MicroRNA Cross-Involvement in Autism Spectrum Disorders and Atopic Dermatitis: A Literature Review

Autism Spectrum Disorder (ASD) is a category of neurodevelopmental disturbances seriously affecting social skills, to which the scientific community has paid great attention in last decades. To date, their pathogenesis is still unknown, but several studies highlighted the relevance of gene-environment interactions in the onset of ASD. In addition, an immune involvement was seen in a wide number of ASD subjects, leading several researchers to hypothesize a possible common pathogenesis between ASD and immune disturbances, including Atopic Dermatitis (AD). In general, among potential contributing factors, microRNAs (miRNAs), small molecules capable of controlling gene expression and targeting mRNA transcripts, might represent one of the major circulating link, possibly unraveling the connections between neurodevelopmental and immune conditions. Under such premises, we conducted a systematic literature review, under the PRISMA guidelines, trying to define the panel of common miRNAs involved in both ASD and AD. The review retrieved articles published between January 1, 2005, and December 13, 2018, in PubMed, ScienceDirect, PsycARTICLES, and Google Scholar. We found a handful of works dealing with miRNAs in ASD and AD, with the most overlapping dysregulated miRNAs being miR-146 and miR-155. Two possible compounds are abnormally regulated in both ASD and AD subjects, possibly cross-contributing to the interactions between the two disorders, setting the basis to investigate more precisely the possible link between ASD and AD from another, not just clinical, perspective.


Introduction
Autism spectrum disorder (ASD) is a heterogeneous group of neurodevelopmental disorders characterized by impairments in social interaction and communication and restricted or stereotyped interests and behaviors [1], typically occurring before the fourth year of life [2]. In the last decades, there has been an increased attention of the scientific community towards this group of conditions, with a consequent improvement in the knowledge about this topic, and contextually, a broadening of the diagnostic criteria [3,4].
Nonetheless, ASD is the most heritable among neuropsychiatric disorders, with genetic contributions accounting for around 80% of ASD risk [5][6][7][8][9], and higher risk is seen in siblings of autistic children [10]. Although epidemiological studies provide information on the genetic contribution to

General Insight into MicroRNAs
It is known that miRNAs are very short (18-25 nucleotides), single-stranded non-coding RNAs, able to control gene expression and to target mRNA transcripts, possibly bringing on their translational degradation or their repression, with particular degrees of complementarity [40]. The targeting of mRNA transcripts by miRNA occurs since one miRNA is able to target a number of mRNA transcripts; conversely, a single mRNA transcript can be targeted by many miRNAs.
Actually, miRNAs have rapidly induced a great interest in humans, being potential biomarkers for diagnostic and prognostic aims [41,42], with the number of classified miRNA increasing to over 2500 potential molecules in the Homo Sapiens genetic makeup [40]. Despite being small molecules not capable of encoding proteins, miRNAs hold important structural, regulatory, and catalytic functions.
The miRNA genes are located in the introns of protein-coding genes or in independent non-coding DNA loci [43], whereas nearly half of the total miRNAs are pooled on chromosomes with a common promoter [44].
The biogenesis of miRNA is extremely complex, consisting of several phases. These include: (i) in the nucleus, the transcription of miRNA genes into primary miRNA transcripts by RNA polymerase II; (ii) the freeing of pre-miRNA hairpin, through the trim of the primary miRNA transcripts by the RNAse III Drosha endonuclease; (iii) the active exportation of the pre-miRNA hairpin out of the nucleus in a process involving the nucleocytoplasmic shuttler Exportin-5; (iv) the final maturation, in the cytoplasm, processed by Dicer RNase III endonuclease, splitting the pre-miRNA into a single-stranded mature miRNA [45]; (v) the binding of the mature miRNA to proteins of the Ago family, and (vi) the assembly of the RNA-induced silencing complex (RISC) together in order to employ its physiological functions.
The mature miRNA, once incorporated into the RISC, induces post-transcriptional gene silencing by binding RISC to be partially complementary to the target mRNA found mainly within the 3'-untranslated region (UTR) [46,47].
Flaws in miRNA expression deeply affect several pathways related to cell regulation, including apoptosis, stress responses, or cell proliferation throughout the human body [48][49][50]. Indeed, a single miRNA could repress around 100 mRNAs, while around 60% of human protein coding-genes are represented by conserved targets of miRNAs, thus a number of mRNA targets are regulated by miRNAs [51].

MiRNAs Linked to Brain Function
MiRNAs approximately regulate two-thirds of human mRNAs [51], and are as much as 70% expressed in the central nervous system (CNS), including the brain and spinal cord [52,53]. Their changes during childhood are different depending on the affected brain region [54].
In particular, miRNAs are abundant in neurons and glia, often placed at the synaptic level, and able to regulate the structure of the dendritic spine, as happens with miR-134 which reduces spine growth by targeting spine growth-promoting kinase Limk1 [55].
Indeed, dendritic spines are bulges on a dendritic tree of a neuron, composing the post-synaptic termination of a synapse, reflecting-through their structure-the degree of brain maturation, and, somehow, brain plasticity.
Their density is abnormal in several conditions, including schizophrenia (dendritic spines loss) and ASD, the latter featuring an increase in the quantity of spines in specific brain areas [56][57][58][59].
Overall, several miRNAs affect brain functions and development, neuronal plasticity, maturation, and differentiation [37,64]. Dysregulation of miRNA expression is particularly frequent within several neurological disorders, including ASD, therefore the association between some common miRNA families and ASD, despite still largely unknown, is nowadays clearer than in the past.

MiRNAs and Skin Disorders
Recently, much evidence has been published about the role of miRNAs in several cellular processes, including immune response, DNA repair, apoptosis, proliferation, and differentiation [65], but also in morphogenesis, differentiation, wound healing, psoriasis, and AD [66][67][68]. Specifically, AD pathogenesis is also associated with a complex gene-environment interaction, as well as with an alteration of the skin barrier function, and a deregulation of the immune system [69]. Several miRNAs, including miR-146a, miR-155, miR-203, and miR-483-5p, are also differentially expressed in AD and in other immunologic and inflammatory disorders.

Materials and Methods
A literature review of the articles published between 1 January 2005, and 13 December 2018, was conducted in PubMed, ScienceDirect, PsycARTICLES, and Google Scholar following the PRISMA guidelines.
Overall, the search was limited to articles describing studies conducted on humans published in peer-reviewed journals. After having discarded duplicates, the obtained results were sorted by relevance and the most significant works related to ASD and miRNAs and to AD and miRNAs were selected. We will first present the results from the literature review and then discuss the possible associations between miRNAs in ASD and AD according to such findings.

Results
The literature search displayed 26 articles directly related to the relevant topics ( Figure 1).

Studies about ASD
According to the literature review, a number of miRNAs were found to be associated to ASD (Table 1). Specifically, the most overlapping dysregulated miRNAs appeared to be let-7, miR-19b, miR-23, miR-106, and miR-146.

Studies about AD
A few works studied microRNAs involvement in AD (Table 2). Here, the main dysregulated miRNAs are miR-146, miR-155, and miR-203.
Summarizing, the association between ASD and AD revealed a common unbalance for miR-146 and miR-155.

Studies about ASD
According to the literature review, a number of miRNAs were found to be associated to ASD (Table 1). Specifically, the most overlapping dysregulated miRNAs appeared to be let-7, miR-19b, miR-23, miR-106, and miR-146.

Studies about AD
A few works studied microRNAs involvement in AD (Table 2). Here, the main dysregulated miRNAs are miR-146, miR-155, and miR-203.

The Overlap between Atopy and Autism
Recent data strongly support the clinical association between atopy and ASD [21,22]. It has been widely demonstrated that allergic diseases, especially food allergies, are more frequent among ASD children [95,96].
Notably, a large observational study, comparing 14,812 atopic subjects with 6944 non-atopic subjects, with no lifetime atopic disease, highlighted a strong association between atopy and the risk of developing ASD [97]. Furthermore, autoimmune disorders, including psoriasis (2-fold risk), are also frequently identified in ASD [98].
Beside the robust clinical evidence for the association between atopy and ASD, an intriguing neuroinflammatory hypothesis has been advanced for ASD, involving the disruption of the brain blood barrier induced by inflammatory molecules, brain mast cell activation, and mast cells-microglia interactions [99].
However, the hypothesis that the pro-inflammatory cascade induced by AD could lead to ASD in the presence of genetic susceptibility, is supported by the clinical association and by a shared pattern of cellular damage, with epigenetic changes as a common pathogenic mechanism.

Role for Overlapping MiRNAs in ASD and AD
The main literature finding concerning miRNAs in AD and ASD is represented by miR-146a. Upregulated in various neurodevelopmental disorders [82], miR-146a was reported to be highly expressed throughout the cortex, hippocampus, and amygdala, key structures for higher cognitive functioning [103]. Furthermore, it was demonstrated that reproducing abnormal miR-146a expression in mouse primary cell cultures leads to impaired neuronal dendritic arborization-producing shriveled dendritic trees with branching points at more proximal levels compared to controls, proving the defective neural connectivity typical of ASD-and to increased astrocyte glutamate uptake capacities [82], in turn modifying fast synaptic transmission at the CNS level. Further, miR-146a, expressed in the developing brain, is enclosed within neurons, with poor expression in the glial lineage in adult mice. However, it generally inhibits the expression of neuron-specific targets, including Nlgn1 and Syt1, preventing glial cells from mistakenly adopting neuron-specific phenotypes.
In addition, the neuron excitation at the cortex is probably affected by miR-146a deregulation through the involvement of potassium two pore domain channel subfamily K member 2 (KCNK2), having a key role in neural excitability and migration at the cortex level of developing mice, a critical issue in ASD.
Furthermore, miR-146a expression contributes to neuroinflammation in the brain of ASD subjects, having a role in immune system regulation.
Its function in the regulation of inflammatory processes could partially fill the gap between ASD (and neurodevelopmental disorders in general) and AD (and atopic conditions in extenso).
It is evident indeed that an increased miR-146a expression is present in the lesional skin of AD patients [104], as it inhibits nuclear factor κ B (NF-κB)-mediated proinflammatory cytokines and chemokines, bringing alleviation of the inflammation directly linked to AD and similar conditions [91].
In AD, during skin inflammation, miR-146a is increased in keratinocytes, controlling chronic inflammatory processes triggered by IFN-γ and the activation of NF-κB. Indeed, the relevance of miR-146a in inflammatory skin disorders is confirmed by evidence from psoriasis research [105]. Furthermore, the expression of miR-146a is strongly dependent on NF-κB, and the miRNA has been shown to suppress the NF-κB signaling pathway through a direct targeting of a number of compounds, including IL-1 receptor-associated kinase 1 (IRAK1), TNF receptor-associated factor 6 [106], v-rel avian reticuloendotheliosis viral oncogene homolog B (RELB) [107], and CARD10 [108].
Moreover, it was discovered that mice with a deficiency of miR-146a develop a late autoimmunity caused by an impaired activation of NF-κB in T cells, and signal transducer and STAT1 activator in regulatory T cells [109].
Of note, previous work demonstrated that both an enhanced opioidergic activity and reduced vitamin D levels could represent shared features of AD [110] and ASD [111], and possibly miR-146a and miR-155 could interact with the genetic milieu in subjects with these disorders. Indeed, both miR-146a and miR-155 have been tested in models of LPS tolerance and miR-146 was able to amplify the severity of morphine-mediated hyper-inflammation [112].
Finally, the regulatory role of miR-146a also occurs at lung alveolar epithelial cells, where the release of IL-8 and CCL5 occurs independently from IL-1β signaling.
Concerning miR-155, its role in ASD is not yet known, whereas in AD it appears to modulate T helper type 17 (Th17) cell differentiation and function [92], and to directly target the suppressor of cytokine signalling-1 (SOCS1) gene, taking part in a negative feedback loop to attenuate cytokine signaling [113]. Interestingly, miR-155 is also linked to inflammation and immunity, thanks to its potent upregulation in immune cell lineages, including lymphocytes, fibroblasts, macrophages, mast cells, and dendritic cells, in turn implicated in the pathogenesis of chronic skin inflammation [114][115][116][117][118][119]. Briefly, miR-155 seems to be also involved in the regulation of T-cell responses through a suppression of cytotoxic T-lymphocyte antigen 4 and by enhancing T-cell proliferation [66]. Its deregulation, seen as increased expression in peripheral CD4 T cells of AD patients, was correlated to disease severity, supporting its role in AD pathogenesis [92].

Conclusions
Beyond the clinical evidence [22,23], a possible, yet speculative, role for genetics (miRNAs in particular) can be hypothesized to justify the clinical association between AD and ASD.
However, both miR-146a and 155 appear to be involved in this common pathogenetic pathway, despite the role of the latter still being poorly known. Several other aspects differentiating these diseases remain elusive, including the identification of putative environmental injuries and the complex role of vitamin D in immune and neurologic disorders. Therefore, further studies focusing on the association between vitamin D and opiod receptors in skin and neurologic disorders should investigate the role of target genes for common dysregulated miRNAs, in order to discover specific overlapping features of these conditions. It remains evident that an inflammatory component is active in both diseases, and the actual data support future applications for miR-146a, both as a biomarker and as a target for therapy.
Interestingly, it can be speculated that a deregulation of miR-146a occurs earlier, during embryonic development, thus participating in the development of ASD. Apart from the well-described effect on NF-κB activity and the associated inflammatory pathways strictly linking miR-146a with AD, the deregulation of miR-146a and miR-155 could influence a wide range of their validated targets (Figure 2), essential for brain development and function. However, it remains complex to determine the major source (skin vs brain) of miR-146a and miR-155, and whether they are potentiating each other or having more organ-or disease-specific effects. These aspects warrant future larger longitudinal studies. Finally, a better understanding of the link between ASD and AD might be useful to investigate whether a specific miRNA could act as a biomarker for the risk of developing ASD for patients with AD, and potentially represent a target for ASD prevention.