Toll-like receptors (TLRs) belong to the family of pattern recognition receptors that recognize specific patterns of microbial components and regulate the activation of both innate and adaptive immune responses.1,2 To date, 10 TLRs and their ligands have been described in humans and are expressed by different cells of the immune system, either at the cell membrane (TLR1, 2, 6, 4, 5, 10) or in endosomes (TLR3, 7, 8, 9). These receptors are structurally characterized by the presence of a Toll/IL-1 receptor motif in the intracellular domain. TLR ligation triggers signalling pathways that activate transcription factors, initiating adaptive immune responses.3,4
Growing data supporting associations between single-nucleotide polymorphisms (SNPs) in TLRs and the increased risk of bacteria and viral infections are being reported. These include TLR2, that recognizes the widest repertoire of pathogen-associated molecular patterns due to its ability to recognize ligands as a heterodimer with TLR1 and TLR6,5 and TLR4, that is highly specific for lypopolysaccharide, the main component of the gram-negative bacteria cell wall6 and respiratory syncytial virus.7 Hence, the association between common SNPs in TLR2 and TLR4 and several infectious diseases has been deeply studied demonstrating that TLR2 Arg753Gln (2258G/A) and Pro631His (1892C/A) and TLR4 Asp299Gly (896A/G) and Thr399Ile (1196C/T) might be involved in the susceptibility to bacterial and/or viral infections.7-12
In the context of HIV infection, the role of SNPs in TLRs may provide relevant information on HIV pathogenesis because the in vitro activation and signalling through TLR4,13 TLR2, and TLR914 might enhance HIV replication. Furthermore, HIV may impair the in vitro immune response after TLR415,16 and TLR216 stimulation. Only 2 recent studies have analyzed the role of SNPs in TLRs in the clinical context of HIV infection. One of them reported a potential association between the TLR4 Asp299Gly SNP and a higher susceptibility to develop active tuberculosis in HIV-infected patients in Tanzania.17 The other study reported an association between CD4 cell depletion and the 1635A/G SNP in the TLR9 gene among rapid progressors, failing to find an association with HIV viral load.18 In this way, given the central role that TLRs play in both innate and adaptive immune responses, understanding the mechanisms and factors involved in HIV clinical progression is of particular importance. Although other different SNPs in these TLRs have previously been associated with other infections,19,20 we have focused the present study in the investigation of the role of SNPs in TLR2 (1892C/A and 2258G/A), TLR4 (896A/G and 1196C/T), and TLR9 (1635A/G) genes on CD4 count, HIV viral load, and clinical disease progression in a cohort of naive HIV-infected patients.
The open, seroprevalent, and dynamic therapeutic cohort from the Viral Hepatitis and AIDS Study Group of Virgen del Rocío Hospital, Seville, Spain, includes 487 HIV-infected patients, who have been enrolled since 1989. Clinical and immunovirological data had been recorded at the beginning of the inclusion of patients in the cohort and afterward on a 12-week basis. Selection criteria to be included in the present study were (1) to be naive for any antiretroviral therapy at the entry in the cohort; (2) to have frozen peripheral blood mononuclear cell samples available for genotyping; and (3) to belong to the white Mediterranean ethnic group. White ethnicity was assessed by personal interview. A total of 369 Spanish naive HIV-infected patients were finally included. Patients started either mono/bitherapy or highly active antiretroviral therapy (HAART) depending on the date of their inclusion, that is, before 1996, they all started mono/bitherapy (n = 192), and afterward, all patients started HAART (n = 142). In addition, 35 patients were only used for the analyses at cohort entry because their follow-up was lost after their first visit. Patients had given written informed consent, and the ethical committee of the hospital approved the study. The study was censored in September 30, 2007, being the median observational period of 7.3 (2.5-12.2) years.
We analyzed TLR2 1892C/A and TLR9 1635A/G alleles in a group of healthy Spanish white Mediterranean (self-reported) volunteers due to the lack of previous data in our country. The frequencies of TLR2 2258G/A, TLR4 896A/G, and TLR4 1196C/T in healthy Spanish white Mediterranean people had been previously reported.21,22
CD4 count and HIV viral load were determined at the entry of patients in the cohort. CD4 count was determined in fresh samples by conventional flow cytometry. Plasma HIV-1 RNA was measured by a quantitative polymerase chain reaction (PCR) method (HIV Monitor Test Kit; Roche Molecular System, Hoffman-La Roche, Basel, Switzerland) according to the manufacturer's instructions and was available in 163 patients (limitation imposed by the time of introduction of the method in the clinical practice).
Total DNA was automatically extracted from peripheral blood mononuclear cells with MagNa Pure LC system (Roche Diagnostics, Mannheim, Germany) using the MagNa Pure LC DNA isolation kit, according to manufacturer's instructions. Ten nanogram per microliters of DNA was used to perform the PCRs. Primers were designed to amplify both polymorphisms within each gene in the same amplicon for TLR2 and TLR4 genes. PCRs were carried out in 96-well plates at a final volume of 20 μL using 10 ng of genomic DNA, 1.5 mM of MgCl2, 125 μM of each dNTP, 2U of Taq polymerase, and 0.25 mM of the primer that generates the strand that binds to the detection probes (forward primer for TLR2 2258G/A and TLR4 896A/G SNPs) and 0.05 mM of the other primer (reverse primer for TLR2 1892C/A, TLR4 1196C/T, and TLR9 1635A/G SNPs). PCR amplification was performed with an initial denaturation at 95°C for 5 minutes, 45 cycles of denaturation (95°C for 30 seconds), annealing (53°C for 30 seconds, for the TLR2 gene; 61°C for 20 seconds, for the TLR4 gene; and 55°C for 30 seconds, for the TLR9 gene), and elongation (72°C for 30 seconds, for the TLR2 and TLR9 genes and 72°C for 45 seconds, for the TLR4 gene).
Genotypes were determined using the LightCycler 480 System (Roche Diagnostics). Detection probes for each polymorphism were designed using the LightCycler Probe Design Software 2.0. Melting curves analyses for TLRs genes were performed using 0.2 μM of each detection probe. After an initial denaturation at 95°C for 2 minutes at a ramp rate of 4.4°C/s, temperature was dropped to 45°C at a ramp rate of 1°C/s and finally led to 80°C with one acquisition per degree Celsius. Primers, detection probes, and melting temperatures for each allele are shown in Table 1.
To analyze deviation from Hardy-Weinberg equilibrium and to compare allelic frequencies between cases and controls, tests adapted from Sasieni23 at the online resource available at the Institute for Human Genetics, Munich, Germany,24 were used. A recessive model for the wild-type allele was used for the comparison between the different genotypes because the likelihood of this model (R2 = 0.4) better fitted than either the additive (R2 = 0.2) or the dominant (R2 = 0.18) models. Continuous variables are expressed as median (interquartile range) and categorical ones as number of cases (percentage). The Mann-Whitney U test was used to analyze differences between continuous variables whereas categorical ones were compared by χ2 test.
To assess the potential variables associated with the CD4 count and the HIV viral load at patients' entry in the cohort, those factors with a P < 0.1 in the univariate analysis were introduced in the multivariate forward stepwise multiple linear regression analysis. The HIV clinical progression was analyzed under 2 points of view: first, the event of interest was defined as progression from clinical stage A or B to C and second, we analyzed the progression from clinical stage A or B to death due to AIDS-related events. Patients who had not died due to AIDS were considered censored at the date of death. For the analyses of the HIV clinical progression, we first performed the Kaplan-Meier curves for the survival analysis using the log-rank test to compare between categories. Continuous variables such as age and CD4 count were categorized by their median value to perform this analysis. Next, those variables significantly associated with the end point were introduced in the univariate Cox regression analysis; and finally, those variables with a P < 0.1 were introduced in the multivariate forward stepwise Cox regression analysis to analyze the independent factors associated with either progression to clinical stage C or death due to AIDS-related events. The statistical analysis was performed using the Statistical Package for the Social Sciences software (SPSS 15.0, Chicago, IL).
The characteristics of the 369 HIV-infected patients at their entry in the cohort are summarized in Table 2. In addition, the number of patients who entered in the cohort by 2 periods of 10 years is also shown (first period: January 1, 1989, to December 31, 1997, and second period: January 1, 1998, to September 30, 2007). Of these 369 patients, DNA amplification was possible in 368 samples for TLR2 and TLR4 genes and in 365 patients for the TLR9 gene. Frequencies of TLR2 2258G/A and TLR4 896A/G and 1196C/T SNPs in our HIV-infected population were similar to the previously reported non-HIV-infected individuals in Spain.21,22 In addition, there were no statistical differences between the control group we had analyzed and the HIV-infected population according to the frequencies of TLR2 1892C/A (n = 214) and TLR9 1635A/G (n = 194) SNPs (Table 3). All genotype frequencies were in accordance with the Hardy-Weinberg equilibrium.
Table 4 shows the univariate analysis of the potential association between patients' characteristics at their entry in the cohort with both the CD4 count and the HIV viral load. The factors included in the univariate analysis were age (as continuous variable) and the categorical variables: sex, route of transmission, and the different SNPs in TLR2, TLR4, and TLR9. In addition, because we only knew the time of infection of a few patients, we tried to minimize this limitation analyzing the potential differences according to the periods of years when patients entered in the cohort. We found that age inversely correlated with CD4 count (P < 0.001, r = −0.26) and positively with HIV viral load (P = 0.004, r = 0.23). In addition, patients who sexually acquired the virus showed higher HIV viral load than injecting drug users (IDUs) (P = 0.011), and also patients who entered in the cohort in the first period showed higher CD4 count than those who entered in the second period (P = 0.04). There were no statistical associations with the end points and the SNPs in TLR2 or TLR4, but patients with the TLR9 1635AA genotype showed statistically lower CD4 count and higher HIV viral load than patients with the other genotypes (P = 0.003, P = 0.018, respectively).
Next, the variables significantly associated either with the CD4 count or the HIV viral load in the univariate analysis were introduced in the univariate and multivariate regression analyses (Table 5). Qualitative variables are represented by their reference category, route of transmission (IDU), periods of patients' inclusion in the cohort (first period), and TLR9 1635AA genotype. For the analysis of the CD4 count, age (P < 0.001), periods of patients' inclusion in the cohort (P = 0.019), and TLR9 1635AA genotype (P = 0.012) were entered in the forward stepwise multiple regression analysis. Only age (P < 0.001) and TLR9 1635AA genotype (P = 0.016) were significantly associated with the CD4 count at patients' entry in the cohort. On the other hand, for the analysis of HIV viral load, age (P = 0.004), IDU (P = 0.006), and TLR9 1635AA genotype (P = 0.013) were required for the multivariate regression analysis. Finally, only age (P = 0.012) and IDU (P = 0.024) showed a significant association with the HIV viral load. Although the TLR9 1635AA genotype showed no statistical association in the multivariate analysis, of note is that only 163 patients had their viral load quantification available, and of them, only 36 harbored the AA genotype.
For all the analyses of the HIV clinical progression, of the 369 HIV-infected patients, those who did not have 2 consecutive visits (35/369) and also those who were in clinical stage C at their entry in the cohort (58/369) were excluded for the studies. Finally, 276 patients were included. Unfortunately, HIV viral load could not be used for these analyses because the sample size did not allow performing them. There were no statistical differences in the frequency of HIV clinical progression according to TLR2 or TLR4 SNPs (P > 0.05). Regarding TLR9 1635A/G SNP, the 33% (20/60) of patients with the AA genotype and the 21% (45/216) of patients with the AG+GG genotypes had progressed from clinical stage A or B to C (P = 0.04). In the same way, the 17% (10/60) of patients with the AA genotype and the 7% (15/216) of patients with AG+GG genotypes had died due to AIDS-related events (P = 0.02).
Kaplan-Meier curves were used to assess the relation between the potential explicative variables of HIV clinical progression during the observational period of our study. As described above, the event of interest was defined as progression from clinical stage A or B to either stage C or death. For the univariate analyses, age, sex, CD4 count, therapy regimen [patients who had only received mono/bitherapy (absence of HAART) vs patients who had received HAART], route of transmission (sexual vs IDU), and the SNPs in TLRs were compared using the log-rank test. Only CD4 count (P = 0.001), therapy regimen (P < 0.001), and the TLR9 SNP (P = 0.031, Fig. 1A) showed statistical differences according to the progression to clinical stage C. In the same way, CD4 count (P = 0.002), therapy regimen (P < 0.001), and the TLR9 SNP (P = 0.017, Fig. 1B) were significantly associated with the progression to death.
Factors with statistical differences in the previous univariate analysis assessed by the log-rank test (CD4 count, therapy regimen, and the TLR9 SNP) were introduced in the univariate and multivariate Cox regression analyses to assess the potential variables independently associated with HIV clinical progression either to stage C or to death due to AIDS-related events during the observational period of our study (Table 6). Qualitative variables are represented by the reference category. The 3 of them showed a P < 0.1 in the univariate Cox analysis, and finally the forward stepwise multiple Cox regression analysis showed that CD4 count (P < 0.001), absence of HAART (P < 0.001), and TLR9 1635AA genotype (P = 0.035) were significantly associated with the probability of progression to clinical stage C. In addition, as commented above, CD4 count, therapy regimen, and the TLR9 SNP showed statistical association with the progression to death in the univariate analyses and were introduced in the Cox regression analyses (Table 6). Only the absence of HAART and the TLR9 1635AA genotype showed a P value lower than 0.1 in the univariate analysis and were introduced in the forward stepwise multiple Cox regression analysis. Finally, the analysis of progression to death showed that both the absence of HAART (P < 0.001) and the TLR9 1635AA genotype (P = 0.017) were significantly associated with the probability of progression to death.
The results presented in this study show that patients with the TLR9 1635AA genotype have lower CD4 count, higher HIV viral load, and a higher probability of HIV clinical disease progression than patients with the other genotypes. In addition, the SNPs analyzed in this study for TLR2 and TLR4 genes may not be involved in HIV clinical progression, although due to the low frequency of these SNPs and the limited coverage of the genes variability, further studies are needed.
The allelic frequencies and the genotypic distribution of the SNPs were similar among healthy controls and HIV-infected individuals in both the previously reported21,22 and in the population we have analyzed.
Interestingly, our results showed that both age, as expected, and the TLR9 1635AA genotype were significantly associated with CD4 count in the multivariate analysis. Patients with the TLR9 1635AA genotype showed both lower CD4 count and higher HIV viral load at their entry in the cohort. However, in the analysis of the covariates for the HIV viral load, although the TLR9 SNP was required for the multivariate analysis, finally only age and the route of transmission were significantly associated with this end point. This lack of association may be probably due to the low number of patients whose HIV viral load was available by the time they were included in the cohort because this quantification was introduced in the clinical practice since mid-1990s. It has been previously described in vitro that HIV infection may lead to a noneffective immune response, in addition to an enhanced HIV replication after TLR2, TLR4, and TLR9 stimulation.13-16 Therefore, our results open the possibility that the TLR9 1635AA genotype could be associated in vivo with a functional impairment in the immune response that could lead to a lower CD4 count and a higher HIV viral load. Further functional studies are needed to assess this hypothesis.
Although there is increasing evidence of the complex manner in which HIV interacts with TLR9,25,26 the underlying mechanisms are not still understood. Although TLR9 recognizes unmethylated cytidine-phosphate-guanosine motifs present in viral and bacterial DNA,27 ssRNA is the natural ligand for TLR7.28 Both receptors are mainly expressed in plasmacytoid dendritic cells,29 which are critical antigen-presenting cells linking innate and adaptive immunity that produce large amounts of type 1 interferon upon viral infection.30 However, it is unlikely to think in a highly specific recognition, and therefore TLR9 might interact with other pattern recognition receptors as other TLRs do.31 In this way, functional interactions between TLR7 and TLR9 have been demonstrated in vitro.32 Other hypotheses to explain the interactions between HIV and TLR9 include the phagocytosis of HIV-infected cells.18
Bochud et al18 have recently reported that the TLR9 1635GG genotype was more frequent among rapid progressors, failing to find an association with HIV viral load. Although we do not have an explanation for the differences in their study and ours, it is important to highlight that the design of both studies was different. They assessed the CD4 count decay among patients naive for HAART whereas we have analyzed the association between TLR9 1635A/G SNP and the development of AIDS-related events and death in treated patients. The differences between both studies may also reflect the variable nature of HIV infection among genetically different individuals. Furthermore, although HIV viral load was only available in 163 patients, and therefore our analyses were restricted in this variable, we also found higher HIV viral load in patients with the AA genotype.
On the other hand, our results also show that HIV clinical progression was associated with CD4 count, the absence of HAART, and the TLR9 1635AA genotype. The CD4 count and the high levels of HIV replication at baseline, among others, have been associated with increased rates of clinical progression.33 Therefore, it cannot be excluded that the earlier progression in patients with the 1635AA genotype may be due to the association with lower the CD4 count and the higher HIV viral load at patients' entry in the cohort, despite its significant association in the multivariate analysis. In addition, this SNP could be a marker of some phenotypic change in other genes, thus further functional studies are needed to assess this hypothesis. It is also possible that patients with the TLR9 1635AA genotype had a higher degree of immune activation that is due to a higher rate of microbial translocation, which has been reported to be a better marker of immune activation than HIV viral load.34 Therefore, it can be hypothesized that patients with the TLR9 1635AA genotype would have a higher injury of the immune component of the gastrointestinal tract because this SNP is associated with lower CD4 count that would lead to a higher microbial translocation. The limitations of our survival analyses include that we do not know the time of infection of all patients, which is an important variable involved in the rate of HIV clinical progression, and also that we do not analyze the natural course of the HIV infection because it has been performed in a therapeutic cohort. However, we included the treatment regimen as covariate to minimize the potential bias.
In conclusion, our study shows for the first time that the TLR9 1635AA genotype is associated with lower CD4 count and higher HIV viral load at patients' entry in the cohort and also with the HIV clinical disease progression. Our results should be taken into account because the role of polymorphisms in the activation of TLR9 might be of importance for cytidine-phosphate-guanosine oligonucleotide-based therapeutic approaches.
The authors would like to thank Sonia Molina-Pinelo for critical reading of the manuscript. Juan Manuel Praena Fernández from the Methodology and Research Evaluation Unit at Virgen del Rocío University Hospital and Antonio González-Pérez from Neocodex S.L for statistical support.
1. Gordon S. Pattern recognition receptors: doubling up for the innate immune response. Cell
2. Akira S, Takeda K, Kaisho T. Toll-like receptors: critical proteins linking innate and acquired immunity. Nat Immunol
3. Takeda K, Kaisho T, Akira S. Toll-like receptors. Annu Rev Immunol
4. Akira S, Uematsu S, Takeuchi O. Pathogen recognition and innate immunity. Cell
5. Hajjar AM, O'Mahony DS, Ozinsky A, et al. Cutting edge: functional interactions between toll-like receptor (TLR)2 and TLR1 or TLR6 in response to phenol soluble modulin. J Immunol
6. Takeuchi O, Hoshino K, Hawai T, et al. Differential roles of TLR2 and TLR4 in recognition of gram-negative and gram-positive bacterial cell wall components. Immunity
7. Tal G, Mandelberg A, Dalal I, et al. Association between common toll-like receptor 4 mutations and severe respiratory syncytial virus disease. J Infect Dis
8. Lorenz E, Mira JP, Cornish KL, et al. A novel polymorphism in the toll-like receptor 2 gene and its potential association with staphylococcal infection. Infect Immun
9. Texereau J, Chiche JD, Taylor W, et al. The importance of toll-like receptor 2 polymorphisms in severe infections. Clin Infect Dis
10. Agnese DM, Calvano JE, Hahm SJ, et al. Human toll-like receptor 4 mutations but not CD14 polymorphisms are associated with an increased risk of gram-negative infections. J Infect Dis
11. Lorenz E, Mira JP, Frees KL, et al. Relevance of mutations in the TLR4 receptor in patients with gram-negative septic shock. Arch Intern Med
12. Smirnova I, Mann N, Dols A, et al. Assay of locus-specific genetic load implicates rare toll-like receptor 4 mutations in meningococcal susceptibility. Proc Natl Acad Sci USA
13. Equils O, Faure E, Thomas L, et al. Bacterial lipopolysaccharide activates HIV long terminal repeat through toll-like receptor 4. J Immunol
14. Equils O, Schito ML, Karahashi H, et al. Toll-like receptor 2 (TLR2) and TLR9 signalling results in HIV-long terminal repeat trans-activation and HIV replication in HIV-1 transgenic mouse spleen cells: implications of simultaneous activation of TLRs on HIV replication. J Immunol
15. Tachado SD, Zhang J, Zhu J, et al. HIV impairs TNF-α release in response to toll-like receptor 4 stimulation in human macrophages in vitro. Am J Respir Cell Mol Biol
16. Equils O, Salehi KK, Cornataeanu R, et al. Repeated lipopolysaccharide (LPS) exposure inhibits HIV replication in primary human macrophages. Microbes Infect
17. Ferweda B, Kibiki GS, Netea MG, et al. The toll-like receptor 4 Asp299Gly variant and tuberculosis susceptibility in HIV-infected patients in Tanzania. AIDS
18. Bochud PY, Hersberger M, Taffe P, et al, and the Swiss HIV cohort Study. Polymorphisms in toll-like receptor 9 influence the clinical course of HIV-1 infection. AIDS
19. Sutherland AM, Walley KR, Russell JA. Polymorphisms in CD14, mannose-binding lectin, and toll-like receptor-2 are associated with increased prevalence of infection in critically ill adults. Crit Care Med
20. Bochud PY, Magaret AS, Koelle DM, et al. Polymorphisms in TLR2 are associated with increased viral shedding and lesional rate in patients with genital herpes simplex virus type 2 infection. J Infect Dis
21. Montes AH, Asensi V, Alvarez V, et al. The toll-like receptor 4 (Asp299Gly) polymorphism is a risk factor for gram-negative and haematogenous osteomielitis. Clin Exp Immunol
22. Sánchez E, Orozco G, Lopez-Nevot MA, et al. Polymorphisms of toll-like receptor 2 and 4 genes in rheumatoid arthritis and systemic lupus erythematosus. Tissue Antigens
23. Sasieni PD. From genotypes to genes: doubling the sample size. Biometrics
25. Del Corno M, Gauzzi MC, Penna G, et al. Human immunodeficiency virus type 1 gp120 and other activation stimuli are highly effective in triggering alpha interferon and CC chemokine production in circulating plasmacytoid but not myeloid dendritic cells. J Virol
26. Martinelli E, Cicala C, Van Ryk D, et al. HIV-1 gp120 inhibits TLR9-mediated activation and IFN-α secretion in plasmacytoid dendritic cells. Proc Natl Acad Sci USA
27. Lund J, Sato A, Akira S, et al. Toll-like receptor 9-mediated recognition of herpes simplex virus-2 by plasmacytoid dendritic cells. J Exp Med
28. Lund JM, Alexopoulo L, Sato A, et al. Recognition of single-stranded RNA viruses by toll-like receptor 7. Proc Natl Acad Sci USA
29. Hornung V, Rothenfusser S, Britsch S, et al. Quantitative expression of toll-like receptor 1-10 mRNA in cellular subsets of human peripheral mononuclear cells and sensitivity to CpG olygodeoxynucleotides. J Immunol
30. Siegal FP, Kadowaki N, Shodell M, et al. The nature of the principal type 1 interferon-producing cells in human blood. Science
31. Ishii KJ, Coban C, Akira S. Manifold mechanisms of TLR-ligand recognition. J Clin Immunol
32. Berghöfer B, Haley G, Frommer T, et al. Natural and synthetic TLR7 ligands inhibit CpG-A- and CpG-C- oligodeoxynucleotide-induced IFN- production. J Immunol
33. Egger M, May M, Chêne G, et al, and the ART Cohort Collaboration. Prognosis of HIV-infected patients starting highly active antiretroviral therapy: a collaborative analysis of prospective studies. Lancet
34. Brenchley JM, Price DA, Schacker TW, et al. Microbial translocation is a cause of systemic immune activation in chronic HIV infection. Nat Med
This article has been cited
Keywords:© 2008 Lippincott Williams & Wilkins, Inc.
toll-like receptors; polymorphisms; HIV clinical progression