Abstract
Background/Aim: Long-term exposure to betel quid (BQ)-, cigarette-, and alcohol-induced chronic inflammation is a crucial risk factor for oral and pharyngeal squamous cell carcinoma (OPSCC) progression. We analyzed the genotypes of stromal-cell-derived factor-1 (SDF-1) and CXC-chemokine receptor-4 (CXCR4) and determined the association between their polymorphisms and the risk of OPSCC. Materials and Methods: This study consisted of 452 patients with pathologically proved OPSCC and 424 sex- and age-matched cancer-free controls. The genotypes of SDF-1 and CXCR4 were detected through the TaqMan real-time polymerase chain reaction (PCR) method. Results: Our data indicated that the C allele and C/C genotypes of CXCR4 were significantly associated with OPSCC [adjusted odds ratio (AOR)=1.41, 95% confidence interval (CI):1.02-1.96, p=0.037 and AOR=1.51, 95% CI:1.05-2.17, p=0.028, respectively] and OSCC (AOR=1.41, 95%CI:1.00-2.00, p=0.049 and AOR=1.49, 95%CI:1.01-2.20, p=0.044, respectively) risk. Patients with genetic polymorphisms of the genotype combination SDF-1/CXCR4 had a higher risk of OSCC (p trend=0.033). We analyzed the effects of CXCR4 genetic variants on susceptibility to OPSCC in patients with different risk habits of BQ chewing, tobacco smoking and alcohol consumption, and revealed that C/T+T/T genotypes exerted an increased risk only in patients with one (AOR=2.68, p=0.036) or two risk habits (AOR=2.02, p=0.027) compared to patients with the C/C genotype. Conclusion: We concluded that CXCR4 C>T can be used as a genetic marker of susceptibility to OPSCC, particularly in OPSCC patients with one or two types of risk habits with a synergistic effect.
- SDF-1
- CXCR4
- oral cancer
- SNP
Oral and pharyngeal cancer is an critical public health issue worldwide and leads to cancer morbidity and mortality worldwide (1). Squamous cell carcinoma (SCC) accounts for >90% of all oral and pharyngeal cancers (2). Three types of substance use, namely betel quid (BQ) chewing, cigarette smoking, and alcohol drinking, have been identified as major risk factors for oral and pharyngeal SCC (OPSCC) (3, 4). BQ ingredients induce oral keratinocyte to secrete tumour necrosis factor (TNF)-alpha and interleukin (IL)-6, which may provoke oral and pharyngeal mucosal inflammation (5). Cigarette smoking causes a drastic change in immunity that leads to increased constitutive inflammation and suppressed antitumoral immune cell responses (6). It also causes inappropriate priming and activation of monocytes and neutrophils (7). Acetaldehyde, rather than alcohol itself, is responsible for the carcinogenic effect of alcohol consumption through induction of inflammation and enhancement of cell injury (8, 9). Therefore, long-term exposure to BQ, cigarette, and alcohol may induce OPSCC through chronic inflammation of oral and pharyngeal mucosa. However, only a fraction of cigarette smokers, alcohol drinkers, or BQ chewers develop OPSCC, suggesting the existence of genetic susceptibility to this type of cancer (10). Our previous studies have shown that the interaction of genetic polymorphisms of inflammation-related genes with substance use were associated with the risk of OPSCC (11, 12).
Chemokines are mediators of acute and chronic inflammation (13). They have low molecular weights (approximately 8-17 kDa) and were originally defined as potent attractants for leukocytes (13). Stromal-cell-derived factor-1 (SDF-1) is produced by stromal cells of mesenchymal origin, including fibroblasts, osteoblasts, and endothelial cells, and by haematopoietic stem/progenitor cells and endothelial stem/progenitor cells of the bone marrow (14-16). SDF-1 is a powerful chemoattractant for monocytes and T cells and regulates the recruitment of lymphocytes from the blood to the endothelium (17, 18). Two studies have demonstrated that SDF-1 expression is correlated with the invasion and metastasis of head and neck SCC (19-21). The AA genotype of the single nucleotide polymorphism (SNP) at position 801, 3’-untranslated region (3’-UTR), leads to higher SDF-1 secretion (22). The chemokine SDF-1 is the sole ligand of CXC-chemokine receptor-4 (CXCR4), which is widely expressed in monocytes, neutrophils, lymphocytes, epithelial cells, and a majority of tumours, including those of epithelial, mesenchymal, and hematopoietic origin (23, 24).
Previous studies have revealed that the SDF-1/CXCR4 axis contributes to cell motility in OSCC and is well correlated with OSCC progression (25-28). An immunohistochemical study has demonstrated that all SCC tissues of the tongue express CXCR4, whereas no CXCR4 expression could be detected in almost any normal epithelial tissues of the tongue (29). Another study has revealed that the expression levels of CXCR4 significantly decreases during cerebellar development and is specifically upregulated in medulloblastomas of the desmoplastic/ nodular subtype (30). These results imply that CXCR4 expression is essential for the development of the cerebellar cortex and in the pathogenesis of medulloblastomas. The same study has also identified two variants that may affect the binding of SDF-1 by changing codons 53 and 97, which are located in the transmembrane region (30). In addition, a silent sequence variant with a C/T substitution (+414 C > T; rs2228014) has been found in medulloblastomas, with no association between the +414 C>T variant and the subtype of medulloblastomas (30). Until now, numerous studies have reported that SDF-1 rs1801157 and CXCR4 rs2228014 are significantly associated with susceptibility to carcinogenesis in humans (28, 31-38). However, the association of SDF-1 and CXCR4 polymorphisms with the risk and the clinicopathological development of OPSCC remain inconclusive and controversial (39, 40). Teng et al., have reported that patients with the A/G genotype of SDF-1 had a 1.86-fold higher risk of oral cancer than those with the G/G genotype (39). Vairaktaris et al., have reported that the G/A genotype of SDF-1 had no significant impact on susceptibility to oral cancer, but patients with this genotype had a 0.33-fold lower risk of oral cancer in the advanced stage (40).
In Taiwan, approximately 2 million people have the habit of chewing BQ. Most BQ chewers are also cigarette smokers (86%) or alcohol drinkers (74%) (41). The effects of the interactions of the SDF-1 and CXCR4 genetic variants with the habits of BQ chewing, tobacco smoking, and alcohol consumption on the risk of OPSCC remain uncertain. Therefore, we investigated the relationship between SDF-1/CXCR4 genetic variants and susceptibility to OPSCC. We also evaluated the effect of interactions between SDF-1/CXCR4 genetic variants and substance use on OPSCC risk.
Materials and Methods
Participants. In total, 452 patients with primary OPSCC were recruited from the Department of ENT and Dentistry at the Kaohsiung Veterans General Hospital (KSVGH). These OPSCCs included 344 (76.1%) oral cavity, 72 (15.9%) oropharynx, and 36 (8.0%) hypopharynx cancers. In addition, 424 non-cancer controls were also recruited in this study. The detail process and information is described in our previous study (12). The study protocol was approved by the Institutional Review Board of KSVGH (VGHKS10-CT11-17). The methods were carried out in accordance with the approved guidelines and all patients provided informed consent.
Polymorphism genotyping. Genomic DNA of whole-blood samples was extracted and purified using the QIAamp® DNA extraction kit (Qiagen Sciences®, Germantown, MD, USA). Genotypes of SDF-1 G>A (rs1801157) and CXCR4 C>T (rs2228014) were detected using the TaqMan real-time polymerase chain reaction (PCR) method and were subsequently analysed using the ABI PRISM 7500 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). The genotyping assay was performed as previously described (12).
Statistical analysis. The Hardy–Weinberg equilibrium of SDF-1 and CXCR4 was evaluated in the controls by using a chi-squared test with 1 df. The crude odds ratios (ORs) and 95% confidence intervals (CIs) of SDF-1 and CXCR4 polymorphisms between the patients and controls were calculated through logistic regression. Following further adjustment of various confounders, the association of allele types and genotypes of the SNPs in SDF-1 and CXCR4 with the risk of OPSCC was evaluated using multiple logistic regression models. Because only one female participant was present in the pharyngeal SCC (PSCC) group, sex was not included in the multiple logistic regression models for analyzing the PSCC risk. The association between clinicopathological outcome and polymorphisms was estimated using the Kaplan–Meier method. The statistical analysis has been described in detail in our previous study (12). The statistical software packages SPSS (version 12.0, SPSS Inc., Chicago, IL, USA) and SAS/Genetics (version 9.1.3, SAS Institute, Inc., Cary, NC, USA) were used for all statistical analyses. Differences with p<0.05 were considered significant.
Results
Demographic and substance-use features of the participants. This case–control study consisted of 452 patients with pathologically proved OPSCC [comprising 344 with oral cavity SCC (OSCC) and 108 with PSCC] and 424 sex- and age-matched cancer-free controls. The controls and three patient groups were well matched in terms of age and sex, but the PSCC group had a slightly higher proportion of male participants compared to the control group (p=0.086) (Table I). As expected, BQ chewing, cigarette smoking, and alcohol drinking were associated with an increased risk of OSCC, PSCC, and OPSCC. The PSCC group contained a considerably greater proportion of heavy smokers and heavy drinkers compared to the control (Table I) and OSCC groups (heavy smoker vs. never smoker, COR=5.07, 95%CI=2.12-12.15, p<0.001; heavy drinker vs. never drinker, COR=6.27, 95%CI=3.15-12.49, p<0.001, data not shown). In addition, most substance users (73.80%) use more than one type of substance. The combined effect of the three types of substance use (namely BQ chewing, cigarette smoking, and alcohol drinking) on OPSCC risk was, therefore, evaluated in this study. Compared to those without any substance use (habit free), participants with two or more types of substance use had an increased risk of OSCC (COR=3.64-9.39, 95%CI=2.28-14.98), PSCC (COR=11.17-50.15, 95%CI=3.26-165.09), and OPSCC (COR=4.22-12.53, 95%CI=2.69-19.66) (Table I). However, participants using only one type of substance did not have an increased risk of various types of OPSCC. The consumption volume was significantly lower in participants using only one type of substance than in those using more than one type of substance (p<0.001 for smoking and p=0.019 for drinking), except for BQ chewing (p=0.57). Only 16 participants had only BQ chewing habit, thus, the statistical power might be insufficient for the difference.
Association of genotype and allele type of SDF-1 G>A (rs1801157) and CXCR4 C>T (rs2228014) with OPSCC risk. The genotypic frequencies of SDF-1 G>A and CXCR4 C>T among the controls did not differ from the expected distributions based on the Hardy–Weinberg equilibrium (p=0.857 and p=0.799, respectively). The genotypic and allelic frequencies of SDF-1 G>A and CXCR4 C>T among the different patient groups and controls are presented in Table II. We found that the genotypic and allelic types of SDF-1 G>A were not correlated with the risk of OSCC, PSCC, and OPSCC. However, compared to the participants with the C/T+T/T genotype or the T allele of CXCR4 C>T, those with the C/C genotypes or the C allele exhibited an approximately 1.41- to 1.49-fold increased risk of OSCC [adjusted OR (AOR)=1.49, 95%CI=1.01-2.20, p=0.044; AOR=1.41, 95%CI=1.00-2.00, p=0.049; respectively] and OPSCC (AOR=1.51, 95%CI=1.05-2.17, p=0.028; AOR=1.41, 95%CI=1.02-1.96, p=0.037; respectively). A significant linear trend was also observed for the increasing risk of OSCC and OPSCC as the number of C alleles at the locus of CXCR4 C>T increased (p for linear trend=0.047 and 0.036, respectively).
Combined genotypes of SDF-1 G>A and CXCR4 C>T between OPSCC patients and controls. When analysing the genotypic combination of two SNPs between patients and controls (Table III), we regarded the G allele of SDF-1 G>A and the C allele of CXCR4 C>T as the risk alleles on the basis of the aforementioned preliminary results. The various genotypic combinations were categorised into three groups according to the number of risk alleles. Group I had a genotypic combination of none, one, and two risk alleles, such as SDF-1-AA/CXCR4-CC, SDF-1-AG/CXCR4-CT, SDF-1-GG/CXCR4-TT, SDF-1-AA/CXCR4-CT, SDF-1-AG/CXCR4-TT, and SDF-1-AA/CXCR4-TT. Group II had a genotypic combination of three risk alleles, such as SDF-1-AG/CXCR4-CC and SDF-1-GG/CXCR4-CT. The genotypic combination of SDF-1-GG/CXCR4-CC with four risk alleles was classified as group III. Participants carrying four risk alleles had an increased susceptibility to OSCC as compared to participants carrying none, one, or two risk alleles (AOR=2.02, 95%CI=1.02-4.00, p=0.044). We also found that participants carrying more risk alleles in the combined genotypes of SDF-1 and CXCR4 SNPs had a higher risk of OSCC (p trend=0.033). However, the combined genotypes for SDF-1 and CXCR4 SNPs were not associated with the risk of PSCC and OPSCC.
Interactions between three types of substance use and CXCR4 C>T. We evaluated whether there was any interaction between the CXCR4 +414 C>T and the three types of substance use, namely BQ chewing, cigarette smoking, and alcohol drinking (Figure 1). Compared to the assumed lowest risk category, namely the carriers of C/T+T/T genotypes and without any substance use, the risk gradually increased for those carrying the C/C genotype and having a habit of one (AOR=1.79, p=0.174), two (AOR=7.68, p<0.001), or three (AOR=20.49, p<0.001) types of substance use. However, for those carrying C/T+T/T genotypes, the risk was increased only in participants with two (AOR=3.81, p=0.005) or three (AOR=17.68, p<0.001) types of substance use. Furthermore, compared to those carrying C/T+T/T genotypes, patients carrying the C/C genotype were at an increased risk only if they engaged in one (AOR=2.68, p=0.036) or two (AOR=2.02, p=0.027) types of substance use. Therefore, CXCR4 +414C>T did not have any effect on OPSCC risk either in participants without any substance use or in those with the three types of substance use.
Association of genotypes of SDF-1 G>A (rs1801157) and CXCR4 C>T (rs2228014) with the clinicopathological outcomes of OPSCC. We further examined the association of SDF-1 G>A (A/A+A/G vs. G/G) and CXCR4 C>T (T/T+C/T vs. C/C) with the clinicopathological characteristics of the OPSCC patients. Our results revealed that the genotypes of SDF-1 and CXCR4 did not significantly affect the clinicopathological features of the patients (Tables IV and V). Furthermore, a multivariate analysis showed that the G/G type of SDF-1 was not significantly correlated with the disease-specific survival (DSF) time of OSCC [A/A+A/G vs. G/G, adjusted hazard ratio (AHR)=0.79, 95%CI=0.51-1.23, p=0.298), PSCC (AHR=1.33, 95%CI=0.60-2.96, p=0.484], and OPSCC (AHR=0.97, 95%CI=0.66-1.42, p=0.858). In addition, no significant difference was also found between DSF and the genotype of CXCR4 C>T in OSCC (T/T+C/T vs. C/C, AHR=0.89, 95%CI=0.53-1.47, p=0.642), PSCC (AHR=1.05, 95%CI=0.42-2.66, p=0.914), and OPSCC (AHR=0.91, 95%CI=0.58-1.41, p=0.660), even in the combined genotypes of SDF-1 and CXCR4 variants (Table VI). No significant correlation was observed between SDF-1 and CXCR4 variants and prognostic outcomes of OPSCC. Based on the present study, CXCR4 C>T can be used as a genetic marker of susceptibility to OPSCC, which exerts a synergistic effect on OPSCC patients with the one or two types of risk habits.
Discussion
In this study, we found that the C allele and C/C genotypes of CXCR4 at locus +414 were significantly associated with OPSCC risk, particularly OSCC. However, the combined genotypic types of SDF-1/CXCR4 were not correlated with OPSCC risk, except for OSCC. In addition, the risk of OPSCC increased gradually in participants with a habit of two or more types of substance use compared to those without a habit of any substance use. Regarding the combined effect of CXCR4 C>T and substance use, we found that the effect of CXCR4 C>T on OPSCC risk increased only when participants had the habit of one or two types of substance use. To the best of our knowledge, the combined effect of CXCR4 C>T and three types of substance use on the susceptibility to OPSCC has not been previously reported.
In the present study we did not observe an association between SDF-1 G>A and OPSCC risk, including OSCC and PSCC. A study conducted in central Taiwan by Teng et al., has showed that the heterozygous genotype of G/A at SDF-1 G>A is associated with an increased risk of oral cancer, but the A/A genotype at SDF-1 G>A did not have any effect on oral cancer risk (39). Analysis based on chi-square test revealed that the genotypic frequencies of the controls are significantly different between the study by Teng et al., and our study. In the study by Teng et al., the frequencies of the G/G, A/G, and A/A genotypes at SDF-1 G>A were 58.5%, 28.5%, and 13.0%, respectively, whereas in our study, they were 51.4%, 40.3%, and 8.3%, respectively (χ2=12.03, p=0.002). However, our genotypic frequencies of the control are consistent with those of Lee et al., in Taiwan (the genotypic frequencies were 52.1%, 41.5%, and 6.4%; X2=0.93, p=0.629) (42). SDF-1 G>A in the 3’-UTR, whose allele A is regarded a target of cis-acting factors, has been shown to have the ability to up-regulate SDF-1 expression in prostate cancer (22). By contrast, we observed no correlation between the serum expression levels of SDF-1 and its genotype, which was consistent with the finding of Viswanathan et al. (43). These contradictory results might be attributed to the different cancer types and sample sources. Viswanathan et al., and us have both examined circulating SDF-1 in the serum through ELISA, but Hirata et al., have analyzed the expression levels of SDF-1 in prostate cancer tissue immunohistochemically (22). These findings must be confirmed by additional experiments in the future.
A silent SNP of CXCR4 (CXCR4 +414C>T) has been revealed to be located at the locus of +41434. Teng et al., have reported that no significant association existed between CXCR4 C>T and oral cancer risk (39). However, the results from our study (in which the number of participants is >2-fold greater compared to the study of Teng et al.) show that the C allele and C/C genotypes of CXCR4 C>T are associated with an increased risk of OPSCC and OSCC. A comparison of genotypic frequencies at CXCR4 C>T among the controls revealed no differences between our results and those of Teng et al., and Chang et al., (39, 44) (the frequencies of the C/C, C/T, and T/T genotypes at CXCR4 +414C>T were 71.5%, 27.1%, and 1.4%, respectively, in Teng et al.; 70.5%, 27.1%, and 2.4%, respectively, in our study; and 72.0%, 26.4%, and 1.6% in Chang et al.). Regarding the genotypic frequencies at CXCR4 C>T among patients, no differences were observed between the results of Teng et al., and our results. The binding of SDF-1 to CXCR4 induces intracellular signalling through several distinct pathways, such as chemotaxis, inflammation, and proliferation (23, 45). CXCR4 C>T has been found to affect codon 97 with a substitution of negatively charged aspartic acid (D) with neutral asparagine (N). This mutant codon is located in the second transmembrane region, a part of the receptor, which is relatively close to the cell surface and possibly important for SDF-1 binding (30). Brelot et al., have demonstrated that substitutions of acidic residues in the membrane-spanning domain (D97N) affects SDF-1 binding and impairs or stops receptor activation (46). Therefore, we speculated that CXCR4 C>T might, at least in part, lead to a decrease in the trend of SDF-1/CXCR4 signalling. Many cancers arise from sites of chronic irritation and inflammation (47). As mentioned, long-term exposure to BQ, cigarette, or alcohol may cause injury to oral-cavity cells. The result attracts diverse leukocyte-expressing CXCR4 populations – for example, neutrophils, macrophages, and lymphocytes – all of which are capable of producing an assorted array of cytokines, cytotoxic mediators, including reactive oxygen species, serine, and cysteine proteases (i.e., MMPs), and soluble mediators of cell killing (i.e., TNF-alpha, ILs, and interferons) (47, 48). It is now becoming clear that the tumour microenvironment, which is largely orchestrated by inflammatory cells, is an indispensable participant of the neoplastic process, fostering proliferation and migration (47). Therefore, CXCR4 on leukocyte population cells plays a crucial role in a developing neoplasm.
Risk habits, such as tobacco use, alcohol consumption, and BQ chewing have been identified as major risk factors for head and neck cancer (4). As described in the previous paragraph, these risk factors exert profound effects on multiple components of the inflammatory system. Chemokines accumulate in the microenvironment of tissues affected by chronic inflammation. If persistent, these inflammatory factors have the capacity to induce cell proliferation and to promote prolonged cell survival through the activation of oncogenes and inactivation of tumour-suppressor genes (49). We speculated that the C/C genotype of CXCR4 C>T might increase tumour development through the modulation of inflammatory cell aggregation and the suspension of cell proliferation. Therefore, we further characterised the combined effect of BQ cigarette, alcohol, and CXCR4 C>T. Our data indicate that compared to the assumed lowest risk category, such as the carriers of C/T+T/T genotypes without the habit of any type of substance use, the risk was gradually increased for those carrying the C/C genotype and with the habit of two or three types of substance use, placing the last ones in the higher risk category. The results were in accordance with our expectations. Compared to the C/T+T/T genotypes, the C/C genotype showed a 2.68-fold and 2.02-fold risk of OPSCC among participants with the habit of one or two types of substance use, respectively. Accordingly, the role of CXCR4 C>T cannot be ignored, particularly in individuals with one or two types of substance use.
In conclusion, our study suggests that SDF-1 G>A and CXCR4 C>T polymorphism plays a role in the development of OSCC and OPSCC. However, the combination of SDF-1 G>A and CXCR4 C>T did not correlate with OPSCC risk, except for OSCC. Moreover, OPSCC risk was gradually increased in participants with a habit of using two or more of the aforementioned substances as compared to those without a habit of using any of these substances. With regards to the combined effect of CXCR4 C>T and substance use, the effect of CXCR4 C>T on OPSCC risk increased only when participants had the habit of using one or two of the aforementioned substances.
Acknowledgements
This work was supported by grants from Kaohsiung Veterans General Hospital (VGHKS 107-137) and Zuoying Branch of Kaohsiung Armed Forces General Hospital (ZBH-106-14).
Footnotes
* These Authors contributed equally to this work.
Authors' Contributions
SJH, YKT and YHL conducted this study and drafted the manuscript. SJH performed the genotype analysis. PCW, JHL, CCL, CMY, CCW, LMY, YCL and MHL assisted in clinical samples' collection and manuscript drafting. LPG and HHL performed the statistical analysis. LPG and KWT supervised the study and edited the manuscript.
Conflicts of Interest
The Authors declare that they have no competing interests.
- Received April 10, 2019.
- Revision received May 16, 2019.
- Accepted May 17, 2019.
- Copyright© 2019, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved