Reversal of Drug Resistance in Ovarian Cancer: Where Do We Go From Here?

Disease management in patients with ovarian cancer exemplifies the fundamental problem underlying much of today’s cancer chemotherapy: an initial excellent response followed by the eventual evolution of fatal drug resistance. It is therefore understandable that clinicians would seek to address this problem in clinical trials based on the evidence available to them regarding the underlying mechanisms of resistance. More than 10 years ago, an intriguing possibility being considered in this context was the pharmacologic reversal of so-called “multidrug resistance.” This is the concept behind the trial reported by Lhomme et al in this month’s issue of Journal of Clinical Oncology, and it involves the agent valspodar, which is a potent inhibitor of the energy-dependent membrane transport protein, P-glycoprotein (P-gp). Overexpression of P-gp (first identified by Victor Ling more than 30 years ago) is responsible for the experimental observation of cross-resistance in cancer cells to various cytotoxic agents, all of which are derived from natural products. Reversal can be achieved in experimental models using several types of P-gp inhibitors. What is much less clear is the relevance of these observations to the clinic, and the terminology is not helpful. Multidrug resistance based on P-gp is widely misinterpreted as underlying clinical resistance to many drugs, and this is not the case. It has no relevance, for example, to platinum, the key drug in ovarian cancer treatment. So why did the investigators in 1997 hope that a P-gp inhibitor might have an impact in this context? In their defense, the positive evidence was the observation that P-gp–positive cells had been clearly identified in patients with drugresistant ovarian cancer; indeed, this was the first tumor type in which they were seen. Secondly, the laboratory data indicating that taxane resistance may be susceptible to this approach were persuasive, and it was not until some years later that the results of randomized clinical trials (Gynecology Oncology Group 132 and International Collaborative Ovarian Neoplasm–3) indicated that, of the two drugs used in first-line treatment, platinum is likely to play a much more important role than paclitaxel. The trial was therefore completed between 1997 and 1999, and so the negative results represent a mature analysis. It was scrupulously conducted, and it incorporated the inevitable dose reduction of paclitaxel, which is required with cotreatment using a potent P-gp inhibitor such as valspodar, through its physiological inhibition of biliary drug excretion. The high level of normal tissue toxicity in the experimental arm indicates that underexposure to paclitaxel is not the reason for the lack of enhancement of efficacy. The more likely explanation is that resistance to chemotherapy in this patient population is not related to multidrug resistance. In retrospect, this could have been predicted, since we now have a better insight into at least some of the molecular mechanisms underlying drug resistance in this disease. Although we are still some way from a full explanation, the availability of modern techniques to analyze the complete genome in cancer cells, and to correlate aberrations with clinical data, is clearly a step forward. The acquisition of appropriate tumor tissue from ovarian cancer patients is the key, and this may be achieved in several ways. A pretreatment sample on its own can be informative, although when correlations with outcome are made, large numbers are required to take account of other prognostic variables in addition to the response to chemotherapy. Paired biopsy samples from the same patient (pretreatment and at relapse) are extremely valuable,buthardertoobtain;inthiscontext,tumorcellspresentinascites areoneobvioussource. Itmayevenbepossible toobtainmeaningfuldata simply through blood sampling, because we know that in the circulation of ovarian cancer patients, free tumor DNA (and RNA) is present, as well as a variable number of circulating tumor cells. This has been utilized in one analysis of paired blood samples in a large first-line clinical trial in ovarian cancer; the conclusion was that the acquisition of methylation of damage recognition genes (specifically hMLH1) was a potential explanation for the resistance to subsequent chemotherapy in patients whose disease had relapsed. Other groups have utilized microarray-based profiling to generate various gene signatures that may correlate with drug resistance or treatment outcome. These studies have been relatively small ( 100 cases), and there is usually a significant degree of heterogeneity in the circumstances of sample collection that makes comparison difficult. International collaboration is likely needed to maximize the information becoming available using various techniques. A particular challenge will be to identify the different factors relating to platinum and taxanes, because it is very likely that they will not be the same, and modulation approaches will ultimately need to recognize that clinical drug resistance is almost certainly multifactorial. Nevertheless, clinical investigators are already exploring alternative pharmacologic means of resistance reversal, based at least partly on data, albeit incomplete, from clinical material. Considering platinum resistance, these include the demethylation approach, which is capable of reversing resistance in an appropriate ovarian cancer xenograft model. A randomized trial involving decitabine is now underway in the United Kingdom following the demonstration in a phase I trial of the feasibility of this approach. Phenoxodiol is an inhibitor of the XIAP family of antiapoptotic proteins, and experimental reversal of resistance to both platinum and taxanes has been demonstrated JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 26 NUMBER 16 JUNE 1 2008