Abstract
Background:
Host genetic factors play a major role with respect to susceptibility to infections. Many polymorphisms of the Toll-like receptors (TLRs), members of the innate immune response, are directly associated with the clinical outcomes following infection. The 2848 G/A variant (rs352140) of the TLR9 gene is associated with increased TLR9 expression. However, the impact of the genotypes of this SNP on HIV+, HCV+, and HCV+/HIV+ individuals is still debated.
Materials and Methods:
This study investigated the 2848 G/A polymorphism in HCV infection, HIV infection, and HCV/HIV co-infection in a large sample of Brazilians (n = 1,182). Groups were initially compared without considering stratification by ethnicity and subsequently stratifying individuals between whites and non-whites.
Results:
Considering non-white individuals, a significant difference between the HIV+/HCV+ group and controls was observed with the GG genotype serving as a protective factor (p = 0.023). Additionally, significant allelic differences were observed between the HCV+ group and controls (p = 0.042); between the HIV+/HCV+ group and controls (p = 0.011); and between the HIV+/HCV+ group and HIV+ individuals (p = 0.047). However, all significant results were lost following adjustment for multiple comparisons (p > 0.05).
Conclusion:
Although our initial results indicated a potential influence of the rs352140 genotype on host altered susceptibility to viral infections, following correction for multiple comparisions the standard (p < 0.05) for statistical association was lost. This may be due to insufficient sample size as we were examining many different associations. Thus, a larger study is warranted to further pursue this topic.
Introduction
Sexually transmitted infections (STIs) affect a significant portion of the world's population, with high incidence rates in general. Among STIs, human immunodeficiency virus (HIV) and hepatitis C virus (HCV) have a significant impact on the health of the human host due to the potential of these viruses to impair immune responses and other body systems. Of note, HIV and HCV are commonly detected together in a single individual, characterizing HCV/HIV coinfection.
In 2020, about 37 million people were living with HIV, with 1.5 million cases in that same year alone (UNAIDS, 2022). In Brazil, from 2007 to June 2021, 381,793 cases of HIV infection were recorded (Ministério da Saúde, 2021a). In 2021, 58 million people were living with chronic HCV infection in the world, with 1.5 million new cases detected per year (WHO, 2021). In Brazil, from 1999 to 2020, 262,815 cases of HCV infection were reported. In addition, in Brazil, 19,924 cases of HCV/HIV coinfection were observed in the period from 2007 to 2020 (Ministério da Saúde, 2021b).
Worldwide, about 6% of individuals living with HIV have already tested positive for HCV/HIV coinfection (Avert, 2019). Concomitant HIV and HCV infection can facilitate progression to AIDS and worsen chronic liver diseases such as cirrhosis and hepatocellular carcinoma. Moreover, there are some suggestions of a connection between the presence of HIV infection and extrahepatic manifestations in HCV+ patients (Piroth et al., 1998; Radkowski et al., 2002; Hernandez and Sherman, 2011).
Toll-like receptor 9 (TLR9) is an essential molecule of the innate immune system, being responsible for recognition of unmethylated CpG DNA and triggering inflammatory responses (Hemmi et al., 2000). TLR9 is expressed by hepatocytes, dendritic cells, B lymphocytes, Kupffer cells, and hepatic stellate cells, among other cell types (Hemmi et al., 2000; Akira et al., 2001; Meli et al., 2014).
The TLR9 gene polymorphisms could affect different aspects of viral infections (Lai et al., 2013; Said et al., 2014; Sánchez-Luquez et al., 2021; Mozer-Lisewska et al., 2010). This receptor can participate in the response to HIV infection by recognizing the DNA formed from the activity of the reverse transcriptase enzyme on viral RNA (Mogensen et al., 2010). Furthermore, interaction of the viral protein gp120 with dendritic cells inhibits innate responses mediated by TLR9, suggesting a role for this receptor in the response against HIV (Martinelli et al., 2007).
In HCV infection, the role of TLR9 is still poorly understood. It is known that only DNA binds to this receptor; considering that the genetic material of HCV is RNA, the recognition and activation of a TLR9-mediated response are intriguing. However, antiviral effects have already been observed from activation of TLR9 in HCV+ individuals (Broering et al., 2008). Furthermore, elevated levels of the receptor were identified in these patients compared with a control group (Hu et al., 2011). Thus, TLR9 indeed participates in the immune response to HCV infection, but the mechanisms of this participation are still unknown and should be explored (Broering et al., 2008).
The 2848 G/A variant (rs352140) is a synonymous single nucleotide polymorphism (SNP) located in exon 2 of the TLR9 gene, characterized by the presence of an A allele instead of the wild-type G allele. Although this SNP does not have a direct impact on protein structure, it has been related to another polymorphism in the promoter region, the -1237 C/T polymorphism (rs5743836). This variation allows increased binding of the NFκB transcription factor and increased TLR9 expression (Hamann et al., 2006; Ng et al., 2010; Roszak et al., 2012).
In addition, the presence of the A allele has been linked to accelerated progression to AIDS (Bochud et al., 2007). In another study, an association was found between endosomal TLR gene polymorphisms and susceptibility to HCV/HIV coinfection (Valverde-Villegas et al., 2017).
Understanding how genetic polymorphisms influence the clinical outcome and susceptibility to infections may indicate paths for development of better therapies (Ellwanger et al., 2018). The TLR9 2848 G/A polymorphism has been explored in HCV+, HIV+, and HCV+/HIV+ individuals, but its role is controversial in these contexts.
Thus, this study investigated the frequency of the TLR9 2848 G/A polymorphism in HCV+, HIV+, and HCV+/HIV+ Brazilian individuals and a control group to assess the potential impact of this polymorphism on susceptibility to or protection from these infections.
Materials and Methods
Samples and ethical aspects
All participants of this study voluntarily donated blood samples for DNA extraction. Samples were collected in Porto Alegre city (State of Rio Grande do Sul, South Brazil) at referral centers for treatment of infectious diseases. Table 1 shows the characteristics of the four groups evaluated in this study, totaling 1182 Brazilian individuals.
Demographic Data of the Individuals Included in This Study
Statistically significant values are shown in bold.
n, sample number.
p-Value observed comparing the HCV+ group and control group; Pearson's χ2 with Yates's correction.
p-Value comparing the HIV+ group and control group; Pearson's χ2 with Yates's correction.
p-Value comparing the HCV+/HIV+ group and control group; Pearson's χ2 with Yates's correction.
Based on color of skin and self-declaration.
HCV, hepatitis C virus; HIV, human immunodeficiency virus.
The control group comprised healthy blood donors (negative for both HIV and HCV). HCV+ and HIV+ individuals were recruited when they presented a positive serological or molecular test for HCV and/or HIV at the referral centers for treatment of infectious diseases involved in this research.
This study was approved by the Ethics Committees of Universidade Federal do Rio Grande do Sul (UFRGS, Brazil), Universidade Luterana do Brasil (ULBRA, Brazil), and Hospital de Clínicas de Porto Alegre (HCPA, Brazil). All participants signed an informed consent following Resolution No. 466 from Ministério da Saúde (2012).
Genotyping
DNA samples were amplified following a PCR-restriction fragment length polymorphism protocol. The PCR was performed as described by Cheng et al. (2007), generating a DNA fragment of 177 bp. The cleavage reaction was executed using Bsh123I (BstUI) restriction endonuclease, generating the following fragment patterns: 177 bp (homozygous variant AA genotype); 177 bp +135 bp (heterozygous GA genotype); and 135 bp (homozygous wild-type GG genotype).
Amplicons were visualized on a 3% agarose gel under UV light. The fragment of 42 bp derived from G allele cleavage was not identifiable under these conditions. Negative controls were included in all reactions as well as fragment length controls (DNA ladder) and controls with predefined genotypes for interpretation of results after cleavage with the restriction enzyme. In addition, our results rely on a check of 10% of the samples, selected at random.
Statistical analyses
We verified Hardy-Weinberg equilibrium using the chi-square (χ2) test. Allele and genotype frequencies were compared between groups through Pearson's χ2 test (2 × 3 tables) and Pearson's χ2 test with Yates's correction (2 × 2 tables). Residual values were used to identify differences between specific genotypes.
Once groups were not homogeneous regarding the ethnic background (p < 0.05, Table 1), we stratified individuals by ethnicity (white or nonwhite based on skin color and self-declaration) and repeated the analyses.
A p-value <0.05 was set as statistically significant. When appropriate, p-values were adjusted for multiple comparisons by Benjamini-Hochberg's step-up false discovery rate. Analyses were performed using WINPEPI, v.11.65 (Abramson, 2011).
Results
Table 1 shows demographic data of the individuals included in this study. The total sex ratio (male/female) of the study was 1.15/1. There was no statistically significant difference in sex ratio between groups (all p > 0.05). When considering ethnicity, however, there was a difference between controls and HCV+ individuals, between controls and HIV+ individuals, and between controls and HCV+/HIV+ individuals, the latter with a significantly larger number of nonwhite individuals (Table 1).
Genotype and allele frequencies of TLR9 2848 G/A polymorphism are shown in Table 2. All groups were in Hardy-Weinberg equilibrium (p > 0.05 in all analyses). Table 3 shows allele and genotype frequencies of individuals stratified by ethnicity (white and nonwhite individuals). Table 4 shows results of comparisons of genotype and allele frequencies between groups without considering ethnicity. No statistically significant difference was observed (all p > 0.05).
Distribution of TLR9 2848 G/A Polymorphism Genotypes and Alleles
Allele frequency = (2 × n individuals with GG or AA)+(n individuals with GA)/(2 × n total number of individuals).
AA, variant homozygous genotype; freq., frequency; GA, heterozygous genotype; GG, homozygous wild-type genotype; n, sample number.
Distribution of TLR9 2848 G/A Polymorphism Genotypes and Alleles (Groups Stratified by Ethnicity)
Allele frequency = (2 × n individuals with GG or AA)+(n individuals with GA)/(2 × n total number of individuals).
Comparisons of Genotype and Allele Frequencies (Detailed in Table 2) Between Groups
Pearson's χ2 for genotypes and Pearson's χ2 with Yates's correction for alleles.
Benjamini-Hochberg step-up false discovery rate (grouping p-values from genotypes or alleles, according to ethnicity).
CI, confidence interval; OR, odds ratio.
Following these analyses, comparisons between groups were performed with stratification by ethnicity (Table 5). At first, we observed some statistically significant differences between groups of nonwhite individuals. Comparing the HIV+/HCV+ group with controls, we found a significant difference considering genotypes (Pearson's χ2 = 7.581, p = 0.023), specifically regarding the GG genotype (residual value = 2.6/−2.6, p = 0.009), and also considering alleles (Pearson's χ2 = 6.507, p = 0.011, odds ratio [OR] = 1.85, confidence interval [95% CI = 1.18-2.93]).
Comparisons of Genotype and Allele Frequencies (Detailed in Table 3) Between Groups (Groups Stratified by Ethnicity)
Statistically significant p-values are shown in bold.
Pearson's χ2 for genotypes and Pearson's χ2 with Yates's correction for alleles.
Benjamini-Hochberg step-up false discovery rate (grouping p-values from genotypes or alleles, according to ethnicity).
Adjusted residual for the comparison considering the GG genotype: 2.6/−2.6, p = 0.009.
In addition, statistically significant differences were observed when comparing allele frequencies between the HCV+ group and controls (Pearson's χ2 = 4.144, p = 0.042, OR = 1.52 [95% CI = 1.03-2.22]) and between the HCV+/HIV+ group and HIV+ group (Pearson's χ2 = 3.933, p = 0.047, OR = 1.61 [95% CI = 1.03-2.52]). However, all statistically significant results are lost if correction of p-values by multiple comparisons is considered (Table 5). No statistically significant difference between other groups was observed (p > 0.05 in all other comparisons, Table 5).
Discussion
This study evaluated the distribution of TLR9 2848 G/A in a noninfected control group and HIV+, HCV+, and HIV+/HCV+ individuals, comparing allele and genotype frequencies between groups. A total of 1182 Brazilian individuals were genotyped. This large sample size gives robustness to the results of our study. Among nonwhite individuals, the GG genotype was more frequent in controls than in the HCV/HIV-coinfected group, with an initial statistically significant p-value. This initial finding could point to the GG genotype as a potential protective factor against HCV/HIV coinfection.
In agreement, the G allele was more frequent in controls and HIV-monoinfected individuals than in HCV/HIV-coinfected and HCV-monoinfected individuals. However, all significant differences were lost when adjustment for multiple comparisons was applied. Thus, we stress that there was no statistical association between the mentioned genotype and allele and HCV and/or HIV infection.
Adjustment of p-values by multiple comparisons is a topic of debate. This adjustment would not be necessary for exploratory, nonconfirmatory, or nonclinical studies, such as our study (Rothman, 1990; Bender and Lange, 2001). However, we decided to show results before and after this statistical adjustment since this is a conservative way of presenting results and because it enhances the reliability of the findings (Jafari and Ansari-Pour, 2019).
Considering these aspects, the results observed before statistical adjustments indicate that the TLR9 2848 G/A could indeed have a biological effect on altered susceptibility to viral infections, but this result was disregarded in our conservative statistical analysis. Furthermore, investigating the effect of this polymorphism on susceptibility to/protection from viral infections in other populations is needed.
Our results raised a potential protective role of the G allele against HCV monoinfection and HCV/HIV coinfection. Conversely, the A allele would act as a risk factor for these infections. In agreement with our results, the AA genotype had already been associated with susceptibility to HCV/HIV coinfection in African descendants (Valverde-Villegas et al., 2017), further supporting the stratification by ethnicity applied to our analysis.
We found a higher frequency of the AA genotype in the HCV+/HIV+ group compared with controls among nonwhite individuals (Table 3), although the difference did not reach statistically significant values. When we stratified individuals according to ethnicity, the coinfected group encompassed only a few individuals, a factor that can potentially explain the lack of statistically significant association observed in this study regarding the AA genotype.
We did not have access to some risk factors that influence susceptibility to HCV or HIV, such as behavioral factors or drug use. This may also have influenced our results, representing an additional limitation of this study.
The TLR9 molecule and TLR9-related polymorphisms have varied impacts on HCV and HIV infections. The viral escape of HCV in patients with chronic hepatitis seems to be related to the lack of IFN production from activation of TLR9 in plasmacytoid dendritic cells, once there is a considerable increase in apoptosis of these cells from the interaction of the HCV core protein with TLR2 in monocytes (Dolganiuc et al., 2006).
In addition, TLR9 appears to be an important receptor in development of autoimmunity in patients with chronic HCV infection. This role is related to overexpression of TNF alpha from activation of TLR9, which promotes activation of effector Th1 cells and reduction in the activity of regulatory T cells (Comarmond et al., 2019). In addition, reduced spontaneous clearance of HCV infection was associated with the TLR9 -1237 C/T polymorphism, which is in low linkage disequilibrium with 2848 G/A in African descendants and modest linkage disequilibrium in European descendants (Fischer et al., 2017; Valverde-Villegas et al., 2017).
It has been suggested that the C allele of -1237 C/T is associated with increased IL-6 production and B cell proliferation, which may be correlated with enhanced liver fibrosis (Carvalho et al., 2011; Fischer et al., 2017). Considering that the chronicity of infection may be related to the decrease in TLR9 activation, the increased expression, associated with the 2848 G/A variant, could have a role in the clinical outcome of HCV infection.
Unfortunately, a limitation of the study is the unavailability of clinical data from patients regarding chronicity, spontaneous clearance, and other outcomes. Thus, although TLR9 appears to be a key receptor in the immune response against HCV, the influence of TLR9 2848 G/A in HCV-related diseases remains a topic to be investigated.
The impact of TLR9 2848 G/A on HIV infection is an open question. In a previous study, the presence of the G allele of this polymorphism was associated with rapid progression of HIV infection (Bochud et al., 2007). Some studies addressed the viral load and CD4+ cell counts in the context of HIV infection in association with evaluation of the TLR9 2848 G/A polymorphism, but results were conflicting (Soriano-Sarabia et al., 2008; Said et al., 2014; Valverde-Villegas et al., 2017).
In this study, no influence of the TLR9 2848 G/A polymorphism on susceptibility to HIV infection in individuals from South Brazil was evidenced, but studies evaluating the role of this polymorphism on the progression of HIV infection are still needed. We highlight that as a limitation of the study, it was not possible to evaluate patients exposed to HCV with differential outcomes. Even so, we emphasize that such an evaluation could bring robustness to the results, and we bring this analysis as a future perspective for construction of the scenario presented here.
Conclusions
This study evaluated a potential association between the TLR9 2848 G/A polymorphism and susceptibility to or protection from HIV infection, HCV infection, and HIV/HCV coinfection in a large sample of Brazilians. Our initial results indicate a potential influence of the TLR9 2848 G/A polymorphism on susceptibility to viral infections, but no statistical effect of the polymorphism on protection from/susceptibility to infections was observed when adjustment by multiple comparisons was applied.
The potential role of the G allele as a protective factor against HCV monoinfection and HCV/HIV coinfection deserves more investigation. Considering the large sample size evaluated, the three infection profiles included in the analysis, and the results before adjustments by multiple comparisons, this study contributes to the understanding of host susceptibility to viral infections.
Footnotes
Authors' Contributions
B.K.-L. was involved in conceptualization, methodology, validation, formal analysis, writing—original draft, writing—review and editing, and visualization; J.H.E. was involved in conceptualization, methodology, formal analysis, writing—review and editing, and project administration; J.M.V.-V. was involved in conceptualization, methodology, investigation, and writing—review and editing; D.S., C.G.M., V.S.M., R.K.L., R.K., and E.S. were involved in investigation and resources; and J.A.B.C. was involved in resources, writing—review and editing, supervision, and funding acquisition.
Author Disclosure Statement
No competing financial interests exist.
Funding Information
B.K.-L. received a scholarship from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Brazil). J.H.E. received a postdoctoral fellowship from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (Programa Nacional de Pós-Doutorado—PNPD/CAPES, Brazil). J.M.V.-V. received a doctoral scholarship from CAPES (Brazil). J.A.B.C. received a research fellowship from CNPq (Brazil) and has research funded by CAPES (Brazil).
