Sex and FOXP3 gene rs2232365 polymorphism may be associated with the clinical and pathological aspects of chronic viral diseases

The forkhead box protein 3 (FOXP3) transcription factor is the main marker of regulatory T-cell (Treg) activation, a subpopulation specialized in the suppression of immune responses and the maintenance of homeostatic tolerance in different microenvironments [1]. The impact of Treg frequency on the progression of chronic diseases is important because these cells mediate the inflammatory balance in an immunologically polarized environment [2].

FOXP3 gene is located on the short arm of the human sex chromosome (X on XY model) (Xp11.23; 21 kb) and consists of 11 exons [3]. The relevance of single nucleotide polymorphisms (SNPs) located in the promoter region of the FOXP3 gene has been investigated due to they are involved in the transcription activation, as well as in the interaction with regulatory elements of the gene expression, possibly reflecting on the level of expression of FOXP3 and, thus, in the activation of Treg cells [4, 5].

In this context, the rs2232365 polymorphism in the FOXP3 gene is related to functional changes that predispose individuals to diseases due to the mutant allele (C) changing the binding site of transcription factors that interact with the region [6]. Although the knowledge about the association of this polymorphism with different diseases is based on the literature [7, 8], the relationship between genetic variants of FOXP3 and sex, remain unclear. Thus, the present study evaluated the association of the rs2232365 polymorphism with clinical and pathological aspects and markers of progression of viral and non-viral chronic diseases between the sexes.

At first, we researched the biological meaning of the rs2232365 polymorphism in public databases of gene expression (GText portal - https://​www.​gtexportal.​org/​home/​snp/​rs2232365). The TT genotype has the highest level of expression of the FOXP3 gene in different tissues investigated (Fig. 1). We collected expression data in whole blood and arterial tissue, the bank does not yet have liver tissue data.

In our results, the allele frequency of the rs2232365 polymorphism was similar between sexes, however statistical differences were observed between genders when assessing viral infections (Table 1). The polymorphism was in Hardy-Weinberg equilibrium (p > 0.05).

In the present study, no significant association between sex and chronic viral liver diseases was observed, although the prevalence of males was higher among those infected. The frequency of polymorphic variants was the same between the sexes when the viral hepatopathy itself was evaluated. (Table 1). In males the allele T was associated with advanced fibrosis, with a relative risk of 4 fold in the bivariate analysis (p:0.05; OR:4.18; CI 95%: 1.13–15.4). In multiple logistic regression the risk of the genetic factor is reduced, however, it still maintains the significance (0.05 ≥ p ≥ 0.01; OR:1.11; CI 9%:0.43–2.90) (Table 1). The viral load (VL) was high among males carrying allele T (p: 0.0007) as compared to carriers of the allele C (Fig. 2a), but the transaminase ratio (AST/ALT) was not significant between alleles and sex, although in males this rate is increased (Fig. 2b). The levels of gamma-glutamyl transferase (GGT) in males with allele C were higher than in those with allele T, but without significant differences (Fig. 2c). These associations were not observed in women. Hepatic inflammation was not associated with the gender or genetic profile of those infected (Table 1).

Our data indicate that women are more associated with HTLV-1 than the rs2232365 polymorphism itself (p:0.001; OR: 1.20; CI 95%:0.63–2.29), although the allele T has been associated as a protective factor against infection, especially in males (p:0.04; OR:0.46; CI 95%:0.18–1.18) (Table 1). Again in men, variant T was also associated with low levels of proviral load (p: 0.0396), T CD4⁺ lymphocytes (p: 0.0483) and IL-8 cytokine (p: 0.0549); the elevation of the levels of the cytokine IL-10 was a trend observed, however, with low statistical power (p: 0.1102) (Fig. 2d-g). The T CD8⁺ lymphocytes count was similar between genotypes/sexes (data not shown). Both gender and polymorphism were not associated with an asymptomatic carrier or with more advanced infection profiles (HAM / TSP) (Table 1).

The risk of coronary heart disease was associated only with males, although the significance was lost when analyzing the history of risk infection by Chlamydia (Table 1). The frequency of polymorphism was not associated with either the pathological profile or the history of risky infection in both sexes (Table 1). Chlamydia pneumoniae was the most frequent species in the study, however, the infection was not associated with the sex and genetic profile of the patients (Table 1). Levels of C-reactive protein (CRP) were low and similar between sexes and alleles (Fig. 2h).

To avoid false positive claims, we infer the value of false positive report probability (FPRP) for all significant associations. All findings remained noteworthy in the prior probability between 0.25 to 0.1; except for CAD (Female vs Male), whose notability is evident from the 0.01 probability (Table 2).

There are controversies regarding the genetic changes caused by the rs2232365 polymorphism [6, 9]; however, based on the analysis of gene expression, it is suggested that the polymorphisms alter the expression of FOXP3, predisposing the individual to different clinical and pathological outcomes related to the disease in question. From an immunological point of view, there is a tendency for the T allele to favor an anti-inflammatory profile, while the C allele favors a pro-inflammatory profile.

It seems counterintuitive to associate an anti-inflammatory factor with the risk of tissue aggression; however, these findings corroborate previous studies that suggest that variant T favors the persistence of viral infection in patients with fibrosis [10]. With the increase in VL, there is a long-term pro-fibrogenic tendency induced by chronic inflammation and continuous response to healing [11], resulting in the fluctuation of pre-cirrhotic biomarkers [12], such as GGT. The transaminase ratio was stable between genotypes and sexes, indicative of chronic infection, but not relevant for estimating the stage of fibrosis [13]. Recent studies by our group show that in liver fibrosis, viral load, in fact, is more associated with the histological profile than liver integrity enzymes [14].

In the multiple logistic regression, the allele T was also associated with advanced fibrosis; however, the decrease in the statistical probability is indicative of other factors influencing liver histopathology. In recent publications, we discard alcoholism as a behavioral factor associated with liver fibrosis [14].

In HTLV-1 infection, the high prevalence of infected women is an epidemiological fact observed in different populations studied [15, 16], and is related to the effectiveness of sexual transmission from male to female. With the present study, we suggest that not only transmissivity, but also sex-linked immunogenetic factors can influence susceptibility to HTLV-1 infection. The observed associations suggest that intrinsic mechanisms regulate the action of FOXP3 variants between sexes. Studies show that, in women, epigenetic processes modulate the expression of FOXP3 and alter the susceptibility to diseases [17]; the normal development of cells carrying mutant genes is a consequence of mixed chimerism [18], which is suggested in the present study due to the marked frequency of heterozygous women (data not shown). In men, changes in FOXP3 tend to be more relevant due to heredity [19], regulation induced by the Y chromosome [20] and sex hormones [21].

Although there are no studies on the association of the polymorphism with HTLV-1 infection, it is suggested that, in men, the anti-inflammatory tendency induced by the allele T reduces chronic immune hyperactivity, typically seen in the pathogenesis of the infection, reflecting the decrease in the proviral load and the cytokines of cellular immunity [22]. The relationship of the polymorphism with the T CD4⁺ lymphocytes and IL-8 is indicative of the constitution of an atypical pro-inflammatory immune network [23], however, already observed in HTLV-1 infection, mainly in patients with Adult T-cell leukemia (ATL) [24].

In the present study, we did not associate the polymorphism with the clinical and pathological aspects of HTLV-1 infection. However, given the prevalence of asymptomatic patients observed, a cohort study in patients with allele T in this clinical group can clarify whether, in the long term, there will be polarization for a pro-inflammatory profile and changes in the pathogenesis of infection.

Only males were associated with the risk of CAD, however, unrelated to the history of Chlamydia infection. The history of Chlamydia penumonie prevailed and CRP levels fluctuated, but both were not associated with sex and polymorphism. Recent reports confirm that men are generally more likely to develop CAD than women, with the highest death rate in middle age. Women, on the other hand, are at greater risk of stroke, which usually occurs at older ages [25].

We defend the legitimacy of our data based on the FPRP values obtained that support the proposed findings. All values remained noteworthy between 0.25 to 0.001 of prior probability, which, according to Wacholder and collaborators, for polymorphisms whose functional data are suggestive of a possible association (as shown in the literature for rs2232365) is the ideal probability range [26].

A total of 323 females and 373 males diagnosed with chronic viral liver disease (CVLD), human T-cell leukemia virus type 1 (HTLV-1) infection or coronary artery disease (CAD) and healthy groups of blood donors, all from regional reference centers in the state of Pará, Brazil (Tropical Medicine Nucleus of the Federal University of Pará, Foundation of Hemotherapy and Hematology Center of Pará (HEMOPA) and Institute of Health Sciences of the Federal University of Pará), were evaluated. Participants were informed about the objectives of the study and, after agreeing to participate, signed a consent form. The study was approved by the ethics committees of the participating entities (CAAE: 73782017.8.0000.0018; CAAE: 0011.0.324.000–09 and CAAE: 31223214.2.0000.0018). All patients were not undergoing treatment.

The inclusion/exclusion criteria and the biological data for patients CVLD were established in our previous study [10, 14].

Patients HTLV-1⁺ were classified as asymptomatic or with paraparesis (HTLV-1-associated myelopathy/tropical spastic paraparesis - HAM/TSP) according to pre-established criteria [27]. The quantification of the proviral load was based on previously described protocols [27, 28]. Quantification of lymphocytes TCD4⁺ and CD8⁺ was performed by immunophenotyping in flow cytometry (FACSCountTM Reagents - TriTEST™/TruCount, BD Biosciences, San Jose, CA, EUA). Pro-inflammatory (TNF-α, TNF-β, IFN-γ, IL6 and IL8) and anti-inflammatory (IL-10) cytokine levels were determined by immunoenzymatic assays (Human ELISAReady-SET-Go, EBioscience, Inc., San Diego, CA, USA).

Genotyping of the polymorphism (rs2232365) was performed by real-time PCR (C_15942641_10 - Applied Biosystems, Foster City, CA, USA) with established temperature and cycling protocols [10]. To explore the biological potential of the polymorphism, we evaluated the gene expression of genotypes based on the public database of the GTex portal (https://​www.​gtexportal.​org/​home/​snp/​rs2232365).

The Kruskal-Wallis test was used to calculate the association of the polymorphism with the quantitative biomarkers of the studied pathologies. The Mann-Whitney test was used as a confirmatory method, this time, comparing two independent samples.

To calculate the probability of false positives for significant associations, we predefine an FPRP limit value of 0.5. The Odds ratio was 1.5 for the calculation of statistical power. Associations that remained noteworthy between the ranges of 0.25 to 0.01 of prior probability were considered true positive, according to Wacholder’s recommendations [26].

Statistical analysis was performed using BioEstat 5.0 [29] and GraphPad Prism 6.1, with a significance level of 95% (p ≤ 0.05). FPRP was calculated using R 3.4.2 software [30].