Figure 4. APP and Aβ40 levels in ASD and FXS brains.

Brain samples were obtained as described in the text. Samples were processed and subject to Western blotting or ELISA as described in the text. In some cases, two brain regions were used for FXS samples, BA21 indicated by solid markers, BA10 indicated by hollow markers. Triangles indicate samples used in the two-brain-region assay. Samples were analyzed either by generalized mixed models (glmm), followed by ANOVA comparing the model with a null model. The glmm was used to take account for assay and brain region as random variables. Models were tested for outliers by Bonferroni-corrected method. No outliers were found in the models. Models with p ≤ 0.05 were further subject to simultaneous pairwise comparisons (Tukey’s). Letters indicate pairwise categories. (A) sAPPα levels differed between control and FXS, while ASD was intermediate between the two. (B) Total APP levels fell into two categories, with control being significantly lower than FXS, but no significant difference between control and ASD. (C) The ratio of sAPPα to total APP did not differ among diagnoses. (D) Levels of DEA-extracted (membrane-bound) Aβ were significantly lower in ASD samples than in either control or FXS. (E) Levels of soluble Aβ were significantly lower in ASD samples than in either control or FXS. (F) Levels of total (soluble + membrane-bound) Aβ40 per diagnosis.