Comment on “Autistic-like phenotypes in Cadps2-knockout mice and aberrant CADPS2 splicing in autistic patients”

Sadakata et al. (1) reported that a CADPS2 isoform lacking exon 3 is aberrantly spliced in the peripheral blood of autistic patients. However, we found this splice isoform in the blood of normal subjects at a similar frequency to that of individuals with autism spectrum disorder (ASD) (95% CI of the difference, –0.06 to 0.1). Moreover, this splice variant exists as a minor isoform in cerebellar RNA of both normal individuals and individuals with ASD. Thus, exon 3 skipping likely represents a minor isoform rather than aberrant splicing and is probably not an underlying mechanism of autism. Defects of CADPS2 function might contribute to autism susceptibility, but likely not through aberrant splicing.

Sadakata et al. (1) reported that 4 of 16 patients with autism expressed an exon 3–skipped variant of CADPS2 mRNA in the blood, while the CADPS2 mRNA of all 24 normal subjects included exon 3. They thus concluded that CADPS2 is aberrantly spliced in autism, and they performed further experiments showing that the subcellular localization of exogenously expressed exon 3–skipped CADPS2 is disturbed in primary cultured neocortical and cerebellar neurons.

We aimed to replicate the CADPS2 findings in an independent set of peripheral blood samples from 41 children with ASD and 39 control children, following the Sadakata et al. protocols (Figure 1A). Furthermore, we performed sequencing (Figure 1B) and nested priming (Figure 1C) to validate the presence or absence of exon 3. Our results showed that, of 39 control samples, 1 was apparently homozygous for the exon 3–skipped allele in peripheral blood, 5 were heterozygous, and 33 were wild type. Of the 41 ASD samples, 5 were heterozygous for the exon 3–skipped isoform, while the rest were wild type. Analysis of these results showed no significant difference in the frequency of the exon 3–skipped allele in ASD versus control samples (P = 0.6, two-proportion z test). Although the samples tested here might differ from those tested by Sadakata et al. in their ethnicity, gender, or age distributions (Supplemental Figure 1 and Supplemental Tables 1 and 2; supplemental materials available online with this article; doi: 10.1172/JCI38620DS1), the finding of exon 3 skipping in healthy controls at a high frequency suggests that this isoform does not represent aberrant splicing and likely is not a mechanism underlying autism.

Figure 1. Exon 3 skipping in CADPS2 mRNA from 41 children with ASD and 39 control children.

(A) RT-PCR of CADPS2 mRNA in blood from subsets of patients with ASD (A1–A14) and control patients (C1–C10) following the Sadakata et al. protocols (1). The 661-bp band represents the full-length exon 1–5 fragment of CADPS2 mRNA, while the 328-bp band is a result of exon 3 skipping. Four control samples (C1, C2, C6, and C8) and 4 ASD samples (A3, A7, A8, and A9) were heterozygous for the exon 3–skipped isoform. The flanking marker is a 50-bp ladder. The remaining samples showed only the 661-bp band (data not shown). (B) Alignment of sequences obtained from the 328-bp bands of samples C1, C2, C6, and A9 to human chromosome 7 showed that all sequences lacked exon 3. Sequencing the 661-bp band of A10 (which was representative of other samples not showing the 328-bp band) demonstrated that this fragment does include exon 3, as expected. (C) RT-PCR of blood CADPS2 mRNA using a nested amplification. A single major band (563 bp), indicating the presence of exons 2–5, is shown in all autistic samples. Control sample C14 was apparently homozygous for a 230-bp band that resulted from skipping of exon 3. (D) RT-PCR of cerebellar CADPS2 mRNA from individuals with ASD and control individuals showed that all cerebella contained the exon 3–skipped splice variant as a minor isoform (230-bp fragment). M, low-DNA-mass ladder (Invitrogen).

Since Sadakata et al. extrapolate function of the exon 3–skipped isoform within the cerebellum, we additionally tested the presence of exon 3 in mRNA extracted from the cerebella of 9 control children and 5 children with ASD. All ASD and control samples were found to contain the exon 3–skipped splice variant as a minor isoform (Figure 1D).

Thus, our experiments suggest that exon 3 skipping represents a normal, minor isoform of CADPS2 in the cerebellum. As we observed no difference in prevalence of this allele between ASD and control samples, we conclude that exon 3 skipping is likely not a mechanism underlying autism susceptibility or pathogenesis.