TRAP labeling of mGCs and birth-dated abDGCs after the TUNL pattern separation task

(A) Top: Schematic illustrating the pattern separation paradigm. Male (n = 9) and female (n = 7) mice (6-8-week-old) received 3 injections of BrdU at D0 and 3 injections of EdU two weeks later (D14) before starting the behavioral paradigm. Mice received one injection of 4-OHT on the day they reached criterion (TRAP) and were perfused 3 days later 90 minutes after one exposure to the behavioral paradigm to allow endogenous c-Fos expression (c-Fos). Control mice (males n = 6, females n = 4) underwent the same treatment but only participated in the Large separation task (L). Mice were water-restricted during the behavioral experiment. Bottom: Cartoon representation of the chamber in TUNL task with examples of Large separation (L) and Small (S) separation configurations. High cognitive demand paradigm consisted of interleaved L and S trials (LS). (B) Number of days for each mouse to reach the 70% success criterion in the Large separation (L). (C) Number of days for mice to reach 70% success criterion in S trials, during the LS separation task. (D) Left: Average percentage of correct choices (% success) in the L separation on the first day of training, as well as during the three days prior to reaching criterion (days before criterion, DBC) and the day of reaching the criterion (crit)(one-sample t-test to chance at criterion Large t(16)=12.04, p<0.0001). Middle: Average percentage of correct choices in the L and S separation on the first day of training, during the three days before criterion (DBC), on the day of reaching criterion and 3 days later (one-sample t-test to chance at criterion Large t(16)=8.38, p<0.0001; Small t(16)=7.37, p<0.0001). Right: Cumulative probability of days to criterion of mice in S for both controls L and LS groups. (E) Example section of BrdU+ labeling (green) EdU+ labeling (blue) and TRAP+ labeling (red) in TRAP2 mice. Calibration bar: 30 μm. (F) Density of TRAP-labeled GCs in mice undergoing LS compared to control mice engaged in either shaping (Shap) or the Large configuration task only (L) (one-way ANOVA, F(2,25) = 7.97, p = 0.0021; Tukey post-hoc test. Groups Shap vs LS and Large vs LS were significantly different). (G) Percentage of BrdU labeled cells (39 days +/-1.23) or Edu labeled cells (25 days +/-1.37 days) also TRAP+ after L or LS training (BrdU+/TRAP+ L: 0.145 ± 0.145 %, 6 mice; LS BrdU+/TRAP+ 0.822 ± 0.474 %, 6 mice; ns; EdU+/TRAP+ L: 0.203 ± 0.144 %, 6 mice; LS BrdU+/TRAP+ 0.304 ± 0.155 %, 7 mice; ns; Mann–Whitney). (H) Percentage of TRAP+ GCs expressing c-Fos+ in Shap, L or LS task after reaching criterion (expressed as a percentage of TRAP+ cells) (one-way ANOVA, F(2,18) = 6.66, p = 0.0069; Tukey post-hoc test. Groups Shap vs LS and Large vs LS were significantly different).

Blade-biased activity of mGCs during a high cognitive demand pattern separation task.

(A) Spatial distribution of TRAP-labeled cells in the infrapyramidal (IB) and suprapyramidal (SB) blades of the dorsal DG in TRAP2 mice performing the LS separation, compared to control mice that performed only the Shap or L configuration (two-way ANOVA (unbalanced) (blade: F(1,38) = 146.34, p = 1.33 × 10⁻¹⁴) (group: F(2,38) = 1.14 × 10⁻⁷, p = 1) (group × blade: F(2,38) = 13.99, p = 2.81 × 10⁻⁵). Tukey post-hoc test: significant differences were observed between IB; LS and Large (p=0.03), LS and Shap (p=0.05), SB; LS and Large (p= 0.03), LS and Shap (p=0.05). (B) Spatial distribution of TRAP+ cells in the hilar to outer radial GCL axes in the SB (0 to 120μm) and the IB (0 to 120μm) of the dorsal DG with 0 μm indicating location at the hilar border (Shap: 5 mice, 255 cells; L: 5 mice, 294 cells; LS: 5 mice, 639 cells). (C) Distribution of TRAP-labeled cells along the apex to blade extremity axes of the SB of the dorsal DG, with 0% indicating location at the dentate apex (Shap: 5 mice, 152 cells; L: 5 mice, 176 cells; LS: 5 mice, 517 cells)(two-way ANOVA (unbalanced) (distance (0%-100%): F(1,26) = 26.05, p = 2.56 × 10⁻⁵), group: F(2,26) = 0.044, p = 0.957) (group× distance (0%-100%): F(2,26) = 23.77, p = 1.35 × 10⁻⁶). Tukey post-hoc test: significant differences were observed between Shap; 0%-50% and 50%-100% (p<0.0001), LS; 0%-50% and 50%-100% (p=0.03). (D) Example section of TRAP+ cells in the dorsal DG illustrating the distinct boundaries between the two blades, the SGZ, and the apex. Calibration bar: 50 μm. (E) Cartoon illustrating the two blades of the DG: IB and SB. Dentate granule cells are depicted in grey while red circles represent TRAP+ cells in the DG. In the L group, a greater number of labeled cells are localized in the SB compared to the IB. This bias of activity is also observed in the LS group, where the bias is more pronounced, and the overall activity of the DG is increased. In the SB, the majority of the activated cells are located closer to the tip of the blade.

Activation of GCs during remote recall in the high cognitive demand task.

(A) Schematic representation of the pattern separation paradigm used in the study. Mice (male n=11, female n=8) performed two trainings sessions: initial (I) and remote (R). The second training (R) took place 3 weeks after the mice reached the S criterion in LS. Mice were perfused when they reached the criterion in S during R. A second group that performed only in the L configuration during both training sessions was included as control (males n=3, females n=3). (B) Learning curves representing the percentage of success during the remote test (R). The control group trained only in the large-separation condition (L group) is shown in green and includes performance in the large-separation trials only. The LS group (orange) was trained in both large-and small-separation conditions and therefore includes two traces: the upper orange trace represents performance during large-separation trials, and the lower orange trace represents performance during small-separation trials. (one-sample t-test to chance at criterion; LS: Large t(14) = 11.8, p < 0.0001, Small t(14) = 10.3, p<0.0001; L: Large t(3) = 6.78, p = 0.0066). (C) Percentage of mice reaching the L and S criterion in LS group, during the remote test (R) represented as cumulative probability. During the remote test, mice reach the L criterion in just a few days but require more time to reach the S criterion. (D) Performance of individual mice indicating a non-significant reduction in the number of days needed to reach the criterion between the initial and remote training sessions. Only 4 mice (red lines) required more days to reach the criterion during the remote test. (I: 8.067 ± 1.173 %, R: 5.93 ± 0.753, 15 mice; p= 0.204; Mann–Whitney). (E) EdU labeled cell that was also TRAP+ labeled after the initial training but was that was not reactivated (c-Fos) during the remote test. EdU (green), TRAP (red), c-Fos (blue) in TRAP2 mice. Calibration bar, 10 μm. (F) Distribution of TRAP-labeled cells in the IB and SB blades of the dorsal DG in TRAP2 mice. Labeling was performed in LS separation during initial training (I) and compared to control mice that performed only the L configuration. Active neurons show a preferential distribution to the SB over the IB in the DG. This bias is more prominent in the LS group of mice (Two-way ANOVA: blade, F(1,12) = 32.98, p < 0.001; group, F(1,12) = 9.2×10⁻³⁰, p = 1; group×blade interaction, F(1,12) = 16.92, p = 0.0014. Tukey post-hoc test: significant differences were observed between IB; LS and Large (p=0.05), SB; LS and Large (p= 0.05). (G) Distribution of c-Fos labeled cells between the IB and SB blades of the dorsal DG in TRAP2 mice performing the LS separation during remote training (R), compared to control mice that performed only the L configuration (Two-way ANOVA: blade, F(1,12) = 56.03, p < 0.00001; group, F(1,12) = 0, p = 1; group×blade interaction, F(1,12) = 7.02, p = 0.021.Tukey post-hoc test: no significant differences were observed between LS and Large for IB or SB). (H) Percentage of newborn neurons (BrdU (7.7 weeks) or EdU (5.7 weeks)) activated during the initial training (I), as indicated by co-labeling of TRAP+(BrdU+ c-Fos+: L: 0.242 ± 0.242 %, 6 mice, LS: 0.170 ± 0.170 %, 6 mice; ns; Mann–Whitney). (I) Percentage of newborn neurons (BrdU or EdU) activated during the remote test (R), as indicated by c-Fos expression (BrdU+ c-Fos+: 0.977 ± 0.426 %, 6 mice, p=0.05; EdU+ c-Fos+: 0.694± 0.317 %, 9 mice; p= 0.0785; Mann–Whitney). (J) Percentage of mGCs activated during the (I) training (TRAP+) that were reactivated during the R test (c-Fos+) (expressed as a percentage of TRAP+ cells) (Unpaired t-test, t(6) = 2.90, p = 0.027).

DREADD inhibition of mGCs during a high cognitive demand task.

(A) Percentage of BrdU-labeled cells that are also Dock10+ at different time points following BrdU injections (in days). (B) Representative section showing BrdU+ labeling (green) and Dock10+ labeling (red) in Dock10 cre;Ai9 mice. Calibration bars: 10 μm. (C) Schematic representation of the pattern separation paradigm. All mice (males n = 17, females n = 27) received DCZ i.p. injections 30 minutes before the beginning of their LS trials. (D) Percentage of mice reaching criterion in (S) and percentage that were not successful (NS) after 14 days of training in S configuration of LS. A majority of mice never succeeded in reaching the criterion in the Cre+ DCZ group (z = 2.00, p = 0.0453, z-test - Group NS). (E) Percentage of mice reaching the criterion in S trials across days represented as the cumulative probability including all successful and not successful mice in the group. (F) Density of c-Fos+ GCs in the dorsal DG across the two groups. A significantly reduced density was observed in the Cre+ DCZ group compared to controls (Unpaired t-test, t(9) = 2.30, p = 0.047). (G) Density of c-Fos+ GCs in the dorsal CA3 across the two groups. A significant reduction was observed in the Cre+ DCZ group compared to the control groups (Unpaired t-test, t(9) = 4.93, p = 0.0008). (H) Blade specific distribution of c-Fos+ GCs in the IB and SB of the dorsal DG in mice performing LS separation (Two-way ANOVA (unbalanced, Type II) revealed a significant main effect of blade (F(1,16) = 315.83, p = 5.87 × 10⁻¹²), but no significant main effect of group (F(1,16) = 0, p = 1) and no significant group × blade interaction (F(1,16) = 0.040, p = 0.845). (I) Example of dorsal DG section from Dock10-cre:hM4Di mice after LS training. c-Fos+ neurons are labeled in red. Calibration bars: 100 μm. (J) Spatial distribution of c-Fos+ along the hilar to outer radial GCL axis of the DG (0 to 120μm), with 0 μm indicating location at the SGZ. (K) Cartoon illustrating the distribution of activity labeled GCs in the SB and IB blades of the DG. GCs are depicted in grey while red circles represent activated cells in the DG. The overall distribution of neurons in mice in which mGCs were inhibited (Cre+ DCZ) and which did not reach criterion was not different to controls even though the total density of activity labeled neurons was lower.

DREADD inhibition of ≤ 7-week abDGCs during a high cognitive demand task.

(A) Schematic representation of the pattern separation paradigm. Prior to onset of the training, mice (males n = 17, females n = 16) were provided tamoxifen-containing food ad libitum in the home cage for 7 weeks to activate TAM inducible Cre recombinase (Ascl1-CreERT2) in abDGCs. All mice received DCZ (deschloroclozapine, 50 μg.kg−1) or DMSO injections 30 minutes before the beginning of LS trials. (B) Example sections from Ascl1-CreERT2; hM4Di mice with immunohistochemical localization of the HA tagged DREADD (top) and after LS training with c-Fos+ neurons labeled in red. Calibration bars: 140 μm. (C) Percentage of mice reaching the criterion in the S trials across days represented as the cumulative probability. The Cre+ DCZ treated group needed additional days to reach criterion compared to controls. (D) Number of days for each mouse to reach the 70% success criterion in S trials, (one-way ANOVA, F(2,31) = 3.77, p = 0.034; Tukey post-hoc test indicate that Cre+ DCZ group is significantly different from Cre+ DMSO, while other pairs are not significantly different). (E) c-Fos+ cell density in all groups of mice. There was an increase in c-Fos+ mGCs in mice in which:: 7 week abDGCs were inhibited by DCZ (H(2) = 7.22, p=0.0271; Kruskal Wallis)(Cre+DCZ comparison to Cre+ DMSO, p = 0.014; Cre+ DCZ comparison to Cre-DCZ, p = 0.027 with Dunn’s multiple comparison post hoc analysis). (F) Example section after c-Fos immunohistochemistry to assess distribution of c-Fos+ mGCs between the IB and SB of the dorsal DG in groups of mice who received DCZ (cre-vs cre+). c-Fos cells are localized closer to the SGZ in the Cre+ group. Calibration bar, 100 μm. (G) c-Fos+ density in the CA3 region of the hippocampus in each of the groups of mice. No significant difference was observed in any of the groups (one-way ANOVA, F(2,13) = 2.27, p = 0.143; Tukey post-hoc test indicates that there is no significant difference between any pairs of groups). (H) Blade distribution of c-Fos+ mGCs in the IB and SB of the dorsal DG in mice performing the LS separation receiving either DCZ or DMSO 30 min prior to task performance. Active neurons are distributed preferentially to the SB than the IB in the DG for the control groups. However, this distribution bias is significantly reduced in Cre+ DCZ group (two-way ANOVA (unbalanced) (blade: F(1,24) = 659.55, p < 2.2 × 10⁻¹⁶; group: F(2,24) = 3.48 × 10⁻¹³, p = 1; group × blade: F(1,16) = 0.040, p = 0.845). Tukey post-hoc test. Significant differences were observed for IB comparisons: Cre+ DCZ vs Cre-DCZ (p=0.000548) and Cre+ DCZ vs Cre+ DMSO (p=0.00765). Significant differences were observed for SB comparisons: Cre+ DCZ vs Cre-DCZ (p=0.000548) and Cre+ DCZ vs Cre+ DMSO (p=0.00765). (I) Spatial distribution of c-Fos+ GCs along hilar to molecular layer axes of the SB (0 to 120μm) and the IB (0 to 120μm) of the dorsal DG, with 0 μm indicating the hilar position. c-Fos+ cells are localized closer to the SGZ in the Cre+ DCZ group. (J) Cartoon illustrating the DG with its two blades: IB and SB. Dentate granule cells are depicted in grey while red circles represent activity labeled cells in the DG. For Cre+ DCZ mice, the distribution of labeled neurons is closer to SGZ, away from outer radial GCL and more evenly distributed in IB and SB.

DREADD inhibition of ≤ 4-week-old abDGCs does not affect performance of mice.

(A) Schematic representation of the pattern separation paradigm used in the study. Prior to the behavioral experiment, mice (males n = 12, females n = 23) were given tamoxifen-enriched food for 4 weeks to activate the Cre recombinase and label newborn neurons. All mice received DCZ or DMSO injections 30 minutes before the beginning of LS trials. (B) Number of days for each mouse to reach the 70% success criterion in S trials (H(2) = 4.21, ns; Kruskal Wallis). (C) Density of c-Fos+ GCs in the dorsal DG across the different groups (one-way ANOVA, F(2,13) = 0.157, p = 0.857). (D) Blade specific distribution of c-Fos+ cells in the IB and SB of the dorsal DG in mice performing LS separation (H(2) = 0.46, ns; Kruskal Wallis). (E) Spatial distribution of c-Fos+ along the hilar to outer radial GCL axis of the DG (0 to 120μm), with 0 μm indicating location at the border of the SGZ and hilus.

Activity labeling of GCs after running

(A) Schematic representation of the running paradigm. Prior to the behavioral experiment, mice (males n = 30, females n = 3) were injected with BrdU. After Δ weeks, the mice were allowed to run for 2 hours before receiving a 4-OHT injection, followed by an additional 2 hours of running. Mice were perfused 3 days later for immunohistochemistry of brain sections. (B) Example section of immunohistochemistry of TRAP+ cells in the dorsal DG illustrating the distinct blades of the DG: SB and IB. Calibration bar: 40 μm. (C) Density of TRAP+ GCs in mice after running compared to home cage controls (H). Each group represents mice that ran after variable time intervals following BrdU injections. (D) Blade specific distribution of TRAP+ cells in the infrapyramidal (IB) and suprapyramidal (SB) blades of the dorsal DG in TRAP2 mice that ran for 4 hours, compared to home cage mice. (IB: home: 42.5 ± 1.43 %, 4 mice; run: 43.9 ± 1.42 %, 33 mice; ns; Mann–Whitney). (E) Example section from mice labeled for BrdU (green), TRAP+ GCs (red) and endogenous c-Fos (blue) in TRAP2 mice. Calibration bar: 20 μm. (F) Percentage of BrdU+ cells (Δ 1-14 weeks old) also labeled with TRAP in the dorsal DG of TRAP2 mice that ran for 4 hours, compared to home cage controls (H). (G) Blade specific distribution of BrdU+ cells in the infrapyramidal (IB) and suprapyramidal (SB) blades of the dorsal DG in TRAP2 mice that ran for 4 hours, compared to home cage mice.

Description of the TUNL protocol and analysis of TRAP+ GCs in mice not meeting the criterion in the pattern separation task.

(A) Overview of successive trials in the Shaping, L and LS steps. Mice must complete 30 trials at each step. The shaping phase involves touching the illuminated square to receive a reward. In the L and LS steps, there are in two phases: the sample (S) phase, where mice must touch the single illuminated square on the panel and the choice test (C), where two squares are illuminated: the previously touched square and a new one. Mice must nose poke the new square to receive the reward. In the L phase, trials consist only of L panels only (see Figure S2B – L). In the LS phase, L and S trials are alternated (see Figure S2B for S panels). (B) Recommended stimulus pairs for the TUNL protocol. Large pairs (green) consist of 2 illuminated squares separated by 4 squares that are turned OFF. Small pairs (orange) consist of two illuminated squares separated by 0 to 1 square turned OFF. (C) Correlation between the performance (= number of days to reach criterion) and the TRAP+ density or TRAP+ distribution in SB (D). Both regression analysis (right tables) show that the slope is not significantly different from zero. (D) Comparison of TRAP+ distribution in IB and SB of the dorsal DG in mice from the home cage (Home), running (Run), large configuration (L) and large and small separation (LS) groups. In the control groups (Home, Run and L) the TRAP+ cells are predominantly distributed in the SB, with a slight preference for the SB over IB. However, in the high cognitive demand (CD), LS takm the distribution pf TRAP+ cells become more balanced, with a marked preference for the SB (SB >>IB), indicating task-specific modulation of activity across the blades. (E) Density of c-Fos-labeled GCs in mice undergoing LS compared to control mice engaged in either shaping (Shap) or the Large configuration task only (L) (one-way ANOVA, F(2,17) = 8.05, p = 0.0035; Tukey post-hoc test. Groups Shap vs LS were significantly different). (F) Percentage of TRAP2 mice reaching the criterion (C) compared to those that did not within 14 days of training (NC). (G) Density of TRAP+ neurons in mice undergoing LS for mice that reached criterion (LS C) with mice that did not reach 70% criterion within 14 days (LS NC) (Unpaired t-test, t(16) = 0.96, p = 0.35). (H) Blade specific distribution of TRAP+ GCs in the SB and the IB of the dorsal DG (IB: LS C: 34 ± 2.79 %, 8 mice; LS NC: 33.8 ± 2.12 %, 12 mice; ns; Mann–Whitney). (I) Percentage of TRAP+ cells also birth-dated with BrdU (39 days +/- 1.23) or Edu cells (25 days +/- 1.37 days) in the dorsal DG of TRAP2 mice that reached criterion (C) and those failed (NC)(BrdU: LS C: 0.71 ± 0.333 %, 7 mice; LS NC: 0.822 ± 0.45 %, 6 mice; ns; EdU: LS C: 0.542 ± 0.33 %, 9 mice; LS NC: 0.304 ± 0.16 %, 7 mice; ns; Mann–Whitney).

Dock10 is not expressed in DCX-positive immature newborn neurons but is expressed in Prox1-positive neurons

(A) Representative section from a Dock10-cre:Ai9 mouse following tdTomato (tdT), Prox1, and DCX immunohistochemistry in the dorsal DG. Calibration bar: 100 μm. (A) 20X magnification of the suprapyramidal layer of the dentate gyrus, showing tdTomato (Dock10) in red, Prox1 in green, and DCX in blue. Calibration bar: 100 μm. (C) tdT (Dock10) labeling. (D) Prox1 (green) labeling. (E) DCX (blue) labeling. Calibration bar: 100 μm.