Phospho-tau immunoreactivity is increased in temporal cortex of Tg2576/Irs2^−/− mice. (A) Phosphospecific anti-tau antibodies were used to probe temporal cortex of age-matched wild-type (WT); Tg2576; Irs2^−/− and Tg2576/Irs2^−/− animals. (B) Data shown is normalised to a phosphorylation-independent tau antibody. (C) Analysis by genotype (ANOVA with post hoc Tukeys test; n = 29) showed significant increases in tau phosphorylation in Tg2576/Irs2^−/− animals relative to WT and Tg2576 animals at epitopes recognised by antibodies MC6, CP13, PG5, PHF1, and TG3 but not at the overlapping AT8 and TAU1 epitopes.

Deletion of Irs2 reduces Aβ deposition and decreases insoluble Aβ. (A) Immunocytochemical labeling with a human specific anti-Aβ antibody of whole brain frontal cortex sections of 12 month-old Tg2576 and Tg2576/Irs2^−/− mice. Scale bar = 200 microns. (B) Immunocytochemical with the anti-Aβ antibody (left panels) and histochemical labeling with Congo red (right panels) of temporal cortex sections of 12 month-old Tg2576 and Tg2576/Irs2^−/− mice, showing the extent of Aβ deposition in representative animals. Scale bars = 50 microns. (C) The mean surface covered by Aβ deposits in the cortex was significantly reduced in Tg2576/Irs2^−/− mice compared with Tg2576 mice (n = 32; ^∗p = 0.01, t-test). (D) APP processing was examined by immunoblotting for holo-APP and APP-CTFs in temporal cortex of Tg2576 and Tg2576/Irs2^−/− animals. Holo-APP values were normlised to actin, CTF values were normalised to holo-APP. No differences were found between genotypes. (E) Human Aβ_1–40 and Aβ_1–42 were measured by ELISA in S1 (soluble) and S2 (detergent soluble) fractions and formic acid extracts of P2 pellets from temporal cortex. Significant reductions in Aβ_1–40 and Aβ_1–42 were found in formic acid extracts from Tg2576/Irs2^−/− compared to Tg2576 animals (n = 21; ^∗p = 0.0005 [Aβ_1–40] and 0.002 [Aβ_1–42], t-test).

GSK-3 is inhibited and PP2a is reduced in mice lacking Irs2. The tau-kinase, GSK-3 and the protein phosphatases PP1 and PP2A were examined in wild-type (WT) and in IRS2−/− mice by Western blot. (A) Animals lacking Irs2 showed no change in total GSK-3 protein or in phosphorylation at the Tyr279 in GSK-3α/GSK-3β 216 epitope. However there was a substantial (p < 0.05) increase in GSK-3 phosphorylation at Ser 21 GSK-3α/Ser 9 GSK-3β epitope reflecting relative inhibition of GSK-3 activity in these animals. (B) There was no change in PP1 but a significant increase in PP2A in animals lacking Irs2. (C) Comparing PP2a in hippocampus across all four genotypes confirmed a reduction in PP2a in all animals lacking Irs2 but no effect of APP expression.

Increase in insulin degrading enzyme and in transthyretin in mice lacking Irs2. Amyloid turnover is regulated by proteolysis through proteases such as insulin degrading enzyme (IDE) and by increased clearance following binding to proteins such as transthyretin; both known to be altered in response to insulin signalling. IDE mRNA was unaltered in IRS2−/− mice (A) but protein levels in the membrane-bound fraction were modestly increased (B). Transthyretin mRNA was increased substantially (3.9-fold; p < 0.02; SEM 1.05) in both genotypes lacking Irs2.