Epithelial ovarian cancer has the highest mortality rate of all gynaecological malignancies. Most women present with advanced disease and develop a recurrence after radical surgery and chemotherapy. Improving the results of first- or subsequent-line chemotherapy has been slow, and novel approaches to systemic treatment are needed. Ovarian cancer is a heterogeneous disease with complex molecular and genetic changes. Understanding these better will provide information on the mechanisms of resistance and opportunities to target therapy more rationally, exploiting specific changes in the tumour. Here we reviewed targeted approaches to therapy, focussing on targeting angiogenesis and inhibition of DNA repair, 2 areas that show promising activity. Additionally, we reviewed studies that are underway, targeting the cell cycle, signalling pathways and immunotherapeutic strategies. Many of these innovative approaches already demonstrate promising activity in ovarian cancer and have the potential to improve the outcome in women with ovarian cancer.
Introduction
Epithelial ovarian cancer (EOC) is one of the most common gynaecological malignancies, and most women present with advanced disease. It has a high mortality rate and is the fifth most common cause of cancer death in women . Although the survival of patients with advanced ovarian cancer has increased over the last 2 decades through better surgery and more chemotherapy options, cytotoxic drug therapy has been non-selective, often resulting in significant toxicity and short-lived anti-tumour responses. Most women with ovarian cancer will suffer a tumour recurrence after the first-line therapy, and in almost all of them, resistance to chemotherapy will eventually develop, leading to death from ovarian cancer. There has been a significant increase in the knowledge of molecular and genetic changes in ovarian cancer, and this has led to the development and evaluation of targeted therapies in this disease. Here we review how these new drugs are being used and investigated in women with ovarian cancer and their contribution thus far to improving the response to therapy and disease outcome.
Angiogenesis
Solid tumour growth and progression is reliant on neovascularisation . Angiogenesis is complex and regulated by several different endogenous pro-angiogenic and anti-angiogenic factors. Key angiogenic molecules include vascular endothelial growth factor (VEGF), platelet-derived growth factor (PDGF), fibroblast growth factor (FGF) and angiopoietin (Ang 1 and 2) . The Ang 1/2–Tie 2 receptor axis is a VEGF-independent signalling pathway that mediates vascular re-modelling . Folkman et al proposed that a fine balance exists between angiogenic inhibitory and stimulatory factors. In normal tissues, angiogenesis is turned off. However, an ‘ angiogenic switch’ can occur in tumours that leads to the production of pro-angiogenic stimuli and causes growth of the tumour and its vasculature . These new blood vessels often have defective basement membranes and are thin walled and leaky, allowing cancer cells to enter the circulation and metastasise .
In EOC, angiogenesis plays a role in tumour growth, formation of ascites and metastasis. The vasculature within the tumour is more structurally and functionally abnormal, with tortuous, leaky, dilated and immature blood vessels with poor flow. The endothelial cells within these vessels are more dependent on VEGF for survival than more mature blood vessels elsewhere in the body .
The VEGF family of growth factors and its receptors are the most important signalling pathways in tumour angiogenesis . VEGF-A is the best-characterised VEGF ligand and appears to play a dominant role in angiogenesis by binding to VEGF receptor tyrosine kinases (VEGFR). In angiogenesis, the 2 most important members of the VEGFR family are VEGFR1 (Flt-1) and VEGFR2 (Flk-1). VEGFR2 has the most direct effect on angiogenesis as it mediates the angiogenic and permeability enhancing effects of VEGF. The role of VEGFR1 is less direct, and it may play a role in angiogenesis by recruiting bone marrow-derived cells and monocytes into the tumour vasculature .
Angiogenesis
Solid tumour growth and progression is reliant on neovascularisation . Angiogenesis is complex and regulated by several different endogenous pro-angiogenic and anti-angiogenic factors. Key angiogenic molecules include vascular endothelial growth factor (VEGF), platelet-derived growth factor (PDGF), fibroblast growth factor (FGF) and angiopoietin (Ang 1 and 2) . The Ang 1/2–Tie 2 receptor axis is a VEGF-independent signalling pathway that mediates vascular re-modelling . Folkman et al proposed that a fine balance exists between angiogenic inhibitory and stimulatory factors. In normal tissues, angiogenesis is turned off. However, an ‘ angiogenic switch’ can occur in tumours that leads to the production of pro-angiogenic stimuli and causes growth of the tumour and its vasculature . These new blood vessels often have defective basement membranes and are thin walled and leaky, allowing cancer cells to enter the circulation and metastasise .
In EOC, angiogenesis plays a role in tumour growth, formation of ascites and metastasis. The vasculature within the tumour is more structurally and functionally abnormal, with tortuous, leaky, dilated and immature blood vessels with poor flow. The endothelial cells within these vessels are more dependent on VEGF for survival than more mature blood vessels elsewhere in the body .
The VEGF family of growth factors and its receptors are the most important signalling pathways in tumour angiogenesis . VEGF-A is the best-characterised VEGF ligand and appears to play a dominant role in angiogenesis by binding to VEGF receptor tyrosine kinases (VEGFR). In angiogenesis, the 2 most important members of the VEGFR family are VEGFR1 (Flt-1) and VEGFR2 (Flk-1). VEGFR2 has the most direct effect on angiogenesis as it mediates the angiogenic and permeability enhancing effects of VEGF. The role of VEGFR1 is less direct, and it may play a role in angiogenesis by recruiting bone marrow-derived cells and monocytes into the tumour vasculature .
Vascular endothelial growth factor inhibition
Inhibition of VEGF restores the balance between pro-angiogenic and anti-angiogenic factors, thereby normalising tumour blood vessel structure and function . This influences tumour growth and is believed to improve the delivery of chemotherapy drugs to tumours and decrease their metastatic potential.
There is a high expression of VEGFR in ovarian cancer, and many tumours produce high levels of VEGF. Inhibitors of VEGF signalling such as the recombinant monoclonal antibody bevacizumab that binds to circulating VEGF-A and aflibercept, a fusion protein that binds directly to VEGF and prevents it from binding to its receptors, have been extensively evaluated in ovarian cancer treatment. Another strategy that has been explored is to use a peptide-Fc fusion protein, trebananib, which binds to Ang 1 and Ang 2, thus preventing its interaction with the Tie 2 receptor .
Bevacizumab
In ovarian cancer, bevacizumab has been explored as a single agent, in combination with chemotherapy and as maintenance treatment post chemotherapy. It has been studied extensively in the setting of first-line platinum-based chemotherapy, ‘platinum-sensitive’ disease (relapse > 6 months following completion of platinum-based treatment) and ‘platinum-resistant’ disease (relapse ≤ 6 months of completing platinum-based treatment). For a summary of the pivotal trials described below, see Table 1 .
| Study | Population | Treatment | Median PFS (months) | Median OS (months) |
|---|---|---|---|---|
| ICON7 | First-line high risk stage I–IIA or stage IIB–IV ovarian cancer (n=1528) | CP vs CP + bevacizumab (7.5 mg/kg) followed by maintenance bevacizumab | 22.4 vs 24.1 (HR 0.87; p=0.04) | 58.6 vs 58.0 (p=0.85) high risk subgroup: 30.2 vs 39.7 (p=0.03) |
| GOG 218 | First-line stage III (incompletely resected) and stage IV ovarian cancer (n=1873) | CP + PL vs CP + bevacizumab (15 mg/kg) + maintenance PL vs CP + bevacizumab (15 mg/kg) + maintenance bevacizumab | 10.3 vs 11.2 vs 14.1 (HR 0.72; p<.001) | 39.3 vs 38.7 vs 39.7 (HR 0.95; p=0.45) |
| OCEANS | PSROC (n=484) | GC + PL vs GC + bevacizumab (15 mg/kg) | 8.4 vs 12.4 (HR 0.48; p<.0001) | 32.9 vs 33.6 (HR 0.95; p=0.65) |
| GOG 0213 | PSROC (n=674) | CP vs CP + bevacizumab (15 mg/kg) | 10.4 vs 13.8 (HR 0.61; p<.0001) | 37.7 vs 42.2 (HR 0.83; p=0.056) |
| AURELIA | PRROC (n= 361) | CT vs CT + bevacizumab (10 mg/kg q2/52 or 15 mg/kg q3/52) | 3.4 vs 6.7 (HR 0.48; p<.001) | 13.3 vs 16.6 (HR 0.85; p<.174) |
ICON7 and GOG 218 are the 2 key first-line studies investigating the addition of bevacizumab to chemotherapy following surgery for advanced ovarian cancer. In ICON7, patients were randomised to standard 6 cycles of 3-weekly carboplatin and paclitaxel with or without intravenous bevacizumab 7.5 mg/kg every 3 weeks followed by maintenance bevacizumab for up to 12 months. The median progression-free survival (PFS) was 17.5 months with standard therapy alone vs 19.9 months in the bevacizumab arm (HR 0.87; p=0.04). The benefit from bevacizumab was greater in women at higher risk of progression because of incomplete cytoreductive surgery (≥1 cm residual disease) or FIGO stage IV disease. In this group, the PFS was 14.5 months compared to 18.1 months in women receiving bevacizumab. No difference was seen in the median overall survival (OS), which was 58 months. However, in the higher risk sub-group Table 1 , there was a difference of 9.5 months in the median OS [30.2 vs 39.7 months for women receiving standard therapy vs bevacizumab (p=0.03)] . The second trial GOG 218 was a US-led 3-arm randomised placebo-controlled study. Each arm received 6 cycles of standard 3-weekly carboplatin and paclitaxel. Arm 1 received chemotherapy with placebo and placebo maintenance for 15 months; arm 2 received chemotherapy plus bevacizumab (15 mg/kg) from cycles 2–6, followed by placebo maintenance as above; and women in arm 3 received chemotherapy plus bevacizumab from cycles 2–6, followed by bevacizumab maintenance. Compared to the control (arm 1), a significant improvement in PFS was seen only in women receiving bevacizumab with chemotherapy and as maintenance (arm 3). The difference in median PFS in this group compared to placebo was 3.9 months, and there was no difference in OS ( Table 1 for details) .
In women with ‘platinum-sensitive’ recurrent disease, the OCEANS trial evaluated the addition of bevacizumab to chemotherapy and then as maintenance post chemotherapy until disease progression. Patients were randomised to 3-weekly chemotherapy (gemcitabine 1000 mg/m 2 D1,8 and carboplatin AUC 4) with bevacizumab 15 mg/kg or placebo. A 4-month improvement in median PFS was seen, which was statistically significant, but there was no OS benefit . Similar results were seen in GOG 0213 that compared 3-weekly carboplatin and paclitaxel with or without bevacizumab. In this trial, there was a 3.4-month significant difference in the median PFS. Preliminary results reported a trend towards an improvement in median OS, and the benefit was unaffected by prior use of bevacizumab .
In women with ‘platinum-resistant’ relapse, the phase III randomised AURELIA trial evaluated the addition of bevacizumab to a choice of chemotherapy regimens [weekly paclitaxel, pegylated liposomal doxorubicin (PLD) or topotecan, given either over 5 days or weekly). Overall, there was a 3.3-month improvement in median PFS when bevacizumab was added to chemotherapy, which was significant ( Table 1 ). Bevacizumab was not continued as maintenance therapy, but the benefit in PFS and quality of life was seen consistently across all sub-groups, particularly in patients with ascites, which is associated with a poorer prognosis . The greatest difference in PFS was seen in patients receiving weekly paclitaxel . There was no statistically significant improvement in OS; however, the trial was not designed to show a difference in OS, and 40% of patients receiving chemotherapy alone were subsequently treated with bevacizumab.
Toxicities of VEGF inhibitors include hypertension, impaired wound healing, proteinuria, increase risk of thromboembolism and gastrointestinal toxicities. This includes the rare but serious complications of perforation and fistulae, and before a decision to use bevacizumab is made, one needs to take note of the volume of serosal disease, particularly thickening of the sigmoid colon because of tumour as the risk of perforation is greater in patients with large amounts of serosal disease.
Despite evidence of activity in different phases of the treatment pathway, there is no international consensus about when the value of bevacizumab therapy is greatest. This is reflected in the license, with indications varying across the world.
Aflibercept
Intravenous aflibercept has been evaluated in recurrent ovarian cancer both as a single agent and in combination with chemotherapy. In ‘platinum-resistant’ disease, the response rate to single-agent aflibercept is <5% , although in combination with docetaxel, the response rate has been reported to be 54%. However, these patients were not as heavily pre-treated as those enrolled in the single-agent phase II study, and 13/46 (28%) patients had ‘platinum-sensitive’ disease . Aflibercept prolongs the time to repeat paracentesis in women with recurrent symptomatic ascites, compared to placebo, with a median time to repeat paracentesis of 55.1 days and 31.8 days, respectively . However, the risk of fatal bowel perforation is increased with aflibercept, and the risk-benefit balance needs to be carefully considered .
Trebananib
Trebananib has been evaluated in a series of trials: TRINOVA-1, a randomised phase III trial combining trebananib with weekly paclitaxel in 919 women with recurrent ovarian cancer relapsing after a platinum-free interval of <12 months. Compared to placebo, trebananib significantly prolonged the median PFS (5.4 vs 7.2 months respectively; HR 0.66; p<0.0001) but no OS advantage was seen (18.3 vs 19.3 months, respectively; HR 0.95; p=0.55). The main toxicity observed was oedema . Results from 2 other phase III studies of trebananib in ovarian cancer are pending: TRINOVA-2 (NCT01281254), a study of PLD with or without trebananib also in recurrent ovarian cancer, and TRINOVA-3 (NCT01493505), a study in which trebananib was combined with first-line chemotherapy.
Multi-targeted anti-angiogenic agents
Tyrosine kinase inhibitors are multi-targeted, low-molecular-weight drugs that bind to the ATP-binding catalytic site of the tyrosine kinase domains of VEGF-R and other tyrosine kinases. These oral agents often target more than one receptor tyrosine kinase . Examples include sorafenib, cediranib, pazopanib and nintedinib, all of which have been shown to have activity in ovarian cancer . It is attractive to use a drug that targets more than one pathway as this could lead to a greater inhibition of angiogenesis than single pathway inhibitors such as bevacizumab. However, although the therapeutic response may be greater, their more complex mechanism of action may increase toxicity ( Table 2 ).
| Agent | Trial | Line of treatment | PFS (months) |
|---|---|---|---|
| Sorafenib | Randomised phase II CP ± sorafenib followed by maintenance sorafenib for a total of 1 year | First line | 15.4 vs 16.3 (p=0.38) |
| Pazopanib | Randomised Phase II: MITO 11 Pazopanib + paclitaxel vs paclitaxel alone | PRROC | 6.3 vs 3.4 |
| Phase III maintenance study: AGO-OVAR-16 Pazopanib vs placebo maintenance | First line | 17.9 vs 12.3 HR 0.77; p=0.0021 | |
| Nintedanib | Phase III: AGO-OVAR-12 CP ± nintedanib followed by nintedinib or placebo maintenance | First line | 17.2 vs 16.6 HR 0.84; p=0.024 |
| Cediranib | Phase III: ICON 6 3-arm study: A: CT with placebo vs B: CT + cediranib followed by placebo maintenance vs C: CT + cediranib followed by cediranib maintenance | First PSROC | 11.0 vs 8.7 HR 0.56; p<0.0001 (A vs C primary analysis) |
Pazopanib
Pazopanib targets several angiogenic receptors including downstream signals of VEGFR-1, 2 and 3; PDGFR; and c-kit . Pazopanib has been evaluated in recurrent ovarian cancer and as maintenance therapy in the first-line setting. In MITO11 , an open-label phase II trial, patients with ‘platinum-resistant’ ovarian cancer were randomised to weekly paclitaxel with or without pazopanib. There was a 2.9-month improvement in PFS in the combination arm (median 6.3 vs 3.4 months) that was statistically significant. In the phase III study AGO-OVAR-16, patients who did not have tumour progression at the end of the first-line therapy, were randomised to maintenance pazopanib or placebo for up to 24 months. There was an improvement of 5.6 months in the median PFS in patients on pazopanib compared to placebo; however, this did not translate into an OS benefit, and there were much higher grade 3/4 adverse event rates in patients on pazopanib, for example, hypertension, neutropenia, liver toxicity, diarrhoea, fatigue, thrombocytopenia and palmar-plantar erythema. The dose of pazopanib was reduced in 58% of the patients, and 33.3% in the treatment discontinued treatment secondary to toxicity compared to 5.6% in the placebo arm . Consequently, the drug has not been submitted for market authorisation.
Nintedanib
Nintedanib (BIBF 1120) is a triple angiokinase inhibitor that inhibits VEGFR, FGFR and PDGFR, which all contribute to tumour angiogenesis . Activity of the drug in a randomised phase II trial led to a phase III study of this drug as first-line therapy.
In the AGO-OVAR 12, nintedinib or placebo was combined with carboplatin and paclitaxel and then continued as maintenance for up to 120 weeks. There was a 1.2-month difference in median PFS, significantly longer in patients receiving nintedanib compared with placebo, but this small difference was believed to be insufficient for further development in this indication . Hypertension is uncommon with this drug, and most of the adverse events were gastrointestinal, with 21% of patients in the nintedanib arm experiencing ≥grade 3 diarrhoea . Trials continue in patients with ‘platinum-resistant’ ovarian cancer and in patients with clear cell tumours of the ovary or endometrium.
Cediranib
Cediranib, an oral inhibitor of VEGFR-1, 2 and 3, is also active in recurrent ovarian cancer. The magnitude of the response of this drug was modest but sufficient for exploration in a phase III trial . Cediranib was evaluated in ICON6, an academic-led randomised, 3-arm, double-blind placebo-controlled trial in ‘platinum-sensitive’ ovarian cancer. The design allowed the comparison of the effect of cediranib with chemotherapy and as maintenance therapy . The original trial design was changed when the manufacturer temporarily ceased the production of the drug, and the main comparison using a smaller number of patients was between chemotherapy and placebo (Group A, Table 2 ) and chemotherapy with concurrent cediranib followed by maintenance cediranib (Group C). There was a significant improvement in median PFS in patients receiving cediranib, from 8.7 to 11 months (HR 0.56; p<0.0001). OS data were immature at the time of publication. Fatigue, diarrhoea and hypertension were the most significant toxicities and 32% of the patients discontinued cediranib during treatment due to side effects. Toxicity was most evident when cediranib was given with chemotherapy . A submission for market authorisation in patients, based on the ICON6 results was recently withdrawn by the manufacturer. However, cediranib is an active drug and continues to be evaluated in ovarian cancer (see below).
Targeting the folate receptor
The folate cycle maintains essential metabolic reactions required for rapidly growing cells. Once folates enter a cell, they have a crucial role in the biosynthesis of purines and thymidine, which are required for DNA synthesis, repair and methylation . The alpha isoform of the FR (αFR) transports folates by receptor-mediated endocytosis and is selectively overexpressed in a number of solid tumours, including non-mucinous ovarian cancers .
Farletuzumab (MORAb-003), a humanised monoclonal antibody to αFR leads to cell-mediated cytotoxicity, complement-dependent killing and non-immune-mediated αFR-dependent inhibition of growth under folate limiting conditions . An initial phase II study assessing at the addition of farletuzumab to chemotherapy followed by maintenance therapy in ‘platinum-sensitive’ ovarian cancer showed promising activity with an improved overall response rate compared to historical controls . However, the benefit was not confirmed in a subsequent phase III trial .
In a different folate-targeting strategy, vintafolide (EC145), a folate-conjugated vinca alkaloid that targets FR-expressing cells, was explored in a randomised phase II trial in ‘platinum-resistant’ ovarian cancer. In the PRECEDENT trial, patients were randomised to PLD alone or in combination with vintafolide . The study met its primary end point, demonstrating a 2.3-month improvement in median PFS in the experimental arm (5.0 vs 2.7 months; HR 0.63; p=.031]), but the subsequent randomised phase III trial (PROCEED) was stopped early because a pre-specified interim analysis showed no benefit.
Homologous recombination deficiency and ovarian cancer
Normal cellular function and genomic stability is maintained by recognition and subsequent repair of DNA damage that occurs in all cells . There are 450 known genes implicated in DNA damage response and repair , and they are sub-divided into 5 distinct pathways that are responsible for repair of specific types of DNA damage . Double-stranded DNA breaks (DSB) are the most lethal insult to the genome, and if left unrepaired, genomic instability and cell death will occur . They can be repaired by several different pathways, the main one being homologous recombination repair (HRR). A large proportion of DSBs arise during DNA replication when a replication fork encounters an unrepaired single strand break. The HRR pathway, together with PARP1, a nuclear enzyme, is especially important in repairing these collapsed replications forks . If HRR is defective (homologous recombination deficiency; HRD), cells are dependent on alternative pathways for repair. These are significantly impaired by the inhibition of PARP enzymes. BRCA1 and BRCA2 (Breast Cancer Associated) proteins play an important role in HRR, and when mutated, cells have HRD (HR deficiency) that can be selectively exploited by PARP inhibitors (PARPi), leading to unrepaired or poorly repaired DNA and tumour synthetic lethality .
Ovarian cancer and PARP inhibitors
Germline mutations of BRCA1 and BRCA2 genes are present in about 15% of high-grade serous ovarian cancer (HGSOC), the most common histological sub-type. Although this is the most common cause of HRD, somatic mutations of the BRCA genes are found in approximately 5% of these tumours and in addition mutations of other key HRR genes, such as RAD51. It has been estimated that approximately 50% of HGSOC tumours have HRD , sometimes referred to as ‘BRCAness’ . Targeting HRD with PARPi in both BRCA mut (BRCA mutated) and BRCA wt (BRCA wild-type) ovarian cancer is an important therapeutic approach that is being developed in high-grade ovarian cancers. Several oral PARPi are being developed and are summarised in Table 3 .
| Agent | Trial | Population | Line of treatment |
|---|---|---|---|
| Olaparib (Astrazeneca) | SOLO-1 (NCT01844986): Phase III maintenance (olaparib vs placebo) | BRCA mut ; AOC | First line |
| SOLO-2 (NCT01874353): Phase III maintenance (olaparib vs placebo) | BRCA mut ; PSROC | After ≥2 lines platinum based CT | |
| SOLO-3 (NCT02282020): Phase III olaparib vs physician choice CT (standard of care non-platinum based) | BRCA mut ; PSROC | After ≥2 lines of platinum based CT | |
| Rucaparib (Clovis) | ARIEL 2 (part 2) (NCT01891344): Single arm study | HGROC | Received ≥2 prior lines of CT |
| ARIEL 3 (NCT01968213): Phase III maintenance (rucaparib vs placebo) | HGROC | After ≥2 lines of platinum based CT | |
| ARIEL 4 (NCT02855944): Phase III rucaparib vs chemotherapy | BRCA mut ; ROC | Received ≥2 lines prior CT | |
| Niraparib (Tesaro) | QUADRA study (NCT02354586): Single arm phase II study | HGROC | Must have had 3/4 prior lines of CT |
| PRIMA study (NCT02655016): Phase III maintenance (niraparib vs placebo) | HRD+ve; AOC | First line | |
| AVANOVA study (NCT02354131): Phase I/II niraparib ± bevacizumab | HRD+ve; PSROC | No limits on number of previous lines of CT | |
| Veliparib (AbbVie) | Phase III combination CT followed by maintenance (NCT02470585) | BRCA mut ; AOC | First line |
| Talazoparib (Medivation/Pfizer) | Phase I/II study (NCT01989546) | BRCA mut solid tumours | After ≥1 standard CT or no standard Rx options |
| Phase II study (NCT02326844) | gBRCA mut ; ROC | Progression following a PARPi |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
