The Interaction between Four Polymorphisms and Haplotype of ABCB1, the Risk of Non-Small Cell Lung Cancer, and the Disease Phenotype

P-glycoprotein, product of the ABCB1 (ATP binding cassette subfamily B member 1) gene, has been reported to play an important role in multiple drug resistance during cancer therapy. However, its influence on non-small cell lung cancer (NSCLC) risk has not been clearly defined. The aim of the present study was to examine the association between clinicopathological factors and SNPs T-129C, C1236T, G2677T/A, and C3435T, as well as its haplotype, and to investigate the role of ABCB1 polymorphisms in NSCLC development. The study included 80 patients who suffered from NSCLC and underwent surgery to remove the tumour and 96 healthy controls. The tissues were genotyped by PCR-RFLP and sequencing methods, and the haplotype frequencies in both groups were estimated. The SNP C3435T was identified as a NSCLC risk factor. The presence of mutated allelic variant T (p=0.0103) or homozygote TT (p=0.0099) was observed significantly more often in cancer patients than in healthy controls. The two groups also demonstrated a highly significant difference in common haplotype frequency (p=0.01). The T−129-T1236-T2677-T3435 haplotype was found to be most closely associated with NSCLC risk. Although the investigated polymorphisms were not related to demographic features, clinicopathological lung tumour characteristics, or blood morphology indices, marginally significant correlations were found with some variables: C1236T with age of disease onset (p=0.0410); C3435T with smoking status (p=0.0561). As the findings indicate, lung cancer and control groups demonstrate significantly different patterns of −129/1236/2677/3435 haplotype distribution; T-T-T-T haplotype contributes to NSCLC susceptibility, and this effect is probably mainly dependent on C3435T. So far, similar studies were published in other populations.


Introduction
Lung cancer is one of the most common cancer types in terms of both incidence and mortality [1]. Worldwide, more than 1.8 million lung cancer deaths occurred in 2020 [2]. In 2019, lung cancer was the second most prevalent newly diagnosed cancer and still the dominant cause of cancer death in both men and women, according to the Polish National Cancer Registry [3].
Lifestyle risk factors play key roles in lung cancer development. While the genome determines susceptibility to cancer or provides protection against the initiation and/or promotion of carcinogenesis, morbidity and mortality rates are usually correlated with exposure to tobacco smoking and outdoor/indoor air pollution [1,2]. In addition, exposure of particular cells to toxic metabolites, drugs, and over-thecounter medications infuences the biological mechanisms contributing to cancer cell development and growth.
One of the most important cellular transporters determining the concentration of organic cations, carbohydrates, amino acids, proteins, polysaccharides, certain cytokines, and antibiotics in the diferent parts of the body is P-glycoprotein (P-gp), encoded by the ABCB1 gene. Te principal role of P-gp is to protect against xenobiotics entering the human body from the environment/diet; as such, the protein is typically localised at sites of exposure to foreign compounds [4,5]. Tumour cells exploit the protective function of P-gp, increasing its level in response to treatment with cytostatic drugs. Moreover, while ABCB1 mRNA level is typically found at an intermediate level in normal lung tissue, and this value is downregulated in lung tumour cells [6]. Te decrease of P-gp may disturb the balance between the extracellular and intracellular environments and exacerbate the changes in the concentrations of compounds which may lead to cancer cell formation. However, as previously shown, patients with lung tumour present a similar level of ABCB1 expression in cancer tissue and blood cells [7]. Hence, it appears that diverse exogenous and endogenous compounds, which have a strong infuence on P-gp function, may infuence the normal, i.e., geneticallydetermined, level of protein activity. In addition, the presence of highly frequent polymorphisms in the ABCB1 sequence may contribute to cancer development and signifcantly determine the disease phenotype.
Te presented study tests the hypothesis that variations (T-129C, C1236T, G2677T/A, and C3435T) in the ABCB1 gene, which forms a biological barrier against toxins/xenobiotics, may infuence the risk of lung cancer and may be a prognostic marker in patients with non-small cell lung cancer (NSCLC). Candidate polymorphisms were selected on the basis of prior literature reporting a signifcant efect of these SNPs on the ABCB1 protein and/or their correlation with the clinical course of many diseases.

Materials and Methods
Te whole blood specimens from patients with non-small cell lung cancer were obtained from Nicolaus Copernicus Regional Specialist Hospital in Lodz (Lodz, Poland). Materials were collected between March 2016 and December 2020. All experiments performed in this study were in accordance with the principles of the Declaration of Helsinki and approved by the Ethical Committee of the Medical University of Lodz (No RNN/87/16/KE). All the included individuals gave informed consent.

Study
Group. Eighty patients with NSCLC (age at diagnosis between 32 and 82 years; median 68 years old) undergoing surgical resection were enrolled in this study. Inclusion criteria for this group are histopathological confrmation indicating NSCLC and its surgical treatment. Te diagnosis of small cell lung cancer or carcinoid recognized after histologic examination were the exclusion criteria. Information concerning age, gender, and smoking status was collected from each patient. At the time of diagnosis, subjects were also clinically staged according to the TNM system. Following analysis of the surgical specimen, the histological type and tumour grade were identifed. Te profle of the investigated group is given in Table 1. Seventeen of the study subjects were treated after surgery with adjuvant chemotherapy consisting of combined cisplatin E and vinorelbine or carboplatin and gemcitabine.

Control Group.
A control group was formed from 96 voluntary blood donors from the regional blood bank who were residing in central Poland. Te samples were only tested for T-129C; they had previously been tested for three ABCB1 polymorphisms (C1236T, G2677T/A, and C3435T) in the Laboratory of Molecular Diagnostics and Pharmacogenomics [8].

Methods
3.1. DNA Isolation. DNA from peripheral blood was isolated according to "Blood Mini" protocol (A&A Biotechnology, Poland). Te concentration and purity of DNA samples were assessed using a UV/VIS nanospectrophotometer (Implen). Until analysis, the DNA samples were stored at −20°C.

Polymerase Chain Reaction (PCR).
For studied polymorphisms, the PCR reactions were conducted according to "2x DreamTaq Green PCR Master Mix" (Termo Fisher Scientifc) protocol in 25 µl volumes of PCR mixture: 12.5 µl of 2x DreamTaq Green PCR Master Mix, 0.5 µl of 10 µM forward and reverse primers specifc for each of the studied SNPs, and 50 ng of DNA. Te mixture was made up to 25 µl with distilled, sterilized water. Each experiment included a negative control, i.e., without DNA. Products of the PCR reactions were assessed by electrophoresis in 2% agarose gel. Te reaction products for the SNPs at positions T-129C, C1236T, G2677T/A, and C3435T were 258, 370, 262, and 208 bp in length, respectively.

Restriction Fragment Length Polymorphism (RFLP).
Te polymorphisms T-129C and 3435 were genotyped by RFLP. Briefy, the PCR products of SNP T-129C were digested by MspA1I. Te reaction mixture consists of 2 µl of 10x bufer, 0.2 µl restriction enzyme 10 U/µl, 1.8 µl of distilled water, and 16 µl of PCR product. Restriction was performed at 37°C for 16 hours. Genotypes for SNP T-129C were identifed by electrophoresis after digestion by MspA1I (one band of 258 bp for genotype TT; two bands of 226 and 32 bp for genotype CC; three bands of 258, 226, and 32 bp for genotype CT). Te RFLP procedure used for genotyping the polymorphism at position 3435 was performed according to Zawadzka et al. [7].

Sequencing.
Genotyping for polymorphism 1236 was performed by Sanger sequencing. Te sequencing-PCR reaction (SeqPCR) was conducted according to the "Termo Sequenase Dye Primer Manual Cycle Sequencing Kit" protocol (Termo Fisher Scientifc). Te reaction mixture for SeqPCR was composed of the following: 2.2 µl of concentrate reaction bufer, 1 µl of Termo Sequenase DNA Polymerase (20 U/μl), 1 µl of 10 µM primer (labeled by IRD800 at 5′end), 1 µl of each dNTP/ddNTP mixture, and 2 µl of amplifed DNA. Te mixture was made up to a fnal volume of 20 µL with distilled water. After sequencing, PCR amplifcation loading bufer was used. Te SeqPCR products were separated into 40% polyacrylamide gels. Sequencing was performed with a LI-COR ® DNA Analyzer 4300 automated sequencer (Supplement Figure 1).
Sequencing for polymorphism 2677 was performed by the GENOMED Company (Warsaw, Poland). In this case, the sequencing reactions were performed using the BigDye ® Terminator v3.1 kit from Applied Biosystems (Termo Fisher Scientifc). Te sequencing reaction products were separated on a 3730x1 DNA Analyzer capillary sequencer (Supplement Figure 2).

Statistics.
A chi-square statistic with or without Yates correction, odds ratios (OR), and 95% confdence intervals (95% CI) were used to evaluate the associations between the investigated SNPs of ABCB1 and NSCLC risk. Te same statistical test was also used to check whether the study group is in equilibrium according to the Hardy-Weinberg principle. Te type of statistical test used for analyzing the other quantitative and qualitative clinical data is given in the corresponding supplementary tables (Supplement Tables A1-A4 and B1-B4).
Linkage disequilibrium (LD) between particular SNPs was estimated based on the r-squared statistic using EMLD software. Te D′ and r 2 parameters were estimated for pairwise combinations of investigated SNPs. Haplotype frequencies from genotyping data were estimated using PHASE, version 2.1 [11,12].
In all conducted tests, a value of p < 0.05 was considered signifcant.

Results
All NSCLC cases were successfully genotyped and included in the analysis. Most were male, and over half were current smokers. Te most common form was squamous cell carcinoma. Almost half of the cancer cases were identifed in TNM stages II or III, and most had moderately welldiferentiated tumour cells (G2). More detailed demographic and clinicopathological characteristics of this group are presented in Table 1.

Association of Genotype and Haplotype with Lung Cancer
Risk. In this group, T-129C, C1236T, G2677T/A, and C3435T were all found to be in line with the Hardy-Weinberg equilibrium (p > 0.05). Te genotyping results were compared to the control values. In both the study and control groups, the −129C allele was identifed in four subjects, none of whom were homozygous ( Table 2).
T-129C did not appear to be related to cancer risk. C1236T was associated with an NSCLC risk, with the 1236 CT genotype demonstrating a lower risk of NSCLC than the CC genotype (p � 0.0480). In contrast, the diference between the frequencies of T and C alleles between the study and control groups was not signifcant (p � 0.7535). For G2677T/A, the frequency of the rare mutant allele A was 5%. Because the study group contained only one TA and three GA carriers, these subjects were included in the mutated homozygous (TT) and heterozygous (GT) groups, respectively, for data analysis. For this SNP, signifcant diferences in neither genotype nor allele frequencies between the study and control groups were stated (p � 0.1116 and p � 0.2104). Finally, for C3435T, signifcant diferences in genotype frequencies and allele distribution were noted between cases and controls. Te lung cancer patients were more likely to harbour a TT genotype, and this was associated with a more than threefold higher risk of NSCLC (OR 3.27; 95% CI 1.33-8.06; p � 0.0099) than that of CC genotype. Accordingly, the T allele was more frequent in cancer patients than in healthy controls and connected with elevated risk of NSCLC (OR 1.75; 95% CI 1.14-2.67; p � 0.0103). Table 2 summarizes the results of four ABCB1 SNPs genotyping in patients with NSCLC and shows also genotypes/alleles distribution in healthy controls.
Te three most extensively studied SNPs of ABCB1 are known to be in strong linkage disequilibrium. Moreover, polymorphisms present in the gene regulatory region directly infuence the expression of the gene and may form important parts of this haplotype. Our present fndings confrm a disequilibrium between every pair of the four investigated ABCB1 SNPs: the D′ values range from 0.999 to 0.700, with the highest disequilibrium observed between T-129C and the other SNPs. Te r 2 values ranged from 0.667 (value between C1236T and G2677T/A) to 0.023 (value between T-129C and C1236T or G2677T/A). T-129C is much rarer than the others, as indicated by the high D′ parameter value together with a low r 2 value ( Figure 1).
To determine the efect of haplotype on NSCLC risk, the haplotype frequencies in lung cancer patients and healthy individuals were estimated, and the two groups were compared. Te most common haplotypes in the study group were T −129 -T 1236 -T 2677 -T 3435 (0.3730) haplotype, followed by T −129 -C 1236 -G 2677 -C 3435 (0.3178). In the control group, the most common was T −129 -C 1236 -G 2677 -C 3435 (0.2087). Both T −129 -C 1236 -T 2677 -T 3435 and T −129 -T 1236 -G 2677 -C 3435 were observed quite frequently in the control group (0.1300 and 0.1496, respectively). All these estimated haplotypes were found to difer signifcantly in frequency between patients and controls (p � 0.01); however, the T −129 -T 1236 -T 2677 -T 3435 haplotype was not only signifcantly diferent in frequency between the two groups, but it was also the most closely associated with NSCLC risk.

Association with Clinicopathological Features of Disease.
We hypothesized that SNPs, which show signifcant association with risk of NSCLC, may also correlate with patient demographics or the pathological features of the tumour. To verify this hypothesis, the clinicopathological data of the lung cancer patients (i.e., age, sex, cigarette smoking status, cancer type, overall stage, and diferentiation grade) were compared with the single SNP genotyping results.
No statistically signifcant correlation was found between the analyzed clinical parameters and the investigated ABCB1 gene polymorphisms (Supplement Tables A1-A4 and B1-B4). However, C1236T was correlated with age of disease onset and C3435T with smoking status. In addition, patients with heterozygous C1236T (CT) tended to be at greater age when the lung cancer was diagnosed and surgically resected compared to homozygous patients (TT) (p � 0.0410). Tis observation was not refected in the analysis of allelic variants of C1236T (Supplement Table B2). Also, nonsmokers with lung cancer were more likely to present at least one 3435 C allele than the tobacco smoking patients (p � 0.0200). Comparisons of the frequencies of the C3435T genotypes between patients stratifed according to smoking status were of only borderline signifcance (Supplement Table A4).

Association with Blood Morphology Indices.
Infammation plays an important role in the initiation, promotion, and progression of cancer. Terefore, blood morphology indices could be prognostic factors in lung cancer patients. Te fnal part of the analysis examined the potential association between the genotypes/allelic variants of ABCB1 SNPs and selected blood indices: red blood cell count, white blood cell count (and its particular components such as neutrophils, lymphocytes, and monocytes), platelet, hematocrit, and immunoglobulin level, and platelet-tolymphocyte (PLR), neutrophil-to-lymphocyte (NLR), and lymphocyte-to-monocyte (LMR) ratio. No association was found between the investigated parameters. A detailed list of results obtained for particular SNPs is shown in the supplementary material (Supplement Tables B1-B4).

Discussion
Te protein encoded by the ABCB1 gene plays a key role in the defence of respiratory tract cells against xenobiotics, as demonstrated by its presence, inter alia, in the bronchial epithelium, mucinous glands, alveolar cells, endothelium, and normal nasal respiratory mucosa [13]. Additionally, ABCB1 gene expression is reduced in lung cancer tissue taken during surgery compared to normal cells [6]. It has been suggested that some allelic variants present in this gene may induce interindividual diferences in P-gp activity, by infuencing mRNA level, protein formation, cellular concentrations, or substrate specifcity [14][15][16][17]. Tis ABCB1 genetic variability can support the process of carcinogenesis at diferent stages, acting as a risk factor or modifying the disease phenotype. Despite the high number of polymorphisms identifed in the ABCB1 gene and the fact that the most frequent allelic variants have been studied in various diseases, the impact of particular SNPs on lung cancer development remains tentative. In response to the limited number of studies addressing the infuence of ABCB1 on NSCLC risk and course of disease, the present study examined the relationships between four selected SNPs and their haplotype with NSCLC susceptibility, patient demographic characteristics, tumour pathomorphological features, and blood morphology. As a result, this study is the frst such comprehensive analysis of these SNPs, lung cancer risk, and disease phenotype in a Polish population.
Our studies found the frequencies of ABCB1 −129C, 1236T, 2677T/A, and 3435T alleles to be 5.0%, 62.5%, 67.9%, and 86.3%, respectively. Te allele frequency distribution did not deviate signifcantly from data published for a similar Caucasian population (https://www.ncbi.nlm.nih.gov/snp/) for lung cancer [18,19] or other malignant tumours [20,21]. Our data indicates that for SNP 3435, the TT genotype and the T allele were more common in lung cancer patients than in healthy subjects, and the presence of this genotype (TT) increased the risk of lung cancer by more than threefold. Our data confrm those of Subhani et al. (2015) in a South Indian lung cancer population or Sheng et al. (2012) in Caucasian and Asian patients with hematologic malignancies, breast cancer, and renal cancer [22,23]. However, the latter metaanalysis based on two studies by Sheng et al. did not confrm any association between lung cancer incidence and C3435T genotype in the Caucasians. A similar analysis in Spain also yielded negative results [24].
Regarding the composition of the patient groups, investigated by Gervasini et al., they were characterised by various histopathologic types of lung cancer, with a signifcant predominance of squamous cell carcinoma and a low rate of adenocarcinoma; this is inconsistent with the general histopathologic pattern of lung cancer types [25]. Tis may indicate the presence of some additional factor/factors with carcinogenic potential in this population. In Gemignani et al., the presence of heterozygous CT at that locus was associated with an increased risk of lung cancer, with equal numbers of adenocarcinoma and squamous cell carcinoma [19]. Finally, Sinués et al. report no correlation between SNP C3435T and lung cancer incidence, neither for the total group of lung cancer patients nor after stratifcation according to histological type [24].
It should be noted that, in line with the fndings from all studies given above, no correlation between the presence of C3435T and NSCLC histological type was found in the present study. Moreover, meta-analysis data indicates that the C3435T polymorphism does not contribute to lung cancer risk among an Asian population [23]. Similar conclusions were drawn for studies based on gastrointestinal or ovarian cancer patients, regardless of ethnic group [16,21,23,26]. Although further analysis is needed to verify whether C3435T afects the risk of lung cancer development in Caucasian populations, it seems that both cancer tissue and the patient population infuence the role of this synonymous polymorphism in the pathogenesis of cancer.
Te C3435T polymorphism demonstrated strong linkage disequilibrium with the other studied forms, one of which is the nonsynonymous G2677T/A, which has been investigated in comparatively small numbers in cancer research. Our fndings suggest that G2677T/A has no efect on NSCLC risk, which is in line with the hypothesis that G2677T/A only enhances the efect observed for synonymous SNPs and has little impact on biological phenotype when analyzed separately. No diference in the frequency of G2677T/a genotypes was observed between patients and controls in a study of lung cancer in Eastern European countries [19]. In contrast, Gervasini et al. suggest that the 2677T variant can be crucial in conferring susceptibility to lung cancer in the area of southwest Spain; however, this group demonstrated a lower presence of G allele than in the population of Polish patients with lung cancer [25].
Te third most frequently observed polymorphism in the ABCB1 gene is the nonsynonymous C1236T. In what appears to be the frst study to examine this SNP in the context of early onset lung cancer, Gemignani et al. report it to not be associated with the susceptibility of lung cancer [19]. Our present data indicate that heterozygous (CT) occurs signifcantly more frequently in healthy subjects than in cancer patients. Unfortunately, due to the limited literature base, it is not possible to gather a sufcient amount of data to verify our fndings. Te present study is also the frst to evaluate the distribution of genotypes and allelic variants of the SNP T-129C in the gene promoter region among NSCLC individuals. Although T-129C was not found to infuence lung cancer pathogenesis, the low frequencies of the C allele in the general population may contribute to negative outcomes. After correlating the ABCB1 genotyping results with age, gender, smoking, histological type of cancer, tumour stage, and grade status, none of the genetic polymorphisms were found to have any infuence on lung cancer phenotype. In contrast, Subhani et al. indicated that the homozygous TT of SNP C3435T showed signifcant association with advanced lung tumour stage, and interestingly, that lung cancer subjects who smoke cigarettes demonstrated more frequent occurrence of 3435TT compared to nonsmokers [22]. Despite this, other studies report no statistically signifcant correlation between this SNP and smoking status [19,24,25]. For the C1236T, a marginally signifcant correlation was observed between 1236 CTgenotype and the age at recruitment to the surgical resection of lung cancer; however, this observation has not been confrmed elsewhere.
In the study presented here, a higher frequency of -T −129 -C 1236 -T 2677 -T 3435 haplotype was noted in NSCLC patients than in healthy controls. Unfortunately, again, due to the paucity of studies currently assessing the ABCB1 haplotype in NSCLC, it is not possible to draw reliable conclusions [25]. However, studies on small cell lung cancer patients have noted that T/A 2677 -T 3435 carriers showed signifcantly higher survival probability after chemotherapy [27], which is consistent with our present fndings. Te prognostic role of ABCB1 polymorphisms and haplotype is also supported by studies on, inter alia, colorectal [28] and breast cancer [29]. Tus, the presence of mutated variants of particular SNPs may be associated with the rick of a malignant neoplasm of epithelial origin.
Te linkage between the C3435T SNP and other mutations in the coding region of the ABCB1 gene or its promoter/enhancer may account for the direct contribution of the silent mutation to disease predisposition. Te T variant of SNP C3435T has been associated with lower mRNA stability, diminished P-gp function, and decreased removal of xenobiotics. For G2677T/A, the presence of the T or A variant was connected with transport-related structural changes of P-gp. Finally, C1236T is reported to afect ABCB1 mRNA stability [17], while the presence of −129TT genotype predisposes to greater DNA damage in subjects exposed to toxic compounds [30]. When combined, such small changes may result in an important reduction of the cellular protective mechanism.
In addition to pumping xenobiotics outside the cell, P-gp inhibits tumourigenesis by working synergistically with immune function and being widely expressed in immune system cells. Tere is evidence that this protein plays a role in the regulation of cell activity, secretion of cytokines, cytotoxic function, and survival by T, natural killer (NK), and B lymphocytes [31,32]. Moreover, cancer patients demonstrate changes in the composition and characteristics of blood immune cell subpopulations as a result of interaction between immune system cells and the tumour microenvironment [32]. Additionally, increased activity of these immune cell subpopulations is commonly associated with improved prognosis [33]. However, neutrophil and platelet activation have been found to correlate with reduced patient survival [33,34]. Terefore, the NLR, LMR, and PLR in cancer patients may be a useful prognostic marker. Terefore, the present study examines whether genotypes/ allelic variants of ABCB1 polymorphisms directly infuence blood morphology parameters or the indicators based on them. Our present fndings indicate that this is probably not the case. Further analyses are essential to better understand the characteristic pattern presented in the immune system of lung cancer patients. PLR and NLR have been shown to be associated with greater survival, including among patients with NSCLC [35].
Te results of the present study should be interpreted with caution because this research is not without limitations. Te frst is the small sample size. Furthermore, in the present study, the majority of patients enrolled to the study were lost to follow-up and the data about survival time after surgery was incomplete. Terefore, it is impossible to assess the infuence of ABCB1 polymorphisms on survival, this being another limitation of this study.
In conclusion, our present fndings indicate that the presence of SNP C3435T of the ABCB1 gene, either combined with G2677T/A, C1236T, or T-129C or as an independent factor, appears to be associated with NSCLC risk. Te remaining three loci in this haplotype appear to be marginally signifcantly associated with the presence of NSCLC in this study population. Our present study also investigated the potential association of ABCB1 -129/1236/ 2677/3435 haplotype with NSCLC development. Our fndings indicate that the lung cancer and control groups demonstrate signifcantly diferent patterns of haplotype distribution. Of particular note is that a higher frequency of -T-129-C1236-T2677-T3435 haplotype was noted in NSCLC patients than in healthy controls. Te investigated polymorphisms were not found to be related to the demographic characteristics of lung cancer patients, tumour histopathological results, or blood morphology indices. However, further studies are needed to better understand the efect of genetic variability and mechanisms on lung cancer development. Nevertheless, the presented results provide further evidence that ABCB1 gene polymorphisms and their haplotype might be genetic risk factors and potential biomarkers for lung cancer.

Data Availability
Te data used to support the fndings of this study are available from the corresponding author on reasonable request.

Conflicts of Interest
Te authors declare that there are no conficts of interest regarding the publication of this article.