Introduction
Crohn's disease (CD) is a chronic inflammatory disorder of the gastrointestinal tract characterised by transmural inflammation and the development of hyperplastic mesenteric adipose tissue (MAT) surrounding affected intestinal segments, commonly referred to as creeping fat (CrF) [1]. CrF is a metabolically and immunologically active tissue that contributes to CD pathophysiology through cytokines, adipokines and other mediators [2].
An unexpected feature of CrF is its capacity to acquire characteristics of beige adipose tissue. Our group first reported the presence of a browning‐associated phenotype in CrF from CD patients, characterised by increased expression of UCP1 and other thermogenesis‐ and mitochondrial biogenesis‐associated markers [3, 4]. Subsequent studies independently confirmed the presence of beige adipocytes within hypertrophic MAT in CD, supporting browning as a previously unrecognised feature of this tissue [4].
Succinate can promote white‐to‐beige adipocyte conversion in mesenteric fat, linking microbial translocation to the CrF phenotype [3, 5]. Kynurenic acid‐mediated signalling has also been implicated in mesenteric adipose tissue browning [6], supporting interacting microbial, inflammatory and metabolic signals within CrF [7]. CrF browning may therefore represent a therapeutic target in CD [8].
Whether currently available therapies influence this phenotype, however, remains largely unknown.
Anti‐tumour necrosis factor (anti‐TNF) agents, including infliximab, are widely used to treat CD [9]. We previously associated anti‐TNF therapy with reduced CrF inflammation, restored adipocyte size and decreased immune‐cell infiltration [9]. Whether anti‐TNF treatment is associated with the CrF beige/browning phenotype remains unknown, as does whether related molecular changes follow direct ex vivo exposure of human CrF to infliximab.
We investigated whether anti‐TNF therapy was associated with browning‐/thermogenesis‐, mitochondrial biogenesis‐ and immune‐related markers in CrF and whether related changes followed direct ex vivo infliximab exposure in paired explants.
Material and Methods
Subjects and Study Design
Patients with CD were recruited at the Joan XXIII Hospital (Tarragona, Spain) and the Vall d’Hebron University Hospital (Barcelona, Spain), and all gave their informed consent. The study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the ethics and research committees of the respective hospitals (references CEIM 233/2024 and PR(CS)383/2021).
CD patients were recruited from those undergoing surgery for symptomatic complications. CrF biopsies were aseptically collected and, when appropriate, obtained prior to bowel opening or resection to avoid contamination.
Clinical, demographic, anthropometric and biochemical variables were collected for both the observational and experimental cohorts, including age, sex, body mass index, smoking status, disease location and behaviour (Montreal classification), C‐reactive protein and/or faecal calprotectin levels.
Due to tissue availability and experimental requirements, not all samples were included in every analysis. The number of samples used in each experiment is indicated in the corresponding figure legends.
Serum Cytokine Measurement
Serum interleukin (IL)‐1β and IL‐6 concentrations were measured in anti‐TNF‐naïve and anti‐TNF‐treated patients using ELISA kits (R&D Systems; DLB50 and HS600C, respectively) according to the manufacturer's instructions.
Isolation of AdMSCs and Mature Adipocytes
CrF samples were washed with phosphate‐buffered saline (PBS), and surgically damaged areas were removed when necessary. Tissue was digested with 0.1% collagenase and 1% BSA in PBS for 60 min at 37°C under gentle agitation, followed by centrifugation at 400 × g for 5 min at 24°C to separate floating mature adipocytes from the stromal vascular fraction (SVF). The SVF pellet was resuspended in DMEM/F‐12 medium (#311320‐074, Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% foetal bovine serum (#A5256801, Gibco, Thermo Fisher Scientific) and 1% penicillin–streptomycin–amphotericin B (#1674049, MP Biomedicals, Fisher Scientific, Madrid, Spain).
Cells were seeded in T75 flasks and maintained at 37°C in a humidified atmosphere containing 5% CO2. After approximately 7 days, cells reaching ∼90% confluence were harvested using trypsin–EDTA, generating the primary AdMSC culture (passage 0) [10]. Cell concentration and viability were determined using the Countess II FL Automated Cell Counter (Invitrogen, Thermo Fisher Scientific) with 0.4% trypan blue, according to the manufacturer's instructions.
AdMSCs Immunophenotyping
AdMSCs (2 × 105) were incubated with a panel of primary antibodies (described in Table S1). The minimal functional and quantitative criteria established by the International Society of Cell Therapy and the International Federation for Adipose Therapeutics and Science were confirmed by flow cytometry [11, 12, 13]. All experiments were performed using passage‐3 cells with > 90% viability and assessed using the Countess II FL Automated Cell Counter (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA).
Proteomics Analysis of AdMSC Conditioned Medium
An exploratory secondary reanalysis was performed on our previously published untargeted proteomic dataset of conditioned medium from CrF‐derived AdMSCs [3]. Samples from patients with active CD were stratified by anti‐TNF exposure, comparing anti‐TNF‐treated patients for ≥ 12 months (n = 3) with anti‐TNF‐naïve patients (n = 5).
Equal numbers of viable passage‐3 AdMSCs (> 90% viability) were seeded for all samples and cultured under serum‐free conditions. Conditioned medium was collected after 9 h, concentrated and analysed by LC–MS/MS as previously described [3]. Further experimental details are provided in the Supporting Information S1: Methods.
Blank or zero intensities were considered non‐detected values. A total of 573 proteins were included in the definitive differential‐abundance analysis. Protein intensities were normalised to the number of viable cells seeded, log2‐transformed and median‐centred.
Differential abundance was assessed using two‐sided Welch's t‐tests without imputation. FDR‐adjusted p < 0.05 was considered statistically significant. Nominal p values were adjusted for multiple testing across all 573 proteins using the Benjamini–Hochberg procedure.
For PCA only, missing values were imputed at one log2 unit below the minimum observed value for each protein. Complete differential‐abundance results are provided in Table S2. Proteomic data are available through ProteomeXchange (PXD030520).
Ex Vivo Infliximab Exposure
CrF explants from anti‐TNF‐naïve patients with CD undergoing surgery were divided into paired untreated and infliximab‐exposed conditions. Infliximab was administered at 5 μg/g tissue (5 mg/kg‐equivalent), while paired control explants from the same patient were maintained under identical conditions without infliximab.
Paired explants were cultured for 24 h at 37°C in a humidified atmosphere containing 5% CO2 in air and subsequently analysed by RT‐qPCR and immunohistochemistry. One explant pair was excluded because it was treated using a different infliximab lot for which visible contamination was subsequently identified in the stored vial. Further experimental details are provided in the Supporting Information S1: Methods.
Immunohistochemistry
CrF samples from both observational and ex vivo cohorts were fixed overnight in 10% neutral‐buffered formalin, paraffin‐embedded, sectioned at 3 μm and mounted onto glass slides. Immunohistochemical staining was performed using a rabbit polyclonal anti‐UCP1 antibody (ab10983; 1:500; Abcam, Cambridge, UK) on a BenchMark ULTRA PLUS platform (Ventana, Roche), following antigen retrieval (Ultra CC1, 64 min) and antibody incubation (48 min).
For quantitative analysis, five non‐overlapping fields were analysed per biological sample using ImageJ version 1.53k (National Institutes of Health, USA). Fields were sampled across the tissue section by a pathologist blinded to the treatment group using a randomised approach to ensure representative tissue coverage and without selection based on UCP1 staining intensity. The percentage of UCP1‐positive area was determined using automatic thresholding, applying the same image‐analysis procedure to all samples without manual adjustment of the threshold.
Values from the five fields were averaged to obtain a single patient‐level value, with the patient/biological sample considered the statistical unit.
Western Blot Analysis of UCP1
CrF samples were lysed in M‐PER supplemented with protease inhibitors (Thermo Fisher Scientific), and protein concentration was determined by BCA assay. Thirty micrograms of total protein per sample was separated by SDS‐PAGE and transferred onto PVDF membranes. Membranes were incubated overnight at 4°C with antibodies against UCP1 (Abcam) and β‐actin (Sigma‐Aldrich), used as a loading control.
Immunoreactive bands were visualised using an iBright 1500 Imaging System (Thermo Fisher Scientific), quantified with ImageJ and UCP1 abundance was normalised to β‐actin. Western blot analysis included CrF samples from anti‐TNF‐naïve (n = 5) and anti‐TNF‐treated (n = 4) patients.
RNA Extraction and Quantitative Real‐Time PCR
Total RNA was extracted from CrF tissue and explants, mature adipocytes and AdMSCs using TRIzol Reagent (Invitrogen, MA, USA). cDNA was synthesised using SuperScript II reverse transcriptase (Invitrogen, MA, USA), and gene expression was quantified by RT‐qPCR using TaqMan Gene Expression Assays on a 7900HT Fast Real‐Time PCR System (Applied Biosystems, MA, USA). 18S ribosomal RNA was used as the reference gene, and relative expression was calculated using the 2‐ΔΔCt method. BAX and BCL2 were additionally analysed as markers of pro‐apoptotic signalling.
Assays are listed in Table S3.
Statistical Analysis
Clinical data were analysed using SPSS v15. Continuous variables are presented as mean ± SD or median (interquartile range), according to their distribution and categorical variables as n (%). Comparisons between anti‐TNF‐treated and anti‐TNF‐naïve patients were performed using Student's t‐test or the Mann–Whitney U test, as appropriate.
Experimental data are presented as mean ± SEM unless otherwise indicated, with the patient considered as the biological and statistical unit. Independent groups were compared using two‐sided Mann–Whitney U tests. Ex vivo untreated and infliximab‐exposed explants were analysed as paired samples using two‐sided paired Student's t‐tests when paired differences were compatible with normality or Wilcoxon matched‐pairs signed‐rank tests otherwise.
For ex vivo RT‐qPCR analyses, comparisons were performed on ΔCt values, whereas 2‐ΔΔCt values were used for graphical presentation. Technical replicates were averaged before analysis.
Proteomic statistics and FDR correction are described in Section 2.5. For other targeted molecular analyses, nominal two‐sided p values are reported without multiple‐testing correction given the exploratory nature and limited sample size of the study. p < 0.05 was considered statistically significant, with p < 0.10 indicated as a trend where specified. GraphPad Prism v8 was used unless otherwise stated.
Results
Clinical Characteristics of the Study Population
Clinical and demographic characteristics of the observational and ex vivo cohorts are summarised in Tables 1 and 2, respectively. The observational cohort comprised 19 anti‐TNF‐naïve and 14 anti‐TNF‐treated patients with CD, including nine receiving adalimumab and five receiving infliximab. The groups showed broadly comparable demographic and clinical characteristics (Table 1).
Serum IL‐1β concentrations were significantly lower in anti‐TNF‐treated patients (p < 0.01), whereas IL‐6 concentrations did not significantly differ between groups. The ex vivo cohort comprised eight anti‐TNF‐naïve patients undergoing surgery for CD (Table 2).
Anti‐TNF Therapy Is Associated With Browning‐, Mitochondrial Biogenesis‐ and Immune‐Related Changes in CrF in CD Patients
Anti‐TNF‐treated patients showed a significantly higher UCP1‐positive area than anti‐TNF‐naïve patients by immunohistochemistry (Figure 1A), together with significantly higher UCP1 protein abundance by Western blotting (n = 4 vs. n = 5; Figure 1B). Whole‐CrF gene‐expression analysis also showed significantly higher TMEM26 and CPT1B expressions in anti‐TNF‐treated patients (Figure 1C).
CrF‐derived AdMSCs from anti‐TNF‐treated patients showed significantly higher UCP1 expression (Figure 1D). Under conditioned‐medium collection conditions, AdMSCs showed no evidence of increased pro‐apoptotic signalling based on the BAX/BCL2 ratio; PCA and volcano‐plot visualisations of the proteomic dataset are shown in Figure S1. Exploratory conditioned‐medium proteomic analysis identified differences in proteins related to inflammatory signalling, adipocyte metabolism, oxidative stress and extracellular matrix/stromal organisation (Figure 1E).
IL6 showed markedly lower abundance in the anti‐TNF‐treated group (log2 fold change −5.35; nominal p = 0.000161; FDR‐adjusted p = 0.092). PRG4 showed higher abundance at the nominal significance level (log2 fold change 3.50; nominal p = 0.033; FDR‐adjusted p = 0.993). No protein remained statistically significant after Benjamini–Hochberg correction across the 573 proteins analysed (Table S2).
Mature adipocytes from anti‐TNF‐treated patients showed significantly higher UCP1 and PPARG expression (Figure 1F), together with higher TFAM and NRF1 expression (Figure 1G). IL4, ADIPOQ and IL13 did not differ significantly between groups (Figure 1H), whereas LEP and TNF expression was significantly lower and IL6 remained unchanged (Figure 1I). Collectively, these findings associate anti‐TNF treatment with a browning‐, mitochondrial biogenesis‐ and immune‐related molecular phenotype in CrF, without demonstrating functional thermogenic or mitochondrial activation.
Ex Vivo Infliximab Exposure Modulates Browning‐, Mitochondrial Biogenesis‐ and Immune‐Related Markers in CrF Explants
To assess whether related molecular changes could be observed following direct infliximab exposure, paired CrF explants from anti‐TNF‐naïve patients were cultured for 24 h with or without infliximab (5 mg/kg‐equivalent). UCP1 immunohistochemistry showed a higher UCP1‐positive area following infliximab exposure (Figure 2A).
RT‐qPCR showed significantly higher expression of UCP1 (p = 0.0023), PPARG (p = 0.0364) and CIDEA (p = 0.0480) following infliximab exposure (Figure 2B). Among mitochondrial biogenesis‐associated markers, TFAM (p = 0.0969) and NRF1 (p = 0.0646) showed upward trends, whereas NAMPT did not differ significantly (Figure 2C). IL13 expression was significantly higher (p = 0.0435), while IL4 and ADIPOQ were unchanged (Figure 2D).
TNF and IL6 showed non‐significant decreases, whereas LEP did not differ significantly (Figure 2E).
The BAX/BCL2 ratio did not differ between paired untreated and infliximab‐exposed explants (n = 4; p = 0.875; Figure S2A), and mean 18S rRNA Ct values were comparable (10.32 ± 0.38 vs. 10.31 ± 0.24; p = 0.875; mean ± SEM). In a preliminary dose‐ranging experiment, all three donors showed increased UCP1, PPARG, ADIPOQ and NAMPT expression at 5 mg/kg‐equivalent, with no consistent further response at 10 or 50 mg/kg‐equivalent (Figure S2B).
Exploratory pooled Pearson analyses suggested positive associations between IL13 and NAMPT (nominal p = 0.0424) and between IL13 and TFAM (nominal p = 0.0425) (Figure S2C–D).
Discussion
We previously described a browning‐associated phenotype in CrF and linked it to microbial translocation and succinate accumulation [3]. Building on these observations, the present study identifies anti‐TNF treatment as a previously unrecognised clinical factor associated with this phenotype. The convergence of higher UCP1 abundance in treated patients with related changes in mature adipocytes and paired ex vivo responses to infliximab supports CrF as a tissue compartment responsive to anti‐TNF exposure.
CrF is a metabolically and immunologically active component of CD [14]. Its cellular complexity enables crosstalk among adipocytes, stromal and immune cells [3, 4, 15]. Zuo et al. independently identified beige adipocytes in hypertrophic MAT and proposed browning as a therapeutic axis [4, 8].
Our previous work associated anti‐TNF therapy with reduced CrF inflammation, restored adipocyte morphology and decreased immune‐cell infiltration [9]. The present findings extend these observations by suggesting that the effects associated with anti‐TNF treatment in CrF may encompass not only inflammatory but also metabolic remodelling. This interpretation is biologically plausible given the established role of TNF‐α in adipocyte differentiation, metabolism and mitochondrial function [16].
Chronic TNF signalling has been associated with impaired mitochondrial function and suppression of thermogenic programmes in adipose tissue, providing a potential framework through which TNF inhibition might favour a browning‐associated phenotype. Overlapping molecular changes following short‐term infliximab exposure in paired ex vivo explants further support a local effect of infliximab within CrF, beyond changes associated with systemic disease control. CrF can influence ileal immune responses and promote angiogenic and fibrotic processes [15].
Adipokines such as adiponectin have also been implicated in intestinal immune homeostasis and have shown protective effects in experimental colitis [17, 18, 19], supporting further investigation of adipocyte–immune crosstalk in CD. The less pro‐inflammatory adipocyte profile observed here is compatible with such immunometabolic interplay. Exploratory associations between IL13 and mitochondrial biogenesis‐associated markers further suggest coordinated immune–metabolic changes although these observations require validation in larger independent cohorts.
The AdMSC conditioned‐medium proteomic analysis provides complementary evidence that anti‐TNF‐associated differences extend beyond mature adipocytes to the stromal compartment. IL6 showed markedly lower abundance and PRG4 showed higher abundance at the nominal significance level; however, neither remained significant after FDR correction. Accordingly, the observed differences in IL6, PPARD, PRG4, FMOD, LUM, COL12A1 and oxidative stress‐related proteins should be regarded as exploratory and hypothesis‐generating.
Extracellular matrix and stromal protein changes may be relevant to intestinal fibrosis and angiogenesis [15]. However, proteins detected in conditioned medium cannot be assumed to be actively secreted and may reflect non‐classical release, extracellular vesicles, or low‐level cellular leakage. Functional conditioned‐medium experiments are needed to determine whether these differences exert paracrine effects on the adipocyte phenotype.
The cellular origin of the observed phenotype is likely to be complex. CrF contains mature adipocytes, adipose‐derived stromal cells, macrophages, endothelial cells and other immune and stromal populations. The findings suggest that the anti‐TNF‐associated remodelling of CrF extends across different stages of the adipocyte lineage.
CrF‐derived AdMSCs from anti‐TNF‐treated patients, which represent adipocyte progenitor cells, showed a browning‐associated molecular phenotype, while mature adipocytes from treated patients displayed a broader beige‐like molecular phenotype characterised by changes in browning‐/thermogenesis‐ and mitochondrial biogenesis‐associated markers. The observation that short‐term infliximab exposure was also followed by browning‐associated changes in whole CrF explants further suggests that this response can occur at the tissue level. Although the cellular contribution to the ex vivo response cannot be resolved in the present study, the concordant findings across AdMSCs, mature adipocytes and intact CrF support the concept that anti‐TNF‐associated remodelling may involve both adipocyte progenitors and differentiated adipocytes.
Several limitations should be acknowledged. The observational cohort was relatively small and cross‐sectional, and residual confounding related to disease characteristics or treatment history cannot be excluded despite broadly comparable clinical parameters between groups. The proteomic analysis involved a particularly small exploratory subgroup, and the targeted molecular analyses included multiple comparisons without formal adjustment.
Browning and mitochondrial phenotypes were assessed predominantly through molecular markers rather than functional measurements, and no dedicated histomorphometric assessment of beige adipocyte morphology was performed. The ex vivo study lacked an isotype control and a dedicated functional tissue‐viability assay; BAX/BCL2 expression and stable 18S Ct values provide only indirect evidence of comparable tissue integrity. The short‐term explant model and limited dose‐ranging experiment also preclude conclusions regarding temporal or dose‐dependent effects.
Finally, CrF heterogeneity could not be resolved. Larger cohorts and functional studies incorporating mitochondrial respiration, mitochondrial mass and complementary metabolic readouts will be required to validate and extend these findings.
In conclusion, anti‐TNF treatment is associated with browning‐, mitochondrial biogenesis‐ and immune‐related molecular remodelling in CrF from patients with CD, while paired ex vivo experiments support a browning‐/thermogenesis‐associated response following direct infliximab exposure. Together, these findings identify CrF as a previously under‐recognised tissue compartment responsive to anti‐TNF exposure and suggest that this remodelling involves both adipocyte progenitors and mature adipocytes. Whether these molecular changes translate into functional metabolic remodelling or contribute to clinical outcomes remains to be established.
Author Contributions
D.M.‐F., I.G., L.C., A.T. and R.S. performed the experiments. C.C., A.C., M.M., M.d.C.L., M.J.M., M.Ma. and J.F. participated in study population selection and human sample processing. M.G. and M.Ga. performed UCP1 immunohistochemistry.
D.M.‐F., I.G., M.M., R.S., J.M. and C.S. contributed to data interpretation and critically reviewed the manuscript. D.M.‐F., I.G. and C.S. conceived the study and discussed the data. D.M.‐F., I.G. and C.S. wrote the manuscript.
C.S. and J.M. are the guarantors of this work.
Funding
This study was supported by the III Basic GETECCU–Galapagos Grant 2023 (awarded to C.S.), the Ministerio de Ciencia, Innovación y Universidades (grant PID2023‐146315OB‐100; awarded to C.S. and J.M.) and the Generalitat de Catalunya (grant PROD2025‐00210; awarded to C.S. and J.M.).
Ethics Statement
This study was approved by the ethics committee of the University Hospital Joan XXIII (CEIM 233/2024) and Hospital Vall d’Hebrón PR(CS)383/2021.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information