Introduction

Neuropeptidergic control of lipid metabolism is increasingly recognized as a conserved feature of energy homeostasis^1–7. Using the C. elegans model system, we had previously delineated a brain-to-gut neuroendocrine axis in which the sensory neuron-derived neuropeptide FLP-7 activates the intestinal G protein-coupled receptor NPR-22 to stimulate fat mobilization via induction of the adipocyte triglyceride lipase ATGL-1 and trigger fat loss via mitochondrial β-oxidation and increased energy expenditure^2,8. Phylogenetic analyses show that FLP-7 sequence resembles the mammalian tachykinin peptide family, and NPR-22 is the C. elegans ortholog of the mammalian neurokinin-2 receptor (NK2R)². We have established NPR-22 as both necessary and sufficient for driving fat loss in the worm intestine². Taken together, these findings are consistent with a conserved mammalian tachykinin-NK2R axis controlling lipid mobilization, prompting our examination of mammalian tachykinins and their receptors.

The mammalian tachykinins, including substance P, neurokinin A, and neurokinin B, constitute a conserved neuropeptide family that has historically been known to mediate nociception, inflammation, cancer progression, and gastrointestinal functions^9–13. These ligands signal through three G protein-coupled receptors (GPCRs): neurokinin-1 receptor (NK1R), neurokinin-2 receptor (NK2R), and neurokinin-3 receptor (NK3R), which are preferentially activated by substance P, neurokinin A, and neurokinin B, respectively. Prior studies have established NK1R as a key regulator of pain, inflammation, and immune responses^14–17. NK1R antagonists have demonstrated therapeutic potential by attenuating blood-brain barrier dysfunction, cerebral edema, and pro-inflammatory cytokine levels in traumatic brain injury models^18–21. In contrast, NK3R has been extensively studied in reproductive biology and neurological disorders^11,22,23. The neurokinin B-NK3R signaling is critical for hypothalamic-pituitary-gonadal axis function, with an NK3R antagonist now clinically deployed to relieve menopausal vasomotor symptoms^24,25. Beyond reproduction, NK3R has also been implicated in addiction, pain, and psychiatric disorders^22,26–28.

By comparison, NK2R regulates diverse physiological processes, including smooth muscle contraction, nociception, intestinal fluid secretion, and reproductive functions^29–34. Although NK2R is expressed in multiple tissues, including the gastrointestinal tract, lung, and immune system^35–41, its specific roles in metabolism and immunity remain poorly defined. A recent study⁴² reported that the administration of a selective NK2R peptide agonist increases energy expenditure, reduces fat mass, and suppresses appetite, highlighting its therapeutic potential for cardiometabolic diseases, including obesity and diabetes. However, the molecular mechanisms engaged by NK2R agonism are largely unknown.

Results

Expression pattern of Tachykinin Receptor 2 (Tacr2)

Our prior work in C. elegans showed that the CeNK2R gene npr-22 is expressed in a few pairs of head neurons and in the intestine, but not in the muscle or hypodermal tissues² (Fig. 1a). We examined Tacr2 expression in mammalian species: the Human Protein Atlas^43,44 RNA-seq datasets revealed that human TACR2 (hTACR2) is preferentially co-expressed with genes enriched in the intestinal “digestion” cluster (Fig. 1b). In contrast, hTACR1 and hTACR3 clustered with transcripts characteristic of connective tissue and retina, respectively (Extended Data Fig. 1a). Consistent with these patterns, quantitative PCR across mouse small-intestinal segments showed that Tacr2 mRNA levels exceeded those of Tacr1 and Tacr3 along the entire proximal-distal axis (Fig. 1c). Immunofluorescence localized the NK2R protein to the basolateral membrane of mouse intestinal epithelial cells (Fig. 1d), indicating a potential role in responsiveness to internal physiological signals rather than luminal nutrients. These results are consistent with our previous findings in C. elegans², in which NPR-22, the invertebrate ortholog of NK2R, is similarly expressed in intestinal cells and responds to internal signals from sensory neurons.

Figure 1.

Generation of Tacr2-null mice and analysis of gene expression in response to dietary challenges

To investigate NK2R function in vivo, we generated global Tacr2-null (Tacr2-/-) mice by CRISPR-Cas9 gene editing (Fig. 1e). qPCR of jejunal epithelium confirmed loss of Tacr2 transcripts (Fig. 1f) without compensatory up-regulation of Tacr1 or Tacr3 (Extended Data Fig. 1b). Tacr2-/- mice of both sexes were grossly normal, displaying wild-type body mass, re-feeding behavior and glucose tolerance (Fig. 1g-l).

Given the role of the C. elegans ortholog NPR-22 in intestinal lipid metabolism², we wished to determine whether NK2R modulates mammalian intestinal lipid handling. Tacr2-/- and wild-type littermates were maintained for one week on either a low-fat diet (LFD) or a western-style high-fat diet (HFD) before jejunal epithelial RNA was profiled by RNA-seq (Fig. 2a). Body mass remained comparable between Tacr2-/- and wild-type littermates of both sexes on either LFD or HFD (Extended Data Fig. 2a, b).

Figure 2.

In males, principal component analysis (PCA) revealed distinct clustering patterns: the wild-type mice clustered closely irrespective of diet; Tacr2-/- mice fed a LFD formed a separate cluster from wild-types; Tacr2-/- mice fed a HFD exhibited markedly different clustering compared to all other groups (Fig. 2b). Only 221 genes were differentially expressed (DE) between genotypes on LFD, whereas 3030 genes were DE on HFD (Fig. 2c, d), suggesting that loss of Tacr2 selectively sensitizes the male intestinal epithelium to dietary fat. Interestingly, Gene Ontology (GO) enrichment linked up-regulated transcripts to lipid-metabolic pathways and down-regulated transcripts to immune and inflammatory processes (Fig. 2e-h), concordant with prior C. elegans work that demonstrates a role for NK2R in intestinal lipid metabolism^2,8.

In females, the PCA was driven principally by genotype rather than diet (Fig. 2i). The number of DE genes remained similar under LFD and HFD (Fig. 2j, k), yet functional annotations mirrored those in males: lipid metabolism among up-regulated genes and immune response among down-regulated genes (Fig. 2l-o). Overlap between diet-specific DE gene sets was modest in males but substantial in females (Fig. 2p, q), indicating a differential response in males to dietary fat in the absence of NK2R.

NK2R governs epithelial lineage allocation and intestinal inflammation

Interestingly, genes downregulated in Tacr2-/- jejunum of both sexes were enriched for GO terms linked to immune response, immune regulation and leukocyte activation (Fig. 2f, h, m, o), indicating an overall dampening of mucosal immunity. Bisque deconvolution⁴⁵ of bulk RNA-seq revealed a basal shift in epithelial lineages in male Tacr2-/- mice (Fig. 3a). Tuft and transit-amplifying (TA) cell signatures were elevated, whereas stem-cell and goblet-cell signatures were reduced; HFD further exacerbated these baseline genotype differences, notably by reducing the relative proportion of enterocytes. Paneth cell signatures declined after HFD in both genotypes, consistent with previous studies^46,47, while endocrine cell abundance was unchanged. In contrast, female mice showed only minor alterations (Extended Data Fig. 3a); in Tacr2-/- females, HFD lowered stem-cell and enterocyte signatures and increased TA cells.

Figure 3.

Building on the transcriptional changes in immune-related genes observed in Tacr2-/- mice, we next assessed functional consequences of Tacr2 loss in vivo by inducing colitis in mice using the well-established dextran sodium sulfate (DSS) model⁴⁸ (Fig. 3b). Male Tacr2-/- mice were protected from the deleterious effects of induced colitis, exhibiting attenuated weight loss, reduced plasma and fecal lipocalin-2 (LCN2) concentration (a marker of intestinal inflammation⁴⁹), ameliorated histopathology with preserved architecture and less immune-cell infiltration. In a second experiment performed under identical conditions, we observed improved epithelial barrier function indicated by reduced FITC-dextran permeability via the intestinal lumen (Fig. 3c-i). In contrast, female Tacr2-/- mice showed no protection in weight loss or plasma LCN2, increased fecal LCN2, and histological injury comparable to controls, although barrier integrity was partially maintained (Extended Data Fig. 3b-h). Of note, male Tacr2-/- mice exhibited increased numbers of differentially expressed genes following HFD, an inflammatory insult^50–53, and demonstrated colitis protection, whereas female Tacr2^-/-mice lacked this expanded gene expression response and did not exhibit protection against DSS-induced colitis. Together, these data indicate a sex-dependent transcriptional and functional responses to inflammatory challenge in the absence of NK2R.

Prompted by this result, we tested whether acute receptor antagonism via a selective NK2R antagonist would phenocopy the genetic absence, and help ascertain translational tractability. Single-dose pharmacological inhibition corroborated the genetic findings: daily administration of the selective NK2R antagonist GR159897 during DSS exposure mitigated weight loss, suppressed plasma and fecal LCN2, and improved histological scores in males (Fig. 3j-n). Together, these data identify NK2R as a sex-dependent regulator of epithelial lineage allocation and susceptibility to mucosal injury.

Role of NK2R in Lipid Metabolism

Guided by our jejunal epithelial RNA-seq results, we next wished to determine the role of NK2R in lipid metabolism. Genes up-regulated in the Tacr2-/- jejunal epithelium were enriched for GO categories related to lipid and small-molecule metabolism, and this signature persisted and was further amplified under HFD challenge, indicating enhanced sensitivity to luminal lipid load. Because enterocyte lipid uptake culminates in cytosolic lipid droplet (LD) storage before export as chylomicrons, we profiled a curated set of 116 LD-associated genes⁵⁴ to examine whether NK2R modulates LD dynamics. In Tacr2-/- males, hierarchical clustering segregated HFD-fed Tacr2-/- samples from all other groups, whereas wild-type (WT) controls clustered together irrespective of diet (Fig. 4a). Tacr2-/- females also exhibited broad up-regulation of LD genes, although clustering by genotype and diet was less pronounced (Extended Data Fig. 5a).

Figure 4.

The enrichment of lipid-metabolic pathways and coordinated upregulation of the enterocyte LD program in the Tacr2-/- jejunal epithelium suggested that NK2R constrains the absorptive phase of dietary lipid handling and therefore regulates luminal lipid uptake. Following oil gavage, Tacr2-/- mice exhibited normal gastrointestinal transit (Extended Data Fig. 4a) but displayed a higher post-prandial triglyceride excursion (oral lipid-tolerance test, OLTT) and increased jejunal lipid retention 4 h after dosing in both sexes (Fig. 4b-h). Importantly, the OLTT reports the appearance of intestinally derived triglycerides in plasma. Thus, the enhanced excursion in Tacr2-/- mice indicates increased intestinal lipid entry into the circulation. The concurrent increase in jejunal lipid stores suggests that luminal uptake and intracellular storage of lipids are augmented, consistent with the observed induction of the LD program. Furthermore, pharmacological modulation recapitulated these effects: in diet-induced obese mice pre-treated with tyloxapol to inhibit lipoprotein lipase (LPL) and to block peripheral triglyceride clearance, the NK2R antagonist GR159897 elevated chylomicron/very-low-density lipoprotein (VLDL) triglycerides, whereas the NK2R agonist GR64349 produced the opposite outcome (Fig. 4i). Chronic HFD feeding for 12 weeks markedly expanded intestinal epithelial LD stores in Tacr2-/- mice of both sexes relative to wild-type controls (∼2-fold in males, ∼5-fold in females; Fig. 4j) without altering food intake, intestinal motility, metabolic rates, or locomotor activity (Extended Data Figs. 4b-k, 5b-k). Together, these data support a model in which NK2R signaling tonically regulates lipid metabolism in enterocytes by tuning luminal lipid uptake, intracellular LD storage and chylomicron secretion.

Since adipose tissue is one of the major depots for chylomicron-derived fatty acids⁵⁵, the intestinal phenotypes above predict altered lipid delivery to fat depots. Moreover, Tacr2 expression in white adipose tissue (WAT) (Extended Data Fig. 6a) raises the possibility of adipose-autonomous NK2R actions independent of the intestine. We therefore asked whether loss of NK2R remodels adipose transcriptional programs and wished to distinguish secondary consequences of increased intestinal export from adipose-intrinsic NK2R effects. Tacr2 is expressed in WAT, with comparable expression in both adipocyte and mesenchymal cell fractions (Extended Data Fig. 6a). To assess potential transcriptional changes, we performed RNA sequencing with perigonadal WAT at early (1 week) and chronic (12 weeks) time points after HFD exposure. (Extended Data Fig. 6b). Remarkably, the PCA revealed no genotype-dependent clustering in either male or female mice at either time point (Extended Data Fig. 6c-j). Furthermore, the number of differentially expressed genes between Tacr2-/- and wild-type mice was negligible, indicating that Tacr2 deletion does not appreciably alter the WAT transcriptome under HFD conditions. Together, these results suggest that, despite robust intestine-specific effects on lipid metabolism, the absence of NK2R has a limited impact on adipose transcriptional remodeling during diet-induced obesity.

Given that acute pharmacological activation of NK2R lowered chylomicron-associated triglycerides, we next asked whether sustained NK2R agonism alters whole-body energy balance. Diet-induced obese (DIO) mice received daily intraperitoneal injections of the NK2R-selective peptide agonist EB1002 (1000 nmol per kg) for 21 days (Fig. 4k). EB1002 is a long-acting NK2R agonist recently reported by the Gerhart-Hines group to improve cardiometabolic parameters in DIO models⁴². In our experiments, EB1002 treatment reduced body weight, perigonadal fat mass, and fasting glycemia of DIO mice relative to vehicle controls, concordant with the recent report⁴² (Fig. 4l, n, o). Food intake was transiently suppressed during the first week of treatment but normalized thereafter (Fig. 4m), indicating that the weight loss was not solely attributable to persistent hypophagia. Together with the acute lipid-handling phenotypes, these data position NK2R as a key direct regulator of intestinal lipid mobilization and long-term systemic energy balance.

Male Tacr2-/- mice have altered microbial populations

To determine whether Tacr2-/- males are uniquely prone to HFD-induced dysbiosis, fecal bacterial communities were profiled by 16S rRNA gene sequencing. α-diversity, assessed by the Shannon index, was reduced in Tacr2-/- mice irrespective of diet (Fig. 5a). β-diversity analysis (weighted UniFrac distances) revealed significant main effects of genotype (PERMANOVA, p=0.002) and diet (p=0.017), as well as a genotype-diet interaction (p=0.001) (Fig. 5b), indicating that both factors independently and synergistically shape the fecal microbiota, with Tacr2 deletion predisposing to larger diet-driven compositional shifts.

Figure 5.

At the phylum level, HFD feeding increased the relative abundance of Firmicutes and concomitantly decreased Bacteroidetes, yielding the characteristic rise in the Firmicutes:Bacteroidetes ratio^46,56,57 (Fig. 5c). A similar shift was observed in Tacr2-/-mice under HFD conditions. Linear discriminant analysis effect size (LEfSe) identified 66 operational taxonomic units (OTUs) that were diet-responsive and 104 that were genotype-responsive (Fig. 5d, e); 18 OTUs were influenced by both variables (Fig. 5f). Notably, OTU 215 was completely absent in Tacr2-/- mice, revealing a genotype-specific microbial deficit.

Discussion

In this report we identify neurokinin-2 receptor (NK2R) signaling as a nexus between intestinal lipid handling, mucosal immunity and host-microbe interactions. Tacr2 deletion or acute pharmacological antagonism augmented chylomicron export and massively expanded lipid-droplet (LD) stores in jejunal enterocytes, whereas NK2R agonism suppressed post-prandial triglyceridemia and reduced adiposity. These bidirectional effects suggest that NK2R serves as a molecular switch for intestinal lipid mobilization. The downstream cascade by which NK2R transduces neuropeptidergic input to LD dynamics remains to be elucidated; however, transcriptomic profiling of Tacr2-/- jejunal epithelium from the present study uncovered the coordinated up-regulation of a core LD gene module, suggesting that tonic NK2R activity normally restrains this program.

Bisque deconvolution revealed that Tacr2-null males, but not females, remodeled goblet and tuft cell lineages when challenged with a HFD. These secretory cell types shape barrier function and epithelial sensing of luminal antigens^58,59 and are associated with intestinal inflammatory responses^60–63, providing a plausible link to the striking male-specific protection from DSS-induced colitis. The absence of similar lineage shifts in females may underlie their failure to gain colitis resistance despite harboring ever larger epithelial lipid stores. Sex hormones, which modulate tachykinin and receptor expression^64,65, may be one contributing factor to this divergence and warrant further study.

Gut-innervating nociceptors regulate mucosal protection against dietary, microbial, and inflammatory insults^66–68. These nociceptors control mucin secretion from goblet cells, to create a physical barrier between the epithelium and the luminal contents⁶⁹. Recent evidence suggests that nociceptor-derived Substance P plays a crucial role in regulating mucosal protection during colitis and dysbiosis⁶⁶. Tac1 excision (which encodes both Substance P and NKA, the endogenous ligand for NK2R) from these neurons aggravates colitis⁶⁶, whereas global NK2R ablation confers male-biased protection, suggesting that proinflammatory signaling through NK2R on immune or stromal cells may outweigh epithelial loss of tachykinin responsiveness during injury. In future efforts, conditional Tacr2 deletion in distinct tissues as well as intestinal epithelial versus hematopoietic compartments will be essential to define the cell- and tissue-specific roles of NK2R.

Consistent with previous studies⁵⁷, the HFD increased the fecal Firmicutes:Bacteroidetes (F:B) ratio after only one week, and this shift was reproduced in Tacr2-/- mice. We note here that our diets are matched for nutritional composition, differing only in fat and sugar sources and percent of total %kcal, limiting confounding effects of dietary fiber in our study. 18 operational OTUs were jointly influenced by genotype and diet; 12 belonged to the Clostridiales order [Families: Lachnospiraceae (5), Ruminococcaceae (3), and unclassified lineages (4)]. Members of Lachnospiraceae have been linked to constipation, whereas Ruminococcaceae abundance correlates with irritable bowel syndrome and type 2 diabetes⁷⁰. Notably, the HFD expanded Ruminococcus gnavus (OTU575) and depleted Akkermansia muciniphila (OTU 1438), a combination associated with impaired barrier function through altered immune tolerance, short-chain fatty acid production and mucin layer maintenance^71–73. Conversely, Bifidobacterium pseudolongum (OUT 9), which is often negatively associated with metabolic and inflammatory disorders^74–76, was increased. Although this composite profile resembles that observed in inflammatory bowel disease, Tacr2 deletion conferred resistance to DSS-induced colitis in males. These findings underscore the complexity of host-microbiota interactions and suggest that NK2R signaling modulates intestinal inflammation through mechanisms that extend beyond simple taxonomic shifts, potentially involving neuropeptide-dependent regulation of epithelial and immune function.

Collectively, our findings position NK2R at the intersection of metabolic and inflammatory pathways in the intestine, supporting the tachykinin-NK2R axis as highly tractable to therapeutic intervention. In principle, NK2R agonists could be leveraged to treat metabolic diseases including obesity, whereas NK2R antagonists may ameliorate mucosal inflammation in inflammatory bowel disease. The pronounced sexual dimorphism underscores the need to consider sex as a biological variable in the development of NK2R-targeted therapeutics. Based on our findings, we expect that delineating the precise mechanisms by which NK2R integrates neuronal, immune, and microbial cues to regulate lipid droplet dynamics and mucosal immunity in the intestine will yield important mechanistic insights and advance the development of targeted therapies for metabolic and inflammatory diseases.

Methods

Animals

Male and female C57BL/6J mice (8-12 weeks old) were group-housed under a 12-h light/dark cycle with free access to food and water. Animals were fed either a low-fat control diet (CD; D14042701, Research Diets, New Brunswick, NJ; 73% kcal carbohydrate, 10% kcal fat, 17% kcal protein) or a Western-style high-fat diet (HFD; D12079B, Research Diets; 43% kcal carbohydrate, 40% kcal fat, 17% kcal protein). Body weights were measured weekly between 0900 and 1100 hours. All animal protocols were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at The Scripps Research Institute.

RNA extraction and RT-qPCR analysis

Mice were euthanized between 0900 and 1100 hours, and tissues were promptly collected and washed with ice-cold phosphate-buffered saline (PBS, 4°C) before snap-freezing in liquid nitrogen. The small intestine was washed with chilled PBS, opened longitudinally on ice, and the contents were removed. Intestinal epithelium was isolated by scraping with glass slides from the underlying submucosal layer and snap-frozen in liquid nitrogen. Samples were stored at −80°C until processing. Tissues were homogenized in TRIzol reagent, followed by phase separation with 1-bromo-3-chloropropane via centrifugation (8000g, 4°C). The upper aqueous phase was collected, and total RNA was purified using an RNeasy Mini Kit (Cat. No. 74104, Qiagen, Redwood City, CA) per the manufacturer’s instructions. cDNA synthesis was performed using iScript Reverse Transcription Supermix (Cat. No. 1708840, Bio-Rad, Hercules, CA). Quantitative PCR was conducted using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad) according to the manufacturer’s protocols. Gene expression was normalized to reference housekeeping genes (Actb, Rplp0, or Hprt). RNA integrity was preserved by sanitizing all surfaces with 70% ethanol and RNase inhibitor (RNase OUT, G-Biosciences, St. Louis, MO). Data normalization was performed against either the small intestine or Vil1 expression. Primer sequences are listed in the Supplementary Table 1.

Immunohistochemistry

The proximal small intestine was excised and flushed with ice-cold modified Bouin’s fixative, then fixed in 10% buffered formalin for 24 hours at room temperature. Tissue was embedded and frozen in optimal cutting temperature (OCT) compound (Fisher Healthcare, Chino, CA) on dry ice. Sections (16 μm) were cut using a cryostat (Leica), mounted onto charged glass slides, permeabilized with 0.025% Triton X-100 in TBS, and blocked with 10% normal donkey serum and 1% bovine serum albumin (BSA) in TBST (Millipore Sigma, Cat. No. 566460). Sections were incubated with primary rabbit antibody against Neurokinin 2 Receptor (NK2R; 20 µg/mL; Cat. No. ATR-002, Alomone Labs, Jerusalem, Israel), followed by incubation with AlexaFluor 488-conjugated donkey anti-rabbit secondary antibody (Cat. No. A-21206, Thermo Fisher Scientific). Sections were washed, mounted with ProLong Gold Antifade reagent containing DAPI (Thermo Fisher Scientific) for nuclear staining, and imaged using a Nikon A1 confocal microscope with a 60x objective at room temperature. Image processing was conducted using ImageJ software version 2.0.0 (NIH, Bethesda, MD).

CRISPR-Cas9 gene editing

Two single-guide RNAs (sgRNAs) were designed to target intronic regions flanking exon 2 of the Tacr2 gene. A single-stranded DNA (ssDNA) repair template (721 nucleotides) containing homology arms (80 nucleotides each) flanking exon 2 and a LoxP sequence within the intronic regions was synthesized. Following Cas9-mediated DNA cleavage and homologous recombination using the ssDNA repair template, loxP sites were integrated flanking exon 2 to generate Tacr2^fl/fl mice. Non-homologous end joining after Cas9 cleavage resulted in the deletion of exon 2, generating Tacr2^−/− mice. Constructs (1 ng/µL) were microinjected into fertilized eggs to produce genetically modified mice, which were validated by sequencing. Genotyping primer sequences are listed in the Supplementary Table 2.

Glucose Tolerance Test

Mice were fasted for 12 h prior to glucose tolerance testing. Fasted mice received an intraperitoneal injection of glucose solution (10% w/v) at a dose of 1 g glucose per kg body weight. Blood glucose concentrations were measured from tail blood samples at 0, 15, 30, 45, 60, and 120 min post-injection using a commercial handheld glucometer.

RNA sequencing and analysis

RNA from intestinal epithelium was extracted as described above. Total RNA samples were prepared into RNA-Seq libraries using the NEBNext® Ultra II Directional RNA Library Prep Kit for Illumina following the manufacturer’s recommended protocol. Briefly, for each sample, 200 ng total RNA was polyA selected, and converted to double-stranded cDNA followed by fragmentation and ligation of sequencing adapters. The libraries were then PCR amplified 12 cycles using barcoded PCR primers, purified, and size selected using AMPure XP Beads before loading onto an Illumina NextSeq 2000 for 100 base single-read sequencing. Raw sequencing reads were quality-assessed using FastQC, and adapters were removed using Cutadapt. Trimmed reads were aligned to the mouse reference genome (ENSEMBL GRCm38) using STAR aligner version 2.6.1d⁷⁷. The distribution of mapped reads across genomic features was analyzed with RSeQC v3.0.1. Differential gene expression analysis was performed using DESeq2 (v2.11.40.8) implemented in R (v4.2.2) via the Galaxy platform. Genes were considered significantly differentially expressed at thresholds of adjusted P < 0.05 and absolute log₂ fold-change >1.0. Gene ontology, pathway enrichment, and network analyses were conducted using ShinyGO v0.80⁷⁸. Cell-type-specific deconvolution was performed using BisqueRNA v1.0.5 against a published single-cell RNA sequencing dataset profiling mouse small intestinal epithelium^79,80. Raw and processed sequencing data will be deposited in the Gene Expression Omnibus (GEO).

Radioactive Tracer Oil challenges

Mice were fasted overnight prior to testing. Fasted mice were pretreated with Poloxamer-407. 10 g of Poloxamer-407 were resuspended in 100 mL of 0.9% NaCl saline and stirred overnight at 4°C. 10 µL/g of body weight was administered by intraperitoneal injection right before the oil gavage with 3 μCi [3H] triolein. Blood was collected at time 0, 1, 2, and 4 h, and the plasma was separated by centrifugation. Radioactivity was measured by scintillation.

C₁₆-BODIPY absorption assay

Mice were fasted overnight before receiving an oral gavage of olive oil containing 2 µg/g body weight of fluorescently labeled fatty acid analog C₁₆-BODIPY (4,4-difluoro-5,7-dimethyl-4-bora-3a,4a-diaza-s-indacene-3-hexadecanoic acid; Thermo Fisher, cat. no. D3821) administered at a volume of 10 µL/g body weight. Mice were euthanized 4 h post-gavage, and intestinal tissues were collected, snap-frozen, sectioned, and imaged as previously described⁸¹.

Gastrointestinal transit assay

Gastrointestinal transit time was assessed by oral gavage of a semi-liquid Evans Blue dye suspension (5% Evans Blue, 0.5% methylcellulose in phosphate-buffered saline; 100 µL per mouse). Following gavage, mice were monitored individually, and the time to appearance of the first blue-colored fecal pellet was recorded. Measurements were conducted in mice maintained on a control diet or after 12 weeks on a high-fat diet (HFD), as previously described⁸².

FPLC assay

Mice were fasted overnight prior to testing and pretreated with either lipoprotein lipase (LPL) inhibitor Tyloxapol (500 µg/g body weight), vehicle (PBS), the NK2R agonist GR64349 (10 mg/kg), or the NK2R antagonist GR159897 (10 mg/kg). Following pretreatment, mice received an oral gavage of olive oil. Blood plasma was collected 2 h post-gavage, separated by fast protein liquid chromatography (FPLC), and triglyceride levels were quantified across collected fractions.

Fecal lipid extraction

Mice were housed in wire-bottom cages for 12 h to collect fecal pellets. Pellets were dried, weighed, and pulverized using a mortar and pestle, then incubated in a 2:1 (v/v) chloroform:methanol solution. Samples were vortexed for 1 min and centrifuged at 1,000g for 10 min. The lower organic phase was collected, dried, and the residual lipids were weighed. Lipid content was normalized to the initial dried fecal weight⁸³.

Feeding behavior (CLAMS)

Indirect calorimetry was performed using a computer-controlled, open-circuit system (Oxymax System) integrated into the Comprehensive Lab Animal Monitoring System (CLAMS; Columbus Instruments, Columbus, OH)^84,85. Single-housed, acclimated mice were placed into clear respiratory chambers (20 × 10 × 12.5 cm) equipped with a water sipper tube, a food tray connected to a balance for continuous food intake measurement, and 16 photobeams arranged in two axes at 0.5-inch intervals to monitor motor activity. Room air was circulated through each chamber at a flow rate of 0.5 L/min, and exhaust air was sampled every 15 min for 1 min. Oxygen consumption (VO₂) and carbon dioxide production (VCO₂) were measured using O₂ and CO₂ sensors (Columbus Instruments), and respiratory exchange ratio (RER) was calculated as VCO₂/VO₂. All metabolic parameters were normalized to lean body mass, which was determined by EchoMRI analysis.

Dextran sulfate sodium (DSS) colitis model

Colitis was induced by providing mice with ad libitum access to drinking water containing dextran sulfate sodium (DSS; 3.0% w/v for males, 5.0% w/v for females; MP Biomedicals, cat. no. MFCD00081551) for 7 consecutive days. Mice were monitored daily for changes in body weight and clinical condition. Animals that lost more than 25% of their initial body weight were euthanized and excluded from further analysis. On day 7, all remaining mice were euthanized, and tissues were collected for downstream analyses⁸⁶.

Tissue scoring parameters

The intact colon was collected, flushed with chilled phosphate-buffered saline (PBS), and fixed overnight in 4% paraformaldehyde (PFA) in PBS. The distal 1 cm segment was cryosectioned, stained with hematoxylin and eosin (H&E), and imaged. Histological sections were scored based on epithelial integrity (0-4) and immune cell infiltration (0-4) for a maximum aggregate score of 0-8⁸⁷. Epithelial integrity was scored as follows: 0, normal morphology; 1, loss of goblet cells; 2, loss of goblet cells in multiple regions; 3, loss of crypts; 4, loss of crypts in multiple regions. Immune cell infiltration was scored as follows: 0, no infiltration; 1, infiltration around the crypt base; 2, infiltration reaching the lamina muscularis; 3, extensive infiltration of the lamina muscularis with mucosal thickening and edema; 4, infiltration extending into the submucosa.

Fecal protein extraction for Lipocalin-2 quantification

Fecal pellets were collected at the time of sacrifice and flash-frozen in liquid nitrogen. Approximately 100 mg of frozen feces per sample was weighed, suspended in 1 mL of 1% Tween-20 in PBS, and vortexed vigorously for 10 min. Samples were then centrifuged at 15,000g for 10 min at 4°C. The resulting supernatant was collected and diluted (1:50 for control samples, 1:2000 for DSS-treated samples) prior to measurement of Lipocalin-2 concentrations using a commercial mouse Lipocalin-2 ELISA kit (R&D Systems, cat. no. DY1857)⁴⁹.

Intestinal permeability assay

Mice were fasted beginning at 09:00 for 4 h prior to oral gavage with fluorescein isothiocyanate-dextran 4000 (FD4; Millipore Sigma, cat. no. 46944) at a dose of 0.6 mg/g body weight from a 100 mg/mL stock solution. Four hours after gavage, blood was collected via cardiac puncture into BD Microtainer tubes containing lithium heparin and maintained on ice. Plasma was isolated by centrifugation at 5,000g for 5 min at 4°C. Plasma from fasted mice that did not receive FD4 was used to generate a standard curve (0-10 µg/mL FD4), and plasma FD4 concentrations in experimental samples were quantified accordingly^88,89.

Microbiome 16S rRNA sequencing and microbial analysis

Fecal pellets were collected from mice, and genomic bacterial DNA was extracted using the DNeasy PowerLyzer PowerSoil Kit (Qiagen, cat. no. 12855) according to the manufacturer’s instructions. DNA concentration was quantified and samples were submitted to GENEWIZ (South Plainfield, NJ, USA) for 16S-EZ sequencing using Illumina paired-end 2 × 250 bp methodology. Raw sequence reads were processed with Cutadapt (v1.9.1) to remove adapter sequences and assigned to operational taxonomic units (OTUs) based on a 97% similarity threshold using VSEARCH (v1.9.1) and QIIME (v1.9.1). Differential microbial abundance analyses were performed using linear discriminant analysis effect size (LEfSe)⁹⁰.

Statistical analysis

Data are presented as mean ± SEM. The number of biological replicates is provided in the figure legends. Comparisons between two groups were performed using unpaired two-tailed Student’s t-tests. Repeated measures two-way ANOVA was used for comparisons across time points, and one-way or two-way ANOVA followed by Sidak’s or Tukey’s post hoc multiple comparisons tests, as appropriate, were used for analyses involving more than two groups. Differences in UniFrac distances were assessed using PERMANOVA. Statistical significance was defined as P < 0.05. Outliers were identified using Grubb’s test. All statistical analyses were performed using GraphPad Prism v10.3.

Data availability

RNA sequence data will be deposited in GEO. We'll submit the data once the repository is operational again, as the GEO FTP server is currently down due to the government shutdown. All data generated or analyzed during this study are included in the manuscript; source data files have been provided for all figures.

Supplementary figures and tables

Extended Data Figure 1.

Extended Data Figure 2.

Extended Data Figure 3.

Extended Data Figure 4.

Extended Data Figure 5.

Extended Data Figure 6.

Supplementary Table 1:

Supplementary Table 2:

Acknowledgements

This work was supported by seed funding from the Scripps Research Institute. P.T. was supported by National Institutes of Health (NIH) grants R01 DK142199 and R01 HL175773. P.P. was supported by the NIH T32 Immunology Training Program (5T32AI007244). C.L. was supported by a Dorris Scholar Award from the Dorris Neuroscience Center at The Scripps Research Institute. We thank members of Enrique Saez’s lab at Scripps Research for technical assistance, support and critical advice during the course of the studies. We also thank Dr. Xinglin Yang (Howard Hang Lab) for technical expertise and guidance on microbiome workflows. We are grateful to Ayub Khan and Aayushi Shah (Srinivasan Lab) for technical assistance with tissue collection, and to Dr. Amanda Roberts (Scripps Research Animal Models Core) for support with mouse experiments. We also acknowledge the resources and staff of the Scripps Research Genomics Core Facility for sequencing services. Elements of Figures 2a, 3b, 4b, 4k, 5f, and Extended Data Figure 6b were created with BioRender.com.