Genetic Evidence Supporting Causal Associations Between Viral Infections and Sjogren's Syndrome

doi:10.21203/rs.3.rs-5006632/v1

Download PDF

Research Article

Genetic Evidence Supporting Causal Associations Between Viral Infections and Sjogren's Syndrome

https://doi.org/10.21203/rs.3.rs-5006632/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Sjogren's syndrome (SS) is a chronic inflammatory autoimmune disorder that mainly targets exocrine glands. Previous studies have suggested potential associations between Epstein-Barr virus (EBV), hepatitis virus (HAV), and other viruses with SS, but the causal nature of these relationships remains uncertain. This study used Mendelian randomisation (MR) to examine the genetic causal association between viral infections and SS.

Methods

Genetic data for SS was sourced from a genome-wide association study (GWAS) database of individuals of European ancestry (1290 patients and 213,415 healthy controls). Genetic data for nine viruses, including EBV, HAV, COVID-19, human immunodeficiency virus(HIV), cytomegalovirus, influenza virus, Coxsackie virus, measles virus, and retrovirus, were obtained from the IEU Open GWAS. Inverse variance weighting (IVW) served as the primary analysis method for MR Analysis, with Wald ratio, MR Egger, and weighted as supplementary analyses.

Results

MR analysis revealed causal associations between SS and five viral infections. Elevated VCA p18 antibodies against EBV, HAV, and COVID-19 were associated with increased SS risk, with respective odds ratios (OR) of 1.270 (95% CI: 1.043–1.550, p = 0.016), 1.163 (95% CI: 1.035–1.317, p = 0.009), and 1.109 (95% CI: 1.024–1.209, p = 0.013). Conversely, IgG antibodies against EBV and human immunodeficiency virus were associated with the reduction of SS risk, with ORs of 0.632 (95% CI: 0.430–0.921, p = 0.016) and 0.875 (95% CI: 0.787–0.972, p = 0.016) respectively. Sensitivity analysis did not reveal significant heterogeneity or horizontal pleiotropy. No statistically significant associations were found between the other four viruses and SS risk (all p > 0.05).

Conclusion

Our findings suggest that genetically predicted elevated levels of VCA p18 antibodies against EBV, HAV, and COVID-19 increase the risk of SS, while IgG antibody levels against EBV and HIV may confer protection. This study provides additional evidence for a link between viral infection and SS, aiding clinicians in identifying potential causative factors and thereby enhancing diagnostic specificity and sensitivity.

Sjogren's syndrome

Mendelian randomization analysis

Viral infection

Causal Association

Risk factor

Sjogren's syndrome (SS) is a chronic inflammatory autoimmune disease marked by the growth of lymphocytes and gradual destruction of exocrine glands, with a particular impact on the salivary and lacrimal glands. The condition presents itself through symptoms such as xerostomia, xerophthalmia, lethargy, arthralgia, and dermatitis (1). Despite extensive research, the precise etiology and pathogenesis of SS remain elusive. Various factors, including infections, genetic predisposition, hormonal imbalances, environmental factors, and immune dysregulation, can trigger SS(2–4).

Several infectious agents, such as Epstein-Barr virus (EBV), hepatitis virus (HAV), COVID-19, human immunodeficiency virus (HIV), cytomegalovirus (CMV), influenza virus (FLU), coxsackie virus (CV), measles virus (MEV), and Retrovirus, have been implicated in the development of SS. Some of these agents, such as EBV, HAV, Human T-lymphotropic virus 1 (HTLV-1), and CMV, establish persistent infections in exocrine glands, leading to organ damage and the development of SS (2, 5). When infecting salivary gland tissues, viral antigens are captured by plasmacytoid dendritic cells, triggering the production of multiple cytokines and chemokines that activate salivary gland epithelial cells, thereby initiating SS. Clinical evidence also indicates the presence of HAV RNA in saliva and salivary gland tissue, with about 50% of HAV-infected patients displaying SS-like inflammatory symptoms. Similar manifestations have been observed in HIV-infected patients, and transgenic mice carrying the HTLV-1 gene have displayed salivary gland inflammation resembling SS (6, 7). Furthermore, studies have shown a higher prevalence of serum antibodies against these infectious agents in SS patients compared to the general population, supporting a potential association (8). However, the exact causal relationship between SS and these infectious agents remains uncertain and requires further investigation.

Observational studies are vulnerable to the effects of reverse causality and unaccounted confounding variables. Conducting randomized controlled trials in this context is expensive and challenging. Therefore, in this study, we performed Mendelian Randomisation (MR) analysis and used aggregated statistics from a Genome-Wide Association Study (GWAS) to investigate the causal connection between infectious agents and SS in two separate samples (9). This approach provides a more robust and accurate means of assessing causality in the complex interplay of factors underlying SS pathogenesis.

2.1 Study design

The flowchart for this Mendelian randomization investigation is shown in Fig. 1. MR study is a data analysis method mostly employed in recent years for epidemiological etiological inference. The GWAS summary results underwent quality control methods to determine the relevant single nucleotide polymorphisms (SNPs). In order to estimate a causal effect, each genetic variant used as an instrumental variable (IV) needs to meet the three requirements listed below: (i) The exposure was linked to the variable in the correlation hypothesis. By analysing the p-value and F-statistics, this assumption was confirmed in the data; (ii) the independence hypothesis, which states that genetic instruments are unaffected by potential confounders; and (iii) the exclusion restriction hypothesis, which states that a genetic instrument can only influence an outcome through exposure. Meanwhile, the Reporting of MR Studies (STROBE-MR) guidelines were used for quality assessment (Table S1).

2.2 Sources of data on exposure and outcomes

Selecting the right genetic variation as an effective IV is a key step in performing MR. We obtained publicly available summary statistics on exposure and outcomes from the complete GWAS summary dataset manually collated by the MRC Integrated Epidemiology Unit (IEU) at the University of Bristol. The genetic background of the data was limited to populations of European descent. All data for SS were collected from the GWAS database and included 1290 cases in the case group and 213,415 in the reference group. We selected nine viruses as exposure factors, including EBV, HAV, COVID-19, HIV, CMV, FLU, CV, MEV, and retroviruses. The background of the exposed genetic data is also limited to European populations, the details of which are shown in Table 1.

Table 1

Detailed information on nine exposures and sources of genetic data on EBV, hepatitis A virus, COVID-19, human immunodeficiency virus, cytomegalovirus, influenza virus, Coxsackie virus, measles virus, and retroviruses.
Exposure/outcome	Dataset	Sample size	Number of SNPs	Population	Consortium	Sex	Year
Anti-epstein-barr virus IgG level	ieu-b-4901	5010	7,002,835	European	NA	Males and Females	2021
Epstein-barr virus VCA p18 antibody level	ebi-a-GCST90006900	8518	9,170,145	European	NA	NA	2020
Hepatitis A virus cell receptor 2	prot-a-1311	3301	10,534,735	European	NA	Males and Females	2018
COVID-19	ebi-a-GCST011075	1,388,342	9,739,225	European	NA	NA	2020
human immunodeficiency virus	finn-b-AB1_HIV	218,435	16,380,466	European	NA	Males and Females	2021
cytomegalovirus	ieu-b-4900	5,010	7,002,835	European	NA	Males and Females	2021
influenza virus	ebi-a-GCST006350	777	5,278,042	European	NA	NA	2018
Coxsackie virus	prot-a-737	3,301	10,534,735	European	NA	Males and Females	2018
measles virus	ebi-a-GCST006351	885	5,278,042	European	NA	NA	2018
retroviruses	prot-a-986	3,301	10,534,735	European	NA	Males and Females	2018

2.3 Selection of IVs

In order to satisfy the three assumptions of the MR Analysis, we performed a series of quality control steps to select qualified IVs. First, SNPs significantly associated with exposure were selected to exclude the interference of relevant factors (P < 1×10 − 5). Second, to reduce the effects of linkage disequilibrium (LD), we set the aggregation threshold to r2 < 0.001 and the aggregation window size to 1000 kb to maintain independence between IVs exposed each time. Furthermore, the PhenoScannerV2 database (http://www.phenoscaner.medschl.cam.ac.uk/) was employed to eliminate the interference of confounding variables and extract potential confounders such as smoking and drinking.

2.4 MR Analysis

MR analysis using inverse variance weighting (IVW) provides the most reliable causal estimates. When exposed to only one of the available SNPs, Wald ratios were used for the study; The results of the IVW method were considered to be the most reliable when there were multiple valid SNPs, and this method provided accurate and reliable estimates under the condition that all SNPs were valid. Therefore, we adopted the IVW method as the main method of MR Analysis and set P < 0.05 as the criterion for having an effect. MR-Egger and weighted median were used as supplementary analysis methods.

Sensitivity analyses were conducted using Cochrane's Q test, MR-Egger intercept test, and leave-one-out analysis. Cochrane's Q was used to test the heterogeneity of the data, and if p > 0.05, there was no heterogeneity. Horizontal pleiotropy was tested by MR-Egger regression intercept, and p > 0.05 indicated no pleiotropy. Leave-one-left analysis was performed to determine whether causality was driven by any single SNP.

All analyses were conducted utilizing R software (version 4.3.3) and the "two-sample MR" package (version 0.5.11).

We observed causal relationships between five types of viral infection and SS, while no causal relationship was found between the remaining four types of viral infection and SS. The detailed results are shown in Fig. 2.

3.1 Genetic IVs extraction of viral infection

To assess the causal relationship between viral infection and SS, we removed confusion and palindromic SNPs from the SS GWAS dataset and identified SNPs independently associated with air pollution: 31 related SNPs to EBV VCA p18, 50 related SNPs to COVID-19, 13 related SNPs to HAV, 4 related SNPs to Anti-EBV IgG, and 4 related SNPs to HIV, 13 related SNPs to CMV, 8 related SNPs to FLU, 20 related SNPs to CV, 17 related SNPs to MEV, and 12 related SNPs to retrovirus. In addition, all IVs were strongly associated with viral infection (F > 10), suggesting that the causal effect estimates of the study were not interfered with by weak instrument bias.

3.2 Mendelian randomization analysis with positive correlation

Through IVW preliminary analysis, genetically predicted viral infection exhibited a significant association with SS risk: EBV VCA p18: OR = 1.270, 95%CI = 1.043–1.550, p = 0.016; COVID-19: R = 1.109, 95%CI = 1.024–1.209, p = 0.013; HAV: OR = 1.16, 95%CI = 1.03–1.31, p = 0.009; Anti-EBV IgG: OR = 0.632, 95%CI = 0.430–0.921, p = 0.016; HIV: OR = 0.875, 95%CI = 0.787–0.972; p = 0.016 (Fig. 2). In MR-Egger and WME, these relationships also exist and are statistically significant, suggesting the robustness of our results (Fig. 3). However, we found that the causal relationship between HIV and SS was not significant in these methods and fortunately, this did not affect our primary results.

3.3 Mendelian randomization analysis with negative correlation

No evidence of causal relationships between CMV, FLU, CV, MEV, Retrovirus, and the risk of SS was found in this study: CMV: OR = 1.117, p = 0.557; FLU: OR = 0.903, p = 0.781; CV: OR = 1.012, p = 0.831; MEV: OR = 1.002, p = 0.811; retrovirus: OR = 0.973, p = 0.767. Given that the previously specified P-value was above 0.05, the findings did not demonstrate statistical significance.

3.4 Sensitivity Analyses

Our results were found to be robust based on sensitivity studies, which included the Egger intercept, Cochran's Q test, and leave-one-out analysis. No evidence of pleiotropy was found, and no heterogeneity of SNPs was detected. The funnel diagram analysis supported the robustness of our findings(Fig. 4). The results of the leave-one-out method suggested no substantial difference in causal estimations of the studied viruses on SS, indicating no single SNP drove the analysis results(Fig. 5). The results of sensitivity analysis prove the robustness of our results.

Infections, particularly viral infections, especially, have been suggested as potential factors contributing to SS. This study is the first to use MR to analyse the causal relationship between viral infections and SS. Our MR analysis identified several infectious agents (EBV VCA p18 antibodies, COVID-19, HAV) associated with an increased risk of SS, while others (anti-EBV IgG, HIV) showed an opposite effect. However, CMV, FLU, CV, MEV, and Retrovirus were not found to be causally associated with SS risk. Our comprehension of the part these viruses play in SS is strengthened by these findings.

Observational epidemiological studies have reported that EBV is one of the most widely accepted potential pathogenic factors(2). Interestingly, our MR analysis showed differing outcomes for different EBV antibodies, indicating a need for further investigation.

There have are some relevant studies have shown that type I interferon (IFN) signaling molecules associated with the activation of pattern recognition receptors (PRR) in innate immune responses play a significant part in the imbalance of SS immune responses, EBV can enhance the expression of IFN by releasing autoantigens (ribonucleoprotein complexes Ro/SSA and La/SSB) to activate the innate immune response, and then affect the onset of SS (Ro/SSA and La/SSB are ANA)(10, 11). Observational study results show that the prevalence rate of anti-nuclear antibody (ANA) in SS is about 80% and ANA-positive patients have a higher concentration of anti-EBV-VCA(12, 13). The aforementioned elucidation relates the occurrence of SS to immunologically relevant content, which may explain the causal relationship between EBV (VCA p18 antibodies) and SS. In addition, the presence of elevated ANA levels and anti-EBV-VCA antibodies in SS patients suggests immunological relevance, possibly explaining the causal relationship between EBV VCA p18 antibodies and SS(12, 14). However, in contrast to VCA p18, the MR Analysis's findings revealed that IgG antibodies played a protective role in the development of SS. This contradicts the previous conclusion that SS patients have a higher proportion of anti-EBV IgG antibodies, This is worth further investigation(13).

COVID-19, a recent global research focus, was found to increase SS prevalence. There has been no prior research suggesting a cause-and-effect connection between COVID-19 and SS. One study concluded that the hub gene MX1 was highly expressed in both the COVID-19 group and the SS group through bioinformatics methods, and concluded that the association between the two may be related to a specific hub gene (15). Our work presents a novel approach to investigating the correlation between COVID-19 and SS. However, further experiments are still needed to prove this association, which will be the main direction of our future research.

The relationship between HAVCR2 and SS remains largely unexplored. A previous study showed that the HAVCR2 gene can encode T-cell immunoglobulin mucin 3 (TIM-3) and mutations in this gene can lead to sustained immune activation and the production of inflammatory cytokines and the mutation of HAVCR2 leads to the high expression of TIM-3, which affects the number and function of T cells and may lead to the reactivation of EBV(16, 17). This suggests that HAV may promote SS through synergistic EBV effect, explaining SS occurrence.

HIV infection, known for its immunological changes, has shown SS-like symptoms but lacks evidence of Ro (SSA) or La (SSB) antibodies(18, 19). Our MR results contradict previous findings, suggesting a reduced risk of SS associated with HIV. This is the first time that a lower risk of SS has been linked to HIV. In a 1997 study, Yamano et al. did not detect HIV target genes in any salivary gland tissue or SS patients' peripheral blood mononuclear cells by PCR. Another study looking at the relationship between HIV and autoimmune diseases also ruled out the presence of HIV by PCR(20, 21). However, the mechanism of HIV and SS still needs to be further explored.

While many studies implicate CMV, FLU, CV, MEV, and Retrovirus in SS, our MR analysis found no causal relationship, indicating the need for comprehensive consideration of various viral factors.

This is the only study that we are aware of that uses MR Methods to examine the causal link between several viral infectious agents and SS. First, the MR Method we used was able to greatly reduce the influence of confounding factors and reverse causation on the outcomes, providing evidence for observational studies. Secondly, while using the traditional IVW method, we also adopted the simple median weighted median and the MR-Egger method, and the one-method analysis was used as the sensitivity analysis to ensure the robustness of our results. Finally, this study was able to make our results more statistically effective by conducting extensive data analysis through aggregated statistics collected from the GWAS database.

However, there are still potential limitations to this MR Study. First, the study's findings cannot be instantly extrapolated to other ethnic groups with distinct lifestyles and cultural backgrounds because the data used in the study originated from individuals of European descent. Second, our results do not yield specific mechanisms of influence, which need to be explored in further clinical trials. Finally, MR Studies require larger sample sizes to ensure adequate power, and sample sizes can affect results.

Our study revealed a causal relationship between SS and five viral infections. Elevated levels of EBV, hepatitis virus, and COVID-19 VCA p18 antibodies may increase SS risk, while anti-EBV IgG and HIV antibodies may confer protection. Our findings shed new light on the relationship between viral infections and SS, suggesting the need for SS risk consideration following viral infections.

Acknowledgments

We thank the IEU open GWAS project for providing summary statistics data of SS and viral infections for our analyses.

Funding statement

This study was supported by grants from the Natural Science Foundation of Shanxi Province (No. 202203021221269) and the National Natural Science Foundation of China (No. 82001740).

Author contributions

X.-F. L. and C.-H. W. designed the study. H.-Y. Z. and Y.-W. Z. collected and analyzed data. T. C. and C.-H. W. contributed to the analysis and interpretation of the data. S.-X. Z., C.-H. F., and Z.-N. J. wrote the manuscript. All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. X.-F. L. had full access to all of the data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis.

Conflict of interest disclosure

The authors declare no conflict of interest.

Ethics approval statement

The original GWAS data have obtained prior approval from relevant ethics review boards and the summary-level statistics don’t contain any personal information, thus no additional ethics approval was required.

Data availability statement

The datasets generated during and/or analyzed during the current study are available in the IEU open GWAS project (https://gwas.mrcieu.ac.uk/) and the FinnGen database (https://www.finngen.fi/).

Bjordal O, Norheim KB, Rodahl E, Jonsson R, Omdal R. Primary Sjogren's syndrome and the eye. Surv Ophthalmol. 2020;65(2):119-32.
Liu Z, Chu A. Sjogren's Syndrome and Viral Infections. Rheumatol Ther. 2021;8(3):1051-9.
Ramos-Casals M, Brito-Zeron P, Bombardieri S, Bootsma H, De Vita S, Dorner T, et al. EULAR recommendations for the management of Sjogren's syndrome with topical and systemic therapies. Ann Rheum Dis. 2020;79(1):3-18.
Vivino FB, Bunya VY, Massaro-Giordano G, Johr CR, Giattino SL, Schorpion A, et al. Sjogren's syndrome: An update on disease pathogenesis, clinical manifestations and treatment. Clin Immunol. 2019;203:81-121.
Utomo SW, Putri JF. Infections as Risk Factor of Sjogren's Syndrome. Open Access Rheumatol. 2020;12:257-66.
Vitali C. Immunopathologic differences of Sjogren's syndrome versus sicca syndrome in HCV and HIV infection. Arthritis Res Ther. 2011;13(4):233.
Nakamura H, Tsukamoto M, Nagasawa Y, Kitamura N, Shimizu T, Kawakami A, et al. Does HTLV-1 Infection Show Phenotypes Found in Sjogren's Syndrome? Viruses. 2022;14(1).
Bartoloni E, Alunno A, Gerli R. The dark side of Sjogren's syndrome: the possible pathogenic role of infections. Curr Opin Rheumatol. 2019;31(5):505-11.
Chang MJ, Liu MT, Chen MR, Li N, Zhao YH, Zhang SX, et al. Mendelian randomization analysis suggests no associations of herpes simplex virus infections with systemic lupus erythematosus. J Med Virol. 2023;95(3):e28649.
Negrini S, Emmi G, Greco M, Borro M, Sardanelli F, Murdaca G, et al. Sjogren's syndrome: a systemic autoimmune disease. Clin Exp Med. 2022;22(1):9-25.
Shimizu T, Nakamura H, Kawakami A. Role of the Innate Immunity Signaling Pathway in the Pathogenesis of Sjogren's Syndrome. Int J Mol Sci. 2021;22(6).
Cuomo L, Cirone M, Di Gregorio AO, Vitillo M, Cattivelli M, Magliocca V, et al. Elevated antinuclear antibodies and altered anti-Epstein-Barr virus immune responses. Virus Res. 2015;195:95-9.
Kivity S, Arango MT, Ehrenfeld M, Tehori O, Shoenfeld Y, Anaya JM, Agmon-Levin N. Infection and autoimmunity in Sjogren's syndrome: a clinical study and comprehensive review. J Autoimmun. 2014;51:17-22.
Scotto di Fazano C, Grilo RM, Vergne P, Coyral D, Inaoui R, Bonnet C, et al. Is the relationship between spondyloarthropathy and Sjogren's syndrome in women coincidental? A study of 13 cases. Joint Bone Spine. 2002;69(4):383-7.
Luo H, Zhou X. Bioinformatics analysis of potential common pathogenic mechanisms for COVID-19 infection and primary Sjogren's syndrome. Front Immunol. 2022;13:938837.
Nie XQ, Huang CF, Yin Z, Yang Y, Zhou X, Fang D, et al. [Two cases of EB virus-positive diffuse large B-cell lymphoma with HAVCR2 mutation]. Zhonghua Nei Ke Za Zhi. 2023;62(7):863-6.
Zhang Q, Zhou CJ, Li DH, Cui L, Li WJ, Ma HH, et al. Efficacy of ruxolitinib for HAVCR2 mutation-associated hemophagocytic lymphohistiocytosis and panniculitis manifestations in children. Br J Haematol. 2023;202(1):135-46.
Nakamura H, Shimizu T, Kawakami A. Role of Viral Infections in the Pathogenesis of Sjogren's Syndrome: Different Characteristics of Epstein-Barr Virus and HTLV-1. J Clin Med. 2020;9(5).
Talal N. AIDS and Sjogren's syndrome. Bull Rheum Dis. 1991;40(6):6-8.
Yamano S, Renard JN, Mizuno F, Narita Y, Uchida Y, Higashiyama H, et al. Retrovirus in salivary glands from patients with Sjogren's syndrome. J Clin Pathol. 1997;50(3):223-30.
Coll J, Palazon J, Yazbeck H, Gutierrez J, Aubo C, Benito P, et al. Antibodies to human immunodeficiency virus (HIV-1) in autoimmune diseases: primary Sjogren's syndrome, systemic lupus erythematosus, rheumatoid arthritis and autoimmune thyroid diseases. Clin Rheumatol. 1995;14(4):451-7.

No competing interests reported.

TableS1.docx
Table S1 The Reporting of MR Studies (STROBE-MR) guidelines.

Download PDF

Editorial decision: Revision requested
07 Oct, 2024
Reviews received at journal
26 Sep, 2024
Reviewers agreed at journal
25 Sep, 2024
Reviewers invited by journal
25 Sep, 2024
Editor assigned by journal
02 Sep, 2024
Submission checks completed at journal
02 Sep, 2024
First submitted to journal
30 Aug, 2024

You are reading this latest preprint version

Genetic Evidence Supporting Causal Associations Between Viral Infections and Sjogren's Syndrome

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

1 INTRODUCTION

2 MATERIALS AND METHODS

2.1 Study design

2.2 Sources of data on exposure and outcomes

2.3 Selection of IVs

2.4 MR Analysis

3 RESULT

3.1 Genetic IVs extraction of viral infection

3.2 Mendelian randomization analysis with positive correlation

3.3 Mendelian randomization analysis with negative correlation

3.4 Sensitivity Analyses

4 DISCUSSION

5 CONCLUSION

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1