Introduction

With 9.9% of worldwide-diagnosed cancers, colorectal cancer (CRC) accounts as the third most common, and one of the leading causes of cancer-related deaths [1]. More than 1 million people are diagnosed with CRC each year, which additionally emphasizes the statement that CRC represents one of the major global health issues [2]. Adenocarcinoma, deriving from colorectal mucosa epithelium, sums approximately 90% of all histological CRC subtypes [3].

While hereditary CRC is characterized by alterations of highly penetrant gene alleles, it accounts for only a small percentage of overall CRC. Majority of CRC cases develop due to sporadic alterations of low penetrant gene alleles, alone or in combination with environmental factors [4]. Apart from genetic, several modifiable risk factors have been recognized in CRC development, including nutritional habits, obesity, smoking and alcohol consumption [5].

Despite being quite common in general population, recognized CRC risk factors might not be single contributors of CRC development, suggesting that carcinogenesis is influenced by inter-individual genetic variations. Indeed, previous studies suggest that glutathione S-transferases could be linked with the development and progression of CRC [6, 7].

Glutathione-S transferases (GSTs) are large family of xenobiotic-conjugation enzymes [8]. GSTs are particularly important in the detoxification process of heterocyclic aromatic amines and polycyclic aromatic hydrocarbons, widely recognized risk factors of CRC development found in processed meat and tobacco [9]. Polymorphism in GSTs often leads to alteration or even complete lack of enzyme activity, which, due to their important role in detoxifying carcinogens, might have an effect on CRC carcinogenesis.

Bearing in mind that GST polymorphisms may play a significant role in the CRC development, a comprehensive study was conducted, aiming to determine the presence of established risk factors and specific GST gene variants in CRC patients, as well as, evaluation whether phenotype changes reflect genotype-associated risk.

Material & Methods

Study Population

Newly diagnosed CRC patients, treated and followed at the Digestive Surgery Clinic, Clinical Center of Serbia, Belgrade, were included in this study between the years of 2014–2016. All 523 patients (313 men, 210 women; average age 62.25 ± 11.38 years) had their diagnosis histologically confirmed in accordance with TMN and Dukes classification [10]. The control group included 400 individuals (203 men, 197 women, average age 60.40 ± 12.31 years) who had undergone surgery for benign conditions, unrelated to both non-malignant and malignant GI conditions at the same clinical center. The structured questionnaire composed at the Institute of Epidemiology, Faculty of Medicine University in Belgrade was used for acquiring patients’ individual level-data. Regarding that, obese patients were classified as individuals with BMI above 25 and smokers as individuals who reported smoking during a minimum of 60-days period before they have entered the study. Additionally, for the purpose of calculating the pack-years data regarding number of smoked cigarettes, as well as, the duration of smoking were obtained. The study was approved by the Institutional Ethical board (approval number 56–6, Clinical center of Serbia) and was performed in accordance with principles of Helsinki declaration. Informed written consent was obtained from all recruited subjects.

DNA Isolation and Genotyping

Genomic DNA was isolated from 200 μl of the whole peripheral blood by QIAamp DNA Blood Mini Kit (Qiagen, USA) according to the manufacture’s protocol. The multiplex PCR technique used to detect homozygous deletions of GSTM1 and GSTT1 included primers for GSTM1, GSTT1 and CYP1A1 housekeeping gene, according to the method by Abdel-Rahman et al. [11].

PCR-restriction fragment length polymorphism (RFLP) method with Eam1104I (Thermo Fisher Scientific, USA) restriction enzyme was used for the analysis of the GSTA1 C69T (rs3957357) SNP according to the method by Ping et al. [12].

The analysis of GSTP1 Ile105Val (rs1695) SNP was performed according to the manufacturer’s instructions using the Applied Biosystems TaqMan® Drug Metabolism Genotyping assay (Life Technologies, Applied Biosystems, USA, assay ID: C__3237198_20).

Statistical Analysis

Differences in investigated parameters were assessed using Student’s T test/ANOVA for continuous data with normal distribution and Mann–Whitney rank-sum test for continuous data with non-normal distribution. χ2 test was used for categorical variables. The genetic variants and their risk for disease were computed by odds ratios (OR) and 95% confidence intervals (CI) by logistic regression analysis. OR was either crude or adjusted by BMI, as well as by age and gender. Results were considered to be statistically significant if p value was ≤0.050.

Results

Baseline and, clinical characteristics of patients and respective controls are summarized in Tables 1 and 2. As presented, CRC group comprised 1.5 times more male than female patients. CRC patients and controls did not differ in terms of age and BMI (p > 0.05). However, more than a half of the patients (58%) suffered from hypertension compared to controls (30%) and the CRC patients’ group had more active smokers (68% vs. 48%). As far as clinical characteristics of CRC, the most frequent location was rectum (55%), whereas the majority of patients were diagnosed with well-differentiated CRC (77) and with T3 stage (49%).

Table 1 Baseline characteristic of CRC patients and controls
Full size table
Table 2 Clinical characteristic of patient’s tumor
Full size table

GST Genotypes and CRC Risk

The frequency of GST genotypes, as well as, the relation to the CRC risk in both CRC patients and controls is presented in Table 3. While individual GSTM1-null and GSTA1 CT + TT (low activity) genotypes did not significantly contribute to the risk of CRC development, GSTT1-null and GSTP1 IleVal + ValVal (variant) genotypes were significantly associated with higher risk of CRC. Namely, the carriers of the GSTT1-null were at 1.35-fold increased risk for CRC development (OR = 1.35, 95%CI: 0.99–1.83, p = 0.050) when compared to individuals with GTST1-active genotype, whereas carriers of GSTP1-variant genotype were more frequent among patients (63% of CRC patients compared to 55% of controls) contributing to 1.39-fold increased risk of CRC development (95%CI: 1.07–1.82, p = 0.016) (Table 3).

Table 3 GST genotypes in relation to the risk of CRC
Full size table

Combined Effect of GST Genotypes on CRC Risk

Combined effect on CRC development for both null and variant GST genotypes was further examined (Table 4). No significant relationship in terms of CRC risk was observed when the combined effect of any two of GSTM1, GSTA1 and GSTT1 genotypes was assessed (p > 0.05, data not shown).

Table 4 Combined effect of GST genotypes and CRC risk
Full size table

On the other hand statistically significant association was found in all combinations of GSTP1-variant genotype and any of other three GST genotypes (GSTP1 variant and GSTM1-null: OR = 1.53, 95%CI: 1.01–2.18, p = 0.03; GSTP1-variant and GSTA1 low-activity: OR = 1.83, 95%CI: 1.22–2.75, p = 0.004; GSTP1-variant and GSTT1-active: OR = 1.95, 95%CI: 1.27–2.99, p = 0.002) when compared to referent genotype combination. Moreover when combined effect of any three GST genotypes was analyzed statistically significant connection regarding GST polymorphism and susceptibility to CRC was noticed for all triple genotype combinations involving GSTP1 genotype (GSTM-null/GSTT1-active/GSTP1-variant OR = 2.07, 95%CI: 1.06–4.02, p = 0.03; GSTM-null/GSTA1 low activity/GSTP1-variant OR = 2.24, 95%CI: 1.24–4.04, p = 0.001; GSTT1-active/GSTA1 low activity/ GSTP1-variant OR = 2.62, 95%CI: 1.41–4.87, p = 0.001) when compared to referent genotype combination (Table 5).

Table 5 Combined effect of three GST genotypes and CRC risk
Full size table

Last but not least, we have examined combined effect of all four variant GST genotypes and observed an increasing trend in OR with the highest risk of 2.71-fold (95%CI: 1.06–6.91, p = 0.037) in individuals carrying all four variant GST genotypes (GSTM1-null, GSTT1-null, GSTA1 low-activity and GSTP1-variant) in comparison with the reference genotype combination (Table 6).

Table 6 Cumulative effect of GST risk-associated genotypes on the risk of CRC development
Full size table

The Association between GST Genotypes and Pathological Differentiation and Stages of CRC

The possible effect of GST polymorphisms on CRC pathological differentiation, as well as, progression was further assesed and no association was observed in case of GSTM1 and GSTT1 genotypes (figures not shown). However, results concerning the GSTA1 low activity genotype showed borderline significant difference with respect to tumor stage (p = 0.054, Fig. 1). Moreover, significant association was found in the case of GSTP1 genotype and the level of tumor pathological differentiation, where the vast majority of patients, diagnosed with poor tumor differentiation (89%) had GSTP1-variant genotype (Fig. 2a). Additionally, an increasing trend in the frequency of GSTP1-variant genotype was noticed when the association between this genotype and tumor stage was investigated in CRC patients. (p = 0.163, Fig. 2b).

Fig. 1
figure 1

The frequency of GSTA1 variant risk-associated genotypes in CRC patients stratified according to tumor grade

Full size image
Fig. 2
figure 2

The frequency of GSTP1 risk-associated genotype in CRC patients stratified according to: a tumor grade; b tumor stage

Full size image

Discussion

The role of GST polymorphisms as potential contributing risk factors in the process of CRC carcinogenesis is disputable. In this study, we have assessed four most common GST polymorphisms in terms of CRC susceptibility and found that individuals carrying GSTT1-null or GSTP1-variant genotypes had significantly higher risk of CRC development, which was even more noticeable in their combination or when third GST genotype, either GSTM1 or GSTA1 was added. Additionally, CRC patients with combined GSTM1-null/GSTT1-null/GSTA1 low-activity/GSTP1-variant genotype had 2.71-fold increased susceptibility to CRC.

Given that modifiable environmental risk factors have an important role in the complex course of CRC development and progression, genetic polymorphisms in xenobiotic metabolyzing enzymes GSTs have been widely investigated [9, 13]. Nevertheless, although vast, existing results on GST polymorphism and CRC are quite debatable.

GSTM1 null genotype is mostly examined GST polymorphism in relation to increased risk to various cancers. However, due to diverse ethnicity and geographical origin of patients included in various studies. Association of GSTM1 polymorphism and CRC risk varies, since GST genotype is known to be influenced by both of these factors [14]. In a large meta-analysis by Economopoulos et al. [15] individuals with GSTM1-null or GSTT1-null genotype were at higher risk for CRC development, while presence of either one or both GSTA1 or GSTP1-variant alleles did not further contribute to risk of CRC development. However, all four GST genotypes were investigated in Caucasian population in only three of the studies included in this meta-analysis [16,17,18]. Still, the results obtained in two of them [15, 17] indicated no significant association between GSTM1-null genotype and CRC risk, which is in accordance with the results obtained in our study. Regarding the distribution of GSTM1 genotype in our study cohort, it was similar with the distribution in other studies conducted on Caucasians [19,20,21], while, as expected, in contrast to numerous studies on Asian populations [22]. Combined effect of GSTM1-null and GSTT1-null analyzed in the study of Martinez et al. showed an increased risk of CRC in carriers og this genotype combination [16], which is partially in agreement with our results, since the observed association in our case was not statistically significant (OR = 1.50, 95%CI: 0.95–2.35, p = 0.080).

Given that that deletion of GST1 gene leads to lack of its expression in intestinal tract and detoxifying ability, GSTT1-null was also GST polymorphism of interest in numerous studies [23,24,25]. It seems that GSTT1-null genotype is recognized as a CRC risk factor, especially in Caucasians [24]. Previous meta- analysis of Qin et al. and Wan et al. [25, 26] suggested that GSTT1-null genotype conferred a 1.21- fold and 1.32-fold increased risk of CRC, respectively. What is more, data from the recent case-control study of Masood et al. [27] also showed that GSTT1 deletion is associated with elevated CRC risk. Results of our study, on 1.35-fold increased CRC risk in carriers of GSTT1-null genotype, are consistent with mentioned reports. This might be of a particular importance in the case of GSTT1-null individuals exposed to compounds recognized as the GSTT1 substrates, that are readily recognized as contributing factors to the CRC development, such as ethylene oxide derived from ethane, abundant in cigarrete smoke [28].

Another examined GST polymorphism in our study was GSTA1. Interestingly, this polymorphism has been investigated to a much lesser degree ili extent compared to other GSTs. Martinez et al. [17] were the first to investigate GSTA1, amoung other examined classes of GSTs in CRC patients. Results of this study suggest that GSTA1 does not affect the predisposition for CRC, although deficiency of GSTA1 enzyme activity might influence carcinogenesis in persons previously exposed to environmental hazards, considering its known role in the process of detoxification. In previously mentioned meta-analysis of Economopoulos et al. [14] individual GSTA1 polymorphism was also not associated to CRC risk. Likewise in the results of Hezova et al. [19] there was no association of GSTA1 polymorphism and CRC. In this line, recent meta-analysis of Deng et al. [28] also suggested that GSTA1 polymorphism has not been recognized as CRC risk factor in majority of included studies [29]. Results of our study are consistent with all reported data. One of the probably most widely investigated GST polymorphisms is GSTP1. GSTP1 is shown to be often excessively expressed in various tumors, including CRC [30], suggesting it’s involvement in metabolism of various carcinogenic substances [19]. Similarly to other GSTs, the distribution of GSTP1 genotypes differs among populations with various geographical and ethnical origin, which significantly contributes to conflicting results regarding the role of GSTP1 polymorphism in CRC development and progression. Recent analysis by Tan et al. [31] and Economopoulos et al. [15] did not show an association between GSTP1-variant genotype and increased risk for CRC, which is not in accordance with our results. However, results of Matakova et al. [32], Wang et al. [33] and Kassab et al. [34], as well as, meta-analysis by Ramsey et al. [6], were all consistent regarding the significant role of GSTP1 polymorphism is susceptibility to CRC development, what our results agree with.

Aside from analyzing the association of independent GST polymorphisms with CRC risk, we have assessed the presence of the combined effects of two, three and all four putative risk genotypes of GST gene variants with respect to CRC risk.

Firstly, we have assessed the combined effects of two presumed risk genotypes with respect to low-risk GST genotypes. Of all enrolled CRC patients, 40% had the combination of GSTT1-null with GSTP1-variant, with a 1.95-fold increased CRC risk. The combination of GSTM1-null and GSTP1-variant was present in 64% CRC patients, whereas GSTA1 low-activity and GSTP1-variant in 76% of all patients, with increased CRC risk 1.53-fold and 1.83-fold, respectively. Majority of the data on combined GST genotypes are related to combination of GSTM1-null and GSTT1-null [23, 33, 34]. The study of Matakova et al. [32] suggested that combination of GSTM1-null and GSTP1- variant genotypes is a predisposing factor for CRC development. Combined GSTT1- null/GSTP1-variant genotype indicated 1.89-fold increased risk for CRC development in the study of Wang et al. showed, which is rather similar to the results of our study. Combined GSTA1 low-activity/GSTP1-variant genotype was of interest in only one study related to colorectal adenoma and cruciferous vegetables [20], whereas this study is the first to asses this gene-gene interaction alone.

Later, we have analyzed three putative risk genotypes in respect to low risk GST genotypes in patients with CRC. Combination of GSTM-null, GSTT1-active and GSTP1-variant genotype showed 2.07-fold increased risk for CRC, which is in agreement with the results of Ates et al. [35] and Wang et al. [33]. However, for the first time to our knowledge, our study has demonstrated an increased risk of CRC development in individuals carrying combined GSTM1-null/ GSTT1-null/ GSTA1 low-activity / GSTP1-variant genotype, since no data regarding the combined effect of all four GST genotypes are available in the literature.

Several limitations of this study need to be addressed. First of all, in order to estimate the association between GST genotypes and the risk of CRC we used a case-control design and therefore the selection bias might influence the results. Additionally, our controls were hospital-based instead of population based, comprising white subjects only, therefore the possible effect of ethnicity could not be assessed. A recall bias regarding the questions from the questionnaire might have influenced the results as well. Furthermore, the data on dietary habits were not validated, hence not used in the adjusted analysis of the obtained results.

Conclusion

This study supports hypothesis that GST polymorphisms might have an important role in the process of the CRC development. Additionally, combinedGSTM1-null,/GSTT1-null/GSTA1 low-activity/GSTP1-variant genotype could be potential combination of four common GST genotypes making its carriers more prone to CRC development. Prospectively, further case-control studies assessing the association of GST gene variants and CRC risk should be preformed on patients with individual-level data being collected especially regarding dietary habits, which might further contribute to the adjusted risk analysis.