Leptin receptor neurons in the dorsomedial hypothalamus input to the circadian feeding network

Salient cues, such as the rising sun or availability of food, entrain biological clocks for behavioral adaptation. The mechanisms underlying entrainment to food availability remain elusive. Using single-nucleus RNA sequencing during scheduled feeding, we identified a dorsomedial hypothalamus leptin receptor–expressing (DMHLepR) neuron population that up-regulates circadian entrainment genes and exhibits calcium activity before an anticipated meal. Exogenous leptin, silencing, or chemogenetic stimulation of DMHLepR neurons disrupts the development of molecular and behavioral food entrainment. Repetitive DMHLepR neuron activation leads to the partitioning of a secondary bout of circadian locomotor activity that is in phase with the stimulation and dependent on an intact suprachiasmatic nucleus (SCN). Last, we found a DMHLepR neuron subpopulation that projects to the SCN with the capacity to influence the phase of the circadian clock. This direct DMHLepR-SCN connection is well situated to integrate the metabolic and circadian systems, facilitating mealtime anticipation.

E-F. Acute calcium signal response to food delivery on the 5th day of SF, in successful (E) and mis-targeted (F) mice.The mis-targeted animals in (D) were excluded for the analysis in Fig. 3-4.n=6 mice/group.

G.
Locomotor activity of mis-targeted mice during SF.These mice developed strong FAA.n=5.
H. Average phasic GCaMP7s signal of DMH LepR neurons in the mis-targeted mice at 2 days before SF (black, ad libitum), 5th day of saline treated SF (magenta, SF), or 6th day of SF where saline injection was withheld and food delivery was delayed for 3.5 hours (green, SF-LF).n=6.The unaltered phasic calcium signal of these mis-targeted mice during SF suggests that the pre-meal elevation of phasic calcium signal is not due to artifact of increased locomotor activity.

I.
Average tonic calcium signal (z-score) of DMH LepR neurons from saline control group 2 days before SF (black, ad libitum), 5th day during treatment (magenta, SF), and 6th day where saline injection was withheld and food delivery was delayed for 3.5 hours (green, SF-LF: late feeding).

K.
Average tonic calcium signal (z-score) of DMH LepR neurons from leptin group 2 days before SF (black, ad libitum), 5th day during treatment (magenta, SF), and 6th day where leptin injection was withheld and food delivery was delayed for 3.5 hours (green, SF-LF: late feeding).O. Average phasic calcium signal (second method) of DMH LepR neurons from saline control group 2 days before SF (black, ad libitum), 5th day during treatment (magenta, SF), and 6th day where saline injection was withheld and food delivery was delayed for 3.5 hours (green, SF-LF: late feeding).Normalized to average dark phase signal (ZT12-ZT0).
P. Average phasic calcium signal (second method) of DMH LepR neurons from leptin group 2 days before SF (black, ad libitum), 5th day during treatment (magenta, SF), and 6th day where leptin injection was withheld and food delivery was delayed for 3.5 hours (green, SF-LF: late feeding).

Figure S1 .
Figure S1.SCN and DMH single nuclei RNA-seq quality metrics and marker expression A. UMAP of SCN neurons, colored by sequencing library identity, after batch correction.B. Phylogenetic tree indicating the relatedness of 8 SCN neuronal clusters.C. Expression level distribution of the number of genes per cluster (left) and number of UMIs (unique molecular identifiers, which represent unique gene transcripts; right) among SCN clusters.D. Correlation matrix of average expression of all genes between SCN neuronal clusters.Values within the boxes are Pearson correlation coefficients.E. UMAP of DMH neurons, colored by sequencing library identity, after batch correction.F. Phylogenetic tree indicating the relatedness of 14 DMH neuronal clusters.G. Expression level distribution of the number of genes per cluster (left) and number of UMIs (right) among DMH clusters.H. Correlation matrix of average expression of all genes within DMH neuronal clusters.Values within the boxes are Pearson correlation coefficients.I. KEGG circadian entrainment pathway map.The genes upregulated (Fig. S4C) in DMH LepR neurons during SF are labeled as pink on the map.

Figure S2 .
Figure S2.Relative expression level of KEGG circadian entrainment pathway across feeding conditions in SCN and DMH A. Heatmap of all the genes in the KEGG circadian entrainment pathway across feeding conditions in SCN.B. Heatmap of all the genes in the KEGG circadian entrainment pathway across feeding conditions in DMH.

Figure S3 .
Figure S3.Cluster specific alteration of pathways in the DMH Gene enrichment analysis comparing pathways upregulated among SF vs ad libitum and SF vs fasted feeding conditions in each DMH neuronal cluster and upregulated pathways in all DMH neurons, using KEGG Mouse 2019 database.Bars are colored by clusters.Inclusion criteria required p-value <0.05 and log2 fold change >0.25.

Figure S4 .
Figure S4.Network analysis of differentially expressed genes in the DMH across feeding conditions A. Heatmap indicating expression level of genes up and downregulated among feeding conditions in the 9_Pdyn cluster.

Figure S5 .
Figure S5.DMH LepR neuron calcium recording during SF A-B.Representative image showing pSTAT3 expression in the DMH of Lepr-cre; tdTomato mice sacrificed 90 minutes after saline (A) or 5 mg/kg leptin (B) administration.

L.
Quantification of the tonic calcium signal from leptin treated mice in the FAA window (average of ZT5-6) from (K).Repeated measures one-way ANOVA with Bonferroni post hoc comparison; n = 8 / group; F (2, 14) = 4.346, p=0.0340.M. Quantification of the development of the tonic calcium signal (z-score) 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; Ftreatment (1, 63) = 16.08,p=0.0002.N. Example data trace illustrating second method of calculating phasic calcium signal.See also the Material and Methods section for more details.

Figure S6 .
Figure S6.Pre-FAA leptin suppresses the development but not maintenance of FAA A-B.Actograms of all animals that received (A) saline then leptin or (B) leptin then saline, during scheduled feeding (SF).The actograms of animals that have minimum sum of the square of the residuals to the average FAA value were used as representative figures in Fig. 5, and depicted with red boxes.C. Absolute locomotor activity during the FAA window, 2 hours prior to food delivery time.Repeated measures two-way ANOVA with Bonferroni post hoc comparison; n = 8-9 / group; F treatment (1, 15) = 22.3, p<0.001.

Figure S7 .
Figure S7.Extended duration of food intake does not confound the development of FAA A-B.Actograms of all animals that received (A) control (whole pellet) then extended (same size pellet split up into 4 pellets) or (B) extended then control SF paradigm.Briefly, mice in the extended SF paradigm received one quarter of a pellet of food every hour for 4 hours, while mice in the control SF paradigm received one whole pellet at the first hour (see Methods and (E) for details).C. Quantification of FAA in extended SF experiment.Illustrated as a percentage of light-phase activity, without excluding any light-phase activity.Repeated measures two-way ANOVA with Bonferroni post hoc comparison; n = 6 / group; Ftreatment (1, 10) = 0.004799, p=0.9461.D. Absolute locomotor activity during the FAA window in extended SF experiment.FAA window is defined as 2 hours prior to food delivery time.Repeated measures two-way ANOVA with Bonferroni post hoc comparison; n = 6 / group; Ftreatment (1, 10) = 1.362, p=0.2703.E. Diagram illustrating the comparison between whole pellet and split pallet groups.Data are represented as mean ± SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ns, not significant.