VEGF pathway inhibition potentiates PARP inhibitor efficacy in ovarian cancer independent of BRCA status

To the Editor,

Olaparib is a first-in-class poly ADP-ribose polymerase inhibitor (PARPi) for treatment of recurrent platinum-responding ovarian cancer (OC) [1] with deleterious mutations in BRCA and other homologous recombination repair (HRR) pathway components which sensitize tumors to PARPi [2]. To increase the clinical therapeutic activity of PARPi, combination potential with vascular endothelial growth factor (VEGF) inhibition has been tested in multiple trials [2, 3]. This has resulted in olaparib plus bevacizumab being approved for maintenance treatment of OC in homologous recombination deficiency (HRD)-positive tumors [4]. Combining olaparib and cediranib, a small molecule receptor tyrosine kinase (RTK) inhibitor targeting VEGFR1, VEGFR2, VEGFR3 and mast/stem cell growth factor receptor c-KIT [5], provided a survival benefit in patients with relapsed platinum-sensitive OC, including BRCA wild-type tumours [6]. Our preclinical study assessed the benefit of combining VEGF signalling and PARP inhibition across a broad panel of patient-derived ovarian cancer xenografts (OC-PDXs). Cediranib potentiated the antitumor activity of olaparib regardless of the tumour’s sensitivity to platinum and olaparib and BRCA1/2 mutational status.

A cohort of subcutaneous OC-PDXs representing diverse OC genotypes and phenotypes (Additional file 1: Figs. S1–S3) [7] were treated with olaparib [1] and cediranib [5], mirroring clinical exposure [8]. Cediranib enhanced olaparib antitumor activity in all OC-PDXs, regardless of tumour histotype, HRR mutational status, expression of cediranib-targeted RTKs, sensitivity to platinum or olaparib monotherapy. As shown in Fig. 1A, the combination therapy showed broader activity versus olaparib monotherapy (81% vs 50%; P < 0.0001; odds ratio of 3.9) across the OC-PDX panel, with greater antitumor benefit in tumours refractory to platinum and olaparib. Tumour control seen at 4 weeks (Fig. 1A) was maintained at 8 weeks (Additional file 1: Fig. S4) with no increased toxicity. Notably, the combination achieved a durable growth stabilization of the platinum/olaparib poorly responsive MNHOC182 and MNHOC18 tumours, which bear heterozygous mutations in BRCA2 (Fig. 1B and Additional file 1: Fig. S2). In BRCA-mutated and platinum/olaparib-sensitive OC-PDXs, the combination achieved rapid and prolonged tumour regression and this was evident also in tumours that no longer responded to olaparib at rechallenge (i.e. MNHOC513, MNHOC511, MNHOC508, Additional file 1: Fig. S6).

Extensive analysis was performed to identify potential drivers of combination efficacy. Consistent with its mode of action [9], cediranib dosed alone and in combination with olaparib significantly reduced OC-PDX tumour-associated vasculature (Fig. 1D and Additional file 1: Figs. S5, S7). Previous studies have suggested that cediranib-induced hypoxia sensitizes tumour cells to PARPi by reducing HRR gene expression [10], and that cediranib inhibits BRCA expression through platelet-derived growth factor receptor (PDGFR) inhibition, thus inducing a contextual tumour BRCAness [11]. Our analysis suggests that downregulation of HRR genes or modulation of PDGFRβ signalling is not a dominant effect underpinning the combination benefit. Specifically, under severe hypoxic conditions in vitro, we did not observe any consistent change in RAD51 expression across a panel of OC cell lines (Additional file 1: Fig. S8). Moreover, downregulation of BRCA or RAD51 in vitro was only achieved at ≥ 3 μM concentrations of cediranib (Additional file 1: Fig. S9A,B), far higher than the clinical exposure. In contrast to other studies BRCA or RAD51 modulation was not prevented by silencing of E2F4 (Additional file 1: Fig. S9B), the transcription factors suggested to regulate expression via PDGFRβ [11]. Consistent with this, HRR gene expression was not modulated in the OV2022 PDX by either short or long term dosing, despite increased combination benefit (Additional file 1: Fig. S10). Likewise in OC-PDXs, no common changes in the expression of a broad panel of genes (including HRR ones) by cediranib or the combination were found (Fig. 1C and Additional file 1: Figs. S5, S11), nor there was any association between RTK basal expression (Additional file 1: Fig. S3A) and sensitivity to treatments at the clinically relevant dose of cediranib (Figs. 1, 2). Indeed, cediranib addition to olaparib significantly impaired the growth of MNHOC182 tumors (Fig. 1B) which have low/null PDGFRB expression (Fig. 1C). Pharmacokinetic modeling of the cediranib exposure profile in patients showed that robust inhibition of PDGFR is not achieved at clinically relevant doses [12] and consistent with our findings, olaparib efficacy is also enhanced by bevacizumab [2], which does not inhibit PDGFR, further supporting the conclusion that targeting PDGFR signalling is not a driver of combination benefit.

The benefit of the combination in controlling tumour progression was confirmed in three disease relevant orthotopic OC-PDXs (Fig. 2) that mimic human disease with dissemination occurring in the mouse peritoneal cavity [7]. Cediranib addition prolonged the lifespan of mice bearing the platinum/olaparib-responsive MNHOC8 (Fig. 2A) lacking BRCA1 expression (Additional file 1: Fig. S3) and most importantly the olaparib-resistant MNHOC506 and MNHOC22 (Fig. 2B, C). Analysis at 4 week of treatment, highlighted the need of both drugs controlling ascites formation and tumour dissemination (Fig. 2, right panels), resulting in survival increase. Tumour regrowth at therapy discontinuation (Additional file 1: Fig. S12) suggests that sustained treatment is required to maximise prolonged tumour control. Collectively these data suggest that PARPi and VEGFi target complementary mechanisms to reduce tumour and that continuous therapy may be clinically important to restrain tumour progression.

In summary, we show that the addition of cediranib enhances the antitumor efficacy of olaparib in both olaparib-sensitive and olaparib-resistant OCs. Pre-clinical activity was not associated with a mechanism priming response to cediranib or olaparib. Rather the agents appear to be complementary, with olaparib impacting DNA repair and cediranib modulating tumour vasculature, to increase response across the entire OC-PDXs panel. Our extensive experimental in vivo efficacy data reproduce clinical observations, while our comprehensive mechanistic analysis suggests the combination effect is largely driven by targeting independent mechanisms, rather than inducing specific changes that prime response to a partner drug.