Hilar mossy cell targeting by viral vector injections in vivo.

A) Schematic of AAV injection coordinates in Calcitonin receptor-like receptor (Crlr-Cre) mice, with description of experimental groups including mouse genotype and viral vector used. B) Representative images of dentate sections stained for the mossy cell (MC) marker calretinin in each condition, 6 weeks after virus injection, demonstrating hilar mossy cell loss in MC ablated mice and preserved mossy cells in MC silenced mice. Yellow dashed lines note hilus in images C) Summary data of hilar mossy cell densities for each group, 6 weeks after virus injection (Virus control n = 7 mice, GFP control n = 4 mice, MC ablated n = 7 mice, and MC silenced n = 6 mice; *** p <0.001, **** p <0.0001).

Enhanced dendritic growth of immature adult-born DGCs in the absence of mossy cell inputs.

A) Representative GFP-expressing adult-born dentate granule cells (DGCs) in virus-injected control and mossy cell (MC) ablated mice, 14 days after retroviral labeling/mitosis. B) Skeleton tracing of cell morphology from the representative images in A, with a dashed line 50µm from cell body center. C&D) Total dendritic length and number of branches per cell of neurons from virus control and MC ablated mice (Virus control n = 17 cells/ 5 mice, MC ablated n = 21 cells/ 6 mice; * p <0.05). E) Sholl analysis demonstrates more extensive proximal dendritic arborization for DGCs born after mossy cell ablation (* p= 0.02 and 0.01 at intervals 10 and 20µm distance from soma, blue shading, all other points not significant). F) Representative mCherry-expressing adult-born (DGCs) in control (AAV5-flex-GFP) virus injected and MC silenced (AAV5-flex-TeLC-injected) mice, 14 days after retroviral labeling. G) Skeleton tracing of cell morphology for the representative images in G, with a dashed line 50µm from cell body center. H&I) Total dendritic length and number of branches per adult-born cell in GFP control and MC silenced mice (GFP control n = 11 cells/ 4 mice, MC silenced n = 15 cells/ 5 mice; *** p < 0.001, **** p < 0.0001). J) Sholl analysis demonstrates more extensive arborization of immature DGCs that developed in the absence of functional mossy cell inputs (*p= 0.03, 0.009, and 0.007 at intervals 20, 30, and 40µm distance from soma, **p= 0.007, 0.004, 0.001, 0.001, 0.001, 0.001, 0.004 at intervals 110, 120, 130, 140, 150, 160, and 170 µm distance from the soma, all other points not significant).

Adult-born DGCs achieve normal dendritic morphology in the absence of functional mossy cell inputs.

A) Representative GFP-expressing adult-born dentate granule cells (DGCs) in virus injected control and mossy cell (MC) ablated mice, 21 days after retroviral labeling/mitosis. B) Skeleton tracing of cell morphology from the representative images in A, with a dashed line 100µm from cell body center. C&D) Total dendritic length and number of branches per adult-born cell in virus control and MC ablated mice (Virus control n = 19 cells/ 5 mice, MC ablated n = 22 cells/ 6 mice; p > 0.05). E) Sholl analysis demonstrates similar branching patterns of 21 day old DGCs in both groups. F) Representative mCherry-expressing adult-born (DGCs) in control (AAV5-flex-GFP) virus-injected and MC silenced (AAV5-flex-TeLC-injected) mice, 21 days after retroviral labeling. G) Skeleton tracing of cell morphology from the representative images in F, with a dashed line 100µm from cell body center. H&I) Total dendritic length and number of branches per cell of GFP control and MC silenced mice (GFP control n = 20 cells/ 6 mice, MC silenced n = 24 cells/ 6 mice; p >0.05). J) Sholl analysis demonstrates similar branching patterns of 21 day old DGCs in both groups.

Differential effects of mossy cell ablation vs silencing on proximal dendritic spine formation by adult-born DGCs.

A) Representative dendritic spines in the IML and OML from 21 day old, GFP-labeled adult born granule cells that developed in virus control and MC ablated conditions. B) IML spine density is not changed in cells born following mossy cell ablation (Virus control n = 11 cells/ 3 mice, MC ablated n = 14 cells/ 3 mice). C) OML spine density is not changed in cells born following mossy cell ablation (Virus control n = 11 cells/ 3 mice, MC ablated n = 14 cells/ 3 mice). D) Representative dendritic spines in the IML and OML from 21 day old, mCherry-labeled adult born granule cells that developed in GFP virus control and MC silenced conditions. E) IML spine density is reduced in cells born following mossy cell ablation (GFP control n = 11 cells/ 3mice, MC silenced n = 12 cells/ 3 mice; * p <0.05). F) OML spine density is not changed in cells born following mossy cell silencing (GFP control n = 11 cells/ 3mice, MC silenced n = 12 cells/ 3 mice).

Altered evoked synaptic responses in both mature and adult-born DGCs following mossy cell ablation but not silencing.

A) Schematic of recording paradigm for mature DGCs, with electrical stimulation of the proximal molecular layer. B) Representative image of the recording configuration for a mature DGC. C) Schematic of recording paradigm for virus-labeled 21 day old adult-born DGCs, with electrical stimulation of the proximal molecular layer. D) Representative image of a GFP+ 21dpi DGC during recording, filled with AlexaFluor 594 (red) to confirm cell specificity. E) Representative traces of total evoked synaptic current (green), AMPAR-mediated current (black), and GABAaR-mediated current (derived by subtraction, red) recorded from mature DGCs from virus control and mossy cell ablated conditions. F) Excitation:inhibition (E:I) ratios for synaptic responses to proximal molecular layer stimulation recorded from mature DGCs (Virus control n = 9 cells/ 7 mice, MC ablated n = 7 cells/ 6 mice, MC silenced n = 7 cells/ 6 mice; ** p<0.01, *** p<0.001). G) Excitation:inhibition (E:I) ratios of synaptic responses to proximal molecular layer stimulation recorded from 21 day old adult-born DGCs (Virus control n = 5 cells/ 3 mice, MC ablated n = 6 cells/ 3 mice, MC silenced n = 5 cells/ 5 mice; * p <0.05, *** p<0.001).

Mossy cell ablation causes collapse of the inner molecular layer.

A) Representative images of VGlut2-expressing perforant path axons in relation to calretinin-expressing mossy cell axons from control, MC ablated, and MC silenced mice. B) IML width, as defined by calretinin staining (Virus control n = 9 mice, MC ablated n = 6 mice, MC silenced n = 5 mice, same mouse numbers for panels B-D; **** p < 0.0001). C) IML width, as defined by the VGlut2-negative region (*** p < 0.001, **** p < 0.0001). D) Total molecular layer (ML) width in control, MC ablated, or MC silenced mice (* p < 0.05). E) Perforant path width (VGlut2-positive middle and outer molecular layer) is unchanged after MC ablation, suggesting IML collapse. F) Schematic of viral injections to selectively label medial entorhinal cortex (MEC) axons in control and MC ablated mice. G) Representative GFP-positive MEC axons from virus control and MC ablated mice. White lines highlight the distance between labeled MEC axons and the granule cell layer (GCL) in control and MC ablated mice. H) Distance between the most proximal MEC axons and the GCL is reduced after MC ablation (Virus control n = 5 mice, MC ablated n = 4 mice; **** p < 0.0001).

Strongly reduced CB1R-mediated inhibition of proximal molecular layer inputs after mossy cell ablation or silencing.

A) Schematic of recording paradigm. A stimulating electrode was used to activate inputs in the proximal molecular layer, while recording from mature or retrovirally-labeled adult born DGCs from control, MC ablated or MC silenced mice. B) Representative AMPAR-mediated evoked excitatory synaptic responses before and after application of WIN 55,212-2 (5µM). C) EPSCs recorded from mature DGCs in both MC ablated and MC silenced mice had minimal WIN sensitivity relative to cells from control mice (Virus control n = 9 cells/ 7 mice, MC ablated n = 7 cells/ 6 mice, MC silenced n = 7 cells/ 6 mice; *** p <0.001, **** p <0.0001). D) EPSCs recorded from 21 day old DGCs in MC ablated and MC silenced mice had minimal WIN sensitivity relative to control cells (Virus control n = 5 cells/ 3 mice, MC ablated n = 6 cells/ 3 mice, MC silenced n = 5 cells/ 5 mice; *p < 0.05; ** p < 0.01).

Basal dentate granule cell activity is not altered by functional mossy cell loss.

A) Representative cFos expression (red) in dentate granule cells in virus injected Cre-negative control, GFP virus injected control, MC ablated or MC silenced mice. Calretinin staining (white) of the IML highlights mossy cell axons, while calretinin staining highlights immature neuroblasts in the subgranular zone and mossy cell somata in the hilus, respectively. B) cFos-positive cell density in the dentate granule cell layer; (Virus control n = 4 mice, GFP control n = 4 mice, MC ablated n = 3 mice, MC silenced n = 3 mice; n.s. all comparisons).

Neither mossy cell loss nor silencing alters seizure susceptibility.

A) Representative cFos expression (red) in dentate granule cells after saline or PTZ injection. B) Granule cell cFos expression is increased after PTZ injection similarly in all groups (Virus control Saline= 5 mice, Virus control PTZ= 8 mice, GFP control PTZ= 7 mice, MC ablated PTZ= 7 mice, and MC silenced PTZ= 5 mice; ** p <0.01, *** p <0.001, and **** p <0.0001 compared to saline; n.s. for all PTZ-treated groups relative to each other). C) Maximum seizure scores (* p<0.05, ** p <0.01, and *** p <0.001 compared to saline; n.s. for all PTZ-treated groups relative to each other). D) Cumulative seizure scores (*** p <0.001, and **** p <0.0001 compared to saline; n.s. for all PTZ-treated groups relative to each other). E) Latency to first seizure (time to S3 or greater on modified Racine scale) (** p <0.01, *** p <0.001, and **** p <0.0001 compared to saline; n.s. for all PTZ-treated groups relative to each other).

Summary data and statistical comparisons used in the Results

Mossy cell ablation by AAV5-flex-taCasp3 in dorsal hippocampus.

A) Representative sections from dorsal hippocampus from virus control and mossy cell (MC) ablated mice, 6 weeks after AAV5-flex-taCasp3 virus injection. Tissue was stained for glutamate receptor AMPA type subunit 2 (GluA2, red) and co-stained for DAPI (blue) to visualize the dentate granule cell layer and CA3 pyramidal cell layer, which also express GluA2. B) GluA2+ hilar cell density is reduced in MC ablated (N=4 mice) mice compared to virus control (N=5 mice) mice (** p>0.01).

Minimal off-target viral expression in CA3 neurons.

A) Representative images of DAPI-stained pyramidal cell nuclei in proximal CA3 from virus control and mossy cell (MC) ablated mice. Yellow dashed lines represent borders of the pyramidal cell layer (SP), stratum oriens (OR) and stratum radiatum (SR). B) CA3 cell densities were similar between virus control (N= 4 mice) and MC ablated conditions (N= 4 mice), suggesting minimal ablation of CA3 pyramidal neurons following AAV5-flex-taCasp3 virus injection (p=0.17). C) Representative images of GFP-expressing neurons in CA3 from Crlr-Cre mice injected with AAV5-flex-GFP or AAV5-flex-TeLC-GFP viruses, co-stained with DAPI. D) Cell densities of both GFP-negative (GFP control, N= 3 mice; MC Silenced, N= 3 mice; p=0.60) and GFP-positive (p=0.57) were not different between experimental groups.

Timecourse of microglial activation after mossy cell ablation.

A) Representative dentate sections 1, 2, or 3 weeks following AAV flex Casp3 injections into Crlr-Cre-negative (Virus control) or Crlr-Cre-positive (MC ablated) mice. Microglia are stained with anti-Iba1 (green), with activated microglia identified by anti-Gal3 (red) staining. B) Iba1-positive cell densities at different timepoints after virus injection; (Virus control n = 4 mice at 1 week post-virus, 4 mice at 2 weeks, and 4 mice at 3 weeks; MC ablated n = 6 mice at 1 week, 5 mice at 2 weeks, and 4 mice at 3 weeks (**** p <0.0001 at 1 week; n.s. at 2 and 3 weeks). C) Gal3-positive cell densities at different timepoints after virus injection (* p <0.05 at 1 week; n.s. at 2 and 3 weeks). D) Representative dentate sections 1, 2, and 3 weeks following AAV flex GFP (GFP control) or AAV flex TeLC-GFP (MC silenced) injections into Crlr-Cre-positive mice. Microglia are stained with anti-Iba1 (green, pseudocolored), with activated microglia identified using anti-Gal3 (red) staining. E) Iba1-positive cell densities at different timepoints after virus injection (GFP control n = 3 mice at 1 week post-virus, 3 mice at 2 weeks, and 4 mice at 3 weeks; MC silenced= 3 mice at 1 week, 3 mice at 2 weeks, and 3 mice at 3 weeks (n.s. at all timepoints). F) Gal3-positive cell densities at different timepoints after control virus injection or MC silencing (n.s. at all timepoints).

Cell proliferation and survival in the dentate granule cell layer are not altered following mossy cell manipulations.

A) Representative mitotic labeling in virus controls, MC ablated, and MC silenced conditions 24hr after BrdU administration to assess cell proliferation. B&C) Cell proliferation and outward migration in virus control and MC ablated conditions (Virus control n = 9 mice, MC ablated n = 5 mice; n.s. both measures). D&E) Cell proliferation and outward migration in GFP control and MC silenced conditions (GFP control n = 8 mice, MC silenced n = 9 mice; n.s. both measures). F) Representative BrdU labeling in virus controls, MC ablated, and MC silenced conditions 21 days after BrdU administration to assess newborn cell survival. G&H) BrdU-positive cell density and outward migration distance in virus control and MC ablated conditions, 21 days after BrdU administration (Virus control n = 6 mice, MC ablated n = 9 mice; n.s. both measures). I&J) BrdU-positive cell density and outward migration distance in GFP control and MC silenced conditions, 21 days after BrdU administration (GFP control n = 6 mice, MC silenced n = 5 mice; n.s. both measures).

Differential effects of mossy cell ablation vs silencing on immature granule cells.

A) Representative immature adult-born granule cells, as assessed by doublecortin (DCX) staining (red) in virus control and MC ablated mice. B&C) DCX-positive cell density and outward migration from the hilus, demonstrating no change following mossy cell ablation (Virus control n = 7 mice, MC ablated n = 8 mice; n.s. both measures). D) Representative immature adult-born granule cells (DCX+; red) in GFP control and MC silenced mice. E&F) DCX staining demonstrates a mildly increased density of immature granule cells and increased outward migration of immature cells following mossy cell silencing (GFP control n = 10 mice, MC silenced n = 10 mice; * = p <0.05).

Reduced spontaneous excitatory synaptic currents in mature DGCs after mossy cell ablation or silencing.

A) Representative spontaneous excitatory post-synaptic current (sEPSC) recordings from mature DGCs from virus control, MC ablated, and MC silenced mice. B) Average sEPSC waveforms from the respective recordings in A. C&D) sEPSC frequencies and amplitudes recorded from mature DGCs from virus control, MC ablated, and MC silenced mice. (sEPSC frequency * = p < 0.05, ** p <0.01; sEPSC amplitude n.s. all comparisons).

sEPSCs are unchanged in immature DGCs generated after mossy cell ablation or silencing.

A) Representative sEPSC recordings from 21 day-old adult-born DGCs from virus control, MC ablated, and MC silenced mice. B) Average sEPSC waveforms from the respective recordings in A. C&D) sEPSC frequencies and amplitudes in 21 day old DGCs from virus control, MC ablated, and MC silenced mice. Neither sEPSC frequency or amplitude was different in adult-born cells in each group. (n.s. all comparisons).

Neither mossy cell ablation or silencing drive recurrent granule cell axon (mossy fiber) sprouting.

A) Representative 5X (top) and 40X images (bottom) of mossy cell axon (calretinin, red) and granule cell mossy fiber bouton (ZnT3, green) staining in virus control, MC ablated, and MC silenced conditions 6 weeks after virus injection. B) Mossy fiber sprouting (MFS) is not detected in either MC ablated or MC silenced groups (Virus control n = 8 mice, MC ablated n = 4 mice, MC silenced n = 8 mice; n.s. all groups).

Calretinin-positive IML width decreases during the first 2 weeks after mossy cell ablation and remains stable at later time points.