Cisplatin was adopted as primary chemotherapy schedule in the 1970s in association with cyclophosphamide [5]. In the 1990s a microtubule stabilizer, paclitaxel, was shown to potentiate cisplatin-based therapy in ovarian cancer patients with better efficacy than cisplatin-cyclophosphamide [8,9]. Since these clinical trials, only minor variations in the standard chemotherapeutic schedule for ovarian cancer patients have been implemented. For instance, in the early 2000s it was demonstrated that carboplatin was equality effective as cisplatin in association with paclitaxel but with much less nephrotoxicity, and better tolerability and quality of life than cisplatin [10]. Thus, cisplatin and paclitaxel, and later carboplatin plus paclitaxel, have been broadly accepted as first-line chemotherapy for advanced epithelial ovarian cancer. Another improvement in overall survival was achieved by the adoption of intraperitoneal delivery of these drugs when compared with intravenous administration [11]. However, data worldwide concur that in the past 20years there has been little change in the 5-year survival rates post-diagnosis of patients with ovarian cancer [1].

Our laboratory provided evidence that the synthetic steroid mifepristone is effective as a single agent in vitro and in vivo blocking the growth of human epithelial ovarian cancer cells [12]. When used at concentrations likely to be achieved in vivo in humans [13-16], mifepristone inhibited cell growth by inducing G1 cell cycle arrest associated with inhibition of DNA synthesis, downregulation of the transcription factor E2F1 needed for S phase progression, and inhibition of the activity of cyclin dependent kinase 2 [12,17], which is critical to promote G1 to S phase transition [18]. We also reported that the growth inhibitory effect of mifepristone in ovarian cancer cells does not require the expression of cognate progesterone receptors [19], and is independent of p53 functionality and platinum sensitivity [20], making mifepristone an even more interesting chemotherapeutic candidate for ovarian cancer as the majority of tumors in relapsing patients are platinum resistant and p53 mutant [7]. Finally, we have shown in ovarian cancer cells that mifepristone potentiates the lethality of otherwise sub-lethal doses of cisplatin, and synergizes with cisplatin growth inhibiting ovarian cancer cells of different genetic backgrounds and platinum sensitivities [21].

We modelled in vitro a scenario in which we utilized supra-pharmacological doses of cisplatin and paclitaxel [25,26] to maximize the cytotoxicity of the drugs, but limited their effect to clinically relevant times. Thus, cells were exposed to cisplatin in the range of 0–20μM for only 1h, to paclitaxel in the range of 0–100 nM for only 3h, and chronically to 10μM mifepristone, likely to be maintained in vivo[13-16]. Using this in vitro approach, we demonstrate that although the combination cisplatin-paclitaxel used at high doses is very efficacious in killing a large proportion of ovarian cancer cells encompassing broad genetic backgrounds, there are cells that escape the therapy and repopulate the culture following an otherwise apparent successful chemotherapeutic round.

The initial efficacy of the therapy with cisplatin and paclitaxel for ovarian cancer was amply validated and translated to the clinic many years ago [8,9]. However, the molecular mechanisms whereby cisplatin and paclitaxel alone or in combination cause cell toxicity necessitate further elucidation. Cisplatin displays cytotoxicity targeting the cytoplasm and the nucleus. In the cytoplasm, cisplatin interacts with a wide number of substrates tilting the redox balance to oxidative stress, which facilitates DNA damage [41]. Cisplatin also causes direct mitochondrial dysfunction [42] and endoplasmic reticulum stress [43]. In the nucleus, cisplatin binds DNA leading to the generation of DNA-DNA inter- and intra-strand adducts. These lesions cause distortions in the DNA that can be recognized by multiple repair pathways. When the extent of damage is limited, cisplatin adducts induce an arrest in the S and G2 phases of the cell cycle to allow the DNA repair mechanisms to re-establish DNA integrity and prevent abortive or abnormal mitosis. In contrast, if DNA damage is beyond repair, cells embark into a delayed death pathway [41,44-49]. Paclitaxel, on the other hand, acts by binding to intracellular β-tubulin leading to microtubule stabilization and G2-M arrest; thereafter, the cells may either die by apoptosis or necrosis immediately after the mitotic arrest, or following an aberrant mitotic exit into a G1-like multinucleated state [50]. Paclitaxel-induced cell death is, at least in part, mediated through the degradation of Bcl-2 [51]. In contrast to cisplatin, sensitivity to paclitaxel is independent of the p53 tissue expression status [52,53], while cisplatin-resistant cells retain sensitivity to paclitaxel [27,36]. There is no doubt of the initial efficacy of the combination cisplatin-paclitaxel; their synergistic pharmacological interaction has been amply proven [28,29]. In our work we substantiated that adding mifepristone after cisplatin-paclitaxel does not interfere with the efficacy of the dual combination, but rather, in some cases enhances it.

Dual cisplatin-paclitaxel therapy is followed by disease remission; however, it rarely provides cure as the disease eventually relapses after remaining dormant as minimal residual for over a year [6]. Such therapeutic failure is recapitulated in our in vitro toxicity model system in which scarce, yet critical cells escape cisplatin-paclitaxel therapy and regrow as demonstrated by their increase in number and by having a clonogenic survival capacity even superior than that of cells never treated with cisplatin-paclitaxel. When mifepristone was added after the initial toxic cisplatin-paclitaxel combination, repopulation was prevented and the clonogenic capacity of the remaining cells in culture was minimal.

The repopulation of cancer cells that escape or relapse after chemo- or radiotherapy accounts for the lack of long-term success of current cancer therapy. Repopulation of tumor cells was defined in 2005 by Kim and Tannock [54] as ‘the continuing proliferation, sometimes at an accelerated pace, of surviving tumor cells with the capacity to regenerate the tumor that can occur during a course of chemotherapy or fractionated radiotherapy.’ The mechanism of repopulation of escape cells, however, is less understood: some recent studies provide insights into it. For instance, cancer cells develop the capacity to escape DNA damage caused by pharmacological doses of platinum-based therapy via reverse polyploidy (a.k.a ‘neosis’ [55-57]), leading to the formation of diploid, rapidly proliferating cells with increased platinum resistance [58]. Such diploid descendants are formed upon reactivation of meiosis-specific genes from a polyploid genome [59,60], and in association with the formation of sub-nuclei that become degraded by autophagy [61]. Thus, a possibility exists that mifepristone blocks repopulation of escape cells by preventing reverse polyploidy. Supporting this hypothesis, we observed that OV2008 cultures that retain certain viability in between courses of cisplatin exposure, show giant cells together with a nascent population of small cells [62] that may originate from the likely polyploid, giant progenitors. Cultures treated with mifepristone after cisplatin do not show this small pool of repopulating cells; instead, they display an overall reduced number of cells, with predominance of a giant phenotype that ends up committing suicide as marked by cleaved PARP positivity [62]. Similar mechanism may be taking place in cells repopulating after cisplatin-paclitaxel combination therapy, because it is known that the driving force behind this therapy is cisplatin, not paclitaxel [44]. Additionally, within cultures repopulating after cisplatin-paclitaxel we observed a population of cells with hyperploid DNA content that is reduced parallel to cell repopulation and increased percentage of G1 cells. In cultured treated with cisplatin-paclitaxel plus mifepristone, however, such hyperploid population disappears in favor of hypodiploid DNA content consisting with cells dying by apoptosis instead of returning to the cell cycle.

Two other survival mechanisms that may explain repopulation after escape to chemotherapy could be the target of mifepristone interference. Firstly, mifepristone can block the release of survival factors from dying cells, because a recent study demonstrated that cells that are dying as a consequence of the chemotherapy, release chemical mediators (prostaglandins) that promote the growth of still surviving cells; more importantly, this mechanism requires caspase-3 activity in the dying cells [63], suggesting that caspase-3 has paradoxical functions, on one hand driving apoptotic cell death and on the other promoting the release of survival factors [64]. Secondly, there is also a possibility that mifepristone blocks the growth of scarce tumor initiating cells with the capacity to regenerate the culture and that may remain in culture because are resistant to cisplatin-paclitaxel in contrast to the bulk of differentiated cancer cells that succumb to the chemotherapy [65]. A genetic evolution study of high-grade serous ovarian adenocarcinomas suggests that resistance to cisplatin may develop from pre-existing minor clones that remain as minimal residual disease and become enriched after initial chemotherapy [66]. Within this scenario, mifepristone may block the repopulation of cells that never responded to cisplatin-paclitaxel therapy since, as we have shown, the drug has similar growth inhibition potency in platinum sensitive and platinum resistant ovarian cancer cells [20].