Immunogenetics of autism spectrum disorder: A systematic literature review

The aetiology of autism spectrum disorder (ASD) is complex and, partly, accounted by genetic factors. Nonetheless, the genetic underpinnings of ASD are poorly defined. The presence of immune dysregulations in autistic individuals


Introduction
Autism spectrum disorder (ASD) is one of the most common neurodevelopmental conditions, affecting approximately 1 in 59 individuals worldwide (Chiarotti & Venerosi, 2020).ASD is characterised by difficulties in social communication and interaction, and repetitive patterns of behaviors and interests (American Psychiatric Association, 2013).In addition, hyper and/or hypo reactivity to sensory stimulation may be present (Hazen et al., 2014).These so-called core symptoms persist often into adulthood and are associated with challenges that impact the life quality of both autistic individuals and their carers (Cadman et al., 2012;van Heijst & Geurts, 2015).Nevertheless, at the moment we lack effective (pharmacological) interventions targeting the 'core' autistic symptoms (Loth et al., 2016), in part as the pathophysiological mechanisms driving them remain unclear.
ASD is heterogeneous, consisting of different mechanisms, phenotypes, and trajectories.A previous meta-analysis of twin studies demonstrated that ASD has a strong genetic component, indexed by an estimated heritability of 64-91% (Tick et al., 2016).Some of the genes that have been linked to ASD are genetic regulators of synaptic formation and signaling.However, these putative genetic 'risk' loci do not fully capture the complex biological landscape of ASD (Wegener Sleeswijk et al., 2019) and this highlights the necessity to explore the role played by additional gene families and their related pathways in ASD.
One potentially relevant mechanism involves the immune system, including its genetic regulators.The contribution of immunity to ASD has been hypothesized and corroborated by recent findings in both animal models and humans (Masi et al., 2017;Onore et al., 2012).In mice, prior studies support an association between maternal immune activation (MIA) during pregnancy and neurodevelopmental phenotypes including schizophrenia-like and autistic-like behaviors in newborns (Choi et al., 2016;Reed et al., 2020;Rudolph et al., 2018); and suggestive evidence highlights a modulatory role of the interleukin (IL)-6 pathway (Choi et al., 2016;Reed et al., 2020).In humans, epidemiological reports describe the presence of immune dysregulations, such as oversensitivity to allergens, auto-antibodies production, and deregulated anti-infectious processes in autistic individuals (Ashwood & Van De Water, 2004;Edmiston et al., 2018;Hughes et al., 2018;Zerbo et al., 2015).
Albeit environmental stressors of immune homeostasis are likely to influence these associations, evidence suggests that immunoregulatory genes are important to ASD (Leboyer et al., 2016;Torres et al., 2012).For instance, candidate gene analyses have reported an association between ASD and specific human leukocytes antigens (HLA) haplotypes (Bennabi et al., 2018;Torres et al., 2012).Pangenomic analyses, like genome-wide association studies (GWAS) of ASD and population-based autistic-like traits, have also identified associations with common variants, or single nucleotide polymorphisms (SNPs), in immune system genes (Arenella et al., 2022;Grove et al., 2019).Of interest, a metaanalysis of GWASs performed in several psychiatric disorders, including ASD, confirmed a strong association signal in the HLA gene cluster region (Lee et al., 2019).Transcriptomic analyses of post-mortem brain tissues of autistic individuals have further indicated several immune gene dysregulations (Gandal et al., 2018).Finally, indirect support of immune-related genetic variability in ASD is offered by observational studies as these describe an association between ASD and family history of autoimmune and inflammatory conditions (Atladóttir et al., 2009;Vinet et al., 2015).
Although these prior findings support the contribution of immune gene to ASD, it is important to acknowledge the complexity of the immune system genetics and the multitude of functionally diverse immune gene pathways (Parkin & Cohen, 2001).Currently, it is unclear if -and whichspecific immunogenetic mechanisms are relevant to ASD.To address this question, it is crucial to gather prior findings of the association between immune genes and ASD and examine the immunological and neurological function of associated genes.
Hence, we conducted the firstto the best of our knowledge -systematic review of studies exploring the association between immune genes and ASD.Our work aggregates findings from both (immunerelated) genetic and transcriptomic studies of ASD spanning the period from 2010 to 2022.Synthesized findings provide an overview of bona fide ASD-related immune genes, along with information about their specific immune function.Additionally, we explore the representation of immune genes among established ASD 'risk' genes and evaluated their potential neurodevelopmental role.This effort is key to better define the potential immunogenetic underpinnings of ASD.

Methods
Following the PRISMA guidelines (Moher et al., 2009), we conducted a systematic literature review of immunogenetic research in ASD.The systematic review was registered on PROSPERO with number CRD42021222673.Our literature search covered the period from January 1, 2010, through August 1, 2022.We excluded articles before 2010 to only include studies using state-of-the-art genomic approaches and analytical guidelines (e.g., updated genome builds and reference genomic panel), and the most updated ASD diagnostic criteria (e.g., based on the Diagnostic and Statistical Manual of Mental Disorder (DSM) -VI or 5 or the International Classification of Disease (ICD)-10).Our literature search was conducted using the scientific literature databases PubMed and Web of Science.For both databases, a search query was created including terms for ASD and immune-related genetics: [(ASD OR autism OR autistic OR autistic disorder OR autism spectrum disorder OR Asperger's syndrome OR Kanner's syndrome OR pervasive neurodevelopmental disorder) AND (immune gene OR immune genetics OR immune genetic polymorphism OR inflammatory gene OR inflammation genetics OR immune RNA))].
We filtered reports to extract original research articles, written in English, and confined to human populations.The literature assessment followed four steps consisting of 1) identification 2) screening 3) eligibility 4) interpretation (see Fig. 1).Two authors (MA and RM) independently performed the search and assessed the resulting articles.
First, results from the initial queries in PubMed and Web of Science were combined and recorded.Articles classified as duplicates were removed.
Second, articles were screened based on title and abstract for their relevance to the purpose of the review.We excluded review articles, as well as articles that did not primarily investigate ASD (e.g., focusing on other neurodevelopmental conditions).
Third, studies were assessed for eligibility.At this stage, full-text articles were retrieved and evaluated using the Newcastle-Ottawa Scale (NOS) for case-control studies (https://www.ohri.ca/programs/clinical_epidemiology/oxford.asp)(Wells et al., 2000).In brief, the NOS allowed us to assess studies based on: a) selection, that refers to the definition/ ascertainment of cases -here based on the consensus diagnostic measures for ASD, as defined by the British Association of Psychopharmacology (Howes et al., 2018) (e.g., DSM, ADOS/ADI-R)and the selection of control groups (i.e., individuals without clinical records and unrelated to the case groups); b) comparability, that refers to the use of strategies to control for potential confounding factors, such as sex, medication use, genetic syndromes, comorbidities and c) exposure, that refers to the adoption of the same ascertainment and analytical approach between cases and controls.Studies were classified as 'eligible' if receiving a NOS-based quality score above 4 at least.We also filtered studies based on statistical power.Hence, we considered studies including appropriate sample size, which and this depended upon the genetic approach used and alignment with previous power calculations (Meurs, 2016;Owzar et al., 2012;Spencer et al., 2009).On average, included studies counted on on average, total (cases and controls) N > 30-40 for brain and blood expression studies; N > 100 for SNP genotyping; N > 4000 for genome-wide association studies).
Eligible studies were thus included in a qualitative synthesis.For these studies, the reported genes were cross-checked for their relevance to the immune system in the ImmGen and innatedb portals, offering curated, comprehensive list of human genes involved in immune functioning (https://www.immgen.org/;https://innatedb.com/).
To estimate whether immune genes were also found among genes implicated in ASD, we examined the most recent list of 1,075 ASD genes (July 2022) from the Simons Foundation Autism Research Initiative (SFARI) (https://gene.sfari.org/).Within these genes, we examined the overlap with genes that are known to be functionally involved in the immune responses and which are annotated in two different immune databases: innatedb (N = 3,714 genes; https://innatedb.com/)and ImmGen (N = 2,483 genes; https://www.immgen.org/),using a Fisher's exact test.Moreover, we used the "GENE2FUNC" function of the webbased platform Functional Mapping and Annotation of Genome-wide association studies (FUMA) (Watanabe et al., 2017) to explore the functional role of the overlapping genes by mapping genes to biological and molecular pathways (i.e., KEGG pathways).By leveraging Brain-Span data and Gtex data, FUMA allowed us to estimate gene expression in the brain throughout the life span and overall gene expression across bulk tissues.These enrichment analyses were performed using all the genome-wide genes as background hence to define the function of the overlapping ASD-immune genes as compared to the rest of the genome.

Results
Our search strategy led to 106 scientific reports of which 28 original research articles were deemed eligible for review after the 4-step selection process described above (Fig. 1).To those, we manually added the most recent GWAS meta-analyses of ASD (Grove et al., 2019), which also provides support for a role of immune genes as illustrated in supplementary analyses.Overall, our review supports the role of immune genetic factors in ASD -as confirmed by both genomic and transcriptomic analyses in autistic individuals.11 of the 29 studies consisted of genotype analyses of specific immune gene polymorphisms, and bioinformatic analyses on genome-wide SNP-based association of ASD (e.g., enrichment tests).Included genotype-based studies are presented in Table 1.In addition, 18 of the 29 studies were based on expression analyses of candidate immune-related genes or immune gene pathways in the blood and post-mortem brain tissues of autistic individuals (Table 2).
The following sections illustrate the findings from, respectively, immune genotype association studies and gene expression studies.Also, the findings from our cross-comparison of ASD-risk genes and immunerelated genes are presented, together with their functional annotation.
1. Immune genes associated with ASD Genotype-based studies indicated associations between ASD and three main classes of immune-related genes involved in 1) the processing and presentation of antigens; 2) immune regulation (e.g., transcription factors); and 3) cytokine signalling.Table 3 provides an overview of the immune genes associated with ASD, with information about their neuro-immune function and the signalling pathway to which they have been annotated.

Expression of immune genes in brain and blood in ASD
Expression changes have been measured either in the peripheral blood or in the post-mortem brains of autistic individuals, with consistent results across studies.Tables 4-5 list the immune genes and gene pathways with altered cross-tissue expression in autistic individuals, respectively, with details about direction of effect (i.e., up/down regulation), and neuro-immune functions.

Immune genes among ASD-related genes
The cross-comparison of annotated immune genes and ASD-related genes (SFARI) indicated an overlap of 98 genes., which did not reach statistical significance (χ 2 = 2.4; p = 0.12).The 98 overlapping immune-ASD genes are presented in Supplementary Table 1 and a description of their immune function, together with the SFARI ASD liability score is provided.Moreover, FUMA-based analyses allowed to deepen the biological characterization of overlapping genes.Based on these analyses, overlapping ASD-immune genes are enriched in networks controlling mTOR signaling, NK and T cell signaling, and allograft rejection (Supplementary Fig. 1).Additional enriched pathways include neurodevelopmental processes regulating axon guidance and neurotrophic signaling, along with the estrogen signaling cascade.The analyses of temporal expression of these genes indicate a predominant expression of these in the late prenatal period and late infancy, whereas cross-tissue expression analyses highlight upregulation of immune-ASD genes in adipose tissues and testis (Supplementary Figure 2).

Discussion
There is increasing evidence supporting the contribution of the immune system to the pathophysiology of ASD (Masi et al., 2017).However, the mechanisms through which the immune system may influence the individual liability to ASD are unclear.One possibility is that immune genetic factors mediate this relationship.This hypothesis is reinforced by recent findings suggesting immune genes as one of the mechanisms contributing to the complex genetic architecture of ASD (Arenella et al., 2022;Grove et al., 2019;Tamouza et al., 2020).
To explore this hypothesis, we reviewed the literature on immune genetic studies in ASD available to date.Collectively, prior studies support an association between ASD and 1) inherited variations in genes controlling both innate and adaptive immune responses (Balta et al., 2018;Bennabi et al., 2018;Chen et al., 2022;Fallah et al., 2020;Guerini et al., 2018a;Guerini et al., 2020;Mo et al., 2018;Pekkoc Uyanik et al., 2021;Saad et al., 2020;Safari et al., 2017); and 2) altered expression of these immune genes along their pathways in both the brain and at the systemic level (Gandal et al., 2018;Gupta et al., 2014;Lombardo et al., 2020;Lombardo et al., 2017;Patel et al., 2016;Pramparo et al., 2015;Sabaie et al., 2021;Voineagu et al., 2011;Wright et al., 2017).By specifically investigating the overlap between genes regulating the immune response and genes that have been linked to ASD liability (https://gene.sfari.org/),we were also able to gain insights on the potential neurodevelopmental functions of immune genes relevant to ASD.The immune genes linked to increased ASD susceptibility support neuronal migration and synaptogenesis and are expressed in key stages of brain development, such as pregnancy and early childhood.Overall, our review suggests that immune genes play a key modulatory role in ASD by affecting early brain development and ontogeny.

Inherited immunogenetic polymorphisms
SNP-based association studies link ASD to diverse immune system genes.Most of the studies reviewed focused on the major histocompatibility complex (MHC) and its hosted human HLA genes including those belonging to the class II cluster, namely HLA-DRB1 and HLA-DQB1 (Bennabi et al., 2018;Guerini et al., 2018b;Guerini et al., 2018a;Guerini et al., 2015;Tamouza et al., 2020).These genes encode molecules that control antigen presentation to CD4 + T helper cells, with consequent induction of the humoral immune response and production of antigen-specific antibodies by the B lymphocyte compartment.In addition, there is also evidence of a role for the non-classical HLA class I gene, HLA-G, a strong immunomodulator which helps to achieve allogenic tolerance thereby ensuring proper fetal development during pregnancy.Dysfunctions at the level at these interfaces may induce the passage of inflammatory molecules through the placenta with autoimmune consequences (Ferreira et al., 2017).
The findings of variations in HLA genes in ASD reconcile with reports of immune dysregulations in autistic individuals and in their family members.For example, HLA class I and class II genes are central to autoimmune pathologies (e.g., systemic lupus erythematosus, rheumatoid arthritis) which have been often diagnosed in the mothers of autistic children (Atladóttir et al., 2009;Vinet et al., 2015).Also, due to their effects on the maternal-fetal interface, variations in HLA class I genes may modulate the previously established link between maternal immune (over)activation and neurodevelopment conditions, including ASD and schizophrenia (Estes & McAllister, 2016;Scola & Duong, 2017).Identified ASD-related immune genes include the CLEC7A gene, which encodes for major molecular censors of the antifungal immune responses and, therefore, potentially contributing to the dysbiosis observed in ASD (Bennabi et al., 2015).Further ASD-immune genes include transcription factor genes, FOXP3, VDR, that manage immunoregulation via T regulatory cells signaling (Guerini et al., 2020;Safari et al., 2017); and genes regulating chemokine and cytokine translation and signal transduction which is essential to T cell response and antiviral, inflammasome response, IL1R, IFNy, IP-10, RANTES (Chen et al., 2022;Fallah et al., 2020;Pekkoc Uyanik et al., 2021).

Increased transcription of immune genes
Transcriptomic studies describe a state of increased expression, or upregulation, of immune genes in ASD (Gandal et al., 2018;Gao et al., 2021;Gupta et al., 2014;Lombardo, 2018;Lombardo et al., 2017;Nazeen et al., 2016;Voineagu et al., 2011).This up-regulation suggests the possibility of an over-active immune system in ASD.Functional assessment of upregulated immune genes highlights the influence of both innate immune regulators, such as genes regulating the NK cell signaling pathway, and mediators of the adaptive immune response, such as genes controlling the T and B cell-mediated responses.However, the drivers of the expression changes reported in these studies are unclear.Potential drivers include external triggers such as viruses and bacteria.For example, the observed immunogenetic dysregulations are similar to expression changes observed in animals following viral (synthetic viral RNA polyinosinic:polycytidylic acid) and/or bacterial (lipopolysaccharide) stimulation in utero (Lombardo et al., 2018).Given the findings of inherited variations in inflammasome genes and transcription factor genes in ASD (Balta et al., 2018;Chen et al., 2022;Guerini et al., 2020;Safari et al., 2017), potential environmental triggers may act on the top of a pre-existing immunogenetic vulnerability in some cases of ASD.
Notably, the upregulation of immune genes appears persistent through life, being reported in autistic children, autistic adults, and in the post-mortem brains of autistic individuals (Fallah et al., 2020;Gupta et al., 2014;Pramparo et al., 2015;Wright et al., 2017).It is possible that these expression changes are initiated at a very early stage of life.Increasing evidence suggests that gestation is a period of increased immune sensitivity, and fine-tuning of immune gene regulation (Hsu & Nanan, 2014;Mandal et al., 2013;Morelli et al., 2015).During gestation, the fetal immune system is shaped, and it is endowed with a range of immune responses to adopt later in life.Factors that interfere with the immune milieu at this stage of life, including immunogenetic background and environmental infectious triggers, could prime the immune system towards a pro-inflammatory state.This phenomenon is so-called 'fetal programming', and it impacts health outcome and susceptibility to disease throughout the entire life span (Mandal et al., 2013).As a proof of this, fetal exposure to immune stimulation has been linked to a lifelong pro-inflammatory state in mice (Kim et al., 2022;Mandal et al., 2013).Some autistic individuals also exhibit signs of inflammationindexed by elevated levels of pro-inflammatory cytokineswhich may originate from gestational immune activation (Edmiston et al., 2018;Xu et al., 2015).
Moreover, our study synthesis demonstrates that immune activation is widespread, as it occurs both peripherally and in the central nervous system (Lombardo et al., 2020).This, therefore, validate the hypothesis that ASD should be regarded as a systemic disorder.Concerning brainbased transcriptomic studies (Gupta et al., 2014;Lombardo et al., 2017;Voineagu et al., 2011;Wright et al., 2017), analyses focused on brain regions that are functionally (Kennedy & Courchesne, 2008) and structurally relevant to ASD (Stanfield et al., 2008;Van Rooij et al., 2018); and, for example, locate the upregulation of cytokine and leukocyte activation genes in the temporal lobe and in the frontal cortex (Voineagu et al., 2011), known to mediate social and cognitive autistic symptoms (Mundy, 2018).Of note, these brain regions overlap with those exhibiting immune gene upregulation in mice exposed to immune challenges (for a review see (Woods et al., 2021)), suggesting that these structures may be particularly sensitive to the effect of immune stimulation.Moreover, recent work demonstrates a significant enrichment of immune/inflammatory genes in brain structures where autistic individuals exhibit deviations from a so-called typical neuroanatomical range (Ecker et al., 2022).From these studies, it is however impossible to disentangle if immune gene upregulation reflects a cause of neural anomalies, or if it is a mere consequence of those.

The neurodevelopmental function of immune genes
To further understand the importance of immune genes in ASD, we explored the most recent list of genes associated with ASD liability and curated by the SFARI initiative (Abrahams et al., 2013).Hence, among these ASD genes, we identified a set of immune genes and explored their immunological and neurobiological functions.The identified immune-ASD genes regulate both innate and adaptive immune response, through NK or T-cell-mediated signalling pathways among others.Notably, these genes are enriched for a breadth of neurodevelopmental processes, from neuronal formation, mTOR signalling to axon guidance and synaptogenesis.
To explore the neurodevelopmental influence of these genes, we also examined their pattern of expression across developmental epochs.Hence, we demonstrated a predominant expression of these in the late prenatal period, and early childhood.Both these periods are critical for brain development.The late prenatal period witnesses the formation and organization of synapses between neurons (Goddings & Giedd, 2014;Tau & Peterson, 2010) and the start of axonal myelination (Tau & Peterson, 2010).Late infancy (up to the beginning of adolescence), in contrast, coincides with the refinement of neural circuits and synaptic pruning that underpins sensory processing and learning (Goddings & Giedd, 2014).Immune activationand up-regulation of immune genesat these life stages may have a cascading effect on the formation of neurons and the connectivity among them.For example, the activation of microglia -and the consequent circulation of pro-inflammatory Abbreviations: ASD: autism spectrum disorder; TD: typically developing; NOS: Newcastle Ottawa Scale; LPS: lipopolysaccharides; M. Arenella et al. cytokines and complement molecules -is known to influence the normal process of engulfment and pruning of disused synapses (Boulanger, 2009;Garay & McAllister, 2010).Moreover, the (over) activation of innate immune cells (e.g., NK cells) -one of the molecular pathways upregulated in ASD -has been linked to damage to the myelin sheath, thereby hampering neuronal transmission in experimental models of multiple sclerosis (Shi et al., 2000).Taken together, these findings suggest that immune genes are critical influences on neurodevelopment.This suggestion is reinforced by prior literature investigating immune genes, also included here (Table 1-2), using animal and in vitro models.For instance, studies in mice indicate that HLA genes aid neuronal migration and are central to the formation of synapses (Elmer & McAllister, 2012;Yirmiya & Goshen, 2011).Evidence also supports the role of cytokine genes in the formation of neuronal and glial cells, and in the establishment of neural connectivity (reviewed by (Deverman & Patterson, 2009)).In vitro analyses indicate an overexpression of CLEC7A in microglia cells during early postnatal stages, consistent with the formation of white matter tracts (Q.Li et al., 2019).In addition, the analyses of main immunogenetic pathways (e.g., TNF and IL-6 signalling) reveal that these comprise a pool of neurotrophic factors, such as STAT3 and AKT, in their signalling cascade (Yang et al., 2018).The activation of these pathways may consequently have downstream effects on neuronal formation and organization (Zegeye et al., 2018).To further support the interaction between immune genes and neurodevelopmental genes, prior transcriptomic findings indicate that immune gene upregulation in autism is coupled with downregulation of neurodevelopmental genes (Lombardo et al., 2017).The opposite directions of these effects suggest that the over-activation of immune genes may have inhibitory effects on the normal execution of neurodevelopmental processes.

Sources of inter-individual immune gene variability
Although most associations between immune genes and autism are replicated across studies, there are some inconsistencies.Variations in some immune genes are observed in autistic individuals of European ancestries but not replicated in other ethnical groups (Kara et al., 2018;Mo et al., 2018;Ramos et al., 2012).For example, in a Chinese cohort, ASD was not associated with variations in the AIM2 and the JARID2 genes (Mo et al., 2018), which respectively mediate the inflammasome activation and the neural tube formation, whereas the JARID2 gene, together with other inflammatory genes, have been associated with ASD in exploratory European-based analyses (Ramos et al., 2012).These discordant genetic effects may be accounted by fluctuations in allele distribution across populations, which may occur randomly (in case of genetic drift) and/or non-randomly (due to in/out-breeding), and they are also affected by a population's geographical relocation and exposure to different environmental challenges (Charlesworth & Charlesworth, 2017).HLA genes, in particular, are known to be both highly polymorphic and highly variable across ethnical groups (Ramos et al., 2015;Shiina et al., 2009).One potential approach that could overcome, even partly, such biases is the study of HLA ancestral haplotypes that are conserved across populations due to their immune properties (Price et al, 1999;Dorak et al, 2006;Bennabi et al, 2018).
Some studies indicate that the effects of immunogenetic variations are male-specific (Safari et al., 2017).The male-specific effects are in line with findings in rodents, which indicate a perturbed antigen response following MIA in male but not in female mice (Carlezon et al., 2019).However, multiple factors may account for these sex differences.First, there may be a statistical power issue.Study populations of ASD are highly male-skewed due to an increased incidence of ASD in males and/or underdiagnosis of ASD in females (Halladay et al., 2015).These Table 3 List of immune genes associated with ASD in genotype analyses with description of their immunological and neurodevelopmental function.Abbreviations: NF-KB: nuclear factor kappa-light-chain enhancer of activated B cells; TLR: toll-like receptor; MIF: macrophage migration inhibitory factor; Ca2+: calcium; IL: interleukin;

Table 4
Immune gene and gene pathways with altered expression in the peripheral blood of autistic individuals.The key immune genes in pathway are reported along with information about their immune and neural functions.Arrows indicate the direction of regulation of given pathways as observed in ASD (↑= increased; ↓= decreased).populations may be, therefore, underpowered to detect significant effects in females.Second, there may be an influence of other sex-related factors.It is, for example, recognized that sex hormones modulate the immune response (Roved et al., 2017).Androgensand particularly testosteronedisplay immuno-suppressive properties, while estrogens act as enhancers and regulators of the immune response (Roved et al., 2017).Given their immunoregulatory role, estrogens may counteract the effect of immunogenetic variations in autistic females.Conversely, testosterone may exacerbate the effects of immune gene variations.One major hypothesis of ASD links intra-uterine testosterone to the onset of autistic behaviors (James, 2014).In this context, our analysis suggests that immunogenetic factors may intervene in the potential relationship between ASD and testosterone, or sex hormones in general.Last, it is important to note that the X chromosome hosts the largest number of immune genes, which therefore makes males more sensitive to the effect of variations affecting any of these X-linked immune genes (Schurz et al., 2019).
Previous studies also indicate that immunogenetic variations may be associated with specific clinical determinants such as regressive autism (Tamouza et al., 2020).In regressive forms of autism, gastrointestinal tract dysfunctions are central (Bennabi et al., 2015;Hughes et al., 2018).By studying the distribution of HLA haplotypes in individuals with and without regression, we identified a protective HLA class II sub haplotype namely, the HLA DPA1-DPB4 (Bennabi et al., 2018).Of interest, this haplotype has also been reported to be protective against intestinal autoimmune disorders such as celiac disease.In line with this, prior work described a positive association between these 'protective' HLA haplotypes and cognitive performance, and a diagnosis of Asperger's disorders (Bennabi et al., 2015(Bennabi et al., , 2018;;Debnath et al., 2018).Consistent with these protective effects, some HLA haplotypes have been regarded as beneficial to human evolution due to their constitutive immunosurveillance role and maintenance of brain homeostasis and preservation of blood-brain-barrier (Debnath et al., 2018).
Although further research is needed to clarify these effects, it is possible thatconsidering the clinical heterogeneity of ASDimmunogenetic variability is confined to specific clinical subgroups.The hypothesis of immune clinical subtypes has been reviewed previously (Jyonouchi et al., 2014) and it is reinforced by the evidence that immune dysregulations are observed in only a portion of autistic individuals (Zerbo et al., 2015).It is crucial to explore this possibility as it may help to define individuals that could potentially benefit the most from interventions aimed at minimizing the likelihood of immune (over-)activation and targeting ongoing immune dysregulations.
Current limitations and potential guidelines for the future.
Our review exposed a number of the limitations of previous immunogenetic research in ASD.First, although prior studies investigated key genes (e.g., HLA alleles) of the immune system (Bennabi et al., 2018;Guerini et al., 2018a;Guerini et al., 2015;Sayad et al., 2018), they did not cover the entire range of immune mechanisms that may influence the autistic phenotype.For a fuller picture, it would be useful to extend immunogenetic investigation in ASD to other, functionally diverse, immune genetic factors.This may ultimately clarify which immune mechanisms are clinically relevant.Second, previous research was mostly based on the samples of European descendants and male individuals.As discussed, it is important to include a wide range of ethnical groups (Ramos et al., 2015;Roved et al., 2017), and potentially investigate immunogenetic factors in autistic men and women separately.Furthermore, previous studies did not systematically assess the presence and/or family history of immune conditions in the examined population.Given the heterogeneity of both ASD and the immune response, detailed immune phenotyping of autistic individuals may help to refine the relationship between different classes of immune genes and autistic symptoms.For example, a recent deep analysis of both phenotype and functions of NK cells in adult with high functioning ASD allowed to identify phenotype specificities along with functional alteration highly suggesting the involvement of a yet to be identified immune trigger (Bennabi et al, 2019).Another limitation was the relatively poor assessment of the effects of age.Most studies include individuals of different age groups, spanning from early childhood to adulthood.Although this allows us to understand the life-long impact of immunogenetic variability, it is likely that, through life, other biological mechanisms intervene to counteract the effect of immune genetic variations.This may be especially the case for transcriptional changes that are known to be dynamic and tuned to environmental signals (M.Li et al., 2018).Future research should, therefore, investigate the effects of immunogenetic variations and immune gene transcription in clinical groups of different ages, for example adopting cross-sectional designs or even longitudinal approaches where possible.This is extremely important, especially considering the neurodevelopmental role of immune genes in early life.Moreover, it is important to acknowledge that external factors like stress, lifestyle, health status and medical history have considerable weight on transcriptional signals across tissues.In our review we selected well-powered studies which relied on standardized protocols and accounted for potential confounders, and often included replication samples.Nonetheless, follow-up transcriptomic research is warranted to validate gene expression findings in ASD.Finally, given the neurodevelopmental involvement of immune genes, it is crucial to investigate the influence of immunogenetic factors on the brain structure and functions in autistic individuals, also longitudinally.This may help to answer the question of whether neural (dys)functions bridge between immunogenetic factors and behavioral alterations.

Conclusions and clinical implications
Genetic factors are likely to be one of the mediators in the relationship between the immune system and ASD.Immune genes influence the autistic phenotype via inherited variations and/or changes in the expression levels of genomic products.These immune genes participate in key neurodevelopmental processes and show upregulation during key stages of neurodevelopment, such as gestation.These findings have valuable clinical implications.First, they may support strategies to optimize outcomes.The findings of increased immunogenetic tuning/ sensitivity in the prenatal and perinatal period highlight the importance of ensuring a (immune-)protected maternal and fetal environment during these developmental windows.Also, they highlight the importance of integrating medical history and clinical assessment of mothers to identify women carrying a higher load of immunogenetic variations and who may therefore be more susceptible to exaggerated immune activation during pregnancy (e.g., viral exposure).These women, for instance, may benefit from preventive strategies, such as protection to common allergens or inactivated vaccine (e.g., inactivated seasonal flu vaccine).Additionally, our findings highlight the importance of integrating clinical observations in childhood with the systematic recording of familial history and episodes of immune disturbances.This information may help to define subgroups of children with a higher chance of immune dysregulation, and to further explore whethergenetic predisposition toimmune dysregulations in these children may be a precursor to behavioral and cognitive alterations typical of ASD.Last, our findings may inform novel intervention approaches.The identification of specific immunogenetic pathways associated with ASD may guide future clinical trials to test the efficacy of these pathways as putative intervention targets.
In sum, we emphasize the need for a more systematic investigation of immune genes in ASD.Previous studies have invested attention on wellcharacterized immune genes and genetic pathways.However, to accommodate the heterogeneity of the immune system and ASD, future research should extend to additional immune mechanisms and investigate these across different clinical profiles, sexes, and age groups.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 1 .
Fig. 1.Flow chart of the study selection based on the 4 steps: identification, screening, eligibility, and inclusion to qualitative synthesis.

Table 1
List of genotype-based studies exploring the association of immune-related genes and ASD and included in the qualitative synthesis.Description about the study sample and identified immune genes are provided together with the review assessment scores.

Table 2
List of transcription (RNA-sequencing)-based studies investigating the expression of immune genes in both blood and post-mortem brain of autistic individuals, included in the qualitative synthesis.Description about the study design and identified immune genes are provided together with the review assessment scores.
(continued on next page) M.Arenella et al.

Table 5
List of immune genes pathways with altered expression in the post-mortem brain of autistic individuals.The key immune genes in pathway are reported along with information about their immune and neural functions.Arrows indicate the direction of regulation of given pathways as observed in ASD (↑= increased; ↓= decreased).