Introduction

Lung cancer, also called primary bronchogenic carcinoma, is a malignant tumor derived from the trachea, bronchial mucosa or glands. Currently, it is the most common malignant tumor in the world [1]. According to statistics, in 2015, lung cancer is the most important cause of death in cancer patients, and the incidences of lung cancer in Chinese men and women were 50.9 per 100,000 person-years and 22.4 per 100,000 person-years, respectively [2]. Surgery is still the preferred treatment for early-stage lung cancer. However, most lung cancer patients are diagnosed at an advanced stage, and their 5-year survival rate is only 19.7%, that is to say, they missed the best treatment opportunity [1, 3]. Most studies have suggested that the occurrence of lung cancer is related to environmental factors, such as smoking, occupational exposure, and air pollution [4, 5]. At the same, genetic factors play an important role in the development of lung cancer, including EGFR [6], CHRNA5 [7], CLPTM1L [8], TP63 [9], and so on.

GTF2H1 (general transcription factor IIH subunit 1) protein plays a vital role in the nucleotide excision repair (NER) pathway, participates in the early damage recognition of XPC-HR23B protein, and recruits endonuclease XPG to the injury site to complete the enzymatic cleavage process [10,11,12]. In addition, GTF2H1 protein is involved in transcription and regulates the transcriptional activation of multiple genes [13]. One research has found that rs3802967 and rs4150606 in the GTF2H1 gene may increase the risk of lung cancer, and rs4150667 in the GTF2H1 gene variant may reduce the risk of lung cancer [14]. So, the genetic polymorphisms of GTF2H1 may be involved in the pathogenesis of lung cancer.

RAD54L2, also known as ARIP4 (androgen receptor-interacting protein 4), is a protein-coding gene belonging to the RAD54 subfamily of SNF2-type chromatin remodeling factor superfamily [15], with double-stranded DNA-dependent ATPase activity. In fact, Rad54 interacts with Rad51, thereby enhancing its ability to form cruciform and D-loops. RAD51 catalyzes DNA repair by homologous recombination to ensure the stability of cell genome. In a study of the effects of RAD51 G135C polymorphism on non-small cell lung cancer patients treated with platinum-paclitaxel/gemcitabine Wrst line chemotherapy, it has been found that the G135C allele is associated with a higher survival time and a better prognosis [16]. However, the effect of the RAD54L2 gene on the occurrence and development of lung cancer is unknown.

Here, nine polymorphisms in GTF2H1 and RAD54L2 were selected and genotyped to explore their impact on the risk of lung cancer in the Chinese Han population.

Materials and methods

Study participants

A total of 506 patients with lung cancer and 510 age-and gender-matched healthy controls were recruited from the Qinghai Province Cancer Hospital. The inclusion criteria for cases were lung cancer confirmed by histopathology and no history of malignant tumors in other organs. Patients with prior cancer history, pulmonary diseases, and serious chronic diseases were excluded. The control group consisted of healthy people with no medical or family history of cancer or any lung disease. Further, each subject was interviewed by trained personnel using a structured questionnaire to obtain information about demographic characteristics [age, gender, smoking and drinking, body mass index (BMI)]. Based on smoking status, participants were classified into non-smokers (never) and smokers (ever or current smokers). Subjects who smoked at least one cigarette per day were classified as current smokers. For drinking status, participants were classified into non-drinkers (never) and drinkers (ever or current alcohol drinkers). Subjects who drank at least 100 g of alcohol per week were considered as drinkers. Pathological data (pathological type, lymph node metastasis, and clinical stage) were obtained via medical records. This study was conducted under the approval of the Institutional Review Boards of Northwest University. All participants were informed of the content of the study and signed informed consent. It was confirmed that all methods were carried out in accordance with relevant guidelines and regulations.

SNP selection and genotyping

The GoldMag-Mini Whole Blood Genomic DNA Purification Kit (GoldMag Co. Ltd. Xi’an City, China) was applied to extract DNA samples from the 5 mL peripheral venous blood, and Nanodrop 2000 (Gene Company Limited) was used to detect the concentration and purity of DNA samples to ensure that samples could be used for subsequent experiments. Four SNPs (rs11720298, rs4687721, rs4687592, and rs9864693) in RAD54L2 and five SNPs (rs4150530, rs3802967, rs4150606, rs4150658, and rs4150667) in GTF2H1 were randomly selected in our study based on the following criteria: (1) minor allele frequency (MAF) > 0.05, min genotype > 75%, and r2 > 0.8 from the Genome Aggregation Database (gnomAD, http://www.gnomad-sg.org/), (2) MAF more than 0.05 in the Chinese Han population from dbSNP database (https://www.ncbi.nlm.nih.gov/snp/), (3) previous studies on these polymorphisms have been related to lung cancer [14], (4) combined with MassARRAY primer design software, Hardy-Weinberg equilibrium (HWE) > 0.05 and the call rate > 95% in our study population. Haploreg (https://pubs.broadinstitute.org/mammals/haploreg/haploreg.php), RegulomeDB (https://regulome.stanford.edu/regulome-search/), and FuncPred (https://manticore.niehs.nih.gov/snpinfo/snpfunc.html) were applied to identify potential functional SNPs in the human RAD54L2 and GTF2H1 genes.

Primer design and SNP genotyping were performed as shown in Suppl_Table 1. The genotyping primers were designed with the Agena MassARRAY Assay Design 3.0 software [17]. The Agena MassARRAY RS1000 system was used for genotyping, and the related data were managed using Agena Typer 4.0 software [17,18,19]. To ensure accuracy, about 5% of the samples were randomly re–genotyped, and the concordance of duplicated genotyping was 100%.

Statistical analysis

Demographic characteristics between cases and controls were compared by Student’s t test and χ2 test. The Hardy-Weinberg equilibrium (HWE) was calculated by χ2 test [20]. Multiple genetic models were used to evaluate the association between gene polymorphisms and lung cancer risk. Odds ratios (ORs) and corresponding 95% confidence intervals (CIs) adjusting for age and gender were estimated using a logistic regression model through the PLINK software [21]. Further analyses based on age, gender, BMI, smoking, drinking, histology, lymph node metastasis, and clinical stages were performed to assess the impact of polymorphisms on lung cancer. Multifactor dimensionality reduction (MDR) (version 3.0.2) was applied to evaluate the impact of SNP–SNP interactions on the risk of lung cancer. The threshold of p was set at 0.05.

Results

The information on patients with lung cancer and healthy participants was presented in Table 1. The mean ages of participants in the case and control groups were 59.80 ± 10.63 years and 59.80 ± 9.08 years, respectively. The case group includes 350 males and 156 females, and the control group consists of 353 males and 157 females. There were no significant differences in terms of age (p = 0.992) and gender (p = 0.987) between the two groups. Of the 506 patients, 269 had lymph node metastasis,103 had no metastasis, and 286 (56.50%) patients were in stage III–IV. And there were 174 (34.4%) cases of squamous cell carcinoma and 212 (41.9%) cases of adenocarcinoma.

Table 1 The information of the participants
Full size table

Table 2 showed the basic information on nine SNPs in the RAD54L2 and GTF2H1 genes, including physical location, chromosome, minor allele frequency, and HWE. And all variants met the HWE. Associations between RAD54L2 and GTF2H1 polymorphisms and the risk of lung cancer were evaluated under different genetic models. In allele model (Table 2), no significant association of SNPs with the genetic susceptibility of lung cancer was found. Database analysis presented that the potential functions of these SNPs might be related to promoter /enhancer histone marks, transcription factor binding, DNAse, proteins binding, changed motifs changed, and selected expression quantitative trait loci (eQTL) hits.

Table 2 The information of nine gene polymorphisms on the RAD54L2 and GTF2H1 gene
Full size table

In genotype model, subjects with RAD54L2 rs9864693 GC heterozygote genotype might have an increased risk of lung cancer compared with individuals with GG wild-type genotype (crude analysis: OR = 1.33, 95%CI: 1.01–1.76, p = 0.046; adjusted analysis: OR = 1.33, 95%CI: 1.01–1.77, p = 0.045) (Table 3). No statistically significant association between other SNPs with lung cancer susceptibility was observed (p > 0.05, Suppl_Table 2).

Table 3 Risk analysis forRAD54L2 rs9864693 and lung cancer in different genetic models by logistic regression analysis
Full size table

Stratified analysis was carried out based on age, gender, BMI, smoking, drinking, pathological type, lymph node metastasis, and clinical stages (Table 4). In subjects aged ≤ 59 years, RAD54L2 rs11720298 was related to a reduced susceptibility to lung cancer, whereas several risk-increasing associations of RAD54L2 rs4687592 (OR = 0.69, p = 0.012), RAD54L2 rs9864693 (OR = 1.64, p = 0.012) and GTF2H1 rs4150667 (OR = 1.32, p = 0.048) with lung cancer were found. Among subjects with BMI ≤ 24 kg/m2, GTF2H1 rs4150667 (OR = 0.18, p = 0.008) contributed to a lower risk of developing lung cancer. Among non-smokers and drinkers, the protective effects of RAD54L2 rs11720298 (OR = 0.62, p = 0.011; and OR = 0.17, p = 0.008) on the occurrence of lung cancer were observed. Besides, an increased risk of lung cancer was observed for RAD54L2 rs4687592 in drinkers (OR = 2.54, p = 0.034). Stratified by pathological type, GTF2H1 rs3802967 was associated with a reduced risk of lung squamous cell carcinoma (OR = 0.68, p = 0.045). However, no significant relationships of selected polymorphisms with lung cancer risk in the stratified analysis by gender, lymph node metastasis, and clinical stages were detected (Suppl_Table 3 and Suppl_Table 4).

Table 4 Stratified analysis for the associations between RAD54L2 and GTF2H1 polymorphisms and the risk of lung cancer
Full size table

FPRP analysis was performed to verify positive findings, as shown in Table 5. At a prior probability level of 0.25, a significant association of RAD54L2 rs9864693 remained noteworthy overall (FPRP = 0.160 and statistical power = 0.795). In subjects aged ≤ 59 years, correlations of rs11720298, rs4687592 and rs9864693 in RAD54L2 with the susceptibility to lung cancer were also positive at a prior probability level of 0.1. Moreover, an association of RAD54L2 rs11720298 with the risk of lung cancer in non-smokers and drinkers was significant at a prior probability level of 0.1.

Table 5 False-positive report probability values for the associations between RAD54L2 and GTF2H1 polymorphisms and the risk of lung cancer
Full size table

MDR was applied to analyze the interactions of these SNPs. The results of MDR model analysis for SNP-SNP interactions were displayed in Table 6; Fig. 1. The best multi-loci model was the eight-locus model, namely, a combination of rs3802967, rs4687721, rs9864693, rs11720298, rs4150606, rs4150658, rs4150667, and rs4687592, with a highest testing accuracy (0.5286) and a perfect cross-validation consistency (10/10). As shown in Fig. 1, the dendrogram plot demonstrated the interactions among these eight SNPs and recapitulated the main and/or interaction effect on each pairwise combination of attributes. Red and orange line indicated synergistic interaction, blue and green color indicated redundant interactions. The result suggested that rs4687721 and rs4150667 had a synergistic interaction sharing the positive information gain with lung cancer.

Table 6 SNP–SNP interaction models of candidate SNPs analyzed by the MDR method
Full size table
Fig. 1
figure 1

Dendrogram for the interactions among SNPs in GTF2H1 and RAD54L2 on the risk of lung cancer.  Red and orange lines indicate synergistic interaction, blue and green color indicated redundant interactions. Short connections among nodes represent stronger interactions

Full size image

Discussion

In our research, we found that RAD54L2 rs9864693 was related to an increased risk of lung cancer in the Chinese Han population. Especially, among subjects aged ≤ 59 years, RAD54L2 rs11720298 was related to a reduced susceptibility to lung cancer, while the risk-increasing associations were found for RAD54L2 rs4687592, RAD54L2 rs9864693 and GTF2H1 rs4150667. Among subjects with BMI ≤ 24 kg/m2, GTF2H1 rs4150667 had the lower risk of developing lung cancer. In non-smokers and drinkers, the protective risk effect of RAD54L2 rs11720298 on the occurrence of lung cancer was observed, respectively. Besides, the increased risk of lung cancer was observed for RAD54L2 rs4687592 in drinkers. GTF2H1 rs3802967 was associated with the reduced risk of lung squamous cell carcinoma. MDR result suggested that rs4687721 and rs4150667 had synergistic interaction sharing the positive information gain with respect to lung cancer. Our findings provide data for constructing a genetic panel to predict lung cancer risk in China.

RAD54L2 (ARIP4), located on human chromosome 3p21.2, is initially identified as an ATPase of Rad54/ATRX subfamily of SNF2-like proteins, which contains chromatin remodeling activity, interacts with AR and regulates androgen-mediated transactivation [22]. SNF2-like proteins are thought to modify the structure of chromatin in a non-covalent manner through rearrangement of nucleosomes. Only RAD54L2 has been found to affect gastrointestinal stromal tumors, and the expression of RAD54L2 has a shorter overall survival time [23]. Rad54 interacts with Rad51, which functions during DNA repair. It has been proved that Rad51 G135C allele is associated with a higher survival time and a better prognosis in lung cancer patients treated with platinum-paclitaxel / gemcitabine Wrst line chemotherapy [16]. RAD54B protein expression is significantly higher in lung adenocarcinoma tissues than that in healthy lung tissues, and inhibition of RAD54B expression in A549 cells can significantly reduce cell proliferation and increase apoptotic rate [24]. However, the effect of RAD54L2 on lung cancer is not clear. In our research, we found that RAD54L2 rs9864693 was associated with an increased risk of lung cancer, especially in subjects aged ≤ 59 years. Besides, RAD54L2 rs11720298 was related to a reduced susceptibility to lung cancer in subjects aged ≤ 59 years, non-smokers and drinkers, while a risk-increasing association was found between rs4687592 and lung cancer risk in subjects aged ≤ 59 years and drinkers. Here, we first displayed the genetic association of RAD54L2 polymorphisms with the susceptibility to lung cancer in the Chinese Han population. However, the role of RAD54L2 gene in the occurrence and development of lung cancer needs to be further clarified.

GTF2H1 interacts with the C- and N-terminus of XPC protein, participates in the recruitment of other protein subunits in TFIIH, and initiates the NER repair process. The study found that the expression of GTF2H1 was down-regulated in lung cancer tissues [25]. Rs3802967 is located in the 5’-UTR region of GTF2H1 gene, and the luciferase activity experiments displayed that rs3802967 T allele was related to the enhanced expression of GTF2H1 in lung cancer cells [14]. A previous study has showed that variants of GTF2H1 rs3802967 and rs4150667 are also significantly associated with the risk of lung cancer in the southern Han Chinese population [14]. In our study, it was found that GTF2H1 rs3802967 CT-TT reduced the risk of lung squamous cell carcinoma. Moreover, a risk-increasing association was found between GTF2H1 rs4150667 and lung cancer risk in subjects aged ≤ 59 years, while GTF2H1 rs4150667 was associated with a reduced risk of developing lung cancer in subjects with BMI ≤ 24 kg/m2. However, such inconsistency might result from different behavioral habits or sample sizes. Besides, rs4150606 on GTF2H1 increased the risk of lung cancer [14]. However, no correlation between the variants of GTF2H1 rs4150606 and the risk of lung cancer was found. This may be due to the false negative results of our small sample size, and whether the association of SNPs with the risk of lung cancer needs to be further confirmed.

Several limitations in this study should be considered. First, the subjects were enrolled from the same hospital; therefore, there is selection bias. Second, after stratification, the sample size of each subgroup is relatively small, thus, a well-designed large sample is needed to further confirm our findings. Third, the mechanism of selected polymorphisms on the occurrence and development of lung cancer is still unclear, and further research is needed. Fourth, due to the absence of information, including BMI, smoking, and drinking, only age- and sex- adjustments were performed in this study. In the follow-up studies, we will improve relevant information and adjust risk factors to further analyze the correlation of GTF2H1 and RAD54L2 with the risk of lung cancer.

Conclusion

To sum up, our study first revealed that RAD54L2 rs9864693 was associated with an increased risk of lung cancer in the Chinese Han population. Our finding will provide evidence that age, BMI, smoking, and drinking might be associated with the effects of RAD54L2 and GTF2H1 variants on lung cancer susceptibility. This study may increase the understanding the effects of of RAD54L2 and GTF2H1 polymorphisms on the occurrence of lung cancer.