Dysregulation of cellular metabolism within the gut-brain axis is associated with behavioural changes in chronic intestinal inflammation

Abstract

Background

Inflammatory bowel disease (IBD) is a chronic debilitating condition significantly affecting patient quality of life. Although the exact aetiology remains unknown, accumulating evidence has shown that disruption of the gut-brain axis may be related to the occurrence and development of chronic intestinal inflammation. Psychological disorders are highly prevalent in patients with IBD. However, an association between altered behaviour and dysregulated metabolic pathways within the gut-brain axis is yet to be explored.

Methods

Metabolic multiplexed phenotyping system involving indirect calorimetry and flow-through respirometry monitors was used to assess energy metabolism in Winnie mice with spontaneous chronic colitis and C57BL/6 littermates. Depressive and anxiety-like behaviours were evaluated with light dark, open field, grooming, elevated plus maze, and forced swimming tests. To investigate underlying mechanisms of the metabolic changes in Winnie mice, glycolysis/gluconeogenesis, fatty acid ß-oxidation, tricarboxylic acid cycle and oxidative phosphorylation gene expressions were determined by transcriptome analysis using high-throughput sequencing of mRNA extracted from the distal colon and brain samples.

Results

Our findings showed that energy metabolism and spontaneous activity were reduced in Winnie mice corresponding to alterations in the expression of cellular metabolism-associated genes in the distal colon. Winnie mice displayed depressive and anxiety-like behaviours reflecting downregulation of glycolysis/gluconeogenesis, fatty acid ß-oxidation, tricarboxylic acid cycle and oxidative phosphorylation in the distal colon and brain. Subsequent analyses showed pro-inflammatory cytokine expression was upregulated in the Winnie mouse brain.

Conclusions

These data provide evidence that the dysregulation of cellular metabolism within the gut-brain axis underlies changes in behaviour and energy metabolism in chronic intestinal inflammation.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12974-025-03536-x.

Introduction

Inflammatory bowel disease (IBD), which includes ulcerative colitis (UC) and Crohn’s disease (CD), is characterised by chronic, relapsing, and remitting inflammation of the gastrointestinal tract [1]. The aetiology of IBD is yet to be fully elucidated, however it is considered to be multifactorial, involving complex interactions between genetic susceptibility, environmental factors, gut dysbiosis, immune response dysregulation, and enteric nervous system (ENS) dysfunction [2, 3]. Common clinical manifestations include diarrhoea, abdominal pain and cramping, blood in stools, weight loss, and chronic fatigue [4, 5]. In addition, IBD patients frequently experience behavioural changes, such as affective disorders, energy levels fluctuations, and sleep disturbances, compared with the general population, especially during active disease [6–8]. These symptoms are associated with anhedonia, dysthymia, restlessness, and hypochondriasis, which are distinct from transient sadness or bereavement [9]. Presentation of anxiety in IBD patients is characterised by worry, intolerance of uncertainty, cognitive avoidance, irrational fears, and corresponding cognitive responses [10, 11]. Furthermore, changes in behaviour associated with IBD can be characterised as sickness behaviour, a set of behaviours involving feeling unwell, increased sensitivity to pain, fever, lack of interest in social interactions, reduced energy, decreased exploration, less grooming, loss of pleasure, drowsiness, reduced appetite and weight, difficulty focusing, and heightened anxiety, commonly triggered by acute infections and tissue injuries [12–15]. Behavioural changes in IBD patients become incipient at disease onset, often remaining unrecognised until they develop into a pernicious, chronic syndrome, influencing social and interpersonal functioning, and reducing the patient’s quality of life (QoL) [13, 14, 16–18]. These comorbid psychological disorder symptoms are further associated with exacerbation of disease activity and course, leading to poorer outcomes and increased risk of relapse, hospitalisation, and surgery [19–21].

Although behaviour changes associated with IBD are known to represent different constructs, they are highly inter-correlated and share biological and inflammatory pathways in the gut and brain [13–15]. Given that psychological comorbidities of IBD significantly reduce patient QoL, as well as impact the course of the disease itself, it is widely considered that disruption of the gut-brain axis, a bi-directional communication network driven by neural, hormonal, metabolic, immunological, and microbial signals, is a significant factor in the association between psychological distress and intestinal inflammation [22, 23]. These pathway disruptions are multifaceted, involving genetic and biochemical factors that are proposed to affect cellular and energy metabolism, as well as neurological function [20, 21]. Hence, although the complex mechanisms underlying the interaction between the pathophysiology of IBD and psychological comorbidities are not yet fully understood, it is plausible to consider the involvement of altered cellular and energy metabolism [24].

Cellular metabolism is reliant upon numerous biochemical pathways, such as glycolysis, fatty acid β-oxidation (FAO), tricarboxylic acid (TCA) cycle, and oxidative phosphorylation (OXPHOS) pathways, vital for generating ATP and regulating cellular functions [25, 26]. The expression of the regulatory enzymes of glycolysis and glycolysis-derived metabolites is reduced in IBD influencing intestinal microbiota, mucosal barrier function, and immune regulation [27, 28]. In the brain, dysregulation of glycolysis affects behaviour by reducing energy expenditure and mood regulation through the metabolism of tryptophan [29]. Glycolysis determines the efficiency of tryptophan metabolism, which has further implications for regulating gut motility and pain [30]. Dysbiosis of the gut microbiota in IBD causes alterations in metabolic pathways, compromising amino acid synthesis, monosaccharide conversion to short-chain fatty acids, and FAO [31]. Subsequent abnormalities in lipid utilisation and metabolism lead to defective mucous production, altered fatty acid uptake, and overproduction of inflammatory signalling molecules, contributing to IBD pathogenesis [32]. Lipidome homeostasis is also fundamental to neural cell function and defects in lipid regulation are associated with neurological diseases [33].

The TCA cycle generates nicotinamide adenine dinucleotide (NAD) and reduced flavin adenine dinucleotide cofactors critical for OXPHOS, a process that generates substantial ATP. Amino acids and fatty acids likewise contribute to OXPHOS, either entering the TCA cycle directly or supporting nucleotide and fatty acid synthesis [34]. Dysregulation of the TCA cycle may be implicated in the impaired immune response in IBD patients since TCA metabolites can alter the response of both the innate and the adaptive immune systems [28, 34]. Alteration of the respiratory chain complexes and impaired OXPHOS leads to increased mitochondrial reactive oxygen species (ROS) production and oxidative stress, which are hallmarks of mitochondrial dysfunction in IBD [35, 36]. Furthermore, dysregulation of the TCA cycle and OXPHOS has been implicated in various brain diseases, including neurodegenerative disorders, and can lead to energy deficits and neuronal dysfunction [37–39].

Despite accumulating evidence on the effects of metabolic dysfunction in the intestine and brain, IBD-induced damage to metabolic and cellular processes that are essential for nervous system function remain largely unexplored in the context of gut-brain axis disruption. Furthermore, the association between depressive and anxiety-like behaviours and alterations in cellular metabolism within the colon and brain in IBD is yet to be elucidated. In this study, we hypothesise that behavioural changes associated with chronic intestinal inflammation are correlated with disruption of the cellular metabolism within the gut-brain axis. Therefore, this study aims to determine colitis-induced changes in cellular metabolism within the gut-brain axis and to investigate the association between metabolic alterations and behavioural changes in IBD using the Winnie mouse model of spontaneous chronic colitis. The Winnie mouse, which carries a missense mutation in the MUC2 mucin gene, is a well-established model for chronic intestinal inflammation demonstrating symptoms closely related to UC [40–42]. To assess metabolic changes, this study utilised metabolic cages for comprehensive measurements of food and water intake, energy expenditure, spontaneous physical activity, and sleeping behaviour observation [43]. Rodent behavioural tests were used to determine depressive and anxiety-like states, which are interpreted by measuring spontaneous behaviour and selected actions to understand these behaviours in a controlled setting [44, 45]. To elucidate changes in transcriptome of glycolysis, FAO, TCA cycle and OXPHOS pathways as well as cytokine activity in the distal colon and brain, RNA-sequencing (RNA-Seq) of the distal colon and brain tissue samples from Winnie mice with chronic intestinal inflammation and wild-type littermates was performed. Increasing understanding of changes to metabolism and behaviour within the gut-brain axis in chronic intestinal inflammation will enhance knowledge of IBD pathophysiology and offer potential avenues for development of novel future treatments.

Materials and methods

Animals

Winnie (14 weeks old, n = 17) and C57BL/6 (14 weeks old, n = 18) mice from the same breeding colony were obtained from Victoria University Werribee Animal Facility (Melbourne, Australia). All mice had ad libitum access to food and water and were housed in a temperature-controlled environment with a 12 h day/night cycle. Animals were acclimatised for one week prior to experiments, after which all mice were euthanised via an overdose (100 mg/kg) of pentobarbitone (Lethabarb, Virbac Australia). The whole distal colon and brain tissues were excised and used for subsequent experiments. C57BL/6 and Winnie mice were randomly selected for all experiments. All experimental procedures adhered to the Australian National Health and Medical Research Council (NHMRC) guidelines and were approved by the Victoria University Animal Experimentation Ethics Committee (AEC-17-016).

Evaluation of colonic inflammation

To evaluate the level of colonic inflammation in Winnie mice, faecal water content, faecal lipocalin (Lcn)−2, and gross morphological damage in sections of the distal colon were assessed. For quantification of faecal water content, each mouse was placed into a single sterile cage without bedding until defecation occurred. Individual faecal pellets were collected and weighed immediately to determine their wet weight. Samples were then placed in an oven at 60 °C for 24 h and weighed again (dry weight) and the difference between wet and dry weight was calculated as a percentage. Quantification of faecal lipocalin (Lcn)−2 by ELISA was used as a non-invasive assessment of distal colonic inflammation in Winnie mice as previously described [46, 47]. Briefly, faecal samples collected from C57BL/6 and Winnie mice were reconstituted in PBS-0.1% Tween 20 (100 mg/mL) then vortexed (20 min) to form a homogenous suspension. Samples were centrifuged for 10 min at 12,000 rpm and 4 °C. Lcn-2 was estimated in the supernatants using a Mouse Lcn-2 ELISA Kit (NGAL; cat. no. ab119601, Abcam, Melbourne, Australia) according to the manufacturer’s instructions. All samples were repeated in duplicate. Lcn-2 protein (pg/mL) was detected by measuring the absorbance at 450 nm on a Varioskan Flash Multimode Reader using Skanlt software v2.4.3 (Thermo Fisher Scientific, Waltham, Massachusetts, USA). For histology, distal colon tissues from C57BL/6 and Winnie mice were paraffin embedded, sectioned at 5 μm, deparaffinised, cleared, rehydrated in graded ethanol concentrations, and stained with haematoxylin and eosin (H&E) as previously described [40, 46]. Sections were imaged using the MetaSystems Metafer program and VSlide image stitching software on a Zeiss AxioImager microscope (Zeiss, Germany). Histological scoring was completed blindly and including the following parameters: aberrant crypt architecture (score range 0–5), increased crypt length (0–5), crypt abscesses (0–5), epithelial damage (0–5), ulceration (0–5), and leukocyte infiltration (0–5) (average of three areas of 500µm² per section).

In vivo analysis of metabolism

C57BL/6 and Winnie mice were placed into Promethion metabolic high-definition multiplexed respirometry cages (Sable Systems International, Nth Las Vegas, NV, USA) for six Days to determine real-time food and water intake and energy metabolism. After five Days acclimatisation, the following final 24 h data were used for analyses in this study. Flow-through respirometry monitors determined metabolic rates, and sensors evaluated caloric food and water uptake, digestion, and energy efficiency via indirect calorimetry (food intake monitor measured by mass (g), water intake monitor measured by mass (g), FAC-1 Access Control Module (access to food and water), WM-1 Running Wheel Module, BXZ-1 Total Activity Monitor, BW-1 Body Weight Module). Metabolic data acquisition and instrument control were coordinated by MetaScreen, while the raw data were processed by ExpeData (Sable Systems International, North Las Vegas, NV, United States) using an analysis script detailing all aspects of data transformation. The recorded data points were summarised by circadian cycle, from 0700 h to 1900 h (light) and 1900 h to 0700 h (dark). The respiratory quotient (RQ) was analysed using a substance-fixed ratio between the quantities of VO2 consumption and VCO2 emission [48]. Energy expenditure (EE) was measured using the modified Weir formula [49] with data normalised by dividing the EE by the body weight to generate values of EE/gram, EE (kcal/kg/body weight) = EE (cal) in 12 h/1000/body weight. The substrate oxidation rates were determined using the calculations of Ferrannini [50]: EE (kcal/kg/body weight)=(3.94*VO2mL in 12 h dark) ₊ (1.106*VCO2mL in 12 h dark); GOX (g/kg/body weight)=(4.55*VCO2mL in 12 h dark) - (3.21*VO2mL in 12 h dark); LOX (g/kg/body weight)=(1.67*VO2mL in 12 h dark) – (VCO2mL in 12 h dark).

Assessment of depressive and anxiety-like behaviours

Behavioural tests were used to phenotype depressive and anxiety-like behaviours of Winnie when compared to C57BL/6 mice. The light dark test (LDT), open field test (OFT), and elevated plus maze (EPM) were used to measure anxiety-like behaviour, whereas the forced swim test (FST) were employed to determine depressive-like behaviour [51–53]. To reduce stress, the mice were acclimatised for one hour before commencing the behavioural tests which were performed once between 09:00 h and 13:00 h. All test equipment was cleaned with 70% ethanol after each test to prevent a bias based on olfactory cues. The behaviour tests were recorded by a video camera (D-link camera) and measurements for individual tests were analysed using in-house-developed Advanced Move Tracker (AMT) software [54].

RNA extraction, quality control, and high-throughput RNA sequencing of distal colon and brain samples

Total RNA was extracted from the whole distal colon and the middle portion of the brain tissues of randomly selected C57BL/6 and Winnie mice using the RNeasy Lipid Tissue Mini Kit (cat. no. 74804, Qiagen, Hilden, Germany) and QIAzol^® Lysis Reagent (Invitrogen, USA) according to the manufacturer’s instructions. The quality of RNA was assessed on a 2100 Bioanalyzer microfluidics platform using the RNA 6000 Nano Kit (Agilent Technologies, USA) to confirm samples were free from contamination of genomic DNA and 16 S ribosomal RNA from bacteria (RNA integrity number: 8.5–9.5/10). The concentration of each RNA sample was measured by a NanoDrop Spectrophotometer (Denovix, Melbourne, Australia). Purity was indicated with 260/280nm and 260/230nm ratios exceeding 1.8. Samples then underwent mRNA quantification and quality control using Qubit, polyA purification, RNA-Seq library construction, and high-throughput sequencing on the MGITech MGISEQ-2000 System, using a 100-bp single-end read protocol at Micromon Genomics (Monash University, Australia). All samples quantified with Qubit and analysed with Bioanalyzer passed the QNA QC standard (Supplementary Tables S1, S2). The raw FASTQC files were analysed with the RNAsik pipeline version 1.5.4 [55], using the STAR aligner and the Mus musculus reference genome GRCm38 (GenBank accession GCA_000001635.2) [56]. Raw read counts were quantified using the FeatureCounts program [57], with successive analysis for counts per million (CPM) library size normalisation using Degust [58]. Differentially expressed genes (DEGs) of the distal colon and brain were identified via the R package, DEGseq v 1.34.0, applying a false discovery rate (FDR) of < 0.001 with Benjamini–Hochberg correction [59]. Genes exhibiting changes in expression beyond ± 0.585 Log₂fold (1.5-fold) were excluded in further analyses. The visualisation of selected pathways was performed using the R package, Pathview, using default parameters [60]. Input data included significance scores of gene expression values between experimental groups and computed as the Log₂FC×Log10P values represented in the Kyoto Encyclopedia of Genes and Genomes (KEGG) graph. Heat maps of gene expression values for each distal colon and brain sample were generated using the Morpheus web-based tool and presented as Z-score distributions (across samples) of the CPM values. Data was hierarchically clustered by Euclidian distance.

STRING protein-protein association networks

The Search Tool for Retrieval of Interacting Genes (STRING) database enables comprehensive characterisation of user gene lists and functional genomics datasets, allowing the creation and sharing of highly customised and augmented protein-protein association networks [61]. STRING sources biological pathway databases, specifically gene ontology (GO). In this study, upregulated and downregulated genes in distal colon and brain samples from Winnie mice were uploaded into STRING for functional analysis, biological processes, molecular functions and cellular pathways. GO biological process (GOBP) comprises extensive cellular compartments related to multiple molecular activities. GO molecular function (GOMF) describes catalytic activity molecularly occurring within the cell. These annotations aid in deciphering the molecular mechanisms and biological processes linked to the specific experimental conditions being investigated [62].

Homology of genes in Winnie mice and IBD patients

To establish homology of DEGs in glycolysis and gluconeogenesis, FAO, TCA, and OXPHOS metabolism in the Winnie mouse colon to IBD patients, gene expression data on the transcriptome of human IBD patients was gathered from the National Centre for Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) public data source [63, 64]. Previously, high-throughput sequencing was used to create the expression profile of the inflamed intestine in IBD patients and the healthy colon regions in uninflamed controls undergoing resection of non-obstructive colorectal adenocarcinoma [65]. These data are available at https://www.ncbi.nlm.nih.gov/geo/ under the GEO series accession number GSE83687. This dataset includes the data from CD patients (n = 42), UC patients (n = 31), and controls (n = 60).

Statistical analysis

Statistical differences were determined by unpaired Student’s t-test with Welch’s correction using GraphPad Prism v9 (GraphPad Software, San Diego, CA, USA). For analyses, p < 0.05 was considered statistically significant. All data are presented as the mean ± standard error of the mean (SEM).

Results

Assessment of distal colonic inflammation in Winnie mice

Colonic inflammation was assessed by quantifying water content and Lcn-2 in faecal samples and evaluating gross morphology of distal colon sections from Winnie mice compared to C57BL/6 mice (Supplementary Fig. S1). Faecal water content was determined by calculating the difference between wet and dry weight in samples from both groups (n = 11/group). The percent faecal water content was higher in Winnie (77.0 ± 1.0%, p < 0.001) when compared to C57BL/6 mice (51.5 ± 7.0%, Fig. S1A). Levels of Lcn-2, a highly-sensitive non-invasive biomarker of intestinal inflammation [66], were increased in faecal samples from Winnie mice (412.4 ± 12.8pg/mL, p < 0.001, n = 10) when compared to samples from C57BL/6 mice (246.9 ± 40.7pg/mL, n = 6, Fig. S1B). Gross morphological damage was assessed in H&E-stained cross sections of the distal colon from C57BL/6 and Winnie mice (n = 5/group, Fig. S1C-C^I). Sections from C57BL/6 mice showed distinct colonic layers and regular arrangements of goblet cell and crypt structures (Fig. S1C). In contrast, damage to the crypts, loss of goblet cells, and leukocyte infiltration was evident in sections from Winnie mice (Fig. S1C^I). In consistency with these observations, histological scores were elevated in Winnie (16.3 ± 0.4, p < 0.001) compared to C57BL/6 mice (0.77 ± 0.2, n = 5/group, Fig. S1D).

Energy metabolism is reduced in Winnie mice

To determine differences in energy metabolism between C57BL/6 and Winnie mice (n = 16/group), indirect calorimetry was used to measure water and food consumption, as well as assess energy output mechanisms via VO₂, VCO₂, respiratory quotient, EE, GOX, and LOX analyses (Fig. 1). In this study, Winnie mice drank less water (1.9 ± 0.2 g, p < 0.001) and attempted fewer water bouts (31.9 ± 3.1, p < 0.05) than C57BL/6 mice (sum of water: 3.1 ± 0.2 g, water bouts: 40.9 ± 4.0, Fig. 1A-B). Similarly, Winnie mice consumed less food (2.4 ± 0.1 g, p < 0.05) and attempted fewer food bouts (16.6 ± 1.9, p < 0.001) when compared to C57BL/6 mice (sum of food: 3.1 ± 0.2 g, food bouts: 26.4 ± 1.2, Fig. 1C-D). VO₂ and VCO₂ were reduced in Winnie (VO₂: 61.8 ± 2.6mL/kg/min/body weight, VCO₂: 59.4 ± 2.6mL/kg/min/body weight, p < 0.001 for both) in contrast to C57BL/6 mice (VO₂: 89.1 ± 2.2mL/kg/min/body weight, VCO₂: 82.3 ± 2.5mL/kg/min/body weight, Fig. 1E-F), however respiratory quotient did not differ between the groups (C57BL/6: 0.95 ± 0.01, Winnie: 0.96 ± 0.01, Fig. 1G). Reduced energy metabolism was further demonstrated in Winnie mice with decreased EE (221 ± 9.3 kcal/kg/body weight, p < 0.001), GOX (31.3 ± 2.6 g/kg/body weight, p < 0.001), and LOX (29.9 ± 2.4 g/kg/body weight, p < 0.001) when compared to C57BL/6 mice (EE: 309 ± 9.2 kcal/kg/body weight, GOX: 51.6 ± 2.7 g/kg/body weight, LOX: 45.1 ± 1.3 g/kg/body weight, Fig. 1H-J).

Fig. 1.

Decreased energy metabolism in Winnie mice. The Promethion system for indirect calorimetry was used to evaluated energy metabolism in C57BL/6 and Winnie mice (n=14-16/group). (A) Sum ofconsumed water (g). (B) Total number of drinking bouts. (C)Sum ofconsumed food (g). (D) Total number offood bouts.(E)Rate of oxygen consumption (VO2, mL/kg/min/body weight). (F) Rate of carbon dioxide emission (VCO2, mL/kg/min/body weight). (G)Respiratory quotient. (H)Energy expenditure (EE, kcal/kg/body weight).(I)Glucose oxidation (GOX, g/kg/body weight).(J)Lipid oxidation (LOX, g/kg/body weight). Data are expressed as mean±SEM. *p<0.05,***p<0.001 when compared to C57BL/6 mice

Since EE, GOX, and LOX were reduced in Winnie mice, we conducted further investigations to identify DEGs associated with glycolysis and gluconeogenesis, FAO, TCA cycle, and OXPHOS in the distal colons from Winnie compared to C57BL/6 mice.

Altered expression of genes related to glycolysis and gluconeogenesis, FAO, TCA cycle, and OXPHOS in the Winnie mouse distal colon

To identify DEGs, high-throughput RNA-Seq was performed on mRNA extracted from the distal colons of Winnie (n = 6) and C57BL/6 (n = 7) mice. KEGG pathway-based data integration and visualisation using Pathview, as well as hierarchical clustering analysis of DEGs using Morpheus, revealed changes to the expression of genes involved in the glycolysis and gluconeogenesis, FAO, TCA cycle, and OXPHOS metabolic pathways in the Winnie mouse colon when compared to colons from C57BL/6 mice (Figs. 2 and 3). Functional analysis was used to explore the expression patterns of glycolysis, FAO, TCA cycle and OXPHOS transcriptomes in Winnie mice colons, identifying significant changes in the metabolic processes as annotated by the STRING gene ontology database [61]. Furthermore, homology in the expression of target genes involved in these metabolic pathways was demonstrated between Winnie mice and patients with IBD relative to their respective uninflamed control groups.

Fig. 2.

DEGs associated with glycolysis and gluconeogenesis and FAO in the Winnie distal colon. KEGG pathways showing upregulated and downregulated genes involved in (A)glycolysis and gluconeogenesis and (B)FAO in the distal colon transcriptome of Winnie (n=6) compared to C57BL/6 (n=7) mice. Data are presented as significance scores of gene expression values between Winnie and C57BL/6 mice. Green (dotted line circle) indicates genes that are downregulated (glycolysis and gluconeogenesis: EC:2.7.1.11, Pfk9; EC:5.1.3.3, galM; EC:3.1.3.9, G6pc; EC:4.2.1.11, Eno; EC:1.1.1.27, Ldh; EC:1.1.1.1, Adh6; EC:1.2.1.3, Aldh2; EC:4.1.1.32, Pck, FAO: EC:CPT1, Cpt1; EC:1.3.8.7, Acadm; EC:1.3.8.9, Acadvl; EC:1.3.8.5, Acadsb; EC:1.3.8.6, Gcdh; EC:2.3.1.16, Hadhb; EC:5.3.3.8, Eci; EC:1.1.1.1, Aldh2; EC:1.2.1.3, Adh; EC:1.14.14.80, Cyp4a10. Red (solid line circle) indicates genes that are upregulated (glycolysis and gluconeogenesis: EC:2.7.1.1, Hk2; EC:2.7.1.147, Adpgk; EC:1.2.1.12, Gapdh; EC:2.7.2.3 Pgk1; EC:5.4.2.2, Pgm; EC:3.1.3.80, Minpp1; EC:5.4.2.1, Pgam; EC:5.4.2.4, Bpgm; EC:2.7.1.40, Pkm; EC:1.1.1.2, Akr1a1; EC:1.2.1.5, Aldh3b2; FAO: EC:6.2.1.3, Acsl; EC:1.3.8.8, Acadl; EC:2.3.1.9, Acat1). Heat map representation of upregulated genes (red), downregulated genes (blue), and unchanged genes (white) associated with (A^I) glycolysis and gluconeogenesis and(B^I) FAO in distal colons from Winnie (n=6) and C57BL/6 (n=7) mice determined by RNA-Seq. The homology in the expression of target genes associated with (A^II) glycolysis and gluconeogenesis and (B^II) FAO metabolism between Winnie mice and IBD patients compared to their respective uninflamed controls

Fig. 3.

DEGs associated with TCA cycle and OXPHOS in the Winnie distal colon. KEGG pathways showing upregulated and downregulated genes involved in (C) TCA cycle and (B)OXPHOS in the distal colon transcriptome of Winnie (n=6) compared to C57BL/6 (n=7) mice. Data are presented as significance scores of gene expression values between Winnie and C57BL/6 mice. Green (dotted line circle) indicates genes that are downregulated (TCA: EC:4.1.1.32, Pepck; EC:1.2.4.2, Ogdh; EC:6.2.1.5, Suclg2 andEC:6.2.1.4, Sucla2; OXPHOS: EC:7.1.1.2 (NADH dehydrogenase (ubiquinone)), Ndufa1, Ndufa2, Ndufa4, plus Nduf1b subcomplex, Ndufb2, Ndufb5, QCR6, Cox6b, Cox7A, Cox8, Cox17, EC: 7.1.1.9, cytochrome c oxidase cbb3-type subunit I, ccoN, EC:7.1.2.2 ATP synthase, F-type H+-transporting ATPase subunit a, ATPF0A, Atp5e (epsilon), AtpeV0A, Atp6N, AtpeV0E, Atp6H;and EC3.6.1.1, Ppa1). Red (solid line circle) indicates genes that are upregulated (TCA: EC:6.4.1.1, Pcx; EC:4.2.1.3, Aco;EC:1.1.1.42,Idh;EC 1.1.1.37, Mdh; OXPHOS: NADH dehydrogenase Nduf Fe-S protein/flavoprotein complex, mitochondria, Ndufs8, Ndufa10; Cytochrome c oxidase, Cox4, Cox15; F-type ATPAse, ATP5A1 (alpha), ATP5G (c)I; and V-type ATPase, Atp6B. Heat map representation of upregulated genes (red), downregulated genes (blue), and unchanged genes (white) associated with (A^I) TCA cycle and (B^I) OXPHOS in distal colons from Winnie (n=6) and C57BL/6 (n=7) mice determined by RNA-Seq. The homology in the expression of target genes associated with (A^II) TCA and (B^II) OXPHOS metabolism between Winnie mice and IBD patients compared to their respective uninflamed controls

Several genes associated with glycolysis and gluconeogenesis were differentially expressed in distal colons from Winnie mice when compared to C57BL/6 mice (Fig. 2A-A^I). STRING analysis revealed downregulation of glucose 6-phosphate (G6pc), phosphoenolpyruvate carboxykinase 1 (Pck1), D-glucose 1-epimerase (galM), enolase (Eno1), and 6–phosphofructokinase (Pfk9), which regulate glucose metabolism, including its breakdown and synthesis, suggesting impaired glucose production pathways and disruptions in energy homeostasis in the Winnie mouse colon. Additionally, reduced expression of alcohol dehydrogenase 1 (Adh1), L-lactate dehydrogenase (Ldh), and aldehyde dehydrogenase (NAD⁺) 2 (Aldh2), involved in aldehyde and ethanol metabolism, and Ldh, Pck1, Eno1, and Pfk9, linked to pyruvate metabolism, indicate compromised energy metabolism and detoxification processes in colons from Winnie compared to colons from C57BL/6 mice (Supplementary Table S3) [67].

In the Winnie mouse colon, upregulated STRING GOBP terms included monosaccharide metabolic process, carbohydrate metabolic process, monocarboxylic acid metabolic process, carbohydrate derivative metabolic process, organophosphate metabolic process, phosphate containing compound metabolic process, monosaccharide biosynthesis process. Upregulated DEGs GOMF terms catalytic activity (Supplementary Table S4). In particular, GOBP upregulation of hexokinase 2 (Hk2), glycerate phosphomutase 1(Pgam1), pyruvate kinase (Pkm), bisphosphoglycerate-phospho-glycerate mutase (Bpgm), phosphoglycerate kinase (Pgk1), glyceraldehyde 3-phosphate dehydrogenase (Gapdh), and ADP: D-glucose 6-phospho-transferase (Adpgk), involved in organophosphate, carbohydrate derivative, and phosphate-containing compound metabolism, reflect extensive metabolic reprogramming during colitis, impacting energy production, as well as the synthesis and breakdown of biomolecules. These genes also contribute to small molecule metabolism and organic substance catabolism, crucial for maintaining cellular balance. Additionally, increased expression of Pgam1, aldo-keto reductase family 1 member A1 (Akr1a1), Pgk1, and Gapdh, implicated in monosaccharide biosynthesis, and Pkm, Pgk1, and Gapdh, specifically involved in canonical glycolysis, suggests a compensatory mechanism to produce glucose from non-carbohydrate sources during glucose scarcity. An overall increase in enzymatic catalytic functions is indicated by STRING analysis within the inflamed colons from Winnie mice via upregulation of Hk2, Pgam1, multiple inositol-polyphosphate phosphatase 1 (Minpp1), Akr1a1, Pkm, Bpgm, Pgk1, Gapdh, aldehyde dehydrogenase (NAD(P)⁺) (Aldh3b2), and Adpgk. Specifically, Pgam1 and Bpgm are involved in bisphosphoglycerate mutase and phosphoglycerate mutase activities,

crucial for regulating 2,3-bisphosphoglycerate levels affecting oxygen release from haemoglobin. These transcriptome also participate in phosphoglycerate mutase activity, essential for glycolysis by converting 3-phosphoglycerate to 2-phosphoglycerate [68] (Supplementary Table S4). Upregulated DEGs GOMF terms included catalytic activity (Table S4). Homology investigations showed a 64% and 52% concordance in the expression of genes associated with glycolysis and gluconeogenesis in Winnie mice when compared to UC and CD patients, respectively (Fig. 2A^II).

FAO was downregulated in the Winnie mouse distal colon when compared to colons from C57BL/6 mice (Fig. 2B-B^I). STRING GOBP analysis (Supplementary Table S5) revealed that downregulation of Adh, Aldh2, and long-chain fatty acid omega-monooxygenase (Cyp4a10) indicate impaired catabolism of organic hydroxy compounds, ethanol metabolism, and detoxification processes. The combined downregulation of carnitine O-palmitoyl-transferase (Cpt1) and medium-chain acyl-CoA dehydrogenase (Acadm) impacts carnitine metabolism, essential for transporting fatty acids into mitochondria for beta-oxidation. Additionally, downregulation of Cyp4a10 and Acadm indicates reduced metabolism of medium-chain fatty acids, which are crucial energy sources [69]. Specific transcripts involved in the catabolism monocarboxylic acids, as well as lipid catabolism were downregulated in the Winnie mouse colon, including glutaryl-CoA dehydrogenase (Gcdh), electron-transfer flavoprotein 2,3-oxidoreductase (Acadsb), Cpt1a, Cyp4a10, Acadm, palmitoyl-CoA: electron-transfer flavoprotein 2,3-oxidoreductase (Acadvl), hydroxyacyl-CoA dehydrogenase trifunctional multienzyme complex subunit β (Hadhb), and enoyl-CoA delta isomerase 1 (Eci1) indicating disruptions to energy equilibrium and utilisation within cells [70]. STRING enrichment pathway functional analysis (Table S6) showed that upregulated transcripts of acyl-CoA synthetase (Acsl), palmitoyl-CoA dehydrogenase (Acadl), and acetyl-CoA acetyltransferase 1 (Acat1), which play pivotal roles in ketone body metabolism, suggest an enhanced reliance on ketone body metabolism in the Winnie mouse colon to generate and utilise ketone bodies as alternative energy sources during periods of low glucose availability. Furthermore, increased breakdown of branched-chain amino acids for energy production is indicated by upregulation of Acadl and Acat1 (Supplementary Table S6). Winnie mice showed a 77% concordance in the expression of FAO associated genes with UC patients and a 69% similarity with patients diagnosed with CD (Fig. 2B^II). When compared to control mice distal colons, transcriptome analysis of distal colon samples from Winnie mice revealed differential expression of genes associated with the TCA cycle, vital for ATP production in the mitochondria (Fig. 3A-A^I). STRING predicted protein-protein interactions (Supplementary Table S7) showed that downregulated transcriptome GOBP of oxoglutarate dehydrogenase (Ogdh), Pck1, Suclg1 and Sucla2, which are involved in metabolism of dicarboxylic acids, isocitrate metabolic processes, oxaloacetate metabolic process, carbohydrates, and monocarboxylic acids, such as lactate and pyruvate, indicates disruption to the TCA cycle’s progression, potentially favouring less efficient energy production pathways [34]. Upregulated Suclg1, Ogdh, Sucla2 involved succinyl-coA metabolic processes. In addition, Ogdh and Pck1 are implicated in pyruvate metabolic process, suggesting possible disruptions in nucleotide synthesis and overall energy equilibrium [28]. Functional analysis GOBP (Supplementary Table S8) showed that genes involved in in the metabolism of dicarboxylic acids (pyruvate carboxylase (Pcx), Idh2, and malate dehydrogenase b (Mdh1b)), such as oxaloacetate and malate, were upregulated in the Winnie mouse colon.

Pcx and Mdh1b are particularly important in the metabolism of oxaloacetate, crucial for the formation of citrate from acetyl-CoA, marking the initial step of the TCA cycle, and contributing to gluconeogenesis. An increased demand for energy production in the Winnie mouse colon is further indicated by upregulation of Pcx, aconitate hydratase (Aco), Idh2, and Mdh1bassociated with carboxylic acid metabolic process, NADH, and nicotinamide adenine dinucleotide phosphate (NADP) metabolism [67, 71]. Homology investigations revealed 42% concordance in the expression of genes associated with the TCA cycle between Winnie mice and IBD patients (Fig. 3A^II).

OXPHOS transcriptomes are downregulated in the Winnie mouse distal colon when compared to distal colons from C57BL/6 mice (Fig. 3B-B^I). STRING predicted protein-protein interactions GOBP (Supplementary Table S9) demonstrated that decreased expression of NADH dehydrogenase (ubiquinone) 1 alpha subcomplex 1 (Ndufa1), Ndufa2, Ndufb2, Ndufb5, and ATP synthase subunit epsilon (Atp5e) suggest impaired electron transport and ATP synthesis in the Winnie mouse distal colon. This is consistent with downregulation of genes involved in assembly of mitochondrial respiratory chain complexes, including Ndufa1, Ndufa2, Ndufb2, Ndufb5, cytochrome c oxidase (Cox7a1), and Cox17 [72]. According to GOMF analysis (Table S9), reduced expression of Ndufa4, Cox17, Cox8a and Cos7a1 in the Winnie mouse distal colon may indicate diminished regulation of cytochrome-c oxidase activity, a key enzyme in the ETC that facilitates the final step of electron transfer to oxygen [73]. Disruption to efficient electron transport and ATP production is further indicated by STRING network analysis via GOBP and GOMF (Supplementary Table S10) upregulation of Atp5a1, Atp5e, Ndufa10, Atp6v1a, Cox15, Cox4i1, and Ndufs8 in the distal colons from Winnie mice. Increased expression of Cox15, Cox15, and Cox4i1 suggest impaired generation of the proton gradient used to synthesise ATP, while upregulation of Cox4i1 and Ndufs8 potentially indicate a compensatory effort to enhance the production of ATP through OXPHOS [73, 74]. Mitochondrial activity is also implicated by upregulated Ndufa10, Cox4i1, Ndufs8 and Cox4. Winnie mice exhibit a 56% similarity in the expression of OXPHOS-associated genes to UC patients and 52% similarity to CD patients (Fig. 3B^II).

Reduced spontaneous physical activity and increased sleep behaviour in Winnie mice

Spontaneous physical activity, including wheel distance travelled, wheel speed, percentage of time wheel running, and percentage of time spent in ambulatory locomotion outside of the wheel were measured in Winnie and C57/BL6 mice (n = 16/group, Fig. 4). Sleep behaviour, which is either a characteristic or a precursor of anxiety and depression [75], was concurrently assessed in the same experiment through measurement of sleep duration, percentage of time sleeping, percentage of still time, and percentage of quiet time without sleep. All parameters were measured during a day (light) and night (dark) cycle (12 h per cycle).

Fig. 4.

Reduced spontaneous physical activity and increased sleep behaviour inWinnie mice. Spontaneous physical activity and sleep behaviour measured in C57BL/6 and Winnie mice (n=16/group) during the dark cycle (12h). (A) Wheel distance travelled (m). (B) Wheel speed (m/s). (C) Percentage of wheel running time (wheel running time %). (D) Percentage of time spent in ambulatory locomotion outside of the wheel (ambulatory locomotion time %). (E)Duration of sleep (h). (F) Percentageof sleeping time (sleep time %). (G) Percentage of still time (still time %). (H)Percentage of quiet time without sleeping (quiet time without sleeping %). ***p<0.001 when compared to C57BL/6 mice

During the dark cycle, Winnie mice travelled less wheel distance (2932 ± 617 m, p < 0.001) at a slower wheel speed (0.3 ± 0.03 m/s, p < 0.001) when compared to C57BL/6 mice (distance: 8512.6 ± 987 m, speed: 0.48 ± 0.04 m/s, Fig. 4A-B). The percentage of wheel running time in the dark cycle was lower in Winnie mice (22.08 ± 3.7%, p < 0.001) than in C57BL/6 mice (50.9 ± 2.6%, Fig. 4 C), however there was no difference in the percentage of time spent in ambulatory locomotion outside of the wheel between groups (C57BL/6: 21.49 ± 5.2%, Winnie: 15.82 ± 1.6%, Fig. 4D). There were no differences in spontaneous physical activity between C57BL/6 and Winnie mice during the light cycle (data not shown).

The duration and percentage of sleep time in the dark cycle was increased in Winnie mice (6.1 ± 0.5 h, 50.5 ± 3.9%, p < 0.001 for both) when compared to C57BL/6 mice (3.5 ± 0.4 h, 28.9 ± 3.2%, Fig. 4E-F). The percentage of dark cycle still time was higher in Winnie mice (53.3 ± 4.0%, p < 0.001) than in C57BL/6 mice (35.4 ± 3.7%, Fig. 4G), while the percentage of quiet time without sleeping was lower (C57BL/6: 6.5 ± 0.6%, Winnie: 2.2 ± 0.5%, p < 0.001, Fig. 4G). There were no differences in sleep behaviour between Winnie and C57BL/6 mice during the light cycle (data not shown).

Increased expression of depressive and anxiety-like behaviours in Winnie mice

The LDT, OFT, EPM test, and FST were used to measure depressive and anxiety-like behaviour in C57BL/6 and Winnie mice (n = 16/group, Fig. 5). In the LDT, increased exploratory behaviour is associated with increased time spent in the white box, whereas anxious mice prefer the black box and make fewer transitions between the white and black box [53]. Winnie mice spent less time (226.7 ± 15.3s, p < 0.01) and made fewer entries (18.6 ± 1.3, p < 0.01) into the white box when compared to C57BL/6 mice (time: 307.2 ± 16.1s, entries: 23.8 ± 1.3, Fig. 5A-B). There was no difference in the distance travelled within the white box between groups (Winnie: 1249.5 ± 94.2 cm, C57BL/6: 1169.5 ± 67.8 cm). Correspondingly, Winnie mice spent more time in the black box (362.1 ± 16.6s, p < 0.01) than C57BL/6 mice (299.7 ± 12.9s, Fig. 5 C). The number of pokes from the black box was higher in Winnie mice (20.2 ± 2.1, p < 0.001) than C57BL/6 mice (7.4 ± 1.0, Fig. 5D). In the OFT, anxiety is measured by overall exploratory and locomotor activity, including the frequency of central square entries, the corners, peripheries of the field, and unprotected areas, while depression is indicated by reductions in grooming behaviours [76]. Winnie mice spent more time in the centre of the field (15.5 ± 2.0s, p < 0.05) and less time in the corners (169.2 ± 7.7s, p < 0.01) when compared to C57BL/6 mice (centre: 9.5 ± 1.9s, corners: 203.9 ± 8.7s, Fig. 5E-F). There was no difference in the number of times moving to the centre position or corners between groups (centre - C57BL/6: 13.6 ± 7.1, Winnie: 13.5 ± 4.2; corners - C57BL/6: 56.7 ± 3.9, Winnie: 47.4 ± 2.3). In Winnie mice, grooming time (48.8 ± 8.1s, p < 0.001) and number of groomings (123.1 ± 13.4, p < 0.001) was lower when compared to C57BL/6 mice (time: 98.9 ± 6.9s, groomings: 197.0 ± 11.2, Fig. 5G-H).

Fig. 5.

Winnie mice exhibit increased depressive and anxiety-like behaviours. The light-dark test (LDT), open field test (OFT), elevated plus maze (EPM) test, and forced swim test (FST) were used to measure depressive and anxiety-like behaviour in C57BL/6 and Winnie mice (n=16/group).Winnie mice exhibited higher levels of depressive and anxiety-like behaviours when compared to C57BL/6 mice as measured by : (A)time spent in the white box (LDT), (B)number of entries into the white box (LDT), (C)time spent in the black box, (D)number of pokes from the black box (LDT): (E)time spent in the centre, (F)time spent in the corners, (G)grooming time, (H)number of groomings: (I)time in the open arm: (J)swimming mobility time, (K)swimming immobility time. *p<0.05, **p<0.01,***p<0.001 when compared to C57BL/6 mice

The EPM determines anxiety-like behaviour based on the natural aversion of mice to exploring open and elevated areas and its instinct to explore novel environments [77]. Throughout the EPM test, Winnie mice spent less time in the open arm (21.1 ± 6.1s, p < 0.01) when compared to C57BL/6 mice (58.0 ± 8.8s, Fig. 5I). There were no differences between groups in the time spent in the closed arm (Winnie: 159.0 ± 11.2s, C57BL/6: 156.3 ± 7.9s) or in the centre (Winnie: 117.0 ± 10.3s, C57BL/6: 104.3 ± 8.4s). In addition, the number of entries into the closed and open arms was similar in Winnie (closed: 11.2 ± 0.8, open: 5.2 ± 1.2) and C57BL/6 (closed: 12.4 ± 0.9, open: 6.1 ± 0.6) mice.

The FST uses immobility as a marker of despair and depressive-like behaviour in rodents [52]. Winnie mice demonstrated less swimming mobility (57.4 ± 4.1s, p < 0.01) and higher swimming immobility (182.5 ± 4.1s, p < 0.01) than C57BL/6 mice (mobility: 103.2 ± 12.2s, immobility: 151.4 ± 9.9s, Fig. 5J-K).

Downregulation of glycolysis and gluconeogenesis, FAO, TCA cycle, and OXPHOS in Winnie mice brains

Downregulation of glycolysis and gluconeogenesis, FAO, TCA cycle, and OXPHOS in Winnie mice brains

The mRNA transcriptome in the brains of Winnie mice were compared to C57BL/6 littermates (n = 10/group) to determine any changes in gene expressions associated with glycolysis and gluconeogenesis, FAO, TCA cycle, and OXPHOS metabolism. Pathway-based data integration and visualisation using Pathview plots rendered from the KEGG database revealed differential expression of several genes involved in these pathways in the Winnie mouse brain (Figs. 6 and 7). STRING analysis identified significance of these altered brain transcriptomes in metabolic processes.

Fig. 6.

DEGs related to glycolysis and gluconeogenesis and FAO in the Winnie brain. Pathview plots for the (A)glycolysis and gluconeogenesis and (B)FAO pathways rendered from the KEGG database. Data are presented as significance scores for analysis of gene expressions associated with these metabolic pathways between the brains of Winnie mice and C57BL/6 littermates (n=10/group). Green (dotted line circle) represents downregulated genes (glycolysis and gluconeogenesis: EC:5.3.1.1, Tpi1; EC:1.2.1.12, Pgk1; EC:5.4.2.4, Bpgm; EC:5.4.2.11, Pgam; EC:1.2.4.1 Pdha1; EC:1.1.1.27, Ldha; EC:1.2.1.3, Aldh3a1; EC:1.2.1.5Aldh3b1; EC:1.1.1.1, Adh 1/7; FAO: EC:1.1.1.35,Hadh; EC:2.3.1.16, Acaa2 or Hadhb; EC:1.1.1.1, Adh1; EC:1.2.1.3; Aldh2). Red (solid line circle) represents the upregulated genes (glycolysis and gluconeogenesis: EC:2.7.1.1, Hk2, FAO: Cpt1; EC:6.2.1.3, Acsl; and EC:1.14.14.80, Cyp4A). Heat map representation of upregulated genes (red), downregulated genes (blue), and unchanged genes (white) associated with (A^I) glycolysis and gluconeogenesis and (B^I) FAO in distal colons from Winnie (n=10) and C57BL/6 (n=10) mice determined by RNA-Seq

Fig. 7.

DEGs related to TCA cycle and OXPHOS in the Winnie brain. Pathview plots for the (A)TCA cycle and (B)OXPHOS pathways rendered from the KEGG database. Data presented as significance scores for analysis of gene expressions associated with these metabolic pathways between the brains of Winnie mice and C57BL/6 littermates (n=10/group). Green (dotted line circle) represents downregulated genes for TCA: EC:1.2.1.4,Pdha, EC:1.3.5.1, Sdha, EC:6.2.1.4 and EC:6.2.1.5,Sucla2; OXPHOS: EC7.1.1.2 NADH ubiquinone oxidoreductase supernumerary subunits(Nduf),Ndufs4, Ndufs5, Ndufv3, Ndufa1, Ndufa2, Ndufa5, Ndufa6, Ndufa7,Ndufab1, Ndufab12, Ndufab13, Ndufb2, Ndufb3, Ndufb6, Ndufb7, Ndufab1, Ndufa12, Ndufa13, Ndufb2, Ndufb3, Ndufb6, Ndufb7, Ndufb11, Ndufc1, Ndufc2; EC1.3.5.1, Sdhc; EC7.1.1.8, fbcH, Cyt1, QCR6, QCR7, QCR8, QCR9, QCR10; EC7.1.1.9,Cox4, Cox5a, Cox5b, Cox6a, Cox6b, Cox7a, Cox7b, Cox8; EC7.1.2.2, Atpfoa; EC7.2.2.19 Atp4a, Atpef1d, Atp5f1e, Atp50, Atp5mc1, Atp5me, Atp5mf,Atp5mg; and EC 3.6.1.1, Ppa1. Heat map representation of upregulated genes (red), downregulated genes (blue), and unchanged genes (white) associated with (A^I) TCA cycle and (B^I) OXPHOS in distal colons from Winnie (n=10) and C57BL/6 (n=10) mice determined by RNA-Seq

The glycolysis and gluconeogenesis transcriptome were downregulated in Winnie mice brains demonstrated by reduced expression of triose phosphate isomerase (Tpi), Pgk1, Bpgm, Pgam, pyruvate dehydrogenase E1 subunit α 1 (Pdha1), lactose dehydrogenase (Ldha), aldehyde dehydrogenase 3 [NAD(P)+] (Aldh3a1), Aldh3b1, and Adh 1/7. Hk2 was the only gene related to glycolysis and gluconeogenesis that was upregulated in Winnie mice brains in this study (Fig. 6A-AI). STRING predicted GOBP showed that glycolysis downregulation may have affected pyruvate metabolic process, cellular aldehyde metabolic process, monocarboxylic acid metabolic process, carbohydrate catabolic process, NAD metabolic process, NADH metabolic process, purine ribonucleotide metabolic activity, aldehyde biosynthetic process, and ethanol catabolic process (Supplementary Table S11). STRING predicted protein-protein interactions demonstrated that downregulated Aldh3a1, Pdha1, Aldh3b1, Pgk1, and Ldha affects molecular function of oxidoreductase activity, whereas Aldh3a1 and Aldh3b1 directly impact aldehyde dehydrogenase (NAD+) activity. In addition, molecular function of catalytic activity in the Winnie mouse brain is affected by Adh5, Aldh3a1, Pdha1, Bpgm, Aldh3b1, Pgk1, Tpi1, and Ldha (Supplementary Table S11a). Only one gene was upregulated in glycolysis, such as Hk2 and affecting carbohydrate catabolic process, glycolytic process, glucose 6-phosphate metabolic process, NAD metabolic process, NADH metabolic process and gluconeogenesis (Supplementary Table S12). In addition, upregulated of Hk2 GOMF correlated to carbohydrate kinase activity, monosaccharide and carbohydrate binding (Supplementary Table S12a).

Differential expression of genes associated with FAO were identified in the Winnie mouse brain. Overall, FAO transcriptome was downregulated by reduced expression of hydroxyacyl-CoA dehydrogenase (Hadh), acetyl-CoA C-acyltransferase Acaa2 (Hadhb), Adh1 and Aldh2 (Fig. 6B-B^I). Concurrently, three genes related to FAO were upregulated in Winnie mice brains, including Acsl, Cpt1, and cytochrome P450, family 1, subfamily a, polypeptide 1 (Cyp4A). STRING downregulated GOBP affected acetaldehyde metabolic process, behavioural response to ethanol, and ethanol catabolic process, whereas Adh1, Hadh, Aldh2 affects NAD binding (Supplementary Table S13a), whereas GOMF correlated to NAD binding (Table S13a). Upregulated FAO GOBP directly affected long chain fatty acid metabolic process (Supplementary Table S14).

According to the RNA-seq data, the TCA cycle transcriptome was downregulated in brains from Winnie mice demonstrated by reduced expression of pyruvate dehydrogenase complex enzyme E1 subunit alpha (Pdha1), succinate-Coenzyme A ligase GDP-forming subunit beta (Suclg2), succinate-Coenzyme A ligase subunit with specificity for ADP (Sucla2), and succinate dehydrogenase subunit A (Sdha) when compared to brains from C57BL/6 mice (Fig. 7A-A^I). No upregulated gene expressions were revealed. Downregulated GOBP showed effects on purine ribonucleotide metabolic process, acyl-CoA metabolic process, and succinate metabolic process (Supplementary Table S15). GOMF downregulated revealed succinate-CoA ligase activity and oxidoreductase activity (Supplementary Table S15a).

The OXPHOS transcriptome was downregulated in Winnie mice brains when compared to brains from C57BL/6 mice (Fig. 7B-B^I). Notably, there was downregulated expression of subunits in Complex (C) I, including various NADH dehydrogenase (ubiquinone) Fe-S protein/flavoprotein (Ndufb4b), as well as ubiquinone cytochrome b560 subunit (SDHC) and the membrane anchor subunit (SDHD) in CII, cytochrome b/c1 subunit (Cyt1), and ubiquinol-cytochrome c reductase subunits 6 through 10 in CIII, cytochrome c oxidase (COX) subunits in CIV, and ATPase H + transporting lysosomal V1 subunit H (Atp6v1h), ATPase, H⁺/K⁺ exchanging gastric alpha polypeptide (Atp4a), and pyrophosphatase (inorganic) 1 (Ppa1) in CV. STRING downregulated GOBP (Supplementary Table S16) affected mitochondrial activity including proton motive force-driven ATP synthesis, carbohydrate derivative biosynthetic process, mitochondrial respiratory chain complex I assembly, electron transport chain, mitochondrion organization (Supplementary Table S16). Downregulated GOMF included transmembrane transporter activity, ATP synthase activity, P-type proton-exporting transporter activity, and cytochrome-c oxidase activity (Supplementary Table S16a).

Upregulation of inflammatory cytokines in Winnie mice brains

Cytokines are involved in the pathophysiology of depression and anxiety, and can be upregulated in response to inflammation [78]. High-throughput RNA-sequencing of mRNA was performed to determine cytokine transcripts in C57BL/6 and Winnie mouse brains (n = 10/group, Fig. 8). Gene expression data was mapped onto the Pathview plot for the IBD pathway from the KEGG database. Analysis demonstrated inflammation in brains from Winnie mice by downregulated expression of interleukin (IL)−22 (Il22), IL12a (Il2a), and IL12 receptor ß2 (Il2rb2) and upregulated expression of transforming growth factor ß1 (Tgfb1), IL2 receptor gamma (IL2rg), IL5 (Il5), mothers against decapentaplegic homolog 2 (Smad2), IL21 (Il21), IL23α (Il23a), IL12 receptor ß1 (Il2rb1), toll-like receptor 5 (Tlr5), toll-like receptor 2 (Tlr2) and tumour necrosis factor (Tnf) when compared to C57BL/6 mice brains. Downregulated GOBP correlated to positive regulation of natural killer cell proliferation, regulation of interleukin-17 production, regulation of defense response, regulation of tyrosine phosphorylation of STAT protein, positive regulation of receptor signaling pathway via JAK-STAT, and positive regulation of interferon-gamma production (Supplementary Table S17), whereas GOMF downregulation highlighted interleukin-12 receptor binding (Supplementary Table S17a). Upregulated GOBP enlisted regulation of interleukin-17 production, Positive regulation of cell differentiation, positive regulation of leukocyte differentiation and activity, positive regulation of immune effector process, positive regulation of lymphocyte mediated immunity, regulation of phagocytosis and as listed in Supplementary Table S18. Upregulated GOMF enlisted cytokine receptor binding, signaling receptor binding, and type I transforming growth factor beta receptor binding (Supplementary Table S18a).

Fig. 8.

Inflammatory cytokine expression in Winnie mice brains. High-throughput RNA-Seq of mRNA was performed to determine inflammatory cytokine transcripts in brains from Winnie when compared to C57BL/6 mice (n=10/group). (A)Pathview plot for the IBD pathway from the KEGG database. Data presented as significance scores for cytokine gene expression analysis between the brains fromWinnie mice and C57BL/6 littermates. Green (dotted line circle) represents downregulated gene expression (Il22, Il12a, Il12rb2). Red (solid line circle) represents upregulated gene expression (Tgfb1,Smad2, Il2rγ, Il5, Il21, Il23a, Il12rb1, Tlr5, Tlr2, Tnf). (B)Heat map representation of inflammatory cytokine gene expression in brains from C57BL/6 and Winnie mice. Red indicates upregulated genes, blue signifies downregulated genes, and white denotes unchanged genes

Discussion

In this study, we investigated the association of depressive and anxiety-like behavioural changes, energy metabolism, and variations of cellular metabolism transcriptome within the gut-brain axis in Winnie mice, an experimental model of spontaneous chronic colitis highly representative of UC. Our results demonstrated reduced energy metabolism, decreased spontaneous activity, and increased sleep were associated with affective disorder behaviours in chronic intestinal inflammation. Furthermore, our study indicates that dysregulation of cellular metabolism along with increased inflammatory cytokine activity within the gut-brain axis may play an underlying role in the connection between IBD and psychological disorders.

Chronic inflammation in the distal colon of Winnie mice was confirmed by increased faecal water content and Lcn-2 protein levels, as well as distinct gross morphological damage to the distal colon, consistent with previous studies [40, 46, 47]. Promethion metabolic studies demonstrated that Winnie mice undertook less food and water hopper interactions and reduced food and water intake when compared to C57BL/6 mice. In consistency with our findings, food avoidance behaviour is reported in patients with IBD, whereby a reduced oral intake is considered a main determinant of malnutrition [79, 80]. Symptoms such nausea, vomiting, abdominal pain, and diarrhea are associated with loss of appetite and ensuing inadequate food/fluid intake in individuals with active IBD [81]. Additionally, energy metabolism analyses revealed reduced VO2 consumption and VCO2 emission, thus diminished EE, GOX and LOX in Winnie mice in this study. Similar to studies in IBD patients, our findings of reduced energy metabolism in Winnie mice may reflect an impaired nutritional status, altered metabolic processes, and diminished physical capacity [82, 83].

We subsequently investigated whether reduced energy metabolism in Winnie mice affected spontaneous physical activity. Our results demonstrated decreased running, walking, and wheel activity in Winnie compared to C57BL/6 mice. In support of our findings, a reduction in physical activity has been reported to be associated with neuroinflammation and serves as a sensitive indicator of disturbed wellbeing in dextran sulfate sodium (DSS)-induced colitis [24, 84]. It is plausible to consider that the inflammatory reaction incited by colitis may instigate changes in neurotransmitter systems and hormonal disparities, potentially influencing motivation and exhibition of sickness behaviour, an adaptive response that enhances recovery by conserving energy to combat inflammation [12, 85–87]. Furthermore, although hyperalgesia was not measured in this study, discomfort and pain associated with chronic intestinal inflammation may discourage Winnie mice from participating in activities that exacerbate symptoms [85, 86].

During the Promethion metabolic study, sleep patterns in Winnie mice were monitored. The results demonstrated that diurnal sleep was increased in Winnie mice without variation in nocturnal sleep patterns. Although these findings contrast with sleep disturbances commonly reported from IBD patients [88], it is proposed that Winnie mice may experience hypersomnia or fatigue [89, 90]. In human psychiatric sleep studies, a bidirectional relationship between hypersomnia, atypical depression, mood reactivity, leaden paralysis, rejection sensitivity and reduced resilience to stress has been demonstrated [89, 91, 92]. In Winnie mice, excessive sleep propensity during typical waking periods may imitate atypical depression symptoms in IBD and depressed patients [88], highlighting the complexity of sleep regulation and its psychopathological associations.

This study then investigated depressive and anxiety-like behaviours in Winnie mice to enhance understanding the neuropsychological impact of gut inflammation. Behavioural analyses delineated notable shifts in Winnie mice behaviours, encompassing a spectrum from anxiety to depressive-like symptoms. During the LDT, Winnie mice exhibited anxiety-like behaviour through an inherent conflict between inclination exploratory drive and risk-avoidance behaviour [53]. Our results concur with previous behavioural studies in experimental models of colitis [93, 94]. In addition, groomings were fewer in Winnie mice, a potential indicator of a diminished drive for self-care and reduced self-efficacy of Bandura’s theory [95]. Depressed mice and those with neurodegenerative disorders typically display reduced self-grooming and grooming pattern rigidity [96]. Depressive-like behaviour was observed during the FST, in which the Winnie mice showed heightened immobility in a stress-induced environment [52]. These behavioural changes are consistent with the phenotypes observed in mouse models of chronic intestinal inflammation and may represent a translational bridge to understanding similar symptoms in IBD patients [24, 93, 94, 97]. Furthermore, these shifts in behaviour are notably antecedent to the development of psychiatric disorders and sickness behaviour [87].

Significant disruptions in glucose metabolism were identified in the Winnie mouse colon in this study. Downregulation of glycolysis and gluconeogenesis-associated genes, including G6pc, Pck1, Galm, Eno1, Pfkp, Adh5, Ldhc, and Aldh2, indicate impaired pyruvate, ethanol, and aldehyde metabolism, reducing capacity for ROS detoxification [67, 98, 99]. Given that increased ROS production is implicated in oxidative stress, a critical factor in colonic inflammation observed in both IBD patients and Winnie mice [47, 100], these findings underscore the relevance of altered glycolysis in the disease pathology. Altered monosaccharide metabolism in the Winnie mouse colon reflect adaptive changes in glucose processing demonstrated in individuals with IBD, suggesting a compensatory mechanism to synthesise glucose during scarcity [68, 101]. A preference for glucose metabolism and upregulation of Pkm2 reported in this study may indicate a protective mechanism for intestinal epithelial cells against apoptosis [102]. However, given the increased risk of colorectal cancer (CRC) associated with IBD [103], the role of Pkm2 upregulation in the tumorigenesis of CRC must be considered [104]. Dysregulation of glycolysis can lead to decreased mitochondrial OXPHOS activity [105]. This metabolic shift can cause disruptions in the TCA cycle and initiate metabolic reprogramming which not merely a consequence of inflammation, but can also contribute to the amplification of inflammatory responses and affect disease progression [101, 105, 106].

Functional analysis of downregulated FAO transcriptome implies an increased ketone body metabolism in the Winnie mouse distal colon. Although further protein level investigation is required, this finding indicates allosteric regulation for generating ketone bodies as alternative energy sources, particularly under conditions of low glucose availability, suggesting an adaptive metabolic response in the inflamed colon [107]. Furthermore, the reduced EE in Winnie mice may be attributed to downregulated Hadhb, a crucial component of the mitochondrial trifunctional protein complex, compounded by a decrease in Acadsb [108]. In consistency with our results, reduced Acadsb correlates with a decrease in the TCA cycle and OXPHOS activity, culminating in a lower ATP yield in colonic tissues from IBD patients [109, 110]. Downregulation of Acadvl associates with hypoglycaemia, muscle weakness, and increased production of hydrogen peroxide [47, 111]. Elevation of hydrogen peroxide levels inhibits critical enzymes of the TCA cycle, including aconitase, alpha-ketoglutarate dehydrogenase, and succinate dehydrogenase, exacerbating the pathological state [112]. Hence, these findings corroborate the association between chronic intestinal inflammation and oxidative stress in IBD patients and Winnie mice.

The results of our study suggest that the TCA cycle in the Winnie mouse distal colon may undergo allosteric self-regulation as a response to changes in the glycolysis pathway [41, 113]. In consistency with our findings downregulation of TCA cycle-associated genes, SUCLG1, OGDH, MDH2, and IDH3 in have been reported in mice with DSS induced-colitis [28]. Hence, it is plausible that downregulated Ogdh, Pck1, and Idh1 contribute to a shift towards less efficient anaerobic pathways and disturbances in the phosphorylation of nucleoside diphosphates during intestinal inflammation [34, 114]. In addition, reductions in Sucla2 and Suclg2 activity may have significant implications for mitochondrial DNA integrity [115–117]. Altered expression of Pcx, Mdh1b, and Idh2 observed in this study are consistent with alterations in NAD metabolism previously reported in Winnie mice [118].

The OXPHOS transcriptome of Winnie mice distal colons showed downregulation of Ndufa2, Ndufa1, Cox17, Cox7a1, Ndufb5, and Ndufb2 which influences the assembly of the mitochondrial respiratory chain complexes [119]. This aligns with previous studies in IBD patients and experimental models of colitis reporting the association between perturbed mitochondrial activity, ATP depletion, enteric inflammation, and mucosal barrier dysfunction [36, 120–123]. Furthermore, STRING analysis determined that reduced expression of AtpeV0A and AtpeV0E implies reduced proton pumping activity of V-ATPase, which may exacerbate the inflammatory conditions in the Winnie mouse colon by impairing ion exchange and mucosal barrier function [124, 125].

To investigate whether inflammation-induced changes in cellular metabolism in the Winnie mouse colon are concomitant with disruption of the gut-brain axis, we performed RNA-Seq on brains from Winnie mice. The middle portion of the brain used in our study contains the limbic system, encompassing critical structures such as the amygdala and hippocampus, which are closely linked to emotional regulation, stress response, and behavioural modification, thus ideal for conducting RNA-Seq studies [126]. Although, changes in gene expressions can impact neuronal function, synaptic activity, and overall brain health [127, 128], neurological disorders involve complex, multi-stage processes that cannot be fully understood through the analysis of single markers alone. In providing a descriptive analysis of transcriptome variations, this study offers some insights on how metabolic dysfunction may contribute to neurological symptoms or disease progression in chronic intestinal inflammation.

We report that dysregulation of cellular metabolism in Winnie mouse colon is associated with downregulation of glycolysis, FAO, TCA cycle and OXPHOS in the brain, implicating disruption of the gut-brain axis. Downregulation of glycolysis-associated genes in the Winnie mice brain may influence the conversion of pyruvate to acetyl-CoA, accumulating mitochondria-toxic compound, which in turn, affect ATP and cellular redox balance previously described in neuropathologic abnormalities [129, 130]. Previous studies have reported Bpgm and Pgam downregulation correlate to reduction in blood oxygen level-dependent signals in the brains of patients with UC, as well as susceptibility of neurodegenerative disorders and neurological deficits [131]. The expression of Aldh3a1 and Aldh2 was reduced in both the brains and distal colon from Winnie mice substantiating the association between oxidative stress, intestinal inflammation, and neurological distress via the gut-brain axis; diminished ROS detoxification renders the host vulnerable to aldehyde-induced damage, potentially generating neurotoxic dopamine metabolites, which are one of the foundations for neurological diseases [132].

During the FAO pathway of the Winnie mouse brain, downregulated Adh1, Hadh, Aldh2, and Hadhb associate with NAD binding, potentially diminishing redox reactions and mitochondrial metabolism [133, 134]. Upregulation of Cyp4a corresponds to behavioural changes in Winnie mice reported in this study as CYP enzymes are implicated in the activation of sensory neurons, potentially affecting autonomic and neuroendocrine responses, as well as pathological pain and stress-associated behaviours [135]. In contrast, neuroprotective properties are imminent from the upregulation of Cpt1a and Acsl3, slowing the advancement of age-related neurodegenerative conditions [136]. The TCA cycle transcriptome in Winnie mice brains could indicate disruptions in mitochondrial function and oxidative stress, all of which are implicated in neurological disorder-like pathology [137]. For instance, the decreased expression of Sucla2, Suclg2, and Sdha may negatively impact the metabolic adaptability of brain cells, ATP production, GTP synthesis, and the conversion of succinate to fumarate [116, 117]. These observations align with recent studies indicating that reduced metabolic flexibility in brain cells may increase susceptibility to metabolic stress, potentially exacerbating neurological disease progression [34, 114].

Studies using murine models of depression have demonstrated impaired mitochondrial structure, diminished respiration rates, and reduced mitochondrial membrane potential in the hippocampus, cortex, and hypothalamus [138]. The transcriptomic analysis of Winnie mice brains revealed significant reductions in the activity of mitochondrial ETC complex (C) I to V, with no compensatory upregulation of gene expression to offset the downregulation in transcriptome. In the absence of sufficient ATP from OXPHOS, cells switch to anaerobic glycolysis, producing lactate which can lead to lactic acidosis, impairing information processing, neuronal function and exacerbating hypoxic damage [139]. In conditions like neuropathic pain, mitochondrial dysfunction in sensory neurons and glial cells increases ROS production and disrupts calcium homeostasis, critical for synaptic activity and neuroplasticity [140]. It is plausible to consider that downregulated OXPHOS in Winnie mice brains may reduce the activity of sodium-potassium pumps that maintain neuronal resting potential, promoting chronic pain signalling [141]. This study demonstrated depressive and anxiety-like behaviours in Winnie mice alongside previous reports in experimental models of colitis [93, 142], which suggests an association between affective disorders and OXPHOS-related mitochondrial activity in the brain with chronic intestinal inflammation. However, further investigations are required to substantiate ATP levels and mitochondrial dysfunction in Winnie mice.

Given the prevalence of psychological disorder comorbidities in colitis, we proposed that cytokine modulating factors may play a role in shaping the clinical presentation of affective disorders. Preliminary findings have indicated a connection between inflammatory reactions and psychological disorders in IBD patients and experimental models of colitis [78, 143–145]. While the role of inflammatory cytokines in the development of these disorders remains unclear, dysregulated immune responses led to heightened permeability of the blood brain barrier in mice with DSS-induced colitis, enabling pro-inflammatory cytokines and immune cells to move from the gut into the brain [146, 147]. In this study, functional studies conferred that downregulated Il12a in Winnie mice brains associates with a previously reported decrease in tryptophan metabolism-associated enzyme, indoleamine 2,3-Dioxygenase (Ido) [118]. Studies show that rodents with reduced IDO exhibit a decrease in exploratory behaviour, which was is consistent with observations of Winnie mice during this study [148, 149]. Likewise, lack of Il2rb is linked to difficulties in spatial learning and memory [150]. In consistency with our findings, upregulation of IL23a and TNF-α is associated with major depressive disorder and neuropathic hyperalgesia [78], while TLR5 is linked to various CNS disorders [151]. This emphasises a pivotal connection between peripheral and central immune-inflammatory responses and their collective impact on mental health.

The results of this study present promising avenues for future research. While RNA-Seq performed in this study has laid the groundwork with descriptive analyses revealing alterations in the regulation of metabolism-related transcriptomes, there is a pressing need to expand upon these findings using diverse biochemical and molecular methodologies. Due to difficulties in acquiring brain tissue samples from individuals with IBD, there is presently insufficient direct evidence to confirm a connection between impaired neurogenesis and symptoms of anxiety and depression. Future studies ought to bridge this gap by exploring non-invasive methods like advanced imaging or biomarker analysis to study neurogenesis in living patients. Therefore, to comprehensively explore the intricate interactions of cellular metabolism and its systemic implications in IBD, future investigations should adopt a multifaceted approach. Conducting a thorough comparative analysis of cellular metabolism across colonic tissue, plasma, and cerebral structures is essential for gaining a holistic understanding of the characteristic metabolic disruptions associated with chronic intestinal inflammation. A valuable contribution would be to integrate the colon and brain transcriptomic data to explore potential correlations or shared pathways by performing correlation analyses, pathway overlap enrichment, or even random forest classification to link molecular and behavioural/physiological data. Systems biology strategies can integrate data from genomics, proteomics, and metabolomics to construct biological networks that underlie cellular metabolism, offering predictive insights and a systemic comprehension of disease pathogenesis. Furthermore, employing multi-omic analyses could unravel complex interactions and feedback mechanisms, providing a more nuanced understanding of disease mechanisms. Techniques such as Western blotting, enzyme assays, and mass spectrometry would facilitate the quantification and validation of target proteins, potentially unveiling post-translational modifications and the impact of gene regulation on protein expression. Additionally, immunohistochemistry could localise these proteins within tissue contexts, providing insights into their cellular and tissue-specific distribution and their pathological relevance. While IBD encompasses both UC and CD, Winnie mice exhibit characteristics more closely resembling UC, particularly in terms of disease location, progression, and immune responses. Therefore, we acknowledge that the findings of this study may not fully generalise to CD.

Conclusion

This study provides critical insights into the multifaceted interplay between chronic intestinal inflammation and behavioural changes observed in IBD patients. Our findings have corroborated that chronic intestinal inflammation is intricately associated with a spectrum of behavioural changes. Importantly, we established that inflammation and oxidative stress in chronic colitis are associated with dysregulated cellular metabolism within the gut-brain axis, underlying behavioural alterations. These observations provide insight into pathological mechanisms contributing to the comorbidity of psychological disorders in UC patients. Although this study shows preliminary findings, understanding these changes is crucial for elucidating how metabolic dysfunction may contribute to neurological symptoms or disease progression during chronic intestinal inflammation, highlighting the need for more longitudinal studies.

Abbreviations

Acadl

Palmitoyl-CoA dehydrogenase

Acadm

Medium-chain acyl-CoA dehydrogenase

Acadsb

Electron-transfer flavoprotein 2,3-oxidoreductase

Acadvl

Palmitoyl-CoA:electron-transfer flavoprotein 2,3-oxidoreductase

Acat1

Acetyl-CoA acetyltransferase 1

Aco

Aconitate hydratase

Acsl

Acyl-CoA synthetase

Adh

Alcohol dehydrogenase

Aldh

Aldehyde dehydrogenase

Adpgk

ADP:D-glucose 6-phospho-transferase

Akr1al

Aldo-keto reductase family 1 member A1

Bpgm

Bisphosphoglycerate-phospho-glycerate mutase

CD

Crohn's disease

CRC

Colorectal cancer

Cox

Cytochrome c oxidase

CPM

Counts per million

Cpt1

Carnitine O-palmitoyl-transferase

Cyp4a10

Long-chain fatty acid omega-monooxygenase

DEG

Differentially expressed gene

DSS

Dextran sulfate sodium

Eci1

Enoyl-CoA delta isomerase 1

EE

Energy expenditure

Eno1

Enolase

EPM

Elevated plus maze

ENS

Enteric nervous system

FAO

Fatty acid β-oxidation

FST

Forced swim test

galM

D-glucose 1-epimerase

Gapdh

Glyceraldehyde 3-phosphate dehydrogenase

Gcdh

Glutaryl-CoA dehydrogenase

GI

Gastrointestinal

G6pc

Glucose 6-phosphate

Hadh

Hydroxyacyl-CoA dehydrogenase

Hk2

Hexokinase 2

IBD

Inflammatory bowel disease

Idh1

Isocitrate dehydrogenase 1

Ido

Indoleamine 2,3-dioxygenase

IL

Interleukin

KEGG

Kyoto encyclopedia of genes and genomes

Lcn

Lipocalin

Ldh

L-lactate dehydrogenase

Ldha

Lactose dehydrogenase

LDT

Light dark test

Mdh1b

Malate dehydrogenase b

Minpp1

Multiple inositol-polyphosphate phosphatase 1

NAD

Nicotinamide adenine dinucleotide

Nduf

NADH dehydrogenase

OFT

Open field test

Ogdh

Oxoglutarate dehydrogenase

OXPHOS

Oxidative phosphorylation

Pck1

Phosphoenolpyruvate carboxykinase 1

Pcx

Pyruvate carboxylase

Pdha

Pyruvate dehydrogenase

Pgk1

Phosphoglycerate kinase

Pfk9

6–phosphofructokinase

Pgam

Glycerate phosphomutase

Pkm

Pyruvate kinase

QoL

Quality of life

ROS

Reactive oxygen species

Sdh

Succinate dehydrogenase

Sucl

Succinate-Coenzyme A ligase GDP-forming subunit beta

TCA

Tricarboxylic acid

TLR

Toll-like receptor

TNF

Tumour necrosis factor

TPI

Triose phosphate isomerase

UC

Ulcerative colitis

Gene symbols abbreviations

Acadl

Palmitoyl-CoA dehydrogenase, long-chain-acyl-CoA dehydrogenase

Acadm

Medium-chain acyl-CoA dehydrogenase

Acadsb

Hexanoyl-CoA: electron-transfer flavoprotein 2,3-oxidoreductase

Acadvl

Palmitoyl-CoA:electron-transfer flavoprotein 2,3-oxidoreductase

Acat

Acetyl-CoA acetyltransferase

Acsl

Acyl-CoA synthetase

Adh

Alcohol dehydrogenase

Adpgk

ADP:D-glucose 6-phospho-transferase

Akr1a1

Alcohol dehydrogenase (NADP⁺)

Aldh

Aldehyde dehydrogenase (NAD⁺)

Bpgm

Bisphosphoglycerate-phospho-glycerate mutase

Cpt

Carnitine O-palmitoyl-transferase

Cd38

Cluster of differentiation 38

Cyp4a10

Long-chain fatty acid omega-monooxygenase cytochrome P450

Eci

Delta (D)3-D2-enoyl-CoA isomerases

Eno

Enolase

G6pc

Glucose 6-phosphate

Gadph

Glyceraldehyde 3-phosphate dehydrogenase

galM

Aldose 1- epimerase

Gcdh

Glutaryl-CoA dehydrogenase

Hadhb

Hydroxyacyl-CoA dehydrogenase trifunctional multienzyme complex subunit b

Hk

Hexokinase

Lcad

Long-chain acyl-CoA dehydrogenase

Ldh

L-lactate dehydrogenase

Mcad

Medium-chain acyl-CoA dehydrogenase

Minpp1

Multiple inositol-polyphosphate phosphatase/2,3-

Muc2

Mucin 2, oligomeric mucus/gel-forming

Oadh

Oxoglutarate dehydrogenase (succinyl-transferring)

Pck

Phosphoenol-pyruvate carboxykinase

Pfk

6-phosphofructokinase

Pgm

Phosphoglucomutase

Pgam

phosphoglycerate mutase

Pgk

Phospho-glycerate kinase

Pyk

Pyruvate kinase

Authors’ contributions

J.D. conception and design, collection, analysis and interpretation of data, manuscript writing; A.M.R. interpretation of data, manuscript writing; R.S. analysis, manuscript revision; M.D. collection of data; R.T.F. collection of data; R.E. collection of data; D.K. interpretation of data, critical revision of the manuscript; V.A. interpretation of data, critical revision of the manuscript; K.N. conception and design, interpretation of data, manuscript writing, and project administration. All authors approved the final version of the manuscript.

Funding

This study is supported by the Crohn’s & Colitis Foundation Senior Research Award (K.N., Award number: 903433) and Victoria University postgraduate research scholarships to J.D and M.D.

Data availability

All datasets generated during the current study are available from the corresponding author upon reasonable request. Sequencing data are deposited at the Gene Expression Omnibus (GEO) repository, with the accession numbers GSE244558 (colon) and GSE264317 (brain).

Declarations

Ethics approval and consent to participate

All animal experimental protocols and procedures were approved by the Victoria University Animal Experimentation Ethics Committee (AEC-17-016) and complied with the guidelines of the National Health and Medical Research Council Australian Code of Practice for the Care and Use of Animals for Scientific Purposes.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.