OUP user menu

Extensive viral mimicry of 22 AIDS-related autoantigens by HIV-1 proteins and pathway analysis of 561 viral/human homologues suggest an initial treatable autoimmune component of AIDS

Chris J. Carter
DOI: http://dx.doi.org/10.1111/j.1574-695X.2011.00848.x 254-268 First published online: 1 November 2011


HIV-1 viral proteins, particularly the env protein, are homologous to 22 AIDS autoantigens, suggesting their creation by antiviral antibodies subsequently targeting human homologues. They include antibodies to T-cell receptors, CD4 and CD95, complement components, IgG, TNF and other immune-related proteins. Autoantibodies may compromise the immune system via knockdown of these key proteins, and autoimmune attack on the immune system itself, as supported by immune activation in early stages of infection and during the transition to AIDS. Over 500 human proteins contain pentapeptides or longer consensi, identical to viral peptides. Such homology explains the extensive viral/human interactome, likely related to the ability of viral homologues to compete with human counterparts as binding partners. Pathway analysis of these homologous proteins revealed their involvement in immune-related networks (e.g. natural killer cell toxicity/toll, T-cell/B-cell receptor signalling/antigen processing) and viral and bacterial entry and defence pathways (phagosome/lysosome pathways, DNA sensing/NOD/RIG-1 pathways) relevant to AIDS pathogenesis. At its inception, AIDS may have an autoimmune component selectively targeting the immune system. Immunosuppressive therapy or antibody removal, which has already achieved some success, might be therapeutically beneficial, particularly if targeted at removal of the culpable antibodies, via affinity dialysis.

  • autoimmunity
  • HIV-1
  • immune system
  • AIDS
  • autoantibody
  • mimicry


The human immunodeficiency virus causes acquired immune deficiency syndrome partly by decreasing the capacity of the immune system to deal with opportunistic pathogens (Kaplan et al., 2009). The virus infects and kills CD4+ T-lymphocytes which play an important role in regulation of immune defence (Alimonti et al., 2003) and also targets B cells, natural killer cells, macrophages and microglia (Unutmaz, 2003). However, immune activation, rather than deficiency, is also an early feature of HIV-1 infection (Fernandez et al., 2009; Chang & Altfeld, 2010; Herbein & Varin, 2010).

It has already been noted that the HIV-1 envelope protein is homologous to several components of the immune system including HLA antigens, T-cell receptors, Fas and immunoglobulins (Silvestris et al., 1995). Several autoantibodies have also been noted in AIDS patients and many of these target the immune network (Table 1). Autoimmune disorders are common in HIV-infected patients [e.g. Sjogrens syndrome (Kordossis et al., 1998) and lupus (Zandman-Goddard & Shoenfeld, 2002)]. HIV-1 has also been associated with multiple sclerosis (Brinar & Habek, 2010) and can also cause dementia with Alzheimer's disease-like pathology (Esiri et al., 1998). These associations may be related to immune deficiency but could also reflect viral/antigen mimicry in these autoimmune and other disorders. To this end, homology searches were undertaken in relation to viral homologues of 22 autoantigens reported in HIV infection and to other relationships between viral and human proteomes.

View this table:
Table 1

Autoantigens described in HIV-1-infected patients with a brief description of function (culled from RefSeq unless stated)

Autoantigens and referenceFunction
T-cell receptor and directly related to the immune system
T-cell receptor alpha AAD15152: antibodies to alpha/beta T-cell receptors that cross-react with HIV-1 gp120 (Lake et al., 1994)T-cell receptor
TRBC1 P01850.3 T-cell receptor beta-1 chain C regionT-cell receptor
CD4 (Wilks et al., 1990)A membrane glycoprotein of T lymphocytes that interacts with major histocompatibility complex class II antigens. Also a receptor for the human immunodeficiency virus. Expressed in T lymphocytes, B cells, macrophages and granulocyte and specific regions of the brain. Initiates or augments the early phase of T-cell activation, and may function as an important mediator of indirect neuronal damage in infectious and immune-mediated diseases of the central nervous system (provided by RefSeq)
SPN Sialophorin CD43 (Ardman et al., 1990)Sialoglycoprotein on the surface of human T and B lymphocytes, monocytes, granulocytes: involved in T-cell activation (supplied by OMIM)
FCGR3A Fc fragment of IgG, low affinity IIIa, receptor (CD16a): IgA rheumatoid factor is the Fc portion of Immunoglobulin G (Procaccia et al., 1991)Removes antibody complexes from the circulation, and controls antibody-dependent responses. Expressed on natural killer cells (RefSeq)
Complement system
C1QA C1q cross reactive with gp120 (Metlas et al., 1994)C1q associates with C1r and C1s to yield the first component of the serum complement system. Deficiency of C1q has been associated with lupus erythematosus and glomerulonephritis (provided by RefSeq)
CR1 complement receptor 1 NP_000642 (Sadallah et al., 2003)A monomeric single-pass type I membrane glycoprotein found on erythrocytes, leucocytes, glomerular podocytes, and splenic follicular dendritic cells. The Knops blood group system is a system of antigens located on this protein. The protein mediates cellular binding to particles and immune complexes that have activated complement (RefSeq)
Tumour necrosis factor and cytokines
TNF antibodies correlated with viral load and with antibodies to nef (Capini et al., 2001)This cytokine is mainly secreted by macrophages. Involved in the activation of lymphocytes in response to viral infection (Capini et al., 2001)
FAS (CD95) (Stricker et al., 1998)Member of the TNF-receptor superfamily, containing a death domain. Plays a central role in the physiological regulation of programmed cell death: involved in transducing the proliferating signals in fibroblasts and T cells (RefSeq)
Growth factors, peptides and related
EGFR: NP_005219.2 epidermal growth factor receptor (Ditzel et al., 1994)Mediates monocyte attraction and macrophage proliferation, neutrophil-related inflammatory processes, leucocyte migration (Hamilton et al., 2003; Lamb et al., 2004; Mikami et al., 2005; Liu et al., 2008). In cancer, EGFR antibodies increase the number of natural killer cells, T lymphocytes and dendritic cells infiltrating the metastatic sites (Garrido et al., 2007)
EPO NP_000790.2 erythropoietin (Sipsas et al., 1994)Regulates red cell production but also inhibits T-cell activation and B- and T-cell proliferation (Imiela et al., 1993)
NGF nerve growth factor (Titanji et al., 2003)Regulates the survival of B cells, and its levels are decreased in AIDS patients an effect correlated with B cell loss (Titanji et al., 2003)
RBPJ NP_005340 recombination signal binding protein for immunoglobulin kappa J region (cerebellar soluble lectin) increased autoantibodies in the CSF of AIDS patients (Hagberg et al., 1992)Component of NOTCH signalling regulating B- and T-cell and dendritic cell development and haematopoiesis (Kojika & Griffin, 2001; Tanigaki et al., 2003; Tanigaki et al., 2004; Feng et al., 2010)
Thyroglobulin autoantibodies observed in HIV infected children (Fundaro et al., 1993)Used in the synthesis of thyroxine and triiodothyronine. Thyroid hormones play an important role in preventing the decline in T- and B-cell efficiency in ageing (El Shaikh et al., 2006; Chen et al., 2010).
VIP antibodies cross reactive with gp120 (Velikovic et al., 1993)A potent anti-inflammatory peptide that induces regulatory T cells (Anderson & Gonzalez-Rey, 2010). Also regulates dendritic cell differentiation and Toll receptor signalling (Chorny et al., 2006; Arranz et al., 2008)
DDX17 NP_001091974.1| DEAD (Asp-Glu-Ala-Asp) box polypeptide 17 P72: cross-reacts with the murine leukaemia virus p15-gag antigen (Rucheton et al., 1992)Regulates oestrogen receptor alpha transcription (Wortham et al., 2009). Oestrogens play an important role in T lymphocyte function and inflammation (Weitzmann & Pacifici, 2005; Arnal et al., 2009; Wortham et al., 2009)
ITGB3 integrin, beta 3 (platelet glycoprotein IIIa, antigen CD61 cross-reactive antibodies between HIV-gp120 and CD61 (Bettaieb et al., 2001)Expressed in dendritic cells and involved in the clearance of apoptotic cells and immune tolerance (Parcina et al., 2009): also expressed on T cells (Wierzbicki et al., 2006)
MBP myelin basic protein NP_001020252 Intrathecal autoantibodies (Maimone et al., 1994)Primary constituent of myelin
SERPINA1 alpha(1) proteinase inhibitor [alpha(1)PI; alpha(1) antitrypsin] (Bristow et al., 2001)Serine protease inhibitor whose targets include elastase, plasmin, thrombin, trypsin, chymotrypsin and plasminogen activator (RefSeq). Also a potent anti-inflammatory agent (Janciauskiene et al., 2007)
SNRPB small nuclear ribonucleoprotein polypeptides B and B1: cross-reactive with the p24 gag protein of HIV-1 (De Keyser et al., 1992)Involved in pre-mRNA splicing and the encoded protein may also play a role in pre-mRNA splicing or snRNP structure. Autoantibodies from patients with systemic lupus erythematosus frequently recognize epitopes on the encoded protein (from RefSeq)
SNRNP70 small nuclear ribonucleoprotein 70kDa (U1) cross-reactivity observed with HIV- gp120 (Douvas & Takehana, 1994)Plays a role in nitric oxide-mediated apoptosis in macrophages (Messmer et al., 1998)
TROVE2 60 kD Sjogrens syndrome antigen A (SSA/Ro antigen: Ro60) (Muller et al., 1992)Receptor for apolipoprotein H implicated in lipoprotein metabolism, coagulation, and the production of antiphospholipid autoantibodies (Reed et al., 2009)
Vimentin (Gentric et al., 1991)Regulates monocyte and macrophage differentiation and leucocyte migration (Benes et al., 2006; Barberis et al., 2009)


The autoantigens reported in HIV-1-infected patients are listed in Table 1 with a brief account of function. Homology searches were undertaken vs. the HIV-1 proteome using the NCBI server (blastp). The blast algorithm is designed to detect overall homology between whole proteins, rather than the short contiguous peptide segments within proteins, and the e value was set to 100 000 for this purpose, and matching sequences identified by eye (Supporting Information, Fig. S1). The HIV-1 translated genome (NC_001802) was screened against the human proteome (blast x) with and without the Entrez filter ‘immune’. Proteins thus identified were assigned to pathways using the KEGG mapper (Homo sapiens) (Goto et al., 1997). Clustal alignment was performed at the Uniprot server http://services.uniprot.org/clustalw/. The blast results are available at http://www.polygenicpathways.co.uk/Blasts/hiv1humangenome.htm (no filter) and http://www.polygenicpathways.co.uk/Blasts/hivimmune.htm (filter = ‘immune’). The results of the Kegg pathway analysis are posted at http://www.polygenicpathways.co.uk/hivkegg.htm.


While none of the 22 autoantigens showed significant overall homology to HIV proteins, each contained multiple homologous regions of contiguous amino acids (penta- to octapeptides) or longer gapped consensi, identical to different regions of the same HIV-1 protein, or proteins. These are illustrated in Fig. S1 and Fig. 1.

Figure 1

Top: clustal alignment of the T-cell receptor with the HIV-1 env glycoprotein, gp160. Subsequent: the distribution of viral sequence consensi within T-cell receptor beta, complement receptor 1. Myelin basic protein and vasoactive intestinal polypeptide.

For small proteins such as myelin basic protein, vasoactive intestinal polypeptide or the C section of T-cell receptor beta, these ‘vatches’ (Viral matches) cover extensive regions of the target human protein (from 32% to 34% coverage despite no overall significant alignment). For longer proteins, such as complement receptor 1, vatch motifs were repeated several times (up to four in this case) over the length of the human target. The number of vatches per antigen and per viral protein is also shown in Table 2 (data from Fig. S1).

View this table:
Table 2

The number of viral vatches within the autoantigens reported in HIV-1 infection. The results are shown for each autoantigen and for each viral protein. Env fold refers to the relative enrichment of vatches in the env glycoprotein (e.g. column mean: env/gag or env/pol)

AutoantigenenvgagpolvpurevviftatnefvprRow total
Column mean13.092.502.361.270.860.450.410.360.3221.64
Env fold15.245.5410.2915.1628.8032.0036.0041.14

The total number of viral vatches per antigen ranged from 9 to 43 with a mean of 21.6. By far the greatest numbers of vatches were those identical to the HIV-1 env glycoprotein, the outermost layer of the virus, by 5- to 41-fold in relation to other HIV-1 proteins.

The unfiltered blast (translated viral genome vs. human proteome) http://www.polygenicpathways.co.uk/Blasts/hiv1humangenome.htm returns the most homologous proteins, some of which show a significant overall homology with viral proteins. These included HIV-1 retroviral proteins and human endogenous retrovirus K (HERV-K) proteins (E = 1e-16) showing a family relationship between retroviruses. Human proteins with significant overall homology to viral proteins include the MHC class I polypeptide-related sequence A (MICA: E = 0.015), a ligand that binds to natural killer cell receptors; disruption of this pathway causes autoreactive T-cell stimulation, promoting autoimmunity (Groh et al., 2003). Four zinc finger proteins, ZCCHC10 (E = 0.004) and ZCHHC13 (E = 0.05), both of unknown function, CNBP (E = 0.028) and ZNF384/CAGH1 (E = 0.005) are also homologous to viral proteins. CNBP (CCHC-type zinc finger, nucleic acid-binding protein) preferentially binds to single stranded DNA and RNA (embracing viral sequences), regulates the activity of ornithine decarboxylase and influences the metabolism of cholesterol via binding to a sterol regulatory element (Rajavashisth et al., 1989; Calcaterra et al., 2010). ZNF384 is involved in osteoblast suppression and parathyroid hormone actions (Childress et al., 2010). Parathyroid hormone receptors are also found on T cells, B cells and neutrophils (Geara et al., 2010), although the role of ZNF384 in this respect has not been examined. A further close homologue, junctophilin 3 (E = 0.006), regulates communication between cell surface and intracellular ion channels, while GOLGA2B (golgin A2 family, member B: E = 0.028) or hCG2010686 (E = 2e-08) is of unknown function.

Although other blast results from the unfiltered and filtered blast (‘immune’), with the exception of MICA, were not significant with respect to overall protein identities, all subsequent proteins contain vatches similar to those in Fig. S1 and Fig. 1. blast results are ranked by E value and the final result (i.e. the least significant) on the unfiltered blast page yielded a matching sequence of 50% identity (QP IPIV I + A LV A + I + A V W) belonging to the HLA-A antigen. The final result on the filtered blast page (query = immune) yielded a 6/7 amino acid matching sequence (Q-RCSSN) belonging to tumour necrosis factor (ligand) superfamily, member 13b. http://www.polygenicpathways.co.uk/blasts/hivimmune.htm.

The KEGG pathway analysis of 561 proteins returned from these two sweeps (http://www.polygenicpathways.co.uk/hivkegg.htm) shows that the HIV-1 human homologues are involved in multiple aspects of the immune network (Table 3: number of proteins in brackets) [cytokine/cytokine interactions (33) and chemokine signalling (17); phagosome (26), lysosome (8) and Fc gamma-mediated phagocytosis (9); natural killer cell toxicity (25); haematopoiesis (21); toll receptor signalling (19); JAK-STAT pathway (18); T- (17) and B-cell (12) receptor signalling; antigen processing (16); leucocyte migration (5), IgA production (13) and primary immunodeficiency (5). Pathogen defence networks, including the DNA sensing (11) and the NOD (8) and RIG-1 (14) signalling networks, were also represented as well as pathogen entry pathways] [toxoplasmosis (23), Staphylococcus aureus (20), hepatitis (15), Leischmaniasis (19), Chaga's disease (12), Helicobacter pylori and malaria (9), Escherichia coli infection (4)]. HIV-1 homologues are also involved in autoimmune disorders [systemic lupus erythematosus (26), type 1 diabetes (18), autoimmune thyroid disease (17) and asthma (6)] and in cancer (11), type 2 diabetes (3), Alzheimer's disease (4), and cardiomyopathy pathways (6). Neural pathways included the neurotrophin (10) and ERBB (7) signalling networks and axon guidance (10) pathways as well as long-term potentiation and depression (2), and the common pathways MAPK (17), Wnt and NOTCH (3), calcium (6), PI signalling and apoptosis (11) (Table 3).

View this table:
Table 3

A summary of the KEGG pathway analysis of the human homologues of HIV-1 viral proteins. The number of proteins involved in each pathway is shown in parentheses

Cytokine-cytokine receptor interaction (33)Toxoplasmosis (23)Pathways in cancer (11)AutoimmunePurine metabolism (5)Neuroactive ligand-receptor interaction (10)MAPK signalling pathway (17)Osteoclast differentiation (27)
Phagosome (26)Staphylococcus aureus infection (20)Chronic myeloid leukaemia (7)Systemic lupus erythematosus (26)Oxidative phosphorylation (5)Axon guidance (10)Apoptosis (11)Cell adhesion molecules (24)
Natural killer cell mediated cytotoxicity (25)Leishmaniasis (19)Prostate cancer (6)Type I diabetes mellitus (18)Inositol phosphate metabolism (3)Neurotrophin signalling pathway (10)VEGF signalling pathway (6)Endocytosis (16)
Hematopoietic cell lineage (21)RIG-I-like receptor signalling pathway (14)Pancreatic cancer (6)Autoimmune thyroid disease (17)Tryptophan metabolism (3)ErbB signalling pathway (7)Insulin signalling pathway (6)Regulation of actin cytoskeleton (11)
Toll-like receptor signalling pathway (19)Chagas disease (12)Small cell lung cancer (5)Allograft rejection (17)Galactose metabolism (2)Long-term potentiation (2)Calcium signalling pathway (6)Focal adhesion (9)
Jak-STAT signalling pathway (18)Cytosolic DNA-sensing pathway (11)Acute myeloid leukaemia (4)Graft vs. host disease (17)Histidine metabolism (2)Long-term depression (2)TGF-beta signalling pathway (5)Spliceosome (6)
T-cell receptor signalling pathway (17)Epithelial cell signalling in Helicobacter pylori infection(9)Non-small cell lung cancer (4)Asthma (6)Phenylalanine metabolism (2)GnRH signalling pathway (4)Olfactory transduction (4)
Chemokine signalling pathway (17)Malaria (9)Glioma (4)Other viral myocarditis (17)Glycerophospholipid metabolism (2)Phosphatidylinositol signalling system (4)Extraxellular matrix receptor interaction (4)
Antigen processing and presentation (16)Shigellosis (8)Colorectal cancer (3)Hepatitis C (15)Nicotinate and nicotinamide metabolism (2)Wnt signalling pathway (3)Protein processing in endoplasmic reticulum (4)
Complement and coagulation cascades (14)NOD-like receptor signalling pathway (8)Renal cell carcinoma (3)Hypertrophic cardiomyopathy (6)Starch and sucrose metabolism (2)Notch signalling pathway (3)Vascular smooth muscle contraction (4)
Intestinal immune network for IgA production (13)Amoebiasis (8)Melanoma (3)Dilated cardiomyopathy (6)Tyrosine metabolism (2)mTOR signalling pathway (2)Adherens junction (3)
B-cell receptor signalling pathway (12)Vibrio cholerae infection (5)Endometrial cancer (3)Prion diseases (5)Cell cycle (2)Ubiquitin mediated proteolysis (3)
Fc gamma R-mediated phagocytosis (9)Bacterial invasion of epithelial cells (5)Bladder cancer (2)Arrhythmogenic right ventricular cardiomyopathy (4)Collecting duct acid secretion (3)
Lysosome (8)Pathogenic Escherichia coli infection (4)Alzheimer's disease (4)Tight junction (3)
Fc epsilon RI signalling pathway (6)Type II diabetes mellitus (3)Pancreatic secretion (3)
Leucocyte transendothelial migration (5)Bile secretion (2)
Primary immunodeficiency (5)Nonhomologous end-joining (2)
Adipocytokine signalling pathway (5)Progesterone-mediated oocyte maturation (2)
Salivary secretion (2)
Gastric acid secretion (2)
Cardiac muscle contraction (2)
Dorso-ventral axis formation (2)


All the 22 AIDS autoantigens tested showed extensive homology with diverse HIV-1 proteins, notably the envelope glycoprotein. This suggests that the autoantigens derive from antibodies raised to the HIV-1 virus, which subsequently target their human homologues. Autoantibodies, to C1Q, ITGB3, T-cell receptors alpha and beta, VIP, SNRNP70 and DDX17 are indeed known to cross react with HIV-1 proteins (Table 1). Apart from their immune role, antibodies are antagonists, as in binding to their target protein they impact upon its function, essentially causing protein knock down. Antibodies are not only able to enter cells (Mallery et al., 2010) but can also traverse the blood brain barrier (Pardridge, 2008). Indeed, a large number of autoantigens in a number of autoimmune diseases including systemic sclerosis, systemic lupus erythematosus, mixed connective tissue disease, Sjogren's syndrome and idiopathic myopathies are directed at intracellular antigens (Racanelli et al., 2011).

These human homologues of the AIDS virus are involved in immune defence and an immediate effect of the antibodies to viral infection would be expected to knock down T-cell alpha and beta receptors and related proteins (CD4, sialophorin), complement (CR1, C1Q) and cytokine (FAS, TNF) pathways as well as growth factor pathways (NGF, EGFR) and haematopoiesis (erythropoietin). This strictly immunopharmacological effect would be expected to severely compromise the immune network, as a direct result of antibodies raised to infection. These antibodies might also be expected to activate the immune system, resulting in the production of further antibodies, setting in motion the chain of events leading to the eventual destruction of CD4+ T-lymphocytes as well as other immunocompetent cells. CD4 is an autoantigen and the presence of CD4 autoantibodies in HIV-1-infected patients is directly correlated with the loss of CD4+ lymphocytes (Muller et al., 1994). This scenario would perhaps account for the immune activation seen in the early stages of HIV-1 infection and in the progression to AIDS (Chang & Altfeld, 2010).

Anti-CD4 and other autoantibodies have also been observed in noninfected patients exposed to the HIV-1 virus. Such patients, although seronegative for HIV-1 (as detected using other antibodies), are able to mount T-cell responses to HIV-1 viral proteins. A number of HIV-infected patients are also resistant to infection, an effect dependent in part upon genetic background (e.g. the CCR5 delta32 deletion) (see Miyazawa et al., 2009, for review). This scenario could suggest that the virus has indeed been present, priming the immune system, but that it has been successfully eliminated in some cases. The presence of autoantibodies, targeted to human proteins that are viral homologues, could perhaps be regarded as fingerprint evidence for prior infection, despite the absence of the thief, particularly when such autoantibodies cross react with viral proteins (see Table 1). Antibodies raised in response to HIV-1 are primarily concerned with viral elimination, as in any infection. Whether such antibodies become potentially deleterious autoantibodies, via the mechanisms described above, would depend on a number of factors. These would include the degree of matching between viral and human proteins, a factor determined by the strain of virus and by human gene variants that determine protein structure. HLA-related gene variants are also important determinants of the progression to AIDS (Limou et al., 2009)and are likely to influence both immune and autoimmune outcomes.

The homology of HIV-1 viral proteins to so many other human proteins suggests the possibility of a number of other potential autoantigens. A contribution of molecular mimicry and autoimmunity to AIDS pathogenesis has already been proposed (Bost et al., 1988; Scheider et al., 1993; Susal et al., 1993). This survey shows the extent of this process, and together with the KEGG pathway analysis, demonstrates an intensive targeting of the immune network.

That so many human proteins resemble fragments of the AIDS virus is perhaps not unexpected, given that the HIV-1 human protein interaction database currently cites 2589 host/viral interactions (Fu et al., 2009). This extensive interactome no doubt relates to this homology which allows the virus to compete with its human homologues in a variety of signalling networks. These viral homologues of human proteins are likely to be able to compete with their human counterparts in a variety of interactomes, many of which are implicated in immune related function. This type of homology is not restricted to the AIDS virus: for example, vatches from 30 viruses, including many nonretroviruses can be found extensively in the human proteome (Kanduc et al., 2008; Carter, 2010) (see also http://www.polygenicpathways.co.uk/blasts.htm). This situation has arisen due to the incorporation of viral DNA, from both RNA and DNA viruses into mammalian genomes (Katzourakis & Gifford, 2010), an effect also observed in plants, arthropods, fungi, nematodes and protozoa (Liu et al., 2010). It is unlikely that the HIV-1 genome is present in a reference human genome in the NCBI database, but the close homology with an already incorporated HERV-K virus suggests that the HIV-1 vatches may belong to this and other retroviral or viral species, or to previously nonpathogenic ancestors already resident in the human genome. For whatever reason, HIV-1 viral proteins closely resemble fragments of a large sample of the human proteome.

The pathway analysis implicated the virus in a number of pathogen-related life cycle pathways and in the pathogen defence network (DNA sensing, RIG-1 and NOD pathways). Impaired immune function caused by AIDS increases susceptibility to many such opportunistic infections (Kaplan et al., 2009). HIV-related proteins may also exert an influence by controlling pathogen entry and defence.

Autoimmune disorders are common in HIV-infected patients, and the pathway analysis highlighted several related networks involved in these diseases, for example, systemic lupus erythematosus, type 1 diabetes, autoimmune thyroid disease and asthma. HIV-1 has also been associated with multiple sclerosis (Brinar & Habek, 2010) and the myelin basic protein autoantigen in AIDS patients is clearly relevant, as are members of the neuregulin/ERBB network which regulates myelination (Brinkmann et al., 2008). HIV-1 infection can also cause dementia with Alzheimer's disease-like pathology (Esiri et al., 1998) and several vatch-containing proteins are observed in the Alzheimer's disease pathway, and in other relevant areas such as apoptosis and oxidative phosphorylation (Onyango & Khan, 2006) or Wnt and NOTCH signalling (Balaraman et al., 2006). Assorted psychiatric conditions including depression, cognitive deficits, psychosis and mania (Owe-Larsson et al., 2009) have also been linked to HIV infection and the axon guidance pathways, neurotrophin, neuregulin/ERBB signalling, phosphatidylinositol signalling and long-term depression or potentiation are all relevant to these conditions (Carter, 2006, 2007). The AIDS virus has also been associated with an increased risk of developing cancer (Caceres et al., 2010) and many such pathways are occupied by HIV-1/human homologues.

Autoantibodies in AIDS are sometimes considered an epiphenomenon (Liberti et al., 1991). This may relate to the heterogeneity in autoantibody studies. However, this analysis suggests that such antibodies must be derived from antibodies initially targeted at the virus, which, because of this homology, are also able to target their human homologues. As illustrated by these data, there are many known and even more potential autoantibodies which target proteins of related immune function. In addition, specific antibody production is genetically conditioned. In Sjogren's syndrome, different HLA alleles are associated with the generation of autoantibodies to the SSA and SSB autoantigens (Jonsson et al., 2000). The propensity for developing autoantibodies to particular proteins is also genetically determined and inherited (Phillips et al., 1991). A genome wide association study in AIDS showed over 300 risk and protective single nucleotide polymorphisms restricted to the MHC region, covering risk and protective HLA alleles (Pereyra et al., 2010). The presence or absence of specific autoantibodies is thus likely to depend upon the genetic background which determines whether the immune system denotes that the HIV-1 homologues are self, or foreign.

That autoantibodies might derive from pathogens expressing proteins homologous to the autoantigens appears to be a widespread phenomenon. For example, the autoantigens in Alzheimer's disease, multiple sclerosis and schizophrenia are homologous to proteins expressed by the viral and bacterial risk factors in these diseases. These autoantibodies target key proteins relevant to the pathology of each disease (Carter, 2010, 2011a, c, d). In cystic fibrosis, the reported autoantigens are homologous to both the CFTR protein, and to proteins expressed by bacteria that hypercolonise the airways in this condition (Carter, 2011b). Autoantibodies are observed in many other diseases where viral and bacterial mimicry of human proteins is extensive (Kanduc et al., 2008; Trost et al., 2011). In such conditions, it has been argued that the slightly different homologues, rather than the precise matches may be the more malignant, as such are more likely to be regarded as nonself, while antibodies raised thereto may still target their human homologues, albeit with lower affinity (Kanduc, 2010; Carter, 2011a). While autoantibodies are often considered epiphenomena, this homology link between pathogen and autoantigen, and the pathologically relevant targets of the autoantibodies, suggests that autoantibodies derived from pathogen encounter may have an important causative contribution in many diseases.

This analysis shows a very high degree of homology between HIV-1 viral proteins and multiple HIV-1 autoantigens. This has previously been reported along with evidence for cross reactivity in some cases, but not on such a scale. This suggests that the autoantigens are the direct result of antiviral antibodies raised in response to infection and also that the autoantibodies play a key role in causing AIDS, partly by targeting key immune-related proteins, acting as antagonists, and in the initial stages of the disease-causing immune activation, as observed in AIDS patients, resulting in the death of T lymphocytes and other cells expressing the autoantigens, a supposition supported by the correlation of CD4 antibodies with CD4+ T-cell destruction (Muller et al., 1994). In addition, the pathway analysis of over 500 vatch-containing human proteins illustrates how the virus could interfere with multiple signalling networks related to the immune system and other relevant pathways.

At its inception, AIDS could thus perhaps be classified as an autoimmune disorder that selectively targets the immune system itself. This has implications for future therapy, including immunosuppression and antibody removal strategies.

Hyperactivation of the immune system is a marker of HIV-1 progression to AIDS, and immunosuppressive therapy has indeed been shown to be of benefit in some cases in conjunction with antiretroviral therapy (Argyropoulos & Mouzaki, 2006). Plasmapheresis or immunoadsorption have also been used in AIDS therapy and have shown significant benefit both in the symptoms associated with AIDS (peripheral polyneuropathy and neuromuscular disorders (Cornblath, 1988; Kiprov et al., 1988; Salim et al., 1989) or vasculitis (Cohen et al., 1993), and in the progression of AIDS itself (Raven, 1994; Bainbridge et al., 1997), including a rise in the CD4+ T-cell count (Tomar et al., 1984; Blick et al., 1998). These plasma exchange techniques remove all antibodies, some of which may well be beneficial, as well as other protein messengers including beneficial or deleterious cytokines, growth factors or peptide messengers. A more targeted approach using affinity columns laced with autoantigens and autoantibodies, perhaps tailored to each individual's antibody profile, might be expected to improve upon the already noticeable benefits.

In summary, HIV-1 vatches exist in hundreds of human proteins, and particularly within HIV-1 antigens and in multiple members of the immune network or in other proteins related to the associated comorbid conditions of HIV-1 or AIDS. A closer understanding of this phenomenon, which likely applies to many other viruses, would help in our understanding of the relationships between pathogen and disease and perhaps result in the development of effective therapies in this and other conditions.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Fig. S1.The alignments of HIV-1 viral proteins with the autoantigens reported in HIV-1-infected patients.


View Abstract