The hallmark of HIV pathogenesis is sustained immune activation and dysfunction.1 Chronic antigen stimulation, microbial translocation, and stimulation of the innate immune response,2 all, may contribute to this phenomenon. In this context, we note sustained interferon (IFN) type I signaling in chronic HIV infection.3 Indeed, IFN plasma levels correlated with the plasma viral load and inversely with the CD4+ T-cell count.4 This goes together with a relative overexpression of IFN-stimulated genes (ISGs). A model has been postulated on the basis of these findings that HIV promotes IFN signaling and, in turn, IFN promotes the activation and proliferation of CD4+ T cells, thereby increasing the number of HIV target cells while resulting in immune dysfunction and exhaustion.3 Alternatively, these findings of increased IFN signaling might simply be interpreted as an epiphenomenon of higher viral replication and advanced disease state. In simian immunodeficiency virus–infected rhesus macaques, the administration of IFN-α2 continued beyond the acute phase resulted in an IFN-desensitized state with decreased antiviral gene expression, increased susceptibility to infection, increased cell-associated virus load, and greater CD4+ T-cell depletion.5 In chronically HIV-infected humanized mice, blocking the IFN receptor (IFNAR) resulted in increased viral replication but rescued both total human T-cell and HIV-specific T-cell numbers.6 Similarly, blocking the IFNAR in HIV-infected humanized mice under suppressive combined antiretroviral treatment restored immune function.6,7 These data would rather favor a causal role of chronic IFN activation in HIV pathogenesis. Notably, chronic IFN activation being at the origin of persistent infections has been convincingly demonstrated in the lymphocytic choriomeningitis model.8
Apart from chronic IFN type I activation, low levels of IFN-γ are detected throughout the course of HIV infection.9 It seems that IFN-γ expression has no predictive value for HIV viral set point, mortality, or disease progression rate. Based on its polyfunctional effects on immune responses, IFN-γ was even explored as a treatment strategy for HIV/AIDS but had no therapeutic efficacy.9 By contrast, IFN-γ–induced protein 10 (IP-10 or CXCL10), which is driven predominantly by IFN-γ, shows increased plasma levels in HIV infection.10 Most studies reported a positive correlation between IP-10 levels and disease progression rate.10 High IP-10 levels suppress the functions of T cells and natural killer cells and promote HIV latency and replication.10 Notably, to what extent IFN-γ adds to the HIV-associated immune activation remains largely unknown.
IFN type I and IFN-γ signal through their JAK/STAT pathway after binding to their cognate IFN receptor, which results in the upregulation of hundreds of ISGs.11 Effective immune responses to pathogens need a fine-tuned regulation of proinflammatory and anti-inflammatory factors. In addition, undamped immune activation results in excessive immunopathology. Thus, already in the first wave of ISGs, negative feedback regulators to the IFN axis are upregulated, among other suppressors of cytokine signaling (SOCS).12 SOCS proteins inhibit signal transducer and activator of transcription (STAT) protein phosphorylation by binding and inhibiting Janus kinases, by their increased proteosomal degradation or by competing with STAT proteins for phosphotyrosine binding sites on cytokine receptors.13,14 The family of SOCS proteins consist of 8 members, including SOCS 1–7 and cytokine-inducible Src homology 2 (SH2)-containing protein. Mice lacking SOCS-1 show features reminiscent of HIV pathogenesis, that is, destruction of the lymphoid organs, a loss of CD4+ T cells, and sustained immune activation. SOCS proteins are very short-lived proteins, and their quantification is very delicate. A number of studies reported that HIV results in increased expression of SOCS at the mRNA level, but they did not measure it at the protein level. They concluded that SOCS is increased in HIV and attenuates the anti-HIV IFN response.15–18 By contrast, we found that HIV-infected patients have lower SOCS protein levels than matched controls and interpreted the data that HIV interferes with negative SOCS feedback mechanisms.19
Here, we hypothesized that functionally different human SOCS alleles underlie the distinct HIV progression rate, that is, SOCS with higher activity might attenuate IFN signaling, thereby lessening the immune activation state. Notably, HIV-infected patients commonly progress from transmission to late-stage disease, with <200 CD4+ T cells/μL within 7.5–12 years. However, HIV-infected patients may show a very rapid disease progression rate (3–4 years from transmission to the AIDS phase) or be long-term nonprogressors.20 Long-term nonprogressors are often asymptomatic for 10–20 years with CD4+ T-cell counts >500 cells/μL. Because the chips used for a genome-wide association study (GWAS) do not cover the closer SOCS-1 and SOCS-3 genetic region, the specific aims of this study were to define the prevalent single-nucleotide polymorphisms (SNPs) present in SOCS-1 and SOCS-3 in a cohort of volunteers donating blood and then to examine whether these polymorphisms may be linked to the HIV progression rate.
MATERIAL AND METHODS
Swiss HIV Cohort Study
The Swiss HIV Cohort Study (SHCS), established in 1988, is a systematic longitudinal study enrolling HIV-infected patients in Switzerland. The patients in the SHCS have given their informed consent for genetic testing, and we have obtained the samples for this study from the biobank of the SHCS.
As we did in a previous study, HIV-infected patients from the SHCS had been categorized according to the decline in CD4 counts as follows: slow progressors, intermediate progressors, and rapid progressors.21,22 The categories were created in 2 steps. First, we estimated for each patient the CD4 decline over time before start of any antiretroviral therapy (ART). Second, we grouped the patients by tertiles of the CD4 decline (top 33% of the steepest decline being the rapid progressors and the lowest 33% in decline being the slow progressors). Thus, the rapid progressor group had an annual decline of CD4 cells larger than 101.5, and the slow progressor group had an annual decline of less than 31.7 cells/yr before ART start. For patients with <2 CD4+ T-cell measurements between baseline and start of any ART, we could not estimate a decline, and we excluded them (n = 69). Additional 24 patients without RNA measurements in 18 months before highly active ART start (or stop date) have been excluded as well, resulting in a total of 1051 patients fulfilling the criteria of ≥2CD4+ T-cell and RNA measurements. The number of patients included in our analyses was 1144 for whom DNA could be obtained. Hence, this sample is a representative of the overall SHCS patient population (Table 1).
SNPs and Haplotype Analysis
Screening for Polymorphisms in SOCS-1 and SOCS-3 in the General Population
We screened 96 DNA samples collected from healthy Swiss blood donors (Swiss Red Cross Zurich, https://www.blutspendezurich.ch/) using denaturing high performance liquid chromatography23 for identifying the variations of SOCS-1 and SOCS-3 genes in our target population using the WAVE DNA fragment analysis system (Transgenomic, Berlin, Germany). These genes are located on chromosomes 16 and 17, SOCS-1: 16p13.13 and SOCS-3: 17q25.3, respectively. Screening covered the coding region and the intron–exon boundaries as well as 1.5 kb upstream and 1 kb downstream of the genes. This screening detected the following variations in the Swiss population: SOCS-1 rs193778, rs243330, rs193779, rs33989964, rs33977706, and rs4780355; and SOCS-3 rs8064821, rs7207782, rs563935021, rs199915361, rs4969169, rs4969168, and rs12185261.
Genotyping and Haplotype Analysis of the Polymorphisms in SOCS-1 and SOCS-3 in the General Population
To genotype the 13 detected variations, we established genotyping assays by high-resolution melting technique and fluorescence resonance energy transfer on a LightCycler 480 II instrument (Roche, Basel, Switzerland) and by allele-specific amplification on agarose gels.24 Haplotype frequencies were estimated with the software Haploview.25
Genotyping of additional 176 DNAs from healthy Swiss blood donors revealed that 2 polymorphisms are inherited together with other polymorphisms in healthy Swiss volunteers and 2 are rare, limiting the informative polymorphisms to 5 polymorphisms for SOCS-1 and 4 polymorphisms for SOCS-3 (see Table, Supplemental Digital Content, https://links.lww.com/QAI/B441).
Statistical Analysis in Rapid and Slow Progressors
In all the conducted analyses, we compared the rapid progressors (n = 318) with the slow progressors (n = 376). For univariate comparisons of these groups, we used χ2 statistics for categorical information and t tests for continuous information.
We performed 2 different analyses, one comparing rapid progressors with slow progressors with regard to SNPs in SOCS-1 and SOCS-3 and the other comparing the 2 groups with regard to haplotypes. In both of these analyses, we fitted multivariate logistic regression models and report odds ratios (ORs) and 95% confidence intervals. The model included sex, binary risk group [intravenous drug user (IDU) vs non-IDU], quintiles of baseline CD4, and quintiles of baseline viral load. The SNP-based analyses contained 3 genotype groups as follows: homozygote genotype (AA—reference), heterozygote (AB or BA), and the less frequent versions of the homozygote genotypes (BB).
In the analysis of haplotypes, we first tested for linkage disequilibrium using Haploview.25 Haplotypes were inferred using an expectation–maximization algorithm implemented in the SNPHAP program (version 1.3.1, developed by David Clayton) (https://www-gene.cimr.cam.ac.uk/staff/clayton/software/snphap-1.3.1.zip).
Sensitivity analyses were performed with rapid progressors versus all others (intermediate and slow progressor groups combined), including only white people and recessive models for the analysis of SNPs [homozygote genotypes BB vs all others (AA, AB/BA)].
All analyses were performed using STATA version 13.1 (College Station, TX: StataCorp LP, 2013), and P values less than 0.05 were considered as statistically significant.
Participants in the genetic study were mainly male whites (74%) with a median age of 36 years, similar to those included in the entire SHCS (Table 1). The 3 main risk groups for HIV infection were homosexuals (41% in the genetic cohort and 32% in the general cohort), heterosexuals (36% and 31%), and IDUs (12% and 23%).
Identification of SOCS-1 and SOCS-3 Polymorphisms
To identify polymorphisms in the SOCS-1 and SOCS-3 genes, we screened the promoter, the coding region, and the 3′ UTR region of SOCS-1 and SOCS-3 in a representative population of 96 healthy Swiss blood donors. Subsequently, we selected 5 SNPs in SOCS-1 and 4 SNPs in SOCS-3 with a minor allele frequency of greater than 0.05 for the analysis with the progressor status in the Swiss HIV Genetic Cohort. The analysis of these SNPs in the Swiss HIV Genetic Cohort revealed that all SNPs are in Hardy–Weinberg equilibrium (Table 2). All SNPs with negative values are located in the promoter region of the gene, whereas rs4780355 (c.*842A>G) and rs4969169 (c.*589A>G) are located in an intron and in the 3′ UTR, respectively.
Association of SOCS-1 With the Rapid Progressor Phenotype
To investigate the association of SNPs in SOCS-1 and SOCS-3 with the HIV progressor phenotype, we divided the Swiss HIV Genetic Cohort into 3 groups of slow progressors, intermediate progressors, and rapid progressors according to the decline in CD4 counts.21,22 Comparison of the rapid and slow progressor subgroups showed clearly the distinct CD4+ T-cell decline rate and that the distribution of the ethnicities is similar (Table 3).
Four SOCS-1 SNPs were associated with the rapid progressor phenotype (Table 4). When using genotypes, the OR was 0.39 (0.16–0.96) for rs193779 TT vs CC, 0.34 (0.15–0.81) for rs33989964 CA/CA vs −/−, 2.22 (1.06–4.65) for rs33977706 TT vs GG, and 0.47 (0.25–0.90) for rs4780355 GG vs AA. Similar associations but with slightly higher significance were obtained when using a recessive allelic mode of inheritance (data not shown). By contrast, no association of these polymorphisms was detected with the viral load at the initial visit. To investigate whether a potential bias in ethnicities could interfere with the analysis, we performed a sensitivity analysis using only whites. We observed that the directions of the associations and the odds ratios are similar when looking at whites alone to the results obtained when including all ethnicities (data not shown). In addition, we performed as sensitivity analysis the multivariate logistic regression of rapid progressors versus all others (intermediate and slow) and also obtained similar results (data not shown).
The 3 SOCS-1 SNPs, rs193779, rs33989964, and rs4780355, show a similar risk reduction for the rapid progressor phenotype of approximately 60%, suggesting that some of these SNPs may be in linkage disequilibrium. We therefore investigated the linkage disequilibrium between all the SNPs (see Table 1, Supplemental Digital Content, https://links.lww.com/QAI/B441) and analyzed the association of the haplotypes with the progressor phenotype (Table 5). We observed a linkage between rs33989964 and rs4780355 with an r2 of 0.7, whereas none of the other SNPs were in a strong linkage disequilibrium. Accordingly, the minor alleles of these 2 SNPs convene on the frequent AC+GG haplotype, which is in homozygous carriers also associated with a risk reduction for the rapid progressor phenotype.
In contrast to the 3 SOCS-1 SNPs associated with a risk reduction for the rapid progressor phenotype, the minor TT genotype of rs33977706 is associated with twice the risk for the rapid progressor phenotype. Haplotype analysis revealed that the minor allele of these SNPs is mainly observed on the GC-TA haplotype, distinct from the AC+GG haplotype associated with the risk reduction. Consequently, we also observed a borderline increased risk for homozygous carriers of this GC-TA haplotype for the rapid progressor group (Table 5).
No associations have been observed for the 4 SOCS-3 polymorphisms or for the major SOCS-3 haplotypes (Tables 4 and 5).
Progressive natural HIV infection is characterized by chronic IFN-α and IFN-γ signaling. This chronic IFN activation is believed to contribute to proliferation and activation of CD4+ T cells, thereby increasing the number of available HIV target cells while promoting T-cell exhaustion, eventually resulting in a perpetuating vicious cycle. Notably, IFN signaling is attenuated by negative feedback regulators, among others (SOCS-1 and SOCS-3). Here, we wondered whether distinct polymorphism(s) in SOCS-1 and SOCS-3 are associated with an HIV progression rate by comparing rapid vs slow progressors. We found (1) several SNPs in the gene region of SOCS-1 and SOCS-3 in healthy volunteers and (2) 3 SNPs to be directly associated and 1 SNP to be inversely associated with an HIV progression rate. These data are consistent with the impact of HIV-mediated IFN signaling being critical in HIV pathogenesis.
SNPs of SOCS-1 and SOCS-3 have not been thoroughly sought for in whites, and only a few SNPs in their gene region have been described but none in the promoter or coding region. Thus, we screened the first 96 volunteers for SNPs in SOCS-1 and SOCS-3 and found a large number of SNPs in the promoter and coding region of SOCS-1 and SOCS-3 (Table 2). We focused on the SNPs with an allelic frequency higher than 5%, which were not inherited together to 100% in the general population to investigate their association with natural HIV disease progression in the genetic cohort study of the SHCS (Table 2).
We found that 4 polymorphisms of SOCS-1 were associated with rapid progression of HIV but none in SOCS-3. However, the polymorphisms did not show any association with the viral load at the initial visit. Thus, the effects by the SOCS-1 and SOCS-3 feedback regulators appear to be uncoupled from the viral load, that is, the SNPs identified may primarily have an effect on the intricate process of CD4+ T-cell depletion but not at firsthand on the extent of viral replication. This comes back to the egg and hen problem whether immune activation drives CD4+ T-cell depletion and viral replication, or vice versa, viral replication drives immune activation and CD4+ T-cell depletion, or whether different mechanism(s) underlie those pathogenic events. It might even be that all 3 settings exist with one being more dominant than the other at different stages in HIV pathogenesis. This would explain the controversies in the literature. The lack of any association with SOCS-3 is rather astonishing in view of the data of SOCS-1. We might explain this selective association with SOCS-1 by the higher hierarchical role of SOCS-1 in attenuating rather all cytokines using Janus kinases.26 These data corroborate the imminent role of IFNs and other pathways in HIV pathogenesis.
The association of allelic variants of SOCS-1 has been explored in a variety of diseases, among others in allergic and inflammatory diseases, metabolic diseases, as well as cancer (see Table 2, Supplemental Digital Content, https://links.lww.com/QAI/B441). However, the results from these studies did not reveal an obvious genetic denominator SNP in SOCS-1 with an effect on disease outcomes but rather indicated that the genetic locus may be associated with the investigated diseases. This is caused by the broad range of SNPs and haplotypes analyzed in these studies covering a large genetic region encompassing SOCS-1, which were often different between studies, and by the limited power of most studies to detect polygenic effects. In lack of corroborating associations of 1 SNP or haplotype with disease, studies to investigate a potential functional effect of 2 SOCS-1 SNPs have been performed. The study by Harada et al showed that the del (−) allele at rs33989963 has increased SOCS-1 protein levels in human primary nasal fibroblasts and heterozygous carriers of the deletion had a higher risk for adult asthma.27 Similarly, the T-allelic variant of rs33977706 within the SOCS-1 promoter increases the transcriptional activity of the SOCS-1 gene in transfection experiments and was associated with lower IgE levels in plasma.29 In our analysis, we observed associations in opposite directions for these 2 SNPs with a potential functional increase in SOCS-1 activity, one associated with the rapid progressor phenotype and the other associated with the slow progressor phenotype. However, there are only mechanistic data for increased SOCS-1 protein activity for rs33989963 where the del (−) allele showed increased SOCS-1 protein levels and reduced STAT-1 phosphorylation.27 Such an increased SOCS-1 activity that results in a more efficient attenuation of the Jak/STAT pathway and is associated with the slow progressor phenotype would argue that preserved triggering through the IFN axis is beneficial to HIV infection.
However, what a less prominent ISG response finally means remains unknown—we might speculate that a less prominent ISG response promotes viral replication; on the other side, a less prominent ISG response should trigger less immune activation. In fact, the role of IFNs in HIV remains highly disputed.3 For further defining the role of SOCS-1 in HIV infection, a quantitative analysis of SOCS-1 protein would need to be performed. This task is very demanding because of their transient nature and their manifold influences they are subjected to—a single snapshot analysis would not be meaningful. Alternatively, an interventional study would be attractive, modulating SOCS-1. This remains hypothetical because such compounds are lacking.27
We are aware of the ongoing debate about the value of candidate gene approaches as opposed to GWAS.28,29 It might be argued that multiple testing in studies using the candidate gene approach might provide false-positive results; thus, the data obtained in our cohort need to be validated in the other cohort of HIV-infected patients. On the other hand, GWAS are as good as the coverage of the annotations of the genes, and both SOCS genes are badly covered on the SNP chips. We consider the approaches complementary, imminent for getting a deeper insight into HIV pathogenesis, and as a first step for explorative studies.
In summary, SNPs in SOCS-1 are associated with HIV disease progression rate, pointing to the imminent role the Jak/STAT axis has in HIV pathogenesis. In addition, allelic variants of SOCS-1, with either decreased or increased transcriptional activity of SOCS-1, go along with slow and rapid progressor status, respectively. The latter findings imply that attenuating the Jak/STAT pathway favors the HIV progression rate. Notwithstanding, we lack clinical studies, which specifically interfere with the IFN axis for ultimate proof of its pathogenic significance.
1. Vidya Vijayan KK, Karthigeyan KP, Tripathi SP, et al. Pathophysiology of CD4+ T-cell depletion in HIV-1 and HIV-2 infections. Front Immunol. 2017;8:580.
2. Baenziger S, Heikenwalder M, Johansen P, et al. Triggering TLR7 in mice induces immune activation and lymphoid system disruption, resembling HIV-mediated pathology. Blood. 2009;113:377–388.
3. Doyle T, Goujon C, Malim MH. HIV-1 and interferons: who's interfering with whom? Nat Rev Microbiol. 2015;13:403–413.
4. Hardy GA, Sieg S, Rodriguez B, et al. Interferon-alpha is the primary plasma type-I IFN in HIV-1 infection and correlates with immune activation and disease markers. PLoS One. 2013;8:e56527.
5. Sandler NG, Bosinger SE, Estes JD, et al. Type I interferon responses in rhesus macaques prevent SIV infection and slow disease progression. Nature. 2014;511:601–605.
6. Cheng L, Ma J, Li J, et al. Blocking type I interferon signaling enhances T cell recovery and reduces HIV-1 reservoirs. J Clin Invest. 2017;127:269–279.
7. Zhen A, Rezek V, Youn C, et al. Targeting type I interferon-mediated activation restores immune function in chronic HIV infection. J Clin Invest. 2017;127:260–268.
8. Wilson EB, Yamada DH, Elsaesser H, et al. Blockade of chronic type I interferon signaling to control persistent LCMV infection. Science. 2013;340:202–207.
9. Roff SR, Noon-Song EN, Yamamoto JK. The significance of interferon-gamma in HIV-1 pathogenesis, therapy, and prophylaxis. Front Immunol. 2014;4:498.
10. Lei J, Yin X, Shang H, et al. IP-10 is highly involved in HIV infection. Cytokine. 2019;115:97–103.
11. Borden EC. Interferons alpha and beta in cancer: therapeutic opportunities from new insights. Nat Rev Drug Discov. 2019;18:219–234.
12. Hertzog PJ, Williams BR. Fine tuning type I interferon responses. Cytokine Growth Factor Rev. 2013;24:217–225.
13. O'Sullivan LA, Liongue C, Lewis RS, et al. Cytokine receptor signaling through the Jak-Stat-Socs pathway in disease. Mol Immunol. 2007;44:2497–2506.
14. Yoshimura A, Nishinakamura H, Matsumura Y, et al. Negative regulation of cytokine signaling and immune responses by SOCS proteins. Arthritis Res Ther. 2005;7:100–110.
15. Moutsopoulos NM, Vazquez N, Greenwell-Wild T, et al. Regulation of the tonsil cytokine milieu favors HIV susceptibility. J Leukoc Biol. 2006;80:1145–1155.
16. Yadav A, Fitzgerald P, Sajadi MM, et al. Increased expression of suppressor of cytokine signaling-1 (SOCS-1): a mechanism for dysregulated T helper-1 responses in HIV-1 disease. Virology. 2009;385:126–133.
17. Akhtar LN, Qin H, Muldowney MT, et al. Suppressor of cytokine signaling 3 inhibits antiviral IFN-beta signaling to enhance HIV-1 replication in macrophages. J Immunol. 2010;185:2393–2404.
18. Sachdeva M, Sharma A, Arora SK. Increased expression of negative regulators of cytokine signaling during chronic HIV disease cause functionally exhausted state of dendritic cells. Cytokine. 2017;91:118–123.
19. Miller RC, Schlaepfer E, Baenziger S, et al. HIV interferes with SOCS-1 and -3 expression levels driving immune activation. Eur J Immunol. 2011;41:1058–1069.
20. Gonzalo-Gil E, Ikediobi U, Sutton RE. Mechanisms of virologic control and clinical Characteristics of HIV+ elite/viremic controllers. Yale J Biol Med. 2017;90:245–259.
21. Audige A, Taffe P, Rickenbach M, et al. Low postseroconversion CD4 count and rapid decrease of CD4 density identify HIV+ fast progressors. AIDS Res Hum Retroviruses. 2010;26:997–1005.
22. Bochud PY, Hersberger M, Taffe P, et al. Polymorphisms in Toll-like receptor 9 influence the clinical course of HIV-1 infection. AIDS. 2007;21:441–446.
23. Liu W, Smith DI, Rechtzigel KJ, et al. Denaturing high performance liquid chromatography (DHPLC) used in the detection of germline and somatic mutations. Nucleic Acids Res. 1998;26:1396–1400.
24. Bühler MM. Suppressors of cytokine signalling (SOCS) in human immunodeficiency virus (HIV) pathogenesis. Dissertation. 2015:12–20.
25. Barrett JC, Fry B, Maller J, et al. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics. 2005;21:263–265.
26. Yoshimura A, Ito M, Chikuma S, et al. Negative regulation of cytokine signaling in immunity. Cold Spring Harb Perspect Biol. 2018;10:a028571.
27. Sharma J, Larkin J III. Therapeutic implication of SOCS1 modulation in the treatment of autoimmunity and cancer. Front Pharmacol. 2019;10:324.
28. Pearson TA, Manolio TA. How to interpret a genome-wide association study. JAMA. 2008;299:1335–1344.
29. Naranbhai V, Carrington M. Host genetic variation and HIV disease: from mapping to mechanism. Immunogenetics. 2017;69:489–498.