Fig 2: ATGs expression is altered in PHG and hippocampus of AD postmortem brains. a Heatmap of the Log FC and adjusted p values of the significantly up- and down-regulated ATG expressions found in PHG (BA36) and their counterpart expression in GSE5281 and other MSBB datasets. b-c Pie charts representing the autophagy gene functions associated to the upregulated or downregulated ATGs identified in (a). d Heatmap representing the mean z-score expression of each significant ATGs at different CDR stages within the PHG (MSBB-BA36). e-h NRBF2, BECN1, PIK3C3 and PIK3R4 gene expressions are progressively reduced in the PHG as a function of CDR score

Fig 3: Injection of rAAV-NRBF2-mCherry into dorsal hippocampus rescues memory impairments and reduces Aβ levels in 5XFAD mice. a-b Y-Maze performance and total number of entries recorded 1 day before (− 1), 30 and 60 days post-viral injection. Each group performance was normalized to their initial performance, i.e. -1 dpi data. Results are mean ± SEM of 4–5 months old WT + mCherry (n = 4), WT + NRBF2 (n = 4), 5XFAD + mCherry (n = 4), 5XFAD + NRBF2 (n = 4) mice. The statistical significance was determined using a repeated two-way ANOVA test. c Immunofluorescence analysis of Aβ and mCherry expression in hippocampus of 5XFAD mice injected with mCherry or NRBF2-mCherry viruses. Scale bars, 500 μm. Images shown are representatives of two different experiments. d Quantification of Aβ signal within CA3 and DG area C. Results were normalized over mCherry control and are mean ± SEM of mCherry or NRBF2-mCherry acquired from three separate mice per group. The statistical significance was determined using regular two-way ANOVA test followed by Bonferroni’s post-test. *p < 0.05

Fig 4: Loss of NRBF2 causes memory and LTP deficits in mice. a Discrimination ratio obtained from OLT task. Results are mean ± SEM of 3 months old WT (n = 15) and KO (n = 14) mice. b Freezing behavior recorded during CFC experiments. Results are mean ± SEM of 3 months old WT (n = 12) and KO (n = 10) mice. c Percentage of errors measured during RAM experiments. Results are mean ± SEM of3 months old WT (n = 15) and KO (n = 14) mice. The statistical significance was determined using row-matched two-way ANOVA test followed by Bonferroni’s post-test. d fEPSP slope measured from Schaffer collateral path. Results are mean ± SEM of 3 months old WT (n = 6) and KO (n = 6) slices from three different mice per group. The statistical significance was determined using regular two-way ANOVA on the last hour of recording. e Basal synaptic transmission measured from WT and KO mice. Results are mean ± SEM of 3 months old WT (n = 3) and KO (n = 3) slices from three different mice per group

Fig 5: Transduction of NRBF2 carrying viruses into dorsal hippocampus rescues memory impairments and restores autophagy in NRBF2-KO mice. a Discrimination ratio obtained on testing day of OLT. b Freezing behavior recorded during CFC experiments. Results are mean ± SEM of 3–4 months old WT + mCherry (n = 14), WT + NRBF2 (n = 14), KO + mCherry (n = 16) and KO + NRBF2 (n = 18) mice. c Immunoblot (IB) analysis of p62 and LC3B. d-e Quantification of C. Results are mean ± SEM of 3–4 months old WT + mCherry (n = 11) and WT + NRBF2 (n = 12) KO + mCherry (n = 11) KO + NRBF2 (n = 11) mice. The statistical significance was determined using regular two-way ANOVA test followed by Bonferroni’s post-test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001