Polymorphisms in the BER and NER pathways and their influence on survival and toxicity in never-smokers with lung cancer

Polymorphisms in DNA repair pathways may play a relevant role in lung cancer survival in never-smokers. Furthermore, they could be implicated in the response to chemotherapy and toxicity of platinum agents. The aim of this study was to evaluate the influence of various genetic polymorphisms in the BER and NER DNA repair pathways on survival and toxicity in never-smoker LC patients. The study included never-smokers LC cases diagnosed from 2011 through 2019, belonging to the Lung Cancer Research In Never Smokers study. A total of 356 never-smokers cases participated (79% women; 83% adenocarcinoma and 65% stage IV). Survival at 3 and 5 years from diagnosis was not associated with genetic polymorphisms, except in the subgroup of patients who received radiotherapy or chemo-radiotherapy, and presented with ERCC1 rs3212986 polymorphism. There was greater toxicity in those presenting OGG1 rs1052133 (CG) and ERCC1 rs11615 polymorphisms among patients treated with radiotherapy or chemo-radiotherapy, respectively. In general, polymorphisms in the BER and NER pathways do not seem to play a relevant role in survival and response to treatment among never-smoker LC patients.


Methods
Study-design and setting. A multicenter, hospital-based, case-control study performed in never-smokers, (Lung Cancer Research In Never Smokers-LCRINS) was conducted at 7 hospitals in Galicia, northwest Spain 23 . Patient recruitment took place from January 2011 through July 2019, and only included cases with histologic confirmation of LC who were never-smokers according to the WHO definition, i.e., anyone who has smoked: (1) fewer than 100 cigarettes in his/her lifetime; or (2) less than 1 cigarette per day during a period of no less than 6 months. All cases were over 18 years, without any upper age limit. Subjects with small cell lung cancer were excluded from this study, since the disease's characteristics are different to those of other types of LC. The study protocol was approved by the Ethics Committee at the Santiago de Compostela Healthcare area (reference 2010/295), and informed consent was obtained from all patients. All methods were performed in accordance with the relevant guidelines and regulations. Data-collection. All patients were interviewed by trained researchers, who administered a questionnaire used in a number of previous studies 23,24 to collect data on LC risk factors. Exposure to environmental tobacco smoke was assessed through personal interview and included all persons living with the participant in the last 20 years. Exposure to residential radon was also measured with alpha-track detectors. The detector readings were processed under standard conditions at the Galician Radon Laboratory (School of Medicine, Santiago de Compostela, Spain), a facility officially certified by the National Accreditation Entity of Spain (https ://www. radon .gal).
Patients follow-up and evaluation of treatments and survival. All patients were followed up until date of death or last medical visit. Follow-up was conducted by consulting the electronic medical records, reviewed until July 2019. Follow-up time was calculated as the period of time between date of diagnosis and date of death or last visit. For all patients, we obtained histologic type, stage at diagnosis, and treatments received (chemotherapy, radiotherapy, chemo/radiotherapy, surgery). Special emphasis was laid on the type of chemotherapy treatment received. Data were also obtained on toxicity (overall and by type). Validated toxicity evaluation scores were used, such as that designed by the National Cancer Institute of Canada and the Eastern Cooperative Oncology Group. www.nature.com/scientificreports/ Laboratory methods. A 3-ml blood sample was obtained from all participants, and used for determination of different genetic polymorphisms. Samples were stored at -84 ºC until time of analysis, which was performed by the National Genotyping Center (Centro Nacional de Genotipado-CeGen) at the University of Santiago de Compostela, Spain . The genetic analyses determined the complete deletion of the GSTM1 and GSTT1 genes,  XRCC1 -rs25487-, ERCC1 -rs11615, rs3212986-, ERCC2 -rs13181-, XRCC3 -rs861539-, OGG1 -rs1052133-and AAT deficiency. The technology used was Agena Bioscience's MassARRAY system, which enables determination of specific DNA variants, single nucleotide polymorphisms (SNPs), as well as small insertions and deletions. Polymerase chain reaction (PCR) was followed by iPLEX extension reactions. Reactions took place in 384 separate wells, and the extension products were transferred to a SpectroCHIP Array and analyzed with MALDI-TOF mass spectrometry.
Statistical analysis. Firstly, a univariate analysis was performed, and survival of never-smoker patients was analyzed using Cox Regression in which the dependent variable was survival at 3 and 5 years of diagnosis taking into account the different polymorphisms analyzed. A multivariate Cox regression was further performed adjusted for age, sex, passive exposure to tobacco smoke, residential radon, and stage at diagnosis. The reference group for analyzing the data on the different genotypes was homozygosity for the most common allele (wildtype). In addition, we calculated survival according to the presence of genetic polymorphisms (survival, both overall and specific of patients in stage IV at diagnosis), as well as the effect of polymorphisms on toxicity to chemotherapy or other treatments. p-value was considered significant when it turned below 0.05. Analysis of each specific type of toxicity was performed with the aid of the Chi-square test, due to the low frequency of some types of toxicity and the polymorphisms analyzed. As multiple comparisons have been done, false discovery rate (FDR) was calculated in order to consider statistical power. We obtained survival graphs for some polymorphisms at 3 and 5-years from diagnosis. All analyses were performed using the IBM SPSS v22 software program (IBM, Armonk, NY, USA; https ://www.ibm.com/analy tics/spss-stati stics -softw are).

Results
The study included 356 never-smoker cases, 79% of whom were women. Mean age at diagnosis was similar in both sexes, the predominant histologic type was adenocarcinoma (83%), and the most frequent stage at diagnosis was IV (65%). 25% of all patients received chemotherapy, 11% received radiotherapy, and 28% received chemo-radiotherapy. The remaining patients received either palliative treatment or treatments directed at specific molecular targets. Table 1 shows the characteristics of the participants along with the distribution of the polymorphisms analyzed. Survival at 5 years from diagnosis according to the treatment received, along with the effect of each of the polymorphisms analyzed, is shown in Table 2 (data for survival at 3 years from diagnosis is shown in supplemental Table S1). In the group of patients who received chemotherapy, no association was observed between any of the polymorphisms analyzed and 5-year survival. In the group of patients who received other types of treatment, there was an association between survival at 5 years and presence of the ERCC1 rs3212986 (TT) polymorphism. Supplemental Fig. 1 shows 5-year lung cancer survival according to the treatment received and polymorphisms analyzed. Table 3 shows survival at 5 years from diagnosis for patients in stage IV according to the treatment received and polymorphisms analyzed (data for survival at 3 years from diagnosis is shown in supplemental: Table S2). In the group of stage IV patients who received chemotherapy, an association between survival at 5 years and presence of the AAT (MZ) polymorphism can be seen, while for the ERCC1 rs3212986 polymorphism the association was close to statistical significance. Table 4 shows toxicity data according to the treatment received for each of the polymorphisms analyzed. No association was observed with the polymorphisms and toxicity due to chemotherapy, or for the group of patients who received platinum-based chemotherapy. A significant association was observed for patients who received radiotherapy and presented with the OGG1 rs1052133 (CG) polymorphism, and for those who received chemoradiotherapy and presented with the ERCC1 rs11615 polymorphism (data shown in supplemental: Table S3). The most frequent toxicities were toxicoderma due to anti-epidermal growth factor receptor (EGFR) treatment, and the presence of nausea and vomiting (Supplemental Table S4). The presence of nausea and vomiting was statistically significantly associated with absence of the GSTT1 gene. As multiple comparisons have been done, false discovery rate (FDR) was calculated in order to consider statistical power (Supplemental Fig. 2).

Discussion
LC survival at 5 years from diagnosis in never-smokers does not seem to be clearly influenced by polymorphisms in genes implicated in DNA repair mediated by the BER or NER pathways, or by deletion of GSTM1 and GSTT1 genes. This apparent absence of effect is maintained for subjects treated with platinum compounds. On the other hand, the ERCC1 rs3212986 polymorphism does show greater toxicity for patients who have received radiotherapy or combined chemo-radiotherapy. Similarly, patients who received these treatments (radiotherapy or combined chemo-radiotherapy) presented with greater toxicity in the presence of the OGG1 rs1052133 (CG) and ERCC1 rs11615 polymorphisms, respectively. There was greater digestive toxicity (nausea and vomiting) in patients with GSTT1 deleted. In terms of sample size, it should be noted that the present study has one of the greatest sample sizes of never-smoker cases in whom these polymorphisms have been analyzed. Moreover, it is the only study to adjust its results for residential radon exposure.
DNA damage repair mechanisms in cancer are of various types, i.e., NER and BER 25 . NER-type repair begins with examination of the area of DNA damaged, and the genes involved in this process are ERCC1, ERCC2, ERCC3, ERCC4, ERCC5 and XPA. Previous studies have shown that the level of expression of these genes is www.nature.com/scientificreports/ important in the response and toxicity to platinum-based chemotherapy 26,27 . BER mechanism is responsible for eliminating base lesions in the genome, such as those caused by platinum-based chemotherapy or radiotherapy. The most studied genes in this mechanism are XRCC1 and XRCC2, and studies have associated the relationship between polymorphisms at this level with survival or response to chemotherapy 28  reported an association between ERCC1 rs11615 polymorphism (T allele) and shorter survival, when compared to patients with the C/C genotype 29 . However, the presence of this ERCC1 polymorphism (albeit in a different variant: rs3212986) in our study was associated with longer survival. This discordance could likely be due to the fact that the type of patients analyzed in our study were different (exclusively never-smoker subjects, all stages at diagnosis, higher percentage of women, etc.). In line with our findings, Ren et al. observed that the presence of ERCC1 rs11615 polymorphism (C/T or T/T) was associated with longer survival (p = 0.014) 30 . Furthermore, Ramos et al. in a recent systematic review, analyzed the association between GSTM1 and GSTT1 deletion and survival, concluding that most studies showed no effect or a slight trend towards a worse survival 13 . As can be seen, the results reported by these papers cited are heterogeneous. The majority of these studies have been conducted on LC patients -smokers as well as nonsmokers-among whom the latter tend to represent only a small percentage, so that the relationship between polymorphisms and survival in never-smokers is mostly unknown. Toxicity due to treatment of LC patients has been associated with different genetic polymorphisms in the different studies published. For instance, Sullivan et al. reported an association between the ERCC2 (rs1799793) polymorphism and radiation pneumonitis 14 . In our study, however, the ERCC1 (rs11615) polymorphism was associated with toxicity in patients treated with chemotherapy plus radiotherapy. Yet in patients treated with platinum-based on chemotherapy, Ludovini et al. found no relationship, between the ERCC1, XRCC3 and XDP polymorphisms and toxicity 31 . In the case of the GSTT1 gene, Kalikaki et al. observed that it was associated with a higher risk of toxicity (76% of patients developed neutropenia vs. 24% of patients who did not have this polymorphism) 32 . In our study, in contrast, absence of the GSTT1 gene was associated with a different toxicity (nausea and vomiting). It should be noted: firstly, that none of the above studies specifically evaluated toxicity in never-smokers; and secondly, that polymorphisms can influence toxicity as well as response to treatment 33 .
There are likely differences in the relationship between genetic polymorphism and survival for smokers vs. never-smokers. The fundamental problem lies in the fact that there are few studies which specifically analyze the influence of polymorphisms among never-smokers. One example of these is Yang et al. 's study, which observed that the GSTP1 polymorphism was only associated with better survival in never-smoker patients 34 . LC in neversmokers is probably a different clinical entity (higher percentage of women, greater number of adenocarcinomas, greater presence of mutations in driver genes, etc.,) than that occurring in ever smokers, making it feasible that polymorphisms play a different role in never-smokers to that seen in smokers.
Similarly, it is known that the GSTM1 gene is implicated in the detoxification of metabolites of polycyclic aromatic hydrocarbons and that the GSTT1 gene is associated with the metabolism of other components of tobacco. This could be one reason why no association with survival is observed in never-smokers. GSTM1 and/or GSTT1 deletion produces an absence of enzymatic activity with an enhanced effect of chemotherapy treatment, greater oxidative damage in tumor cells and ensuingly higher toxicity, causing an expected probable increase in survival 13 .
This study has some advantages: it is multicentric and hospital-based, which lends it external validity. It is one of few studies that analyze the relationship between LC survival and the abovementioned polymorphisms in never-smokers, adjusting the results for residential radon concentration. In addition, it has a relatively large sample size. Since data-collection was based on electronic medical records, there were negligible losses to follow-up, www.nature.com/scientificreports/ and given that all the patients belonged to a given geographic area (Galicia), there was homogeneity in the treatments administered. This study also has some limitations. Firstly, it analyzes a high number of low-frequency genetic polymorphisms, thereby reducing the study's statistical power. Furthermore, the patients included few males (21%), most patients presented with stage IV at diagnosis (65%), and the majority of cancers were adenocarcinomas (83%), so that comparison with the rest of the literature published to date should be approached with caution. Secondly, we reported toxicity data according to electronical medical records and not to Common Toxicity Criteria. Finally, not all genes involved in BER and NER pathways have been analyzed and therefore a total picture on the involvement of these DNA repairing mechanisms could not be obtained.
In conclusion, survival at 5 years of diagnosis among never-smoker patients with LC was not observed to be associated with presence of genetic polymorphisms in genes that participate in the BER and NER pathways, except in the subgroup of patients who received radiotherapy or combined chemo-radiotherapy and presented www.nature.com/scientificreports/ with the ERCC1 rs3212986 polymorphism. Among the patients who received these treatments, higher toxicity was also found in the presence of OGG1 rs1052133 (CG) and ERCC1 rs11615 polymorphisms. Further studies are called for to elucidate the mechanisms whereby genetic polymorphisms can influence toxicity or survival among never-smoker LC patients.