Correlation of Tumor Necrosis Factor-α −308G>A Polymorphism with Susceptibility,Clinical Manifestations,and Severity in Behçet Syndrome: Evidences from an Italian Genetic Case–Control Study

Introduction

Behçet syndrome (BS) is a chronic vasculitis characterized by a wide spectrum of clinical manifestations that shares some clinical features with well-recognized autoinflammatory diseases (AIDs) (Konè-Paut et al., 2007; Yüksel et al., 2013; Lin et al., 2016; Burillo-Sanz et al., 2017). Some evidences of the key role of inflammasome-related genes in AIDs and BS were underlined by genetic studies (Gul, 2015; Takeuchi et al., 2015; Deng et al., 2018). In this context, the tumor necrosis factor (TNF)α is located on chromosome 6p21.3 and encodes a 233-amino acid type II transmembrane protein involved in the pathogenesis of BS. TNFα is a multifunctional pro-inflammatory cytokine belonging to the TNF superfamily, implicated in a variety of diseases, ranging from autoimmune diseases, cardiovascular disorders, insulin resistance, and cancer. Anti-TNFα therapies are used for the treatment of various inflammatory and autoimmune rheumatic diseases with positive but variable outcomes for the patients. The causes of inadequate response or treatment failure remain unknown and can be related to several factors, such as an alternative non-TNFα related pathway of inflammation and/or antidrug antibody presence or development (Marotte et al., 2008; O'Rielly et al., 2009; Pavy et al., 2010; Zeng et al., 2013; Bek et al., 2017). In addition, the alterations of TNFα expression can be associated with polymorphic alleles of TNFα gene having a pathogenetic role. TNFα −308A gene variation (rs1800629; NG_007462.1:g.4682G>A; HGVS nomenclature) has been associated with high production of TNFα and poor response to anti-TNF therapy, but the literature evidences showed inconsistent and/or conflicting data in various rheumatic diseases. In fact, rs1800629 was previously associated with increased susceptibility to and severity of rheumatoid arthritis, psoriatic arthritis, ankylosing spondylitis, inflammatory bowel disease, vitiligo, multiple sclerosis, cardiomyopathy, and malignancies, as well as of preeclampsia and celiac disease. TNFα −308A is a single nucleotide polymorphism (SNP) located within the gene promoter consisting in the guanine (G) to adenine (A) substitution.

It was reported that patients with the AA genotype and inflammatory diseases who are treated with anti-TNFα therapies may be less likely to have improvement in symptoms compared to patients with the GG genotype (Marotte et al., 2008; O'Rielly et al., 2009; Pavy et al.; 2010; Badano et al., 2012; Liang et al., 2013; Zeng et al., 2013; Zhang et al., 2013, 2018; Al-Okaily et al., 2015; Wang et al., 2015, 2018; Bek et al., 2017; Aflatoonian et al., 2019; Yang et al., 2019). Conflicting data are now available about the association of the polymorphism and both the susceptibility and the clinical patterns of BS (Duymaz-Tozkir et al., 2003; Lee et al., 2003; Ateş et al., 2006, 2010; Park et al., 2006; Kamoun et al., 2007; Storz et al., 2008; Bonyadi et al., 2009; Touma et al., 2010; Radouane et al., 2012; Al-Okaily et al., 2015; Abdolmohammadi and Bonyadi, 2017).

To date, these aspects were not analyzed in Italian BS population. The aim of this study was to investigate the frequency of rs1800629 and the possible association between the SNP and both clinical features and disease severity in a cohort of Italian BS patients compared with healthy controls (HC).

Materials and Methods

Results

A total of 230 Italian subjects were included in the study. Of which, 130 were consecutive patients with BS and 100 were matched HC. Table 1 reported the demographic and the clinical features of BS patients. The mean age of the patients at disease onset was equal to 45.8 ± 12.3 years. Male/Female ratio was 64:66, while for HC sex ratio was 48M:52F and mean age was 44.1 ± 12.0 years, without statistically significant differences between patients and HC (p > 0.05). BS patients' predominant lesions were oral ulcers (100%), eye lesions (86.2%), skin lesions (72.3%), and genital ulcers (57.7%), followed by joint involvement (50.0%). HLA-B51 positivity was found in 81/130 patients (62.3%).

Genotyping of TNFα rs1800629 showed a higher percentage of wild-type GG genotype in controls group (106/130 patients, 81.5%) than in HC (91/100 subjects, 91%) (p < 0.05). The heterozygous genotype (GA) was identified in 24/130 patients (18.5% of cases) and in 9/100 HC (9%) (p < 0.05). This genotype was significantly associated with the disease (OR = 2.29, 95% CI 1.01–5.18). No AA genotype was found in patients group and in controls group (Table 2). Detailed demographic, genetic, and clinical features for each of BS patients were reported in Supplementary Table S1.

We compared the frequency of rs1800629 genotypes with the distribution recognized in other populations. The data now available demonstrated that the genotype frequency of rs1800629 varied substantially among different ethnic groups of BS patients (Table 3).

A subset analysis was carried out for investigating the difference in genotype frequencies in the clinical subsets of BS. No statistically significant differences were found when both GG and GA frequencies were analyzed comparing the presence and the absence of each clinical manifestation (p > 0.05) (Table 4). We also compared the SNP genotype frequencies according to Krause’ severity index: no statistically significant differences were found when the score was calculated for the patient group with GG genotype (Krause's score = 5.095 ± 2.528) and for the patient group with GA genotype (Krause's score = 5.081 ± 2.430) (p > 0.05).

Discussion

This is the first report of assessment of TNFα −308G/A allele promoter SNP distribution and its association with BS clinical manifestations in patients from South Italy compared with sex-age and ethnically-matched HC. Our sample size is higher than other populations analyzed in European studies.

We found a statistically significant higher frequency of GA genotype in patients group than in HC. No association between rs1800629 genotypes and any clinical hallmark was recognized, as well as no relationship was found considering the disease severity according to Krause's index.

We focused our attention on TNFα due to the role of its polymorphisms as genetic contributors to BS immunopathology. Genetic factors that predispose to BS are considered to have a significant role in the disease development, so various loci were described as susceptibility markers, beside the HLA-B51, known as the major BS risk allele (Khaib et al., 2013; Kirino et al., 2013; Gul, 2014, 2015; Ombrello et al., 2015; Sousa et al., 2015; Takeuchi et al., 2015; Burillo-Sanz et al., 2017; Deng et al., 2018; Leccese et al., 2019; Padula et al., 2019). TNFα is encoded in the HLA complex on chromosome 6, a region that has been known to be associated with BS. In particular, TNFα rs1800629 is localized on the gene transcriptional start site and is related to the specific regulation of TNFα synthesis. In fact, the TNFα regulatory regions may play a role in cytokine production, and the binding of inducible cytoplasmatic factors may result in the translational blockade, leading to the dysregulated cytokine production. Genotype–phenotype correlation studies reported that G allele conferred twofold lower effects on the transcription level than A allele (Duymaz-Tozkir et al., 2003; Marotte et al., 2008; O'Rielly et al., 2009; Pavy et al., 2010; Badano et al., 2012; Liang et al., 2013; Zeng et al., 2013; Zhang et al., 2013, 2018; Al-Okaily et al., 2015; Wang et al., 2015, 2018; Bek et al., 2017; Yang et al., 2019; Aflatoonian et al., 2019). In vitro studies showed that the −308A allele is a more powerful transcriptional activator than the −308G allele (Ateş et al., 2006).

Previous results about the functional role of the SNP are few and contradictory and recently discussed by same group (Ateş et al., 2006). The authors did not observe a significant difference in serum TNFα levels between the subjects with −308 GG and GA genotypes in both patient and control groups. In our cohort, the association between GA genotype and disease risk could be explained taking into account that the presence of A allele could enhance the inflammatory reactivity influencing the binding of the transcriptional factors and contribute to the inflammation signature of BS.

Similar to our investigation, several studies assessed the influence of the promoter polymorphism on the expression of TNFα in BS or its susceptibility or its severity and clinical features, as well as its role as drug-response allele (Duymaz-Tozkir et al., 2003; Lee et al., 2003; Ateş et al., 2006, 2010; Park et al., 2006; Kamoun et al., 2007; Bonyadi et al., 2009; Storz et al., 2008; Touma et al., 2010; Radouane et al., 2012; Al-Okaily et al., 2015; Abdolmohammadi and Bonyadi, 2017). The effects of geographical differences, rather than the high genetic variation, are important factors for the adaptability and the evolution of any population.

rs1800629 genotype distribution was very different in the ethnic groups that we analyzed. GG wild-type frequency was between 3.3% of Saudi population (Al-Okaily et al., 2015) and 96.7% of Turkish population (Storz et al., 2008) in patients group and between 6.2% of Iranian Azer Turkish (Abdolmohammadi and Bonyadi, 2017) and 90.0% of Turkish population (Storz et al., 2008) in controls group. No statistically significant differences between patients and HC (p > 0.05) of the different ethnic groups were considered, except for Saudi (Al-Okaily et al., 2015) and Korean populations (Park et al., 2006) (p < 0.05), in addition to the findings of the present study.

Comparing the study of other populations with our Italian data, we observed that the low frequency of AA genotype is a very common finding.

The frequency of rs1800629 found in our cohort was very similar to the distribution reported in Turkish population by Duymaz-Tozkir et al. (2003) (79.8% for GG genotype, 18.2% for GA genotype) and Ateş et al. (2010) (80.4% for GG genotype, 19.6% for GA genotype). AA genotype was absent in both groups. The SNP frequency was previously studied by another research group in the same Turkish population with comparable results (74.5% for GG genotype, 25.5% for GA genotype, and 0.0% for AA genotype).

The frequencies of our group of controls were very similar to the distribution found in the Korean (91.5% for GG genotype, 8.5% for GA genotype, and 0.0% for AA genotype) (Lee et al., 2003) and Turkish (90.0% for GG genotype, 10.0% for GA genotype, and 0.0% for AA genotype) (Storz et al., 2008) populations.

About the correlation between the SNP and the clinical manifestations of BS, few studies are now available. In agreement with our data, Ateş et al. (2006) reported no association of the polymorphism and both BS clinical hallmark and severity. The polymorphism did not show any association with clinical findings also in Turkish population (Bonyadi et al., 2009; Abdolmohammadi and Bonyadi, 2017).

Lee et al. (2003) studied the SNP genotypes according to the presence of uveitis, recognizing that the presence of rs1800629 in patients with uveitis was not different from those without uveitis. The comparison between patients with severe and mild uveitis was also investigated in Duymaz-Tozkir et al.'s (2003) article; the authors reported that the haplotypic distribution (−308G/−376A) in the disease severity groups did not show significant differences, while the comparison between patients with severe and mild uveitis showed higher frequency of the same haplotype in severe cases than mild cases.

Conclusions

Here we studied the distribution of TNFα rs1800629 in a group of Italian BS patients compared with HC and the relation of the SNP with disease clinical hallmarks and severity. We found statistically significant higher frequency of GA genotype in patients' group than in controls. No significant association was recognized between the polymorphism and clinical parameters, as well as disease severity. About the clinical significance of our findings, we underline that the SNP is a promoter variation that could affect the expression of TNFα and the auto-inflammatory response with implication in BS etiopathogenesis.

The polymorphism could be also associated with anti-TNFα drug nonresponder pattern; in this context, larger molecular analysis results could provide a tool for a genetically-guided treatment. In addition, analyses of a larger cohort of patients and matched controls, as well as functional studies, are needed to confirm our preliminary data, to clarify the SNP role and to achieve future goals in the field of molecular personalized medicine.