The safety and efficacy of olaparib therapy in patients with relapsed ovarian cancer.
1Rachel Nirsimloo, 2Charlie Gourley
1Edinburgh Cancer Centre, Western General Hospital, Crewe Road South, Edinburgh, EH4 2XU, UK
2University of Edinburgh Cancer Research UK Centre, MRC IGMM, Crewe Road South, Edinburgh, EH4 2XR, UK
Author for correspondence: Prof Charlie Gourley; tel: +441316518510; fax: +441316518711 ; email:
Summary
PARP inhibition is an exciting new anticancer strategy. Olaparib has recently obtained a first in class license in Europe and the USA for the treatment of relapsed BRCA-mutant ovarian cancer. We review the key preclinical and clinical data surrounding its use in the maintenance setting. We also consider the market profile, regulatory issues surrounding the agent and offer a five year speculative viewpoint of its future development in ovarian cancer.
Keywords
ovarian cancer; PARP inhibitor; olaparib; BRCA; maintenance; synthetic lethality; homologous recombination
Introduction
Epithelial ovarian cancer is the fifth most common female cancer in the UK [101] and the 7th most common worldwide [101]. In 2011 there were 7116 new ovarian cancer diagnoses in the UK and in 2012 there were 4721 deaths, with an estimated 10 year survival (based on data from 2010/11) of only 35%. [101] Over the last decade it has become clear that the histological subtypes of ovarian cancer differ in terms of their driver mutations [1,2] stage of presentation, [3] tissue of origin, [4] clinical outcome [5] and chemosensitivity. [6] The most common histological subtype is high grade serous ovarian cancer (HGSOC) which accounts for approximately 70% of all ovarian cancer, [7] almost always presents with disease that has spread beyond the ovary and is characterised biologically by genomic heterogeneity and p53 mutations. [8] Patients are treated first line with surgery and combination platinum and taxane chemotherapy and despite a high percentage of initial responders the majority of patients relapse with disease that can be re-challenged with variable numbers of lines of platinum-based combination chemotherapy before platinum resistance ultimately develops. [9] Following the development of platinum resistance, single agent chemotherapy with weekly paclitaxel [10,11] or pegylated liposomal doxorubicin (either in combination with bevacizumab if indicated) are the systemic therapies of choice.[12]
In this profile we evaluate the current evidence and give our own opinion on the safety and efficacy of olaparib (AstraZeneca) monotherapy in relapsed ovarian cancer.
Body of review
Despite the clear heterogeneity of epithelial ovarian cancer at the histological and molecular level, until now there have been no licensed therapies that target a specific subset of ovarian cancer patients. Pivotal pre-clinical studies published in Nature in 2005 demonstrated the potential of harnessing the phenomenon known as synthetic lethality, whereby a tumor cell has utilised a particular genetic change during carcinogenesis but that change has made it more susceptible if another different (but functionally related) pathway is targeted. [13,14] The conceptional advantage of this approach is that if tumor cells are targeted on the basis of a genetic abnormality that their development depended upon, then non-tumor cells would be less susceptible to this therapy (as they would not harbor this genetic change), allowing an element of selectivity that has largely been lacking in cancer therapies to date. The examples used in the iconic 2005 studies were of PARP inhibition and its ability to target cancer cells that were homologous recombination (HR) deficient by virtue of having no functional BRCA1 or BRCA2. These cells were 1000 times more sensitive to PARP inhibition than analogous cells that had either one or two proficient copies of the corresponding BRCA1 or BRCA2 gene. [13]
DNA is constantly going through a cycle of damage and repair. PARP1 is an enzyme involved in one of the repair processes of single strand DNA breaks (base excision repair) and is a member of a largergroup of multifunctional (poly(adenosinediphosphate [ADP]-ribose) polymerases.[15] Double stranded DNA (dsDNA) breaks can be repaired by a number of processes, the most accurate of whichis HR. This pathway relies upon a number of proteins including BRCA1, BRCA2, the Fanconi anaemia proteins, ATM, CHEK2, BARD1, BRIP1, MRE11, RAD50, NSB1, RAD51C, RAD51D and PALB2. [16] Approximately 50% of HGSOC have defects in homologous recombination due to germline mutation in BRCA1 or BRCA2 (inherited), somatic mutations in BRCA1 or BRCA2(arising de novo in tumour tissue), epigenetic inactivation of BRCA1(by promoter methylation for example), EMSY amplification or mutational inactivation of other genes involved in homologous recombination listed above. [8] In ovarian (or fallopian tube) cells which harbor heterozygous BRCA1 mutations for example, a ‘second hit’ is then sustained resulting in the lossof the remaining wild type allele and the inability to repair double stranded breaks via HR.[13,14,17] Instead other low-fidelity pathways of dsDNA repair, such as non-homologous end-joining are employed. This results in increased DNA damage, genomic instability and the development of an ovarian (or fallopian tube) cancer.
The consequences of inhibiting the single strand DNA (ssDNA) break repairs using PARP inhibitors in an HR proficient cell (such as non-tumor cells) is that this can generally be tolerated because while many of the ssDNA breaks progress to double strand breaks, these can be accurately repaired by HR. However, in HR deficient cells, the increased burden of ssDNA breaks that results from PARP inhibition, combined with the inability to repair these by HR when they are converted to dsDNA breaks results in overwhelming genomic disorder and cellular death.
While the initial hypothesis that led to the development of PARP inhibitors was based around their properties as catalytic inhibitors of ssDNA repair in a setting where synthetic lethality was likely to operate, it has recently become clear that this is an oversimplification of their mechanism of action. A number of studies have demonstrated that PARP1 deletion in murine or avian cells (the latter lack PARP2) [18] is not as cytotoxic as PARP inhibition in PARP1 wild type cells, suggesting that PARP1 is required for PARP inhibitor-mediated cytotoxicity and the efficacy is not purely attributable to catalytic inhibition. [19,20,21] Murai et al demonstrated that PARP inhibitors trap PARP1 and PARP2 on regions of damaged DNA and that co-transfection of PARP1 siRNA reduced sensitivity to olaparib, again demonstrating the relative importance of PARP1–DNA complex trapping compared to simple PARP inhibition in the mechanism of action. [21] Interestingly this study also showed significant differences in the in vitroantiproliferative activity of the PARP inhibitors niraparib, veliparib and olaparib that correlated with their ability to trap PARP-DNA complexes rather than their potency as catalytic PARP inhibitors (niraparib and olaparib trapped PARP at significantly lower concentrations than veliparib).
Clinical Efficacy
These convincing preclinical data paved the way for phase 1 trials to test Olaparib(4-[3-{[4-cyclopropylcarbonyl)piperazin-1-yl]carbonyl}-4-flurophenyl)methyl]phthalazin-1(2H)-one,a potent orally active inhibitor of PARP1. [15]
Fong et al evaluated 60 patients with histologically/cytologically confirmed advanced tumours, of which 22 of these were BRCA1/2 mutation (BRCA1/2m) carriers. [15] Starting at 10mg once daily for 2 of every 3 weeks they escalated the dose using a modified accelerated-titration design, up to 600mg twice daily. At this dose 600mg twice daily, two of the five patients had dose limiting toxicity (grade 4 thrombocytopenia and grade 3 somnolence) which fully resolved after drug discontinuation. This was the maximum administered dose. The maximum tolerated dose was 400mg twice daily with only 1 out of 8 patients suffering dose limiting toxicity (grade 3 mood alteration and fatigue), which again fully resolved after withdrawing the drug. Grade 1-2 adverse events were nausea in 19 (32%) patients, fatigue in 18 (30%), vomiting in 12 (20%), taste alteration in 8 (13%) and anorexia in 7 (12%). There was a low incidence of haematological complications with anaemia in three patients, grade 4 thrombocytopenia in two patients and no recorded neutropenia. [15]
There were no deaths related to treatment or serious adverse events (SAEs) attributed to the drug. Pharmacokinetic studies showed rapid absorption with peak plasma concentration 1-3 hours after dosing. The drug has a half-life of 5-7 hours and a mean plasma clearance of 4.6l/hr. [15] Inhibition of PARP by more than 90% was measured in patients receiving ≥60mg BD. [15] Pharmacodynamic analysis (measuring formation of ƴH2AX foci after treatment) confirmed that inhibiton of PARP rapidly causes dsDNA breaks and faulty replication as described in preclinical studies. [13]
Clinically meaningful responses to PARP inhibition were only seen in confirmed BRCA1 or BRCA2 mutation carriers in cancer types usually associated with BRCA mutation status. Of 19 evaluable patients in this category, 12 had a tumour marker or radiological response to treatment or stabilisation for >4 months. [15] There were unexpected responses with remission of 60 weeks in a breast cancer patient carrying a deleterious germline BRCA2 mutation who had pulmonary and lymph node metastases and had progressed through chemotherapy. [15] Another patient with a germline BRCA2 mutation and castration-resistant prostate cancer had resolution of his bone metastases and a 50% reduction in PSA level and was still receiving treatment at 58 weeks. [15] This early trial demonstrated olaparib to actively inhibit PARP (proven with biochemical testing), be safe with acceptable side effects and provided early proof of efficacy even with the small numbers studied.
In order to explore the signal of efficacy in a larger patient population this phase I study was expanded to include a cohort of 50 ovarian cancer patients, 48 of whom had a known deleterious germline BRCA1/2 mutation, one who had a germline BRCA2 sequence change of unknown significance and one who had a strong family history of BRCA1/2-associated cancers but who declined mutation testing. [22] Twenty of these patients (40%) had a radiological or tumour marker response to single agent olaparib. A strong association between platinum sensitivity (as judged by preceding platinum-free interval) and olaparib sensitivity was demonstrated.
In 2011, Gelmon et al published data from a phase 2 trial which investigated activity in sporadic and BRCA1/2m high grade serous ovarian and triple negative breast cancer. [23] The response rate in the BRCAm group and BRCA wild type (BRCAwt) groups (ovarian cancer) were 41% and 24% respectively. [23] In these groups platinum sensitivity was an important factor in determining response. Of note the response in BRCAwt (sporadic) patients was only seen in the ovarian cancers, there was no response seen in sporadic triple negative breast cancers. [23]
In 2012 Kaye et al compared the safety and efficacy of two doses of olaparib (200mg BD and 400mg BD) with pegylated liposomal doxorubucin (PLD) 50mg/m2 every 28 days in platinum resistant BRCA1/2 mutation carriers. No significant PFS difference was demonstrated between the treatment groups. [24] Although this was initially interpreted as a negative study, the performance of the PLD arm was greater than expected, a factor that is almost certainly due to the greater efficacy of PLD in BRCA1/2 mutation carriers compared to non-mutation carriers. [25] Some patients with platinum resistant ovarian cancer are suitable for multiple lines of therapy but quality of life remains the priority in this setting. In the PLD arm there were 50% more grade 3 toxicities than in the olaparib arms suggesting olaparib could be a less toxic cost effective alternative to standard chemotherapy. [24]
As a result of the phase 1B data showing a greater response rate to olaparib in platinum sensitive patients [22] and the evidence of efficacy in patients with sporadic (non-BRCAm) high grade serous ovarian cancer [23] it was decided to compare olaparib vs placebo for maintenance therapy in patients with relapsed platinum sensitive high grade serous ovarian cancer. [26] Phase 2 data from this study showed that PFS was greater in the olaparib (400mg BD capsule formation) group than placebo (HR = 0.35, 95% CI 0.25-0.49, median 8.4 vs 4.8 months). It also showed that time to progression was longer in those who received olaparib. [26] At this point median OS appeared the same in both groups, however this was interim data with only 38% maturity. Again toxicity was as described from phase 1 data. Nausea, vomiting, fatigue and anaemia were the main toxicities with the majority grade 1 or 2, with events more common in the olaparib group than placebo. [26]
This data was presented at the American Society of Clinical Oncology (ASCO) in 2011. Shortly after this AstraZeneca decided to discontinue the development of olaparib as (although the data was yet to reach maturity) they did not anticipate a statistically significant improvement in OS.
In a planned retrospective analysis of these data PFS was even greater in those identified with a BRCAm.[27] (HR = 0.18 [95% CI 0.10-0.31]; median 11.2 [95% CI 8.3-not calculable] vs 4.3 months [3.0-5.4] and even showed a statistical benefit in the BRCAwt patients (HR 0.54 [95% CI 0.34-0.85]; median 7.4 [95% CI 5.5-10.3] vs 5.5 months) [27] This again demonstrated that non BRCA-deficient patients can have alternative mechanisms/mutations leading to defects in the HR pathway and that clinical selection (on the basis of having high grade serous ovarian cancer and being sensitive to multiple lines of platinum based chemotherapy) can be used to provide some enrichment for these patients. Fewer patients were randomised to either arm with somatic BRCA mutations (as opposed to germline) and although this wasn’t directly analysed due to the small numbers it appears olaparib may also be effective in this group. [27] At the time of more mature analyses of study 19 (for time to first subsequent therapy (TFST) or death) PFS remained significantly in favour of the olaparib arm. [27] Analyses of time to second subsequent therapy or death (TSST) conferred a significant advantage over the placebo group suggesting benefit of maintenance therapy beyond even first subsequent treatment. [27]All of these significant improvements in outcome in the olaparib arm were more marked in the BRCAmsubgroup than in the intention to treat population. No statistically significant differences in quality of life were observed between the two groups prior to disease progression, measured by performance status and quality of life scores. Olaparib does not therefore appear to impact on quality of life. [27]
At this point OS data had only reached 58% maturity and no statistical benefit was demonstrated in the olaparib group for the overall population. [27] Planned final OS analysis will be conducted at 80% maturity. Crossover to olaparib was not permitted within the study. However other trials did offer access to PARP inhibitors and clinicians could ask for patients to be unblinded making OS data harder to interpret. [27]
The side effect profile remained the same in long term follow up and low numbers stopped the drug due to adverse events, again supporting the assertion that olaparib is a tolerable maintenance treatment. 24% of patients have received >3yrs of olaparib treatment of which 67% have a known BRCA mutation. [27]
AstraZeneca reversed their decision not to continue olaparib development after the post-progression data was presented taking into account BRCA1/2 mutation status at ASCO 2013.
Around the same time as study 19 was being conducted, the question of whether olaparib could be given concomitantly with platinum-based chemotherapy and then continued as a maintenance was addressed by Oza et al. [9] They conducted a phase 2 study in patients with platinum sensitive recurrent high grade serous ovarian cancer who had received up to three previous courses of platinum-based chemotherapy and had not progressed for 6 months. They were randomised to olaparib 200mg BD days 1-10 of each cycle plus paclitaxel (175mg/m(2) day 1) and carboplatin (AUC4 day 1) then maintenance olaparib at 400mg BD until progression vs paclitaxel (175mg/m(2) day 1) and carboplatin (AUC6 day 1) and no maintenance treatment. [9] As hypothesized PFS was greater in the olaparib plus chemotherapy group (12.2 months vs 9.9 months) especially in BRCA germline mutated patients (P=0.0015). [9]
Adverse events in the chemotherapy plus olaparib arm were 10% higher and consisted of alopecia, nausea, neutropenia, diarrhoea, headache, peripheral neuropathy and dyspepsia. However SAEs were higher in the chemotherapy alone arm. [9] Again PFS was greater (particularly in the BRCAm patients) in the olaparib treated group, this is despite a longer median follow up for these patients and also a lower carboplatin dose (which was based on preclinical data).[9] The late separation of the PFS curves does suggest the maintenance phase provides the biggest contribution to PFS. On this basis concomitant therapy is not felt to confer an advantage over maintenance alone. [9]
There was no significant difference in OS although again this data has not yet reached maturity.[9] Again in this study, as in study19, in the patients with BRCA mutations TFST and TSST were greater in the olaparib group suggesting that thebenefit is maintained beyond first progression. [9] In the maintenance phase tolerability of olaparib was the same as previously reported. [9]
Phase 3 trials of olaparib maintenance therapy in patients with germline or somatic BRCA1/2 mutations are currently being conducted in the first line (SOLO1[NCT01844986]) and recurrent platinum sensitive disease (SOLO2[NCT01874353]) settings. The former should address the question of whether olaparib maintenance can improve time to first relapse or even result in a higher percentage of patients who do not relapse and the latter should confirm the findings of study 19 with greater power to demonstrate an OS advantage..
Preclinical data suggested that use of PARP inhibition in platinum sensitive patients may render them insensitive in the future when they have progressed and need re-challenging. To answer this question Ang and colleagues performed a retrospective analysis of patients with germline BRCAm and ovarian cancer who had objective evidence of progressive disease having had olaparib for more than month.[28] They analysed patients over a fixed 6 year period from 8 different hospitals worldwide.[28] Their data showed that patients continue to have the potential to respond to further chemotherapy.[28] Overall objective response rates (ORRs) to first subsequent chemotherapy following PARPi were 36% by RECIST criteria and 40% for platinum based regimes. An increased platinum-platinum interval was associated with an increased OS and likelihood of response following post olaparib platinum. [28] These data support the continued use of olaparib in treating platinum sensitive ovarian cancer patients.
Regulatory Affairs
The European Medicines Agency licensed olaparib as monotherapy for the maintenance treatment of adult patients with platinumsensitive relapsedBRCAmutated (germline and/or somatic) high grade serous epithelial ovarian, fallopian tube, or primary peritoneal cancer who are in response (complete response or partial response) to platinum-based chemotherapy on the 18th of December 2014. This approval was based upon the progression free survival benefit seen in study 19 [26] with the apparent enrichment of effect when patients with germline BRCA mutations were considered. [27]