Background & Summary

Inflammatory bowel disease (IBD), comprising ulcerative colitis (UC) and Crohn’s disease (CD), is a persistent immune-related inflammatory ailment of the gastrointestinal system1,2,3, involving in complicated interactions of genetic factors4, gut microbiota5,6,7,8,9, immune imbalance10 and environmental risk factors11,12. Thus, a precise comprehension of the pathogenic mechanisms underlying IBD is crucial for the development of effective treatment strategies13. Genome-wide screening studies indicated that multiple susceptibility loci were associated with the pathogenesis of IBD, such as interleukin-10 (IL-10)14. The serum and inflamed intestinal mucosa levels of IL-10 in patients with UC and CD were demonstrated to be lower than that of healthy individuals15,16. Diverse functional abnormalities of IL-10 lead to the occurrence of IBD. For example, first, IL-10 contributes to the protection of intestinal epithelium, with its role being dependent on the autocrine loop involving CD4+ T cells17. Second, IL-10 can further induce the production of IL-10 in macrophages and train the polarization of M2 macrophages by Fcγ-receptor (FcγR) signaling to exert anti-inflammatory effects18. Third, IL-10 also can display an anti-inflammatory profile through distinct metabolism pathways, such as sphingolipid metabolism and metabolic reprogramming involving in the pathogenesis of intestinal inflammation19,20.

Although the administration of IL-10 has been tested as a potential therapy for IBD in preclinical studies, recombinant IL-10 (rIL-10) has not demonstrated significant efficacy for the treatment of IBD16. Multiple elements possibly give rise to the failure of rIL-10 in clinical trials, including the short half-life of IL-10 in the serum21,22, the weak targeting of IL-10 action at sites of intestinal inflammation and the production of IFNγ associated with the IL-10-sensitized proinflammatory reaction22,23. Nevertheless, other factors that probably influence the efficacy of IL-10 for the treatment of IBD have yet to be identified. Consequently, it is imperative to conduct a more in-depth investigation into the specific signaling pathways associated with IL-10 in IBD. Furthermore, the utilization of IL-10−/− mice is essential for advancing fundamental research in IBD.

The colitis observed in IL-10−/− mice closely resembles the colitis associated with human CD and the similarity is likely attributable to the deficiency of IL-10 that results in inappropriate immune responses of the murine gut to commensal intestinal microbiota. Therefore, this model is frequently employed as an mouse model for the study of CD24. Prior research has conducted comparative analyses of bulk RNA sequencing data between IL-10−/− mice and IBD patients, revealing a high degree of similarity. In this study, we replicated this approach (data not shown), further corroborating the validity of IL-10−/− mice as a quintessential model for IBD research25. IL-10−/− mice exhibit a gradual development of spontaneous enterocolitis over the age of 4 weeks26,27,28, which is associated with the infiltration of inflammatory cytokines, deposition of fibrin-like material and IgA, and the atypical expression of major histocompatibility complex class II (MHCII) molecules in the intestinal epithelium28. Several prior investigations have conducted RNA or proteomic sequencing on the colonic tissue of IL-10−/− mice, with the objective of elucidating the pivotal genes or proteins potentially implicated in the pathogenesis of colitis in these mice. Yet, our study aims to concurrently sequence the RNA and proteome of the colon from the same cohort of wild-type (WT) and IL-10−/− mice. This approach enhances the consistency in identifying differentially expressed genes (DEGs) and proteins (DEPs) implicated in the pathogenesis of colitis, thereby providing further evidence supporting the role of IL-10 deficiency in the development of this condition.

This study enables us to perform a detailed comparative analysis of RNA and proteome sequencing results from the intestines of IL-10−/− mice in subsequent experiments. This analysis will allow for a comprehensive screening of the signaling pathways implicated in IL-10 deficiency-induced colitis (Fig. 1), aiming to further elucidate the potential clinical significance of overlapping target genes.

Fig. 1
figure 1

Workflow of colonic samples preparation and data analysis. (a) The establishment of IL-10−/− chronic colitis mice model and acquirement of the colons from WT and IL-10−/− mice. Two groups of mice were both executed at 24 weeks of age (n = 3). (b) The process of colonic tissue preparation for proteomic and transcriptomic sequencing. Subsequently, the integrated analysis of proteomic and transcriptomic data were performed.

Full size image

Collectively, our data have provided a more precise characterization of the immunomodulatory role of IL-10 in the pathogenesis of colitis. Our findings have enhanced the understanding of the mechanisms underlying the progression of IBD.

Methods

Evaluation of genotype and colitis in IL-10−/− mice

The experimental mice were allocated into two cohorts, comprising WT and the IL-10−/− mice group, with each group consisting of three mice. The male C57BL/6 J mice with IL-10 deficiency (4 weeks old) were acquired from Shanghai Model Organisms Center of China. Male WT and IL-10−/− mice were selected by genotype identification at approximately 4 weeks old (Fig. 2a). The IL-10−/− mice exhibited a range of symptoms and pathological changes associated with intestinal inflammation, including weight loss (Fig. 2b), stool consistency and hematochezia, which were scored to represent the disease activity index (DAI) (Fig. 2c). Afterward, the mice were euthanized through cervical dislocation, and their colons were extracted. The length of colons from mice in each group was measured and statistically compared (Fig. 2d). The IL-10−/− mice colons exhibited a notable infiltration of inflammatory cells across various structural layers of the intestinal wall compared to that of WT mice by hematoxylin-eosin (HE) staining (Fig. 2e). The remaining colonic tissues were placed in liquid nitrogen for subsequent transcriptomic and proteomic assays. All mice research proposals were licensed by Ethics Committee of West China Hospital (approval number: 20240918004).

Fig. 2
figure 2

Assessment of chronic colitis in IL-10−/− mice. (a) PCR-based identification of WT and IL-10−/− mice. (b) Changes in body weight of two groups of mice during 24 weeks. (c) Changes in disease activity index (DAI) of two groups of mice during 24 weeks. (d) Comparison of colon length between two groups of mice at 24 weeks of age. (e) Histological scoring of colons in two groups of mice. n = 3, All values represent the mean with SD; a two-sided t test. *p < 0.05.

Full size image

RNA extraction and quality inspection

Standard protocols were followed to extract total RNA from colonic tissue using TRIzol reagent (Thermo, USA). Specifically, 100 mg of colonic tissue was placed in a 1.5 mL Eppendorf (EP) tube, followed by the addition of 1 mL of TRIzol. The mixture was incubated for 10 minutes at room temperature (25 °C). Subsequently, each tube received 0.2 mL of chloroform, and the contents were shaken vigorously for 10 seconds. The samples were then left at room temperature for 5 more minutes before being centrifuged at 12,000 g for 10 minutes at 4 °C. The liquid phase was gently moved to a fresh EP tube for additional processing. Following this, the supernatant was gently taken out of the tube, and the RNA precipitate was left to dry at room temperature before being dissolved in RNase-free water. Samples of each group were analyzed for RNA integrity and DNA contamination by gel electrophoresis. RNA purity was measured based on OD260/OD280 and OD260/OD230 ratio by Nanodrop. Then, initial libraries were constructed using at least 1 µg total RNA. The NEBNext UltraTM RNA Library Prep Kit for Illumina (NEB, USA) was used to create libraries, following the manufacturer’s instructions, with index codes added for sequence assignment. Next, mRNA was purified with magnetic beads attached to poly-T oligos. The mRNA was then randomly broken into multiple fragments using divalent cations in a 6 × NEB fragmentation buffer. Using mRNA as a template, one strand of cDNA was synthesized in segments with the help of M-MuLV reverse transcriptase. The synthesis of another cDNA strand was completed utilizing four types of deoxynucleoside triphosphate (dNTP) in DNA polymerase I system. The newly synthesized complementary DNA (cDNA) required an end repair process, during which cDNA fragments of 250–300 base pairs were initially isolated using AMPure XP beads (Beckman Coulter, Beverly, USA) as well as later amplified through polymerase chain reaction (PCR). Finally, PCR produces were re-purified to produce the final library. A fluorometer was initially used for quantification, and quantitative realtime PCR (qRT-PCR) was employed to confirm the final library met the experimental standards.

Transcriptomics sequencing

After library quality assessment by qRT-PCR, library was pooled in accordance with the effective concentration and downstream data volume required for Illumina sequencing and the library needed to meet the standards required for analysis. The amplification process during sequencing required multiple components, including DNA polymerase, joint primers and four types of fluorescently labelled dNTP. During the synthesis of the complementary chains in the sequencing cluster, the fluorescent labeled dNTPs involved in this process released corresponding fluorescence and the intensity of fluorescence was converted into distinct sequencing peaks. A diverse range of sequence information was obtained for subsequent data analysis.

Protein sample extraction

Mouse colon tissue of each group previously stored in liquid nitrogen tank was taken out into a mortar successively, ground with a small amount of liquid nitrogen, lysed with SDT lysis buffer containing 4% sodium dodecyl sulfate (SDS) as well as 100 mM Tris-HCl (pH 7.6), and then transferred into EP tubes. The mixture in each EP tube was homogenized with a MP homogenizer (6.0 M/S, 30 s, twice), boiled for 3 minutes, sonicated (40 W, 5 s, 6 times), and then centrifuged at 12000 g for 8 minutes. After filtering with a 0.22 μm filter screen, the supernatant was collected, and the protein concentration of each sample was measured using the bicinchoninic acid (BCA) protein assay kit (P0012, Beyotime) (Table 1). For subsequent experiments, samples were distributed and stored at −80 °C.

Table 1 Protein concentration for mass spectrometry.
Full size table

SDS-PAGE

To ensure rigor, we also used Bradford’s method to determine protein concentration. Each supernatant sample containing 20 µg protein was taken into a tube, mixed with 6 × sample buffer, boiled for 5 minutes and segregated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). After electrophoresis, the gels were stained with Coomassie Brilliant Blue to reveal protein bands. The protein quality met the requirements of proteomics. Total protein content met the requirements for two or more experiments (Fig. 3).

Fig. 3
figure 3

Protein gel stained with Coomassie Brilliant Blue.

Full size image

FASP Method

During the FASP method, 100 µg of protein from each sample’s supernatant was moved to an EP tube. Afterward, it was mixed with dithiothreitol (DTT) to reach a final concentration of 120 mM, heated at boiling for 3 minutes, and then cooled to room temperature. Each sample was then added to 200 µL of UA buffer, which contains 8 M urea and 150 mM Tris-HCl at a pH of 8.5. The samples were then transferred into a 30 kDa ultrafiltration centrifuge tube (Sartorius, VN01H22) and subjected to centrifugation with gradual acceleration at 12,500 g for 25 minutes. Each concentrated solution was combined with 100 μL of iodoacetic acid (IAA) buffer at a concentration of 100 mM. The blend was stirred at 600 rpm for 1 minute, kept at normal temperature without light exposure for 30 minutes, and centrifuged at 12,500 g for 15 minutes. The process from applying UA buffer to the final step needed to be repeated twice. Next, each sample was mixed with 100 μL of 50 mM NH4HCO3 (Sigma, A6141-25G) and centrifuged; these procedures also needed to be repeated twice. Then, 40 μL trypsin (Promega, V5117) was added to a new collection tube which was shaken gently for 1 minute before incubating at 37 °C for 18 hours. Then, each sample was combined with 40 μL of NH4HCO3 solution, centrifuged once more, and the supernatant was gathered. The resulting peptides were purified using a C18 Cartridge, freeze-dried, dissolved in 40 μL of formic acid (0.1%), and ultimately measured by ultraviolet spectroscopy.

Mass spectrometry analysis

Mass spectrometry (MS) technology is integral to proteomic research, particularly in the precise identification and quantification of proteins. In this study, each peptide sample was fractionated using a nanoliter-scale NanoElute system (Bruker, Bremen, Germany), which was interfaced with a timsTOF Pro mass spectrometer (Bruker, Bremen, Germany) featuring a CaptiveSpray ion source. The column was balanced with an aqueous solution containing 0.1% formic acid and an acetonitrile solution with 0.1% formic acid. For impurity removal and separation, samples were automatically introduced to the analytical column (IonOpticks, Australia) at a flow rate of 300 nL/min. Each peptide sample underwent chromatographic separation to facilitate mass spectrometry detection and enhance detection sensitivity. The samples were later examined using mass spectrometry with the timsTOF Pro in PASEF mode. The operational parameters for the timsTOF Pro were as follows: an analysis duration of 90 minutes, positive ion detection mode, a parent ion scanning range of 100–1700 m/z, an ion mobility range of 1/K0 from 0.75 to 1.4 V.s/cm², an ion accumulation or release time of 100 ms, an ion utilization rate of 100%, a capillary voltage set at 1500 V, a drying gas flow rate of 3 L/min, and a drying temperature of 180 °C.

Transcriptomic data processing

The core roles of transcriptome sequencing are mainly used to screen DEGs and perform relevant analysis. Firstly, genes were compared between the two groups of colon samples. Then, the DEGs associated with IL-10 deficiency were efficaciously filtered and the biological significance of them was identified. The analysis process involved quality control (QC), type comparison, quantitative expression, identification of significant differences, and enrichment analysis. Furthermore, personalized transcriptomic analysis mainly focused on inflammation related biological functions and signaling pathways.

Proteomics data processing

For proteomics, Maxquant is the premier algorithm used for both qualitative and quantitative analysis. Currently, the non-labelled quantitative computation of proteomics data using label-free algorithm in Maxquant has become one of the major applications of this algorithm29,30.

In our study, we used a high-resolution mass spectrometer to collect raw data using data-dependent acquisition (DDA); then, we applied a non-labelled quantitative method based on the integration of MS1 data. The Maxquant software calculates the intensity values of the results by first identifying the peptide feature followed by graphical integration. The database used in this study was Uniprot_MusMusculus_17082_20210928_swissprot (http://www.uniprot.org). Software MaxQuant 1.6.17.0 (https://maxquant.net/) was used for database searches. Label free quantitation (LFQ)30 is considered to be an accurate measurement of the mass and abundance of proteins in each group of samples based on mass spectrometry. Additional pertinent parameters and descriptions encompassed the following: a maximum of two missed cleavages, fixed modifications such as carbamidomethylation of cysteine residues, flexible modifications such as methionine oxidation and acetylation at the protein’s N-terminus, a target-reverse database pattern, the inclusion of contaminants set to true, a peptide FDR of 0.01 or below, along with a protein FDR of 0.01 or below.

Data Records

Under the accession number PXD05180631, the raw proteomics data and corresponding result files have been archived in the ProteomeXchange Consortium via the iProX partner repository32,33. Concurrently, the transcriptomic dataset, comprising the raw data file, gene counts file, and FPKM (Fragments Per Kilobase of exon model per Million mapped fragments) file, has been deposited in the Gene Expression Omnibus (GEO) under the accession number GSE26429434. Tables 2, 3 presented comprehensive data pertaining to the two sample groups.

Table 2 Proteomic datasets stored in iProX.
Full size table
Table 3 Transcriptomic datasets stored in GEO.
Full size table

Technical Validation

RNA-seq quality control

The image data generated from high-throughput sequencing of the sequenced fragments was processed using CASAVA base calling to convert it into sequence data (reads), resulting in the acquisition of files in FASTQ format. It was essential for the accuracy of follow-up analysis to filter the raw data by removing three types of invalid data, including reads with adapters, reads containing N representing indeterminate bases, and low-quality reads. Clean reads for subsequent analysis were summarized in Table 4. Figure 4a illustrated how principal component analysis (PCA) employed dimensionality reduction to demonstrate the differences between sample groups and the uniformity within them. Researchers identified 635 differentially expressed genes (DEGs), with 273 genes showing increased expression and 362 showing decreased expression (Fig. 4b). Venn diagram analysis demonstrated that 1,282 and 753 genes were expressed only in the WT or IL-10 groups, respectively, while 11,479 genes were co-expressed in both groups (Fig. 4c). In addition, Pearson’s test was used to evaluate correlation between samples (Fig. 4d). Thus, functional differences identified for the co-expressed genes could be re-analyzed in downstream validation experiments, which could better explain the pivotal roles involved in IL-10-related pathways in colitis.

Table 4 Quality control of transcriptomic data.
Full size table
Fig. 4
figure 4

Quality control of proteomic and transcriptomic data. (a) Transcriptomic principal component analysis (PCA) of the six samples in two groups. (b) Differential expression genes (DEGs) between two groups. (c) Venn diagram of overlapping DEGs. (d) Pearson correlation between samples. Two groups contained six samples, and every two samples were compared. (e) The distribution of peptide counts and protein molecular weight. (f) The cluster heatmap of DEPs. (g) The Pearson’s test of proteomics. (h) Distribution of quantitative proteomic data before missing values imputation in each sample. Y-axis represented quantitative values exponentially. (i) Distribution of quantitative proteomic data after missing values imputation in each sample following normalization and logarithmic transformation. (j,k) The intensity of peptides and proteins.

Full size image

Proteomic quality control

The 4D label-free technique adds an ion mobility dimension to the traditional label-free approach, which comprises evaluation indicators with three dimensions, including ionic intensity, retention time and mass/charge ratio (m/z). This enables better separation of the peptide segments and consequently, greater sensitivity for the timsTOF Pro, making it more suitable for high-throughput proteomics research. The peptide count and protein molecular weight distribution were shown in Fig. 4e. The cluster heatmap of DEPs was shown in Fig. 4f. As in the transcriptome, Pearson’s test of proteomics also demonstrated a high degree of similarity between the samples (Fig. 4g). Furthermore, the distribution of value for each sample was comparable (Fig. 4h). The data following standardization was presented in Fig. 4i. The intensity of peptides and proteins was shown in Fig. 4j,k. Finally, we computed the coefficients for two samples within each group and observed a significant degree of reproducibility (Fig. 5a–f).

Fig. 5
figure 5

Comparison of quantitative information between two samples within each group. In the same group, every two replicates were compared.

Full size image