Human endogenous retroviruses (HERVs) are ancient retroviral DNA sequences established into germline. They contain regulatory elements and encoded proteins few of which may provide benefits to hosts when co-opted as cellular genes. Their tight regulation is mainly achieved by epigenetic mechanisms, which can be altered by environmental factors, e.g., viral infections, leading to HERV activation. This review summarizes the recent advances on the epigenetic mechanisms controlling HERV expression and the pathogenic effects triggered by HERV de-repression leading to neurological diseases, inflammatory processes and neurodegeneration.
HERVs are remnants of ancient exogenous viral infections that established the Vertebrates germline millions of years ago
. They classify into different families named after the amino acid one-letter code of the aminoacyl-tRNA used to prime their retrotranscription
[5]
. HERVs typically encompass three main proviral genes:
pol
,
gag
and
env,
flanked by two LTR sequences. The
gag
gene encodes structural proteins; the
pol
gene, a reverse transcriptase, a protease, and an integrase; and the
env
gene, an envelope protein. Since their introduction into the germline, some HERVs have accumulated mutations in their proviral genes, making them replication-defective because of frameshifts, the appearance of early stop codons or deletions/insertions of sequences. The homologous recombination between both LTR sequences has also led to a complete loss of proviral genes in some cases, leaving the HERV represented by its LTR only
. Through these mechanisms, evolution has left copies of TEs with variable genomic contents in the human population.
HERVs have occasionally provided their hosts with beneficial effects. They contribute to genomic variability, provide additional gene regulatory elements, such as alternative promoters or alternative splice sites
, or rule important physiological processes. HERVs also participate in transcriptional regulation with noncoding RNAs
or in the expression of “domesticated” viral proteins
. Nonetheless, if uncontrolled and/or abnormally activated, HERVs can lead to detrimental effects in their host organisms by altering gene expression profiles, expressing pathogenic nucleic acids/proteins or even inducing deleterious mutations
[23]
[24]
[30]
[31]
[34]
[21]
[35]
[36]
[37]
[38]
[29]
[39]
[40]
[41]
[37]
[48]
[49]
Figure 1.
1
[30]
[31]
[41]
2
[36]
3
4
[44]
[50]
[51]
[58]
[59]
[33]
N
[62]
[63]
[64]
Despite their tightly controlled regulation, pathogenic HERV proteins have been detected in chronic inflammatory or degenerative diseases of the nervous system, such as multiple sclerosis (MS)
[67]
[71]
env
[15]
[84]
[85]
Herpesviridae
[66]
[98]
[99][100]. It is tempting to think that HERV activation may represent the common denominator to those risk factors, with particular combinations leading to the reactivation of specific HERV elements. In this hypothetical scenario, depending on the HERV(s) element(s) de-repressed and the cell type(s) involved, the resulting symptoms and, thereby, the disease triggered would be one or another.
One of the most studied pathologies linked to abnormal HERV expression is MS, an autoimmune neurodegenerative disease that affects myelin sheaths, leading to neuronal loss [65][101]. The commonly used experimental model for MS is the mouse autoimmune model EAE (Experimental Autoimmune Encephalomyelitis) [102], from which most of the knowledge about neuroinflammation and neurodegeneration derives. EAE is induced by immunization with myelin or central nervous system (CNS)-derived peptides, and the disease is mediated by activated CNS-specific CD4+ T cells in the periphery, which enter the CNS by crossing the endothelial blood–brain barrier (BBB). By contrast, neuroinflammation in MS has now been revealed to be initiated in the CNS with the activation of microglia [103][104]. Under physiological conditions, microglia protect the brain by sensing pathogens, phagocytosing cell debris and promoting remyelination. However, under MS pathological conditions, microglia become activated and secrete proinflammatory molecules that directly damage myelin sheaths and oligodendrocytes, causing inflammation and neurodegeneration [105][106] (Figure 2, number 1). This proinflammatory environment attracts peripheral monocyte-derived macrophages and dendritic cells (DC), as well as peripheral autoreactive T-lymphocytes, which enter the CNS through the BBB localized in CNS microvessels [103][107][108][109] (Figure 2, number 2). Infiltrating T-lymphocytes become activated by microglia, which also act as antigen-presenting cells (APCs), phagocytosing myelin debris and expressing major histocompatibility complexes (MHC) I and II with costimulatory molecules (Figure 2, number 3). Infiltrated monocyte-derived DCs also act as APCs, presenting myelin antigen to T cells, along with microglia [110] (Figure 2, number 3). Activated T-lymphocytes differentiate into T-helper cell-type 1 (Th1) (producing IFN-γ) and Th17 cells (producing IL-17, IL-22 and IL-21) [103][105][106][111][112] (Figure 2, number 3). The crosstalk between neuroimmune and peripheral immune systems exacerbates neuroinflammation, further activating microglia and astrocytes through cytokine production, as well as facilitating the infiltration of other immune cells, such as activated B cells, monocytes and macrophages. The release of proinflammatory cytokines, nitric oxide, reactive oxygen and nitrogen species [113], together with some other processes like phagocytosis of myelin by macrophages and antibody-dependent cytotoxicity, contribute to inflammation within the CNS, oligodendrocyte loss, and demyelination of the white and gray matter [114].
Figure 2. pHERV-W-ENV drives inflammation and neurodegeneration in MS. In MS, inflammation starts in the CNS with the activation of microglia, which release proinflammatory cytokines (PICs) that directly damage the myelin sheaths (1). Cell debris and PICs attract peripheral immune cells to the CNS (2); among them are T cells, which are then presented by APCs like microglia or infiltrating monocyte-derived dendritic cells (3) [103]. In this sense, pHERV-W ENV stimulates the innate immune system by engaging TLR4/CD14 receptors in microglia (1) monocytes (4), dendritic cells (5) and macrophages (6), inducing the release of PICs. In addition, it stimulates specific T-cell activation (3) and abnormal polyclonal T-cell activation (7) [19][18], acting as a superantigen. Furthermore, pHERV-W ENV activates vascular endothelial cells (8), making the migration of circulating cells possible at the level of ongoing brain lesions initiated by perivascular macrophages and/or microglia [115]. On the other hand, pHERV-WENV impairs the remyelination process by inducing the release of PICs and nitrosative stress in oligodendrocyte precursor cells (OPCs) (9) and the microglia (1), inhibiting the differentiation of OPCs to oligodendrocytes and damaging the myelin sheath, ultimately leading to neuronal loss [116]. pHERV-W ENV also drives microglial cells to physically interact with myelinated axons and induces the leakage of intra-axonal and myelin proteins (10) [117].
Although neurodegeneration has traditionally been thought to be a consequence of inflammation in the CNS, this concept has been challenged by the poor efficacy of immunomodulatory treatments on preventing demyelination and axonal degeneration [115]. Now, we know that pHERV-W ENV may participate in the development of MS by triggering a potent immune response, as well as by impairing the myelin-repairing process [18][19][115][117][118][119][120], which could explain why immunomodulatory treatments failed at preventing neurodegeneration. Furthermore, pHERV-W ENV contribution to MS is supported by the observation that administered pHERV-W ENV with myelin oligodendrocyte glycoprotein (MOG) 35–55 peptide triggers EAE in mice [119].
pHERV-W ENV RNA and protein have been detected in PBMCs of MS patients [74][76][81][121], showing their highest RNA levels in natural killer (NK) cells, followed by B cells and monocytes, the last increasing their expression by differentiation to macrophages. By contrast, no ENV-encoding RNA or protein is detected in T cells [66][122]. pHERV-W ENV has also been detected in astrocytes [66][75], infiltrated macrophages and activated microglia in the brains [123][121][124] of MS patients, and its soluble form in their serum, plasma, and CSF [65][76][125][126][127][128]. Indeed, the presence of HERV-W RNA in CSF is considered as a negative prognostic marker of MS. Its load increases with MS duration and parallels the clinical stages [126][127][128][129][130].
The SU domain of the pHERV-W ENV, either associated to viral particles or in its soluble form has been shown to trigger proinflammatory responses in vitro, in human PBMCs and dendritic cells (DCs) [19] and, in vivo, in humanized severe combined immunodeficiency (SCID) mice, which die from brain hemorrhages with an overexpression of TNF-α after pHERVW-ENV injection [18][131]. The inflammatory reaction induced by pHERV-W ENV is mediated by the activation of Toll-like receptor (TLR) 4 and its coreceptor CD14 [19][119]. TLRs are expressed by different cell types, such as B cells [132], macrophages, monocytes, DCs [133], oligodendrocytes [134], astrocytes [135] or microglia [136], among other, where they mediate part of the innate immune response against invading pathogens [137][138].
On the one hand, pHERV-W ENV activates monocytes, DCs and macrophages (Figure 2, number 4–6). Activated monocytes produce proinflammatory cytokines, including TNFα, IL-1β, IL-6 and IL-12p40 (Figure 2, number 4). Interestingly, the levels of IL-6 and IL-12p40 induced in cultured PBMCs directly correlate with the MS disease severity [19]. Activated DCs not only release the proinflammatory cytokines IL-6, TNFα, IL-12p40 and IL-12p70 but, also, upregulate antigen presentation, activating antigen-specific T cells. The release of IL-12 by DCs activates naïve T cells and promotes their differentiation into IFN γ-secreting Th1 cells [19] (Figure 2, number 3). In turn, released INF-γ and TNF-α upregulate pHERV-W ENV secretion by PBMCs [81]. Thus, pHERV-W ENV not only activates the innate immune system but, also, the adaptative immune system. Moreover, it also acts like a superantigen [18], stimulating abnormal polyclonal T-cell activation through its binding to Vβ chains, and the expansion of T-cell-receptor β-chain (TCR Vβ) cells, leading to nonspecific oligoclonal Vβ+T cell activation [18][139] (Figure 2, number 7).
In addition, pHERV-W ENV activates vascular endothelial cells within the BBB, leading to the release of proinflammatory cytokines IL-6 and IL-8 and to the overexpression of ICAM-1, thus enhancing the migration of circulating cells towards the brain (Figure 2, number 8) [115]. In the brain, the pHERV-W ENV-SU domain triggers CNS innate immunity through TLR4 binding, activating microglia and perivascular macrophages, which release proinflammatory cytokines, leading to neuroinflammation and neurodegeneration [140] (Figure 2, number 1). Additionally, pHERV-W ENV produced by the surrounding cells like microglia or macrophages [74] engages with, and activates, TLR4 receptors present on the brain’s resident oligodendrocyte precursor cells (OPCs). The OPCs’ role is to migrate to the lesion sites where oligodendrocytes are depleted, and differentiate into mature oligodendrocytes to remyelinate axons after an injury [141]. After migration, OPCs transiently express TLR4 receptors, which become downregulated during the OPCs’ differentiation process to oligodendrocytes [116]. TLR4 activation does not affect OPC survival; however, it induces the release of proinflammatory cytokines such as TNF-α, IL-1β and IL-6, as well as the overexpression of inducible nitric oxide synthase (iNOS), leading to an increase in nitric oxide (NO) levels that triggers nitrosative stress, which, in turn, affects the myelin protein expression and reduces OPCs’ differentiation capacity [116] (Figure 2, number 9). NO and TNF-α are well-established mediators of axonal injury and demyelination [142][143]. Interestingly, TLR4 activation in OPCs seems to be sensitive to the cyclin-dependent kinase inhibitor p57kip2, which participates in controlling TLR4’s surface expression [118].
On the other hand, pHERV-W ENV induces a degenerative phenotype in microglial cells characterized by the expression of the same set of proinflammatory cytokines detected on pHERV-W ENV-activated OPCs (TNF-α, IL-1β and IL-6) and by the overexpression of iNOS [117] (Figure 2, number 1). While reducing anti-inflammatory and neuroprotective proteins and downregulating the expression of genes involved in myelin debris phagocytosis, it affects neurorepair and inhibits OPC differentiation. Furthermore, the pHERV-W ENV protein stimulates microglia proliferation and drives physical interaction between microglial cells and myelinated axons to induce leakage of intra-axonal and myelin proteins [117] (Figure 2, number 10). Of note, the multifaceted effects of pHERV-W ENV are often due to its ability to strongly bind TLR4 receptors on expressing cells, thereby activating underlying signaling pathways and leading to various effects with a proinflammatory common point in different cells [144].
Furthermore, pHERV-W-ENV inflammatory and demyelinating consequences can also be observed in chronic inflammatory demyelinating polyradiculoneuropathy (CIDP), a rare immune disease of the peripheral nervous system (PNS) [120].In these patients, pHERV-W ENV is found in PBMCs and peripheral nerve lesions, especially in Human Schwann Cells (HSCs), in which it exerts pathogenic effects by TLR4 engagement, inducing the release of proinflammatory IL-6 and the macrophage and T-cell chemoattractant CXCL10 [145], triggering activation of the innate immune system.
Another HERV-W-related disease is schizophrenia [146][147], a psychiatric disorder that highly impacts patients quality of life. In reference to the mechanisms underlying this disease, several are the hypotheses raised, including the disturbance of the dopaminergic neurotransmission or its relationship with neurodevelopmental problems. However, the idea that autoimmune processes within the nervous system could be playing an important role in its pathophysiology has been gaining more and more acceptance [148]. Some studies have shown the elevated expression of TLRs [149] and abnormal expression of proinflammatory cytokines, like IL-6, IL-8 and the C-reactive protein (CRP), in the serum of these patients [150][151]. CRP is released to blood under inflammatory conditions, and its overexpression supports that immune mechanisms play a role in schizophrenia symptoms beyond cognitive decline [152]. Interestingly, the levels of HERV-W mRNA were found to associate with a proinflammatory phenotype in these patients [153].
HERV-W expression is particularly associated with recent-onset schizophrenia [68][146][154][155]. Both HERV-W/MSRV and Syncytin-1 were reported to be increased in schizophrenia [84][156], and HERV-W/MSRV was shown to be unequally represented in the genome of an affected and its nonaffected monozygotic twin [157]. HERV-W GAG and ENV RNA or protein have been found in the circulating blood of these patients [147][154][155][156][158], correlating GAG protein levels with disease severity [158] and Syncytin-1 or (MSRV) pHERV-W ENV with CRP levels [155], providing a link between HERV-W envelopes and the status of systemic inflammation in these patients [84][147][155]. Activation of CRP in microglia and astrocytes, and the consequent inflammation, can be triggered by Syncytin-1 through direct interaction with TLR3 receptors [84] and IL-6 release. Furthermore, HERV-W presence has also been detected in CSF [68] and in cortical tissue [159], indicating a pathogenic role in the brain. In fact, HERV-W/MSRV induces activation of the Brain-derived neurotrophic factor (BDNF) and Dopamine receptor D3 (DRD3) genes [156], both implicated in providing a higher risk for suffering schizophrenia [160]. Thus, HERV-W Syncytin-1 and MSRV may determine and/or promote the development of schizophrenia. In this regard, it should be mentioned that, by combining single-molecule tracking, calcium imaging and behavioral approaches, Johansson et al. [70] showed that the mechanism used by HERV-W-ENV to alter glutamate synapse maturation and generate behavioral deficits depends on the N-methyl-d-aspartate receptor (NMDAR) organization and cytokine-dependent changes, further supporting an etiological role for this HERV product in the development of psychosis.