Although there are multiple studies describing networks of mRNAs, miRNAs, and lncRNAs to find autism-associated non-coding genes, these characterized genes are only the tip of an iceberg, as most of the non-coding RNAs, including lncRNAs, have not been investigated so far.
Based on SFARI data, we constructed an autism-specific mRNA-miRNA-lncRNA network revealing a complex regulatory relationship between different types of genes. Using a bioinformatics approach, four candidate lncRNAs, comprising Gm10033, 1500011B03Rik, A930005H10Rik, and Gas5, were screened out for further analysis. So far, there have been no detailed reports on studying these lncRNAs in the context of autism. KEGG pathway and GO analysis indicated that mRNAs and miRNAs interacted with the candidate lncRNAs mainly function in synapse organization, synaptic transmission, synapse assembly, and regulation of nervous system development. Based on KEGG pathway analysis, several pathways were enriched, including glutamatergic synapse, Wnt signaling pathway, neurotrophin signaling pathway, Hippo signaling pathway, and PI3K-Akt signaling pathway which were previously reported as dysregulated in autism [42-45]. Furthermore, concerning the disease ontology enrichment, we revealed the association between mRNAs in the network with PDD and SDD. According to the previous studies, PDD refers to the group of heterogeneous conditions that formed a continuum disability with sharing a core triad symptom including qualitative impairments in reciprocal social interactions, verbal and nonverbal communication, and repetitive and stereotyped patterns of interests and behavior which autism is classified as a subset to these group of disorders [46,50]. Moreover, several symptoms that are seen in patients or mouse models of autism including delays in developmental domains such as motor function, language and speech development, and uncanny intelligence or poor educational conditions, are coordinated to SDD definition [51]. Therefore, these candidates’ lncRNAs might play important and functional roles by interacting with key mRNA(s) and miRNA(s) that in turn may contribute to autism.
Exposure to valproic acid, a widely used anticonvulsant drug, during pregnancy has been associated with an increased incidence of autism [47]. VPA-exposed rodents showed the core behavioral signs of autism as well as molecular changes associated with the disorder [48]. In the current study, reduced sociability and social novelty, increased locomotor activity, self-grooming, anxiety-like behaviors, and impairment of long-term memory in VPA-exposed mice confirmed the validity of the animal model.
Then, we speculated that the four candidates lncRNAs initially screened out from a three-component regulatory network may show a link with autism because our enrichment analysis predicted that these lncRNAs and their associated mRNAs and miRNAs are involved in multiple autism and neurological pathways. Therefore, we performed a quantitative real time RT-PCR to examine the expression of these four candidate lncRNAs in the frontal lobe and cerebellum of VPA-exposed mice. Increased expression of Gm10033, 1500011B03Rik, A930005H10Rik, and Gas5 were observed in both frontal lobe and cerebral brain samples of VPA-exposed mice. Although differential expression of these lncRNAs was previously reported in other physiological and pathological states [49-51], there are no reports of involving these lncRNAs in autism or other neurological disorders. However, decreased expression of Gas5 was reported in the whole brain tissue of mouse embryos treated with VPA on GD 8, as a chemically induced neural tube defects (NTDs) mouse model [52]. As several differences exist in our experimental design with theirs, such as VPA dose, day of VPA injection, and the tissue investigated, the results cannot simply be compared between these two studies.
Moreover, in the present study, we identified a novel splice variant of Gm10033, designated as Gm10033-ΔEx2. Expression of Gm10033-ΔEx2 was observed in various mouse tissues, including liver, kidney, heart, lung, testis, and brain. The presence of distinct variants of lncRNA with unknown functions, Gm10033, suggests that it may play a more complex regulatory role as lncRNA in a wide range of biological processes.
By utilizing an autism-specific three-component network comprising mRNAs, lncRNAs, and miRNAs, we have identified a list of four candidate lncRNAs. These lncRNAs and their associated mRNAs and miRNAs were found to be associated with neurologically relevant pathways and functions. Experimental validation of Gm10033, 1500011B03Rik, A930005H10Rik, and Gas5 in the brain of VPA-exposed mice introduces them as potentially novel autism-associated lncRNAs. Further studies are required to unravel the functional role of these lncRNAs in the pathobiology of autism. The present work suggests opportunities to define subgroups of patients within a large heterogeneous clinical category and to examine common treatment targets across distinct neurodevelopmental trajectories [53].