Fig. 4

Photoactivating cholinergic neuron at theta-rhythm within 3 h- consolidation time window efficiently rescues hTau-induced spatial memory deficit. A, B Cholinergic-specific expression of ChR2 by Cre-LoxP strategy, and ~ 97% ChR2 + neurons were ChAT + measured by immunofluorescence staining of ChAT. N = 6 mice per group. Scale bar, 20 μm. C, D Cholinergic neurons expressing ChR2 fired upon 473 nm blue light stimulation at different frequencies. N = 3 mice per group. Scale bar, 20 μm. E Diagram shown the traditional BM test. We photostimulated cholinergic neurons (473 nm, 20 Hz) during probe test at day 5. F-K MS-ChAT-hTau mice showed comparable spatial learning ability to controls during 4-day learning trials (two–way ANOVA group × days, F escape latency: F [6,132] = 0.1181, P > 0.05; G number of errors: F [6,132] = 0.2038, P > 0.05). Photoactivation of cholinergic neurons at 20 Hz could not reverse the increase of latency (H One–way ANOVA group, F [2, 33] = 6.379, P < 0.01), decrease of %correct poke (I One–way ANOVA group, F [2, 33] = 6.341, P < 0.01) and reduction of %time in target (J One–way ANOVA group, F [2, 33] = 5.296, P < 0.05) in hTau-ChR2-light off group. No difference of distance moved was among groups (K One–way ANOVA group, F [2, 33] = 0.7610, P > 0.05). *P < 0.05 vs Vec: ChR2 light off. N = 10–13 mice per group. L Diagram shown the modified BM test (MBM): mice take 4 training trials for spatial task in day 1 and a test on day 3. Photostimulating cholinergic neurons for 15 min was conducted after the forth training trial immediately. M-Q Photoactivating cholinergic neurons at 8 Hz during memory consolidation rescued spatial memory loss in MS-ChAT-hTau mice. During spatial training, escape latency of MS-ChAT-hTau group was not different from controls (M two–way ANOVA group × days, escape latency: F [6,96] = 0.3494, P > 0.05). In memory test, photostimulating at 8 Hz reversed the increased latency (N One–way ANOVA group, F [2, 24] = 6.16, P < 0.01), the decreased correct poke (O One–way ANOVA group, F [2, 24] = 7.793, P < 0.01) and the reduced time in target (P One–way ANOVA group, F [2, 24] = 5.85, P < 0.01) in hTau-ChR2-light off group. *P < 0.05, **P < 0.01 vs Vec: ChR2 light off; ^#P < 0.05 vs hTau: ChR2 light off. N = 9 mice per group. R-U Photoactivating cholinergic neurons at 20 Hz or 40 Hz immediately after training on day 1 in MBM could not reverse spatial memory impairment in MS-ChAT-hTau mice. No differences in latency to reach the target area (R 20 Hz, One–way ANOVA group, F [2, 24] = 6.045, P < 0.01; T 40 Hz, One–way ANOVA group, F [2, 24] = 4.984, P < 0.01) and %correct poke (S 20 Hz, One–way ANOVA group, F [2, 24] = 8.682, P < 0.01; U 40 Hz, One–way ANOVA group, F [2, 24] = 6.877, P < 0.01) were found between light on and light off groups. *P < 0.05, **P < 0.01 vs Vec: ChR2 light off. N = 9 mice per group. V-Z At 3 h after last training, photoactivating cholinergic neurons no longer improved hTau-induced spatial memory deficits (V light panel, two–way ANOVA group × days, escape latency: F (6,108) = 0.5270, P > 0.05; W One–way ANOVA group, F [2, 27] = 5.788, P < 0.01; X One–way ANOVA group, F [2, 27] = 5.001, P < 0.01; Y One–way ANOVA group, F [2, 27] = 5.252, P < 0.01). There were no differences in distance between light on and light off groups (Q One–way ANOVA group, F [2, 33] = 0.7610, P > 0.05; Z One–way ANOVA group, F [2, 27] = 0.4052, P > 0.05). *P < 0.05 vs Vec: ChR2 light off; ^#P < 0.05 vs hTau: ChR2 light off. N = 9–12 mice per group. Data were presented as mean ± SEM