Since 1980, connections have been sought between chronic—especially autoimmune—diseases and possible microbial actions that might promote or suppress immune responses. Here, we propose that the pathogenesis of some chronic diseases is linked to ancestral infections or exposure to noxious agents: Some modern-day diseases reflect the capacity of organisms to “memorize” responses to external signals and transmit them across generations; the resulting information can be subsequently made functional under certain conditions, through stimulation by a relevant secondary influence. The proposal is supported by observations of transgenerational epigenetic inheritance. Therefore, autoimmune hepatitis (AIH) could be a recollection of our ancestors' long-term affliction with viral hepatitis; the original causative agent may not be extant today, but “memory” of the infection has persisted. A similar argument could apply to other chronic diseases. In this article, we discuss evidence supporting this idea, with an emphasis on the exemplar pair of viral hepatitis and AIH, and outline a mechanistic hypothesis.
Several idiopathic or multicausal human diseases may have clinical or pathophysiological mirrors among diseases with known causes ranging from microbial to chemical or lifestyle-related etiologies. Examples include autoimmune and viral hepatitis, ulcerative colitis and amebiasis or chronic dysentery, and sarcoidosis and tuberculosis. Even if the two diseases in such a pair are clinically dissimilar, the primary pathophysiological events could overlap substantially (e.g., increased intestinal iron absorption in nutritional anemia and hemochromatosis).
In this article, we explore an evolutionary explanation for such overlaps, focusing on autoimmune and related infectious diseases, particularly autoimmune hepatitis (AIH) and viral hepatitis. This explanation does not apply to all known human diseases. Rather, we argue that certain chronic (especially autoimmune or multifactorial) diseases reflect “memories” of interactions between the biological system and environmental factors over time. If the pathogenicity were high and the time of interaction were short (a few generations), the environmental factor could radically change the biological system—for example, by inducing a mutation. However, if the pathogenicity were low and the time of interaction were long (several generations), the changes in the biological system would be pervasive, but often subtle, with minimal or no obvious functional derangement on exposure.
The present article is structured as follows. We review the evidence for transgenerational epigenetic inheritance, the relationship between chronic diseases and infections, and the growing recognition that evolutionary theory is important for medicine. We then address general considerations about autoimmune diseases and their proposed relationship to infections and examine recent evidence for the role in both infectious and autoimmune disease of signaling lymphocytic activation molecule and associated proteins and of pattern-recognition receptors, specifically Toll-like receptors (TLRs). Thereafter, we focus mainly on aspects of AIH: the relevance of human leukocyte antigen (HLA) polymorphism, the known autoantigens in the condition, and the role of sex hormones in its pathogenesis. Finally, using AIH as an example, we expand our broad explanation into a mechanistic hypothesis of the evolutionary origins of autoimmune (and other chronic) diseases.
Transgenerational epigenetic inheritance
The idea that the environment can induce heritable changes in the germline seems to connote Lamarckism and has therefore met with skepticism (Morgan and Whitelaw 2008). However, there is now abundant evidence that epigenetic inheritance occurs (Jablonka and Raz 2009). In humans and other mammals, embryonic and fetal development can be influenced by transplacental communication or by effects of the egg cytoplasm (Drake and Walker 2004), but that does not necessarily entail genomic changes that can be transferred to later generations. More interestingly, it has been shown (a) that male gametes can transfer genetic information from prions and probably from viruses to the offspring (Shorter and Lindquist 2005); (b) that DNA methylation patterns, which are instrumental in the negative control of gene expression, can be transmitted to the fetus from both parents (Rakyan et al. 2003, Anway et al. 2005); (c) that toxicants and other environmental factors, including cigarette smoking, can exert transgenerational effects (Franklin and Mansuy 2010); and (d) that such transmission is likely to elucidate the etiologies of complex diseases (Johannes et al. 2009, Petronis 2010).
Although germline cells generally erase epigenetic memory during gametogenesis and preimplantation development, some memories escape this reprogramming and are retained transgenerationally (Lang and Schneider 2010), possibly because of histone methylation or acetylation or other chromatin modifications (Kaufman and Rando 2010). Particularly interesting in the context of the present article is the evidence implicating heritable changes in the major histocompatibility complex (MHC) II in the pathogenesis of multiple sclerosis (Chao et al. 2010) and the heritability of environmental factors promoting type 1 diabetes (MacFarlane et al. 2009).
There has been some debate about whether metastable epialleles (i.e., genes that are identical except in the extent of methylation), which are variably expressed under epigenetic influence in the absence of genetic heterogeneity (Rakyan et al. 2002), have significant effects in humans, as they do in mice. This is an important issue, but we will not focus on it in the present article.
Infection and chronic disease
Marshall and Marshall (2004) and then Proal and colleagues (2009) suggested that many chronic diseases, including inflammatory autoimmune diseases, result from infection with L-form and biofilm bacteria that persist and slowly proliferate, suppressing the innate immune response. They argued that long-term treatment with low-dose antibiotics and withdrawal of vitamin D is therapeutically beneficial. However, this proposal, originally introduced by Trevor G. Marshall, has been supported only by anecdotal evidence and computer simulations and has not been evaluated in randomized controlled trials, so it has not been universally accepted.
Ewald and colleagues have advanced a broadly similar idea (Ewald 2004, Cochran et al. 2000). Drawing attention to the association of several chronic disorders with viral or bacterial infections (e.g., liver cancer with hepatitis viruses B [HBV] and C [HCV]), stomach ulcers with Helicobacter pylori, amyotrophic lateral sclerosis with echovirus, multiple sclerosis with Chlamydiapneumoniae, and so on, Ewald (2004) confuted the widely accepted theory of the genetic causation of chronic diseases. He argued that low-grade bacterial, viral, and other infections are causal in many chronic diseases, emphasizing that at least half of the infections associated with chronic diseases are sexually transmitted (Ewald 2004). In Ewald's (2004) doctrine, two major determinants of the evolution of virulence are the chronicity of infection and the density of the afflicted human population; in contrast, disease-causing genes are likely to be eliminated from the population over a series of generations. In terms of evolutionary theory, this view is simplistic, unless the genes in question decrease the probability of survival and reproduction. However, this does not invalidate the idea that many chronic diseases are consequences of infection rather than genetic predisposition.
Childhood infections may also contribute to the maintenance of immune regulation and, according to the hygiene hypothesis (von Hertzen 2000), may help to protect the adult against autoimmunity. Strachan (1989) linked the heightened expression of allergic and atopic diseases to the reduction of cross-infection in families, a result of improved hygiene and diminished family size. In its most likely interpretation, the hygiene hypothesis states that early-life viral and bacterial infections may suppress T helper 2–mediated proallergic immune responses (Yazdanbakhsh et al. 2002). Animal studies also indicate that early-life viral infections may protect against the development of type 1 diabetes mellitus (Strom 2009). However, the hygiene hypothesis has been challenged by a series of conflicting findings. In an attempt to resolve the controversies, Matricardi and Bonini (2000) proposed that “high turnover of appropriate bacteria at the mucosal level” rather than “stable colonization by certain species” is necessary for the “harmonious development and maintenance of the immune system” (p. 1509). Therefore, the inhibitory effect of early infections on subsequent immune-mediated diseases becomes species specific and is determined by the interaction between particular pathogens and the immune system.
The hygiene hypothesis is appealing, although it is currently supported only by qualitative evidence (such as the higher incidence of allergies and asthma among urban than among rural populations) rather than by thorough epidemiological surveys. The theory predicts that as childhood infections are eliminated, as they have been in developed countries during the past 50–60 years, the incidence of autoimmunity will increase. Indeed, for some autoimmune diseases, such as type 1 diabetes, the worldwide incidence is rising more rapidly than genetic change can explain (Cooke 2009), and no specific initiating infection has been identified. However, noninfectious environmental triggers, such as psychological stress or overfeeding during childhood, could also be implicated in the rise in type 1 diabetes.
Contributions such as those of Ewald and Marshall and the hygiene hypothesis are somewhat outside the current mainstream. Nevertheless, they are interesting because they challenge the consensus that all or most chronic diseases have a genetic basis, and they invite further consideration of the role of infection in chronic conditions such as autoimmunity. Moreover, the association between the infections and chronic diseases has to be reexamined both cross-sectionally and longitudinally. Granted such a role, ancestral infections could effect the development of chronic and autoimmune diseases in later generations by virtue of transgenerational epigenetic inheritance.
Gluckman and colleagues (2007) proposed a developmental-origins-of-disease paradigm, according to which the risk for disease in postnatal life is to some extent established in utero. During development, the organism adapts homeostatically to environmental influences, which, in mammals, means changes in the composition of the placental blood but also makes predictive adaptive responses: That is, changes in the placental blood are interpreted by the developing organism as indicators of its future (postnatal) environment. Such adaptations could have an obvious evolutionary advantage, but if they fail to match actual postnatal conditions, the adult organism could be at increased risk for certain diseases. This, according to Gluckman and colleagues (2007), is the case for humans in the modern developed world, where the balance between energy intake and expenditure is very different from its ancestral value and therefore from the proposed in utero prediction. This, it is argued, is the cause of the present epidemic of chronic noncommunicable diseases.
This idea—that a mismatch between the modern-day human condition and that of our hunter–gatherer forebears has implications for our propensity to certain diseases—has been explored by a number of other authors, although they did not emphasize developmental plasticity as Gluckman and colleagues (2007) did (Eaton et al. 1988, Nesse and Williams 1994, O'Keefe and Cordain 2004, Varki 2009). The past two decades have seen a number of powerful arguments favoring the value of an evolutionary perspective for medicine. Lappé (1994) was primarily concerned with the evolution of antibiotic resistance among bacteria, but he emphasized the more general need for the medical community to take cognizance of evolutionary theory. Nesse and Williams (1994) clarified a number of misunderstandings about evolutionary medicine and cautioned that attempts to improve the species by eugenics are misguided; some of our vulnerabilities to disease may be favored by natural selection because apparently deleterious genes can have unidentified advantages (the well-known case of the sickle-cell trait is far from unique), and there is no such thing as a “normal” human genome.
This view received surprising support in 2008, when a genomic phylostratigraphic study of disease-related genes revealed that most such genes in humans date back to, or before, the earliest eukaryote, and many of the remainder date to the origin of multicellularity (Domazet-Lošo and Tautz 2008). Some disease-related genes that have arisen by gene duplication and mutation may be of relatively recent origin, but they appear to represent a minority. According to Domazet-Lošo and Tautz (2008), very few disease-related genes postdate the origin of mammals.
Autoimmune diseases: General considerations
Five potential etiological and pathophysiological processes have been implicated in autoimmune diseases: inflammation, apoptosis, genetic predisposition, infection, and environmental triggers (Mackay et al. 2008). Inflammation is closely linked to autoimmunity; to date, this relationship has underpinned the standard approach to the treatment of most such conditions, including AIH (e.g., Theve et al. 2008, Testro and Visvanathan 2009, Meyer 2009). Apoptosis in excess may increase the exposure of autoantigens to the immune system, but if it is insufficient, it may fail to eliminate autoreactive lymphocyte clones (Mackay et al. 2008, Meyer 2009). Studies of genetic predisposition have been focused mainly on diversity in the HLA locus (Béland et al. 2009). Infection can be associated with increased levels of autoantibodies. Although few human infections have been definitively linked to autoimmune diseases so far (Mackay et al. 2008), there is evidence for a causal connection in some cases (e.g., Epstein–Barr virus [EBV] causing infectious mononucleosis in young adults appears to be a risk factor for multiple sclerosis; Thacker et al. 2006). Moreover, the environmental triggers for autoimmune disorders identified to date, apart from xenobiotics, are mostly infectious agents, notably viruses (Béland et al. 2009). Autoimmune disorders affecting distinct organs and loss of self-tolerance may be explained by molecular mimicry between foreign and self-antigens (Fujinami 1988).
However, the relationship between autoimmunity and infection is complicated. It often seems that the overall burden of infections dating from childhood is instrumental, along with genetic predisposition, in inducing autoimmunity; conversely, according to the hygiene hypothesis, infections during childhood might protect the patient from autoimmune diseases (Liu and Murphy 2003).
Involvement of SLAM and SAP in both infectious and autoimmune conditions
The role of the signaling lymphocytic activation molecule (SLAM) and related proteins in the pathogenesis of autoimmune diseases has been studied in animal models. For example, a murine study showed that the SLAM and Fc-gamma receptor intervals in the lupus-susceptibility locus Nba2 on chromosome 1 act cooperatively on the differentiation and survival of autoantibody-producing cells (Jørgensen et al. 2010). In particular, SLAM-associated protein (SAP), encoded by the X-chromosomal gene SH2D1A, has been implicated in X-linked lymphoproliferative disease (Veillette et al. 2007), which makes patients abnormally susceptible to infection with EBV (Hislop et al. 2010). SAP binds to SLAM-family receptors, and the cytoplasmic Src-related protein tyrosine kinase FynT is recruited (Latour et al. 2003), initiating a signal sequence that plays a crucial role in regulating T-cell and natural killer (NK)–cell function. This signaling pathway appears to have a role in various autoimmune conditions, including systemic lupus erythematosus (SLE) and AIH (Furukawa et al. 2010).
For the purposes of the present article, the most important implication of these studies is that a specific gene product—SAP—has been implicated in both viral infection and autoimmune disease. However, this entails the involvement of the adaptive immune response. It is well known that most autoimmune disorders involve recognition of autoantibodies by B cells, with (usually) no evident T-cell involvement, so the innate immune response would seem to be more important. In this context, the possible role of TLRs must be considered.
Toll-like receptors and other pattern-recognition receptors
Several pattern-recognition receptors (PRRs), a network of germline-encoded receptors, have been identified during the past two decades. PRRs recognize molecular motifs characteristic of infectious microorganisms and also endogenous molecules produced by injured tissues (Montero Vega and de Andrés Martín 2009). They control many aspects of both innate and adaptive immunity (Joffre et al. 2009) and regulate tissue repair and regeneration (Rakoff-Nahoum and Medzhitov 2008). They have been implicated in a wide variety of pathological conditions, suggesting that a universal immunobiological model could explain a multiplicity of human diseases (Atkinson 2008).
The best studied of these sensors are the TLRs, 10 of which are known to be expressed in humans; their structures and functions have been reviewed by Montero Vega and de Andrés Martín (2009) and Testro and Visvanathan (2009). The recent identification of mutations and common polymorphisms in the TLRs has revealed their importance not only in susceptibility to infection but also in several noninfectious diseases (Misch and Hawn 2008). In particular, in some autoimmune disorders, recognition of endogenous ligands by TLRs leads to sustained responses by innate immune cells that contribute to a loss of self-tolerance (Wagner 2006). Many autoantigens exposed by tissue injury stimulate innate immunity through TLRs (Marshak-Rothstein 2006). In vivo and in vitro studies have revealed that self-DNA and self-RNA translocated by endosomes stimulate plasmacytoid dendritic cells via TLR9- and TLR7-dependent pathways, just as microbial nucleic acids do (Montero Vega and de Andrés Martín 2009). TLR4 recognizes molecular patterns on Gram-negative bacteria and also on ligands released from injured host cells (Testro and Visvanathan 2009).
TLR activation initiates a signaling cascade that up-regulates inflammatory genes, recruiting myeloid cells and enhancing the secretion of proinflammatory cytokines (interleukin-1 and tumor necrosis factor [TNF]–∝) and interferons (IFNs), and promoting the expression on antigen-presenting cells (APCs) of molecules required to induce an adaptive immune response (Testro and Visvanathan 2009). The stimulation of dendritic cells requires a type I IFN. When recombinant IFN is used persistently to treat patients with tumors or chronic viral infections, SLE is often induced; aberrant IFN-∝ production induced by endocytosed self-DNA and self-RNA through TLRs is now considered pivotal to the pathogenesis of SLE (Marshak-Rothstein 2006).
Interferons, defective apoptosis, and telomerase deficiency
Current knowledge about the IFNs—ubiquitous cytokines secreted by all mononuclear cell types infected with a DNA or RNA virus—was reviewed by Meyer (2009). There are three major classes: Type I or nonimmune IFNs are mostly IFNs ∝ (from leukocytes) and β (from fibroblasts), although there are other minor variants; type II or immune IFN is IFN γ, which is mainly produced by NK and T cells; and type III comprises the λ IFNs. IFNs change the expression of up to 300 genes that encode proteins involved in antiviral defense mechanisms, inflammation, adaptive immunity and angiogenesis. SLE patients have overactive IFN pathways that are associated with inadequate clearance of apoptotic particles, so apoptosis products (DNA-CpG motifs and U-RNA) accumulate (Antwi-Baffour et al. 2010). Similar abnormalities have been found in patients with other chronic conditions, such as primary Sjögren's syndrome (Routsias and Tzioufas 2010); they appear to be associated with telomerase deficiency and the resulting abnormalities in telomere length, and this may be characteristic of autoimmune conditions in general (Georgin-Lavialle et al. 2010).
Viral hepatitis and autoimmune hepatitis: Historical aspects
What is now recognized as viral hepatitis is an ancient disease and appears to correspond to descriptions in Classical Greek texts, although a viral etiology was only established after the work of Findlay and MacCallum in the late 1930s (Snell 1947). Soon after World War II, it had become clear that at least two different strains of virus were hepatopathogenic (Snell 1947). Shortly afterward, HBV was recognized as the causal agent in infectious hepatitis and was distinguished from hepatitis A virus (Havens 1948).
Waldenström (1950) described a chronic form of hepatitis in young women. The disease was later found to be associated with other autoimmune syndromes and acquired the name lupoid hepatitis, because circulating antinuclear antibodies were identified in the patients. Mackay and colleagues (1965) renamed it autoimmune hepatitis when autoimmunity was first accepted as a general concept. The association between AIH and HLA alleles was first described in 1972 (Manns and Vogel 2006), and during the succeeding two decades, autoantibodies in AIH patients were found to be directed against endoplasmatic reticulum proteins in the liver and kidney and against soluble liver antigens. Since 1990, it has become evident that AIH is immunoserologically and genetically heterogeneous; there are racial, geographic, and genetic variations in clinical manifestations, disease progression, and treatment outcome (Manns and Vogel 2006, Czaja 2008). The onset of the disease is usually during the first two decades of life and women are more susceptible than men (Béland et al. 2009).
Human leukocyte antigen polymorphism
It is well established that HLA class I and II polymorphisms are associated with autoimmune diseases, such as autoimmune thyroid disease (Kim et al. 2003). It is now evident that HLA diversity also influences susceptibility to viral diseases, including HBV and HCV infections (Shichi et al. 2008). HLA class I diversity seems to have a role in the innate immune response as well as in the acquired immune response.
The susceptibility alleles (e.g., DRB1*0301 and DRB1*0401) for AIH in white North American and northern European patients have been identified on the gene DRB1 (Czaja 2008). They encode a six-residue sequence, associated with susceptibility, at positions 67–72 in the DRβ chain of the major MHC class II molecules; the key determinant is lys-71 (Czaja and Donaldson 2000). Factors influencing the severity of the disease include the DRβ71 gene dose and polymorphisms in genes that can modify the immune response (e.g., TNF-∝, cytotoxic T lymphocyte antigen-4, the tumor necrosis factor-receptor superfamily; Czaja 2008). Other susceptibility alleles reflecting indigenous triggering antigens may be significant in different geographic regions, and these may provide clues to the environmental triggers involved in AIH.
Autoantigens in autoimmune hepatitis
Most liver-specific autoantigens are enzymes with key roles in cellular homeostasis. In AIH patients, many of the disease-specific autoantibodies are directed against cytochrome P450 2D6 (CYP2D6; Bogdanos and Dalekos 2008). Antibodies against the same or functionally related enzymes have been recognized in patients chronically infected with hepatitis C (Sutti et al. 2010) and drug-induced autoimmunity (Obermayer-Straub et al. 2000).
Intermediate filament autoantibodies have been reported in viral hepatitis patients; Murota and colleagues (2001) found that serum levels of anti-cytokeratin [CK] 8, anti-CK18, and anti-CK19 antibodies were significantly higher in AIH patients than in healthy controls and were normalized by steroid treatment. Once again, there are similarities in the pathophysiology of both viral hepatitis and AIH at the molecular level.
Although none of the autoantigens so far identified has significant sequence homology with HBV or HCV antigens, studies on a mouse model infected with the murine hepatitis virus A59 indicated a close similarity between the three-dimensional structures of the viral and autoantigen epitopes (Duhalde-Vega et al. 2009). It is not yet known whether the same applies to human disease, but in view of the foregoing evidence, the idea is at least plausible.
Hepatitis and sex hormones
Liver-specific expression of the HBV surface antigen gene was investigated in transgenic mice by Farza and colleagues (1987). Expression rose dramatically in males at puberty but not in females, and this rise was reversed by castration; nevertheless, not only testosterone but also both estrogen and glucocorticoids stimulated expression of the gene. A mouse model study by Theve and colleagues (2008) corroborated the particular susceptibility to infection of males before puberty. The transcription of both feminine (cytochrome P450 4a14) and masculine (cytochrome P450 4a12 and trefoil factor 3) genes was abnormal in the infected animals, and several gender-neutral genes were also affected: H19 fetal liver messenger RNA, lipocalin 2 and ubiquitin D. The unsaturated:saturated fatty acid ratio was higher in infected animals. In young human males chronically infected with HBV, earlier-onset puberty and increased steroid 5∝-reductase type II activity are associated with earlier hepatitis B e antigen seroconversion, higher serum alanine aminotransferase levels and a more marked decrease in HBV load (Wu et al. 2010).
Sexual contact is a common route of viral hepatitis infection, and infection is accordingly more likely to occur during the period of sexual activity. Accordingly, the genes involved in the reaction to the virus are activated concomitantly with those involved in sexual activity and sex-hormone secretion. The hypothesis discussed in the present article is that in our ancestors, such simultaneity could have resulted in the linkage of the above-mentioned changes; thus, the memory of ancestral reactions to viral hepatitis is activated at the onset or during the period of sexual activity or on physiological secretion of sex hormones and is manifest as AIH. This view is consistent with the typical age of onset of AIH (15–40 years old).
We suggest that an ancestral form of a human hepatitis virus antigenically related to modern-day HBV and HCV was instrumental in transmitting the risk for AIH to subsequent generations (figure 1). Similar associations can be suggested between rheumatoid arthritis and ancestral microbe-induced reactive arthritis; obesity and poor ancestral lifestyle; multiple myeloma and a virus- associated encephalomyelitis; periodic fever and malarial infection; achalasia and trypanosome-related esophagopathy; sarcoidosis and mycobacterial infection; SLE and syphilitic infection; and so on.
A mechanistic hypothesis: Ancestral origins and manifestation of autoimmune hepatitis
The mechanism proposed in this section is based on the foregoing evidence. We set it out in numbered stages, but this does not imply a linear sequence, or that every stage must be implicated in all cases of the disease. For example, either or both of the innate and acquired immune response mechanisms could be involved in pathogenesis in a particular patient. (1) An ancestral human male was infected with a hepatitis virus. (2) Part of the viral genome entered gametes and passed to the zygote at fertilization. (3) During embryogenesis, the epigenetic memory escaped erasure. (4) In subsequent generations, virus-like antigens were synthesized postnatally and became endogenous (EVLA [endogenous virus-like antigens]). (5) Childhood infections have become much less frequent in developed countries (hygiene hypothesis); therefore, T helper 2–mediated immune responses are not suppressed. (6) At puberty, EVLA is activated concomitantly with genes involved in sexual activity and sex-hormone secretion. (7) On—for example—tissue injury, TLRs recognize EVLA and its associated autoantigens (cytochrome P450 2D6, CKs). This is more probable in individuals with susceptibility alleles of HLA genes (e.g., DRB1*0301). (8) The inflammatory response is stimulated, and EVLA fragment and associated autoantigens are presented on APCs. (9) IFN overproduction suppresses apoptosis (which is perhaps associated with telomerase deficiency). (10) SAP may potentiate the autoimmune response through effects on T-cell and NK-cell function. (11) Exposure to an antigen (or perhaps hapten) similar to EVLA stimulates immune responses and the response becomes autoimmune through molecular mimicry.
More generally, we propose that biological systems can memorize reactions or responses to a variety of external signals and transmit them, genetically or epigenetically, across generations. The influence of such memory and its penetration into system function is likely to depend on the duration, intensity, and pathogenicity of the external influence, as well as the density of the system affected and its vulnerability or ability to accommodate and withstand the challenge. If n units of a biological system receive a given amount of external signal for a sufficient time, the interaction between the pathogenicity of the signal and the vulnerability of the system determines the fate of that system. If it survives, its reaction might be integrated and stored in memory, pending recovery and activation, which may be primary or secondary to a factor resembling the original challenge. With reference to the hygiene hypothesis: Early-life infections may inhibit the system that expresses data preserved in the genome from ancestral infections.
According to this view, AIH is a recollection of our ancestors' long-term affliction with viral hepatitis; the original causative virus may or may not be extant today, but the memory of the infection has persisted and can present as AIH. AIH could be just one example of a general relationship between infectious and autoimmune disorders. Similarly, sarcoidosis could recall ancestral tuberculosis, ulcerative colitis could recall chronic dysentery, and so on. Table 1 depicts several pairs of infection and chronic autoimmune diseases, which potentially exhibit an evolutionary linkage.
|Viral polyarthritis||Rheumatoid arthritis|
|Amebiasis or chronic dysentery||Inflammatory bowel diseases|
|Viral hepatitis||Autoimmune hepatitis|
|Syphilis||Systemic lupus erythematosus|
|Viral encephalomyelitis||Multiple sclerosis|
|Helicobacter pylori infection||Idiopathic peptic ulcer|
|Viral polyarthritis||Rheumatoid arthritis|
|Amebiasis or chronic dysentery||Inflammatory bowel diseases|
|Viral hepatitis||Autoimmune hepatitis|
|Syphilis||Systemic lupus erythematosus|
|Viral encephalomyelitis||Multiple sclerosis|
|Helicobacter pylori infection||Idiopathic peptic ulcer|
A skeptical perspective
In principle, the hypothesis proposed in this article can be tested critically in specific applications, such as the possible relationship between AIH and an ancestral viral infection. As a general account of the origin of autoimmune and other chronic diseases, it is not amenable to refutation. The infection–AIH proposal should be testable in mice, some strains of which are especially susceptible to viral hepatitis and to murine AIH; to date, however, no evidence for the epigenetic inheritance of viral genes in infected mice has been reported, although a large-scale study would be required to detect what must—if it occurs at all—be a low-probability event.
Moreover, the hypothesis comprises a combination of nonmainstream and, in some cases, controversial ideas from a variety of different fields. It is hard to reconcile it with mainstream hypotheses about autoimmunity (e.g., clonal deletion, clonal anergy, idiotype network theory), and that in itself must invite skepticism. Moreover, it does not always provide convincing explanations in areas where the mainstream hypotheses appear weak; for example, it could barely give a clear explanation for the absence of T-cell responses that are typical of autoimmune conditions. The established connection between infection and autoimmunity is also commonly explained by reference to superantigens, which is a much simpler explanation than the hypothesis provides, although superantigens have not been identified in all relevant cases. However, the induction of autoimmune conditions by chemical agents, as in the case of drug-induced SLE, which is difficult to explain by mainstream theory, could potentially be elucidated by our hypothesis; that suggests a worthwhile program of experiments using suitable animal models.
The hypothesis that we have outlined here implies that a long-lasting external influence can be coded by a biological system into a persistent set of data, which can remain silent until it is functionalized by appropriate conditions and provoked by a relevant secondary influence. If ancestral microbial infections are the origin of an autoimmune disease, the extraction and use of the original or related antigens to induce immune tolerance is, theoretically, a plausible therapeutic approach.
Moreover, the hypothesis predicts that the status of diseases of interest is dynamic and depends on the variability of external influences over time. This provides a context for predicting future illnesses by estimating the burden and penetrability of recent influences. For example, cigarette smoking is a known cause of obstructive pulmonary disease and lung cancer. If this effect continues for a sufficient time, the hypothesis predicts that future generations may suffer from a primary obstructive lung disease, similar in nature to the smoking-related lung diseases, without exposure to cigarette smoke. Similar arguments might apply to the association between atherosclerosis and its risk factors; if the risk factors remain effective, a primary atherosclerosis may emerge at some time in the future. The validity of this assertion remains to be studied in longitudinal population-based studies; the follow-up of generations of individuals exposed to significant environmental factors or microbial agents at a particular time may shed light on potential mechanisms by which these exposures can be encoded into heritable characteristics.
This study was supported by a grant from the Tuberculosis and Lung Disease Research Center, at Tabriz University of Medical Sciences, in Tabriz, Iran.