Brain growth trajectories in mouse strains with central and peripheral serotonin differences: relevance to autism models.

The genetic heterogeneity of autism spectrum disorders (ASDs) suggests that their underlying neurobiology involves dysfunction at the neural network level. Understanding these neural networks will require a major collaborative effort and will depend on validated and widely accepted animal models. Many mouse models have been proposed in autism research, but the assessment of their validity often has been limited to measuring social interactions. However, two other well-replicated findings have been reported in ASDs: transient brain overgrowth in early postnatal life and elevated 5-HT (serotonin) levels in blood platelets (platelet hyperserotonemia). We examined two inbred mouse strains (C57BL/6 and BALB/c) with respect to these phenomena. The BALB/c strain is less social and exhibits some other autistic-like behaviors. In addition, it has a lower 5-HT synthesis rate in the central nervous system due to a single-nucleotide polymorphism in the tryptophan hydroxylase 2 (Tph2) gene. The postnatal growth of brain mass was analyzed with mixed-effects models that included litter effects. The volume of the hippocampal complex and the thickness of the somatosensory cortex were measured in 3D-brain reconstructions from serial sections. The postnatal whole-blood 5-HT levels were assessed with high-performance liquid chromatography. With respect to the BALB/c strain, the C57BL/6 strain showed transient brain overgrowth and persistent blood hyperserotonemia. The hippocampal volume was permanently enlarged in the C57BL/6 strain, with no change in the adult brain mass. These results indicate that, in mice, autistic-like shifts in the brain and periphery may be associated with less autistic-like behaviors. Importantly, they suggest that consistency among behavioral, anatomical, and physiological measures may expedite the validation of new and previously proposed mouse models of autism, and that the construct validity of models should be demonstrated when these measures are inconsistent.