alone could kill cancer cells with pre-existing DNA repair defects, specifically mutations in the BRCA1 and 2 genes, which are involved in double stranded DNA repair through homologous recombination (HR). Cell killing that occurs after combining two conditions, neither in itself sufficient to affect viability, is known as synthetic lethality. PARP inhibition and mutant BRCA were synthetically lethal in preclinical models, suggesting an elegant, targeted and minimally toxic way to treat patients.

Dramatic Phase I results for AstraZeneca’s olaparib in BRCA-positive breast and ovarian cancers, published in 2009, energized clinical oncologists. (New Engl. J. Med. 361, 123-34; 2009.) “The idea that you could take these individuals, who were at such elevated risk for cancer because of these mutations, and then use that as an Achilles heel, was just tremendously exciting,” says Domchek.

The field, though, overreached when it went after a wider breast cancer population. Tumor cells with intact BRCA genes but harboring other HR defects, a condition dubbed “BRCAness,” should in theory be vulnerable to PARP inhibitors, hypothesized some researchers, since such cells are similarly unable to repair double strand breaks. Triple negative breast cancer, a particularly lethal subtype, was identified with “BRCAness”, and outstanding Phase II results for iniparib in this population, published in January 2011 (New Engl. J. Med. 364, 205-14; 2011) seemed to confirm that hypothesis . The drug, combined with chemotherapy, doubled progression-free survival (PFS) compared to chemotherapy alone, without inducing myelosuppression seen with other PARP inhibitors.

The elation was shortlived. Just a few weeks later, Sanofi reported that iniparib, inexplicably, had failed in a Phase III trial in the same patient population. The announcement staggered the entire field. Then AstraZeneca’s olaparip failed to produce any response whatsoever in 16 patients with triple negative breast cancer lacking BRCA mutations. (Lancet Oncol. 12, 852-61, 2011.) Sanofi abandoned iniparib for breast cancer, and olaparib clinical activity in the disease “came to a screeching halt,” says one outside clinical investigator.

PARP inhibitors suffered a second, near-fatal blow that December when AstraZenecadiscontinued development of olaparib for ovarian cancer. One Phase II trial of the drug had failed to show a difference in PFS compared to doxorubicin chemotherapy (J. Clin. Oncol. 30, 372-9; 2012), while a second Phase II trial, known as “study 19,” had shown benefit over placebo on PFS, but not on overall survival. (N. Engl. J. Med. 366, 1382-92; 2012.) Following these perceived setbacks, with all PARP inhibitors now widely viewed as ineffective, “the field kind of evaporated,” says Andrew Allen, chief medical officer of Clovis Oncology. Companies cancelled clinical trial plans, and Pfizer and Merck outlicensed their PARP inhibitors to biotech startups Clovis and Tesaro, respectively.

But not everyone lost faith in PARP inhibitors. “For those of us who were in the know, we recognized very early on that that [iniparib] was not a bona fide PARP inhibitor,” says Philip Jones, former PARP inhibitor medicinal chemistry team leader at Merck. “It is a prodrug that generates reactive species that have multiple other activities, even if it is a weak PARP inhibitor. When that compound failed in the phase III study, we were not surprised, but it created a lot of negative press for this very promising class of compounds.” Jones is now head of drug discovery at the Institute for Applied Cancer Science at the M.D. Anderson Cancer Center.

Reassessment and Resurrection

Two recent developments have rehabilitated PARP inhibitors. First, independent investigators confirmed that iniparib and its metabolite were not functional PARP inhibitors. (Clin. Cancer Res. 18, 510-23, 2012 and 18, 1655-62, 2012) “We took the view that the iniparib data were completely irrelevant to the class,” says Allen.

Around the same time, AstraZeneca re-examined study 19, after retrospectively genotyping almost all the patients. The results were presented at the annual meeting of the American Society for Clinical Oncology in June. In the BRCA-mutant group, a trend towards improved overall survival began to emerge. Most dramatically, patients with germline or somatic tumor BRCA mutations had a median PFS of 11.2 months on olaparib, compared with 4.3 months for placebo—double the difference reported earlier in the overall patient population. Many BRCA mutation-negative patients also benefitted, in line with findings by researchers showing that high-grade serous ovarian cancer, the most common disease subtype, frequently harbors non-BRCA “BRCAness” defects in HR. On the strength of this new data, AstraZeneca recently restarted its PARP inhibitor program in ovarian cancer after reformulating olaparib for more convenient dosing.

But AstraZeneca wasted at least two years. Stopping olaparib development “was just such a big mistake,” says Jonathan Ledermann, the oncologist at the University College London Cancer Institute who conceived and led study 19. “We lost a lot of time. And it allowed competitors to move forward quite quickly.”

Jones agrees that AstraZeneca blundered. “They were clearly leading the race at that time,” he says. “If they had continued they probably would have had marketing approval by now, if they had done the right clinical study. That’s really opened the field to the likes of Clovis, Tesaro and BioMarin.”

Who to treat?

So far, the pharmacological properties of the four leading candidates suggest little differentiation. “We all have very similar drugs from an efficacy point of view,” says Clovis’s Allen. BioMarin’s BMN 673 is more potent than the others (Clin. Cancer Res., 5 Sep 2013), but that does not necessarily translate into a clinical advantage. Like the other PARP inhibitors, its dose-limiting toxicity is myelosuppression, probably mechanism-based. “The one advantage with a more potent compound is that it can be used at a lower dose, but ultimately if the clinical activity is the same that doesn’t really help much,” explains Jones.

The compounds do have somewhat different specificity profiles. While they all target PARPs 1 and 2, they have different activity against the 15 other PARP family members, some of which have unknown function. Off-target toxicity may as a result vary enough to make a difference. “Tolerability matters,” says Allen, especially in the maintenance setting, where women start out free of measurable disease and don’t expect to get sick from their medication.

The main determinant of success may instead be Phase III trial design for high-grade serous ovarian cancer, where the drugs still have the potential to help the “BRCAness” population. Three companies – AstraZeneca, Tesaro and Clovis – are adopting the study 19 strategy, testing their candidates as single agents for maintenance therapy, with the goal of delaying or prevent disease relapse following chemotherapy treatment. Although 50-60% of high-grade serous ovarian cancer patients achieve a complete response to initial therapy, most eventually relapse; and relapsed ovarian cancer is incurable.

Differences appear in the trial populations. AstraZeneca is targeting the narrowest but least risky population, the small fraction of serous ovarian cancers that harbor BRCA mutations. (One of Astra Zeneca’s two phase III trials is in the frontline setting, where it could contribute to disease cure.) Tesaro, by contrast, will test its PARP inhibitor niraparib in relapsed patients with high-grade serous ovarian cancer regardless of BRCA status, a broader but riskier strategy that could dilute the overall clinical benefit. “The data… suggests that patients who are non-germline BRCA may benefit,” argues Tesaro president Mary Lynne Hadley. Not offering PARP inhibitors to such patients “would be a travesty,” she adds.

Clovis, too, is offering rucaparib, its PARP inhibitor, to the larger population of relapsed high-grade serous ovarian cancer patients. But, in an effort to identify the subpopulation most likely to benefit, Clovis is collaborating with cancer genomics company Foundation Medicine on a predictive molecular signature of homologous recombination deficiency (HRD) The signature, whichwill be developed from Phase II data, will not be used to select phase III patients, but will instead be tested atthe trial’s end for predictive value. This unusual approach is acceptable to the FDA and EMA, says Clovis’s Allen. “Both are content that this is a true prospective test of the BRCAness signature,” he says.

Since Clovis announced its strategy, other companies have also started to strike up agreements to develop HRD companion diagnostics, says Robert Coleman, an oncologist at M.D. Anderson. Myriad spokesperson Ron Rogers says his company has two ongoing collaborations with undisclosed drug companies to develop its HRD test as a companion diagnostic for PARP inhibitors.

Domchek cautions that these diagnostics will need to be carefully validated to make sure that patients who might benefit from PARP inhibitors aren’t excluded. “We want to make sure that we’re not throwing the baby out with the bathwater, if we’re narrowing in,” she says. Already the neediest patients, women with ovarian cancer who have failed chemotherapy, have been left out of phase III.

AstraZeneca, Tesaro and Clovis are committed to the maintenance strategyas the quickest route to FDA and EMA approval. But PARP inhibitors will have a shot in breast cancer, too. Tesaro and BioMarin are each on the verge of launching Phase III trials in the 5-10% of metastatic breast cancer patients with germline mutations in BRCA1 or BRCA2, a narrow but correct approach, says Domchek. And AbbVie, which has yet to disclose its phase III plans for its candidate veliparib, is positioning the agent for combination therapy, perhaps because of its lower potency.

PARP inhibitors have been in human trials for cancer since 2003, so industry needs to start recouping its investment soon. Other indications can be explored later, says Ledermann. “The most important thing right now is to get the drug licensed,” he says.

In the United States that has become more likely, because the FDA seems prepared to relax its usual insistence on an overall survival benefit, in part because PARP inhibitors are so well tolerated. “The whole issue of whether this is an exciting drug or not doesn’t necessarily have to rest in overall survival,” says Coleman. “It just has to rest in the overall balance between delaying progression and toxicity.”