Mitochondrial dysfunction in patients with celiac disease: A multidimensional study

Abstract

BACKGROUND

Celiac disease (CeD) is an autoimmune enteropathy triggered by gluten, but its pathogenesis remains unclear. One possible mechanism is mitochondrial dysfunction; however, no structural or functional evidence of mitochondrial damage in the intestinal epithelium of CeD patients has been reported.

AIM

To identify whether there is any structural or functional mitochondrial damage in the duodenal mucosa of active CeD patients.

METHODS

In this pilot study, 18 individuals with active CeD were enrolled alongside 18 age- and sex-matched healthy controls. Markers of metabolic energy (adenosine triphosphate, L-lactate), apoptosis (caspase 9), and oxidative stress (malondialdehyde, 4-hydroxynonenal, superoxide dismutase, glutathione peroxidase) were quantified using commercial-grade biochemical assay kits. Transmission electron microscopy was utilized to examine ultrastructural variations of mitochondria in intestinal epithelial cells. The expression levels of mitochondrial variations-related proteins (MFN1, MFN2, OPA1, DRP1, FIS1) were assessed using western blotting in duodenal mucosal samples. Quantitative real-time polymerase chain reaction was utilized to calculate mitochondrial DNA copy number in duodenal tissue samples. Clinical hematological and biochemical parameters were also assessed for exploratory correlation analyses.

RESULTS

Transmission electron microscopy revealed notable ultrastructural changes in the mitochondria of intestinal epithelial cells in CeD patients, including swelling of the organelles, disorganized or disrupted cristae, and vacuolization. In comparison to healthy subjects, individuals with CeD showed pronounced alterations in intestinal mucosal mitochondrial energy balance, characterized by decreased adenosine triphosphate production alongside increased L-lactate concentrations. Superoxide dismutase and glutathione peroxidase activities were significantly reduced, and malondialdehyde and 4-hydroxynonenal levels were significantly elevated, indicating a dramatic change towards oxidative stress. Western blot analysis simultaneously identified changed expression of mitochondrial dynamics proteins in CeD patients with a significant decrease in the expression of MFN1 and MFN2 (fusion proteins) and a significant increase in the expression of DRP1, p-DRP1 (active isoform of DRP1), and FIS1 (fission proteins). The copy number of the mitochondrial DNA using quantitative real-time polymerase chain reaction was also considerably lower among CeD patients than their healthy counterparts (P < 0.05). Correlation analysis suggests potential associations between mitochondrial parameters and hematological abnormalities.

CONCLUSION

This work is the first comprehensive evidence of structural and functional mitochondrial damage in the duodenal mucosa of individuals with CeD. The results offer new experimental data on the study of the pathogenesis of CeD and a useful theoretical basis to develop new therapeutic options that will target the mitochondria.

Key Words: Celiac disease; Mitochondria; Oxidative stress; Energy metabolism; Mitochondrial dynamics proteins

Core Tip: This study provides the first multidimensional evidence of structural and functional mitochondrial damage in the duodenal mucosa of patients with active celiac disease, including ultrastructural abnormalities, impaired energy metabolism, increased oxidative stress, imbalanced mitochondrial dynamics, and reduced mitochondrial DNA copy number. These findings offer new experimental insights into the pathogenesis of celiac disease.

INTRODUCTION

Celiac disease (CeD), also known as gluten-sensitive enteropathy, is an immune-mediated disorder triggered by gluten intake in genetically susceptible individuals. Key histopathological changes include villous atrophy, enlarged crypts, and infiltration of intraepithelial lymphocytes within the small intestinal lining[1,2]. Epidemiological studies have shown that CeD is a significant public health problem worldwide, affecting almost one percent of the global population[3,4]. A lifelong gluten-free diet (GFD) helps in the alleviation of most symptoms[1], but the underlying pathophysiological mechanisms of CeD are still not fully understood. This incomplete and insufficient understanding not only leads to suboptimal and variable clinical responses to a GFD in a subset of patients but is also associated with a higher incidence and increased risk of complications such as anemia, osteoporosis, and intestinal T-cell lymphoma, all of which severely and profoundly impact quality of life and long-term prognosis in individuals with CeD.

The pathogenesis of CeD is thought to result from the interplay of genetic, environmental, and immune factors[5]. The HLA-DQ2 and HLA-DQ8 genotypes are strongly associated with susceptibility to disease among these[6]. The peptides that result from the digestion of gluten by enzymes in the intestine activate the immune system in the gut and have been implicated in autoimmune responses[7]. There is also increasing evidence that disturbances in metabolic homeostasis and the functional integrity of intestinal epithelial cells may contribute to the initiation and progression of immune-mediated injury[8,9]. The intestinal epithelium, characterized by high energy demands and rapid turnover, relies on a continuous and efficient supply of adenosine triphosphate (ATP) to sustain its homeostatic functions[10]. Mitochondria function as central energy-metabolic hubs in eukaryotic cells, serving as primary powerhouses that generate ATP while also regulating critical biological processes such as redox homeostasis, calcium balance, and apoptosis control[11]. It has been previously reported that mitochondria have a significant role in the damage of the epithelial barrier and chronic inflammation in intestinal diseases like inflammatory bowel disease (IBD)[12,13]. In the rapidly renewing intestinal epithelium, mitochondrial integrity is essential for maintaining tight junctions, supporting nutrient absorption, and resisting inflammatory stress. Proper mitochondrial function is especially important in the fast-renewing intestinal epithelium, where it supports tight junctions, nutrient uptake, and resistance to inflammatory stress. Damaged mitochondria often generate excessive reactive oxygen species (ROS), show reduced OXPHOS activity, and release pro-apoptotic molecules[14,15]. This vulnerability likely stems from the lack of histone protection and limited DNA repair mechanisms in mitochondrial DNA (mtDNA)[16].

Several studies have reported an increase in oxidative stress and alterations in the quantity of mtDNA in patients with CeD[17,18]. It has been demonstrated that peptides, especially the immunodominant 33-mer α-gliadin fragment, induce ROS production and endoplasmic reticulum (ER) stress in intestinal epithelial cells, which leads to mitochondrial damage[19]. Patients with CeD consuming gluten often show enhanced lipid peroxidation, an elevated oxidized glutathione to reduced glutathione ratio, and decreased protein thiol concentrations, reflecting intracellular redox dysregulation[20-22]. Moreover, elevated serum lactate levels represent a shift in metabolism to glycolysis, which could be a marker of reduced mitochondrial respiratory function[23]. However, it remains unclear whether these abnormalities are a direct result of mitochondrial abnormalities in intestinal epithelial cells or are linked to the typical intestinal pathology of CeD.

The objective of this study was to systematically review mitochondrial injury in patients with CeD and explore the mechanisms of this injury. Transmission electron microscopy was used to evaluate the ultrastructural alterations of mitochondria in the duodenal epithelial cells in CeD patients and controls. We also evaluated alterations in energy metabolism, oxidative stress, mitochondrial dynamics, and mtDNA copy number of duodenal tissues. These results offer vital experimental data for the involvement of mitochondrial dysfunction in the pathogenesis of CeD and indicate that targeting mitochondria could be a future therapeutic approach in CeD.

MATERIALS AND METHODS

Study participants

Patients with CeD who visited the Department of Gastroenterology at the People’s Hospital of Xinjiang Uygur Autonomous Region between December 2024 and December 2025 were enrolled. All patients were diagnosed according to the 2017 World Gastroenterology Organization Global Guidelines for Celiac Disease[24], were aged over 18 years, and were matched 1:1 with healthy subjects of comparable age (± 3 years) and sex. All enrolled healthy subjects, recruited from the Health Examination Center of the same hospital, had no CeD-related clinical or gastrointestinal symptoms. Endoscopy was performed solely to evaluate gastrointestinal health status. It was confirmed that these subjects had normal endoscopic and histological findings and tested negative for serum anti-tissue transglutaminase immunoglobulin A antibodies (anti-tTG IgA < 20 CU). Tissue samples were collected after communication with and informed consent from all healthy subjects. All participants provided written informed consent after consultation with their attending physicians. Following informed consent, we collected general and clinical data, established complete records, and obtained intestinal tissue samples for subsequent analyses. All collected tissue specimens were properly preserved and stored at -80 °C until further processing.

Exclusion criteria were as follows: (1) Presence of other active gastrointestinal diseases, including: Biopsy-confirmed IBD; giardia lamblia infection or bacterial/viral enteritis (within the preceding 4 weeks); lactose intolerance (confirmed by hydrogen breath test or genetic testing); (2) Receipt of the following treatments within the preceding 3 months: Glucocorticoids, immunosuppressants, biological agents, antibiotics, probiotics, or prebiotics; (3) Use of the following medications within 2 weeks prior to biopsy: Drugs known to directly damage mitochondria (e.g., statins, metformin, antiretroviral drugs, valproic acid); nonsteroidal anti-inflammatory drugs, due to their potential to cause small intestinal mucosal injury; (4) History of upper gastrointestinal surgery (e.g., gastrectomy, gastroenterostomy, or other procedures that may affect duodenal anatomy or perfusion); and (5) Pregnancy or lactation.

At biopsy, all enrolled CeD patients presented endoscopic features of villous atrophy and had histopathological Marsh grades ≥ II (the 18 CeD patients in this study had Marsh grade ≥ III), all consistent with active CeD.

This study was approved by the Ethics Committee of Xinjiang Uygur Autonomous Region People’s Hospital (approval No. KY2024120189). All participating patients provided written informed consent. The hospital ethics committee waived the requirement for informed consent for the use of other clinical data. This study was conducted in accordance with the STROBE guidelines.

Sample collection and processing

Blood sample collection: Fasting venous blood samples were collected. Hemoglobin, mean corpuscular volume, albumin, serum ferritin, and 25-hydroxyvitamin D [25(OH)D] levels were measured on the day of collection.

Tissue specimen collection: All tissue specimens were obtained during esophagogastroduodenoscopy under direct visualization using standardized biopsy forceps. Mucosal samples were taken from duodenal lesions and placed in 10% neutral buffered formalin for routine histopathological examination. Pathological grading was performed according to the Marsh-Oberhuber classification for CeD[25]. Additionally, two mucosal tissue specimens were collected from the duodenal bulb and distal duodenum of each patient with CeD. For each control subject, two normal mucosal tissue specimens were similarly obtained. During collection, at least three biopsy samples were taken from each site using sterile biopsy forceps to ensure sufficient material for analysis. After collection, samples were immediately snap-frozen in liquid nitrogen and transferred to -80 °C for storage. Alternatively, samples were fixed in glutaraldehyde solution and stored at 0 °C to prevent tissue degradation. Preparation of tissue homogenates: Tissue homogenates were prepared using mechanical homogenization. Frozen tissues were added to pre-chilled phosphate-buffered saline and thoroughly homogenized on ice. The homogenate was centrifuged and the clear supernatant subjected to biochemical tests. All procedures were performed under aseptic conditions to minimize external contamination.

Sample pretreatment

Tissue samples used in this study were processed according to the specific assay requirements. The pretreatment steps prior to detection are detailed in Supplementary material.

Transmission electron microscopy observation

Transmission electron microscopy was used to assess mitochondrial morphological changes. Experimental samples were selected from three patients in the case group and three subjects in the control group. This subset analysis was performed as an exploratory morphological assessment. The three CeD patients were selected as representative cases based on the following criteria: (1) Elevated serum anti-tTG-IgA titers (> 300 CU); (2) Marsh IIIa or IIIb histology; and (3) Presence of gastrointestinal symptoms at enrollment, to maximize the likelihood of detecting mitochondrial abnormalities in this exploratory analysis. Freshly obtained tissue samples were cut into 1 mm³ pieces and immediately placed in 2.5% glutaraldehyde fixative at 4 °C overnight. Subsequently, samples were post-fixed in 1% osmium tetroxide for 1 hour. Dehydration was carried out in a graded ethanol series, and then infiltrated with epoxy resin. Later followed by curing at 60 °C for 48 hours to generate embedded blocks. The embedded blocks were cut into 70 nm thick sections with an ultramicrotome and placed on copper grids. The sections were double stained with uranyl acetate and lead citrate, and examined by a transmission electron microscope. The following observation parameters were used: Acceleration voltage of 80 kV and magnification range of (5000-20000) ×. Mitochondrial morphological changes like swelling, disruption of cristae and vacuolization were noted and photographs were taken for further analysis.

Biochemical assays

Biochemical assays were analyzed by commercial kits and all were performed strictly according to the manufacturers’ instructions. The metabolism energy indicators including L-lactate and ATP were calculated using the lactate dehydrogenase assay and luciferase method, respectively. Oxidative stress markers 4-hydroxynonenal (4-HNE), malondialdehyde (MDA), glutathione peroxidase (GSH-Px) and superoxide dismutase (SOD), were determined using the thiobarbituric acid method, xanthine oxidase method, enzyme-linked immunosorbent assay, and DTNB assay, respectively. The activity of caspase 9 was assessed spectrophotometrically using absorbance of p-nitroaniline produced by the particular substrate Ac-LEHD-pNA at 405 nm. All measurements were taken using the MFlexStation 3 multi-mode microplate reader, which was calibrated before use to ensure accuracy and reliability.

Western blot analysis

Three CeD patients with high Marsh grade and high serum anti-tTG-IgA levels, and with prominent gastrointestinal symptoms, and three matched healthy controls were selected for the study. RIPA lysis buffer with PMSF and phosphatase inhibitors was used to extract total protein and determine the protein concentration by using the BCA protein assay kit. Next, 40 μg of protein for each well was loaded, denatured at 95 °C for 5 minutes, and isolated by sodium-dodecyl sulfate gel electrophoresis using 5% stacking gel and 12% or 15% resolving gel as per the MW of the protein of interest. The isolated protein bands were then transferred onto polyvinylidene fluoride membranes in wet form at 1.5 A for 300 seconds for < 30 kDa proteins, including FIS1, and at 1.5 A for 420 seconds for 30-150 kDa proteins (e.g., β-actin, DRP1, p-DRP1). When transfer was completed, the membranes were blocked with 5% non-fat milk in Tris-buffered saline with Tween for 2 hours, and then primary antibodies (DRP1, p-DRP1, FIS1, OPA1, MFN1, MFN2) were added and incubated at 4 °C overnight. The following day, membranes were washed with Tris-buffered saline with Tween and incubated with secondary antibodies at room temperature for 1 hour. Signals were visualized using ECL chemiluminescent substrate and imaged. Band gray values were analyzed using ImageJ software and normalized to β-actin.

Quantitative real-time polymerase chain reaction analysis

Relative mtDNA copy number in duodenal tissues was determined using quantitative real-time polymerase chain reaction (qPCR). Genomic DNA was extracted from frozen tissues using the Tiangen Genomic DNA Extraction Kit (Tiangen Biotech, Catalog No. DP304) following the manufacturer’s instructions. Briefly, approximately 20-30 mg of tissue was lysed with Buffer GA and proteinase K, purified using a silica membrane spin column, and eluted with TE buffer. DNA concentration and purity were measured using a NanoDrop spectrophotometer, with an A260/A280 ratio between 1.8 and 2.0 considered acceptable.

qPCR reactions were performed on an ABI ViiA™ 7 Real-Time PCR System using AceQ qPCR SYBR Green Master Mix (Vazyme, Catalog No. Q111). Amplification targets included the mitochondrial gene*MT-ND1*and the single-copy nuclear gene β-actin: MT-ND1: Forward 5’-CATACCCATGGCCAACCTCCTACTCCTCATT-3’, reverse 5’-TGCTCGCAGTGCGCCGATCAGGGCGTAGT-3’ (amplicon length: 406 bp); β-actin: Forward 5’-AGTCCTCTCCCAAGTCCACACA-3’, reverse 5’-CACGAAGGCTCATCATTCAAAA-3’ (amplicon length: 130 bp).

The 20 μL reaction mixture contained 4 μL of genomic DNA template (diluted to approximately 10 ng/μL), 0.4 μL each of forward and reverse primers (10 μM), 10 μL of 2 × SYBR Green Master Mix, and nuclease-free water. The amplification protocol was as follows: Initial denaturation at 95 °C for 5 minutes, followed by 40 cycles of 95 °C for 15 seconds and 60 °C for 30 seconds. The melting curve analysis was conducted after amplification to confirm specificity. The 2^-∆∆Ct method (ΔCt = Ct_MT-ND1 - Ct_β-actin) was used to calculate relative mtDNA copy number. The results were presented as relative fold change to the mean value of healthy control group.

Statistical analysis

GraphPad Prism 10.0 was used to analyze data. Normality was determined by using the Shapiro-Wilk test, and homogeneity of variance was determined using Levene’s test. For data meeting parametric assumptions, results were expressed as mean ± SD, and comparisons between paired groups were analyzed using paired t-tests. For non-normally distributed data, results were expressed as median (interquartile range) and analyzed using the Wilcoxon signed-rank test. Correlation analyses were performed using Pearson or Spearman correlation, depending on data distribution. To account for multiple comparisons, false discovery rate (FDR) correction using the Benjamini-Hochberg method was applied to all correlation analyses, with a threshold of q < 0.10 considered statistically significant for these exploratory analyses. All tests were two-sided, and a P value < 0.05 was considered statistically significant for primary analyses, while FDR-adjusted P values were reported for correlation analyses.

RESULTS

Baseline characteristics of study participants

A total of 18 confirmed patients with CeD and 18 age- and sex-matched healthy subjects were enrolled. As shown in Table 1, the two groups were well balanced for basic demographic characteristics. The mean age of the CeD group was 51.33 ± 8.89 years, compared with 50.11 ± 8.20 years in the control group (P = 0.112). Female participants accounted for 11 cases (61.1%) in both groups (P = 1.000). Body mass index also showed no statistically significant difference between the two groups (CeD: 21.59 ± 3.86; control: 22.41 ± 2.84; P = 0.4365). Among the 18 CeD patients, the median symptom duration was 15 months (interquartile range: 1.75-48.5 months). All CeD patients exhibited modified Marsh type III lesions on small intestinal mucosal pathology, including 15 cases (83.3%) with type IIIa and 3 cases (16.7%) with type IIIb. Serum anti-tTG-IgA antibody titers in CeD patients were markedly elevated, with a median value of 479.3 U/mL (interquartile range: 292.4-3021 U/mL).

Table 1 Demographic and disease characteristics of celiac disease patients and healthy controls, mean ± SD/median (interquartile range)/n (%).

Characteristic	Parameter	CeD group (n = 18)	Control group (n = 18)	P value
Demographic characteristics	Age (years)	51.33 ± 8.89	50.11 ± 8.20	0.11241
	Female	11 (61.1)	11 (61.1)	12
	BMI (kg/m2)	21.59 ± 3.86	22.41 ± 2.843	0.43651
	Smoking history	6 (33.3)	7 (38.9)	12
	Alcohol consumption history	0	1 (5.6)	12
Disease characteristics	Symptom duration (months)	15 (1.75-48.5)	-	-
	Modified Marsh grade
	IIIa	15 (83.3)	0	-
	IIIb	3 (16.7)	0	-
	IIIc	0	0	-
	Serum anti-tTG-IgA titer (CU)	479.3 (292.4-3021)	1.9 (lower detection limit)	-

¹P values were calculated using the paired t-test (for age and body mass index).

²P values were calculated using the McNemar’s exact test (for sex).

CeD: Celiac disease; BMI: Body mass index; tTG: Tissue transglutaminase; IgA: Immunoglobulin A.

Blood test parameters in CeD patients

Routine blood test and biochemical parameters in CeD patients are presented in Table 2. Patients exhibited typical nutritional deficiencies associated with malabsorption. The mean hemoglobin level in the CeD cohort was 129.4 ± 22.6 g/L, falling near the lower boundary of the normal reference intervals (male: 130-175 g/L; female: 115-150 g/L). A subset of patients had hemoglobin values below the sex-specific lower limits. Mean corpuscular volume was 84.62 ± 11.93 fL, within the normal range (82-100 fL). Mean serum albumin concentration was 41.84 ± 4.47 g/L, within the normal reference range (35-50 g/L), though values clustered near the lower limit. Iron stores were reduced, with a median serum ferritin level of 12.9 μg/L (interquartile range: 6.86-61.22), below the normal reference range (11-306.8 μg/L), consistent with iron deficiency. Furthermore, all patients exhibited vitamin D insufficiency or deficiency. Mean serum 25(OH)D concentration was 14.29 ± 5.97 nmol/L, far below the sufficient level (> 50 nmol/L). This falls into the category of severe deficiency (< 25 nmol/L) according to Endocrine Society guidelines[26].

Table 2 Comparison of blood test results in celiac disease patients with reference ranges, mean ± SD/median (interquartile range).

Variable	CeD group (n = 18)	Reference range
Hemoglobin (g/L)	129.4 ± 22.6	130-175 (male), 115-150 (female)
MCV (fL)	84.62 ± 11.93	82-100
Albumin (g/L)	41.84 ± 4.468	35-50
Serum ferritin (μg/L)	12.9 (6.86-61.22)	11-306.8
25(OH)D (nmol/L)	14.3 (8.7-19.6)	> 50 (sufficient), 25-50 (insufficient), < 25 (deficient)
Erythropoietin (mIU/mL)	12.76 (9.54-24.07)	1.48-31.88

Reference ranges are based on the clinical laboratory standards of our hospital; vitamin D status is defined according to the Endocrine Society guidelines[26]. CeD: Celiac disease; MCV: Mean corpuscular volume; 25(OH)D: 25-hydroxyvitamin D.

Mitochondrial ultrastructural abnormalities: Transmission electron microscopy was performed on duodenal intestinal epithelial cells from three pairs of matched samples to directly compare morphological differences in mitochondria between CeD patients and healthy controls. Transmission electron microscopy revealed that, in the healthy control group, mitochondria exhibited regular morphology, appearing predominantly elliptical or short rod-shaped. They displayed clear, dense, and well-organized cristae structures, intact outer and inner membranes, and uniformly distributed matrix electron density (Figure 1). However, the CeD group showed considerable and uniform changes in the structure of mitochondria. The main changes observed were mitochondrial swelling and vacuolization, with most mitochondria being markedly enlarged and exhibiting irregular shapes, and in some instances, the presence of obvious internal vacuoles; cristae disruption, with the loss of cristae from mitochondria, some cristae being sparse, shortened and disorganized, or even completely absent, a phenomenon known as “bare cristae”; and membrane integrity loss, with some mitochondria showing a blurred or disrupted outer mitochondrial membrane (Figure 1). Such characteristic transformations suggest that the intestinal mucosa of the CeD patients is severely damaged and dysfunctional in terms of mitochondrial structure and function.

Figure 1 Transmission electron microscopy images of duodenal intestinal epithelial cells from celiac disease patients and healthy controls (n = 3 per group). A-F: Healthy control group (A-C: × 12000; scale bars = 2 μm. D-F: × 20000; scale bars = 1 μm); G-L: Celiac disease group (G-I: × 12000; scale bars = 2 μm. J-L: × 20000; scale bars = 1 μm). Blue arrows indicate swollen mitochondrial matrix; pink arrows indicate cristae disintegration, disruption, or loss; yellow arrows indicate pleomorphic mitochondria. HC: Healthy control; CeD: Celiac disease.

Quantitative analysis of mitochondrial number and morphological parameters: To further quantify the extent of mitochondrial damage, we performed mitochondrial counting and morphometric measurements in duodenal intestinal epithelial cells from three CeD patients and three healthy controls. Five random fields of view at 12000 × magnification were analyzed per sample. Mitochondrial morphometry data are provided in Supplementary Table 1.

Results of biochemical assays

Mitochondrial energy metabolism impairment and oxidative stress in CeD mucosa: Comparison of energy metabolism parameters between CeD patients and healthy controls revealed that ATP content in the intestinal mucosa of CeD patients was significantly lower than that in the control group (P = 0.0028; Figure 2A), indicating impaired mitochondrial oxidative phosphorylation. Conversely, L-lactate levels were significantly elevated in the CeD group (P = 0.0017; Figure 2B). Our results suggest a compensatory shift in cellular energy metabolism from oxidative phosphorylation toward glycolysis.

Figure 2 Changes in mitochondrial biochemical indicators, apoptosis, and mitochondrial DNA copy number. A: Adenosine triphosphate content; B: L-lactate levels; C: 4-hydroxynonenal content; D: Malondialdehyde content; E: Superoxide dismutase activity; F: Glutathione peroxidase activity; G: Mitochondrial DNA copy number; H: Caspase 9 activity. Data are shown as mean ± SD or median (IQR) as appropriate. P values were determined by paired t-test or Wilcoxon signed-rank test. Group sizes: n = 15 per group. ATP: Adenosine triphosphate; 4-HNE: 4-hydroxynonenal; MDA: Malondialdehyde; SOD: Superoxide dismutase; GSH-PX: Glutathione peroxidase; mtDNA: Mitochondrial DNA; HC: Healthy control; CeD: Celiac disease.

Regarding oxidative stress, levels of 4-HNE (P = 0.0002) and MDA (P = 0.0001) were significantly higher in the CeD group than in the control group (Figure 2C and D). Meanwhile, antioxidant enzyme system function was impaired, as evidenced by significantly reduced activities of SOD (P = 0.0043) and GSH-Px (P < 0.0001) in the CeD group (Figure 2E and F). qPCR results showed that mtDNA copy number in the intestinal mucosa of the CeD group was significantly lower than that in the control group (P = 0.0097; Figure 2G), suggesting a possible reduction in mitochondrial number or genomic damage. Sustained oxidative stress and energy depletion can trigger apoptosis. We assessed the activity of caspase 9, a key executor of the mitochondrial apoptotic pathway. The results showed that caspase 9 activity in the intestinal mucosa of CeD patients was significantly higher than that in the healthy control group (P < 0.0001, Figure 2H), indicating that the mitochondria-dependent apoptotic pathway was activated in the intestinal mucosal injury of CeD.

Imbalanced expression of mitochondrial dynamics-related proteins: Western blot analysis showed imbalanced expression of mitochondrial dynamics-related proteins. For fusion proteins, the corresponding western blot bands are shown in Figure 3A. OPA1 showed a decreasing trend but did not reach statistical significance (P = 0.0621; Figure 3B), while the expression levels of MFN1 (P = 0.0133; Figure 3C) and MFN2 (P = 0.0335; Figure 3D) were significantly downregulated in the CeD group. For fission proteins, the western blot bands are shown in Figure 3E. The expression of p-DRP1 (P = 0.0393; Figure 3F), DRP1 (P = 0.0103; Figure 3G), and FIS1 (P = 0.0049; Figure 3H) were significantly upregulated.

Figure 3 Expression of mitochondrial fusion and fission proteins (n = 3 per group for western blotting). A: Western blot analysis showed that the fusion proteins OPA1, MFN1, and MFN2 were expressed in the intestinal mucosa of the celiac disease group; B-D: Quantitative analysis of the fusion proteins OPA1 (B), MFN1 (C), and MFN2 (D); E: Western blot analysis showed that the fission-promoting proteins DRP1, p-DRP1 (the activated form), and FIS1 were expressed in the intestinal mucosa of the celiac disease group; F-H: Quantitative analysis of the fission-promoting proteins p-DRP1 (F), DRP1 (G), and FIS1 (H). Bar charts represent relative protein expression normalized to β-actin (n = 3 per group). Data are presented as mean ± SD. β-actin was used as a loading control. HC: Healthy control; CeD: Celiac disease.

Correlation between mitochondrial damage indicators and their association with clinical features

To investigate the relationships among different aspects of mitochondrial dysfunction and their associations with clinical characteristics, we performed Spearman correlation analysis on key indicators derived from the quantitative data presented in Figure 4. All correlation findings should be considered hypothesis-generating and require validation in larger cohorts. ATP content was positively correlated with SOD activity (r = 0.853, P < 0.001, FDR-adjusted P = 0.006; Figure 4A). L-lactate levels showed a nominal positive correlation with caspase 9 activity (r = 0.546, P = 0.038, FDR-adjusted P = 0.045; Figure 4B).

Figure 4 Correlation between mitochondrial damage indicators and their association with clinical features. Scatter plots show the following significant correlations (after Benjamini-Hochberg false discovery rate correction, q < 0.10). A: Adenosine triphosphate levels with superoxide dismutase activity; B: L-lactate levels with caspase 9 activity; C: 4-hydroxynonenal concentration with serum 25-hydroxyvitamin D levels; D: L-lactate levels with erythropoietin; E: L-lactate levels with serum ferritin; F: Caspase 9 activity with erythropoietin. The Spearman correlation coefficient (r) and P value are indicated in each plot, with the solid line representing the linear regression fit (n = 15 per group). SOD: Superoxide dismutase; ATP: Adenosine triphosphate; 4-HNE: 4-hydroxynonenal; 25(OH)D: 25-hydroxyvitamin D.

Regarding associations between mitochondrial parameters and clinical hematological indicators, 4-HNE concentration showed a positive correlation with serum 25(OH)D levels (r = 0.663, P = 0.009, FDR-adjusted P = 0.027; Figure 4C). L-lactate exhibited a nominal negative correlation with erythropoietin (EPO) levels (r = -0.588, P = 0.04, FDR-adjusted P = 0.048; Figure 4D) and a nominal positive correlation with ferritin levels (r = 0.566, P = 0.05, FDR-adjusted P = 0.050; Figure 4E). Caspase 9 activity showed a significant negative correlation with EPO levels (r = -0.615, P = 0.03, FDR-adjusted P = 0.045; Figure 4F). No notable correlations were observed between ATP or 4-HNE and hemoglobin, albumin, or other clinical parameters (all P > 0.05). A full summary of all exploratory correlations, including raw and FDR-adjusted P values, is provided in Supplementary Table 2.

DISCUSSION

In IBD, mitochondrial dysfunction in intestinal epithelial cells has been linked to barrier disruption and persistent inflammation[12,13,27]. However, whether intestinal epithelial cell-specific mitochondrial damage occurs in CeD, and how such damage relates to the characteristic pathological features of villous atrophy and crypt hyperplasia, remains poorly understood. To address this gap, the present study systematically evaluated the structural and functional integrity of mitochondria in the duodenal mucosa of patients with active CeD. We identified severe, multi-level mitochondrial dysfunction in the CeD duodenal mucosa, characterized by structural damage, energy metabolism impairment, exacerbated oxidative stress, imbalanced mitochondrial dynamics, and activation of the intrinsic apoptotic pathway. These findings establish that mitochondrial dysfunction is a prominent pathological feature of the CeD intestinal epithelium.

In the present study, transmission electron microscopy revealed pronounced mitochondrial damage in the duodenal epithelium of CeD patients, including swelling, disrupted cristae, and vacuolization, accompanied by a reduced mitochondrial number per field. We report these ultrastructural changes for the first time in CeD. Although transmission electron microscopy revealed striking qualitative abnormalities (swelling, cristae disruption) in CeD mitochondria, quantitative morphometric parameters (area, length, width, aspect ratio, circularity) did not differ significantly between groups. This discrepancy likely reflects the very small sample size (n = 3 per group) and high inter-individual variability. Additionally, two-dimensional morphometric measures may not fully capture the organization of three-dimensional cristae, and these parameters do not adequately quantify qualitative features such as cristae disorganization. Mitochondrial ultrastructural damage has been shown to lead to OXPHOS impairment, thereby exacerbating intestinal inflammation and mucosal injury[28-30]. Furthermore, in this study, we observed that mitochondrial ultrastructural damage was consistent with a significant decrease in tissue ATP levels and compensatory accumulation of L-lactate, indicating that impaired OXPHOS had forced intestinal epithelial cells to shift toward inefficient glycolysis, i.e., metabolic reprogramming. In the highly energy-dependent intestinal epithelium, such a metabolic shift may compromise ATP-dependent processes required for nutrient transport and barrier maintenance. The accompanying lactate accumulation could also contribute to microenvironmental acidification and activation of pro-inflammatory pathways, although this remains speculative pending direct measurement in future studies.

In particular, all the CeD patients in this study had a severe vitamin D deficiency (mean 14.29 nmol/L) and low iron stores (median serum ferritin 12.9 μg/L). Considering that the vitamin D receptor takes part in regulating mitochondrial gene transcription and OXPHOS function[31], and while iron serves as an essential cofactor for the mitochondrial respiratory chain complexes I-III[32,33], the profound deficiencies of these micronutrients are likely to collectively impair mitochondrial energy metabolism. This is an interpretive issue since it is hard to distinguish the effects of CeD per se from secondary effects caused by nutritional deficiencies. We cannot determine whether the mitochondrial dysfunction is a cause or result of the inflammatory environment in CeD in our cross-sectional data. Longitudinal studies of mitochondrial parameters from diagnosis to gluten withdrawal and nutrition repletion are needed. The intestinal epithelial cell functions in nutrient transport and barrier integrity are dependent on mitochondrial ATP generation[34]. Thus, mitochondrial dysfunction can result in greater malabsorption, which will exacerbate micronutrient deficiencies, furthering the cycle of mitochondrial damage. Earlier research has shown that vitamin D possesses antioxidant and immune-regulating properties, whereas its deficiency may compromise the intestinal mucosal barrier against oxidative damage[35,36]. In this study, there was a strong positive correlation between the concentration of 4-HNE and serum 25(OH)D level. This association was surprising but could represent a compensatory mechanism in response to mobilization of the remaining vitamin D when oxidative stress is very high. Nevertheless, due to the limited sample size, the observed finding should be considered preliminary and validated through larger-scale studies.

The intestinal epithelial cells of patients with CeD exhibited markedly elevated levels of lipid peroxidation products (including MDA and 4-HNE). In contrast, significantly decreased activities of antioxidant enzymes (SOD and GSH-Px) were observed in CeD intestinal epithelial cells. These observations indicate increased oxidative stress and compromised antioxidant protection in the intestinal epithelium of CeD patients[37-40]. Additionally, excessive accumulation of MDA and 4-HNE may directly interfere with mitochondrial membrane integrity and respiratory chain function[41-44]. Together with the significant reduction in mtDNA copy number observed in our cohort, this could be a vicious feedback cycle: MtDNA is not protected by histones and is very close to the respiratory chain, and thus is very susceptible to damage by ROS[16]; damage to mtDNA may lead to a reduction in expression of respiratory chain subunits, which may increase the leakage of electrons and ROS production[45].

Changes in mitochondrial morphology and cristae are primarily regulated by proteins involved in mitochondrial fission and fusion and are also influenced by cell type, energy demand, and metabolic status[46,47]. In this study, we found that the expression of fusion-promoting proteins (MFN1, MFN2) was significantly downregulated, while the expression of fission-promoting proteins (DRP1, FIS1) was significantly upregulated in the duodenal mucosa of CeD patients. These findings indicate a shift in mitochondrial dynamics balance toward excessive fission (P < 0.05). This is consistent with the abnormal mitochondrial morphology observed under transmission electron microscopy, characterized by scattered, swollen mitochondria with reduced or absent cristae structures. These findings indicate that disturbances in mitochondrial fission-fusion can contribute to the deterioration of mitochondrial structural integrity.

Moreover, excessive mitochondrial fission is associated not only with morphological alterations but also with several functional defects. In particular, fragmented mitochondria demonstrate reduced OXPHOS efficiency, ultimately leading to reduced ATP synthesis[48], and mitochondrial fragmentation is also a crucial signal for mitophagy[49] (damaged mitochondria clearance). But once it is beyond the clearance limit, the damaged mitochondria continue to generate ROS, which further increases the damage to mtDNA and dysfunction in the energy metabolism. This results in activation of the mitochondria-dependent intrinsic apoptotic pathway and loss of intestinal epithelial cells and villous atrophy. Villous atrophy is the characteristic pathological finding of CeD and it is indicative of an imbalance between epithelial apoptosis and regeneration. The observed enhancement in caspase 9 activity detected in this study suggests the presence of mitochondrial outer membrane permeabilization, a defining event in intrinsic apoptotic signaling. However, interpretation at the molecular level cannot be conclusively determined from tissue-level caspase 9 activity data. These findings support the hypothesis that mitochondrial dysfunction may represent a central mechanism connecting autoimmune injury to the histopathological manifestations of CeD.

Correlation analysis revealed a positive relationship between ATP levels and SOD activity, implying coordinated functioning of cellular energy metabolism and antioxidant defense mechanisms. The ATP-dependent glutathione regeneration system is important for maintaining the stability of the SOD and antioxidant defense is important for maintaining the function of the respiratory chain[50,51]. Chronic gluten-induced immune activation in CeD causes mitochondria to be under simultaneous metabolic and oxidative stress. Normally, production of energy and antioxidant protection would rise in parallel to preserve the homeostasis. In this study, however, ATP and SOD were found to decrease simultaneously, indicating that a decrease in energy metabolism would lead to a decrease in antioxidant defense and vice versa, thus forming a vicious circle. These results show that the mitochondrial dysfunction in intestinal epithelial cells in CeD patients is a systemic disorder, and not restricted to energy metabolism or oxidative stress.

In addition, the positive association between L-lactate and caspase 9 activity indicates that the glycolytic switch in CeD intestinal mucosa is initially likely to serve as a mechanism to overcome the lack of energy, but can eventually lead to apoptosis. Furthermore, there was a negative correlation between L-lactate and EPO and between caspase 9 and EPO, and a positive correlation between L-lactate and ferritin. These associations link mitochondrial dysfunction to anemia, which is a common extraintestinal symptom of CeD. During chronic inflammation, decreased EPO production and impaired iron metabolism may overlap with the pathways (e.g., hypoxia-inducible factor-1 alpha and nuclear factor kappa B) that are upstream of mitochondrial oxidative stress and metabolic abnormalities[52-54]. However, these preliminary associations should be considered hypothesis-generating given the small sample size and borderline significance levels, and require validation in larger cohorts. The concentration of 4-HNE showed a positive correlation with serum 25(OH)D levels. Although the underlying reason remains unclear, studies have indicated that severe oxidative stress can interfere with vitamin D metabolism, and vitamin D deficiency can weaken the antioxidant capacity of the intestinal epithelium. The concentration of 4-HNE showed a positive correlation with serum 25(OH)D levels[55]. Furthermore, deficiency of vitamin D and its receptor can lead to mitochondrial dysfunction in epithelial cells, resulting in reduced energy production, compromised barrier integrity, and impaired epithelial cell development[56,57]. Whether vitamin D supplementation could exert mitochondrioprotective effects in CeD is an interesting hypothesis that warrants dedicated interventional studies, but it cannot be inferred from our observational data.

This study has several limitations. First, the cross-sectional design precludes any inference regarding the temporal sequence or causality of the observed mitochondrial changes. Second, the study was conducted at a single center with highly selected CeD patients (all Marsh III), which limits generalizability. Third, duodenal mucosal biopsies contain heterogeneous cell populations; some molecular signals may originate from infiltrating immune cells rather than epithelial mitochondria. However, the transmission electron microscopy findings confirm that epithelial cell mitochondria are directly affected. Finally, the exploratory correlation analyses should be interpreted with caution given the borderline significance of several associations and the small sample size; these findings are hypothesis-generating and require validation in larger, multicenter cohorts.

CONCLUSION

This pilot study is the first to show structural and functional mitochondrial damage in the duodenal mucosa of patients with active CeD. Our findings suggest that mitochondrial dysfunction may play a significant role in CeD, but larger multicenter studies are needed to confirm this. We also found reduced SOD and GSH-Px activity and a shift toward mitochondrial fission. These changes raise the possibility that targeting mitochondria (for example, with MitoQ or nicotinamide riboside) could supplement a GFD in future preclinical and clinical studies. Any therapeutic potential, however, remains speculative for now.

ACKNOWLEDGEMENTS

The authors thank all patients and healthy volunteers who participated in this study for their invaluable cooperation. We are grateful to the staff of the Endoscopy Center and the Department of Clinical Laboratory at the People’s Hospital of Xinjiang Uygur Autonomous Region for their technical assistance in sample collection and biochemical measurements. We also acknowledge the support of the Xinjiang Clinical Research Center for Digestive Diseases. Finally, we thank the members of our research team for their helpful discussions and critical review of the manuscript.