Location: Home >> Detail
This work is licensed under aCreative Commons Attribution 4.0 International License
J Psychiatry Brain Sci. 2018; 3(6): 14. https://doi.org/10.20900/jpbs.20180014
1 Department of Biosciences, Biotechnologies and Biopharmaceutics, University of Bari, Bari 70125, Italy;
2 2 Complex Structure of Neuropsychiatry Childhood-Adolescence, United Hospitals of Foggia, Foggia 71122, Italy.
* Corresponding Author: Darja Kanduc.
Objective: Genetic, epigenetic, and environmental factors such as infections have been proposed as potential causes of autism spectrum disorder (ASD). Searching for the molecular mechanism by which infections might contribute to the etiopathogenesis of ASD, we analyze here the hypothesis that immune responses to infectious agents may cross-react with human proteins that, when altered, relate to autistic neurodevelopmental spectrum disorders.
Methods: Viral and human proteins were analyzed for peptide sharing using the Pir Peptide Match resource.
Results: We find that: (i) an intense peptide overlap occurs between ASD–related viruses and ASD–related human proteins, and might underlie cross-reactivity scenarios following viral infections; (ii) viral peptide sharing also occurs with Y-linked proteins, in this way highlighting an additional potential cross-reactivity burden that would involve male subjects only; (iii) many shared peptides are also part of epitopes experimentally validated as immunopositive in the human host.
Conclusion: This study offers a cohesive set of data that suggests a contribution of immune cross-reactivity to the genesis of ASD.
There is a clinical and epidemiological consensus that infections may be causally involved in the neurodevelopmental and behavioural disturbances that characterize ASD [1–6]. Specifically, a main role in the pathophysiology of ASD has been repeatedly suggested for the activation of the immune system against the infectious agents [7–21] rather than to viral/microbial activities per se, such as, for example, subversion of the host protein synthesis machinery , manipulation of membrane transport pathways , and up-regulation of HLA-E expression with, consequently, suppression of NK cell recognition [24–26].
However, how the immune system may be involved in ASD remains undetermined. The issue is further complicated by the fact that ASD are biased towards males, with ratios of 4:1 or higher [27–29], so that analyses of or hypotheses on ASD have to contemplate the male bias.
During last decades, in an attempt to further our understanding of infection-induced diseases, we analyzed sequence identities between viruses and humans [30,31]. We pursued the hypothesis that peptide commonality between microbial and human proteins might have the potential to trigger cross-reactions in the human host during infection, thus inducing autoimmune pathologic sequelae [30–37]. Here, we test such a hypothesis by analyzing viral pathogens that have been related to ASD [38–51]—namely Borna disease virus, Rubella virus, Measles virus, Influenza A virus, and Mumps virus—and searching for amino acid (aa) sequences common to (i) human proteins that, when altered, have been associated with autistic disorders, and (ii) proteins expressed by Y-linked genes.
The viral proteomes analyzed in the present study are as follows, in order of aa length and with abbreviations, Taxonomy ID, number of proteins, and number of aa in parentheses: Parvovirus B19 ( B19; 10798, 3 proteins, 2006 aa); Borna disease virus (BDV; 928296; 6 proteins; 3014 aa); Rubella virus (RUBV; 11041; 2 proteins; 3179 aa); Measles virus (MeV; 11235; 7 proteins; 4680 aa); Influenza A virus, H1N1 (211044; 13 proteins; 4788 aa); Mumps virus (MuV; 11171; 8 proteins; 4977 aa). Proteomes are described in detail at http://www.uniprot.org .
The primary sequence of viral proteins was dissected into hexapeptides overlapped by five residues each other. For example, BDV Envelope glycoprotein p57 (UniProt: P52638; 503 aa) was sequentially dissected into MQPSMS, QPSMSF, PSMSFL, SMSFLI, and so forth until its last hexapeptide LGRWQE, for a total of 498 hexapeptides. Then, each viral hexamer was probed for occurrences within human proteins characterized by being related to ASD or encoded by Y-linked genes.
A set of 138 ASD-related human proteins was randomly retrieved from UniProt database and NCBI (https://www.ncbi.nlm.nih.gov/gene) using ‘autism’ and ‘autistic’ as keywords and consisted of 138 proteins listed in Table S1. A set of 44 proteins expressed by Y-linked genes was assembled using data from Skaletsky et al.  and UniProt database, and are listed in Table S2. Proteins are indicated by UniProt entry and names.
The immunological potential of the peptide matching was analyzed using the Immune Epitope Database (IEDB; www.iedb.org) database . Only epitopes that had been experimentally validated as immunopositive in the human host were considered. Data on brain protein expression were retrieved from https://www.proteinatlas.org/humanproteome [55,56].
We selected and analyzed five proteomes belonging to infectious agents that have been reported as related to or concomitant with ASD. That is, BDV [38–40], RUBV [41–46], MeV [41,47,48], Influenza A virus [49–51], MuV . As a control, we used Parvovirus B19. B19 is the etiological agent of the infantile fifth disease, preferentially targets the erythroblasts in the bone, and does not appear to be related to ASD .
Hexapeptides were used as operational minimal immune determinants in light of a vast scientific literature that documents the crucial roles exerted by peptides 5–6 aa long in immunogenicity and antigenicity [58–85].3.1 Hexapeptide sharing between B19, BDV, RUBV, MeV, influenza A virus, and MuV proteomes, and human proteins related to ASD
Table 1 quantitatively describes the hexapeptide sharing between B19, BDV, RUBV, MeV, Influenza A virus, and MuV, and the set of the 138 human proteins related to autism (see Table S1). It can be seen that all of the analyzed viruses share hexamers with human proteins related to ASD. Even if at a lesser extent, the peptide commonality also involves the control B19 virus.
Qualitatively, the viral hexapeptide distribution among the ASD-related proteins is described in Table 2. At first glance, space does not permit a match-by-match discussion of the vast peptide sharing illustrated in Tables 1 and 2. In synthesis, three main points emerge. Firstly, 96 hexapeptides belonging to the 6 analyzed viral pathogens also occur in 76 out of 138 human proteins associated with ASD, in this way indicating a non-stochastical clustering of peptide matches in 55% of the analyzed human proteins.
Secondly, in light of the fact that the probability of a hexapeptide occurring once in a protein is 1 out of 206, the viral vs human peptide overlap reported in Table 2 largely exceeds mathematical expectation. As a note a latere, we observe that this unexpected high peptide matching may be explained by the evolutionary role played by viruses in the origin of the eukaryotic nucleus .
Then, as a third point, it was found that human proteins related to ASD and sharing peptides with the analyzed viruses are mostly expressed in the brain. Limiting our analysis to a few examples—i.e., ARI1B, CTTB2, HUWE1, SETD2, and SHAN3 proteins—we find that:
The human Y chromosome contains a male-specific non-recombining region with 27 protein-coding genes (Table S2) . The peptide sharing between the analyzed viruses and the Y-chromosomal proteins is shown in Table 3. It can be seen that 7 Y-linked proteins namely: KDM5D (SMCY), PC11Y, TBL1Y, TXNG2, USP9Y, UTY (KDM6C), and ZFY—share hexa-/heptapeptides with all of the potential viral pathogens analyzed here, B19 excluded (Table 3).
The 7 Y-linked proteins are widely expressed in the brain. Specifically:
The viral vs human peptide overlap illustrated in Tables 1–3 also has an immunologic potential. Indeed, Table 4 shows that many shared peptides are part of immunopositive epitopes cataloged at IEDB .
Again, the high number of epitopes containing the shared peptide sequences precludes a detailed epitope-by-epitope discussion. However, a special attention has to be drawn to the peptide RLLDRLVR shared between MeV and the human Deleted in Autism protein 1 (DIA1). In fact, the peptide RLLDRLVR corresponds to the epitope IEDB ID 54638 (Table 4) that was shown to be responsive in 80% of 5 HLA-A2–positive adults revaccinated with measles-mumps-rubella vaccine .
Another point calling for attention is that hexapeptide analyses underestimate by one order of magnitude the potential cross-reactivity that may be evoked by immune responses following infections. As a matter of fact, also a pentapeptide can represent a minimal immune unit endowed with immunogenicity and antigenicity [58–86]. And, in addition, discontinuous pentapeptide epitopes have been reported in IEDB database such as the Influenza A hemagglutinin conformational epitope P134S137K177Y180T183 (IEDB ID: 164481). Hence, expanding the similarity analyses to (dis)continuous pentapeptides would generate a viral vs human cross-reactivity scenario even more massive than that displayed in Tables 1-4.3.4 ASD-related and Y-chromosomal proteins involved in the viral peptide overlap: expression in the human brain
Taken together, Tables 1-4 factually support the hypothesis that, following active infections by the viral pathogens analyzed here, the consequent anti-viral immune responses might cross-react with ASD-related and/or Y-linked proteins expressed in the human brain. However, it is incumbent to observe that the brain expression data reported above have been mainly obtained in animal models, and using microarray analyses, quantitative real-time PCR, and in situ hybridization technologies. Actually, it is well-known that transcript abundances only partially predict protein abundances [119,120]. Consequently, since a conditio sine qua non for a cross-reaction to occur is a sufficient level of antigenemia, the Human Protein Atlas resource (https://www.proteinatlas.org/) [55,56] was searched for data on the expression level in the brain of the proteins discussed above. Results are reported in Fig. 1 that shows that the ASD-related proteins ARI1B, CTTB2, HUWE1, and SHAN3 have an expression level from medium to high in the nervous system cells (panel A) and that, among the Y-chromosomal proteins, TBL1Y has a high protein expression level in almost all nervous system cells (panel B). Expression in peripheral nerve cells was low or absent. Data on KDM5D, USP9Y, and TXNG2 proteins were pending or not available at the time of the present study.
ASD is of unknown etiology. Genetic components such as mutations , epigenetic disorders such as altered methylaton , abnormal cytokine profile where inflammatory signals dominate , and environmental factors such as pollutants  or immune responses to infections [7–21] appear to contribute to ASD. However, whichever it may be the invoked causal factor, the mechanism(s) at the basis of ASD remain unsettled.
Here, we hypothesized that immune responses against infectious viral agents might have the potential to cross-react with proteins that, when altered, are related to autism. Actually, Tables 1, 2 and 4, and Fig. 1A document an ample and potentially immunologic peptide matching of B19, BDV, RUBV, MeV, Influenza A virus, and MuV with ASD-related proteins, thus supporting the possibility of a causal connection between infection and neurodevelopmental diseases through cross-reactivity. Very much the same consideration applies to data from Tables 3 and 4, and Fig. 1B that highlight peptide overlaps between viral and Y-chromosomal proteins. In this case, the potential cross-reactivity burden specifically involves male subjects, so determining a higher male susceptibility to neurodevelopmental disorders.
It has to be underlined that ASD comprehends autism, childhood disintegrative disorder and Asperger syndrome that are characterized, in different combinations and at various level of intensity, by symptoms such as impaired capacity for interactions, a restricted repertoire of interests, stereotyped repetitive activities, and decreased intellectual ability . Hence, the here described numerous brain proteins involved in the peptide matching and in the consequent potential cross-reactions might explain the multitude of symptoms that characterize ASD. In addition, the ASD symptomatology and severity may have spatial-temporal patterns, with, for example, in utero infections involving the maternal immune system. In this regard, as a final caveat, it has to be kept in the due account the observation that the maternal immune response in the absence of virus and obtained by using the synthetic double-stranded RNA poly (I:C) is sufficient to cause behavioral changes in the offspring . Moreover, the infection outcome in children and adults may depend on previous immune responses following previous encounters with the pathogens [125,126].
In sum, the data support our previous studies [30–37], offer the immune cross-reactivity paradigm as a possible approach for studying autism and neuropsychiatric disorders, and strongly warrant further collaborative research efforts to determine the impact of viral vs human cross-reactivity in the etiology of ASD.
Supplementary File 1: PDF file
DK proposed the original idea, developed sequence analyses and wrote the manuscript. AP contributed to the clinical analysis and discussion of the data, and to the writing of the manuscript.
The authors declare that they have no conflicts of interest.
Kanduc D, Polito A. From Viral Infections to Autistic Neurodevelopmental Disorders via Cross-Reactivity. J Psychiatry Brain Sci. 2018; 3(6): 14. https://doi.org/10.20900/jpbs.20180014