The effect of long-term HFD on gene expression in BXD colons.

(A) Graphical representation of a pipeline of a previously described BXD mouse study (Williams et al., 2016). Mice were fed HFD or CD starting from 8 weeks of age and metabolic phenotyping was performed as indicated. Mice were sacrificed at 29 weeks of age and multiple organs were collected and frozen for further analyses. BXD colon transcriptomes were analyzed in this study. CD: chow diet indicated in blue, HFD: high-fat diet indicated in red. P values were calculated by two-tailed Student’s t-test and indicated as follows: * P <0.05; ** P <0.01; *** P <0.001; **** P <0.0001. (B) Volcano plot showing the HFD effect on BXD colon transcriptomes compared to CD and the up-and down-regulated differentially expressed genes (DEGs, absolute Log2(Fold Change) > 0.5 and BH-adjusted P value (adj. P) < 0.05) were highlighted in red and blue, respectively. (C) Gene set enrichment analysis showing the effect of HFD on gene expression in BXD colons. Gene sets were grouped into five categories: Translation, Inflammation, Mitochondria, Stress, and Intermediate filament. Normalized enrichment scores (NES) were represented by color and -Log10 (BH-adj. P) were represented by dot size and indicated as follows: * Adjusted P value <0.05; ** Adjusted P value <0.01; *** Adjusted P value <0.001. (D) Enrichment analysis of molecular signatures of mouse and human IBD on the transcriptome of BXD colons. UC: Ulcerative colitis, CDs: Crohn’s disease

Identifying susceptible strains to HFD-induced IBD-like inflammation and their effect on plasma cytokines.

(A) Enrichment analysis of the molecular signatures in human (bottom panel) and mouse IBD (top panel) models on the gene expression of individual BXD colons. Dendrogram showing the propensity to develop a mouse IBD-like signature among BXD strains upon HFD. BXD strains were divided into three clusters: susceptible (in red), intermediate (in blue), and resistant strains (in green). Normalized enrichment scores (NES) were represented by color and -Log10(BH-adjusted P values) were represented by dot size and indicated as follows: * Adjusted P value <0.05; ** Adjusted P value <0.01; *** Adjusted P value <0.001. UC: Ulcerative colitis, CDs: Crohn’s disease (B, C) Boxplots showing the effect of susceptible strains on plasma IL-10 (B) and IL-15 (C) level compared to resistant strains. P values were calculated by two-tailed Student’s t-test and indicated as follows: * P <0.05; ** P <0.01; *** P <0.001; **** P <0.0001.

Identifying IBD-related gene modules.

(A) Pipeline for exploring the IBD-associated gene modules. A co-expression gene network was constructed based on the transcriptome of BXD colons under HFD. IBD-associated modules were then defined as gene modules under HFD that are significantly clustered in mouse DSS-induced UC signatures. (B) UMAP representation of the co-expression gene network under HFD. 39 co-expression modules are represented in the corresponding color and the correlated modules (Spearman correlation coefficient between the eigengene of modules > 0.7) were linked by a grey line. (C) Heatmap showing the enrichment of co-expression modules in mouse and human IBD gene signatures and the number and percentage of enriched genes were labeled. The number of enriched genes divided by the number of genes involved in the respective module was defined as the percentage of enriched genes. Signed –Log10(adj. P) indicated BH-adjusted P value and the enriched gene set. Enriched modules of up-and down-regulated genes upon disease were highlighted in red and blue, respectively. UC: Ulcerative colitis, CDs: Crohn’s disease.

Biological interrogation of identified IBD-related modules.

(A, B) Dot plots showing the enrichment of IBD-related modules in hallmark genesets (A) and cell-type gene signatures of inflamed colon in Crohn’s disease patients (B). Gene ratios higher than 0.1 are shown and represented by dot size. Dots are colored by -log10(BH-adjusted P values). (C, D) The enriched motifs for promoters of the genes involved in module HFD_M9 (C) and HFD_M28 (D). The significantly enriched motifs (P value < 0.001) were ranked based on the percentage of enriched promoters (In top motifs) and then the top five TFs were selected. TF: Transcription factor. PWM: Positional weight matrix.

ModQTL mapping for two IBD-related modules and the prioritization of candidate genes.

(A) Manhattan plot showing the ModQTL mapping result for disease-related modules HFD_M9 and HFD_M28. ModQTL maps of HFD_M9 and HFD_M28 are indicated in orange and purple, respectively. The threshold calculated by permutation test (P < 0.1) for HFD_M28 is represented by a purple dashed line. (B) The filtering criteria for selecting candidate genes under the ModQTL peak for HFD_M28. Genes with 2 of the described criteria are considered as candidate genes. (C) The most significant associations between 27 candidate genes under the ModQTL peak and Crohn’s disease or UC identified through GWAS according to whole genome sequence in the human UKBB are shown in the scatter plot (top panel). Crohn’s disease and UC are indicated by pink circle and green triangle, respectively. The threshold (-Log10P value = 5) is represented by a red dashed line. Heatmap showing the identified 27 candidate genes of module HFD_M28 (bottom panel). Variants colored in blue indicate genes with high-impact genetic variants in BXD mice (including missense, frameshift, initiator codon, splice donor, splice acceptor, in-frame deletion, in-frame insertion, stop lost, stop gained). Inflammation is indicated in red and represents genes associated with inflammation based on literature mining. Cis-eQTL colored in purple indicates genes with Cis-eQTLs. Tissue specificity colored in orange means genes that are highly expressed in human intestine (data were downloaded from human protein atlas, https://www.proteinatlas.org/humanproteome/tissue/intestine (Uhlén et al., 2015)). GWAS result of UC in UKBB colored in green indicates that genes are significantly associated with human UC. (D, E) Manhattan plots showing the associated gene expression modules of Epha6 in mouse gastro-intestinal tract (D) and that of MUC4 in human gastro-intestinal tract (E) (data from https://systems-genetics.org/gmad (Li et al., 2019)). The threshold is represented by the red dashed line (absolute Gene-Module Association Score (GMAS) >= 0.268). Terms above the threshold are identified as the significant associated terms. GO terms or gene modules are ranked by similarity. Known associated terms are shown as red dots and new significant associated terms are colored in black. (F) Dot plot showing that the expression of MUC4 was higher in four cell types of human inflamed colon with Crohn’s disease.

Transcriptome profile in BXD colons.

(A) Principal-component analysis (PCA) of the microarray profiles of BXD colons under high-fat (HFD, indicated in red) or chow diet (CD, represented in blue). (B) Bar plot showing the primary principle (PC1) calculated by PCA of the colon transcriptomes in each BXD strain fed with CD (blue) or HFD (red). BXD strains highlighted in bold mean they are more resistant to dietary challenges. (C) Heatmap showing unsupervised hierarchical clustering of colon transcriptome in both the CD and HFD fed BXDs. (D) Enrichment analysis of inflammation-related genesets showing the effect of HFD on gene expression in individual BXD colon. Normalized enrichment scores (NES) were represented by color and -Log10(BH-adjusted P values) were represented by dot size and indicated as follows: * Adjusted P value < 0.05; ** Adjusted P value < 0.01; *** Adjusted P value < 0.001. BXD strains highlighted in red represent the 3 most susceptible strains to gut inflammation upon HFD. BXD strains colored in green show no significant enrichment in gut inflammation.

Exploring IBD-associated co-expression modules under CD.

(A) Heatmap showing modules enriched in mouse IBD gene signatures and the number and percentage of enriched genes were labeled. The number of enriched genes divided by the number of genes involved in the respective module was defined as the percentage of enriched genes. Signed –Log10(adj. P) indicated BH-adjusted P value and the enriched gene set. Enriched modules of up-and down-regulated genes upon disease were highlighted in red and blue, respectively. (B) Heatmap showing the similarity of co-expression modules identified under CD or HFD. The number and percentage of overlapped genes were labeled. The number of overlapped genes divided by the number of genes involved in the respective module in HFD was defined as the percentage of overlapped genes. Adjusted P values calculated by BH were indicated by color. (C) Heatmap showing the modules under CD enriched in human IBD gene signatures and the number and percentage of enriched genes were labeled. UC: Ulcerative colitis, CDs: Crohn’s disease.

Genome-wide association studies (GWAS) for Ulcerative colitis (UC) and Crohn’s disease (CDs) in humans.

(A, B) Lollipop plots showing UC-or CDs-associated genetic variants of EPHA6 (A) and MUC4 (B) based on UKBB whole genome sequence data. Variants effect were predicted and their classification are represented by color. VEP: Variant Effect Prediction.

Gene signatures of mouse and human IBDs