DNA repair and cisplatin resistance in non-small-cell lung cancer

Abstract

The results of cisplatin-based chemotherapy seem to have reached a plateau, and empirical approaches are targeting the inclusion of novel biological agents with different mechanisms of action, but their clinical benefit is still unknown. In preparing this review of cisplatin resistance, we posed two questions: Who are we writing for and why? We believe that medical oncologists should be involved in the reality of the growing list of genetic mechanisms of cancer and chemoresistance. Only by becoming familiar with these mechanisms will we be able to circumvent them. In this review, we provide some insight into DNA repair defects involved in non-small-cell lung cancer (NSCLC) and cisplatin effect. Some DNA repair genes, like ERCC1, have been shown to be crucial in predicting cisplatin resistance and can be used for tailoring cisplatin-based chemotherapy.

Introduction

Cell repair capacity is stored in the linear sequence of approximately 3×10⁹ copies of the four bases guanine, cytosine, adenine and thymine, aligned in the DNA. A growing list of reports identify DNA damage with the regulation of DNA repair gene transcription and the control of cell cycle progression and apoptosis via DNA damage checkpoints. Different pathways of DNA repair are polymorphic and vary interindividually and with age. These features influence the chemosensitivity of tumor cells toward DNA-reactive cytotoxic drugs [1]. DNA repair is a counteragent in carcinogenesis and an accomplice in cancer therapy resistance [1]. There are several major DNA repair pathways. Excision repair, including nucleotide excision repair (NER) has been strongly linked to cisplatin resistance. Base excision repair (BER) also plays an important role in chemotherapy resistance. Finally, the repair of double-strand breaks, induced by cytotoxic agents, radiotherapy, and reactive oxygen species, is carried out by homologous recombination and non-homologous DNA end joining. Other pathways are mismatch repair (MMR) and one-step repair (OSR), meaning the direct reversal of DNA damage. The repair protein O⁶-alkylguanine-DNA alkyltransferase, also known as O⁶-methylguanine-DNA methyltransferase (MGMT), intervenes in OSR through removal of an alkyl group from the O⁶-atom of guanine in the DNA of cells exposed to alkylating agents. With increasing size of the alkyl group, the relative contribution of MGMT to the repair of O⁶-alkylguanines in DNA decreases and excision repair becomes more relevant (reviewed in [1]). As an example of OSR, treatment with chloroethylnitrosoureas (BCNU) correlates with MGMT activity; in the process of cytotoxic interstrand cross-links in target cell DNA, BCNU initially alkylate the O⁶-atom of guanine. Intriguingly, MGMT levels vary greatly between tumors, which has been used in pharmacogenomic interpretation. Hypermethylation of MGMT (abrogating OSR) was observed in 40% of brain tumors treated with BCNU and was related to significantly better survival [2]. Interestingly, the activity of temozolomide has been linked to tumor MGMT. However, when temozolomide was combined with CPT-11, this mechanism of resistance was circumvented in tumor cells that were either MGMT proficient or MMR deficient [3].

The possibility of individualizing DNA repair profiles is becoming a central issue in the search for improved chemotherapy results. However, while gene expression profiling with cDNA arrays is on the cutting edge of technology, it has no immediate application in the clinical setting. Cancer cells accumulate multiple genetic abnormalities in signal transduction pathways during carcinogenesis and cancer progression. One Achilles’ heel of tumor cells is defects in DNA repair, which can be responsible for lung oncogenesis while simultaneously conferring a chemotherapy advantage, especially with cisplatin, the drug that has been studied to the greatest extent. Cisplatin is the paradigm of cytotoxic drugs used in the treatment of non-small-cell lung cancer (NSCLC), and like many DNA alkylators, acts as a cross-linker, inhibiting DNA replication, which is the critical target in cancer chemotherapy. Cross-links between guanine bases are induced by cisplatin, carboplatin and oxaliplatin. Cisplatin and carboplatin form an identical cross-link, while the oxaliplatin cross-link is structurally very different due to the bulky 1,2-diaminocyclohexane group in the adduct (reviewed in [4]).

After more than two decades of clinical trials, we have reached a therapeutic plateau that is often interpreted nihilistically by non-experts in lung cancer treatment, who argue that nothing is better than the archaic cisplatin/etoposide regimen and who place their hopes in new orally-administered drugs. Table 1 shows some studies in which response rate, time to progression and median survival have been similar across different chemotherapy regimens. Cisplatin, 40 or 120 mg/m², in combination with either cyclophosphamide and doxorubicin or vindesine, respectively, resulted in a 10% improvement in 1-year survival [5]. Pooled data from eight cisplatin trials (778 patients) showed a hazard ratio of 0.73, with an absolute improvement of 10% at 1 year and a 1–1/2-month increase in median survival time [6]. An almost identical number of patients (797) were randomized in a British trial comparing MIC (mitomycin/ifosfamide/cisplatin), radiotherapy alone and palliative care [7]; the magnitude of the benefit was similar to that reported in the meta-analysis [6]. In recent years, new cisplatin combinations have been tested, including vinorelbine [8]. The Le Chevalier et al. study [8] showed a better median survival time for vinorelbine/cisplatin than for vindesine/cisplatin or vinorelbine alone. More recently, a Southwest Oncology Group (SWOG) study [9] found that the median survival time with vinorelbine/cisplatin was similar to that obtained with paclitaxel/carboplatin (8.1 and 8.6 months, respectively). When looking at clinical parameters as prognostic factors in multivariate analyses, cisplatin and female gender emerge as independent variables in addition to performance status, as demonstrated in a review of SWOG trials including a total of 2531 patients [10]. The European Lung Cancer Working Party trial also identified female gender as an independent prognostic variable [11]. Deficient DNA repair capacity enhances cisplatin activity and females have less DNA repair capacity than males. Finally, the yardstick in chemotherapy, the four-arm Eastern Cooperative Oncology Group (ECOG) trial [12], included 1155 eligible patients and observed an overall 19% response rate, 7.9-month median survival time, 33% 1-year survival, and 11% 2-year survival. No differences were found between paclitaxel/cisplatin and the three experimental arms. However, median time to progression was significantly better in the cisplatin/gemcitabine arm than in the cisplatin/paclitaxel arm (P=0.001). These findings seem to indicate that cisplatin resistance is linked to relative levels of excision repair cross-complementing 1 (ERCC1) mRNA expression. ERCC1 is one of the master genes involved in the NER pathway and is required for synergism between cisplatin and gemcitabine, indicating that the threshold of ERCC1 expression leading to cisplatin failure may be higher when cisplatin is combined with gemcitabine.