Skip to main content
NCBI home page
CNS Neuroscience & Therapeutics logoLink to CNS Neuroscience & Therapeutics
. 2026 Oct 1;32(10):e71185. doi: 10.1002/cns.71185

Brain‐Wide Connectivity and Topographic Organization of Leptin Receptor Neurons in the Dorsomedial Hypothalamus

Yaxin Hao 1, Xiang Zhang 1, Danyang Zhao 1, Haixing Zhong 2, Yongqiang Chen 1, Heming Li 1, Minghui Ren 1, Mengchu Zhu 1,3,✉, Fang Yuan 1,4,✉
PMCID: PMC13629605  PMID: 42820923

ABSTRACT

Objective

The dorsomedial hypothalamus (DMH) integrates metabolism and autonomic homeostasis, but whether subnuclear functional topography exists remains unclear. This study compared whole‐brain input–output connectomes of leptin receptor (LepRb) neurons in dorsal DMH (DMD) versus DMH to test for connection‐biased segregation.

Methods

Rabies monosynaptic retrograde tracing and Cre‐dependent AAV anterograde tracing in LepRb‐Cre mice mapped and quantified input–output projections of DMDLepRb and DMHLepRb neurons.

Results

Both DMDLepRb and DMHLepRb neurons received inputs from 58 upstream nuclei and projected to 43 downstream targets across the telencephalon, diencephalon, midbrain, pons, and medulla, forming extensive overlapping networks with similar global architectures. Nevertheless, they showed significant connection weight biases: DMDLepRb preferentially integrated arousal, defensive, and emotional signals and projected to stress/defense‐related nuclei, whereas DMHLepRb preferentially integrated viscerosensory and reward‐aversion signals and projected to metabolic/circadian nuclei.

Conclusion

DMDLepRb and DMHLepRb neurons share highly homologous whole‐brain input–output architecture but exhibit local connection‐weight biases, revealing conserved architecture with local specialization and providing a circuit basis for functional heterogeneity.

Keywords: connectivity bias, dorsal part of the DMH, dorsomedial hypothalamus, leptin receptor, neural circuitry

Whole‐brain input–output mapping reveals that DMDLepRb and DMHLepRb neurons share a highly conserved global connectivity architecture yet exhibit local connection biases: DMDLepRb neurons preferentially integrate arousal and defensive signals, whereas DMHLepRb neurons favor viscerosensory and metabolic inputs, uncovering a principle of conserved global topology with local functional specialization within the DMH.

graphic file with name CNS-32-e71185-g005.webp

1. Introduction

The hypothalamus maintains metabolic and autonomic homeostasis. Among its subregions, the dorsomedial hypothalamus (DMH) links metabolic status with behavioral and autonomic outputs and is associated with obesity [1], stress [2], hypertension [3], and sleep disorders [4]. The DMH is functionally heterogeneous, with distinct subregions and neuronal subtypes [5], regulating energy expenditure [6], feeding [7], thermogenesis [8], autonomic activity [9], and arousal [10]. Its dorsal part (DMD) is more associated with defensive responses, stress‐induced sympathetic activation, and thermogenesis, whereas the ventral part (DMV) is more neuroendocrine [11, 12], suggesting a subnuclear functional topography within the DMH, though the circuit basis remains unclear.

Leptin receptor (LepRb)‐expressing neurons represent a major functional population within the DMH. DMHLepRb neurons regulate feeding, energy metabolism, thermoregulation [13, 14, 15], respiration [16], autonomic activity [17], and sleep–wake regulation [18]. They integrate metabolic and environmental information from the arcuate nucleus (Arc) and preoptic area (POA), and project to autonomic/behavioral centers [19, 20]. Distinct DMH subtypes regulate thermogenesis [21], respiration [16], arousal [22], and stress responses [23] via specific projections. However, most studies treat DMHLepRb neurons as homogeneous, overlooking subregional differences. Functional studies show DMVLepRb neurons rapidly inhibit feeding via GABAergic projections [24], while DMDLepRb neurons specialize in thermogenesis and energy expenditure [15, 25]. Whether DMDLepRb neurons have distinct input–output patterns remains unknown.

Neuronal function depends on connectivity: afferent inputs define the information available to neurons, while efferent outputs determine the downstream systems they regulate. Comparing whole‐brain input–output connectivity of DMDLepRb and DMHLepRb neurons is critical for understanding whether connection‐biased topography exists within the DMH. If both populations share a common global framework but exhibit local biases, it would suggest that the DMH achieves refined regulation through conserved architecture with local specialization. This principle is validated in other nuclei, such as the ventromedial hypothalamus (VMH) [26], and limbic system regions [27] as a key mechanism for generating diverse behavioral outputs. The DMD is a well‐characterized DMH subdivision with documented roles in defense, stress‐evoked sympathetic activation, and thermoregulation—functions distinct from the metabolic and neuroendocrine roles of the DMV. This functional divergence makes the DMD a suitable model to test whether subregional specialization arises from segregated circuits or weighted biases within a shared network, an established strategy in hypothalamic circuit research. This study thus focuses on how DMDLepRb neurons integrate into the broader DMH connectome, rather than comprehensively comparing all DMH subnuclei.

Here, we used Cre‐dependent viral tracing, monosynaptic rabies tracing, anterograde tracing, RNAscope, and whole‐brain quantitative analysis to map and compare the input–output networks of DMDLepRb and DMHLepRb neurons. Our results show that both populations share similar global architectures centered on the diencephalon as a bidirectional hub, yet exhibit significant biases in connection weights: DMDLepRb neurons connect more to autonomic/behavior‐related networks, whereas DMHLepRb neurons connect to metabolic networks. These findings suggest a subnuclear topographic organization within the DMH, where distinct subregions regulate different physiological systems via differential connectivity. Collectively, this study reveals the whole‐brain connectivity principles of DMHLepRb neurons and provides a circuit‐based foundation for understanding DMH functional heterogeneity in metabolism, autonomic functions, and behavior.

2. Method

2.1. Animals

Adult LepRb‐Cre (Jackson Laboratory, #008320) and C57BL/6J mice (Beijing Vital River), both sexes, 8–10 weeks old, were housed on a 12‐h light/dark cycle (24°C–25°C) with food/water ad libitum.

2.2. Viruses and Surgeries

AAV‐EF1α‐DIO‐TVA‐GFP (1.7 × 1013 genome copies/mL), AAV‐EF1α‐DIO‐RVG (6.8 × 1013 genome copies/mL), RV‐EnvA‐ΔG‐dsRed (5.0 × 1013 fluorescence‐forming units/mL) (Brain Case, Wuhan), and AAV‐EF1α‐DIO‐EYFP (1.17 × 1013 genome copies/mL, Genechem #19870) were used. Mice under isoflurane were unilaterally injected via glass pipette (25 μm) into DMD (AP −1.94 mm, ML ±0.3 mm, DV −4.9 mm) or DMH (AP −1.94 mm, ML ±0.3 mm, DV −5.25 mm). For retrograde tracing, AAV mixture (1:2, 100 nL, 10 nL/min) was injected; 2 weeks later, 100 nL RV‐EnvA‐ΔG‐dsRed was injected, perfused after 1 week. For anterograde tracing, 100 nL AAV‐EF1α‐DIO‐EYFP was injected, perfused after 4 weeks.

2.3. Histology and Immunostaining

Anesthetized mice (urethane 1.8 g/kg i.p.) were perfused with saline then 4% PFA. Brains were post‐fixed (24 h, 4°C), cryoprotected in 30% sucrose (24–36 h), and sectioned at 25 μm (CM1950, Leica). Sections were blocked (5% BSA, 0.25% Triton X‐100 in PBS, 30 min, RT), incubated with chicken anti‐GFP (1:1000, Abcam ab13970) overnight at 4°C, then with goat anti‐chicken Alexa Fluor 488 (1:500, Abcam ab150169) for 2 h at RT, and mounted with Vectashield.

2.4. RNAscope Fluorescent In Situ Hybridization

Animals were perfused and brains processed as above. Dehydrated tissues were embedded in O.C.T., snap‐frozen, and cryosectioned. RNAscope Multiplex Fluorescent v2 was performed on 16‐μm sections with antibody co‐detection. After H2O2 pretreatment and PBS washes, sections were incubated overnight at 4°C with primary antibody in Co‐Detection Antibody Diluent, then treated with Protease Plus. Probes for Slc17a6‐C2 (#319171‐C2) and Slc32a1‐C3 (#319191‐C3) were hybridized, followed by amplification. Sections were then incubated with goat polyclonal secondary antibody in Co‐Detection Diluent for 30 min at RT, washed, and mounted with Fluoromount‐G (SouthernBiotech).

2.5. Single‐Cell RNA Sequencing

Fifteen 8‐week‐old male C57BL/6J mice were anesthetized with urethane (1.8 g/kg, i.p.) and euthanized. Only males were used to minimize variability. Coronal brain slices (200 μm) containing the DMH were prepared using a vibrating microtome (VT1200S, Leica Biosystems, Germany). The DMH was bilaterally microdissected, and pooled tissues were stored at −80°C. Samples were processed using the 10X Genomics Chromium Single Cell 3′ v3.1 platform (OmicsMaster BIOTECHNOLOGY, Guangzhou, China), including: (1) single‐cell suspension preparation; (2) gel beads in emulsion generation and reverse transcription; (3) cDNA amplification and purification; (4) library construction; (5) sequencing on a NovaSeq X Plus platform (PE150).

Subsequent analyses were performed using Seurat (v.5.0.2) in R (v.4.2). Quality control excluded cells with < 500 or > 17,000 unique molecular identifiers (UMIs), > 10% mitochondrial reads, or doublets, retaining 5288 cells (median genes: 2533; median UMIs: 4486.5; median mito%: 0.32%). After normalization (LogNormalize, factor 10,000), the top 2000 variable genes were identified (vst) and scaled while regressing out UMI counts, mito%, and cell‐cycle, followed by principal component analysis (PCA). Using the top 30 principal components, Louvain clustering (resolution 0.1–1.0; resolution 0.5 selected) and UMAP were applied. Markers were defined by Wilcoxon test (log2FC > 0.25, adjusted p < 0.05).

2.6. Cluster Analysis

To examine projection pattern similarity, we performed hierarchical cluster analysis on input/output intensity matrices. Projection intensities were Z‐score normalized. A similarity matrix was constructed using Euclidean distances (smaller distance = higher similarity) and visualized as a heatmap. Hierarchical clustering used the complete linkage method, with results displayed as a dendrogram.

2.7. Imaging and Data Analysis

Images were captured using a confocal microscope (LSM 800, Carl Zeiss, Germany) and processed with ZEN software (Zeiss, Germany). Neural somata and axonal varicosities were quantified semi‐automatically using ImageJ (U.S. National Institutes of Health, Bethesda, MD, USA). Input ratio = afferent cells in a nucleus/total dsRed+ cells (including bilateral afferents). Axonal projection proportion = regional varicosity count/whole‐brain varicosities (reflects relative anatomical varicosity, not functional). Afferent and efferent preference ratios were calculated as the proportion of each input or output to DMD divided by that to DMH. Values > 1, < 1, and = 1 indicate bias toward DMD, DMH, or balanced distribution, respectively. This ratio is a descriptive measure to rank regional biases and is not a formal statistical comparison. Axons with transverse diameters exceeding 0.5 μm were classified as varicose. Data were analyzed using Prism 9.0 (GraphPad Software, San Diego, CA, USA) and presented as mean ± standard error of the mean (SEM). Whole‐brain DMD versus DMH comparisons were assessed by multiple unpaired t‐tests, with Benjamini–Krieger–Yekutieli false discovery rate (FDR) correction for multiple comparisons (Q = 5%). Differences with q < 0.05 were considered significant.

3. Results

3.1. Anatomical Distribution and Neurochemical Phenotype of LepRb Neurons in the DMD or DMH

To investigate LepRb neuron distribution, we injected AAV‐EF1α‐DIO‐EYFP into the DMD or DMH of LepRb‐Cre mice (Figure 1A), enabling specific labeling (Figure 1B). EYFP‐labeled neurons distributed widely along the DMH rostrocaudal axis, with highest density in the intermediate part; DMDLepRb neurons similarly concentrated in the intermediate region (Figure 1C). Although DMD is considered part of the DMH, its LepRb neurons displayed a more clustered and sharply demarcated pattern, suggesting DMD as a structurally distinct subregion.

FIGURE 1.

FIGURE 1

Open in a new tab

Distribution and neurochemical identity of DMDLepRb and DMHLepRb neurons. (A) Schematic diagram of the viral labeling strategy. (B) Schematic illustrating the rostrocaudal distribution of LepRb‐expressing neurons in the DMD and DMH. Neurons were mapped in coronal sections following immunohistochemical detection. (C) Quantitative rostrocaudal distribution of EYFP+ neurons in the DMD and DMH. (D) Violin plots showing the expression of Slc17a6 (Vglut2), Slc32a1 (Vgat) and Lepr across DMH neuronal clusters. (E) In situ detection of Slc32a1 and Slc17a6 in the DMDLepRb and DMHLepRb neurons. White arrows indicate Vgat+ LepRb+ neurons. Yellow arrows indicate Vglut2+ LepRb+ neurons. Scale bars, 20 μm. (F) Proportion of Vgat+ and Vglut2+ in all DMDLepRb and DMHLepRb neurons (n = 3). Data represent mean ± SEM.

TABLE 1.

List of brain region abbreviations.

Abbreviation Nucleus name Brain region
Acb Accumbens nucleus Telencephalon
AH Anterior hypothalamic area Diencephalon
AHi Amygdalohippocampal area Telencephalon
AI Agranular insular cortex Telencephalon
Arc Arcuate hypothalamic nucleus Diencephalon
Bar Barrington's nucleus Pons
BL Basolateral amygdaloid nucleus Telencephalon
BM Basomedial amygdaloid nucleus Telencephalon
BST Bed nucleus of the stria terminalis Telencephalon
CA1 Field CA1 of hippocampus Telencephalon
CeA Central amygdaloid nucleus Telencephalon
Cg1/2 Cingulate cortex, area 1/2 Telencephalon
DpG Deep gray layer of the superior colliculus Midbrain
DPGi Dorsal paragigantocellular nucleus Medulla
DpMe Deep mesencephalic nucleus Midbrain
DR Dorsal raphe nucleus Midbrain
Gi Gigantocellular reticular nucleus Medulla
Hb Habenular nucleus Diencephalon
HDB Nucleus of the horizontal limb of the diagonal band Telencephalon
IRt Intermediate reticular nucleus Medulla
KF Kolliker‐Fuse nucleus Pons
LA Lateroanterior hypothalamic nucleus Diencephalon
LC Locus coeruleus Pons
LDTg Laterodorsal tegmental nucleus Pons
LEnt Lateral entorhinal cortex Telencephalon
LH Lateral hypothalamic area Diencephalon
LPB Lateral parabrachial nucleus Pons
LRt Lateral reticular nucleus Medulla
LS Lateral septal nucleus Telencephalon
M1/2 Primary/secondary motor cortex Telencephalon
MeA Medial amygdaloid nucleus Telencephalon
MM Medial mammillary nucleus, medial part Diencephalon
MnR Median raphe nucleus Midbrain
MPB Medial parabrachial nucleus Pons
MS Medial septal nucleus Telencephalon
MTu Medial tuberal nucleus Diencephalon
NTS Nucleus of the solitary tract Medulla
PAG Periaqueductal gray Midbrain
Pe Periventricular hypothalamic nucleus Diencephalon
PH Posterior hypothalamic area Diencephalon
PMD Premammillary nucleus, dorsal part Diencephalon
PMV Premammillary nucleus, ventral part Diencephalon
Pn Pontine reticular nucleus Pons
POA Preoptic area Diencephalon
PPTg Pedunculopontine tegmental nucleus Pons
PSTh Parasubthalamic nucleus Diencephalon
PV Paraventricular thalamic nucleus Diencephalon
PVN Paraventricular hypothalamic nucleus Diencephalon
RLi Rostral linear nucleus of the raphe Midbrain
RMg Raphe magnus nucleus Pons
ROb Raphe obscurus nucleus Medulla
RPa Raphe pallidus nucleus Medulla
RRF Retrorubral field Midbrain
RS Retrosplenial cortex Telencephalon
RtTg Reticulotegmental nucleus of the pons Pons
SCh Suprachiasmatic nucleus Diencephalon
SHy Septohypothalamic nucleus Diencephalon
SN Substantia nigra Midbrain
SuM Supramammillary nucleus Diencephalon
VDB Nucleus of the vertical limb of the diagonal band Telencephalon
VMH Ventromedial hypothalamic nucleus Diencephalon
VOLT Vascular organ of the lamina terminalis Diencephalon
VP Ventral pallidum Telencephalon
VTA Ventral tegmental area Midbrain
VTM Ventral tuberomammillary nucleus Diencephalon
ZI Zona incerta Diencephalon
Open in a new tab

To characterize neurochemical composition, single‐cell RNA sequencing was performed on DMH tissue. Quality control metrics of the sequencing dataset were evaluated (Figure S1A). Based on neurotransmitter markers, neurons were classified into inhibitory (Vgat+/Slc32a1 +) and excitatory (Vglut2+/Slc17a6 +) populations (Figure S1B). Most DMHLepRb neurons expressed Slc32a1 (Figure 1D and Figure S1C), indicating a predominantly inhibitory phenotype. RNAscope showed that in DMDLepRb neurons, 82.1% co‐expressed Slc32a1, 14.7% co‐expressed Slc17a6, and 3.2% co‐expressed both. In DMHLepRb neurons, proportions were 79.9%, 17.2%, and 2.9%, respectively (Figure 1E,F). Thus, LepRb neurons in both regions are predominantly GABAergic with similar proportions, suggesting functional differences likely arise from distinct connectivity, not neuronal phenotype.

3.2. Rabies Virus Tracing Reveals Monosynaptic Inputs to DMDLepRb and DMHLepRb Neurons

To establish afferent connectivity, we performed cell‐type‐specific monosynaptic rabies tracing in LepRb‐Cre mice. AAV‐EF1α‐DIO‐TVA‐GFP and AAV‐EF1α‐DIO‐RVG were unilaterally injected into the DMD or DMH, followed two weeks later by EnvA‐pseudotyped glycoprotein‐deleted rabies virus (RV‐EnvA‐ΔG‐dsRed) at the same sites (Figure 2A). This distinguished GFP+ LepRb neurons (green), dsRed+ input neurons (red), and double‐labeled starter neurons (Figure 2B). Both GFP and dsRed signals were detected unilaterally (Figure 2C–E), with no expression in wild‐type mice (Figure 2C, right), confirming Cre‐dependent specificity. These methods enabled visualization and dissection of whole‐brain monosynaptic upstream inputs to DMDLepRb and DMHLepRb neurons.

FIGURE 2.

FIGURE 2

Open in a new tab

Monosynaptic inputs to DMDLepRb and DMHLepRb neurons. (A) Schematic of AAV helper viruses and RV injection timeline. (B) Coronal schematic of DMD and DMH viral infection: Red for RV labeling, green for AAV helper virus infection, and yellow for starter neurons. (C) Representative images of the DMD (left) and DMH (middle) show GFP‐ and dsRed‐expressing neurons in LepRb‐Cre mice, but not in wild‐type mice (right), following helper and RV injection. Scale bars, 50 μm. (D, E) Magnified views of DMD and DMH showing GFP‐expressing LepRb neurons (green), input neurons (red), and starter neurons (yellow, white arrowheads) in LepRb‐Cre mice. Scale bars, 10 μm. (F) Whole‐brain input counts of DMDLepRb and DMHLepRb neurons (n = 4; multiple unpaired t‐tests followed by Benjamini–Krieger–Yekutieli FDR correction for multiple comparisons). (G) Whole‐brain input proportions to DMDLepRb and DMHLepRb neurons (n = 4; multiple unpaired t‐tests followed by Benjamini–Krieger–Yekutieli FDR correction for multiple comparisons). Brain region abbreviations are detailed in Table 1. Data represent mean ± SEM. *q < 0.05, **q < 0.01.

3.3. Whole‐Brain Input Connectivity Features of DMDLepRb and DMHLepRb Neurons

To characterize whole‐brain input patterns of DMDLepRb and DMHLepRb neurons, we quantified dsRed‐labeled inputs on serial sections. Absolute numbers (Figure 2F) and proportion (Figure 2G) showed that both populations received bilateral monosynaptic inputs from 58 nuclei spanning the telencephalon, diencephalon, midbrain, pons, and medulla, with largely shared major sources. Dense input (> 5%) to DMDLepRb neurons came from the lateral hypothalamic area (LH, 13.91%), POA (10.43%), Arc (6.43%), and periaqueductal gray (PAG, 5.13%); for DMHLepRb neurons, from POA (12.57%), LH (10.37%), and Arc (5.58%). Moderate inputs (2%–5%) to DMDLepRb neurons arose from the bed nucleus of the stria terminalis (BST, 4.56%), posterior hypothalamic area (PH, 4.52%), lateral septal nucleus (LS, 4.36%), zona incerta (ZI, 3.91%), VMH (3.59%), anterior hypothalamic area (AH, 3.01%), deep mesencephalic nucleus (DpMe, 2.86%), lateral parabrachial nucleus (LPB, 2.85%), and paraventricular hypothalamic nucleus (PVN, 2.51%); and to DMHLepRb neurons from PAG (4.72%), LS (4.60%), BST (4.55%), VMH (4.32%), AH (4.15%), PH (4.18%), ZI (4.06%), LPB (2.76%), and PVN (2.16%). Sparse inputs (< 2%) came from 45 (DMD) and 46 (DMH) other regions (Figure 2G).

Despite shared input framework, further analysis revealed notable biases in the weights of specific connections. Although DMDLepRb and DMHLepRb neurons share largely overlapping input sources, we examined whether DMD differentially weights these shared afferents. We thus calculated a preference ratio for each input nucleus (DMD versus DMH). A ratio > 1 indicates that a given upstream nucleus provides a higher proportion of inputs to DMDLepRb neurons relative to the overall DMHLepRb population, suggesting that this pathway is more likely dominated by DMD. Conversely, a ratio < 1 indicates that the upstream nucleus provides a lower proportion of inputs to DMD relative to the overall DMH value, implying that these nuclei more likely project predominantly to the remaining DMH subdivisions. We calculated the input ratio between DMDLepRb and DMHLepRb neurons and categorized these regions into three types: DMD‐biased, balanced, and DMH‐biased. This approach identified input nodes specialized for DMD within the shared global afferent network (Figure 3A). A connectivity map shows representative DMD‐biased and DMH‐biased regions (Figure 3B). The results showed that DMDLepRb neurons received stronger inputs from the basolateral amygdala (BL), retrosplenial cortex (RS), medial mammillary nucleus (MM), LH, DpMe, and Kölliker‐Fuse nucleus (KF); conversely, DMHLepRb neurons received stronger inputs from the ventral pallidum (VP), dorsal premammillary nucleus (PMD), agranular insular cortex (AI), lateral reticular nucleus (LRt), lateroanterior hypothalamic nucleus (LA), and intermediate reticular nucleus (IRt). DMD‐biased inputs mainly involved emotional/arousal/defense regions, whereas DMH‐biased inputs enriched in visceral sensation, reward, and innate behaviors (Figure 3C,D). Together, DMDLepRb and DMHLepRb neurons share a conserved global input architecture, while differential connection weights of specific upstream nuclei form the structural basis for subregional connectional specialization.

FIGURE 3.

FIGURE 3

Open in a new tab

Preference analysis of afferent nuclei to DMDLepRb and DMHLepRb neurons. (A) Preference ratio of afferent inputs to DMD over DMH. The preference ratio was calculated as the proportion of each afferent input to DMD divided by that to DMH. Values > 1 indicate a bias toward DMD, whereas values < 1 indicate a bias toward DMH. These ratios are intended only to describe relative connectivity patterns and are not used to assign statistically significant bias categories. (B) Sankey diagram showing biased afferent inputs to DMD versus DMH. The width of each curve reflects the preference bias, while the color intensity corresponds to the input proportion. (C) Immunofluorescence of DMDLepRb‐biased afferent nuclei. Scale bars, 100 μm. (D) Immuno‐fluorescence of DMHLepRb‐biased afferent nuclei. Scale bars, 100 μm. Brain region abbreviations are detailed in Table 1.

3.4. Axonal Projection Patterns of DMDLepRb and DMHLepRb Neurons

To map downstream projections, AAV‐EF1α‐DIO‐EYFP was unilaterally injected into DMD or DMH (Figure 4A,C). Four weeks after injection, strong EYFP signals were detected in the somata (Figure 4B,D). Quantification focused on axonal boutons, excluding non‐bouton segments. We then quantified absolute bouton numbers (Figure 4E) and proportions (Figure 4F) per region. Both populations sent bilateral projections to 43 brain regions across the telencephalon, diencephalon, midbrain, pons, and medulla, with axons predominantly ipsilateral. Dense projections (> 5%) from DMDLepRb neurons targeted Arc (27.28%), POA (13.62%), LH (8.05%), and PAG (5.33%). For DMHLepRb neurons, dense targets included Arc (26.01%), POA (11.49%), LH (10.51%), PAG (9.68%), PVN (5.05%), and VMH (5.01%). Moderate projections (2%–5%) from DMDLepRb neurons were observed in eight regions: PVN, AH, VMH, LA, paraventricular thalamic nucleus (PV), PH, periventricular hypothalamic nucleus (Pe), and BST. DMHLepRb neurons targeted seven regions: premammillary nucleus, ventral part (PMV), PH, Pe, AH, PV, suprachiasmatic nucleus (SCh), and ventral tegmental area (VTA). Sparse projections (< 2%) were found in 31 and 30 regions, respectively (Figure 4F). These findings indicate that DMDLepRb and DMHLepRb neurons share major downstream targets within a conserved output framework.

FIGURE 4.

FIGURE 4

Open in a new tab

Efferent projections of DMDLepRb and DMHLepRb neurons. (A, C) Schematic diagram of anterograde tracing virus strategy. (B, D) Representative images of EYFP expression in the DMD and DMH. Scale bars, 100 μm. (E) Whole‐brain output counts of DMDLepRb and DMHLepRb neurons (n = 4; multiple unpaired t‐tests followed by Benjamini–Krieger–Yekutieli FDR correction for multiple comparisons). (F) Whole‐brain output proportions from DMDLepRb and DMHLepRb neurons (n = 4; multiple unpaired t‐tests followed by Benjamini–Krieger–Yekutieli FDR correction for multiple comparisons). Brain region abbreviations are detailed in Table 1. Data represent mean ± SEM. *q < 0.05, **q < 0.01, ***q < 0.001.

To further quantify projection preference, we applied the same ratio to identify DMD‐specialized output targets (Figure 5A,B). Despite sharing similar overall targets, this analysis revealed that DMD preferentially projected to the basomedial amygdaloid nucleus (BM), horizontal limb of the diagonal band (HDB), LA, DpMe, raphe pallidus nucleus (RPa), and IRt—regions involved in sleep–wake regulation, thermoregulation, and autonomic control. In contrast, DMH preferentially projected to the SCh, PAG, pontine reticular nucleus (Pn), dorsal raphe nucleus (DR), median raphe nucleus (MnR), and nucleus of the solitary tract (NTS)—regions associated with circadian rhythms, instinctive behaviors, and visceral sensory integration. Representative biased regions are shown in Figure 5C (DMD) and Figure 5D (DMH). These findings indicate that DMDLepRb and DMHLepRb neurons share a conserved global output framework, and subregional projection biases are implemented through differential weighting of shared downstream targets.

FIGURE 5.

FIGURE 5

Open in a new tab

Preference analysis of efferent projections from DMDLepRb and DMHLepRb neurons. (A) Preference ratio of outputs from DMD over DMH. The preference ratio was calculated as the proportion of outputs from DMD divided by those from DMH. Values > 1 indicate a bias toward DMD output, whereas values < 1 indicate a bias toward DMH output. These ratios are intended only to describe relative connectivity patterns and are not used to assign statistically significant bias categories. (B) Sankey diagram showing biased efferent projections from DMD versus DMH. The width of each curve reflects the projection bias, while the color intensity corresponds to the output proportion. (C) Immunofluorescence of DMDLepRb‐biased efferent nuclei. Scale bars, 50 μm. (D) Immunofluorescence of DMHLepRb‐biased efferent nuclei. Scale bars, 50 μm. Brain region abbreviations are detailed in Table 1.

3.5. Comparison of Input–Output Distribution of DMDLepRb and DMHLepRb Neurons

To determine whether connectivity differences reflect distinct networks, we integrated whole‐brain input and output data for DMDLepRb and DMHLepRb neurons (Figure 6A,B). Both populations formed predominant bidirectional connections with the diencephalon, acting as hypothalamic hubs. However, regional weighting differed between the two populations in their outputs: DMDLepRb neurons preferentially targeted the telencephalon, whereas DMHLepRb neurons preferentially targeted the midbrain (Figure 6B). In contrast, no significant differences were observed in their input patterns across major brain divisions (Figure 6A). Correlation analyses of major input–output nuclei (Figure 6C) showed strong positive correlations (r ≥ 0.95) between DMDLepRb and DMHLepRb, with highly consistent key data. Thus, both populations share a highly homologous global connectivity architecture, operating on a shared core framework rather than as independent parallel circuits.

FIGURE 6.

FIGURE 6

Open in a new tab

Comparative analysis of whole‐brain input and output of DMDLepRb and DMHLepRb neurons. (A) Comparison of input distributions to DMDLepRb and DMHLepRb neurons across five major brain subdivisions (n = 4, multiple unpaired t‐tests followed by Benjamini–Krieger–Yekutieli FDR correction for multiple comparisons). (B) Comparison of output distributions from DMDLepRb and DMHLepRb neurons across five major brain subdivisions (n = 4, multiple unpaired t‐tests followed by Benjamini–Krieger–Yekutieli FDR correction for multiple comparisons). (C) Comparison of inputs (left) and outputs (right) between the DMDLepRb and DMHLepRb neurons. Nuclei with a ratio greater than 1 were selected for analysis. (D) Input and output proportions of nuclei that are bidirectionally connected with both DMDLepRb (left) and DMHLepRb (right) neurons. (E, F) Hierarchical clustering analysis (Euclidean distance) of all inputs to DMDLepRb and DMHLepRb neurons. (G, H) Hierarchical clustering analysis (Euclidean distance) of all output nuclei from DMDLepRb and DMHLepRb neurons. The color scales in E–H indicate Euclidean distance, where darker colors represent smaller distances (higher similarity) and lighter colors represent larger distances (lower similarity). Brain region abbreviations are detailed in Table 1. Data represent mean ± SEM. *q < 0.05, **q < 0.01.

Given this global homology, we examined specific‐nucleus connection strengths via input–output bias scatter plots (Figure 6D). Some regions showed similar input–output ratios across groups: Arc consistently above the diagonal (strongest common output target), while BST, PH, and LS remained below (predominant input‐receiving). Others displayed distinct distributions: LH showed input bias in DMDLepRb, balanced in DMHLepRb; POA and AH exhibited the opposite pattern (output bias in DMDLepRb versus input bias in DMHLepRb); and PAG was balanced in DMDLepRb but output‐biased in DMHLepRb. These varied directions reveal input–output polarization across the two populations, indicating that different nuclei exhibit distinct connection biases.

To further assess spatial heterogeneity, we performed hierarchical clustering on input (Figure 6E,F) and output nuclei (Figure 6G,H) of DMDLepRb and DMHLepRb neurons. On the input side, PVN, LPB, DpMe, and AH formed tight, dark clusters in DMDLepRb but dispersed into distinct, lighter branches in DMHLepRb, indicating reduced input homogeneity in DMHLepRb; however, LH and POA remained in consistent clusters. On the output side, DR, medial amygdaloid nucleus (MeA), supramammillary nucleus (SuM), and LS exhibited high similarity in DMDLepRb neurons but scattered in DMHLepRb neurons, whereas LH and PAG shared common clusters. These results confirm that DMDLepRb neurons possess a more compact and coherent intrinsic network organization than DMHLepRb neurons. Collectively, these data demonstrate that DMDLepRb neurons belong to the unified DMH network system rather than forming an isolated independent circuit, and that local connection‐weight differences underlie the connectional divergence between subdivisions.

4. Discussion

4.1. Comparison With Previous Tracing Studies

Classical anatomical studies using conventional tracers, including cholera toxin subunit B (CTb) [28], biotinylated dextran amine (BDA) [29], Fluoro‐Gold (FG), and Phaseolus Vulgaris leucoagglutinin (PHA‐L) [30, 31], revealed DMH connectivity with hypothalamic, midbrain, and brainstem nuclei, but lacked cell‐type specificity and were susceptible to passing fibers and polysynaptic contamination. Cre‐loxP viral tracing enabled cell‐type‐specific analyses. Zhang et al. demonstrated that DMHLepRb neurons receive POA inputs and project to POA, BST, Arc, PAG, and RPa [32], however, a cell‐type‐specific connectomic comparison across DMH subregions remains lacking. Here, we combined monosynaptic rabies retrograde tracing, Cre‐dependent anterograde tracing, and whole‐brain quantitative analysis to compare the input–output organization of DMDLepRb neurons with that of the DMHLepRb population. This approach provides cell‐type selectivity, monosynaptic specificity, and a brain‐wide assessment. Rather than defining DMD and DMH as separate systems, our findings reveal that connectional diversification within the DMH emerges through locally biased connectivity embedded within a conserved homeostatic network architecture.

4.2. DMHLepRb Neurons Form a Broad Brain‐Wide Homeostatic Integration Network

Our results demonstrate that DMHLepRb neurons establish extensive connections across the telencephalon, diencephalon, midbrain, pons, and medulla, with the diencephalon as the major bidirectional hub, aligning with classical DMH connectivity studies [30, 31]. At the input level, DMHLepRb neurons receive prominent monosynaptic inputs from POA, LH, Arc, PAG, BST, and VMH—regions involved in thermoregulation, energy balance, feeding, arousal, defensive behaviors, and autonomic control [33, 34, 35, 36, 37, 38]. At the output level, DMHLepRb neurons project extensively to Arc, POA, LH, PAG, and PVN, with the PVN serving as a major relay for hypothalamic–pituitary–adrenal axis activation and autonomic output [39]. This extensive connectivity with both the limbic system and brainstem provides a structural substrate for translating emotional states into visceral responses. Previous studies show that DMHLepRb neurons regulate feeding via GABAergic projections to the Arc [24]; coordinate metabolism through distinct subpopulations projecting to the Arc and RPa [15]; and modulate sleep‐related respiration, REM sleep [18], and circadian feeding rhythms [13]. Thus, DMHLepRb neurons constitute a multimodal integration network linking metabolic status with autonomic, respiratory, sleep, and behavioral states.

4.3. Conserved Global Architecture With Local Specialization: An Organizational Principle of the DMH

As the DMD is an anatomical subregion of the DMH, overlapping upstream and downstream nuclei are expected, confirming that DMD circuits are embedded in the shared DMH‐wide connective framework rather than forming independent parallel pathways. Despite this conserved global architecture, quantitative analyses revealed marked connection‐weight biases across whole‐brain input–output networks. DMDLepRb neurons preferentially connect with nuclei regulating arousal, defense, and sympathetic thermogenesis, whereas DMHLepRb neurons show stronger weighting in metabolic, neuroendocrine, and circadian regions. Within the DMH, which comprises the DMD, dorsomedial hypothalamus compact part (DMC), and DMV subdivisions, this divergence indicates that stress defense and sympathetic thermogenesis are preferentially mediated by the DMD, while feeding and neuroendocrine control are likely dominated by the DMC and DMV subdivisions.

This pattern of a shared network backbone with subregion‐specific weight biases mirrors organizational principles described in the VMH [26], PAG [40], and limbic circuits [27] where functional diversity arises from redistributed information flow within a common anatomical framework. Extending this principle to the DMH, our data demonstrate that hypothalamic subnuclear specialization does not require fully segregated circuits: conserved architecture ensures coordinated homeostatic regulation, while differential weighting enables subregions to govern distinct physiological processes. We therefore propose a hierarchical DMH organization: a shared connectivity framework supports global integrative function, and local connection‐weight biases drive subregional specialization. Our data validate this principle using the functionally well‐defined DMD as a test case.

4.4. Input–Output Connectivity Biases and Complementary Functional Organization of DMD and Remaining DMH Subdivisions

As the whole‐DMH dataset covers all three subdivisions (DMD, DMC, and DMV), weight differences between DMDLepRb and DMHLepRb neurons indirectly reflect connectivity preferences of the remaining subdivisions (predominantly DMC and DMV). At the input level, the DMD receives enriched afferents from emotion/arousal/defense regions (BL, RS, MM, LH, DpMe, KF) [41, 42, 43, 44], while remaining subdivisions show stronger inputs linked to instinctive behavior, reward, and visceral sensation (VP, PMD, IRt, AI, LA, LRt) [45, 46, 47]. Efferent projections follow a parallel complementary pattern: the DMD preferentially targets sleep–wake, thermoregulatory, and autonomic nuclei (BM, HDB, LA, DpMe, RPa, IRt) [48, 49, 50, 51], whereas remaining subdivisions weight more heavily toward circadian regulation, neuroendocrine control, and visceral sensory integration regions (SCh, PAG, Pn, DR, MnR, NTS) [52, 53, 54].

Functionally, this profile defines the DMD as a specialized gateway for higher‐order stress signals, integrating arousal/defense inputs to autonomic effectors without disrupting DMH network coherence. This arrangement offers a network‐topological mechanism for hypothalamic subnuclear diversity: functional partitioning arises from differential weighting of shared circuits, independent of full anatomical segregation. Notably, these connectivity features of non‐DMD subdivisions are inferred from group comparisons rather than direct anatomical tracing. Future studies with subregion‐specific labeling of DMV and DMC neurons are required for validation.

4.5. Limitations and Future Directions

Several limitations should be acknowledged. First, rabies tracing reveals anatomical connectivity, yet does not confirm functional transmission or physiological significance. Moreover, we did not directly test the behavioral, autonomic, or metabolic roles of DMDLepRb neurons. Future studies combining projection‐specific manipulations (e.g., optogenetics, electrophysiology) with behavioral and physiological assays are needed to establish the precise functional relevance of these circuits. Second, while we quantified axonal varicosities as candidate presynaptic sites, varicosities do not always equate to functional synapses, necessitating future optogenetic‐electrophysiological and presynaptic marker validation. Third, although both sexes were included, systematic quantitative comparisons between sexes were not performed; whether these circuits exhibit sex‐dependent differences requires future investigation. Fourth, the DMH contains three subnuclei, but our comparison was between DMDLepRb and whole‐DMHLepRb populations. Thus, connectivity features of DMVLepRb and DMCLepRb neurons may be diluted in the whole‐DMH dataset. Subnucleus‐specific tracing is needed to fully resolve DMH heterogeneity. Finally, unilateral viral injections primarily reveal ipsilateral projections; given that many DMH‐regulated functions involve bilateral coordination, contralateral pathways may be underestimated and warrant bilateral tracing.

5. Conclusions

This study systematically maps whole‐brain input–output organization of DMDLepRb and DMHLepRb neurons, showing that the DMH achieves functional diversification through local connectivity biases within a conserved global architecture. Our findings provide an anatomical framework for DMH functional heterogeneity and circuit mechanisms coordinating metabolic, autonomic, and behavioral states.

Author Contributions

Study design and manuscript drafting/revising: Fang Yuan and Mengchu Zhu. Running of the experiments, acquisition of the data, and figure drawing: Yaxin Hao, Xiang Zhang, and Yongqiang Chen. Acquisition of the data and statistical analysis: Danyang Zhao, Haixing Zhong, Heming Li, and Minghui Ren. All authors have read and approved the final version of the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China, 32671478 and 82170104. Hebei Province Yanzhao Golden Platform Project, HJZD202510. Natural Science Foundation of Hebei Province, H2025206880. Graduate Innovation Funding Program of Hebei Province, CXZZBS2024114.

Ethics Statement

All experimental procedures were approved by the Animal Care and Ethics Committee of Hebei Medical University (Approval No. Hebmu‐P2023052) and were conducted in accordance with the Guide for the Care and Use of Laboratory Animals.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Quality control and cell clustering characterization of single‐cell transcriptomic profiling in the DMH nucleus. (A) Violin plots illustrating the distribution of three quality control parameters for individual cells isolated from the DMH: number of detected genes per cell (left), UMI counts per cell (middle), and mitochondrial gene transcript percentage (right). (B) UMAP‐based dimensionality reduction plot of single cells. Each dot represents an individual cell, and distinct colors denote different cell subclusters. (C) Feature expression plots of representative marker genes projected onto the UMAP space. From left to right: Slc17a6 (excitatory neuronal marker), Slc32a1 (inhibitory neuronal marker), and Lepr. Color intensity from light gray to red represents gene expression levels from low to high (red, high expression).

CNS-32-e71185-s001.tif (1.2MB, tif)

Acknowledgments

We thank the support from the Core Facilities and Centers, Institute of Medicine and Health, Hebei Medical University for technical support. This research was funded by National Natural Science Foundation of China (32671478 to F.Y.; 82170104 to F.Y.), the Hebei Province Yanzhao Golden Platform Project (HJZD202510 to F.Y.), the Natural Science Foundation of Hebei Province (H2025206880 to F.Y.), and Graduate Innovation Funding Program of Hebei Province (CXZZBS2024114 to X.Z.).

Contributor Information

Mengchu Zhu, Email: zhumengchu@hebmu.edu.cn.

Fang Yuan, Email: yuanfang@hebmu.edu.cn.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

  • 1. Maejima Y., Yokota S., Shimizu M., et al., “The Deletion of Glucagon‐Like Peptide‐1 Receptors Expressing Neurons in the Dorsomedial Hypothalamic Nucleus Disrupts the Diurnal Feeding Pattern and Induces Hyperphagia and Obesity,” Nutrition & Metabolism 18, no. 1 (2021): 58. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. DiMicco J. A., Samuels B. C., Zaretskaia M. V., and Zaretsky D. V., “The Dorsomedial Hypothalamus and the Response to Stress: Part Renaissance, Part Revolution,” Pharmacology Biochemistry and Behavior 71, no. 3 (2002): 469–480. [DOI] [PubMed] [Google Scholar]
  • 3. Chaar L. J., Coelho A., Silva N. M., Festuccia W. L., and Antunes V. R., “High‐Fat Diet‐Induced Hypertension and Autonomic Imbalance Are Associated With an Upregulation of Cart in the Dorsomedial Hypothalamus of Mice,” Physiological Reports 4, no. 11 (2016): e12811. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Tsuji S., Brace C. S., Yao R., et al., “Sleep‐Wake Patterns Are Altered With Age, Prdm13 Signaling in the DMH, and Diet Restriction in Mice,” Life Science Alliance 6, no. 6 (2023): e202301992. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Lee S., Bookout A. L., Lee C. E., et al., “Laser‐Capture Microdissection and Transcriptional Profiling of the Dorsomedial Nucleus of the Hypothalamus,” Journal of Comparative Neurology 520, no. 16 (2012): 3617–3632. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Piñol R. A., Zahler S. H., Li C., et al., “Brs3 Neurons in the Mouse Dorsomedial Hypothalamus Regulate Body Temperature, Energy Expenditure, and Heart Rate, but Not Food Intake,” Nature Neuroscience 21, no. 11 (2018): 1530–1540. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Jeong J. H., Lee D. K., and Jo Y. H., “Cholinergic Neurons in the Dorsomedial Hypothalamus Regulate Food Intake,” Molecular Metabolism 6, no. 3 (2017): 306–312. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Su Z., Hou Y., Wang B., et al., “The DMH(GABA) Neurons Play a Crucial Role in the Regulation of Cold‐Induced White Adipose Browning Through DMH to LPO Projection,” Advanced Science 13, no. 5 (2025): e08513. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Carrillo‐Franco L., González‐García M., Morales‐Luque C., Dawid‐Milner M. S., and López‐González M. V., “Hypothalamic Regulation of Cardiorespiratory Functions: Insights Into the Dorsomedial and Perifornical Pathways,” Biology 13, no. 11 (2024): 933. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Liu Q., Bell B. J., Kim D. W., et al., “A Clock‐Dependent Brake for Rhythmic Arousal in the Dorsomedial Hypothalamus,” Nature Communications 14, no. 1 (2023): 6381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Tanaka M. and McAllen R. M., “Functional Topography of the Dorsomedial Hypothalamus,” American Journal of Physiology. Regulatory, Integrative and Comparative Physiology 294, no. 2 (2008): R477–R486. [DOI] [PubMed] [Google Scholar]
  • 12. Kataoka N., Hioki H., Kaneko T., and Nakamura K., “Psychological Stress Activates a Dorsomedial Hypothalamus‐Medullary Raphe Circuit Driving Brown Adipose Tissue Thermogenesis and Hyperthermia,” Cell Metabolism 20, no. 2 (2014): 346–358. [DOI] [PubMed] [Google Scholar]
  • 13. Faber C. L., Deem J. D., Phan B. A., et al., “Leptin Receptor Neurons in the Dorsomedial Hypothalamus Regulate Diurnal Patterns of Feeding, Locomotion, and Metabolism,” eLife 10 (2021): e63671. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Tang Q., Godschall E., Brennan C. D., et al., “Leptin Receptor Neurons in the Dorsomedial Hypothalamus Input to the Circadian Feeding Network,” Science Advances 9, no. 34 (2023): eadh9570. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Francois M., Kaiser L., He Y., et al., “Leptin Receptor Neurons in the Dorsomedial Hypothalamus Require Distinct Neuronal Subsets for Thermogenesis and Weight Loss,” Metabolism 163 (2025): 156100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Amorim M. R., Wang X., Aung O., et al., “Leptin Signaling in the Dorsomedial Hypothalamus Couples Breathing and Metabolism in Obesity,” Cell Reports 42, no. 12 (2023): 113512. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Simonds S. E., Pryor J. T., Ravussin E., et al., “Leptin Mediates the Increase in Blood Pressure Associated With Obesity,” Cell 159, no. 6 (2014): 1404–1416. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Pho H., Berger S., Freire C., et al., “Leptin Receptor Expression in the Dorsomedial Hypothalamus Stimulates Breathing During NREM Sleep in Db/Db Mice,” Sleep 44, no. 6 (2021): zsab046. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Wu J., Liu D., Li J., et al., “Central Neural Circuits Orchestrating Thermogenesis, Sleep‐Wakefulness States and General Anesthesia States,” Current Neuropharmacology 20, no. 1 (2022): 223–253. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Rezai‐Zadeh K. and Münzberg H., “Integration of Sensory Information via Central Thermoregulatory Leptin Targets,” Physiology & Behavior 121 (2013): 49–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Yang W. Z., Xie H., Du X., et al., “A Parabrachial‐Hypothalamic Parallel Circuit Governs Cold Defense in Mice,” Nature Communications 14, no. 1 (2023): 4924. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Wu Y. E., Luca R. D., Broadhurst R. Y., et al., “Suprachiasmatic Neuromedin‐S Neurons Regulate Arousal,” bioRxiv [Preprint] (2025): 2025.02.22.639648, 10.1101/2025.02.22.639648. [DOI] [Google Scholar]
  • 23. Ahmadlou M., Giannouli M., van Vierbergen J. F. M., et al., “Cell‐Type‐Specific Hypothalamic Pathways to Brainstem Drive Context‐Dependent Strategies in Response to Stressors,” Current Biology 34, no. 11 (2024): 2448–2459.e2444. [DOI] [PubMed] [Google Scholar]
  • 24. Garfield A. S., Shah B. P., Burgess C. R., et al., “Dynamic GABAergic Afferent Modulation of AgRP Neurons,” Nature Neuroscience 19, no. 12 (2016): 1628–1635. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Rezai‐Zadeh K., Yu S., Jiang Y., et al., “Leptin Receptor Neurons in the Dorsomedial Hypothalamus Are Key Regulators of Energy Expenditure and Body Weight, but Not Food Intake,” Molecular Metabolism 3, no. 7 (2014): 681–693. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Gutierrez‐Castellanos N., Dias I. C., Husain B. F. A., and Lima S., “Functional Diversity Along the Anteroposterior Axis of the Ventromedial Hypothalamus,” Journal of Neuroendocrinology 37, no. 7 (2025): e13447. [DOI] [PubMed] [Google Scholar]
  • 27. Morgane P. J., Galler J. R., and Mokler D. J., “A Review of Systems and Networks of the Limbic Forebrain/Limbic Midbrain,” Progress in Neurobiology 75, no. 2 (2005): 143–160. [DOI] [PubMed] [Google Scholar]
  • 28. Chou T. C., Scammell T. E., Gooley J. J., Gaus S. E., Saper C. B., and Lu J., “Critical Role of Dorsomedial Hypothalamic Nucleus in a Wide Range of Behavioral Circadian Rhythms,” Journal of Neuroscience 23, no. 33 (2003): 10691–10702. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Papp R. S. and Palkovits M., “Brainstem Projections of Neurons Located in Various Subdivisions of the Dorsolateral Hypothalamic Area‐An Anterograde Tract‐Tracing Study,” Frontiers in Neuroanatomy 8 (2014): 34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Thompson R. H., Canteras N. S., and Swanson L. W., “Organization of Projections From the Dorsomedial Nucleus of the Hypothalamus: A Pha‐L Study in the Rat,” Journal of Comparative Neurology 376, no. 1 (1996): 143–173. [DOI] [PubMed] [Google Scholar]
  • 31. Thompson R. H. and Swanson L. W., “Organization of Inputs to the Dorsomedial Nucleus of the Hypothalamus: A Reexamination With Fluorogold and Phal in the Rat,” Brain Research Reviews 27, no. 2 (1998): 89–118. [DOI] [PubMed] [Google Scholar]
  • 32. Zhang Y., Kerman I. A., Laque A., et al., “Leptin‐Receptor‐Expressing Neurons in the Dorsomedial Hypothalamus and Median Preoptic Area Regulate Sympathetic Brown Adipose Tissue Circuits,” Journal of Neuroscience 31, no. 5 (2011): 1873–1884. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Nakamura K., “Central Mechanisms of Thermoregulation and Fever in Mammals,” in Advances in Experimental Medicine and Biology, vol. 1461 (Springer Nature Singapore, 2024), 141–159. [DOI] [PubMed] [Google Scholar]
  • 34. Vohra M. S., Benchoula K., Serpell C. J., and Hwa W. E., “AgRP/NPY and POMC Neurons in the Arcuate Nucleus and Their Potential Role in Treatment of Obesity,” European Journal of Pharmacology 915 (2022): 174611. [DOI] [PubMed] [Google Scholar]
  • 35. Arrigoni E., Chee M. J. S., and Fuller P. M., “To Eat or to Sleep: That Is a Lateral Hypothalamic Question,” Neuropharmacology 154 (2019): 34–39. [DOI] [PubMed] [Google Scholar]
  • 36. Barcelon E., Noh K., Hwang M., et al., “Midbrain PAG Astrocytes Modulate Mouse Defensive and Panic‐Like Behaviors,” Advanced Science 13, no. 19 (2026): e06062. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. restani C. C. C., Alves F. H., Gomes F. V., Resstel L. B., Correa F. M., and Herman J. P., “Mechanisms in the Bed Nucleus of the Stria Terminalis Involved in Control of Autonomic and Neuroendocrine Functions: A Review,” Current Neuropharmacology 11, no. 2 (2013): 141–159. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. de Jesus A. A., Dos‐Santos R. C., Rodrigues‐Santos I., et al., “PI3K in the VMH Attenuates Diet‐Induced Obesity and Participates in the Effects of E2 on Energy Expenditure in Mice,” Journal of the Endocrine Society 9, no. 7 (2025): bvaf080. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Elsaafien K., Kirchner M. K., Mohammed M., et al., “Identification of Novel Cross‐Talk Between the Neuroendocrine and Autonomic Stress Axes Controlling Blood Pressure,” Journal of Neuroscience 41, no. 21 (2021): 4641–4657. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Zhang H., Zhu Z., Ma W. X., et al., “The Contribution of Periaqueductal Gray in the Regulation of Physiological and Pathological Behaviors,” Frontiers in Neuroscience 18 (2024): 1380171. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Asim M., Wang H., Waris A., and He J., “Basolateral Amygdala Parvalbumin and Cholecystokinin‐Expressing GABAergic Neurons Modulate Depressive and Anxiety‐Like Behaviors,” Translational Psychiatry 14, no. 1 (2024): 418. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Liu H., Shi Y., Zhang Q., et al., “Two Distinct Cell Types of the Medial Mammillary Body Forming Segregated Subcircuits,” Molecular Psychiatry 30, no. 10 (2025): 4845–4858. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43. Naganuma F., Kroeger D., Bandaru S. S., Absi G., Madara J. C., and Vetrivelan R., “Lateral Hypothalamic Neurotensin Neurons Promote Arousal and Hyperthermia,” PLoS Biology 17, no. 3 (2019): e3000172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. Chen L., Cai P., Wang R. F., et al., “Glutamatergic Lateral Hypothalamus Promotes Defensive Behaviors,” Neuropharmacology 178 (2020): 108239. [DOI] [PubMed] [Google Scholar]
  • 45. Smith K. S., Tindell A. J., Aldridge J. W., and Berridge K. C., “Ventral Pallidum Roles in Reward and Motivation,” Behavioural Brain Research 196, no. 2 (2009): 155–167. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46. Cezario A. F., Ribeiro‐Barbosa E. R., Baldo M. V. C., and Canteras N. S., “Hypothalamic Sites Responding to Predator Threats—The Role of the Dorsal Premammillary Nucleus in Unconditioned and Conditioned Antipredatory Defensive Behavior,” European Journal of Neuroscience 28, no. 5 (2008): 1003–1015. [DOI] [PubMed] [Google Scholar]
  • 47. Immordino‐Yang M. H., Yang X. F., and Damasio H., “Correlations Between Social‐Emotional Feelings and Anterior Insula Activity Are Independent From Visceral States but Influenced by Culture,” Frontiers in Human Neuroscience 8 (2014): 728. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48. Liang C. L., Nguyen T. Q., and Marks G. A., “Inhibitory and Excitatory Amino Acid Neurotransmitters Are Utilized by the Projection From the Dorsal Deep Mesencephalic Nucleus to the Sublaterodorsal Nucleus REM Sleep Induction Zone,” Brain Research 1567 (2014): 1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49. Hitrec T., Vecchio F. D., Alberti L., et al., “Activation of Orexin‐A (Hypocretin‐1) Receptors in the Raphe Pallidus at Different Ambient Temperatures in the Rat: Effects on Thermoregulation, Cardiovascular Control, Sleep, and Feeding Behavior,” Frontiers in Neuroscience 18 (2024): 1458437. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50. Ajayi I. E., McGovern A. E., Driessen A. K., Kerr N. F., Mills P. C., and Mazzone S. B., “Hippocampal Modulation of Cardiorespiratory Function,” Respiratory Physiology & Neurobiology 252‐253 (2018): 18–27. [DOI] [PubMed] [Google Scholar]
  • 51. Chen L., Chen C., Jin Q., et al., “Efferent Pathways From the Suprachiasmatic Nucleus to the Horizontal Limbs of Diagonal Band Promote NREM Sleep During the Dark Phase in Mice,” BMC Neuroscience 25, no. 1 (2024): 34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52. Liu Y., Huo R., and Zhang E. E., “Evolving Perspectives on the Molecular and Neural Foundations of Mammalian Circadian Rhythms,” Trends in Neurosciences 48, no. 11 (2025): 904–918. [DOI] [PubMed] [Google Scholar]
  • 53. Reis F. M. C. V., Mobbs D., Canteras N. S., and Adhikari A., “Orchestration of Innate and Conditioned Defensive Actions by the Periaqueductal Gray,” Neuropharmacology 228 (2023): 109458. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54. Huang T. X., Wang S., and Ran C., “Interoceptive Processing in the Nucleus of the Solitary Tract,” Current Opinion in Neurobiology 93 (2025): 103021. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S1: Quality control and cell clustering characterization of single‐cell transcriptomic profiling in the DMH nucleus. (A) Violin plots illustrating the distribution of three quality control parameters for individual cells isolated from the DMH: number of detected genes per cell (left), UMI counts per cell (middle), and mitochondrial gene transcript percentage (right). (B) UMAP‐based dimensionality reduction plot of single cells. Each dot represents an individual cell, and distinct colors denote different cell subclusters. (C) Feature expression plots of representative marker genes projected onto the UMAP space. From left to right: Slc17a6 (excitatory neuronal marker), Slc32a1 (inhibitory neuronal marker), and Lepr. Color intensity from light gray to red represents gene expression levels from low to high (red, high expression).

CNS-32-e71185-s001.tif (1.2MB, tif)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Articles from CNS Neuroscience & Therapeutics are provided here courtesy of Wiley

RESOURCES