Over the last three decades there has been an interest in human endogenous retroviruses (HERVs) as potential contributors towards autoimmune disease, made all the more important given that the incidence of autoimmunity is increasing worldwide [1]. Endogenous retroviruses (ERVs) and other retroviral elements have been found in all vertebrates investigated [2]. Much work has been done to elucidate the numerous retroviral families that exist within the human genome, of which ERVs make up 8% [3]. HERVs are believed to be pathogenic in several autoimmune diseases, but especially the rheumatic diseases such as rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE) [4, 5]. The growing body of evidence suggests RA and SLE are complex diseases which require genetic susceptibility, environmental factors and opportune circumstances operating together to induce disease [6]. Previous studies have explored the possibility of induction of SLE by Epstein - Barr virus (EBV) and HERVs [7-9].

The aim of this review is to summarise the existing evidence for the role of HERVs in autoimmune rheumatic disease, focussing specifically on RA and SLE. This paper describes several hypotheses on the mechanisms by which HERVs are thought to be pathogenic. This is important as the link between HERVs and rheumatic disease could lead to the discovery of novel therapeutic agents – and even cures - for these chronic diseases.

Retroviruses are small viruses which replicate by reversing the normal flow of genetic information from DNA to RNA, known as reverse transcription. Exogenous retroviruses such as human immunodeficiency virus (HIV) and human T cell leukaemia virus (HTLV) can reproduce viral RNA from their pro-viral DNA, and hence retain infectivity, a feature lost by HERVs. HERVs were first integrated into the human genome 30–40 million years ago and are believed to be the key molecular link between the host genome and exogenous viral particles [10, 11]. Over millions of years, HERVs (coined “fossil viruses”) have become trapped within the human genome. They are transmitted genetically in a Mendelian fashion and are found within the DNA of all cells. HERVs have similar gene structures to exogenous retroviruses, and are composed of gag, pol, and env regions sandwiched between two long terminal repeats (LTRs) (see Fig. 1) [12].

Fig. (1). Structure of human endogenous retroviruses (HERVs). HERV products may be generated using different open reading frames (ORFs)

Fig. (2). Molecular models of HERV-K10/IgG1Fc epitopes. Epitopes A and B taken from Westwood [18] and modeled using PyMOL (Molecular Graphics System, Version 1.2r3pre, Schrodinger, LLC: http://www.pymol.org/pymol) and enhanced using UCSF Chimera (http://www.cgl.ucsf.edu/chimera). Molecular models were displayed in ‘mesh’ mode. Arrows point to amino acid residues highlighted and displayed on an IgGFc molecule (PDB file: 3AVE) using UCSF Chimera. CH: constant heavy 2 and 3 domains of IgG. In B: despite the amino acid substitution, there appears to be little change in molecular shape which may facilitate antibody cross-reactivity.

Fig. (3). Schematic interactions with human endogenous retroviruses. The interaction of HERVs on the immune system may be complex, compounded by external and internal factors to augment detrimental effects at the molecular and protein level.

LTRs are sequences of DNA that repeat hundreds or thousands of times and are used by viruses to insert their genetic sequences into the host genomes. Gag is a polyprotein and is an acronym for Group Antigens; pol is the reverse transcriptase; and env is the envelope protein that determines viral tropism. Retroviral transcription is regulated by promoter and enhancer regions in the LTR present at both ends of the retroviral genome.

Over 26 families of HERVs have been identified [3] and, although most are defective through mutation and deletion signals, a small number have the ability to produce viral products [12]. The HERV-K family is the only family of HERVs that is able to produce intact viral particles. It is one of the most transcriptionally active families as its members retain intact open reading frames (ORFs) that encode viral particles [13]. HERVs have been broadly classified into three categories, based on their genetic similarity in the pol region (see Table 1).

Class I contains HERVs related to Gammaretroviruses; class II viruses are related to Betaretroviruses; while Class III viruses are related to Spumaviruses [5]. Confusingly, some HERVs are known by several names, for example, HERV-K10 is also known as HML-2. A complete database of HERVs can be found online (http://herv.img.cas.cz/). It should also be noted that mobile genetic elements that behave as HERVs, such as LINE-1, have been implicated in RA [14]. LINE-1 is a DNA sequence that can change its relative position within the genome of a single cell, leading to phenotypically significant mutations. However, LINE-1 is not thought to be a true HERV by some authors as it has no LTR.

Bioinformatics is a rapidly-growing interdisciplinary field which harnesses computer science, mathematics, physics and biology to capture and interpret biological data [15]. Bioinformatic computer programmes can provide information about HERV protein sequences that are structurally similar (have sequence homology) to known antigenic epitopes (epitope recognition) [16]. For example, bioinformatics has been used to demonstrate that HERV-K10 shares sequence homology with epitopes of rheumatoid factor (RF) [17] (Fig. 2 and Table 2), implicating molecular mimicry (see section on “Molecular mimicry”). Short peptides that reflected these epitopes were then synthesized and an enzyme-linked immunosorbent assay (ELISA) was used to test their reactivity to patient serum. This demonstrated significant up-regulation of HERV-K mRNA levels in RA patients compared to inflammatory and healthy controls. Other molecular techniques can be used to harness HERV proteins, including: immunoblot, ELISA, nucleic acid sequence-based amplification (NASBA), polymerase chain reaction (PCR), multiplex PCR and quantitative reverse transcription–polymerase chain reaction (RT-PCR) (see Table 3). It is beyond the scope of this article to look at these in detail but it is prudent to point out that quantitative RT-PCR provides a highly sensitive technique to detect even a very low amount of RNA, found in all HERVs. Furthermore, antigen microarray profiling in RA can provide diagnostic information and allow stratification of patients with early RA into disease subsets [20].

Autoimmune diseases are complex diseases, whereby genetic background confers susceptibility to disease onset, but is neither sufficient nor causative for disease development [21]. The role of HERVs in disease pathogenesis is not confined to rheumatic disease, for example, it is known that expression of HERV-Fc1 is increased in patients with active multiple sclerosis [22] while in psoriatic lesions, HERV sequences of the W, K and E families are expressed and a new variant of the ERV-9/HERV-W family has been characterized [23].

The importance of genetics in the pathogenesis of autoimmune disease is supported by their clustering in families [24]. In RA and SLE a concordance rate in monozygotic twins of only 15% [25] and 25% [26] respectively is seen, emphasizing the importance of environmental factors upon the susceptibility to autoimmune disease. Genetic susceptibility might explain why only a subgroup of individuals will develop autoimmunity after infections [24]. Moreover, genome-wide analyses in lupus-prone mice have shown that major histocompatibility complex (MHC)-linked and multiple non-MHC-linked genetic factors contribute to the overall susceptibility and progression of SLE [27].

Retroviruses have been repeatedly discussed as etiological factors of autoimmune rheumatic diseases and antibodies to gag and env regions of retroviruses, including HERVs, have been reported in patients with autoimmune disease [28-31], but the role of retroviruses in these diseases remains unclear. It is known that autoimmune rheumatic disease is more prevalent in countries where retroviral infection is endemic, for example, HTLV in Japan and Central America [32, 33]. Furthermore, the observation that other types of chronic arthritis occur with an increased prevalence in HIV-1-infected individuals also lends weight to this assumption [34]. In a pivotal study of 23 patients it was demonstrated that 45% of synovial tissues obtained from patients with RA were positive for HTLV-1 p19 and p24 antigens but, intriguingly, no antibodies to HTLV-1 were detected in the sera [35]. In addition, another study showed 55% of patients with RA had p17 and p24 HIV antigens in synovial tissues but, yet again, their sera tested negative for antibodies to HIV [36]. It was therefore likely that an alternative retrovirus, with sequence homology to HTLV-1 (in the former) or HIV-1 (in the latter), was the cause of the immune response. Other studies revealed that SLE patients make antibodies to p24 gag of HIV-1 and that anti-Sm antibody can cross-react with p24 gag [37, 38]. Moreover, HRES-1 (Human T-cell lymphotropic virus-related Endogenous Sequence) is a HERV sequence capable of protein expression, and is expressed in higher quantities in patients with varying autoimmune disease than healthy controls [39, 40]. Thus, HERVs have been implicated as contributors to autoimmune rheumatic disease. Several studies have since demonstrated evidence of increased HERV presence in autoimmune rheumatic disease, summarized in Table 3.

Tolerance is the process that eliminates or neutralises self-reactive immune cells [50]. Autoimmunity is caused by loss of tolerance, leading to tissue damage and destruction [51]. Tolerance is achieved in the thymus during the ontogeny of the immune system when cells that react against self are deleted (central tolerance) [52, 53]. Some potentially self-reactive cells survive these measures and are subsequently held in abeyance (peripheral tolerance) [54]. HERV-encoded proteins should be considered as self-antigens and be tolerated by the immune system but may trigger the breakdown of tolerance.

The interaction of HERVs on the immune system is likely to be complex, particularly if associated with an autoimmune disease (Fig. 3). For the endogenous virus, mechanisms such as molecular mimicry, the role of superantigen motifs and modulation via elements within LTRs have been proposed that primarily focus on branches of the adaptive immune system (B cells plus antibodies and T cells). Undoubtedly, moieties such as RNA/DNA may also contribute via toll-like receptors (TLR’s) of the innate immune system. Added to these processes, is the premise of stimulating potentially auto-reactive immune cells which have been held in abeyance through peripheral tolerance. Could this be dependent on HLA type? The breakdown of tolerance is of course a key element in autoimmunity since a number of host components may be compromised through molecular mimicry, physiochemical damage and exposure of cryptic epitopes/previously sequestered epitopes. In RA for example, IgGFc (present in all individuals) becomes a major target for autoantibodies in patients classified ‘RF-positive’, yet remains innocuous for those deemed ‘RF- negative. Other agents such as helper viruses (e.g. EBV, cytomegalovirus (CMV)) or hormones/cytokines can also facilitate the stimulation of HERVs. Presently the role of epigenetics further compounds the situation since environmental agents may surreptitiously modulate gene expression. Consequently, the interaction of external and internal factors on HERVs at the molecular level and the generation of protein products in disrupting the immune system to effect change, damage and clinical consequence will require further research.

Epigenetics refers to heritable changes in genes, regulating gene expression without changing the DNA sequence [10]. DNA methylation is an essential process for normal development and cellular differentiation. It consists of the addition of a methyl group to cytosine residues and helps to stabilize chromatin in an inactive configuration, thus inhibiting gene transcription [26, 83]. Hypomethylation can cause increased expression of given genes that are normally silenced [84]. It is thought that hypomethylation may alter synovial fibroblasts, which triggers RA development [85-86]. Altered methylation has been explored in B cells [87] and T cells [88] while changes in the pattern of DNA methylation have been found to associate with twin discordance in SLE [89]. Murine models have clearly demonstrated reduced methylation levels in the thymus of lupus-prone mice compared to non-prone mice [90]. Several other human studies have corroborated the importance of DNA hypomethylation in SLE etiology [91-93]. Expression of HERV clone 4-1 (HERV-E family) is significantly increased in SLE patients compared with healthy controls, while serum autoantibodies to this HERV and expression of its antigens on lymphocytes are detected in SLE patients, but not in healthy subjects [83, 94-96]. Furthermore, certain drugs are known to be inhibitors of DNA methylation, such as procainamide, and when demethylating agents have been given, HERV 4-1 expression is markedly increased in healthy controls, but not in SLE patients [93], which suggests involvement of demethylation of normally methylated sequences in SLE. As yet, there is no concrete link between HERVS and hypomethylation but the possibility exists. No doubt, the expanding area of epigenetics will provide the evidence required.

This paper has reviewed the evidence for HERVs as contributors to autoimmune rheumatic disease. Progress in accrual of evidence has been slow but major advances in serological, molecular and viral load have occurred. Suggested mechanisms for disease pathogenesis include molecular mimicry, immune dysregulation and superantigen motifs. The evidence appears to be mounting for HERV-K10 in RA pathogenesis and for HRES-1 and ERV-3 in SLE, but there is a clear need for animal models to corroborate this, and a call for large multicentre studies to examine the percentage of patients that exhibit high reactivity to these viruses. This may reveal other differences between populations and different races. Toll-like receptors and their possible interaction with viral material is also an emerging area of importance in both SLE and RA, which introduces the possibility that the innate system may be involved in pathogenesis. Bioinformatic and molecular modeling studies to investigate potential molecular mimicry would similarly be useful.

Given the debilitating nature and chronicity of these conditions, prevention of autoimmune rheumatic disease is the ‘holy grail’. Just as vaccination against Human Papilloma Virus is routinely given in some countries to protect females from development of cervical carcinoma, perhaps one day a similar programme may exist for those patients identified to be at high risk of developing RA (eg. individuals who are strongly RF+ and HERV-K10+). Moreover, HERV titers may one day be used as markers of disease activity or as diagnostic markers, just as RF is used today. Looking ahead, there may be a need for larger studies of antiretroviral treatment in patients with autoimmune rheumatic disease that investigate whether eliminating the viral proteins leads to disease remission. Clearly, there are numerous questions to be answered and we remain some way off this stage. National and/or international collaboration would add to the depth of quality research and help to unlock the answers within this fascinating field.