Identification of metabolism-related subtypes and feature genes in Crohn’s disease

doi:10.21203/rs.3.rs-3982266/v1

Download PDF

Research Article

Identification of metabolism-related subtypes and feature genes in Crohn’s disease

https://doi.org/10.21203/rs.3.rs-3982266/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Crohn's disease is a chronic inflammatory disorder of the gastrointestinal tract, capable of causing disruptions in gastrointestinal function. However, the precise etiology of the disease remains unclear at present. This study aims to analyses metabolism-related signatures to identify feature genes. To investigate potential treatment targets to improve patient prognosis. We downloaded Crohn's disease (CD) datasets from the GEO database, we identified three subtypes (MCA, MCB, MCC) through consensus cluster, as well as the correlation with immune infiltration. We selected the MCA cluster for weighted gene correlation network analysis (WGCNA). After selecting the strongly correlated red module, we identified 20 hub genes. We conducted KEGG, GO functional enrichment analyses. Further analysis using LASSO, SVM and protein-protein interaction (PPI) methods narrowed down to 11 hub genes. Taking the intersection with differentially expressed genes, we finally obtained 4 feature genes: ACTN1, MMP2, THY1, and ZCCHC24. We constructed a nomogram and used the DGIdb database for candidate drug prediction. Molecular docking validation was performed using Schrödinger software to calculate and visualize the interactions between MMP2 and 19 drugs. Through the analysis of metabolism-related signatures, we have identified 4 feature genes and used them to construct a nomogram. This model is expected to offer novel insights for clinical treatment.

Crohn’s disease

Metabolic subtypes

Immune infiltration

Molecular docking

Crohn's disease is a chronic inflammatory condition affecting the gastrointestinal tract, classified along with ulcerative colitis as inflammatory bowel disease (IBD). It is a progressive disorder that disrupts gastrointestinal function, with the most commonly affected the terminal ileum and colon (Torres et al. 2017). However, the exact cause of Crohn's disease is not yet clear. It is believed to result from the complex interplay between genetic susceptibility, environmental factors, and alterations in the intestinal microbiota. This interaction leads to abnormal mucosal immune responses and impaired epithelial barrier function, contributing to the development of this autoimmune disease (Bernstein 2015). Therefore, revealing the complexity behind this interaction may provide crucial insights into the pathogenesis of CD and offer potential targets for therapeutic interventions and disease prevention.

Metabolomics has currently being a popular topic for exploring human health conditions (Surendran et al. 2022). Integrated metabolite analysis and metabolomics define the chemical phenotypes of human subjects and animal models. Therefore, it holds unique potential in defining biomarkers for predicting disease incidence, severity, and progression, providing new perspectives on underlying mechanistic abnormalities (Newgard 2017). Humans primarily absorb energy through carbohydrates, fats, and proteins, involving specific biochemical cascades with various metabolic enzymes. Physiologically, metabolism-related signatures are crucial for cellular metabolic functions. Consequently, disruptions or imbalances in these signatures may lead to cellular dysfunction and various metabolic disorders (Suzuki et al. 2020). Metabolic dysregulation plays a crucial role in the pathogenesis and progression of many diseases, including tumors, obesity, diabetes, and cardiovascular diseases (Tong et al. 2020) (Newgard and Attie 2010). Certainly, some studies has demonstrated a close connection between metabolomics and the gut microbiota (Wang et al. 2011). Crohn's disease, being an autoimmune disorder, also exhibits unique characteristics of gut microbiota dysbiosis (Wolter et al. 2021). We hypothesize that metabolomics is associated with the occurrence and development of Crohn's disease. To validate this hypothesis, this study conducts a correlation analysis between metabolomics and Crohn's disease.

Data collection

We obtained two datasets (GSE186582, GSE112366) from the GEO database (https://www.ncbi.nlm.nih.gov/geo/) for analysis. The data were merged, batch effects were removed, and corrections were applied.

Identification of CD subtypes

We obtained metabolic processes and genes from the GSEA database (https://www.gsea-msigdb.org/gsea/index.jsp). Using the Consensus Cluster package, we divided CD samples into different subtypes. We set the maximum number of clusters to 10 and determined the optimal number of clusters using consensus index and cumulative distribution function (CDF). To ensure the stability of our classification, we repeated the process 100 times.

Functional enrichment analysis

Gene Set Variation Analysis (GSVA) is a non-parametric, unsupervised analytical method primarily used to assess gene set enrichment results in chip-based transcriptomics. Its main principle involves transforming the expression matrix of genes across different samples into an expression matrix of gene sets among samples. We obtained metabolism-related processes and genes from the KEGG website and utilized GSVA to analyze the enrichment status and score of different clusters across 81 metabolism-related processes.

We utilized the clusterProfiler package to conduct GO and KEGG functional enrichment analysis for hub genes. Additionally, we employed the clusterProfiler package for Gene Set Enrichment Analysis (GSEA) functional enrichment to analyze the of feature genes.

Weighted correlation network analysis

WGCNA is employed to identify modules of highly correlated genes, summarize the interconnections among modules, explore their associations with external sample features, and identify candidate biomarkers or therapeutic targets. WGCNA software package is used to build a WGCNA to identify the most correlated modules in CD patients. The optimal soft threshold is selected to transform the correlation matrix into the adjacency matrix, and the topological overlap matrix is created from the adjacency matrix, and the one module with the strongest correlation is selected for subsequent analysis.

Machine learning

We employed LASSO regression and SVM two machine learning models to filter hub genes. LASSO regression can reduce the influence of multicollinearity on the regression results by changing the coefficient of the relevant independent variables to zero through the correlation between the independent variables. SVM is a powerful supervised learning algorithm widely used for classification and regression problems. Its uniqueness lies in its optimization objective, which aims to find a hyperplane that maximizes the margin between the classification boundary and the nearest training samples. We performed LASSO and SVM algorithms on the gene sets respectively and then took the intersection to obtain the hub genes.

We utilized the "limma" R package with criteria of | log2 fold change (FC)| > 0.5 and p < 0.0001 to filter out differentially expressed genes (DEGs) between the CD group and the normal group.

Immune infiltration analysis

We employed the ESTIMATE package to assess the degree of immune cell and stromal cell infiltration in CD tissue and assigned scores. Additionally, we utilized CIBERSORT to deconvolute the immune gene expression profile, estimating the proportions of different immune cells in CD.

We used the ggpubr package to perform significant differential analysis of immune checkpoint gene expression matrices between different subtypes. Mcpcounter package was utilized to quantify the absolute abundance of eight immune cell types and two stromal cell types in CD subtypes.

XCell is an immune cell infiltration assessment tool based on gene expression data, capable of identifying potential immune cell subtypes and calculating their relative abundance in tissues. We used the XCell package to quantify the relative abundance of immune cells and stromal cells between CD subtypes.

Establishment and assessment of a nomogram

After merging the two datasets, a total of 826 CD samples and 51 normal samples were obtained. We randomly divided the CD samples into training and validation cohorts in 8:2 ratio to build a nomogram. The model was evaluated using the area under the curve (AUC) values and validation curves.

PPI and Molecular docking verification

PPI network is composed of interactions between proteins, participating in various aspects of life processes such as biological signal transduction, gene expression regulation, energy and substance metabolism, and cell cycle control. We used PPI network to explore the interconnections between proteins.

The DGIdb database (https://dgidb.genome.wustl.edu/)is a drug-gene interaction database that provides information on the associations between genes and drugs. We searched the database for information on drugs potentially correlated with the four feature genes. Schrödinger software was used for visualizing and analyzing the interactions between proteins and drugs, as well as binding patterns.

Consensus clustering identified three subtypes of CD

To explore the metabolic mechanisms in CD, we selected two microarray datasets (GSE112366, GSE186582) from the GEO database, comprising a total of 826 CD samples and 51 normal samples. Principal Component Analysis (PCA) before (Fig. 1A) and after removing the batch effect (Fig. 1B). Based on comprehensive consideration, k = 3 was considered as the optimal number of clusters (Fig. 1C). A heatmap of metabolism-related genes revealed significant differences among the three subtypes (Fig. 1D, E).

Screening for metabolism-related signatures among CD subtypes

Based on metabolism-related genes, we established three subtypes of CD. We explored whether there were distinct metabolic signatures among different subtypes. Using the "GSVA" R package, we detected 81 metabolic processes, conducted differential analysis to determine the metabolic characteristics of different subtypes. The enrichment scores of metabolic processes showed differences between the three subtypes. The results revealed that MCA exhibited distinct metabolic features.

Lipid metabolism of MCA mainly included fatty and biosythesis, arachidonic acid metabolism. MCA was primarily associated with glycan biosynthesis metabolism, mainly consisted of genes related to Glycosphingolipid biosynthesis-ganglio series, Glycosaminoglycan biosynthesis-chondroitin sulfate/dermatan sulfate, N-Glycan biosythersis, various types of N glycan biosynthesis. While genes associated with the Glycosylphosphatidylinositol (GPI) -anchor biosynthesis and Glycosphingolipid biosynthesis-globo and isoglobo series metabolism were downregulated compared to the other two subtypes. MCB and MCC were mainly associated with carbohydrate and nucleotide metabolism (Fig. 2A).

The enrichment score of metabolism-related processes in the three subtypes can be seen from the diagram as parts with differences (Fig. 2B). The diagram showed that MCA had a higher enrichment score in the 19 processes, including citrate cycle, fructose and mannose metabolism.

Correlation analysis between CD subtypes and immune status

We applied the ESTIMATE algorithm to perform immune-related analyses on three subtypes of CD. All three subtypes exhibited significant differences in immune scores, with MCA showing higher stromal and immune scores (Fig. 3A). Using CIBERSORT, we analyzed the 22 immune cells infiltration in the immune microenvironment, revealing differences among the three subtypes. In MCA cluster, B cells memory, Macrophages M0, Macrophages M1, and Neutrophils exhibited higher levels of infiltration (Fig. 3B, C). Subsequently, we analyzed the status of immune checkpoint genes. Compared to other subtypes, MCA showed higher expression levels in most immune checkpoint genes, especially CD274 (PDL1) and TNFRSF9 (Fig. 3C, D).

We used the MCPcounter package to calculate the score of immune cells and stromal cells, finding that MCA had higher infiltration levels of T cells, B lineage cells, Monocytic lineage cells, Neutrophils, endothelial cells, and fibroblasts (Fig. 3E). Additionally, we utilized the Xcell package to identify potential immune cell subtypes and calculate their relative abundance. Compared to the other subtypes, MCA exhibited higher infiltration levels of immune cells and stromal cells (Fig. 3F).

WGCNA WGCNA WGCNA WGCNA

Screening for hub genes

We use WGCNA to study metabolism-related hub genes. When the soft-threshold is set to 9, results in a scale-free network with the highest connectivity efficiency (Fig. 4A). The hierarchical clustering algorithm is employed to divide the clustering tree into 19 gene modules, with each module assigned a unique color (Fig. 4B). To identify hub genes, we analyzed the module-trait relationships between modules and subtypes. The red module, composed of 332 genes, is positively correlated with MCA (Fig. 4C).

Therefore, we choose the red module as the hub module using filtering criteria of cor.MM > 0.8 and cor.GS > 0.5 to select hub genes (Fig. 4D), resulting in a total of 20 hub genes being filtered (Table S1). We conducted GO and KEGG functional enrichment analysis for the hub genes. The GO analysis indicated significant enrichment in the extracellular matrix structural constituent, collagen fibril organization, and collagen-containing extracellular matrix (Fig. 4E). KEGG functional enrichment analysis reveals a significant association of hub genes with focal adhesion, ECM-receptor interaction, protein digestion and absorption (Fig. 4F). The circular heatmap and Manhattan plot depict the genomic positions of various genes, with ACTN1 located on chr14, MMP2 on chr16, THY1 on chr11, and ZCCHC24 on chr10 (Fig. 4G, H).

Development and validation of the feature genes

We performed LASSO (Fig. 5A, B) and SVM analysis (Fig. 5C, D) on the hub genes, respectively. Obtaining the intersection genes from both computational methods (Fig. 5E), a total of 15 hub genes (ACTN1, BGN, COL14A1, COL1A1, COL6A1, GJC1, HTRA1, MMP2, MYADM, PLXND1, PRKCDBP, RBPMS, THBS2, THY1, ZCCHC24) were included. Utilizing the PPI network for protein correlation analysis of these 15 hub genes, we found that only 11 genes exhibit interaction relationships among them (Fig. 5F).

We used the limma package to analyze CD and normal samples, obtaining differentially expressed genes and their expression profiles (Table S2). By taking the intersection of the 11 hub genes and differentially expressed genes, we identified 4 feature genes (ACTN1, MMP2, THY1, ZCCHC24) (Fig. 6A). We conducted immune cell infiltration analysis for the 4 feature genes, revealing positive correlations with Dendritic cells activated, Macrophages M0, Neutrophils, and T cells CD4 memory activated, while showing negative correlations with Dendritic cells resting, Macrophages M2, and CD8 + T cells (Fig. 6B). The lollipop plot demonstrates positive correlations of the 4 feature genes with Natural killer cells and plasmacytoid dendritic cells, and negative correlation with CD56dim natural killer cells (Fig. 6C-F). We analyzed the relationship between the feature genes and found that these feature genes were positively correlated with each other (Fig. 6G).

To explore the biological functions of the feature genes, we conducted GSEA functional analysis. ACTN1 showed a positive correlation with the chemokine signaling pathway and a negative correlation with drug metabolism cytochrome P450 and intestinal absorption (Fig. 6H). MMP2 exhibited a positive correlation with the chemokine signaling pathway and a negative correlation with drug metabolism cytochrome P450 (Fig. 6I). THY1 showed a negative correlation with the drug metabolism cytochrome P450 pathway (Fig. 6J). ZCCHC24 demonstrated a positive correlation with the extracellular matrix structural constituent pathway and a negative correlation with metabolism of xenobiotics by cytochrome P450 (Fig. 6K).

Eatablishment and validation of nomogram

We included the 4 feature genes (ACTN1, MMP2, THY1, ZCCHC24) to establish a nomogram. The 826 CD samples were randomly divided into training and validation cohorts in an 8:2 ratio. A nomogram was constructed to predict the risk of CD patients (Fig. 7A). The predictive value of the model was evaluated using ROC curves and calibration curves. The AUC for the training and validation cohorts was 0.708 and 0.704, respectively (Fig. 7B, C). The calibration curve results for both the training and validation cohorts showed good agreement (Fig. 7D, E). ROC curves for the 4 feature genes were also plotted, demonstrating a good predictive performance (Fig. 7F). Subsequently, we compared the expression of these feature genes between CD and normal samples. ACTN1, MMP2, THY1, and ZCCHC24 showed higher expression levels in CD samples (Fig. 7G-J).

Molecular docking and validation

We screened candidate drugs associated with feature genes from the DGIdb database and found that only MMP2 had 20 potential candidate drugs. We conducted Schrödinger software to score these drugs and found that 19 drugs have interactions with MMP2, and we visualized their interactions (Fig. 8).

Crohn's disease has a multifactorial etiology, involving various factors such as genetic family history, the microbiome, and environmental factors. Currently, there is limited research on the association between Crohn's disease and metabolism. Literature reviews indicate that the human host's intestinal microbiota, particularly bacteria and archaea, play a significant role in human metabolism (Fan and Pedersen 2021). There are some studies reporting a close correlation between changes in the gut microbiota and autoimmune responses (De Luca and Shoenfeld 2019). Dysbiosis of the gut microbiome has been observed in many studies on autoimmune diseases (López et al. 2016). Crohn's disease, being an autoimmune disorder, is intricately linked to the gut microbiota. Therefore, we conducted an in-depth investigation into the metabolic processes associated with Crohn's disease. CD and normal samples were collected from the GEO database. Through GSVA and WGCNA analysis, we identified hub genes. Subsequently, we conducted immune infiltration and functional enrichment analysis on these genes. After screening using machine learning methods, we finally obtained four feature genes and established a predictive model for Crohn's disease. GO and KEGG results indicate that hub genes are associated with extracellular matrix structural constituent, focal adhesion, and ECM-receptor interaction. Crohn's disease is a chronic inflammation. Long-term stimulation of the intestine leads to the release of a large number of inflammatory mediators, which stimulate the activation of fibroblasts and aggregation at the injured site.

Some reports has shown that ACTN1 is associated with the effectiveness of anti-tumor necrosis factor (TNF) therapy, where higher expression of ACTN1 is correlated with better treatment outcomes (Iacucci et al. 2023). Matrix metalloproteinases (MMPs) constitute a family of zinc-dependent extracellular matrix remodeling endopeptidases (Cabral-Pacheco et al. 2020), capable of collectively degrading all components of the extracellular matrix at neutral pH (Ravi et al. 2007). Early studies have found upregulation of MMPs (MMP-1, MMP-2, MMP-3, MMP-8, MMP-9, MMP-12, and MT1-MMP-1) in various colitis animal models and IBD patients (Baugh et al. 1999). MMPs levels are directly associated with disease activity (von Lampe et al. 2000). MMP-2 is expressed in the normal colon, mainly in epithelial cells and the lamina propria. In IBD patients, MMP-2 is highly upregulated and localized to fibroblasts/myofibroblasts around the epithelium and crypts, mononuclear cells (macrophages and lymphocytes), epithelial cells, and vascular endothelial cells (Kirkegaard et al. 2004). MMP-2 is typically involved in regulating physiological processes such as barrier and mucosal defense (Baugh et al. 1999). Studies have shown a significant increase in MMP2 expression in CD, which may contribute to impaired intestinal barrier function by increasing FITC glucose translocation, thus exacerbating the disease (Brazil et al. 2013). THY1 initially considered a thymocyte marker, has diverse biological functions, including tumor suppression, regulation of adipogenesis, cell-cell and cell-matrix adhesion, cell migration, cell differentiation, and T cell activation. Research has reported its upregulation in active Crohn's disease (Arijs et al. 2011), which may be linked to the pathogenesis of CD. Currently, there are few reports on ZCCHC24. Some literature associates it with smoking (Markunas et al. 2021) and alternative splicing (Cieply et al. 2016). There is no report on ZCCHC24 in Crohn's disease, so further exploration is needed to determine whether ZCCHC24 is a potential target gene in CD.

Crohn's disease is an autoimmune disorder, and its pathogenesis is closely related to innate immunity. Innate immune responses constitute the first line of defense against diseases in humans, involving various cells such as immune cells (neutrophils, monocytes, macrophages, and dendritic cells) and non-immune cells (epithelial cells, endothelial cells, and mesenchymal cells) (Brazil et al. 2013). In the context of IBD, neutrophils are the first immune cells to infiltrate the intestinal mucosa, and as long as active inflammation, these white blood cells persist throughout the entire IBD process. Neutrophils contribute to the onset and progression of IBD through various mechanisms, including compromised epithelial barrier function, tissue damage caused by oxidative stress and protein hydrolysis, and the sustained presence of inflammation through the release of various inflammatory mediators (de Souza and Fiocchi 2016).

In comparison to normal tissue, dendritic cells in the intestinal mucosa of CD are in an activated state (Hart et al. 2005), consistent with the results of our immune infiltration analysis. Our four feature genes are positively correlated with the activation of neutrophils and dendritic cells. It is speculated that these four feature genes may participate in the innate immunity of CD by regulating neutrophils or dendritic cells, contributing to the onset and development of CD. The products secreted by immune cells, such as cytokines and chemokines, play a significant role in initiating, mediating, and sustaining inflammation and tissue damage (Neurath 2014b). In Crohn's disease, dendritic cells produce IL-12 and IL-6, mediating immune responses (Neurath 2014a). Elevated levels of IL-8 have been observed in patients with inflammatory bowel disease (Mahida et al. 1992), consistent with our research findings. The results of GSEA enrichment analysis in this study indicate that characteristic genes are associated with pathways regulating chemokines and cytokines.

Crohn's disease is not only characterized by complex pathological mechanisms but also imposes long-term challenges for patients. One-third of patients require multiple surgeries, and approximately 14% of severe cases, especially those with rectal involvement, may require permanent ostomy, leading to various serious complications (Cosnes et al. 2012). With the development of tumor necrosis factor inhibitors, the treatment landscape of CD has evolved beyond conventional approaches. Although these drugs exhibit long-term efficacy, some patients do not respond to them or cannot sustain the treatment. Therefore, through bioinformatic analysis, we have identified metabolism-related feature genes and conducted molecular docking analysis to clarify the interaction targets of drugs and characteristic genes, providing new avenues for clinical treatment.

By analyzing Crohn's disease samples from the GEO database, we conducted functional enrichment analysis and immune cell infiltration analysis on metabolism-related signatures. Subsequently, we identified four feature genes, established a predictive model, and performed molecular docking analysis to elucidate the interaction targets between drugs and these feature genes. This approach offers new therapeutic insights for clinical treatment of Crohn's disease.

Acknowledgements Not applicable.

Author contributions XZL and YJZ were in charge of designing the research. XZL and XHZ were responsible for data analysis and manuscript writing. XZL and SZG mainly revised the article. All the authors gave final approval of the published version, and agreed to be responsible for all aspects of the work.

Funding None.

Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. The data that support the results of current study is available on Gene Expression Omnibus (GEO) websites

Ethics approval and consent to participate The GEO database is an open database and the information is freely available. Ethics statement is not required.

Consent for publication All authors agree to publish

Competing interests The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Arijs I, De Hertogh G, Machiels K, et al (2011) Mucosal gene expression of cell adhesion molecules, chemokines, and chemokine receptors in patients with inflammatory bowel disease before and after infliximab treatment. Am J Gastroenterol 106:748–761. https://doi.org/10.1038/ajg.2011.27
Baugh MD, Perry MJ, Hollander AP, et al (1999) Matrix metalloproteinase levels are elevated in inflammatory bowel disease. Gastroenterology 117:814–822. https://doi.org/10.1016/s0016-5085(99)70339-2
Bernstein CN (2015) Treatment of IBD: where we are and where we are going. Am J Gastroenterol 110:114–126. https://doi.org/10.1038/ajg.2014.357
Brazil JC, Louis NA, Parkos CA (2013) The role of polymorphonuclear leukocyte trafficking in the perpetuation of inflammation during inflammatory bowel disease. Inflamm Bowel Dis 19:1556–1565. https://doi.org/10.1097/MIB.0b013e318281f54e
Cabral-Pacheco GA, Garza-Veloz I, Castruita-De la Rosa C, et al (2020) The Roles of Matrix Metalloproteinases and Their Inhibitors in Human Diseases. Int J Mol Sci 21:9739. https://doi.org/10.3390/ijms21249739
Cieply B, Park JW, Nakauka-Ddamba A, et al (2016) Multiphasic and Dynamic Changes in Alternative Splicing during Induction of Pluripotency Are Coordinated by Numerous RNA-Binding Proteins. Cell Rep 15:247–255. https://doi.org/10.1016/j.celrep.2016.03.025
Cosnes J, Bourrier A, Nion-Larmurier I, et al (2012) Factors affecting outcomes in Crohn’s disease over 15 years. Gut 61:1140–1145. https://doi.org/10.1136/gutjnl-2011-301971
De Luca F, Shoenfeld Y (2019) The microbiome in autoimmune diseases. Clin Exp Immunol 195:74–85. https://doi.org/10.1111/cei.13158
de Souza HSP, Fiocchi C (2016) Immunopathogenesis of IBD: current state of the art. Nat Rev Gastroenterol Hepatol 13:13–27. https://doi.org/10.1038/nrgastro.2015.186
Fan Y, Pedersen O (2021) Gut microbiota in human metabolic health and disease. Nat Rev Microbiol 19:55–71. https://doi.org/10.1038/s41579-020-0433-9
Hart AL, Al-Hassi HO, Rigby RJ, et al (2005) Characteristics of intestinal dendritic cells in inflammatory bowel diseases. Gastroenterology 129:50–65. https://doi.org/10.1053/j.gastro.2005.05.013
Iacucci M, Jeffery L, Acharjee A, et al (2023) Computer-Aided Imaging Analysis of Probe-Based Confocal Laser Endomicroscopy With Molecular Labeling and Gene Expression Identifies Markers of Response to Biological Therapy in IBD Patients: The Endo-Omics Study. Inflamm Bowel Dis 29:1409–1420. https://doi.org/10.1093/ibd/izac233
Kirkegaard T, Hansen A, Bruun E, Brynskov J (2004) Expression and localisation of matrix metalloproteinases and their natural inhibitors in fistulae of patients with Crohn’s disease. Gut 53:701–709. https://doi.org/10.1136/gut.2003.017442
López P, de Paz B, Rodríguez-Carrio J, et al (2016) Th17 responses and natural IgM antibodies are related to gut microbiota composition in systemic lupus erythematosus patients. Sci Rep 6:24072. https://doi.org/10.1038/srep24072
Mahida YR, Ceska M, Effenberger F, et al (1992) Enhanced synthesis of neutrophil-activating peptide-1/interleukin-8 in active ulcerative colitis. Clin Sci (Lond) 82:273–275. https://doi.org/10.1042/cs0820273
Markunas CA, Semick SA, Quach BC, et al (2021) Genome-wide DNA methylation differences in nucleus accumbens of smokers vs. nonsmokers. Neuropsychopharmacology 46:554–560. https://doi.org/10.1038/s41386-020-0782-0
Neurath MF (2014a) New targets for mucosal healing and therapy in inflammatory bowel diseases. Mucosal Immunol 7:6–19. https://doi.org/10.1038/mi.2013.73
Neurath MF (2014b) Cytokines in inflammatory bowel disease. Nat Rev Immunol 14:329–342. https://doi.org/10.1038/nri3661
Newgard CB (2017) Metabolomics and Metabolic Diseases: Where Do We Stand? Cell Metab 25:43–56. https://doi.org/10.1016/j.cmet.2016.09.018
Newgard CB, Attie AD (2010) Getting biological about the genetics of diabetes. Nat Med 16:388–391. https://doi.org/10.1038/nm0410-388
Ravi A, Garg P, Sitaraman SV (2007) Matrix metalloproteinases in inflammatory bowel disease: boon or a bane? Inflamm Bowel Dis 13:97–107. https://doi.org/10.1002/ibd.20011
Surendran P, Stewart ID, Au Yeung VPW, et al (2022) Rare and common genetic determinants of metabolic individuality and their effects on human health. Nat Med 28:2321–2332. https://doi.org/10.1038/s41591-022-02046-0
Suzuki A, Minamide M, Iwaya C, et al (2020) Role of Metabolism in Bone Development and Homeostasis. Int J Mol Sci 21:8992. https://doi.org/10.3390/ijms21238992
Tong Y, Gao W-Q, Liu Y (2020) Metabolic heterogeneity in cancer: An overview and therapeutic implications. Biochim Biophys Acta Rev Cancer 1874:188421. https://doi.org/10.1016/j.bbcan.2020.188421
Torres J, Mehandru S, Colombel J-F, Peyrin-Biroulet L (2017) Crohn’s disease. Lancet 389:1741–1755. https://doi.org/10.1016/S0140-6736(16)31711-1
von Lampe B, Barthel B, Coupland SE, et al (2000) Differential expression of matrix metalloproteinases and their tissue inhibitors in colon mucosa of patients with inflammatory bowel disease. Gut 47:63–73. https://doi.org/10.1136/gut.47.1.63
Wang Z, Klipfell E, Bennett BJ, et al (2011) Gut flora metabolism of phosphatidylcholine promotes cardiovascular disease. Nature 472:57–63. https://doi.org/10.1038/nature09922
Wolter M, Grant ET, Boudaud M, et al (2021) Leveraging diet to engineer the gut microbiome. Nat Rev Gastroenterol Hepatol 18:885–902. https://doi.org/10.1038/s41575-021-00512-7

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Identification of metabolism-related subtypes and feature genes in Crohn’s disease

Status:

Version 1

Abstract

Figures

Introduction

Materials and methods

Data collection

Identification of CD subtypes

Functional enrichment analysis

Weighted correlation network analysis

Machine learning

Immune infiltration analysis

Establishment and assessment of a nomogram

PPI and Molecular docking verification

Results

Consensus clustering identified three subtypes of CD

Screening for metabolism-related signatures among CD subtypes

Correlation analysis between CD subtypes and immune status

WGCNA WGCNA WGCNA WGCNA

Screening for hub genes

Development and validation of the feature genes

Eatablishment and validation of nomogram

Molecular docking and validation

Discussion

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1