Colonic metabolomic and transcriptomic alterations in a mouse model of metabolic syndrome

Animals.

Male MS NASH/PcoJ (030888) (7–11) and C57BL/6J (000664) mice were purchased from Jackson Laboratory. All mice were 14–16 weeks of age. Male MS NASH mice were specifically used because this model has been extensively characterized in males and reproducibly develop obesity, insulin resistance, steatosis, and steatohepatitis, whereas females do not develop overt diabetes or systemic insulin resistance to the same degree as male mice. All experiments were approved by the Institutional Animal Care and Use Committee of the University of New Mexico Health Sciences Center, in accordance with the National Institutes of Health guidelines for use of live animals. The University of New Mexico Health Sciences Center is accredited by the American Association for Accreditation of Laboratory Animal Care.

Cells and Tissue Preparation.

Upon necropsy, tissues were weighed, measured, sectioned out for various procedures, and flash frozen. Samples collected for downstream analysis included colon, liver, perigonadal fat pads, and stool. These samples were either flash frozen in liquid nitrogen and stored in −80C, fixed with formalin, or put into 70% ethanol (EtOH). Stool samples collected at euthanasia were processed and normalized to 90mg/mL using PBS + 0.1% TWEEN to analyze inflammatory markers of the colon such as Lipocalin-2 (LCN-2)(12) using R&D Systems Quantikine ELISA Mouse Lipocalin-2/NGAL Immunoassay (Catalog no. MLCN20). A section of the liver was stored in 10% formalin for 48 hours, then washed and stored in 70% EtOH prior to processing for tissue sectioning. Histopathological evaluation was performed on mouse liver tissue. Formalin-fixed, paraffin-embedded (FFPE) sections of livers were prepared at the University of New Mexico Comprehensive Cancer Center (UNMCCC) Human Tissue Repository (HTR) Core Facility. Sections (5 μm) were cut from blocks fixed in 10% neutral-buffered formalin and stained with hematoxylin and eosin (H&E) using standard protocols. The degree of intestinal inflammation scoring was performed based on the guidelines provided by Erben et al (13). In brief, colon sections were evaluated based on the guidelines for chemically-induced colonic inflammation specifically. Inflammatory status was graded based on two major criteria a) inflammatory cells infiltrate and b) intestinal architecture. Inflammatory cell infiltrate was assessed for the severity (mild, moderate and marked) and the extent (mucosa, mucosa and submucosa and transmural) and received a score of 0–3. The intestinal architecture was assessed for epithelial changes (focal and generalized erosions) and mucosal architecture (focal and extended ulcerations +/− granulation tissue +/− pseudopolyps) and received a score of 0–3. Scores in both categories were summed together to provide a final overall score of 0–6. For the generation of colonic organoids (colonoids), media was comprised of Advanced Dulbecco’s modified Eagle medium/Ham’s F-12 (ThermoFisher, Waltham, MA), 100 U/mL penicillin/streptomycin (Quality Biological, Gaithersburg, MD), WNT surrogate-Fc fusion protein (ImmunoPrecise Antibodies Ltd, Utrecht, The Netherlands) 15% v/v R-spondin1 conditioned medium (cell line kindly provided by Calvin Kuo, Stanford University), 10% v/v Noggin conditioned medium (cell line kindly provided by Gijs van den Brink, Tytgat Institute for Liver and Intestinal Research), 1X B27 supplement (ThermoFisher), 10 mM HEPES (ThermoFisher), 1X GlutaMAX (ThermoFisher), 1mM N-acetylcysteine (MilliporeSigma), 50 ng/mL human epidermal growth factor (ThermoFisher), 10 nM [Leu-15] gastrin (AnaSpec, Fremont, CA), 500 nM A83–01 (Tocris, Bristol, United Kingdom), 10 μM SB202190 (MilliporeSigma), 100 mg/mL primocin (InvivoGen, San Diego, CA). Base media, for differentiation of colonoids, had the same composition of media but lacked WNT surrogate Fc fusion protein, Rspo-1 and SB202190. Colonic crypt isolation and colonoid generation were prepared from proximal colons as previously reported (14–17). Isolated crypts were resuspended in Matrigel (Corning) and 25 μL droplets were plated in a 24-well tissue culture plate (Corning). After polymerization at 37°C, 500μL of colonoid expansion media was added for 2 d. After 2 d, the colonoid expansion media was replaced every other day. Colonoids were passaged every 3–7 d by harvesting in Cultrex Organoid Harvesting Solution (Bio-Techne, Minneapolis, MN) at 4°C with shaking for 45 min, as previously described (16, 17). All colonoid cultures were maintained at 37°C and 5% CO2. Unless noted, colonoid lines have been passaged >10 times. Colonoids were harvested from Matrigel using Cultrex Organoid Harvesting Solution

Microbiota analysis.

Fresh fecal samples from C57BL/6J and MS NASH mice were collected fresh in sterile tubes and flash frozen. Microbial communities were determined by sequencing of the 16S rRNA as previously reported (12, 18) with minor modifications described below. Microbial DNA was isolated from feces using the ZymoBIOMICS DNA Miniprep Kit (Zymo Research) following manufacturer’s recommendations. Variable regions V-3 through V-4 of the 16S rRNA gene were amplified by PCR using 100 ng input of DNA for each sample in duplicate using primers 319F- (5'-ACTCCTRCGGGAGGCAGCAG-3') and 806R- (5'-GACAGGACTACHVGGGTATCTAATCC-3') containing Nextera adapter overhangs. A second PCR was performed with Nextera^® XT Index Kit v2 Set A (Illumina) to complete the adapter and add a unique sample-specific barcodes. After each PCR, a clean-up with AxyPrep Fragment Select-I magnetic beads (Axygen Biosciences) was completed, and all PCR reactions were run on an Applied Biosystems 2720 Thermal Cycler. Indexed samples were combined to yield duplicate 300 ng pools, followed by the creation and denaturation of a 4 nM library, and paired 250bp sequencing runs were completed on the Illumina MiSeq using v3 sequencing chemistry (Illumina).

NanoString nCounter Gene Expression Analysis.

Gene expression profiling was performed on proximal colon tissue using the NanoString nCounter XT platform (NanoString Technologies, Seattle, WA). Total RNA was isolated from proximal colon samples (n=12) and subjected to quality control prior to analysis. RNA quantity was initially assessed by NanoDrop spectrophotometry. Fluorometric quantification was performed using the Qubit RNA High Sensitivity assay (Thermo Fisher Scientific). RNA integrity and fragmentation profiles were evaluated using a Fragment Analyzer (Agilent Technologies). RNA QC summary metrics included total concentration, fragment distribution profiles, and assessment of 18S and 28S ribosomal RNA peaks. Because samples exhibited moderate RNA fragmentation, input RNA amounts were adjusted above the standard 100 ng recommended for intact RNA. Based on QC recommendations, 150–300 ng of total RNA per sample (typically 200–250 ng depending on degradation level) was used for hybridization. The same RNA input amount for each sample was maintained across both NanoString panels. NanoString Hybridization and Data Acquisition Gene expression profiling was performed at the University of Arizona Genetics Core using the NanoString nCounter XT platform (NanoString Technologies, Seattle, WA). Gene expression was assessed using a pre-designed NanoString mouse CodeSet panel: XT_Mm_Metabolic Pathways_CSO (Catalog no. 115000374). Reactions were performed using the nCounter XT Master Kit for 12 reactions (Catalog no. 100052). Hybridization was conducted at 65°C for 17 hours (overnight) according to manufacturer instructions. Following hybridization, samples were processed on the NanoString nCounter Prep Station and analyzed on the nCounter Digital Analyzer to obtain raw reporter code counts (RCC files). Reporter Library Files (RLFs) corresponding to each panel were used to map barcode counts to target transcripts during downstream analysis. For data processing and quality control, raw RCC files were imported into NanoString nSolver Analysis Software (NanoString Technologies). All samples passed initial NanoString quality control metrics, including imaging quality, binding density, positive control linearity, and limit of detection. Data normalization was performed using standard NanoString workflows, including: a) Positive control normalization to adjust for system variation and b) CodeSet content normalization using housekeeping genes. Normalized counts were used for downstream differential expression and pathway analyses.

Metabolomics.

Metabolomic profiling of colon tissue from C57BL/6J and MS NASH mice was performed by Human Metabolome Technologies, Inc. (HMT; Tsuruoka, Japan) using capillary electrophoresis–time-of-flight mass spectrometry (CE-TOFMS) (Supplementary Table S1). Eight colon samples (n = 4 C57BL/6J, n = 4 MS NASH) were homogenized in 50% acetonitrile (v/v) containing 20 μM internal standards, centrifuged, and filtered through 5-kDa cut-off membranes to remove macromolecules. The filtrates were concentrated and resuspended in ultrapure water prior to analysis. Both cationic and anionic metabolites were analyzed under positive and negative electrospray ionization modes, respectively, using Agilent CE-TOFMS instruments (Agilent Technologies, Palo Alto, CA) equipped with fused-silica capillaries (50 μm × 80 cm). Analytical parameters included a capillary voltage of 30 kV, detector voltage of 4.0 kV, and a mass scan range of m/z 50–1000. Data were processed using MasterHands software (version 2.17.11; Keio University) for automated peak extraction, alignment, and normalization to internal standards. Metabolites were annotated by migration time (±0.5 min) and mass accuracy (±10 ppm) against HMT’s standard library. Relative peak areas were log₂-transformed and mean-centered, and metabolites detected in fewer than 75% of samples were excluded. Statistical analyses were performed using unpaired two-tailed Student’s t-tests (p < 0.05), and multivariate analysis by principal component analysis (PCA) and hierarchical clustering (Ward’s linkage, Euclidean distance) was used to visualize global metabolic differences between groups. Human stool metabolomics from individuals with and without metabolic syndrome was conducted by the NYU Langone Metabolomics Core Resource Laboratory (New York, NY) using a hybrid liquid chromatography–mass spectrometry (LC-MS) platform (Supplementary Table S2). Approximately 10 mg of stool per subject was extracted in aqueous-organic solvent containing internal standards and analyzed on an Agilent 6545 Q-TOF mass spectrometer operating in both ESI+ and ESI− modes. Mass accuracy was maintained within ±1.2 ppm and retention-time stability within 0.1 min across all runs. Internal quality-control samples were used to monitor instrument precision and reproducibility. Data acquisition and peak processing were performed using Agilent MassHunter Qualitative Analysis and Profinder Suite (version 10.0). Metabolite identities were confirmed by accurate mass, isotopic distribution, and retention-time matching to NYU’s in-house MS/MS spectral library. Processed data were log₂-transformed, normalized to internal standards, and analyzed by unpaired two-tailed Student’s t-tests (p < 0.05). PCA and hierarchical clustering (Ward’s linkage, Euclidean distance) were used to assess group separation between healthy and metabolic-syndrome samples. For both mouse and human datasets, we used MetaboAnalyst v6.0 to run pathway enrichment analysis and check disease-signature enrichment using the feces metabolite-set library (Supplementary Table S3). Enrichment testing was conducted using the well-established globaltest algorithm, which calculates a Q-statistic for each metabolite set, incorporating generalized linear modeling to account for pathway topology. To directly compare directionality across species, we plotted log₂ fold-change (mouse: MS NASH vs C57BL/6J on the x-axis; human: MetS vs Control on the y-axis) for metabolites shared between datasets; per-metabolite p values were obtained with unpaired two-tailed t-tests and adjusted by Benjamini–Hochberg FDR, and metabolites were classified as increased or decreased when |log₂FC| ≥ 0.10 in both species. Only metabolite sets with five or more members were evaluated.

RNA isolation, quantification, and RT-qPCR.

RNA isolation was performed using the RNA Purelink Minikit (Invitrogen) according to the manufacturer’s protocols on colonoids at passage 12 where either kept at intestinal stem cell state or allowed to differentiate for 3 days and then harvested from Matrigel using Cultrex Organoid Harvesting Solution as previously described [5]. RNA was quantified on Nanodrop2000 and all samples yielded a 260/280 of 2 ± 0.15. cDNA synthesis was performed using SuperScript IV VILO Master Mix (ThermoFisher). Reverse transcription reaction utilized the viiA7 Thermal Cycler (ThermoFisher and RT-qPCR runs utilized Taqman MasterMix (ThermoFisher) using the QuantStudio 3. RT-qPCR primers (TaqMan Assays) and reagents used are listed in Supplementary Table S4.

Microscopy and image analysis.

For confocal microscopy, colonoids at passage 15 were differentiated for 3 days and harvested from Matrigel using Cultrex Organoid Harvesting Solution. Cells were then fixed with 4% PFA followed by a wash with 1x PBS. Blocking buffer contained PBS with 15% FBS, 2% BSA and 0.1% saponin. After a 1-hour stain in primary antibody, cells were washed with PBS then followed by 1- hour stain in blocking buffer containing secondary antibody. Cells were mounted using ProLong Gold Antifade and imaged using the Zeiss LSM 800 Airyscan Confocal microscope with a 20x objective lens. Images were processed using Zen lite Software. All primary and secondary antibodies used for confocal staining are listed in Table S5.

Statistical Analysis.

Statistical analysis was performed as described in figure legends and graphs generated display mean (± SD) and were obtained using Prism software. Sample sizes differed between experiments depending on tissue availability and the requirements of each assay. Larger cohorts were used for phenotypic characterization, whereas transcriptomic, metabolomic, microbiome, and organoid experiments were designed as exploratory discovery studies using sample sizes commonly employed for these methodologies. Sample sizes for each experiment are reported in the corresponding figure legends. Microbiome data was sequenced and processed by Illumina’s service lab using their in-house analysis pipeline. Cluster analysis was performed using heatmap3 (19) package in R. T-test was used to measure specific microbiome species abundance between conditions. Adjusted P-value < 0.05 was used as significant threshold. Principle coordinate analysis was conducted in R. All other data were analyzed using one -way ANOVA or two-tailed unpaired Student’s t test (Prism).

MS NASH mice show phenotypic and microbial alterations in absence of intestinal inflammation.

MS NASH mice are a polygenic mouse model of obesity, metabolic syndrome, MASLD and diabetes with an intact leptin pathway that have proven to be a translationally-relevant model (7–11, 20). Male MS NASH mice have been found to progressively develop several features of metabolic disorders that are similar to humans including steatosis, obesity, dyslipidemia, and insulin resistance (7–11, 20). Therefore, we examined both MS NASH and control (C57BL/6J; B6) mice after being on a standard diet for 10-weeks. MS NASH mice were significantly heavier than controls and exhibited increased perigonadal fat mass (Fig. 1A–B). MS NASH mice also displayed hepatomegaly (Fig. 1C, H–I) with histological evidence of macrovesicular steatosis (Fig. 1F–G), consistent with MASLD progression. Interestingly, MS NASH mice had significantly elongated colons compared to controls (Fig. 1D).

Histological examination of the colon revealed prominent goblet cell hyperplasia in MS NASH mice, characterized by increased goblet cell abundance with visibly denser mucus within the colonic epithelium (Supplementary Fig. S1A-D). However, no significant inflammatory cell infiltrate or major mucosal architectural abnormalities were identified (Supplementary Table S6). Consistent with these histological findings, fecal lipocalin-2 (LCN-2) concentrations were unchanged between groups (Fig. 1E). Furthermore, colonic expression of inflammatory cytokines (Ifng, Tnfa, Il12b, and Il1b) and epithelial tight junction-associated genes (Ocln, F11r, Tjp1, and Cldn2) did not differ between MS NASH and control mice (Supplementary Fig. S1E-L). Collectively, these findings indicate that the metabolic and epithelial alterations observed in MS NASH mice occur in the absence of overt intestinal inflammation or major disruption of epithelial barrier-associated gene expression.

To determine whether metabolic dysfunction was associated with alterations in the gut microbiota, we performed 16S rRNA sequencing on fecal samples from MS NASH and control mice. Principal coordinate analysis (PCoA) revealed clear separation between microbial communities, with PC1 and PC2 explaining 37% and 29% of the variance, respectively (Fig. 1J). Hierarchical clustering and genus-level heatmap analysis further demonstrated distinct microbial signatures between groups (Fig. 1K). Several taxa were differentially abundant between groups. Notably, members of the genus Lactobacillus and Limosilactobacillus were increased in MS NASH mice, while multiple taxa including Blautia, Parabacteroides, Prevotella, and Roseburia were relatively enriched in controls. These alterations indicate a distinct shift in gut microbial composition associated with metabolic syndrome. This is characterized by enrichment of Lactobacillus and Limosilactobacillus and reduced relative abundance of several commensal anaerobic genera including Blautia, Parabacteroides, Prevotella, and Roseburia. Together, the PCA and hierarchical clustering results demonstrate that MS NASH mice exhibit significant gut microbiota remodeling, which may contribute to the metabolic and colonic alterations observed in this model.

Colonic gene expression analysis reveals metabolic pathways reprogramming in MS NASH mice.

To examine alterations in genes involved in core metabolic processes and signaling pathways, proximal colon tissues from B6 and MS NASH mice were profiled using the NanoString nCounter Metabolic Pathways gene expression panel. Genes were defined as upregulated or downregulated based on a ≥2-fold change in expression with a nominal P < 0.1 compared to control mice. Using these criteria, 298 genes were upregulated and 131 genes were downregulated in MS NASH mice (Fig. 2 A–E). Unsupervised hierarchical clustering of all 768 panel genes revealed clear segregation of MS NASH mice from controls, indicating widespread transcriptional remodeling of colonic metabolic programs (Fig. 2F). Pathway-level aggregation indicated increased expression of genes associated with carbon metabolism pathways, including glycolysis and the pentose phosphate pathway, as well as increased activity of the tryptophan-kynurenine metabolic axis (Fig. 2G). Complementary pathway enrichment analysis identified significant overrepresentation of gene sets associated with non-alcoholic fatty liver disease (NAFLD, currently renamed MASLD), oxidative phosphorylation, mitochondrial respiration, and carbon metabolism (Fig. 2H), consistent with transcriptional changes associated with altered bioenergetic demand and metabolic stress in the MS NASH colon. Collectively, these data demonstrate broad alterations in colonic metabolic gene expression that are consistent with metabolic remodeling of the colonic epithelium.

Colonic metabolomic profiling reveals conserved pathway alterations between MS NASH mice and human metabolic syndrome.

To determine whether the metabolic pathway alterations identified by NanoString transcriptomic profiling were also reflected at the metabolite level, we performed targeted metabolomic profiling of colon tissue from B6 and MS NASH mice. This analysis was designed to provide an independent assessment of the transcriptional findings by evaluating whether coordinated changes in metabolic pathways were accompanied by corresponding alterations in metabolite abundance. Unsupervised clustering of metabolite abundances revealed clear separation between control and MS NASH colons, indicating widespread metabolic reprogramming in the MS NASH colon (Fig. 3A). MetaboAnalyst pathway enrichment analysis using the RaMP-DB metabolite and lipid library identified significant alterations in glycolysis/gluconeogenesis, the pentose phosphate pathway, tryptophan metabolism, and sphingolipid metabolism-pathways central to cellular energy homeostasis, redox balance, and inflammatory signaling (Fig. 3B). To explore whether metabolic alterations observed in the MS NASH colon resemble metabolic signatures reported in human metabolic syndrome, we performed metabolite-set enrichment analysis using the MetaboAnalyst fecal metabolite disease library. Because comparable human colonic tissue metabolomic datasets are not currently available, this analysis was intended to identify shared disease-associated metabolic patterns rather than directly compare equivalent biological specimens. The top 25 enriched disease-associated metabolite sets in MS NASH mouse colons closely mirrored those observed in human metabolic syndrome fecal metabolomics datasets (5), including signatures linked to inflammatory bowel disease, colorectal cancer, and metabolic dysfunction (Fig. 3C–D). Enrichment testing was conducted using the globaltest algorithm (21), which applies generalized linear modeling to evaluate coordinated metabolite-set behavior. Comparison of metabolites represented in both datasets identified several metabolites that changed in the same direction in MS NASH colon tissue and published human fecal metabolomics datasets (Fig. 3E). Several metabolites were increased in both species, including 3-hydroxybutyric acid, uric acid, succinic acid, guanidinoacetic acid, N-acetylglutamic acid, and xanthosine. Notably, multiple concordantly elevated metabolites such as 3-hydroxybutyric acid and uric acid are established markers of mitochondrial dysfunction, altered energy metabolism, and MetS. Although different sample types were analyzed, this comparison was intended to identify shared metabolic patterns. These findings suggest partial conservation of metabolic signatures across species despite the different biological matrices analyzed.

Colonoids derived from MS NASH mice exhibit altered epithelial marker expression under stem cell and differentiated conditions.

To determine whether metabolic disease is associated with intrinsic changes in colonic epithelial cell phenotype, we generated colonic organoids (colonoids) from B6 and MS NASH mice. Colonoids were maintained under intestinal stem cell (ISC) conditions or induced to differentiate by withdrawal of WNT3A and R-spondin (Fig. 4A–E). Gene expression analysis demonstrated genotype-specific differences in epithelial lineage markers. Under ISC conditions, transcripts encoding the ISC marker Lgr5 (Fig. 4A) and the goblet cell marker Muc2 (Fig. 4D) differed between control and MS NASH colonoids. Upon differentiation following WNT3A and R-spondin withdrawal, Lgr5 expression was significantly reduced in both control and MS NASH colonoids (Fig. 4C–D), confirming effective induction of differentiation. In contrast, Muc2 transcript expression was significantly reduced in differentiated MS NASH colonoids relative to controls (Fig. 4E). Given our histological observation of goblet cell hyperplasia in the colons of MS NASH mice (Supplementary Fig. S1A-D), we performed immunofluorescence staining for MUC2 in differentiated colonoids. Representative images demonstrated the presence of MUC2-positive cells in both groups (Fig. 4F–G). Collectively, these findings indicate that epithelial cells from MS NASH mice exhibit intrinsic alterations in epithelial lineage marker expression that persist following ex vivo culture, consistent with persistent epithelial remodeling.