647 Basis for the link between atopic dermatitis and autism

Abstract

A subset of patients with autism spectrum disorders (ASD) have a prior history of atopic dermatitis (AD). Moreover, the prevalence of ASD increases dramatically as AD phenotypes worsen. In a standard ASD animal model (parenteral administration of valproic acid [VPA] to second trimester pregnant mice), we explored the basis for the baseline association, and second, for the AD phenotype-dependent increase in ASD prevalence. At birth, VPA-treated mice demonstrated extensive neuro- and epidermal cytotoxicity, but no evidence of inflammation. Yet, by one-day post-partum, cutaneous inflammation became apparent, while in contrast, neuroinflammation did not appear until approximately 4 days post-partum. These histopathologic changes were paralleled by alterations in cutaneous, as well as brain lipid content and ultrastructural morphology, which mimicked comparable, well-known abnormalities in AD. The initial emergence of cytotoxicity was paralleled by elevated levels of TNFα, IL-17 and IFNγ in both tissues, but Th2 markers appeared initially in skin, and only later in brain, suggesting a skin⇒brain putative sequence. We conclude first, that the parallel susceptibility in both the brain and epidermis of VPA-induced cytotoxicity reflects their shared origin in the primitive neuroectoderm. Second, that the AD phenotype-‘driven’ increase in ASD prevalence reflects the prior development of Th2 inflammation in the skin, linking the additional, superimposed burden of xeric stress that occurs at birth, further amplifying cutaneous cytokine production. These studies could also explain why evidence of AD (1-3 months) precedes the earliest evidence of ASD (≥ 6 months). Finally, if ASD represents another end-point of the ‘atopic march;’ then it is possible that early treatment of subjects at risk for AD could prevent/mitigate the downstream development of ASD.