ELMOD3‐SH2D6 gene fusion as a possible co‐star actor in autism spectrum disorder scenario

Abstract Autism spectrum disorder (ASD) is a group of neurodevelopmental disorders characterized by high heritability. It is known that genetic factors contribute to ASD pathogenesis. In particular, copy number variants (CNVs) are involved in ASD susceptibility and can affect gene expression regulation. 2p11.2 microdeletions encompassing ELMOD3, CAPG and SH2D6 genes have been described in four unrelated ASD families. The present study revealed that this microdeletion is responsible for the production of a chimeric transcript generated from the fusion between ELMOD3 and SH2D6. The identified transcript showed significantly higher expression levels in subjects carrying the deletion compared to control subjects, suggesting that it is not subjected to nonsense‐mediated decay and might encode for a chimeric protein. In conclusion, this study suggests the possible involvement of this gene fusion, together with the other previously identified variants, in ASD.

Gene fusions are well studied in cancer genetics, as chimeric genes have an important role in different cancers. They can occur when CNV breakpoints disrupt two genes encoded on the same chromosomal strand.
Some studies have previously investigated whether CNVs generating fusion transcripts may have an impact on ASD susceptibility 6 obtaining negative results. However, fusion transcripts were not detected for 80% of CNVs tested and the only CNV resulting in a fusion transcript was present in ASD and control subjects at similar rate. 6 In another study, Pagnamenta et al detected a DOCK4-IMMP2L fusion transcript in both ASD subjects and their non-affected family members. This fusion transcript was expressed at very low levels compared to the wild-type transcript. In fact, IMMP2L portion of the fusion gene was out of frame leading to a premature stop codon and likely to nonsense-mediated decay of the fusion transcript. 7 Another indirect way, by which CNVs can alter gene expression, is the disruption of a gene regulatory landscape. Gene transcription is regulated by cis-acting elements that can be separated from their target gene by the coding sequences of neighbouring genes. 8 We conducted a comprehensive genetic analysis of a Sardinian family (AUT003) consisting of two siblings affected by ASD and their unaffected parents. 9 Genetic studies conducted on Sardinian population have been proven fundamental for the identification of genetic loci linked to several complex diseases. [10][11][12][13] We identified a rare heterozygous 2p11.2 deletion, including the last coding exons of ELMOD3, the entire CAPG gene and the first non-coding exon of SH2D6, in both ASD siblings inherited from their unaffected mother ( Figure 1A). Interestingly, two similar heterozygous 14

| RE SULTS
We evaluated SH2D6 gene expression in AUT003 family and 17 healthy individuals using two different assays as described in Materials and Methods. The first assay did not reveal any detectable SH2D6 product in all the tested samples. Since the forward primer maps on the deleted region, the absence of SH2D6 expression in the individuals with the deletion implies that the wild-type SH2D6 locus is not transcribed. In fact, SH2D6 is expressed at very low levels in whole blood (GTEx data, Figure 2A). The second assay revealed that, as expected, SH2D6 was undetectable in all subjects without the deletion, including the unaffected father (AUT003.1) ( Figure 1B). Interestingly, a significant increase of SH2D6 mRNA ex-   ). B, SH2D6 regulatory region analysis using chromatin state segmentation analysis data from ENCODE/Broad, CpG methylation data from ENCODE/HAIB and GeneHancer data A PCR product of the expected size was obtained for ASD probands and their unaffected mother while, as expected, no amplification was detected in controls.
The expression levels of this fusion transcript (evaluated using primers mapping in ELMOD3 exon 10 and SH2D6 exon 2-3) were compared to that of SH2D6 (SH2D6 exon 13 and exon 14 primers, PCR product: 99bp) in deletion carriers by qRT-PCR. Delta Cts ( Figure 1C) revealed that transcript expression levels detected using both assays do not differ implying that the increased expression of SH2D6 in subjects with the deletion is only due to this fusion transcript ( Figure 1C and 1).
Sequencing analysis confirmed the presence of this fusion transcript in both ASD siblings and their unaffected mother ( Figure 1E and 1). ExPASy Translate tool predicted that the cDNA sequence re- did not reveal any protein product since the antibody recognizes an epitope which is not present in the predicted chimeric protein sequence (data not shown).

| D ISCUSS I ON
The complex ASD phenotype implied that multiple gene variants participate to ASD phenotypic manifestations. Previous studies have investigated the presence of fusion transcripts in ASD. 6,7 However, the detected fusion transcripts were present at similar levels in both ASD patients and healthy individual or were expressed at much lower levels compared to wild-type transcript suggesting that they are subjected to nonsense-mediated decay.
To our knowledge, our study is the first that highlights the possible contribute of a chimeric gene in ASD phenotype. In fact, the so evident and detectable expression of the transcript produced by the identified fusion gene in the cells of deletion carriers suggests that this transcript does not undergo mechanisms of nonsense-mediated decay. Further investigations will be necessary to reveal the possible translation of a chimeric protein and its possible function, perhaps contributing together with other variants (as already suggested in our previous work, 9 ) to generate the ASD phenotype.

ACK N OWLED G EM ENTS
We gratefully acknowledge the AUT003 family and all the subjects who have participated in the study. We would also acknowledge Prof. Andrea Perra for the blood draws of all the control subjects.

CO N FLI C T O F I NTE R E S T
The authors declare no potential conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data generated in this study are available from the corresponding author on reasonable request.