Ventral hippocampus neurons encode meal-related memory

Abstract

The ability to encode and retrieve meal-related information is critical to efficiently guide energy acquisition and consumption, yet the underlying neural processes remain elusive. Here we reveal that ventral hippocampus (HPCv) neuronal activity dynamically elevates between eating bouts during meal consumption and this response is predictive of performance in a foraging-related memory test for the spatial location of a previously consumed meal. Targeted recombination-mediated ablation of HPCv meal-responsive neurons impairs meal location memory without influencing food motivation or spatial memory for escape location. These HPCv meal-responsive neurons project to the lateral hypothalamic area (LHA) and are enriched in serotonin 2a receptors (5HT2aR). Either chemogenetic silencing of HPCv-to-LHA projections or intra-HPCv 5HT2aR antagonist yielded meal location memory deficits, as well as increased caloric intake driven by shorter temporal intervals between meals. Collective results identify a population of HPCv neurons in male rats that dynamically respond during eating to encode meal-related memories.

Fig. 1. Dynamic elevations in ventral hippocampus calcium-dependent activity between eating bouts during meal consumption are predictive of performance during a foraging-related spatial memory task.

a Diagram of viral approach for fiber photometry, adapted from the Brain Maps 4.0 structure of the rat brain. b Representative photomicrograph of viral expression and optic fiber placement in the ventral CA1 relative to the alveus (alv); scale bar 100 μm (representative pattern was observed in all animals used in the analyses, n = 5). c Representative trace of a single animal of the increase in ventral CA1 calcium-dependent activity during the interbout intervals (purple). d Average change in z-score for fluorescence over the course of an eating bout versus an interval (n = 5 animals)(p = 0.0341). e Foraging-related spatial memory task apparatus (created in BioRender. Kanoski, S. (2025) https://BioRender.com/5vog18n). f Simple linear regression of the increase during interbout intervals predictive of subsequent performance in a separate foraging-related spatial memory task. g Diagram of the viral approach for 4OHT-inducible expression of green fluorescent protein (GFP) in ventral CA1 neurons active in a Fasted or Fed state, adapted from the Brain Maps 4.0 structure of the rat brain. h Timeline for intraperitoneal (ip) injection of 4-hydroxitamoxifen (4OHT) under the Fasted, Fed or Coyote state. i Representative photomicrographs of the pyramidal layer of the ventral CA1 (CA1v(sp)) with GFP⁺ cell bodies from neurons that were active in the Fasted, Fed or Coyote state, relative to the alveus (alv); scare bar 100 μm. j Quantification of ventral CA1 cell bodies labeled with GFP using the TRAP method under the Fasted (n = 5), Fed (n = 4) or Coyote (n = 5) state (p value for Group = 0.0130; p value for Fasted^GFP vs. Fed^GFP = 0.0120). Data are presented as mean ± SEM. For Fig. 1d, two-tailed paired t-test (n = 5). For Fig. 1j, one-way ANOVA, Tukey post hoc *p < 0.05. No symbol indicates lack of significance.

Fig. 2. Ablation of ventral hippocampus meal-responsive neurons selectively impairs foraging-related spatial memory.

a Diagram of viral approach for 4-hydroxytamoxifen (4OHT)-inducible Cre-dependent expression of diphteria toxin (dTA) (Fasted^dTA and Fed^dTA) or green fluorescent protein (Control^GFP) in ventral CA1 neurons active in the Fasted or Fed state, adapted from the Brain Maps 4.0 structure of the rat brain. b Diagram of Barnes maze apparatus for the foraging-related spatial memory task for the Control^GFPvs Fasted^dTA vs Fed^dTA groups (created in BioRender. Kanoski, S. (2025) https://BioRender.com/5vog18n). c Diagram of Barnes maze apparatus for the escape-based spatial memory task for the Control^GFP vs Fed^dTA groups (created in BioRender. Kanoski, S. (2025) https://BioRender.com/ani3u4u). d Average number of errors and e latency to find the food during training for the foraging-related spatial memory task for the Control^GFP (n = 8) vs Fasted^dTA (n = 9) vs Fed^dTA (n = 10) groups. f Performance index during the probe for the foraging-related spatial memory task for the Control^GFP (n = 8) vs Fasted^dTA (n = 9) vs Fed^dTA (n = 10) groups (p value for Group = 0.0009; p value for Control^GFP vs. Fed^dTA = 0.0239; p value for Fasted^dTA vs. Fed^dTA = 0.0011; p value for 1-sample t-test for Control^GFP = 0.0023; p value for 1-sample t-test for Fasted^dTA < 0.0001; p value for 1-sample t-test for Fed^dTA = 0.0009). g Average number of errors and h latency to find the food during training for the foraging-related spatial memory task for the Coyote^GFP (n = 7) vs Coyote^dTA (n = 7) groups. i Performance index during the probe for the foraging-related spatial memory task for the Coyote^GFP (n = 7) vs Coyote^dTA (n = 7) groups (p value for 1-sample t-test for Coyote^GFP = 0.0225; p value for 1-sample t-test for Coyote^dTA = 0.0183). j Average number of errors and k latency to find the escape box during training for the escape-based spatial memory task for the Control^GFP (n = 7) vs Fed^dTA (n = 6) groups. l Performance index during the probe for the escape-based spatial memory task for the Fed^GFP (n = 7) vs Fed^dTA (n = 6) groups (p value for 1-sample t-test for Control^GFP < 0.0001; p value for 1-sample t-test for Fed^dTA = 0.0020). Data are presented as mean ± SEM. For d, e, g, h and j, k, two-way ANOVA. For f, Kruskal-Wallis, multiple comparisons. For i and l, two-tailed unpaired t-test; *p < 0.05, **p < 0.01. For f, i and l, one-sample t-test, different from chance set at 0.1667; ^#p < 0.05, ^##p < 0.01, ^###p < 0.005, ^####p < 0.001. No symbol indicates lack of significance.

Fig. 3. Ventral hippocampus meal-responsive neurons project to the lateral hypothalamic area.

a Left: Diagram of a coronal section of the nucleus accumbens (ACB), adapted from the Brain Maps 4.0 structure of the rat brain. Middle: Representative photomicrographs of axonal green fluorescent protein (GFP)⁺ expression in the ACB from ventral CA1 neurons active in the Fasted (Fasted^GFP)(n = 5) or Fed (Fed^GFP)(n = 4) state, relative to the anterior commissure (ac); scale bar 100 μm. Right: Average number of GFP⁺ pixels in the ACB. b Left: Diagram of a coronal section of the lateral septum (LS), adapted from the Brain Maps 4.0 structure of the rat brain. Middle: Representative photomicrographs of axonal GFP⁺ expression in the LS from ventral CA1 neurons active in the Fasted (Fasted^GFP)(n = 5) or Fed (Fed^GFP)(n = 4) state, relative to the lateral ventricle (lv); scale bar 100 μm. Right: Average number of GFP⁺ pixels in the LS. c Left: Diagram of a coronal section of the lateral hypothalamic area (LHA), adapted from the Brain Maps 4.0 structure of the rat brain. Middle: Representative photomicrographs of axonal GFP⁺ expression in the LHA from ventral CA1 neurons active in the Fasted (Fasted^GFP)(n = 5) or Fed (Fed^GFP)(n = 4) state, relative to the fornix (fx) and 3rd ventricle (3 v); scale bar 100 μm. Right: Average number of GFP⁺ pixels in the LHA (p value = 0.0159). d Left: Representative photomicrograph of retrograde GFP viral expression in the LS, relative to the lv; scale bar 200 μm. Right: Representative photomicrograph of retrograde mCherry viral expression in the LHA, relative to the fx; scale bar 200 μm. e Average number of LS (GFP)(n = 4) and LHA (mCherry)(n = 4)-projecting neurons labeled in the CA1v. f Average percentage of CA1v neurons sending axonal projections to both the LS (GFP)(n = 4) and LHA (mCherry)(n = 4). g Average percentage of LS (GFP)(n = 4) and LHA (mCherry)(n = 4)-projecting CA1v neurons expressing cFos in the Fed state (p value < 0.0001). h Representative photomicrograph of fluorescent in situ hybridization for Fos (cyan) in LS (GFP) and LHA (mCherry)-projecting neurons of the CA1v, relative to the alveus (alv); scale bar 100 μm. Data are presented as mean ± SEM. For a–c and d–g, two-tailed unpaired t-test; *p < 0.05, ****p < 0.0001. No symbol indicates lack of significance.

Fig. 4. Ventral hippocampus neurons projecting to the lateral hypothalamic area increase activity between eating bouts and promote foraging-related spatial memory.

a Diagram of viral approach for expression of GCaMP7s in ventral CA1 neurons projecting to the lateral hypothalamic area and implantation of the optic fiber, adapted from the Brain Maps 4.0 structure of the rat brain. b Representative photomicrograph of GCaMP7s expression for placement validation of viral injections, relative to the alveus (alv); scale bar 100 μm (representative pattern was observed in all animals used in the analyses, n = 3). c Representative trace of a single animal of the increase in calcium-dependent activity during the interbout intervals (purple) in ventral hippocampus neurons projecting to the lateral hypothalamic area. d Average change in z-score for fluorescence over the course of an eating bout versus during an interbout interval (n = 3 animals)(p value = 0.0165). e Diagram of viral approach for expression of hM4Di receptors in ventral CA1 neurons projecting to the lateral hypothalamic area and administration of vehicle (VEH) or clozapine-N-oxyde (CNO) through a lateral ventricle (LV) cannula, adapted from the Brain Maps 4.0 structure of the rat brain. f Representative photomicrograph of mCherry expression for placement validation of viral injections, relative to the alveus (alv); scale bar 100 μm (representative pattern was observed in all animals used in the analyses, n = 10). g Average number of errors and h latency to find the food during training for the foraging-related spatial memory task in animals assigned to VEH (n = 6) or CNO (n = 4) groups. i Performance index during the probe for the foraging-related spatial memory task, 1 h following LV administration of VEH (n = 6) or CNO (n = 4)(Group comparison p value = 0.0024; p value for 1-sample t-test = 0.0014 for VEH group). Data are presented as mean ± SEM. For d, two-tailed paired t-test. For g–h, two-way ANOVA. For i, two-tailed unpaired t-test (n = 4–6/group); *p < 0.05, **p < 0.01. For i, one-sample t-test, different from chance set at 0.1667; ^##p < 0.01. No symbol indicates lack of significance.

Fig. 5. Meal consumption alters the transcriptional profile of ventral hippocampus endothelial cells and CA1v excitatory neurons and engages 5HT2aR expressing cells.

a Uniform manifold approximation and projection (UMAP) of the ventral hippocampus (HPCv) samples identifying 17 clusters. b Annotation of cellular subtypes with known makers of HPCv cellular subtypes. The size and color of dots are proportional to the percentage of cells expressing the gene (Pct. Exp) and the average expression levels of the gene (Avg. Exp.), respectively. The cluster numbers and colors are matched to that of the UMAP. c Volcano plot depicting the number of significant differential expression events induced by meal consumption (Wilcoxon rank sum test, bonferroni correction; p values adjusted for multiple testing). d Number of genes with meal consumption-altered increased (right) or decreased (left) expression per cluster. e Heatmaps of select genes enriched in Fos+ cells under a Fasted or Fed state.

Fig. 6. Ventral hippocampus 5HT2aR expressing neurons are engaged by a meal, project to the lateral hypothalamic rea, and are functionally required for foraging-related spatial memory.

a Representative photomicrograph of fluorescent in situ hybridization for Fos (red) and Htr2a (green) in the ventral CA1, relative to the alveus (alv); scale bar 100 μm. b Percentage of ventral CA1 Fos+ cells that co-express Htr2a in rats perfused under a Fasted (n = 3) or Fed state (n = 4)(p value = 0.0463). c Diagram of approach for ventral CA1 administration of vehicle (VEH) or the 5HT2aR antagonist M100907 (1 g/hemisphere), adapted from the Brain Maps 4.0 structure of the rat brain. d Average number of errors and e latency to find food during the training for the foraging-related spatial memory task for animals assigned to group VEH (n = 9) or M100907 (n = 7). f Performance during the probe for the foraging-related spatial memory task, 5 min following ventral CA1 infusion of VEH (n = 9) or M100907 (n = 7)(Group comparison p value = 0.0453; p value for 1-sample t-test = 0.0064 for Group VEH). Data are presented as mean ± SEM. For b, f, two-tailed unpaired t-test. For d, e, two-way ANOVA; *p < 0.05. For f, one-sample t-test, different from chance set at 0.1667; ^##p < 0.01. No symbol indicates lack of significance.

Fig. 7. Blockade of ventral hippocampus to lateral hypothalamic area signaling or 5HT2aR signaling increases food intake by reducing temporal intervals between meals during spontaneous feeding.

a Average 2 h chow intake (p = 0.0290), b meal size, c meal frequency and d inter-meal intervals following ventral CA1 administration of vehicle (VEH)(n = 6) or M100907 (n = 6)(p = 0.0171). e Average 6 h chow intake (p = 0.0052), f meal size, g meal frequency (p = 0.0010), and h inter-meal intervals following lateral ventricle administration of vehicle (VEH)(n = 7) or clozapine-N-oxide (CNO)(n = 7) (p = 0.0081) in rats expressing hM4Di receptors in ventral CA1 neurons projecting to the lateral hypothalamic area. Data are presented as mean ± SEM. Two-tailed paired t-test; *p < 0.05, **p < 0.01, ***p < 0.001. No symbol indicates lack of significance.