Background
Psoriasis is a common T-cell mediated autoimmune disease, affecting 2-3% of the world’s population. Genetic studies of psoriasis have identified over 40 susceptibility loci [
1]-[
3]. One of the interesting findings of these studies is the observation that many of these genetic variants involve genes that are known to play important roles in anti-viral defense mechanisms. Notable among these are IL28RA, IFIH1, DDX58 [
3], and RNF114 [
4]. IL28RA codes for the alpha subunit of IL-28 receptor and forms a complex with the IL-10 receptor, IL10RB. This receptor complex interacts with three closely related virus-induced cytokines, IL28A, IL28B and IL29 and plays role in antiviral defense [
5]. IFIH1 has been known to play an important role in sensing viral nucleic acids and in activation of anti-viral immune responses. DDX58 encodes the RIG-I innate antiviral receptor, which recognizes cytosolic double stranded RNA. The exact function of RNF114 is not known but its paralog RNF125 is involved in ubiquitination of the innate anti-viral receptors, RIG-I and MDA5 [
6]. Additionally, certain HLA alleles associated with psoriasis such as HLA-B*57 and HLA-B*27, have been associated with robust viral control of HIV-1 [
7],[
8].
Physiological studies on psoriatic skin have also indicated a role of anti-viral restriction factors or anti-viral proteins (AVP) in psoriasis. A previous study has found that AVPs such as MX1, BST2, ISG15 and OAS2 are strongly elevated in the skin of psoriatic patients in comparison to healthy controls [
9]. The authors further observed that IL29 might be responsible for the antiviral milieu in psoriasis as its expression correlated with AVP levels. Recently, we have performed whole transcriptome analysis of psoriasis skin and found that antiviral restriction factors are strongly upregulated in psoriatic skin and not in atopic dermatitis skin (manuscript in preparation). Altogether, these results suggest that psoriasis patients might have a strong cutaneous anti-viral immunity. However, the inciting factors and consequences of this anti-viral immunity are not known.
Possible triggers for psoriasis have been attributed to drugs [
10] or pathogens such as bacteria and possibly virus [
11]. Human endogenous retroviruses (HERVs) might play a role in triggering these anti-viral immune responses, since the role of HERVs in the pathogenesis of autoimmune diseases has generated considerable interest [
12]. HERVs exist as proviruses in the human genome and consist of 5-10 kb of sequence encoding three genes,
Gag, Pol and
Env, flanked on both sides by long terminal repeats (LTR), which are 300-1200 nucleotides in length. They are estimated to have integrated into human genome starting 30-40 million years ago and as recently as 150,000 years ago [
13]. Most HERVs have undergone significant mutational changes and are not thought to encode infectious virus, although there may be exceptions [
14]. There are many families of HERVs but the most recent and widespread entrants into the human genome belong to the HERV-K family of endogenous retrovirus (HML-2, human MMTV like family), which are believed to have integrated into the human genome 200,000 to 5 million years ago [
15],[
16]. Studies also report that it as one of the most transcriptionally active HERV in human genome [
17],[
18].
Expression of HERV-K is known to be up-regulated in several diseases, including breast cancer, ovarian cancer, during HIV infection, and rheumatoid arthritis (RA), where an increase in viral mRNA and viral load of HERV-K was observed in active disease [
19]-[
23]. However, functional consequences of this expression are not known. It has been proposed that HERVs can trigger immune responses directly by acting as super-antigens, by encoding auto-antigens or by mimicking the self-proteins. Indeed, antibodies against HERVs are often observed in the sera of patients, resulting from an increase of transcriptional and translational activity [
24],[
25]. Alternatively, HERVs could indirectly affect immune responses by influencing expression of genes, regulating immune responses, or facilitating tolerance.
The first evidence of role of HERVs in psoriasis came from the observation of retrovirus like particles from the skin of psoriasis patients [
26]. Bessis
et al.[
27] observed that most of psoriasis patients showed positive immunofluorescence staining for HERV-E transmembrane envelope glycoprotein while only 15% of normal skin samples were positive. Furthermore, using a pan-retroviral detection system, it was observed that endogenous retroviral sequences for HERV-K, HERV-E and ERV-9 are expressed both in psoriatic and in normal skin [
28]. These authors then used specific primers for ERV-9 and saw a significant increased expression in lesional skin compared to controls, but did not report quantitative measurements on HERV-K or HERV-E. However, the biological significance and the level of expression of HERV in psoriasis are not entirely known.
In the present study, we examined the humoral immune response against proteins coded by HERV-K Gag and Env gene in psoriasis patients and controls. We further used a sensitive approach, quantitative reverse transcription PCR (RT-qPCR), to measure the expression level of a comprehensive panel of HERV-K sequences (gag, env, pol and rec) in lesional and non-lesional skin from psoriatic patients compared to normal skin from healthy controls.
Discussion
Previous studies on HERVs indicate an association with several autoimmune diseases, cancers and even several infectious diseases. Some studies have also reported an association of HERVs and psoriasis [
27],[
28]. However, these studies did not quantify the level of expression of HERV-K, which is known to have entered into human genome very recently and which is also believed to most functionally active of all endogenous retroviruses. To understand the association of HERV-K expression with psoriasis, we first examined the humoral immune response against HERV-K recombinant proteins in sera of psoriasis patients and controls. We observed that psoriasis patients showed a significant decrease in IgM responses against HERV-K Env-recTM, Env-recSU, and Gag-recCA. Coherent with this observation, the IgG response against those proteins were either not modified or decreased in psoriasis patients.
When we measured the responses against the more immunogenic linear peptide epitopes of HERV-K (HML-2) Gag (see Methods), we detected an increase in antibody titer against a single linear epitope belonging to the nucleocapsid (Gag 137). This epitope is also the most immunogenic in healthy donors. Although we detected a decreased antibody response against HERV-K recombinant proteins and an increased antibody response against Gag 137, discordant antibody responses against HERV-K (HML-2) have previously been reported. For instance, although being derived from a single transcript, SU and TM [
37],[
38] show discordant transcriptional regulation in HIV-1 patients resulting in different specific humoral responses [
24]. We can thus speculate that, as observed during HIV-1 infection, HERV-K (HML-2) transcriptional activity of different transcripts might be differentially modified during psoriasis. Alternatively, high specificity of B-cell receptor or MHC Class II binding to a particular epitope may lead to heightened antibody responses. Of note, we have previously shown that psoriasis patients have stronger antibody responses against HERV-K dUTPase recombinant protein compared to healthy controls [
39]. It is noteworthy that only peptide 137 elicited responses in either controls or patients, and this 15mer has a 12/15 amino acid identity with the amino acid positions 74–88 of the predicted CaO19.10692 protein from
Candida albicans, which is known to be more frequent in psoriasis patients [
40] and could potentially cause cross reaction.
We further analyzed the expression of various HERV-families in a pooled set of cDNAs derived from 11 lesional and 4 non-lesional skin samples. However, we only observed expression of HERV-K genes (Env, Gag, Pol and Rec) and ERV-9. We then compared the expression level of HERV-K (Env, Gag, Pol and Rec) and ERV-9 genes in a larger set of patients, 14 pairs of lesional and non-lesional skin biopsies from psoriasis patients and 27 biopsies from healthy controls. In accordance with our ELISA findings, we observed a decreased expression of all HERV-K genes tested as well as ERV-9 in comparison to healthy controls.
Regarding the relationship of our ELISA results and our RT-qPCR results, there was a partial overlap between mRNA sequences examined by RT-qPCR and the proteins used for antibody testing. Our primer/probe set for the Env gene amplifies a part of SU gene, thus the decrease in Env expression seen by RT-qPCR directly correlates to the decrease in IgG and IgM responses against SU by ELISA. In contrast, our primer/probes for Gag amplify part of NC and thus do not directly correlate with the decreased antibody response seen here in case of CA. These results might be verified by either using primer/probes that amplify CA or using recombinant NC for analysis of antibody responses.
Taken together, our results demonstrate that decreased expression of HERV-K Gag and Env in psoriasis patients correlate with a decreased antibody response against these proteins. Although the decreased expression levels are not necessarily causal for the decreased antibody responses, the relationship is suggestive. Interestingly, a similar decrease in ERV expression was observed in a recent study on systemic lupus erythematosus (SLE) using RNA sequencing to characterize SLE transcriptome [
41]. Thus, certain inflammatory diseases may be associated with suppression of HERV transcription.
There are several possible mechanisms to explain the observed decreased expression of HERV-K in psoriasis. It has been observed that global control of HERV expression can occur by heritable changes in gene expression without any change in underlying DNA sequences. These include alternation of DNA methylation or histone modification [
42]. Past studies have reported that ERV transcription can be controlled by the methylation state of genomic DNA [
43]. Furthermore, it has been shown previously that treatment of cells with agents promoting demethylation of genomic DNA could result in ERV induction [
44]. All these findings have suggested that methylation of genomic DNA could be a way to control and regulate HERVs. Methylation studies done on psoriasis skin samples have observed differentially methylated regions (DMRs) covering a large part of the genome [
45],[
46]. Furthermore, Zhang
et al.[
45] observed that the number of hypermethylated DMRs was considerably higher than that of hypomethylated DMRs in lesional samples form psoriasis patients. Whether these hypermethylated sites correspond to the genomic locations of HERVs would be interesting to determine.
Another mechanism that might affect HERV expression is RNA degradation of HERVs at the level of post-transcription. In fact, recent studies done in this regard indicate that control of HERVs can occur both at the level of transcript repression by methylation and RNA degradation at post-transcription and these two mechanisms can be interrelated. A third possibility is RNA interference. It has been speculated that dsRNA derived from retrotransposons and other retroelements may induce both transcript degradation and DNA methylation by using RNA interference (RNAi) pathways [
43] and this can be driven by small interfering (siRNA) and Piwi-interacting RNA. Furthermore, these small RNAs can also help in targeting of repeats for DNA methylation and other chromatin modifications mechanisms.
Our present results contrast the results from Bessis
et al., and Moles
et al., [
27],[
28] in that they found increased HERV expression and we have found decreased HERV expression. However, Bessis
et al. only focused on HERV-E, and as mentioned earlier not all HERVs will be expressed in the same manner and this may account for a difference between HERV-E and HERV-K expression. Moles
et al. did examine HERV-K sequences, but only measured ERV-9. Even so, we did find ERV-9 with a lower expression level in psoriasis patients, whereas they found a higher level of expression. More studies are needed to explain this discrepancy, but our ERV-9 data is consistent with our HERV-K results.
Conclusions
In summary, we have shown that psoriasis patients show a decrease in antibody response to HERV-K proteins as compared to healthy controls. Furthermore, we also observed a decrease in expression of HERV-K genes Env, Gag, Pol and Rec and ERV-9 in psoriasis skin as compared to healthy skin. The reasons for the unexpected, low levels of HERV expression in psoriatic lesions are unclear and might be in part a consequence of anti-viral defense mechanisms. Whether these are triggered by HERV expression, expression of other antigens, and/or antigen independent mechanisms pose interesting questions for future study. Suppression of HERV-K expression across the genome may also indicate the possibility that the inflammatory state of psoriasis is associated with epigenetic changes leading to the observed decrease. However, further studies are warranted to explore this hypothesis.
Study limitations
A limitation of this study is the moderate sample size examined. Some degree of variability was observed in mRNA expression across different genes. Independent replication of our findings in other cohorts is warranted.
Competing interests
The authors declare that they have no competing interests.