Patients and colonic biopsy

Colonic biopsies from 8 young UC patients and 8 elderly UC patients were performed in this study, and the biopsy site was assessed as endoscopic Mayo score ranging from 2–3. We defined patients diagnosed with UC after 60 years of age as the elderly group, whereas patients with a diagnosis of UC before 40 years of age were defined as the young group. The diagnosis of UC was fulfilled according to the third European Crohn’s and Colitis Organization (ECCO) consensus guidelines [15]. Meanwhile, healthy controls who met the same criteria for age were included, and normal colonic biopsy specimens were collected from these controls. All included cases had no history of comorbidities including malignant tumors, severe hepatic and renal insufficiency, severe cardiovascular and cerebrovascular diseases. Clinical data and samples were collected after patients signed written informed consent.

RNA-sequencing analysis and identification of differential genes

Colonic mucosal tissues from 3 young and 3 elderly patients with UC were collected for RNA extraction using RNeasy Plus Mini Kit (Qiagen), followed by RNA-sequencing (RNA-seq) analysis. Library preparation was performed using the NEBNext® UltraTM RNA Library Prep Kit on an Illumina NovaSeq 6000 platform. HISAT2 (v.2.0.5) was used to map the clean reads to each gene, with raw data normalized to fragments Per kilobase of exon model per million mapped fragments (FPKM) for subsequent analyses. Differentially expressed genes (DEGs) were identified with the DESeq2 (v1.20.0) package with a cutoff of P < 0.05 and |log(fold change)|> 1.0, where Log2(FC) > 1.0 indicated upregulated genes, and log2(FC) < − 1.0 indicated downregulated genes. The DEGs between the young and elderly group were illustrated with a volcano plot generated via the R package “ggpubr”. Functional enrichment analyses of the DEGs were performed with the clusterProfiler package, with the terms of Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) and gene set enrichment analysis (GSEA) identified with a cutoff of P < 0.05.

Identification of mitochondria-related DEGs and functional enrichment analysis

The Integrated Mitochondrial Protein Index (IMPI) of the MitoMiner database (http:// mitominer.mrc-mbu.cam.ac.uk/) provides 1626 human mitochondria-related genes. Overlapping genes among DEGs from RNA-seq and mitochondrial-related genes from the MitoMiner database were defined as mitochondria-related DEGs (mDEGs). A heatmap of mDEGs expression in adult- and elderly-onset UC patients was drawn based on log2-transformed FPKM values from RNA-seq using the "pheatmap" package in R. Different color indicated the expression of the genes in different samples. As mentioned above, the list of mDEGs was used for GO and KEGG enrichment analyses via the clusterProfiler package.

Immune infiltration analysis

Based on the RNA-seq results, we further used the CIBERSORT algorithm to evaluate immune cell infiltration and its relationship with mDEGs in adult- and elderly-onset UC. The analysis was performed using the "CIBERSORT" package in R [16]. The FPKM data from RNA-seq was used for the CIBERSORT algorithm. The validated leukocyte gene signature matrix (LM22) was applied to quantify 22 phenotypes of human infiltrating immune cells (https://www.nature.com/articles/nmeth.3337#MOESM207) [17]. The proportions of 22 immune cells in each tissue sample were then calculated. As a reference expression profile of human tissue, the LM22 signature matrix defined 22 infiltrating immune cell components.

A heatmap was generated using R package “pheatmap” to illustrate the proportions of different immune cells in each sample. Comparative analysis of immune cell profiles between adult- and elderly-onset UC tissues was conducted using the Kruskal test, and the results were visualized as the boxplot using R package “ggplot2”. The stacked bar plot figure was generated using R package “ggplot2” to illustrate the composition of 22 immune cell subsets in each sample. We subsequently evaluated the correlations between the proportions of infiltrating immune cells and the 30 mDEGs expression in the adult- and elderly-onset UC group respectively using Spearman’s correlation coefficient with a significance threshold of P < 0.05. Similarly, the correlation between immune cells was also analyzed. The results were visualized as a correlation heatmap using R package “corrplot”.

Protein–Protein Interaction (PPI) networks and hub gene identification

A protein–protein interaction (PPI) network of the mDEGs was constructed via STRING (version 12.0), followed by visualization through Cytoscape (version 3.8.2). Then, cytoHubba and MCODE plugins were subsequently used for the identification of gene clusters. The hub gene was identified based on multiple factors including the fold change and significance of differential expression, the degree of connectivity in the PPI network, and subsequent validation of expression in tissue samples.

Young and aged DSS-colitis mouse model

Male C57BL/6N mice aged 6–8 weeks and 20–24 months, purchased from Vital River Laboratory Animal Technology Co. Ltd. (Beijing, China), were maintained in a specific pathogen-free animal laboratory at a constant temperature of 22℃ with a fixed 12-h light–dark cycle. After 1 week of adaptive feeding, the young and aged mice were utilized to establish a colitis model. As we previously reported [3], colitis was induced in mice via the oral delivery of 2% dextran sodium sulfate (DSS) (MP Biomedicals, USA) in drinking water for 7 days. At the end of the treatments, mice were euthanized by cervical dislocation. Deaths of mice not associated with experimental intervention were excluded. All animal procedures were performed in accordance with the guiding principles for the care and use of laboratory animals approved by the Animal Care Committee of Peking Union Medical College Hospital.

During DSS administration, body weight, stool features and hematochezia were recorded daily, and the disease activity index (DAI) was evaluated as previously reported [18], as shown in the Supplementary Material. The mice were sacrificed on day 8, and the length of the colon was measured. The colon tissues were sampled using a Swiss roll technique [19] and fixed in 10% formalin for 48 h and then embedded in paraffin. 4-μm slides were stained with hematoxylin and eosin (H&E) for histopathological analysis to assess the pathological score of experimental colitis as previously described [18] (Supplementary Material).

Quantitative real-time PCR

The colonic samples collected from patients with UC and DSS-colitis mice were immersed in the RNAlater (Thermofisher) for RNA extraction. Among these 30 mDEGs, we further selected 14 mDEGs (HK3, DACT2, SUGCT, ECHDC3, ALDH1L1, FASTKD1, PYCR1, HMGCS2, TIMM23B, TOMM20L, MTERF1, CLUH, PLIN5, P2RY12) that were reported in previous studies (not limited to mitochondrial or inflammation-related studies) and 8 highly reported genes related to NAD(P) + /NAD(P)H homeostasis (ME1, ME2), mitochondrial respiratory complex (Prkaa2, Pdha, NDUFS4, SDHA), mitochondrial biogenesis (PGC1α, MFF) for qRT-PCR. Total RNA was extracted and first-strand cDNA synthesis was performed using FastPure® Tissue Total RNA Isolation Kit (Vazyme) and the PrimeScriptTMRT Reagent Kit with gDNA Eraser (Takara). Then, qRT-PCR was performed with TB Green Premix Ex TaqTMII (Takara). The primer sequences are summarized in Supplementary Table 1 and GAPDH was used as an internal reference. The relative expression of target genes was calculated by the 2−△△Ct method. All PCR reactions were conducted in triplicate.

Transmission electron microscopy

Fresh distal colon tissues were fixed with 2.5% glutaraldehyde for 4 h, followed by dehydration in increasing concentrations of ethanol and propylene oxide. The samples were finally embedded in epoxy resin. Ultrathin sections were prepared following standard protocols and mitochondrial morphology was observed with transmission electron microscope (TEM 1400 plus, Japan). Five colonocytes were randomly selected from each sample and the number of mitochondria in each colonocyte was counted manually by 2 investigators blinded to the groups. The average number of mitochondria in each colonocyte was recorded.

Assessment of tissue ATP content

Approximately 20 mg of fresh colon tissue was obtained to measure the ATP concentration. According to Enhanced ATP Assay Kit (Beyotime), colon tissue was homogenized and the supernatant was collected to determine the ATP level. The detection solution was added to a 96-well plate and incubated at room temperature for 5 min. The supernatants were then added to the wells and mixed quickly before the luminescence signals were determined via Varioskan Flash (Thermofisher). Total ATP levels were subsequently calculated on the basis of the luminescence signals.

Cell culture and treatment

The human colon cancer cell line HCT-116 was obtained from the American Type Culture Collection (ATCC) and was cultured in RPMI 1640 supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin under standard culture conditions (5% CO2, 37 °C). HCT116 cells were seeded into 6-well culture plates for 24 h to reach 70–80% confluence, after which the cells were separated into two groups: the non-targeting scrambled control siRNA (NC) group and the ALDH1L1-siRNA (siALDH1L1) group. These reagents were transfected into cells via Lipofectamine™ RNAiMAX (Thermofisher scientific) according to the manufacturer’s protocol. Following siRNA transfection and 24 h incubation for 24 h, the medium was changed and the cells were treated with TNFα (100 ng/ml) for 24h and collected for RNA and protein extraction. The targeted sequences for the siRNA experiments were as follows: siALDH1L1: 5’- GACCCTCATTCACGGAGATAA −3’; negative control (NC) siRNA: 5’-TTCTCCGAACGTGTCACGT-3’. The efficacy of gene knockdown was validated through qRT-PCR and Western blot analysis (Supplementary Fig. 1).

Tumor necrosis factor (TNF) is a major pathogenic mediator and therapeutic target in IBD, and TNF-α can be used to treat intestinal epithelial cells to induce cellular inflammatory processes, as previously mentioned [20, 21].

Mitochondrial ROS detection

Under exogenous TNF-α stimulation, cells treated with siRNAs were assessed for reactive oxygen species (ROS) production with MitoSOX Red (Invitrogen, Cat.no. M36008), a fluorescent superoxide indicator that selectively targets mitochondria. HCT116 cells were stained with 5 μM MitoSOX Red for 30 min at 37 °C in a humidified environment with 5% CO2. The proportion of MitoSOX-positive cells and the fluorescence intensity were detected via flow cytometry after the cells were washed and suspended.

Measurement of Oxygen Consumption Rate (OCR)

HCT116 cells transfected with negative non-targeting scrambled control siRNA or ALDH1L1-siRNA were seeded in an XF 24-well cell culture microplate (Seahorse Bioscience, Copenhagen, Denmark) at a density of (6 ~ 7 × 104) and incubated overnight, followed by TNFα (100 ng/ml) treatment for 6 h. The hydration of the sensor probe plate and preparation of the mitochondrial stress test kit (Cat.103015–100, Agilent) were then performed according to the manufacturer’s recommendations. The final concentrations of oligomycin, the uncoupler carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone (FCCP), the complex III inhibitor rotenone, and antimycin A (Rot/AA) were respectively 1.5 μM, 0.5 μM, 0.5 μM, respectively. The oxygen consumption rate (OCR) was measured using Seahorse XFe24 extracellular flux analyzer (Agilent, Santa Clara, CA, USA).

Western blotting

Total protein was extracted from tissue and HCT116 cells treated with siRNA with RIPA buffer (C1053, APPLYGEN, China), and the protein concentration was determined via BCA protein assay (A53226, Pierce). Loading buffer was added to protein samples, followed by denaturation in a water bath at 100 °C for 10 min. Samples containing the same amount of protein (30 μg) were separated with NuPAGE Bis–Tris gels (Thermofisher). The separated proteins were subsequently transferred to polyvinylidene difluoride (PVDF) membranes. The PVDF membranes were blocked in 5% bovine serum albumin at room temperature and shaken gently for 1 h. Then, the membranes were incubated at 4 °C overnight with primary antibodies diluted with 5% BSA. These primary antibodies included anti-ALDH1L1 (17,390–1-AP, Proteintech, 1:2000), anti-NLRP3 (19,771, Proteintech, 1:1000) and anti-IL1β (12,242, CST, 1:1000) antibodies. The membranes were subsequently washed with Tris-buffered saline-containing tween three times and incubated with appropriate horseradish peroxidase (HRP)-conjugated secondary antibodies at room temperature for 1 h. The protein bands were detected by a chemiluminescence detection system, and GAPDH was used as the internal standard.

Statistical analysis

Experimental data are expressed as the means ± standard errors of the mean (SEMs). The significance of data among the 4 groups was tested using one-way ANOVA test followed by Tukey’s multiple comparisons test for normally distributed data, and the Kruskal–Wallis test for nonnormally distributed data. P < 0.05 was considered statistically significant. GraphPad Prism (version 10.0, San Diego, USA) was used for statistical tests, and R (3.6.2, R Foundation for Statistical Computing, Vienna, Austria) was used for data processing and visualization as described above.