A leptin-responsive hypothalamic circuit inputs to the circadian feeding network

Abstract

Salient cues, such as the rising sun or the availability of food, play a crucial role in entraining biological clocks, allowing for effective behavioral adaptation and ultimately, survival. While the light-dependent entrainment of the central circadian pacemaker (suprachiasmatic nucleus, SCN) is relatively well defined, the molecular and neural mechanisms underlying entrainment associated with food availability remains elusive. Using single nucleus RNA sequencing during scheduled feeding (SF), we identified a leptin receptor (LepR) expressing neuron population in the dorsomedial hypothalamus (DMH) that upregulates circadian entrainment genes and exhibits rhythmic calcium activity prior to an anticipated meal. We found that disrupting DMH^LepR neuron activity had a profound impact on both molecular and behavioral food entrainment. Specifically, silencing DMH^LepR neurons, mis-timed exogenous leptin administration, or mis-timed chemogenetic stimulation of these neurons all interfered with the development of food entrainment. In a state of energy abundance, repetitive activation of DMH^LepR neurons led to the partitioning of a secondary bout of circadian locomotor activity that was in phase with the stimulation and dependent on an intact SCN. Lastly, we discovered that a subpopulation of DMH^LepR neurons project to the SCN with the capacity to influence the phase of the circadian clock. This leptin regulated circuit serves as a point of integration between the metabolic and circadian systems, facilitating the anticipation of meal times.

Keywords: Calcium imaging; Chemogenetics; Circadian; Dorsomedial hypothalamus; Food entrainment; Leptin; RNA sequencing; Scheduled feeding.

Figure 1.. SCN snRNAseq íeveals minimal alteíation of ciícadian genes duíing SÏ

A. Diagram illustrating that food timing as a potent zeitgeber entraining an oscillatory network system in the brain and peripheral organs, relaying rhythmic behavior outputs. B. The ZT phase of the first bioluminescence peak of SCN and liver from PER2::Luciferase mice that are either provided with scheduled food access for 4 days at ZT6 or ad libitum fed controls (untreated or given ZT6 saline injections). Two-way ANOVA with Bonferroni post hoc comparison; n = 10–11 / group; F_treatment (1, 38) = 19.05, p<0.001. C. Schematic of experimental design. Mice were housed on a 12–12 light-dark cycle and either fed ad libitum, overnight fasted, or provided a scheduled meal for 10 days at ZT6. Blue shading denotes food access. All mice were sacrificed for tissue collection at ZT5. D. Normalized locomotor activity starting 29 hours before tissue collection. n=4 mice / condition. Data are represented as mean ± SEM. E. Schematic of single nuclei RNA sequencing (snRNAseq) workflow using 10X Genomics. F. Representative images illustrating the area of dissection in the SCN for snRNAseq. G. Uniform Manifold Approximation and Projection (UMAP) plot of 8 molecularly distinct SCN neuron subtypes (n=8,957 neurons). H. Heatmap of cluster-average marker gene expression, scaled by gene. I. Dot plot of average expression level (dot color) and percent expression (dot size) for each SCN neuron cluster. Genes shown were previously defined as SCN markers ^, and validated based on Allen Brain Atlas Mouse Brain in situ hybridization data . J. Kyoto Encyclopedia of Genes and Genomes (KEGG) from the Mouse 2019 database comparing top 5 pathways up- and down-regulated among feeding conditions in all SCN neurons. Inclusion criteria required p-value <0.05 and log2 fold change >0.25. See also supplemental figure 1.

Figure 2.. SF alters circadian entrainment genes in specific DMH neuron subtypes

A. Representative images illustrating the DMH area dissected for snRNAseq. B. UMAP of 14 defined DMH neuron subtypes (n=16,281 neurons). C. Average gene expression heatmap labeled by cluster-specific markers in the DMH. D. Dot plot of average expression level (dot color) and percent expression (dot size) of genes of interest within DMH clusters. These genes were either previously identified in DMH ^,,, or validated as DMH markers by the Allen Brain Atlas Mouse Brain in situ hybridization data. E. KEGG from the Mouse 2019 database comparing top 5 pathways up- and down-regulated across feeding conditions in all DMH neurons. Inclusion criteria required p-value <0.05 and log2 fold change >0.25. F. DMH clusters with differentially regulated circadian entrainment pathways in at least one scheduled feeding comparison. G. Feature plots indicating spatial expression of Lepr (left), Glra2 (middle), Pdyn (right), in DMH clusters. H. Representative coronal section image localizing expression of LepR, Pdyn, and Glra2 in the DMH. LepR cells were marked by LepR-Cre;TdTomato protein, whereas Pdyn and Glra2 transcripts were visualized by RNA FISH. See also supplemental Fig 3G for zoomed-out view of the same brain section. I. Representative RNA FISH coronal section image showing Lepr and Pdyn transcripts in the DMH. Quantification of Lepr and Pdyn co-expressing cells is depicted at the bottom. n=3 mice. J. Heatmap of select genes that were differentially expressed across feeding conditions in DMH^LepR neurons. See also supplemental figure 1 and 2.

Figure 3-. Leptin sensitive DMH^LepR neurons exhibit food entrainable calcium activity patterns which correlate with FAA

A. Schematic diagram illustrating unilateral injection of AAV-hSyn-DIO-GCaMP7s and fiber optic cannula implantation to the DMH of LepR Cre mice. B. Example data trace illustrating two readouts of long-term fiber photometry calcium imaging. “Tonic calcium signal” represents the total fluorophore brightness. “Phasic calcium signal” represents the intracellular calcium activity over baseline activity in a given recording session. C. Representative locomotor actogram of single animals treated with saline (left) or leptin (right) during DMH^LepR neuron GCaMP7s fiber photometry recording. Mice are housed in 12:12 LD, fasted at lights off on day 5 (solid red line), injected with saline or leptin at 2.5 hours after lights on (ZT2.5, solid blue line), and fed at ZT6 (2 g on days 1 & 2, 2.5 g on remaining days). Red shaded area is the FAA time window 2 hours pre-meal time (ZT4–6), and blue shaded area is the first 2 hours after food delivery (ZT6–8). Food is restored at ZT10 on day 6 of SF. Color coded arrows indicate three days that are selected for quantification in panels E-H. D. Quantification of FAA during long-term DMH^LepR neuron GCaMP7s recording. FAA is defined as the locomotor activity in the two-hour window prior to food delivery as a percentage of 24-hour activity. AL indicates ad libitum condition two days prior to initiation of drug administration. Mixed-effects (REML) analysis with Bonferroni post hoc comparison; n = 5–8 / group; F_treatment (1, 52) = 25.44, p<0.001. E. Average tonic calcium signal of DMH^LepR neurons from saline control group 2 days before SF (black, ad libitum), 5th day during treatment (red, SF), and 6th day where saline injection was withheld and food delivery was delayed for 3.5 hours (green, SF-LF: late feeding). F. Quantification of the tonic calcium signal from saline treated mice in the FAA window (average of ZT5–6) from (E). Mixed-effects (REML) analysis with Bonferroni post hoc comparison; n = 6–7 / group; F (2, 16) = 12.27, p=0.0006. G. Average tonic calcium signal of DMH^LepR neurons from leptin group 2 days before SF (black, ad libitum), 5th day during treatment (red, SF), and 6th day where leptin injection was withheld and food delivery was delayed for 3.5 hours (green, SF-LF: late feeding). H. Quantification of the tonic calcium signal from leptin treated mice in the FAA window (average of ZT5–6) from (G). Repeated measures one-way ANOVA with Bonferroni post hoc comparison; n = 8 / group; F (2, 14) = 3.596, p=0.0549. I. Quantification of the development of the tonic calcium signal during FAA. AL indicates ad libitum condition two days prior to initiation of drug administration. Mixed-effects (REML) analysis with Bonferroni post hoc comparison; n = 6–8 / group; F_treatment (1, 13) = 4.744, p=0.0484. J. Average phasic calcium signal of DMH^LepR neurons from saline control group 2 days before SF (black, ad libitum), 5th day during treatment (red, SF), and 6th day where saline injection was withheld and food delivery was delayed for 3.5 hours (green, SF-LF: late feeding). K. Quantification of the phasic calcium signal from saline treated mice in the FAA window (average of ZT5–6) from (J). Mixed-effects (REML) analysis with Bonferroni post hoc comparison; n = 6–7 / group; F (2, 10) = 15.01, p=0.0010. L. Average phasic calcium signal of DMH^LepR neurons from leptin group 2 days before SF (black, ad libitum), 5th day during treatment (red, SF), and 6th day where leptin injection was withheld and food delivery was delayed for 3.5 hours (green, SF-LF: late feeding). M. Quantification of the phasic calcium signal from leptin treated mice in the FAA window (average of ZT5–6) from (L). Repeated measures one-way ANOVA with Bonferroni post hoc comparison; n = 8 / group; F (2, 14) = 2.508, p=0.1172. N. Quantification of the development of the phasic calcium signal during FAA. AL indicates ad libitum condition two days píioí to initiation of díug administíation. Mixed-effects (REML) analysis with Bonferroni post hoc comparison; n = 6–8 / group; F_{treatment * time} (4, 50) = 8.834, p<0.0001. O. Summary diagram illustrating the observation of calcium activity pattern in DMH^LepR neurons during SF, in animals treated with saline control or mis-timed leptin. Data are represented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant. See also supplemental figure 3.

Figure 4-. Silencing of DMH^LepR neurons impairs FAA

A. Schematic diagram illustrating DMH^LepR neuron silencing by bilateral injection of AAV-CBA-DIO-GFP-TeTx to the DMH of LepR Cre mice. B. Representative images showing the expression of GFP-TeTx in DMH^LepR neurons. C-D. Representative actograms of (C) DMH^LepR mCherry and (D) DMH^LepR TeTx mice on SF. Shading color scheme is described in Fig 3C. See supplemental Fig 3 for actograms of all animals. E. Quantification of FAA. FAA is defined as the locomotor activity in the two-hour window prior to food delivery as a percentage of light-phase activity. Repeated measures two-way ANOVA with Bonferroni post hoc comparison; n = 5–6 / group; F_virus (1, 9) = 33.00, p=0.0003. Data are represented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant. See also supplemental figure 3.

Figure 5-. Mis-timed leptin or activation of DMH^LepR neurons suppresses the development but not maintenance of food entrainment

A-B. Representative actograms of two mice on a 12:12 L:D cycle under ad libitum conditions that are then subjected to scheduled feeding (SF) beginning at lights off on SF day 0 and receiving either (A) saline (SF days 1–5) then 5mg/kg leptin (SF days 6–10) or (B) 5mg/kg leptin (SF days 1–5) then saline (SF days 6–10). Shading color scheme is described in Fig 3C. See supplemental Fig 4A-B for actograms of all animals. C. Quantification of FAA. FAA is defined as the locomotor activity in the two-hour window prior to food delivery as a percentage of light-phase activity excluding the activity one hour post-injection. Note that red (leptin) or black (saline) data markers indicate the treatment for the day while the data connecting lines identify the initial leptin (red) or saline (black) treatment groups. Repeated measures two-way ANOVA with Bonferroni post hoc comparison; n = 8–9 / group; F_treatment (1, 15) = 22.96, p<0.001. D. Food intake of mice during SF 1 hour after food delivery. Blue shading indicates the total amount of food that was available for mice to consume on each day. Repeated measures two-way ANOVA with Bonferroni post hoc comparison; n = 9 / group; F_{time*treatment} (9, 144) = 10.15, p<0.0001. E-F. First hour food intake on day 1 (E) and day 6 (F) from D. n = 9 / group, Unpaired Student’s t test. Notably, the same treatment groups are illustrated by the same color on the same side of the graphs. G. Proposed model illustrating the suppressive effect of leptin on food entrainment, which in turn leads to both decreased locomotion and appetite. H. Schematic diagram illustrating bilateral injection of AAV-hSyn-DIO-hM3Dq-mCherry or AAV-hSyn-DIO-mCherry to the DMH of LepR Cre mice. I. Representative images showing the expression of mCherry (top) or hM3Dq-mCherry (bottom) in DMH^LepR neurons and c-Fos response 2 hours after CNO injection. J-K. Representative actograms of (J) DMH^LepR mCherry and (K) DMH^LepR hM3Dq mice on SF that received 0.3mg/kg CNO (SF days 1–5), saline (SF days 6–10), and 0.3 mg/kg CNO (SF days 11–15) injection at ZT2.5. Shading color scheme is described in Fig 3C. See supplemental Fig 5 for actograms of all animals. L. Quantification of FAA. Pink shading indicates days with CNO injection. No shading indicates saline injection. Repeated measures two-way ANOVA with Bonferroni post hoc comparison; n = 8–10 / group; F_virus (1, 16) = 15.98, p=0.0010. Data are represented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant. See also supplemental figure 4, 5 and 6.

Figure 6-. Repetitive activation of DMH^LepR neurons alters circadian locomotor activity

A-B. Representative actograms of LepR Cre animals bilaterally injected with (A) AAV-hSyn-DIO-mCherry or (B) AAV-hSyn-DIO-hM3Dq-mCherry and injected with 0.3mg/kg CNO at ZT6 (solid blue line) in 12–12 LD with access to a running wheel. Pink shading represents one hour prior to injection, and blue shading 3 hours after injection. See supplemental Fig 7A-B for actograms of all animals. C. Light phase (Day) wheel revolutions of DMH^LepR mCherry and DMH^LepR hM3Dq animals before, during, and after CNO injections at ZT6 in 12–12 LD. Pink shading represents days with CNO injection. Repeated measures two-way ANOVA; n = 5–6 / group; F_{virust * time}(19, 171) = 3.302, p<0.001. D. Dark phase (Night) wheel revolutions of DMH^LepR mCherry and DMH^LepR hM3Dq animals before, during, and after CNO injections at ZT6 in 12–12 LD. Pink shading represents days with CNO injection. Repeated measures two-way ANOVA; n = 5–6 / group; F_{virust * time}(19, 171) = 1.707, p=0.0390. E. Average wheel running activity induced by chemogenetic activation of DMH^LepR neurons at ZT6 in 12–12 LD. Color coded time window is indicated in (A-B). Repeated measures two-way ANOVA; n = 5–6 / group; F_{virust * time}(79, 711) = 2.519, p<0.001. F. Quantification of activity for 5 days after cessation of CNO injections. Color coded time window is indicated in (A-B). Repeated measures two-way ANOVA; n = 5–6 / group; Fvirus (1, 9) = 0.9519, p=0.3547; Fvirus*time (79, 711) = 1.202, p=0.1215. G-H. Representative actograms of LepR Cre animals bilaterally injected with (G) AAV-hSyn-DIO-mCherry or (H) AAV-hSyn-DIO-hM3Dq-mCherry and injected with 0.3mg/kg CNO at ~CT14 (solid blue line), in ad libitum, constant dark conditions with access to a running wheel. Pink shading represents one hour prior to injection, and blue shading 6 hours after injection. See supplemental Fig 7D-E for actograms of all animals. I. 24-hour total wheel revolutions of DMH^LepR mCherry and DMH^LepR hM3Dq animals before, during, and after CNO injections at ~CT14. Animals were housed under ad libitum, constant dark conditions. Pink shading represents days with CNO injection. Repeated measures two-way ANOVA; n = 9–12 / group; F_{virust * time} (29, 551) = 2.468, p<0.001. J. Average wheel running activity induced by chemogenetic activation of DMH^LepR neurons at ~CT14. Color coded time window is indicated in (E-F). Repeated measures two-way ANOVA; n = 9–12 / group; F_{virust * time}(140, 2660) = 2.114, p<0.001. K-L. Representative actograms of LepR Cre animals bilaterally injected with (K) AAV-hSyn-DIO-mCherry or (L) AAV-hSyn-DIO-hM3Dq-mCherry and injected with 0.3mg/kg CNO at ~CT22 (solid blue line), in ad libitum, constant dark conditions with access to a running wheel. Pink shading represents one hour prior to injection, and blue shading 3 hours after injection. See supplemental Fig 7F-G for actograms of all animals. M. 24-hour total wheel revolutions of DMH^LepR mCherry and DMH^LepR hM3Dq animals before, during, and after CNO injections at ~CT22. Animals were housed under ad libitum, constant dark conditions. Pink shading represents days with CNO injection. Repeated measures two-way ANOVA; n = 9–12 / group; F_virus (1, 19) = 0.2061, p=0.6550; F_virus*time (29, 551) = 0.7120, p=0.8676. N. Average wheel running activity induced by chemogenetic activation of DMH^LepR neurons at ~CT22. Color coded time window is indicated in (K-L). Repeated measures two-way ANOVA; n = 9–12 / group; Fvirus (1, 18) = 8.580, p=0.0090; Fvirus*time (80, 1440) = 2.029, p<0.001. O. Quantification of sustained activity for 5 days after cessation of CNO injections. Color coded time window is indicated in (K-L). Repeated measures two-way ANOVA; n = 8–12 / group; Fvirus*time (80, 1440) = 1.745, p<0.001. Data are represented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant.

Figure 7-. DMH^LepR neurons alter circadian behavior via the SCN

A. Schematic diagram illustrating bilateral injection of AAV-hSyn-DIO-hM3Dq-mCherry or AAV-hSyn-DIO-mCherry to the DMH of LepR Cre mice coupled with electrolytic lesioning of the SCN in the same mice. B-C. Representative DAPI staining images of (B) intact and (C) electrolytic lesioned SCN. D-E. Representative actograms of an SCN lesioned DMH^LepR (D) mCherry, and (E) DMH^LepR hM3Dq animal injected with CNO every 24 hours for 10 days in constant darkness. Solid blue line indicates CNO injection. See supplemental Fig 7H-I for actograms of all animals. F. Average wheel running activity induced by chemogenetic activation of DMH^LepR neurons from (D-E). Color coded time window is indicated in (D-E). Repeated measures two-way ANOVA; n = 5–6 / group; Fvirus (1, 9) = 8.245, p=0.0184; F_virust*time (79, 711) = 2.432, p<0.0001. G. Quantification of activity for 5 days after cessation of CNO injections. Color coded time window is indicated in (D-E). Repeated measures two-way ANOVA; n = 5–6 / group; Fvirus (1, 9) = 0.3450, p=0.5714; Fvirus*time (79, 711) = 1.047, p=0.3746. H. Schematic illustration of NMS Cre dependent rabies virus monosynaptic retrograde tracing. I-J. Representative images of (I) SCN and (J) DMH from retrograde tracing strategy in H. K. The ZT phase of the first bioluminescence peak of SCN and Liver from PER2::luciferase mice injected with CNO at ZT6 (Control), or PER2::luciferase;LepR-Cre mice with hM3Dq expressed in DMH^LepR neurons and injected with CNO at ZT6 (hM3Dq). Two-way ANOVA with Bonferroni post hoc comparison; n = 5–6 / group; F_treatment(1, 19) = 5.361, p=0.0319. L. The ZT phase of the first bioluminescence peak of SCN and Liver from control (mixture of ZT6 saline or untreated) or ZT6 leptin injected PER2::luciferase mice. Two-way ANOVA with Bonferroni post hoc comparison; n = 6–11 / group; F_treatment (1, 30) = 4.772, p=0.0369. Control group is the same dataset as in fig 1B, re-plotted and analyzed with ZT6 leptin treated animals. Data are represented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant.