Plasma metabolomic and lipidomic profiling reveals distinct molecular signatures and time-dependent remodeling in ulcerative colitis Authors: Gülden Bilican, Nurulamin M. Noor, Stephan L. Haas, Grigorios Petrousis, Derya Kirman, Kenan Moral, Engin Koçak, Sevilay Erdoğan Kablan, Emirhan Nemutlu, Çağdaş Kalkan, Mehmet Cindoruk, Murat Kekilli, Tarkan Karakan Published: 2026 ABSTRACT BackgroundUlcerative colitis (UC) is a chronic inflammatory bowel disease marked by relapsing mucosal inflammation. Recent data indicate that subclinical metabolic disturbances persist even in patients who have achieved remission. However, the temporal evolution of these molecular alterations remains poorly defined. This study examined how disease duration shapes systemic metabolic and lipidomic remodeling in UC.MethodsPlasma samples from 58 UC patients in clinical remission and 70 age- and sex- matched healthy controls (HC) were analyzed using integrated untargeted gas chromatography–mass spectrometry (GC–MS) and liquid chromatography–mass spectrometry (LC–MS). Patients were stratified by disease duration (≤ 2 years vs. > 2 years). Multivariate and univariate analyses, including partial least squares discriminant analysis (PLS-DA), variable importance in projection (VIP) scoring, and t-tests or Mann–Whitney U test (p < 0.05), were applied, and pathway enrichment was performed using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database.ResultsUC patients showed distinct plasma metabolomic and lipidomic signatures compared with HC, with 15 metabolites and 126 lipid species significantly altered. Reduced pyruvic and lactic acids, together with disrupted fatty-acid and amino-acid metabolism, suggested impaired energy and redox homeostasis. Lipidomic profiling revealed broad disturbances across phosphatidylcholines, lysophosphatidylcholines, phosphatidylethanolamines, sphingomyelins, diacylglycerols, and triglycerides, reflecting persistent oxidative and inflammatory imbalance. Stratification by disease duration revealed progressive remodeling, particularly within amino-acid pathways and sphingolipid metabolism.ConclusionsEven during remission, UC exhibits systemic metabolic disequilibrium that evolves with disease chronicity. These time-dependent changes could guide the identification of metabolic biomarkers for disease monitoring and personalized therapeutic strategies.Graphical Abstract Supplementary InformationThe online version contains supplementary material available at 10.1186/s12876-026-04944-w. Introduction Ulcerative colitis (UC) is a chronic inflammatory bowel disease (IBD) characterized by continuous mucosal inflammation beginning in the rectum and extending proximally through the colon. It typically follows a relapsing–remitting course. UC incidence has plateaued in high-income countries but is rising rapidly in newly industrialized regions, contributing to a growing global disease burden [1]. The etiology is multifactorial, involving genetic susceptibility, immune dysregulation, environmental factors, and alterations in the gut microbiome [2]. Despite significant progress in medical therapy, a considerable proportion of patients continue to experience flares or an eventual loss of therapeutic response. Current data show that approximately 10% of patients with UC still require a colectomy at some stage during their disease course [3]. Given the variability in disease trajectory and treatment response, there has been a growing interest in personalized treatment strategies in UC. Precision medicine approaches aim to optimize therapeutic selection and timing for individual patients, guided by a mechanistic understanding of disease pathways. In this context, omics-based technologies, including metabolomics and lipidomics, have emerged as promising tools for more tailored therapeutic approaches [4, 5]. While early investigations largely emphasized genetic and genomic approaches for treatment personalization, recent research has increasingly focused on the functional biology downstream of genomic alterations, with growing recognition of the roles of metabolites and lipids in modulating inflammatory processes. Metabolomics seeks to evaluate small molecules that mirror how cellular metabolism changes under disease conditions [6, 7]. Meanwhile, lipidomics focuses on lipids that sustain energy balance, maintain membrane integrity, and regulate cell signaling. Disruption of lipid pathways has been increasingly recognized as a central feature of UC, contributing to inflammation, epithelial damage, and immune activation [8]. Moreover, growing evidence indicates that patients with UC display distinct metabolic patterns compared with healthy individuals, particularly in pathways related to amino acid, fatty acid, and bile acid metabolism, as well as in metabolites linked to oxidative stress [9–11]. These patterns appear to reflect not only ongoing inflammatory activity but also adaptive host–microbiota interactions. Certain metabolic shifts have also been linked to treatment response, suggesting potential use as biomarkers for therapeutic monitoring [12]. However, it remains unclear how disease duration influences metabolic and lipidomic patterns in UC. Chronic inflammation can induce metabolic adaptations—some protective, others deleterious—that may contribute to disease persistence and inter-patient variability. It also remains unclear whether these alterations persist during remission or how they differ from those observed in healthy individuals. Accordingly, this study aimed to characterize plasma metabolomic and lipidomic profiles in patients with UC in clinical remission and to assess whether disease duration and chronicity influence these metabolic patterns. We focused specifically on patients in clinical remission, as this approach allows metabolic alterations to be examined without the confounding influence of active inflammation, thereby revealing adaptations that may persist as part of the disease’s biological footprint. Findings were also compared with healthy controls (HCs) to determine whether UC-specific metabolomic and lipidomic changes could be identified. Materials and methods Study population This cross-sectional study included 58 patients with a confirmed diagnosis of UC in clinical remission and 70 age- and sex-matched HCs. UC was diagnosed based on standard clinical, endoscopic, and histopathological criteria. Clinical remission was defined as a total or partial. Mayo score of ≤ 1. Patients who had received corticosteroid therapy within the previous 3 months were excluded. Patients were classified as having “early” UC if the duration of time since diagnosis was ≤ 2 years and “late” UC if there was more than 2 years since diagnosis. HCs had no history of inflammatory, metabolic, or gastrointestinal disorders and were not taking any medications at the time of sample collection. Sample collection and preparation Peripheral blood samples were collected in EDTA-containing tubes following an overnight fast and centrifuged at 10,000 rpm for 10 min at 4 °C to separate plasma, which was then aliquoted and stored at − 80 °C until analysis. To minimize pre-analytical variability, all samples were collected between 08:00 and 10:00 a.m. and processed using standardized protocols. Metabolites and lipids were extracted using a co-solvent methanol: water: chloroform (4:3:3, v/v) system, consistent with our previously established protocols for plasma analysis [13]. Briefly, 400 µL of methanol was added to 100 µL of plasma, followed by 300 µL of water and 300 µL of chloroform. The mixture was vortexed and then centrifuged at 10,000 rpm for 10 min, resulting in three distinct layers: a polar, metabolite-rich upper phase; a lower lipid-containing chloroform phase; and a protein-rich interphase. The upper and lower layers were carefully separated and stored at − 80 °C prior to analysis. Gas Chromatography–Mass spectrometry (GC-MS) based metabolomics analysis The separated supernatants were fully evaporated overnight at 4 °C using a vacuum centrifuge. The resulting dried residues were subjected to derivatization via methoxylation and trimethylsilylation, following previously described protocols [13]. Briefly, 20 µL of methoxyamine hydrochloride (20 mg/mL in pyridine) was added to each sample and incubated at 30 °C for 90 min. Subsequently, N-methyl-N-(trimethylsilyl)-trifluoroacetamide containing 1% trimethylchlorosilane was added, and samples were further incubated at 37 °C for 30 min. GC–MS was performed using a Shimadzu GCMS-QP2010 Ultra system (Kyoto, Japan) equipped with a DB-5MS capillary column (30 m × 0.25 mm i.d., 0.25 μm film thickness). Samples were injected in splitless mode. The GC oven temperature program consisted of an initial hold at 60 °C for 1 min, followed by a linear increase to 325 °C at 10 °C/min, with a final hold for 10 min. Electron ionization mass spectrometry (EI-MS) was conducted at 70 eV in full scan mode, acquiring data across an m/z range of 50–650. Data acquisition was performed in positive ionization mode. Metabolite identification was performed using MS-DIAL version 2.56. Data preprocessing included chromatogram deconvolution, peak alignment, and generation of data matrix. Metabolites were detected in the mass range of 50–650 Da, with a minimum peak height threshold of 1000. Retention time tolerance was set to 0.05 min. A metabolite identification confidence threshold of 70% relative to the Fiehn Retention Index database was applied. Liquid Chromatography–Mass spectrometry (LC-MS) based lipidomics analysis The lipid-containing supernatants were evaporated overnight at 4 °C using a vacuum centrifuge. The dried samples were reconstituted in 300 µL of isopropanol: acetonitrile (IPA: ACN, 7:3, v/v), gently vortexed, and subsequently filtered. Lipidomic analysis was conducted using an Agilent 6530 LC/Q-TOF-MS system (California, USA). Chromatographic separation was achieved on a Poroshell HPH C18 column (100 × 2.1 mm, 2.7 μm particle size). The mobile phases consisted of (A) water: acetonitrile (6:4, v/v) with 0.1% formic acid and 10 mM ammonium formate, and (B) isopropanol: acetonitrile (9:1, v/v) with 0.1% formic acid and 10 mM ammonium formate. The flow rate was maintained at. 0.25 mL/min, and the column temperature was set at 60 °C. Data acquisition was performed in both positive and negative ionization modes over a mass-to-charge ratio (m/z) range of 100–1700. Tandem mass spectrometry (MS/MS) fragmentation was carried out using medium isolation with collision energies of 10, 20, and 40 V. Quality control (QC) samples were processed and analyzed in parallel using the same protocol. Raw data were processed using MS-DIAL version 4.92 for feature detection, deconvolution, alignment, gap filling, and normalization. Lipid identification was performed by matching. MS/MS spectra (acquired at 10 V, 20 V, and 40 V) to the MS-DIAL LipidBlast database, applying a 70% similarity threshold. Statistical analysis Baseline characteristics are presented as mean ± SD for continuous data and number (%) for categorical data. Group comparisons were made with the t-test or Mann–Whitney U test for continuous variables and the chi-square test for categorical variables. P < 0.05 was considered significant. Analyses were performed using SPSS v26.0 (IBM, Armonk, NY). For metabolomic and lipidomic datasets, the MetaboAnalyst 5.0 platform was used. Partial Least Squares Discriminant Analysis (PLS-DA) was applied to identify group-specific patterns and assess differences across experimental groups. This supervised multivariate method allows visualization of group separation and identifies the metabolites or lipid species that contribute most to the observed variation. The Variable Importance in Projection (VIP) score reflects the relative contribution of each variable to the PLS-DA model, with values greater than 1 considered significant for class discrimination. Volcano plot analysis was performed to identify significantly altered metabolites and lipids. Student’s t-test (p < 0.05) was used together with predefined fold-change thresholds (1.2 for metabolites and 1.5 for lipids). Moreover, false discovery rate (FDR)-adjusted q-values were calculated. For metabolomic pathway enrichment analysis, a minimum of three metabolites per pathway was required to ensure reliable pathway identification. Results A total of 128 individuals were included: 58 patients with UC in clinical remission and 70 age- and sex-matched HC (Table 1). Plasma metabolomic and lipidomic profiles were analyzed, and patients were stratified by disease duration (≤ 2 years vs. >2 years) for subgroup analyses. Table 1Basic characteristics of UC and HC subjectsVariableUC (n = 58)HC (n = 70)P-valueSex, n (%)0.12 Male, n (%)32 (55.2%)29 (41.4%) Female, n (%)26 (44.8%)41 (58.6%)Age, years44.3 ± 16.243.9 ± 13.90.895BMI, kg/m²25.3 ± 4.325.1 ± 4.20.82Smoking status, n (%)0.691 Yes17 (29.3%)24 (34.3%) No31 (53.4%)35 (50.0%) Former10 (17.2%)11 (15.7%)Disease extent, n (%)— E1: Proctitis10— E2: Left-sided colitis29— E3: Extensive colitis19—Extraintestinal manifestations, n (%)— Skin involvement2 (3.4%)— Arthralgia4 (6.9%)— Ocularinvolvement1 (1.7%)— Primary sclerosing cholangitis3 (5.2%)— None48 (82.8%)— Hypertension6 (10.3%) Diabetes mellitus7 (12.1%)Medication at the time of sample collection, n (%) 5-ASA monotherapy25 (43.1%)— 5-ASA+ Azathioprine9 (15.5%)— Biologics*23 (39.7%)— None1 (1.7%)—TG (mg/dL)139.6 ± 91.7130.9 ± 72.80.40LDL (mg/dL)122.6 ± 32.9118.2 ± 29.40.35HDL (mg/dL)48.9 ± 12.049.8 ± 11.60.65T-Cho (mg/dL)196.5 ± 40.4195.0 ± 36.10.81Glucose (mg/dL)98.7 ± 14.295.5 ± 10.70.21Albumin (mg/dL)4.0 ± 0.5—Data are presented as numbers (n) with corresponding percentages (%) for categorical variables and as mean ± standard deviation (SD) for continuous variables, unless otherwise specifiedAbbreviations: 5-ASA 5-aminosalicylates, BMI Body mass index, HDL High-density lipoprotein, LDL Low-density lipoprotein, T-Cho Total cholesterol, TG Triglyceride, UC Ulcerative colitis, HC Healthy control*Biologics include the use of adalimumab, infliximab, vedolizumab, and ustekinumab Table 1Basic characteristics of UC and HC subjectsVariableUC (n = 58)HC (n = 70)P-valueSex, n (%)0.12 Male, n (%)32 (55.2%)29 (41.4%) Female, n (%)26 (44.8%)41 (58.6%)Age, years44.3 ± 16.243.9 ± 13.90.895BMI, kg/m²25.3 ± 4.325.1 ± 4.20.82Smoking status, n (%)0.691 Yes17 (29.3%)24 (34.3%) No31 (53.4%)35 (50.0%) Former10 (17.2%)11 (15.7%)Disease extent, n (%)— E1: Proctitis10— E2: Left-sided colitis29— E3: Extensive colitis19—Extraintestinal manifestations, n (%)— Skin involvement2 (3.4%)— Arthralgia4 (6.9%)— Ocularinvolvement1 (1.7%)— Primary sclerosing cholangitis3 (5.2%)— None48 (82.8%)— Hypertension6 (10.3%) Diabetes mellitus7 (12.1%)Medication at the time of sample collection, n (%) 5-ASA monotherapy25 (43.1%)— 5-ASA+ Azathioprine9 (15.5%)— Biologics*23 (39.7%)— None1 (1.7%)—TG (mg/dL)139.6 ± 91.7130.9 ± 72.80.40LDL (mg/dL)122.6 ± 32.9118.2 ± 29.40.35HDL (mg/dL)48.9 ± 12.049.8 ± 11.60.65T-Cho (mg/dL)196.5 ± 40.4195.0 ± 36.10.81Glucose (mg/dL)98.7 ± 14.295.5 ± 10.70.21Albumin (mg/dL)4.0 ± 0.5—Data are presented as numbers (n) with corresponding percentages (%) for categorical variables and as mean ± standard deviation (SD) for continuous variables, unless otherwise specifiedAbbreviations: 5-ASA 5-aminosalicylates, BMI Body mass index, HDL High-density lipoprotein, LDL Low-density lipoprotein, T-Cho Total cholesterol, TG Triglyceride, UC Ulcerative colitis, HC Healthy control*Biologics include the use of adalimumab, infliximab, vedolizumab, and ustekinumab Basic characteristics of UC and HC subjects Data are presented as numbers (n) with corresponding percentages (%) for categorical variables and as mean ± standard deviation (SD) for continuous variables, unless otherwise specified Abbreviations: 5-ASA 5-aminosalicylates, BMI Body mass index, HDL High-density lipoprotein, LDL Low-density lipoprotein, T-Cho Total cholesterol, TG Triglyceride, UC Ulcerative colitis, HC Healthy control *Biologics include the use of adalimumab, infliximab, vedolizumab, and ustekinumab Metabolomic profile Untargeted GC–MS analysis identified approximately 140 metabolites. Multivariate analysis using PLS-DA demonstrated clear separation between UC patients and HC (Fig. 1A). The top 15 discriminatory metabolites are shown in the VIP score plot (Fig. 1B). Volcano plot analysis identified significantly altered metabolites between groups (Fig. 1C; p < 0.05, fold-change cutoff = 1.2). Fig. 1Plasma metabolomic profiling distinguishing ulcerative colitis (UC) patients from healthy controls (HC). A Partial least squares discriminant analysis (PLS-DA) score plot showing separation between UC (n = 58) and HC (n = 70). B Variable importance in projection (VIP) score plot contributing to group discrimination (VIP > 1). C Volcano plot illustrating significantly altered metabolites between UC and HC. Significance was determined using Student’s t-test (p < 0.05) with a fold-change cutoff of 1.2. Multiple testing correction was applied using the Benjamini–Hochberg method (FDR-adjusted q-values). Data are presented as mean ± standard deviation (SD) where applicable. Abbreviations: C, healthy control; U, ulcerative colitis; PLS-DA, partial least squares discriminant analysis; VIP, variable importance in projection, FC, fold change Plasma metabolomic profiling distinguishing ulcerative colitis (UC) patients from healthy controls (HC). A Partial least squares discriminant analysis (PLS-DA) score plot showing separation between UC (n = 58) and HC (n = 70). B Variable importance in projection (VIP) score plot contributing to group discrimination (VIP > 1). C Volcano plot illustrating significantly altered metabolites between UC and HC. Significance was determined using Student’s t-test (p < 0.05) with a fold-change cutoff of 1.2. Multiple testing correction was applied using the Benjamini–Hochberg method (FDR-adjusted q-values). Data are presented as mean ± standard deviation (SD) where applicable. Abbreviations: C, healthy control; U, ulcerative colitis; PLS-DA, partial least squares discriminant analysis; VIP, variable importance in projection, FC, fold change Volcano plot analysis showed significant differences in multiple metabolites (p < 0.05, FC cutoff = 1.2). Compared with HC, UC had lower levels of β-alanine (C/UC = 2.99), lactic acid (1.46), pyruvic acid (1.46), and saturated fatty acids—octanoic, dodecanoic, tetradecanoic, and hexadecanoic acids (1.21–1.25). By contrast, UC had higher levels of N2-(D-1-carboxyethyl)-L-arginine (0.31), D- erythrose-4-phosphate (0.54), and 2-hydroxybutyric acid (0.74). The complete list is shown in Table 2. Table 2Significantly altered metabolites between UC and healthy controls. (Data are presented as mean values ± standard deviation (SD). Fold change is expressed as control/UC. Statistical significance was determined using Student’s t-test. False discovery rate (FDR)-adjusted q-values were calculated using the Benjamini–Hochberg method)Metabolite nameFold Change (C/UC)Average UCAverage CSD (UC)SD (C)P valueQ valueBeta-Alanine2.990.000482560.001444870.000180010.000720042.2643E-73.40E-06Lactic acid1.460.006660050.009734610.002541530.004882773.25E-052.44E-04Pyruvic acid1.460.006840590.009964490.002597680.002168045.73E-052.87E-04Octanoic acid1.250.000496810.000618850.000137070.000237950.000790090.00296284N2-(D-1-Carboxyethyl)-L-arginine0.310.000112453.4644E-050.000174619.2962E-050.00194390.00380359Tetradecanoic acid1.210.004511560.005466340.001326330.001957430.00246960.00380359Hexadecanoic acid1.220.005733430.007004680.001935570.002562360.00210440.00380359Dodecanoic acid1.228.7267E-050.000106622.3951E-054.2169E-050.00247460.003803595-aminoimidazole-4-carboxamide0.700.062692390.041966570.048507720.024887860.00268710.00380359D-erythrose-4-phosphate0.546.4577E-053.4601E-057.01E-053.6696E-050.00278930.00380359Hydroxypropionic acid1.280.000899950.001148110.000427740.000520750.0046470.00580875Phosphoserine0.760.000100527.6665E-055.4325E-054.0671E-050.00593070.00684312Uric acid0.690.001038810.000711820.000716760.000696560.0109830.0117675N-Acetyl-glucosamine 1-phosphate1.413.2417E-054.5781E-052.1817E-053.6202E-050.0156270.0156272-hydroxybutyric acid0.740.000371820.000274850.000236550.00014230.0023310.00380359Abbreviations: C Healthy control, UC Ulcerative colitis, SD Standard deviation Table 2Significantly altered metabolites between UC and healthy controls. (Data are presented as mean values ± standard deviation (SD). Fold change is expressed as control/UC. Statistical significance was determined using Student’s t-test. False discovery rate (FDR)-adjusted q-values were calculated using the Benjamini–Hochberg method)Metabolite nameFold Change (C/UC)Average UCAverage CSD (UC)SD (C)P valueQ valueBeta-Alanine2.990.000482560.001444870.000180010.000720042.2643E-73.40E-06Lactic acid1.460.006660050.009734610.002541530.004882773.25E-052.44E-04Pyruvic acid1.460.006840590.009964490.002597680.002168045.73E-052.87E-04Octanoic acid1.250.000496810.000618850.000137070.000237950.000790090.00296284N2-(D-1-Carboxyethyl)-L-arginine0.310.000112453.4644E-050.000174619.2962E-050.00194390.00380359Tetradecanoic acid1.210.004511560.005466340.001326330.001957430.00246960.00380359Hexadecanoic acid1.220.005733430.007004680.001935570.002562360.00210440.00380359Dodecanoic acid1.228.7267E-050.000106622.3951E-054.2169E-050.00247460.003803595-aminoimidazole-4-carboxamide0.700.062692390.041966570.048507720.024887860.00268710.00380359D-erythrose-4-phosphate0.546.4577E-053.4601E-057.01E-053.6696E-050.00278930.00380359Hydroxypropionic acid1.280.000899950.001148110.000427740.000520750.0046470.00580875Phosphoserine0.760.000100527.6665E-055.4325E-054.0671E-050.00593070.00684312Uric acid0.690.001038810.000711820.000716760.000696560.0109830.0117675N-Acetyl-glucosamine 1-phosphate1.413.2417E-054.5781E-052.1817E-053.6202E-050.0156270.0156272-hydroxybutyric acid0.740.000371820.000274850.000236550.00014230.0023310.00380359Abbreviations: C Healthy control, UC Ulcerative colitis, SD Standard deviation Significantly altered metabolites between UC and healthy controls. (Data are presented as mean values ± standard deviation (SD). Fold change is expressed as control/UC. Statistical significance was determined using Student’s t-test. False discovery rate (FDR)-adjusted q-values were calculated using the Benjamini–Hochberg method) Abbreviations: C Healthy control, UC Ulcerative colitis, SD Standard deviation Pathway enrichment analysis identified four major pathways with significant alterations: biosynthesis of cofactors (four metabolites), propanoate metabolism (four metabolites), fatty acid biosynthesis (four metabolites), and biosynthesis of amino acids (three metabolites) (Supplementary Table S1). Lipidomic profile PLS-DA demonstrated clear separation between the lipidomic profiles of UC patients and HC (Fig. 2A), with the most discriminant lipid species identified by VIP analysis (Fig. 2B). Volcano plot analysis identified significantly altered lipids, with a total of 126 lipid species differing between groups (Fig. 2C). For each lipid species, mean values, SD, and FDR-adjusted q-values were calculated (Supplementary Table S2). Fig. 2Plasma lipidomic profiling distinguishing UC patients from HC. A PLS-DA score plot demonstrating separation between UC and HC groups. B VIP score plot identifying lipid species contributing most to group discrimination (VIP > 1). C Volcano plot displaying significantly altered lipid species between UC and HC. Statistical significance was defined as p < 0.05 with a fold-change cutoff of 1.5. Multiple testing correction was performed using the Benjamini–Hochberg method (FDR-adjusted q-values). Data are presented as mean ± SD where applicable. Abbreviations: C, healthy control; U, ulcerative colitis; PLS-DA, partial least squares discriminant analysis; VIP, variable importance in projection, TAG, triacylglycerol; DAG, diacylglycerol; PC, phosphatidylcholine; LPC, lysophosphatidylcholine; SM, sphingomyelin; FC, fold change Plasma lipidomic profiling distinguishing UC patients from HC. A PLS-DA score plot demonstrating separation between UC and HC groups. B VIP score plot identifying lipid species contributing most to group discrimination (VIP > 1). C Volcano plot displaying significantly altered lipid species between UC and HC. Statistical significance was defined as p < 0.05 with a fold-change cutoff of 1.5. Multiple testing correction was performed using the Benjamini–Hochberg method (FDR-adjusted q-values). Data are presented as mean ± SD where applicable. Abbreviations: C, healthy control; U, ulcerative colitis; PLS-DA, partial least squares discriminant analysis; VIP, variable importance in projection, TAG, triacylglycerol; DAG, diacylglycerol; PC, phosphatidylcholine; LPC, lysophosphatidylcholine; SM, sphingomyelin; FC, fold change At the class level, multiple phosphatidylcholines (PC) differed between UC and controls (Fig. 3A). Lysophosphatidylcholines (LPCs) also exhibited significant alterations, with five species affected (Fig. 3B). Phosphatidylethanolamines (PE) were consistently downregulated, with seven species reduced in UC (Fig. 3C). Several oxidized phospholipids (OxPLs), including oxidized PE and phosphatidylglycerol (PG) species, were predominantly downregulated (Fig. 3D). Fig. 3Class-level lipid alterations in UC compared with HCs. A Phosphatidylcholines (PC) showing species-specific increases and decreases. B Lysophosphatidylcholines (LPC) demonstrating class-level differences. C Phosphatidylethanolamines (PE) predominantly reduced in UC. D Oxidized phospholipids (OxPL) and phosphatidylglycerols (PG) showing downregulation patterns in UC. Data represent fold-change differences between groups. Statistical significance was determined using p < 0.05 with FDR-adjusted q-values. Data are presented as mean ± SD where applicable. Abbreviations: HC, healthy control; UC, ulcerative colitis; PC, phosphatidylcholine; LPC, lysophosphatidylcholine; PE, phosphatidylethanolamine; PG, phosphatidylglycerol Class-level lipid alterations in UC compared with HCs. A Phosphatidylcholines (PC) showing species-specific increases and decreases. B Lysophosphatidylcholines (LPC) demonstrating class-level differences. C Phosphatidylethanolamines (PE) predominantly reduced in UC. D Oxidized phospholipids (OxPL) and phosphatidylglycerols (PG) showing downregulation patterns in UC. Data represent fold-change differences between groups. Statistical significance was determined using p < 0.05 with FDR-adjusted q-values. Data are presented as mean ± SD where applicable. Abbreviations: HC, healthy control; UC, ulcerative colitis; PC, phosphatidylcholine; LPC, lysophosphatidylcholine; PE, phosphatidylethanolamine; PG, phosphatidylglycerol Sphingomyelins (SM) exhibited heterogeneous alterations, with 14 species decreased and 16 species increased in UC patients (Fig. 4A). Diacylglycerols (DG) were elevated in UC (Fig. 4B), while triacylglycerols (TG) were predominantly reduced (Fig. 4C). Among detected fatty acids, three of four species were downregulated in UC (Fig. 4D). Routine biochemical measurements of plasma triglycerides, LDL, HDL, total cholesterol, and glucose did not differ significantly between UC and HC (Table 1). Fig. 4Class-specific lipid alterations in UC and HCs. A Sphingomyelins (SM) showing both increased and decreased species. B Diacylglycerols (DG) elevated in UC patients. C Triacylglycerols (TG) predominantly reduced in UC. D Fatty acid (FA) profiles showing differential regulation. Statistical significance was determined using Student’s t-test (p < 0.05) with FDR-adjusted q-values calculated using the Benjamini–Hochberg method. Data are presented as mean ± standard deviation (SD) where applicable. Abbreviations: HC, healthy control; UC, ulcerative colitis; SM, sphingomyelin; DG, diacylglycerol; TG, triacylglycerol; FA, fatty acid Class-specific lipid alterations in UC and HCs. A Sphingomyelins (SM) showing both increased and decreased species. B Diacylglycerols (DG) elevated in UC patients. C Triacylglycerols (TG) predominantly reduced in UC. D Fatty acid (FA) profiles showing differential regulation. Statistical significance was determined using Student’s t-test (p < 0.05) with FDR-adjusted q-values calculated using the Benjamini–Hochberg method. Data are presented as mean ± standard deviation (SD) where applicable. Abbreviations: HC, healthy control; UC, ulcerative colitis; SM, sphingomyelin; DG, diacylglycerol; TG, triacylglycerol; FA, fatty acid Disease duration analysis Baseline demographic and clinical characteristics stratified by disease duration (≤ 2 years vs. >2 years) are presented in Supplementary Table S3. When stratified by disease duration, metabolomic profiles demonstrated clear separation between patients with ≤ 2 years and those with > 2 years of disease (Fig. 5A). Amino acids were among the most discriminant metabolites in VIP analysis (Fig. 5B). Volcano plot analysis identified significantly altered metabolites between groups (Fig. 5C; p < 0.05, fold-change cutoff = 1.2). Amino acid profiles were particularly affected by disease duration (Fig. 5D). All significantly altered metabolites are provided in Supplementary Table S4, including mean values, SD, p-values, and FDR-adjusted q-values. Fig. 5Metabolomic differences according to disease duration in UC. A PLS-DA score plot showing separation between early (≤ 2 years) and long-standing (> 2 years) UC. B VIP plot highlighting amino acids as key discriminant metabolites. C Volcano plot illustrating significantly altered metabolites (p < 0.05, fold-change cutoff = 1.2). D Fold-change analysis showing differences in amino acid levels between early and long-standing UC. Statistical significance was determined using Student’s t-test with FDR-adjusted q-values calculated using the Benjamini–Hochberg method. Data are presented as mean ± SD where applicable (see Supplementary Table S4). Abbreviations: UC, ulcerative colitis; U1, UC patients with disease duration ≤ 2 years (early UC); U2, UC patients with disease duration > 2 years (long-standing UC); PLS-DA, partial least squares discriminant analysis; VIP, variable importance in projection; FDR, false discovery rate; SD, standard deviation Metabolomic differences according to disease duration in UC. A PLS-DA score plot showing separation between early (≤ 2 years) and long-standing (> 2 years) UC. B VIP plot highlighting amino acids as key discriminant metabolites. C Volcano plot illustrating significantly altered metabolites (p < 0.05, fold-change cutoff = 1.2). D Fold-change analysis showing differences in amino acid levels between early and long-standing UC. Statistical significance was determined using Student’s t-test with FDR-adjusted q-values calculated using the Benjamini–Hochberg method. Data are presented as mean ± SD where applicable (see Supplementary Table S4). Abbreviations: UC, ulcerative colitis; U1, UC patients with disease duration ≤ 2 years (early UC); U2, UC patients with disease duration > 2 years (long-standing UC); PLS-DA, partial least squares discriminant analysis; VIP, variable importance in projection; FDR, false discovery rate; SD, standard deviation Pathway enrichment analysis confirmed significant alterations in amino acid-related pathways, including alanine, aspartate, and glutamate metabolism; glycine–serine–threonine metabolism; and cysteine–methionine metabolism, as well as branched-chain amino acid biosynthesis (valine, leucine, and isoleucine). Additional alterations were observed in purine, pyrimidine, and 2-oxocarboxylic acid metabolism. The complete list of significantly affected pathways is provided in Supplementary Table S5. Fold-change analysis comparing early and long-standing UC revealed distinct differences in amino acid concentrations. Metabolites including L-alanine, glutamine, L-aspartic acid, serine, methionine, L-threonine, 4-hydroxyproline, and phenylalanine exhibited positive fold-change values (log₂ FC = 0.3–1.0) in early disease, indicating higher relative abundance compared with late disease. In contrast, leucine showed a negative fold-change (log₂ FC ≈ − 0.3), reflecting lower levels in early compared with late disease. In lipidomic analyses, differences according to disease duration were also observed. SM were increased in patients with longer disease duration (Fig. 6A). TG exhibited species-level differences, with several species elevated in early disease and others reduced in patients with disease duration exceeding two years (Fig. 6B). DG were higher in patients with disease duration greater than two years (Fig. 6C). PC levels were also upregulated with increasing disease duration. The full list of significantly altered lipid species according to disease duration, including SM, TG, DG, and PC, is provided in Supplementary Table S6. Fig. 6Lipidomic differences according to disease duration in UC. A Sphingomyelins (SM) increased in long-standing UC (> 2 years). B Triacylglycerols (TG) showing species-level differences between early and long-standing disease. C Diacylglycerols (DG) elevated in long-standing UC. D Additional lipid class alterations associated with disease duration. Statistical significance was determined using Student’s t-test (p < 0.05) with FDR-adjusted q-values calculated using the Benjamini–Hochberg method. Data are presented as mean ± SD where applicable. Abbreviations: UC, ulcerative colitis; U1, UC patients with disease duration ≤ 2 years (early UC); U2, UC patients with disease duration > 2 years (long-standing UC); SM, sphingomyelin; TG, triacylglycerol; DG, diacylglycerol; PC, phosphatidylcholine; FDR, false discovery rate; SD, standard deviation Lipidomic differences according to disease duration in UC. A Sphingomyelins (SM) increased in long-standing UC (> 2 years). B Triacylglycerols (TG) showing species-level differences between early and long-standing disease. C Diacylglycerols (DG) elevated in long-standing UC. D Additional lipid class alterations associated with disease duration. Statistical significance was determined using Student’s t-test (p < 0.05) with FDR-adjusted q-values calculated using the Benjamini–Hochberg method. Data are presented as mean ± SD where applicable. Abbreviations: UC, ulcerative colitis; U1, UC patients with disease duration ≤ 2 years (early UC); U2, UC patients with disease duration > 2 years (long-standing UC); SM, sphingomyelin; TG, triacylglycerol; DG, diacylglycerol; PC, phosphatidylcholine; FDR, false discovery rate; SD, standard deviation Discussion In this study, we analyzed plasma metabolomic and lipidomic profiles in UC patients in clinical remission and compared them with those of HCs. We observed distinct clustering between UC patients and HCs, reflecting broad alterations in both metabolites and lipid species. Furthermore, stratification by disease duration (≤ 2 years vs. >2 years) revealed additional differences, suggesting that these molecular changes may evolve over time and contribute to disease chronicity. Collectively, these findings suggest that UC may be associated with systemic metabolic alterations even during clinical remission. IBD studies have primarily focused on identifying diagnostic biomarkers that differentiate Crohn’s disease from UC or from HCs, as well as on exploring markers of disease activity and therapeutic response [9, 10, 14–16]. However, rather few studies have investigated the impact. of disease duration on metabolic remodeling. Understanding temporal changes in the metabolome and lipidome is critical, as these alterations may reflect the combined effects of inflammation, therapeutic interventions, and microbial adaptation. In this study, we characterized these changes in plasma profiles of UC patients at different stages of disease duration to provide novel insights into their evolution. We specifically focused on patients in clinical remission to minimize the confounding effects of active inflammation and to better capture underlying metabolic adaptations associated with disease chronicity. Recent large-scale and translational studies increasingly show that chronic inflammation is closely linked with systemic metabolic imbalance. The UK Biobank atlas demonstrated that alterations in circulating lipids and amino acids occur across a wide range of inflammatory diseases, not solely those of metabolic origin [17]. In treatment-naive IBD patients, an inception cohort study demonstrated early alterations in tryptophan, nitrogen, and short-chain fatty acid metabolism, indicating that these metabolic changes appear before therapeutic interventions and parallel microbial shifts [18]. Another study confirmed enhanced tryptophan breakdown through the kynurenine pathway as a common signature in chronic inflammatory disorders [19]. Together, these findings emphasize that altered amino acid and lipid metabolism are closely linked to immune dysregulation in UC. In our cohort, pyruvic acid and lactic acid were reduced in UC compared with HC. As key intermediates of glycolysis and mitochondrial respiration, this may reflect altered cellular energy dynamics and possible changes in mitochondrial metabolism. Prior studies have reported variable results, likely due to differences in sample type and disease activity. For example, serum pyruvate has been reported as increased in treatment-naïve IBD, whereas a Crohn’s disease serum study showed decreased pyruvate, which aligns with our findings [10, 14]. Tissue metabolomics from UC/CD samples have shown reduced mucosal lactate that persists even during remission, suggesting diminished mucosal glucose utilization or increased peripheral uptake [20]. These variations highlight the importance of sample type when interpreting metabolic data, as blood reflects systemic balance while tissue and feces provide more localized metabolic information. Propionate metabolism, which depends on gut microbial fermentation, was also altered. Lower 3-hydroxypropionic acid in UC is consistent with alterations in short-chain fatty acid-related pathways, potentially affecting epithelial energy supply, barrier integrity, and anti-inflammatory signaling. In contrast, higher 2-hydroxybutyrate likely reflects increased oxidative stress and altered glutathione metabolism, a pattern described in chronic inflammatory states, including IBD [21, 22]. Pathway analysis also suggested impaired cofactor biosynthesis, which may influence mitochondrial energy production and host–microbial metabolic interactions [23]. Alterations in fatty acid biosynthesis were consistent with previous IBD findings [24, 25]. Medium-chain fatty acids with anti-inflammatory properties and long-chain fatty acids relevant for membrane stability were reduced, suggesting impaired microbial lipid metabolism and weakened anti-inflammatory capacity. Amino acid metabolism was similarly affected, which is notable given its roles in immune regulation, redox control, and mucosal repair [26, 27]. The changes observed in amino acid biosynthesis likely reflect ongoing host–microbe metabolic disturbance. Our lipidomic analysis showed broad shifts across multiple lipid classes. PCs and LPCs, critical for membrane structure and immune signaling, demonstrated altered composition, consistent with altered membrane composition and immune signaling [28–33, 35]. PE species were reduced, in line with previous IBD data [32], which may reflect impaired microbial lipid synthesis and compromised epithelial membrane stability [36, 37]. Oxidized phospholipids were largely decreased, suggesting oxidative imbalance and reduced antioxidant buffering capacity [38, 39]. Sphingomyelins showed mixed directional changes, possibly indicating an active but insufficient compensatory response to chronic inflammatory stress [40]. DG levels were elevated, consistent with enhanced protein kinase C mediated inflammatory signaling, while TG levels were reduced, which may indicate altered lipid utilization under prolonged inflammation [35, 41, 42]. In our study, several fatty acid species were altered between groups, with most (three out of four) showing a decrease in UC. Previous studies show that plasma fatty acid composition is strongly influenced by the gut microbiome [43]. Our findings likely reflect dysbiosis disrupting microbial fatty acid synthesis and signaling, altering the host–microbial metabolic balance in UC. Taken together, the observed alterations across PC, LPC, PE, OxPL, SM, DG, and TG classes depict a consistent pattern of oxidative stress, disrupted lipid turnover, and impaired membrane integrity. These findings suggest that metabolic and lipidomic alterations may persist even in the absence of clinical symptoms. The persistent lipidomic imbalance may reflect ongoing subclinical metabolic activity. Importantly, several of the metabolic and lipidomic alterations identified in our study are consistent with findings reported in independent UC cohorts, including disruptions in amino acid metabolism, fatty acid pathways, and phospholipid profiles [9, 10, 15, 31, 32, 35]. This concordance across studies provides indirect external validation and supports the robustness of our observations despite differences in study design and analytical platforms. We also observed differences in metabolic and lipidomic patterns when comparing patients with shorter versus longer disease duration. In patients with more than two years since diagnosis, amino acid metabolism was more markedly affected, including pathways important for epithelial repair and antioxidant defense. Declines in threonine and glutamine levels parallel earlier findings linking these metabolites to mucosal barrier support [44, 46]. Alterations in purine and pyrimidine metabolism further suggest sustained metabolic strain [47]. In the lipidome, increases in PC and SM together with higher DG and lower TG after longer disease duration point toward ongoing oxidative stress and persistent inflammatory signaling [28, 29, 39, 40, 42]. Overall, the time-dependent changes support the idea that UC may show persistent metabolic alterations during remission, with patterns that appear to vary according to disease duration. This study has important strengths. It combines metabolomic and lipidomic profiling and includes patients in clinical remission, which helps reveal ongoing molecular activity beyond inflammation. Disease-duration stratification adds a valuable temporal perspective that has been rarely addressed in previous UC metabolome studies. There are also some limitations. The cross-sectional design of this study limits causal interpretation of the observed metabolomic and lipidomic alterations.Treatment responses were not assessed, and the potential confounding effect of biologic therapies represents an important limitation, as treatment distribution differed between disease duration groups. Since these agents may directly influence metabolic and lipidomic profiles, this may confound the observed differences. Although a subgroup analysis according to treatment type would be informative, the number of patients in each treatment subgroup was not sufficient to allow a statistically robust comparison. Future studies incorporating treatment-stratified analyses would be valuable to better distinguish disease-driven from treatment-driven molecular changes. Furthermore, pathway enrichment analysis was not performed for lipidomic data due to the limited annotation of lipid metabolic pathways and the structural heterogeneity of lipid species. Additionally, as this was an untargeted metabolomic and lipidomic analysis, key pathways such as tryptophan metabolism and bile acid profiles were not specifically assessed, which are areas of growing interest in IBD research. Another limitation is the potential influence of sample handling and storage on metabolite stability, which highlights the need for standardized protocols in future metabolome studies [48]. Future research should include longitudinal designs and integrate treatment outcomes to better understand how metabolic profiles change over time in UC. Integrating metabolomic and lipidomic data with transcriptomic approaches, such as RNA sequencing, may further illuminate the molecular mechanisms underlying these findings. Conclusions Despite the growing global prevalence of UC, the functional pathways driving disease biology, particularly early in the disease course, remain incompletely understood. A deeper understanding of these molecular mechanisms is essential for developing more targeted and potentially preventive therapeutic strategies. In this study, plasma metabolomic and lipidomic analyses demonstrated that metabolic alterations persist even during clinical remission and become more pronounced with longer disease duration. These findings suggest ongoing metabolic adaptation over time and emphasize the importance of clearly defining disease phase when interpreting molecular profiles. Ultimately, such insights may help explain inter-patient heterogeneity and support progress toward more personalized treatment approaches for individuals living with UC. Supplementary Information Supplementary Material 1. Supplementary Material 1.