Abstract
BACKGROUND: Autism, a highly heritable psychiatric condition, is hypothesised believed to be associated with altered synaptic structure and function, as indicated by findings from genetic, post-mortem, and pre-clinical studies. AIMS & OBJECTIVES: In this study, we used positron emission tomography (PET) to investigate synaptic density in vivo. Utilizing a twin design, our objective was to isolate autism specific variation. METHODS: We examined seven twin pairs discordant for autism (aged 22-29; three male, two female, and two mixex pairs; five monozygotic, two dizygotic pairs; five qualitatively, two quantitatively discordant pairs), alongside eight independent controls (aged 20-27). Radioligand [11 C]UCB-J binds to synaptic vesicle 2A (SV2A). To estimate synaptic density, we employed volume of distribution (VT) using radioactivity measured in blood and binding potential (BPND) with SV2A devoid centrum semi-ovale (CSO) as reference region. To compare the twin pairs, we conducted paired t-tests; for the entire sample, we ran generalized estimating equation (GEE) models. RESULTS: Neither the t-tests nor the GEE models revealed statistically significant differences in synaptic density between autistic twins and their co-twins. While we observed lower synaptic density in autistic twins compared to controls in some brain regions using BPND, no significant differences were observed using VT. Although not statistically significant, we did observe group differences in VT of CSO (which served as the reference region). DISCUSSION & CONCLUSION: We did not find any differences in synaptic density between autistic twins and their non-autistic (or less autistic) co-twins. The intra-pair similarities imply that although synaptic density is highly heritable, it may reflect genetic liability for autism, rather than being associated with an autistic phenotype display. We did not find any differences in VT between groups, whereas our collaborators did observe lower mean VT in the autistic group, compared to controls (preliminary findings, presented as a poster, not published). When modelling synaptic density with CSO as a reference region, differences in BPND were found between autistic twins and non-autistic controls. Noteworthy, this analysis rests on the premise that there should be no discrepancies in the reference region between the groups. However, this assumption is contravened by the presence of the observed distinctions in CSO VT between the groups, as well as documented variations in white matter within autistic brains4. Consequently, it is plausible that the observed differences in BPND do not directly relate to differences in synaptic density, but rather reflect variations in white matter composition (and thus the behaviour of the radiotracer within the CSO). This consideration underscores the importance of cautious interpretation of results and accounting for potential confounders when interpreting differences in synaptic density. REFERENCES: 1. Tang et al. Neuron (2014) 2. Bagni et al. Neuron (2019) 3. Bourgeron et al. Nature Reviews Neuroscience (2015) 4. Faraji et al. Pscyhiatry Research: Neuroimaging (2023)