Progressive remote memory decline in APP/PS1 mice.

a. Experimental design. WT and APP/PS1 mice underwent contextual fear conditioning (CFC) at 12 and 16 weeks of age, and memory retrieval 30 days later at 16 weeks and 20 weeks of age, respectively. b. At 16 weeks old, APP/PS1 mice did not differ in freezing levels during memory retrieval compared to WT controls. Unpaired t-test: t13 = 0.52, p = 0.61, WT (n = 8), APP/PS1 (n = 7). c. At 20 weeks old, APP/PS1 mice showed reduced freezing levels during memory retrieval compared to WT controls. Unpaired t-test: t15= 3.88, *p = 0.002, WT (n = 8), APP/PS1 (n = 9). Graphs show mean ± s.e.m.

APP/PS1 mice show an age-dependent increase in PV interneuron excitability.

a. Schematic coronal brain section indicating the mPFC prelimbic region in dark grey, where tdTomato+ parvalbumin (PV) cells and pyramidal (PYR) cells were recorded in APP/PS1 PV-Cre tdTomato (APP/PS1) and PV-Cre tdTomato (control) mice. Representative differential interference contrast images (left) and a corresponding fluorescent image (right). fmi= forceps minor of the corpus callosum. ML= midline. Recordings were performed in 16 (b-i) and 20 (j-q) week-old mice. b. Resting membrane potential was unaltered in PV cells at 16 weeks of age. Mann-Whitney test: U = 433, p = 0.98, n = 29/30 cells, N = 4 mice/genotype. Ctr= control. c. Action potential (AP) firing of PV cells upon a depolarizing current step (250 pA). d. AP frequency in PV cells in response to 0-425 pA depolarizing current steps did not differ between genotypes. Two-way repeated measures ANOVA genotype x current F(10,390) = 0.44, p = 0.93, n = 29/30 cells, N = 4 mice/genotype. e. Rheobase was unchanged in PV cells. Unpaired t-test: t57 = 0.08, p = 0.93, n = 29/30 cells, N = 4 mice/genotype. Ctr= control. f. Resting membrane potential of PYR cells did not differ between genotypes. Unpaired t-test: t36 = 0.51, p = 0.61, n = 17/21 cells from N = 4/6 control vs. APP/PS1 mice, respectively. Ctr= control. g. AP firing of PYR cells upon a depolarizing current step (250 pA). h. AP frequency in PYR cells in response to 0-250 pA depolarizing current steps did not differ between genotypes. Two-way repeated measures ANOVA genotype x current F(10,370) = 0.25, p = 0.99, n = 17/21 from N = 4/6 control vs. APP/PS1 mice, respectively. i. APP/PS1 mice show a lower rheobase in PYR cells. Mann-Whitney test: U = 111, *p = 0.048, n = 17/21 cells from N = 4/6 control vs. APP/PS1 mice, respectively. Ctr= control. j. Resting membrane potential was unaltered in PV cells at 20 weeks of age. Unpaired t-test: t35 = 0.93, p = 0.36, n = 22/15 cells from N = 6/7 control and APP/PS1 mice, respectively. Ctr= control. k. AP firing of PV cells upon a depolarizing current step (250 pA). l. APP/PS1 mice show an increased AP frequency in PV cells in response to 0-425 pA depolarizing current steps. Two-way repeated measures ANOVA genotype x current F(17,595) = 4.05, *p < 0.0001, n = 22/15 cells, N = 6/7 control and APP/PS1 mice, respectively. m. Rheobase was unchanged in PV cells. Unpaired t-test: t35 = 1.67, p = 0.10, 22/15 cells, N = 6/7 control and APP/PS1 mice, respectively. Ctr= control. n. Resting membrane potential of PYR cells did not differ between genotypes. Unpaired t-test: t74 = 0.92, p = 0.36, n = 37/39 cells, N = 9 mice/genotype. Ctr= control. o. AP firing of PYR cells upon a depolarizing current step (250 pA). p. AP frequency in PYR cells in response to 0-250 pA depolarizing current steps did not differ between genotypes. Two-way repeated measures ANOVA genotype x current F(10,740) = 1.80, p = 0.08, n = 37/39 cells, N = 9 mice/genotype. q. APP/PS1 mice show a decrease in PYR cell rheobase. Unpaired t-test: t74 = 2.34, *p = 0.022, n = 37/39 cells, N = 9 mice/genotype. Ctr= control. Graphs show mean ± s.e.m.

Size and reactivation of the mPFC engram ensemble, as well as PV interneuron (re)activation, are unaffected in APP/PS1 mice.

a. Schematic representation of the viral-TRAP method. A viral cocktail of AAV-Fos::CreERT2 and Cre-dependent AAV-hSyn::DIO-mCherry were injected into the mPFC, allowing irreversible expression of mCherry upon neuronal activity and systemic injection of 4-hydroxytamoxifen (4TM). b. Schematic timeline depicting CFC and engram tagging at 12 weeks (WT n = 8, APP/PS1 n = 11 mice) or 16 weeks (WT n = 10, APP/PS1 n = 9 mice) old. On d30 after training, mice underwent a remote memory test at 16 or 20 weeks old and were perfused 90 min later. c. Representative images at 12-16w showing PV+, mCherry+ and Fos+ cells in WT (top row) and APP/PS1 (bottom row) mice. Nissl staining (general neuronal marker) is not shown. White arrowheads indicate reactivated neurons (Fos+/mCherry+ cells). Gray arrowheads indicate PV cells part of the engram (PV+/mCherry+ cells). Empty arrowheads indicate reactivated PV neurons (Fos+/mCherry+/PV+ cells). Scale bar = 50 µm. Colocalization of cell-type markers is shown for 16 (d-j) and 20 (l-r) week-old mice d. Percentage of PV cells did not differ between genotypes. Unpaired t-test: t17 = 2.04, p = 0.06. e. Percentage of mCherry+ cells did not differ between genotypes. Unpaired t-test: t = 1.02, p = 0.32. f. Percentage of Fos+ cells did not differ between genotypes. Unpaired t-test: t = 0.53, p = 0.60. g. In both genotypes, Fos and mCherry overlapped above chance level. One sample t-test: WT t7 = 3.59, * p = 0.009; APP/PS1 t10 = 3.02, * p = 0.013. h. Both genotypes showed above chance overlap of PV and mCherry. One sample t-test: WT t7 = 4.27, * p = 0.004; APP/PS1 t10 = 6.69, * p < 0.0001. i. Both genotypes showed below chance colocalization of Fos and PV. One sample t-test: WT t7 = 5.06, * p = 0.002; APP/PS1 t10 = 12.99, * p < 0.0001. j. Fos, PV, and mCherry+ overlap did not differ from chance level in both genotypes. One sample t-test: WT t = 4.05, * p = 0.005; APP/PS1 t = 4.60, * p = 0.001. k. Representative images at 16-20w showing PV+, mCherry+ and Fos+ cells in WT (top row) and APP/PS1 (bottom row) mice. l. Percentage of PV cells did not differ between genotypes. Unpaired t-test: t17 = 0.67, p = 0.51. m. Percentage of mCherry+ cells did not differ between genotypes. Unpaired t-test: t = 0.42, p = 0.68. n. Percentage of Fos+ cells did not differ between genotypes. Unpaired t-test: t17 = 0.55, p = 0.59. o. In both genotypes Fos and mCherry colocalized above chance level. One sample t-test: WT t9 = 5.97, * p = 0.0002; APP/PS1 t8 = 6.63, * p = 0.0002. p. Both genotypes showed above chance overlap of PV and mCherry. One sample t-test: WT t9 = 6.73, * p < 0.0001; APP/PS1 t8 = 7.05, * p = 0.0001. q. Both genotypes showed below chance colocalization of Fos and PV. One sample t-test: WT t9 = 6.04, * p = 0.002; APP/PS1 t8 = 4.07, * p = 0.004. r. Fos, PV, and mCherry overlap did not differ from chance level in both genotypes. One sample t-test: WT t8 = 3.45, * p = 0.009; APP/PS1 t8 = 4.00, * p = 0.004. Graphs show mean ± s.e.m.

Perisomatic PV labelling is increased on engram cells in 20-week-old APP/PS1 mice.

a-b. Left: Representative image of WT (a) and APP/PS1 (b) mice from the 12-16 week groups showing PV staining and mCherry+ cells in the mPFC. Scale bar = 50 µm. Right: Examples of a mCherry- (top row) and mCherry+ (bottom row) cell. Soma outline is based on Nissl. PV signal was masked and measured inside the ring surrounding the soma. Scale bar = 10 µm. c. PV labelling around mCherry+ cells did not differ from mCherry- cells in WT and APP/PS1 mice. Two-way repeated measure ANOVA cell population: F(1,17) = 2.25; p = 0.15; WT (n = 8), APP/PS1 (n = 11) d. Left: Representative image of WT (d) and APP/PS1 (e) mice from the 16-20 week groups showing PV staining and mCherry+ cells in the mPFC. Scale bar = 50 µm. Right: Examples of a mCherry- (top row) and mCherry+ (bottom row) cell. Soma outline is based on Nissl. PV signal was masked and measured inside the ring surrounding the soma. Scale bar = 10 µm. f. An increased amount of PV labelling around mCherry+ cells was found compared to mCherry- cells in APP/PS1 mice but not control mice. Two-way repeated measure ANOVA cell population: F(1,17) = 21.74; *p = 0.0002. Post-hoc Bonferroni test: APP/PS1 mCherry+ vs. mCherry-*p = 0.0015. Graphs show mean ± s.e.m.

Engram cells of 20-week-old APP/PS1 mice receive increased inhibitory input.

a. Coronal brain section indicating the mPFC region (dark grey) where AAV-Fos::CreERT2 and Cre-dependent AAV-hSyn::DIO-mCherry were injected. Mice underwent CFC at 16 weeks old and engram cells were tagged. Thirty days after CFC, mice were re-exposed to the training context and then immediately sacrificed for whole-cell patch-clamp electrophysiology. b. Left: Representative image showing labelled mCherry+ engram cells in the mPFC. Right: recordings were made from mCherry+ and mCherry- pyramidal cells. fmi= forceps minor of the corpus callosum. ML= midline. c. Example sIPSC traces of mCherry+ and mCherry- for WT control and APP/PS1 mice d. Example sEPSC traces of mCherry+ and mCherry- cells for WT and APP/PS1 mice e. Frequency of sIPSCs differed between mCherry+ and mCherry- cells in APP/PS1, but not WT, mice. Two-way repeated measure ANOVA genotype x cell-type F(1,43) = 5.44, *p = 0.024. Post-hoc Bonferroni APP/PS1 mCherry+ vs. mCherry- *p = 0.011. n = 22 per cell-type from N = 6 WT mice, n = 23 per cell-type from N = 7 APP/PS1 mice. f. sIPSC amplitude did not differ between cell-type and genotype. g. Frequency of sEPSCs was enhanced in mCherry+ cells compared to mCherry- cells in both genotypes. Two-way repeated measure ANOVA cell-type F(1,36) = 7.26, *p = 0.011, n = 20 per cell type from N = 6 WT mice, n = 23 per cell-type from N = 7 APP/PS1 mice h. sEPSC amplitude did not differ between cell-type and genotype. Graphs show mean ± s.e.m.