Article Text

Download PDFPDF

Original article
Sex-specific effects of TLR9 promoter variants on spontaneous clearance of HCV infection
  1. Janett Fischer1,
  2. Alexander N R Weber2,
  3. Stephan Böhm1,
  4. Sabine Dickhöfer2,
  5. Souhayla El Maadidi1,2,
  6. Danilo Deichsel1,
  7. Viola Knop3,
  8. Hartwig Klinker4,
  9. Bernd Möller5,
  10. Jens Rasenack6,
  11. Lisa Wang7,
  12. Manu Sharma7,
  13. Holger Hinrichsen8,
  14. Ulrich Spengler9,
  15. Peter Buggisch10,
  16. Christoph Sarrazin3,
  17. Michael Pawlita11,
  18. Tim Waterboer11,
  19. Manfred Wiese1,
  20. Elsbeth Probst-Müller12,
  21. Raffaele Malinverni13,
  22. Pierre-Yves Bochud14,
  23. Clair Gardiner15,
  24. Cliona O'Farrelly15,
  25. Thomas Berg1
  1. 1 Department of Gastroenterology and Rheumatology, Section of Hepatology, University Hospital, Leipzig, Germany
  2. 2 Department of Immunology, Interfaculty Institute for Cell Biology, University of Tübingen, Tübingen, Germany
  3. 3 Medical Department 1, Goethe-University Hospital Frankfurt/Main, Frankfurt, Germany
  4. 4 Department of Internal Medicine II, University of Würzburg, Würzburg, Germany
  5. 5 Department of Medical Practice, Charlottenstraße 81, Berlin, Germany
  6. 6 Medical Department, Albert-Ludwigs University Freiburg, Freiburg, Germany
  7. 7 Division of Genetic Epidemiology, Institute for Clinical Epidemiology and Applied Biometry, University of Tübingen, Tübingen, Germany
  8. 8 Department of Gastroenterology, Gastroenterologische Schwerpunkt-Praxis, Kiel, Germany
  9. 9 Department of Internal Medicine I, University of Bonn, Bonn, Germany
  10. 10 Liver Unit, IFI Institute for Interdisciplinary Medicine, Asklepios Klinik St. Georg Hamburg, Hamburg, Germany
  11. 11 Department of Genome Modifications and Carcinogenesis (F020), Research Program Infection and Cancer, German Cancer Research Center (DKFZ), Heidelberg, Germany
  12. 12 University Hospital, Zürich, Switzerland
  13. 13 Pourtalès Hospital, Neuchâtel, Switzerland
  14. 14 Infectious Diseases Service, University Hospital and University of Lausanne, Lausanne, Switzerland
  15. 15 School of Biochemistry and Immunology, Trinity College Dublin, Dublin, Ireland
  1. Correspondence to Dr Janett Fischer, Department of Gastroenterology and Rheumatology, Section of Hepatology, University Hospital, Leipzig, Germany, Liebigstraße 21, Leipzig 04103, Germany; janett.fischer{at}

Statistics from

Significance of this study

What is already known on this subject?

  • Toll-like receptors (TLR) as a part of the innate immunity play an essential role in the control of HCV infection.

  • Genetic variants of the TLRs can affect the immune response and disease progression.

  • The impact of TLR9 in a setting of natural HCV infection is unknown.

What are the new findings?

  • In the study, we identified a gender-related association of the polymorphism rs187084 within the TLR9 gene with the natural course of HCV infection.

  • The combined determination of interferon lambda 4 (IFNL4) rs12979860 and TLR9 rs187084 improved the identification of women who have higher likelihood of spontaneous viral clearance (SC).

  • Functional analyses revealed that the single nucleotide polymorphisms affect the regulation of TLR9 expression.

  • TLR9 might play a greater role in HCV infection than previously expected.

How might it impact on clinical practice in the foreseeable future?

  • The inclusion of the genetic pattern defined by TLR9 rs187084 and IFLN4 rs12979860 into stratification algorithm will increase the identification of women with higher likelihood of SC.


The innate immune system forms the first defence line during the early phase of HCV infection.1 The viral genome and viral proteins are detected by three main classes of pattern recognition receptors (PRRs): the Toll-like receptors (TLR), the retinoic acid-inducible gene-I (RIG-I)-like receptors (RLRs) and the nucleotide oligomerisation domain-like receptors (NLRs). The PRRs trigger downstream signalling cascades to activate innate and adaptive immune responses. Downstream molecules such as the adaptor myeloid differentiation primary response 88 (MyD88) or kinases of the interleukin (IL)-1 receptor-associated kinase (IRAK) family induce the production of type I and III interferons (IFNs), pro-inflammatory cytokines and chemokines.2 Furthermore, the TLR-mediated signalling pathway represents an important link between innate and adapted immunity.3

HCV proteins can be sensed by TLR2 and TLR4 located on the cell surface and HCV RNA is detected by the intracellular TLR3 and TLR7.4 However, based on evidence from cell culture systems, it has been proposed that TLR9, a DNA-sensing TLR expressed on human plasmacytoid dendritic cells (pDCs), B cells, hepatocytes, Kupffer cells and hepatic stellate cells,5–8 takes part in HCV antiviral effects, although viral RNA is no typical substrate of this receptor.9 ,10 For example, in peripheral blood mononuclear cells of HCV-infected patients, the TLR9 mRNA and protein concentration were impaired and inversely correlated with the level of viral copies in the serum.11 Other studies showed inhibitory effects of HCV particles on pDC-associated IFN alpha production in response to TLR9 activation.12 Furthermore, antiviral effects were determined in HCV-infected patients upon stimulation of TLR9.13 However, the importance of TLR9 in a setting of natural HCV infection has not been demonstrated.

The innate immune response inhibits HCV replication only to a certain level, resulting in a lifelong chronic infection in approximately 70–80% of HCV-infected patients. It has been reported that HCV uses several mechanisms to evade host innate immunity contributing to viral persistence and chronicity.14 Interestingly, genetic variations within the host's innate immunity can affect the immune response leading to spontaneous or treatment-induced clearance of HCV infection. For example, several independent genome-wide association studies (GWAS) identified strong associations of single nucleotide polymorphisms (SNPs) near the interferon lambda 3 (IFNL3) (commonly known as IL28B) and in the interferon lambda A4 (IFNLA4) gene loci with spontaneous viral clearance (SC) and treatment outcome.15–18 Especially the SNP rs12979860 is suggested to be a marker for the functionally relevant dinucleotide variant interferon lambda 4 (IFNL4)-ΔG (rs368234815). Variants of melanoma differentiation-associated protein 5 (Mda5), the human leucocyte antigen C (HLA-C) and of TLRs were also found to be associated with increased risk of viral persistence and development of a severe liver disease.19–21

Due to the enigmatic role of TLR9, in the present study, we undertook a multicentre, retrospective association study of candidate polymorphisms within the TLR9 gene on the natural course of HCV infection to assess if this would epidemiologically confirm a role for TLR9 in HCV infection. The TLR9 gene is located on chromosome 3p21.3 and spans approximately 5 kb.22 The TLR9 polymorphisms rs187084 (−1486T/C) and rs5743836 (−1237T/C) are located within the putative promoter region that may influence transcriptional regulation of the TLR9 gene. Therefore, we subsequently performed functional analysis of the TLR9 SNPs to assess their impact on TLR9 transcriptional activity.

Patients and methods

Ethics statement

All patients and volunteers provided written informed consent.


The evaluation cohort (EC) comprised 494 patients of European descent with HCV type 1 infection from Germany; 161 (33%) patients spontaneously cleared the virus. Parts of the cohort were included in the response-guided individualised tailored treatment regimen of the INDIV-1 study.23 Further information on patients' recruitment is given in the online supplementary material. Parts of the Swiss HCV cohort (The Swiss Hepatitis C Cohort Study (SCCS); n=1138), which were used for GWAS15 and already genotyped for TLR9 rs187084, served as replication cohort. Eighty-one (7%) patients of the SCCS were coinfected with HIV and were excluded from further analysis. The characteristics are shown in table 1.

Supplementary data

Table 1

Patients' characteristics of the evaluation and replication cohorts

Furthermore, the results were validated in two independent cohorts only composed of women of the well-documented iatrogenic single-source HCV outbreaks in recipients of HCV-contaminated anti-D IgG during pregnancy in Germany (AD1: n=305)24 and Ireland (AD2: n=198)25 in 1977–1979. The samples of the patients from anti-D cohorts were collected by the treating hepatologists in the original referral centres from 2011 to 2012.

SC was defined as undetectable HCV RNA in serum within the first 12 months of follow-up after the estimated date of infection in the absence of treatment and presence of anti-HCV antibodies. Chronic HCV infection was diagnosed by a positive anti-HCV test in routine diagnostic (Architect Anti-HCV, ABBOTT, Wiesbaden, Germany) and by the presence of HCV RNA in serum (COBAS AmpliPrep/COBAS TaqMan HCV Roche Diagnostics, Mannheim, Germany) for more than 6 months.

Healthy volunteers

Healthy female volunteers were recruited at the Department of Immunology, Tübingen, following written ethical approval of the study and after obtaining written informed consent. All volunteers reported to be healthy at the time of blood sampling.


All reagents were purchased from Sigma, unless otherwise stated. Estradiol (E2758) and estrone (3-hydroxy-estra-1,3,5(10)-trien-17-one, E9750) were dissolved in dimethyl sulfoxide (DMSO) to 1 mg/mL and prediluted in phosphate-buffered saline before stimulation.



The DNA samples were analysed for the TLR9 SNPs rs187084 (−1486T/C), rs5743836 (−1237T/C) and IFNL4 rs12979860. Genotyping involved real-time PCR and melting curve analysis in a Light Cycler 480 System (Roche) using fluorescence resonance energy transfer (FRET) probes (TIB MOLBIOL, Berlin, Germany) (see online supplementary table S1). Genomic DNA from healthy volunteers was conducted using allele-specific TaqMan-based assays (Life Technologies, Darmstadt, Germany) for rs187084 and rs5743836, as described.26

Plasmids, cloning and site-directed mutagenesis

Mutations corresponding to the TLR9 promoter SNPs were introduced into a TLR9 promoter reporter construct in the pGL3 vector22 (Promega, Mannheim, Germany) using a QuickChange XL Kit (Agilent, California, USA) and verified by DNA sequencing. Plasmids were always prepped from the same bacterial strain to rule out differences in methylation.

Reporter gene assay

The assay was performed in Nawalma R20 B cells as described in the online supplementary material. B cells are a professional TLR9+human cell type and represent a homogeneous cell system to explore regulation of TLR9 expression and signalling.27 They express between twofold and fourfold more TLR9 mRNA than other cell lines.28 TLR9 signals lead to activation of the B cells initiating the production of pro-inflammatory cytokines.29 Furthermore, B cells are suggested to be important for viral transmission and as a reservoir for persistent infection.30

TLR9 mRNA qPCR analysis

After stimulation with estrogen receptor alpha (ERα) modulators (see figure legend for concentrations) for 3 h, mRNA was isolated from 1.5 mL whole blood samples (QIAamp RNA Blood Mini Kit; Qiagen), genomic DNA was digested and transcribed to cDNA (High Capacity RNA-to-cDNA Kit; Life Technologies). The expression of TLR9 was studied using a TaqMan Gene Expression Assay (Life Technologies #Hs00152973_m1). Data were normalised to the TATA-binding protein housekeeper. The samples were analysed on a real-time cycler (Applied Biosystems; 7500 fast) in triplicates. All donors were analysed in parallel.

Statistical analysis

Statistical analysis for epidemiological associations was performed with SPSS 20.0 (SPSS, Chicago, Illinois, USA). The significance of associations between dichotomous data was assessed by Pearson's χ2 test and Fisher's exact test. Univariate and multivariate logistic regression analyses were performed to determine factors associated with SC and to estimate interactions between the SNPs. We used the following variables: TLR9 rs187084, rs5743836, IFNL4 rs12979860, sex and age at study time. They were categorised as follows: rs187084 CT/CC versus TT (reference), rs5743836CT/CC versus TT (reference), rs12979860 CT/TT versus CC (reference), female versus male (reference) and age as a continuous variable. The model was controlled for cohort impact. All tests were two-sided and p values less than 0.05 were considered statistically significant. The OR and the 95% CI were calculated. Structure of linkage disequilibrium (LD) was analysed with Haploview 4.2 (Broad Institute, Cambridge, USA) by using expectation-maximization (EM) algorithm.

For functional experiments, p values were determined using an unpaired Student's t test due to Gaussian distribution (as assessed by Kolmogorov–Smirnov test). p Values were calculated in GraphPad Prism 6.0 and considered significant at p<0.05 (denoted by asterisk *).


SNP selection

In previous experiments, DNA samples from patients of a model hepatitis C cohort (n=50), which included patients chronically infected with HCV and those who cleared the virus, were sequenced by using a 48-capillary 3730 DNA analyzer (Applied Biosystems). Three polymorphisms in the TLR9 gene (rs187084, rs5743839 and rs352140) and one in the IRF7 gene (rs1061502) appeared to have an effect on the course of HCV infection. Moreover, studies identified association with susceptibility to other diseases,31–34 location in the regulatory region22 and a minor allele frequency (MAF) >0.15 in Caucasians (dbSNP35 and HAPMAP Databases36). Those four SNPs and the IFNL4 SNP rs12979860 were examined in the EC. After Bonferroni correction for five SNPs (p<0.01), the TLR9 rs187084 and the IFNL4 SNP rs12979860 were still significantly associated with SC in women.

Association of TLR9 rs187084 and IFNL4 rs12979860 with SC

The SC rates were higher in women than in men in the EC (n=494, EC: 37% vs 28%, p=0.045) and in the Swiss HCV cohort (n=1057, SCCS: 12% vs 7%, p=0.011). There was no deviation from Hardy–Weinberg equilibrium for TLR9 rs187084 (EC: p=0.43, SCCS: p=0.84), rs5743836 (EC: p=0.85) and IFNL4 rs12979860 (EC: p=0.64; SCCS: p=0.10). The overall genotype distribution of TLR9 rs187084, rs5743736 and IFNL4 rs12979860 in the EC, in the SCCS and the replication cohorts (AD1, n=305 and AD2, n=198) was similar. Genotype analysis in the EC and SCCS was performed with regard to SC, which generally had been observed in 161 patients (33%) of the EC and 94 patients (9%) of the SCCS. Clearance rates are given in table 2.

Table 2

Genotype distribution of TLR9 rs187084 and rs5743836 and IFNL4 rs12979860 and rates of spontaneous viral clearance (SC) (%)

Significant differences in virus clearance rates between men and women were observed for TLR9 rs187084 in the EC and in the SCCS. The C allele was more frequent in women able to clear the virus (EC: 75% SC vs 58% CH, p=0.006; SCCS: 78% SC vs 62% CH, p=0.039) than in men (EC: 70% SC vs 68% CH, p=0.879; SCCS: 73% SC vs 63% CH, p=0.333).

In univariate analyses, the C allele of TLR9 rs187084 showed a high-level association with SC in a dominant manner for the evaluation cohort and Swiss HCV cohort (EC: CC/CT vs TT: OR=1.56 (95% CI 1.03 to 2.35) p=0.035; SCCS: CC/CT vs TT: OR=1.70 (95% CI 1.04 to 2.77) p=0.033). Stratifying for gender revealed that this association only occurs in women (EC: CC/CT vs TT: OR=2.21 (95% CI 1.25 to 3.90) p=0.006; SCCS: CC/CT vs TT: OR=2.13 (95% CI 1.05 to 4.30) p=0.035) but not in men (EC: CC/CT vs TT: OR=1.08 (95% CI 0.59 to 2.57) p=0.813; SCCS: OR=1.41 (95% CI 0.71 to 2.79) p=0.324). The results could be replicated in both anti-D cohorts. Meta-analysis of all cohorts revealed no cohort heterogeneity (I2=0%) and a pooled OR of 2.07 (95% CI 1.56 to 2.77) p<0.0001 (see online supplementary figure S1). Since the association was only seen in women, further analyses focused on effects in women (table 3). As expected,18 we found strong differences in genotype distribution of the SNP IFNL4 rs12979860. In all cohorts but the Swiss HCV cohort, the T allele of IFNL4 rs12979860 was significantly associated with reduced likelihood of SC.

Table 3

Univariate and multivariate analyses of factors associated with spontaneous viral clearance in women

To ensure that the results for TLR9 rs187084 were not due to a coincidental over-representation or under-representation of genotypes at the IFNL4 locus, a multivariate logistic regression model was performed, including the covariates; IFNL4 rs12979860; TLR9 rs187084 and rs5743836 and age. In women, the T allele of IFNL4 rs12979860 (OR=0.42 (95% CI 0.24 to 0.72) p=0.002) and the C allele of TLR9 rs187084 (OR=2.15 (95% CI 1.18 to 3.90) p=0.012) were independently associated with SC, which was confirmed in both anti-D cohorts (table 3). After combination of the EC and SCCS, both the C allele of TLR9 rs187084 (OR=1.99 (95% CI 1.30 to 3.05) p=0.002) and the T allele of IFNL4 rs12979860 (OR=0.54 (95% CI 0.37 to 0.79) p=0.002) were still independent predictors of SC in women but not in men (rs187084 C allele: OR=1.19 (95% CI 0.763 to 1.86) p=0.440; rs12979860T allele: 0.25 (95% CI 0.16 to 0.38) p=1.19×10−10).

Haplotype analysis of TLR9 rs187084 and rs5743836

Since the rs187084 and rs5743836 are in LD (D′=0.89, r2=0.09), four haplotypes exist: rs187084T/rs5743836T (46.0%), rs187084C/rs5743836T (38.8%), rs187084 T/rs5743836C (14.6%) and rs187084C/rs5743836C (0.4%). Overall, in women, the presence of at least C allele of rs187084 was associated with a higher likelihood of SC in the overall cohort (CT vs TT: OR=1.39 (95% CI 1.11 to 1.75) p=0.005). In contrast, the C allele of rs5743836 reduced the chance of SC (table 4).

Table 4

Main haplotypes of TLR9 rs187084/rs5743836 associated with spontaneous viral clearance (SC) in women of the overall cohort

Combination of TLR9 rs187084 and IFNL4 rs12979860

Overall combined analysis of TLR9 rs187084 and IFNL4 rs12979860 of the EC and of both anti-D cohorts was performed. In the overall cohort, the TLR9 rs187084 C allele (42% SC, OR=2.03 (95% CI 1.48 to 2.79) p=0.00001) and the rs12979860CC genotype (56% SC, OR=4.38 (95% CI 3.19 to 6.01) p=8.14×10−20) were independently associated with SC in women.

Genotype analysis revealed that in women with rs12979860CC genotype, the additional determination of rs187084 significantly affected clearance rates. Increased SC rates were observed after adjusting for rs12979860CC genotype compared with no adjustment (rs187084CC vs rs12979860CC/rs187084CC: 39% SC vs 58% SC, p=0.035; rs187084CT vs rs12979860CC/rs187084CT: 43% SC vs 62% SC, p=0.0002; rs187084TT vs rs12979860CC/187084TT: 26% SC vs 46% SC, p=6.34×10−11). Moreover, the presence of the C allele of rs187084 was associated with more SC (see online supplementary table S2). Multivariate logistic regression analysis including TLR9 rs5743836 revealed that the presence of the CC genotype of IFNL4 rs12979860 and at least one C allele of rs187084 increased the association with SC (rs187084CC/CT/rs12979860CC vs rs187084TT/rs12979860CT/TT: OR=4.42 (95% CI 2.93 to 6.66) p=1.37×10−12). In patients with the heterozygous variants of the rs12979860T non-responder allele, the pattern of the rs187084 affected the chances of achieving SC. There were significant differences in SC rates between carriers of rs12979860CT/rs187084CC/CT and carriers of rs12979860CT/rs187084TT (15% SC vs 29% SC, p=0.001).

In the Quick, Unbiased, Efficient Statistical Tree (QUEST) algorithm,37 ,38 the combination of TLR9 rs187084 and IFNL4 rs12979860 identified 62% of women with SC and 73% with chronic hepatitis C (figure 1).

Figure 1

Classification tree for spontaneous viral clearance (SC) of HCV infection in women of the overall cohort. TLR9 rs187084, rs5743836 and interferon lambda 4 (IFNL4) rs12979860 were included in analysis. Data were analysed with the Quick, Unbiased, Efficient Statistical Tree (QUEST) algorithm. The split points were found with quadratic discriminant analysis (QDA). In total, the model correctly identified 62% of women with SC and 73% with chronic hepatitis C. The false positive rate was 30%. To significantly maximise the likelihood of clearance, the cohort was split according to IFNL4 rs12979860 genotype followed by the TLR9 rs187084 genotype. In carriers of the rs12979860T allele, the additional determination of rs187084 improved the prediction of SC. TLR, Toll-like receptors.

Differences in TLR9 promoter activity

As the genetic association data indicated the base substitution within the TLR9 promoter might impact on gene expression, we performed functional assays to assess the impact of these polymorphisms in signalling. Using site-directed mutagenesis, we introduced the SNPs rs1807084 or rs5743836 into a reporter construct carrying the human TLR9 promoter.33 We transfected these TLR9 variants into Nawalma R20 B cells. Analysis of promoter activation revealed that the rs187084 (p=0.041) and the rs5743836 (p=0.012) variants exhibit significantly increased baseline gene expression compared with the wild type (WT) promoter (figure 2A). Thus, both TLR9 SNPs appear to regulate promoter activity which may underpin the observed effects on viral clearance.

Figure 2

Transcriptional activity of the TLR9 promoter variants and regulation by estrogen receptor alpha (ERα) ligands. (A) Basal transcriptional activity of pGL3_basic and TLR9 wild type (WT), rs187084 and rs5743836 encoding the Firefly luciferase; in Nawalma R20 B cells. Control plasmid pRL-TK (Renilla) was used for transfection efficiency normalisation of luciferase activity. The results are reported as the relative luciferase units (RLU). The figure shows the mean±SEM of results obtained from six experiments, each performed in triplicate. (B) The 432 bp sequence of the TLR9 promoter construct with mapped rs187084, rs5743836, nuclear factor ‘kappa-light-chain-enhancer’ of activated B cells (NF-κB)-binding and estrogen receptor alpha (ERα)-binding sites according to Hasan et al.41 (C) Transcriptional activity of TLR9 WT, rs187084 and rs5743836 in whole blood assay of healthy female volunteers after 3 h stimulation with 1 µmol and 10 µmol of estrone and β-estradiol. The TLR9 mRNA was measured using a TaqMan Gene Expression Assay. The statistical significance of differences between TLR9 WT and the indicated TLR9 mutant plasmids was assessed by using a Student's t test due to Gaussian distribution (* denotes p<0.05). TLR, Toll-like receptors.

In the process of exploring putative regulatory elements in the TLR9 promoter region, we noted that coregulation of TLR9 by nuclear factor ‘kappa-light-chain-enhancer’ of activated B cells (NF-κB) and ERα had been reported in chronic human papillomavirus (HPV) 16 infection38 and that rs5743836 maps in close proximity to this NF-κB/ERα regulatory motif (figure 2B). To explore whether this had any functional significance and given that some of the genetic associations were linked to the female sex, we stimulated whole blood from healthy female volunteers, homozygous for the T allele (WT), heterozygous or homozygous for the minor C allele with two ERα activators and naturally occurring estrogens, β-estradiol and estrone.39 Whereas homozygous WT carriers strongly downregulated TLR9 within 3 h in response to treatment with both β-oestradiol and oestrone, downregulation was significantly lower in heterozygotes and the single homozygote carrier (figure 2C). This suggests that rs5743836 may have a functional influence on TLR9 mRNA levels during the natural cycle of estrogen levels in women.


Several independent GWAS proved that the variation in genes coding for regulators of the innate immune system significantly affects the natural course of HCV infection as well as the therapy response of chronically infected patients and may promote the progression of liver disease. The most important finding was the strong impact of polymorphisms near the IFNL4 gene locus with SC and treatment-induced outcome.15–18

While these polymorphisms in the IFNL3/4 gene have been considered most relevant in studies investigating the association with hepatitis C, our data suggest that the TLR9 SNPs may be considered as additional, and possibly hitherto overlooked, host genetic factors affecting the course of infection. We investigated the effect of polymorphisms within the TLR9 gene on the natural course of HCV infection since TLR-mediated signalling pathways play a crucial role in HCV innate immunity. There is evidence that TLR9 contributes to HCV antiviral effects and TLR9 SNPs have been significantly associated with other diseases such as rheumatoid arthritis, systemic lupus erythematosus, gastric cancer and cervical cancer.31–34 The major finding of our study was that the TLR9 rs187084 polymorphism appears to be genetically associated with gender-specific SC of HCV infection.

In our study, we found that the minor C allele of the TLR9 rs187084 was associated with a 1.9-fold to 2.2-fold increase in likelihood of SC of HCV infection in women compared with carriers of the TT genotype suggesting rs187084 is a novel prognostic factor for SC. Although Wei et al 40 conducted a similar study in chronically infected Chinese patients with HCV, they did not observe an association which may be due to their small cohort size.

Combining of TLR9 rs187084 and IFNL4 rs12979860, the rs187084 pattern had a significant impact on women with SC. In homozygous carriers of the genotype rs12979860CC, the presence of the rs187084 C allele had a marginal effect on SC rate. However, in patients with a heterozygous variant for the non-responder rs12979860 allele, the additional genotyping of rs187084 significantly affected SC prediction. Here, the additional presence of the rs187084 C allele increased the chance of SC by about twofold.

Collectively, this means that the inclusion of the genetic pattern defined by TLR9 rs187084 and IFLN4 rs12979860 into stratification algorithm may increase the identification of patients who have higher likelihood of SC. Furthermore, the study allows new insights into gender-dependent epidemiological factors affecting the natural course of hepatitis C.

Functional analyses of the impact of the TLR9 polymorphisms on basal promoter activity revealed that both SNPs rs187084 and rs5743836 provoke a 25% higher gene expression compared with the WT promoter. Increased promoter activity of TLR9 rs573836 might be explained by the findings of Ng et al.41 They showed that the variant C allele at rs5743836 creates a potential NF-κB-binding site that increased the transcriptional activity of the gene. The presence of this extra putative NF-κB-binding site promotes TLR9 transcription in response to various stimuli more effectively than the WT TLR9 rs5743836T sequence. On the other hand, Hasan et al 38 reported that this NF-κB, in cooperation with the aforementioned ERα element, can be involved in TLR9 downregulation in keratinocytes. Furthermore, in silico investigations have indicated that rs187084 T>C substitution generates a new binding site of the transcription factor specificity protein 1 (Sp-1), which harbours a zinc finger motif and binds to GC-rich motifs of many promoters,42 a ubiquitous transcription factor which regulates constitutive transcription from many eukaryotic promoters but can also activate or repress stimulus-induced transcription.43 Others have suggested that the C allele of rs5743836 generates a regulatory site for a IL-6-responding element and observed an IL-6-mediated upregulation of TLR9 expression—presumably via STAT3 although this was not proven—in mononuclear cells with rs5743836CT genotype.44 This led to increased IL-6 production and B-cell proliferation. Interestingly, apart from its systemic proinflammatory properties, IL-6 was reported to promote accelerated hepatic inflammation and development of liver fibrosis.45 Another ERα regulatory element that may be affected by this variant was first reported by Hasan et al 38 and they observed that the addition of the ERα ligand inhibitor melatonin relieved HPV-induced downregulation and consequently increased TLR9 mRNA concentration in vitro due to interference with NF-κB/ERα-mediated suppression.

Indeed, we observed genotype-dependent differences in the strength of the TLR9 regulation in response to the ERα activators, β-estradiol and estrone. The SNP rs187084, which is located 250 bp upstream from rs5743836, did not affect TLR9 regulation by oestrogens (data not shown).

The concentrations of both naturally occurring oestrogens fluctuate according to menstrual cycle.39 Thus, heterozygous or C homozygote women would be expected to be less sensitive to menstrual cycle-linked TLR9 downregulation and maintain higher TLR9 levels, which might reflect positively on HCV clearance since heterozygous SNP carriers show better clearance of HCV on an epidemiological (table 2) level.

This is in agreement with the recent studies which provide intriguing evidence for a strong impact of sex hormones on the natural course of HCV infection.46 ,47 The results elucidate the sex-specific differences during infection and clarify the reason for the higher rates of SC rates in HCV-infected women compared with men.48 ,49

It will be interesting to monitor the fluctuation of TLR9 levels over the menstrual cycle and during pregnancy and also to re-examine the effect of HPV in different allele carriers. Collectively, whereas the impact on basal transcription for rs187084 awaits further investigation, rs5743836 might thus have an effect on both IL-6-dependent signalling and combined ERα/NF-κB chromatin immunoprecipitation (IP) analysis in cells from different allele carriers following stimulation/infection with HCV, IL-6, purified TLR ligands and/or estrogens may be informative to dissect the relative contributions of these different afferent signals and their net effect on the clearly complex regulation of TLR9 in human cells. The mechanism by which TLR9 contributes to antiviral signalling during HCV infection remains poorly understood. It has been proposed that pDCs can be activated via TLR9 through the HCV RNA polyuridine tail,13 although this awaits further confirmation and would be informative to verify in liver cells which also express TLR9.7 ,8 In any case, it appears plausible that modulation of TLR9 mRNA levels may enhance or suppress the functionality of any directly TLR9-dependent response to HCV components or possibly endogenously generated Damage-associated molecular pattern molecules (eg, high-mobility group protein 1 DNA complexes50) released during the infectious cycle. Alternatively, it has been shown in mice that TLR9 levels impact on TLR7 recognition by affecting the latter receptor's trafficking.51 Thus, it would also be informative to compare TLR7 responsiveness in different TLR9 allele carriers stimulated with the above modifiers of TLR9 transcription.

Our study has some limitations. The sample number in this study was moderate and all chronic subjects were from hospitals and may not be representative of the target population. Selection bias cannot be excluded. Thus, other large population-based studies are needed to validate the findings further. On the other hand, the analysis of four independent cohorts argues against the observed effects resulting from selection bias only. The functional analyses of the TLR9 polymorphisms were performed in B cells as professional TLR9+human cell type. Therefore, confirmation of the differences in promoter activity in human primary hepatocytes, which serve as target cells for HCV infection, or pDCs, which are involved in HCV surveillance, is required but unfortunately, it is technically highly challenging. Additionally, the mechanisms by which TLR9 is regulated and involved in recognition and clearance of HCV need to be further elucidated.

In conclusion, our data provide strong indirect evidence that TLR9 might play a greater role in HCV infection than previously expected. We identified an association of the polymorphism rs187084 within the TLR9 gene with the natural course of HCV infection in women. Functional analyses revealed that the SNPs affect the basal TLR9 expression. The combined determination of the well-known IFNL4 rs12979860 and TLR9 rs187084 might serve as a further tool to identify patients with an increased likelihood of chronic manifestation of HCV infection.


The authors thank the members of the East German HCV study group, the Irish HCV Research Consortium, the Swiss Hepatitis C Cohort Study (SCCS, members see online supplementary material), the investigators, technical staff, healthy volunteers and patients involved in this study.


View Abstract

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.