Introduction

Clostridioides difficile is an anaerobic, Gram-positive bacterium that causes severe and life-threatening diarrhea1. Virulent C. difficile strains are highly toxigenic, they secrete two primary enterotoxins; TcdA and TcdB2,3. These toxins deactivate small GTPases within the host cells after penetrating them, causing disturbances in the actin cytoskeleton, and resulting in the loss of tight junctions in the intestinal epithelial layer4. This increases permeability and ultimately results in cell death (cytopathic effect)5. Neutrophil infiltration, a distinctive characteristic of CDI, results in inflammation and mucosal tissue damage6. CDI occurs when C. difficile spores are introduced into the intestines of vulnerable individuals7. Spores of C. difficile can originate either within the body or from external sources, such as contact with contaminated fecal matter8. In the intestine, the presence of bile salts triggers the germination of C. difficile spores into toxin-producing vegetative cells9. Disturbances in the normal intestinal flora (often due to broad-spectrum antibiotic use, hospitalization, old age, and comorbidities such as HIV or cancer) can lead to C. difficile infection (CDI)10. Although CDI is frequently associated with hospitalization; its pathogenesis remains incompletely understood11. The Infectious Diseases Society of America (IDSA) and the Society for Healthcare Epidemiology of America (SHEA) limit CDI treatment guidelines to two antibiotics, vancomycin and fidaxomicin. At the same time, metronidazole is only recommended without other agents12,13,14,15. As mentioned, enterotoxins are pivotal in the pathogenesis of CDI as they induce a robust inflammatory response in the host16. Toxin A (TcdA) and toxin B (TcdB) disrupt the cytoskeleton of intestinal epithelial cells by glycosylating Rho family GTPases, leading to cell rounding, loss of barrier function, and cell death17. This cellular damage triggers the release of pro-inflammatory cytokines and chemokines, which recruit neutrophils and monocytes to the site of infection further perpetuating tissue damage and inflammation. The resulting colonic inflammation manifests as pseudomembranous colitis, characterized by the formation of diphtheritic membrane composed of inflammatory cells, dead epithelial cells, and fibrin18,19.

Most host cells possess uniform DNA. Analyzing cells at the protein level rather than the DNA level is more effective in evaluating their variety and functions20. Current protein profiling technologies offer limited insights into proteins expressed in low abundance and the kinetics of protein expression. Spatial and sequential cross talk between cells at the site of infection can be comprehensively evaluated using mRNA expression signatures (transcriptome)21. Investigating the contribution of individual cell types could provide novel approaches for preventing and treating C. difficile infections. In several previous mouse studies conducted in our and other laboratories, up to 30% of the mice infected with C. difficile do not show clinical signs typical to CDI post-infection even though they carry C. difficile in their intestinal contents (asymptomatic mice)22,23,24,25,26,27,28,29,30. The specific objective of the present study was to determine the differential gene expression in the large intestinal mucosa of the symptomatic mice and compare it to asymptomatic mice using spatial transcriptomics technology. Spatial transcriptomics and direct sequencing of transcripts from the tissue generates a detailed map of diverse cell types and corresponding gene expression patterns in their natural spatial context31.

Materials and methods

Materials and reagents

C. difficile strain ATCC 43255 was obtained from Microbiologics (St. Cloud, MN, USA). C. difficile ATCC 43255 (VPI 10463) is a ribotype 087, toxigenic strain that produces both toxins A and B. This strain is commonly used in mouse infection models of CDI. Vancomycin hydrochloride (Gold Biotechnology, St. Louis, MO, USA), metronidazole (Beantown Chemical Corporation, Hudson, NH, USA), kanamycin, gentamycin, and colistin (TCI America, Portland, OR, USA) were procured from commercial vendors. Brain Heart Infusion (BHI) was purchased from Becton, Dickinson, and Company (Cockeysville, MD, USA). Phosphate-buffered saline (PBS), fetal bovine serum, and non-essential amino acids (NEAA) were purchased from Fisher Scientific (Waltham, MA, USA). Yeast extract, L-cysteine, vitamin K, and hemin were obtained from Sigma-Aldrich (St. Louis, MO, USA). The RNA extraction kit (RNeasy® Mini), DNase, and reverse transcription kit (QuantiTect®) were purchased from Qiagen (Germantown, MD, USA). PowerUp™ SYBR™ Green Master Mix (Applied Biosystems, Waltham, MA, USA) was purchased from Fisher Scientific.

Mouse model of C. difficile infection, tissue sample collection and processing

Mouse studies were conducted by the American Veterinary Medical Association (AVMA) guidelines and were approved by the Purdue Institutional Animal Care and Use Committee (IACUC) under protocol number 2008002068. Intestinal tissue samples were obtained from a prior animal study conducted in our laboratory27. All methods involved in the study were performed in accordance with relevant guidelines and regulations. All methods involved in the study comply with ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines and regulations and with the relevant ethical guidelines32. Male and female C57BL/6 mice (n = 6 [3 Males and 3 females], Jackson Laboratory, Bar Harbor, ME, USA) were acclimatized for one week before oral treatment with a five-antibiotic cocktail (kanamycin [0.4 mg/mL], gentamicin [0.035 mg/mL], vancomycin [0.045 mg/mL], metronidazole [0.215 mg/mL], and colistin [850 U/mL]) for three days. Antibiotic treatment was then ceased for two days to allow for drug clearance before injecting the mice intraperitoneally with clindamycin (10 mg/kg). One day later, mice were orally infected with 5 × 10⁵ CFU/mL C. difficile ATCC 43255 spores. Uninfected mice were used as a control. Mice were monitored for signs of CDI (including diarrhea, scuffed coat, hunching, inability to eat or drink, lethargy, and unresponsiveness) and were euthanized immediately upon the appearance of severe signs. Untreated and asymptomatic mice were monitored for an additional week and then euthanized at the end of the experiment using CO₂ asphyxiation, and organs were collected for histopathological examination and spatial transcriptomic analysis. This difference in timing was necessary to comply with ethical guidelines and humane endpoints, as animals displaying severe disease signs could not be kept alive for extended observation. The experiment was repeated three times independently, and samples and results were pooled from them. Thirty colonic and cecal tissues were preserved in 10% neutral buffered formalin, transferred to 70% ethanol, and cleared in xylene. Subsequently, the tissues were embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H&E) for histological evaluation. Tissue processing was conducted by the Histology Research Laboratory at Purdue University.

Spatial transcriptomics analysis

NanoString GeoMx spatial profiling platform was utilized. Sections of paraffin-embedded FFPE tissues at 5 μm thickness were placed in a positively charged slide (PathSupply, Wilmington, DE, USA). Sections were mounted on microscope slides and dried.​ The slides were deparaffinized in antigen retrieval solution (Thermo Fisher Scientific) at 100 °C for 20 min and treated with 1 µg/mL proteinase K at 37 °C for 15 min (Thermo). Mouse transcriptomics was visualized using the GeoMx mouse whole transcriptome atlas (WTA, NanoString, Bruker Spatial Biology, Inc,). An overnight in situ hybridization was performed with a mouse WTA probe and the next day the slides were washed twice at 37 °C for 25 min with 50% formamide/2X saline-sodium citrate (SSC) buffer to remove unbound probes. The slides were then incubated with morphological markers. Anti-pancytokeratin (Pan-CK) antibody at 1:200 (Thermo) tagged to Alexa Fluor 532, anti-CD45 antibody at 1:200 (CST) tagged to Alexa Fluor 594, and anti-CD3 antibody at 1:150 (OriGene) tagged to AlexaFluor 647 antibodies were used for immunostaining of slides. Nuclei were stained with Syto13 fluorescent dye (Fluorescence 532, Thermo). Stained slides were loaded onto the GeoMx Digital Spatial Profiler (DSP; NanoString) and scanned. After visual inspection using the GeoMx DSP, geometrical regions of interest (ROI) were selected for oligonucleotide collection. Photocleaved oligonucleotides from each spatially resolved ROI were sequenced with the Illumina NextSeq2000 sequencer (Illumina) and downstream analyses were performed using GeoMx Server (NanoString).

Quality control and data normalization

Quality control, normalization, and data analysis were conducted using the GeoMx DSP Analysis Suite (DAS) online portal. After quantifying the probes, the RNA analysis data (DCC files) were transferred to the Nanostring GeoMx platform for detailed QC, normalization, and subsequent analysis. This step facilitated a thorough quality control (QC) assessment using set criteria, the necessary normalization, and analytical procedures. Precise measurement and high-quality data of the RNA samples under study were ensured. As a quality control criterion, the raw read threshold and total number of barcodes were counted by the Illumina sequencer. Only regions with > 1000 reads were used for the analysis. Percent aligned reads segments with < 80% alignment were flagged for evaluation, assessing potential contamination sources. Sequencing saturation of < 50% was flagged. Negative probe count geomean (RNA-NGS) segments of negative probes < 10 were flagged, likely to be influenced by slide preparation and wash steps. No template control (NTC) counts of > 1000 were analyzed if the DCC file for NTC has low, even count distribution, segments do not have similar deduplicated counts as their NTC, and segments have adequate downstream gene detection. Segments with a cell number or surface area below the threshold were flagged. QC values for % trimmed and % stitched were > 90%. Biological QC aimed to remove low-performing probes using one specific probe per gene, focusing on negative probes. A probe was excluded if it was a high or low outlier in > 20% of segments. The limit of quantification (LOQ) was defined as two standard deviations above the geometric mean of the negative probes. Segments with low signals were removed retaining segments where 5% of targets were above the expression threshold, set at LOQ below 2. Low-detected genes were removed retaining targets detected in at least 5% of segments as defined by an expression threshold. Technical effects, like segment size and tissue quality, were normalized using third quartile (Q3) normalization. The 75th percentile (Q3) of gene counts was calculated within each segment, and all gene counts were divided by this Q3 value. This step was repeated for each segment, and then all gene counts from all segments were multiplied by the geometric mean of all Q3 counts. Screening genes of interest was performed to eliminate anomalies before consolidating the data into one count per sample and target.

Gene expression analysis

Genes were analyzed across the following comparisons; symptomatic versus asymptomatic mice, superficial layer of symptomatic versus asymptomatic mice, deep layer of symptomatic versus asymptomatic mice, and the superficial versus deep layers of colon, cecum and colon + cecum in both symptomatic and asymptomatic mice. Genes with a Log2 fold change (LFC) of ≥ 1 were considered significantly upregulated, while genes with an LFC ≤ −1 were considered significantly downregulated. The functions of these significantly dysregulated genes were verified using GeneCards (https://www.genecards.org) and relevant literature.

Pathway analysis and differential expression analysis of data

Pathway and differential expression analyses and differential expression analysis were conducted using the GeoMx Digital Analysis Suite (DAS). The online databases Reactome (https://reactome.org/) and KEGG (https://www.genome.jp/kegg/pathway.html) databases were utilized to analyze and illustrate pathways and pathway maps33,34. Pathways were sorted based on adjusted p-value, coverage, and normalized enrichment score (NES). Pathways with an adjusted p-value of < 0.05, coverage equal to or greater than 70%, and a significant NES were prioritized. NES accounts for differences in gene set sizes and variations in enrichment scores across gene sets, providing a normalized measure of pathway enrichment. NES was considered significant if the adjusted p-value was less than 0.05.

Raw datasets generated and analyzed during this study are available through supplementary Tables 1–6 and through the Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/) with the accession numbers provided in the supplementary files.

Quantitative PCR (qPCR) validation of Spatial transcriptomic results

To validate the differentially expressed genes identified by spatial transcriptomics, we performed qPCR using RNA extracted from colon tissues of symptomatic and asymptomatic mice. Total RNA was extracted using the RNeasy Mini Kit (Qiagen), and on-column DNase digestion was performed using the RNase-Free DNase Set (Qiagen) to eliminate genomic DNA contamination as per manufacturer’s instructions. Reverse transcription was carried out using the QuantiTect Reverse Transcription Kit (Qiagen). Gene-specific primers were designed using NCBI Primer-BLAST and synthesized by Integrated DNA Technologies (IDT, Ann Arbor, MI). Beta-actin was used as the reference housekeeping gene. Quantitative PCR was performed using SYBR Green PCR Master Mix (Thermo Fisher Scientific) on a QuantStudio Real-Time PCR System (Applied Biosystems). Colon samples from six symptomatic and six asymptomatic mice (n = 6, collected from independent biological replicates under identical experimental conditions) were tested. All reactions were run in technical duplicates. Gene expression levels were calculated using the –ΔΔCt method and log2 transformed to present data as log2 fold change (–ΔΔCt).

Statistical analysis

Following normalization, we conducted all statistical analyses and created visualizations based on the normalized Log2 counts. To determine expression correlations, we calculated the Pearson correlation coefficient (R) for paired groups, where R > 0 indicated a positive correlation and R < 0 as a negative correlation. We used linear mixed modeling to assess expression differences between two groups and performed one-way analysis of variance (ANOVA) for comparisons across multiple groups. For the generation of heatmaps, the normalized Log2 counts were zero-centered, and the Euclidean distance method was used to quantify distances. Average hierarchical clustering was then employed to identify expression clusters. For qPCR validation, unpaired t-tests were conducted to compare gene expression between symptomatic and asymptomatic mice. GraphPad Prism was used to generate dot plots with mean ± SEM, focusing on significantly altered genes. The average ΔCt of each gene from asymptomatic mice was used as the reference to compute the ΔΔCt values for all the mice.

Results

Selection of rois between symptomatic and asymptomatic mice

Histological differences in the tissue between symptomatic and asymptomatic mice in a CDI mouse model were reported previously27. Briefly, symptomatic mice exhibited significant edema, neutrophil-driven inflammation, necrosis, and ulceration. The asymptomatic mice had tissue morphology comparable to the uninfected control mice27. The variation in tissue morphology between symptomatic and asymptomatic mice was analyzed by selecting 25 regions of interest (ROIs) from each sample, using hematoxylin and eosin (H&E), pancytokeratin (PanCK), and CD45 staining. CD3 staining was also performed; however, it did not label sufficient cells to reach the threshold nucleus count, whereas CD45 provided the necessary cellular coverage. ROIs stained positive for pancytokeratin (PanCK+) were classified as areas of illumination (AOIs), indicating epithelial regions. Furthermore, targeted AOIs were analyzed for CD45, a surface marker specific to immune cells such as T cells, B cells, neutrophils, and macrophages35. Anti-CD45 allowed for the identification and localization of inflammatory cell infiltration within these epithelial regions. CD45 was selected for further analysis to assess immune activity and its correlation with symptomatic versus asymptomatic tissue changes.

Figure 1 shows example ROIs selected from symptomatic and asymptomatic mice samples. These ROIs were chosen from superficial and deep layers of the colon and cecum.

Fig. 1
figure 1

A, B, and C depict the regions of interest (ROIs) selected from the superficial and deep layers of the colon and cecum in three symptomatic mice. D and E show the ROIs from the superficial and deep layers of the colon and cecum in two asymptomatic mice. In F, PanCK is shown in red, and CD45 appears in light purple within each selected ROI.

Full size image

Heat map and cluster analysis result

Figure 2 shows a heatmap displaying cluster analysis of gene expression in C. difficile-infected mice, comparing symptomatic and asymptomatic groups across the superficial and deep layers of the cecum and colon. The analysis highlights the significant upregulation and downregulation of genes involved in activated inflammatory pathways in symptomatic mice, including Lcn2, Dusp1, Cxcl2, S100a8, S100a9, Defa2, Gstm1, H3c2, H4c17, Cd74, and B2m.

Fig. 2
figure 2

Cluster analysis heatmap displays gene expression in Clostridioides difficile-infected mice, comparing symptomatic and asymptomatic groups across superficial and deep layers of the intestine. This visualization distinguishes between significantly upregulated and downregulated genes pivotal to activated pathways in symptomatic mice (Lcn2, Dusp1, Cxcl2, S100a8, S100a9, Defa2, Gstm1, H3c2, H4c17, Cd74, and B2m). Colors range from red (high expression) to blue (low expression), indicating distinct expression profiles between the groups.

Full size image

Spatial transcriptomics analysis

Differential transcriptomic analysis of symptomatic versus asymptomatic mice

Significantly upregulated or downregulated genes with known functions were investigated, which was followed by a pathway analysis. The dysregulation of genes was confirmed by performing gene analysis based on the results from pathway analysis. This sequential approach allowed us to transition from pathway analysis to gene-level investigation. Differential gene expression analysis was conducted on RNA samples from selected ROIs within mouse large intestinal mucosa. Upregulated pathways predominantly displayed pro-inflammatory characteristics when comparing all the layers (superficial, middle, and deep) of the colon and cecum of the symptomatic CDI mice to asymptomatic. The interleukin-17 signaling pathway emerged as one of the most upregulated pathways. This pro-inflammatory pathway stimulates other pathways, such as MAPK and NF-κB signaling. Our results indicate a significant upregulation of genes involved in MAPK and NF-κB signaling pathways namely, Cd14, Jun, Areg, Dusp1, Mapk8, Tlr4, Nfkbia, and Il1b.

Figure 3 depicts a volcano plot illustrating the upregulated and downregulated genes in symptomatic CDI mice compared to infected asymptomatic mice. This plot provides a visual representation of the differential gene expression, highlighting the genes that are significantly altered in response to symptomatic infection.

Fig. 3
figure 3

Volcano plot: symptomatic CDI mice upregulated and downregulated genes compared with infected asymptomatic mice, identified 13,801 genes when comparing symptomatic CDI mice to infected asymptomatic mice. Of these, 227 genes had at least a two-fold upregulation in symptomatic mice as evidenced by Log2 fold change (LFC) of ≥ 1 and 173 genes were significantly downregulated with LFC of ≤−1.

Full size image

Among the most upregulated genes were those associated with proinflammatory pathways, particularly the IL-17 signaling pathway. Table 1 shows the Log2 fold-change of selected upregulated genes between infected symptomatic and asymptomatic mice. S100a8 and S100a9, crucial genes in IL-17 signaling pathway, were significantly upregulated (LFC = 3.13 and 2.55, respectively). Other upregulated genes include, Lcn2, Cxcl2, and Dusp1 (LFC = 3.56, 2.67 and 3.4, respectively). Genes play a role in gut integrity and mucosal immunity were not significantly upregulated in symptomatic mice, including Muc1, Ocln, and Reg3g. Table 1 also displays downregulated genes between infected symptomatic and asymptomatic mice. A total of 173 genes were significantly downregulated in symptomatic mice with an LFC of ≤−1, indicating at least a two-fold decrease in expression. These values highlight the genes significantly less expressed in symptomatic mice, offering insights into the molecular mechanisms that may contribute to the development of symptoms in CDI. Genes significantly downregulated in comparing symptomatic mice versus asymptomatic mice are mainly associated with crucial cellular processes such as antigen processing and presentation and various metabolic pathways encompassing lipid, amino acid metabolism and PPAR (Peroxisome Proliferator-Activated Receptor) signaling. Within the antigen presentation and processing pathway, several genes including Cd74, Ctss, Psme2, and B2m showed notable decreases in expression (LFC= −3.56, −1.33, −1.22, and − 1.39, respectively). Additionally, significant downregulation was observed in metabolism-related genes, including Fabp6, Fads2, Hao2, and Gstm1 (LFC= −2.46, −1.87, −1.95, and − 1.49 respectively). Moreover, genes involved in the neutrophil extracellular trap (NET) formation pathway, including H3c2 and H4c17, also exhibited significant downregulation (LFC= −2.28 and − 1.64) in symptomatic mice compared to asymptomatic mice. This suggests a decrease in histone gene expression critical for NET formation. Another significantly downregulated group of genes is the defensin family, including Defa2, Defa17, Defa34, and Defa40 (LFC= −1.46, −1.21, −1.14, and − 1.07, respectively), which are part of the IL-17 signaling pathway and contribute to host defense against pathogens.

Table 1 Log2 fold change of selected upregulated and downregulated genes in symptomatic mice. These genes were chosen for their roles in the inflammatory processes, immune response, immune defense, antigen processing and presentation during Clostridioides difficile infection (CDI), including those related to the IL-17 signaling pathway and its downstream effects.
Full size table

Differential gene expression in superficial and deep intestinal layers between symptomatic and asymptomatic mice

Gene expression was compared in the superficial layer of the colon and cecum of symptomatic mice versus the superficial layers of asymptomatic mice. A similar comparison was made for the deep layer of the colon and cecum. Figure 4 illustrates differential gene expression in the superficial layers and Fig. 5 depicts differential gene expression in the deep layers.

Fig. 4
figure 4

Volcano plot illustrates differential gene expression in the superficial layers of the intestine between symptomatic and asymptomatic mice, identifying the expression of 13,801 genes, 444 genes were significantly upregulated with an LFC of ≥ 1. Conversely, 506 genes were significantly downregulated with an LFC of ≤−1.

Full size image
Fig. 5
figure 5

Volcano plot illustrates differential gene expression in the deep layer of the intestine between symptomatic and asymptomatic mice, identifying the expression of 13,801 genes, 663 genes were significantly upregulated with an LFC of ≥ 1. Conversely, 1833 genes were significantly downregulated with an LFC of ≤−1.

Full size image

The analysis of differential gene expression between the superficial and deep layers of the colon and cecum in symptomatic mice compared to asymptomatic mice revealed distinct patterns of expression of several genes. The S100a8 gene was significantly upregulated in both the superficial (LFC = 3.30) and deep layers (LFC = 2.50) of the intestine in symptomatic mice, with higher expression in the superficial layer. S100a9 also showed significant upregulation in the superficial (LFC = 3.26) and deep layers (LFC = 2.42), with more pronounced expression in the superficial layer. Dusp1 was significantly upregulated in both layers (superficial LFC = 3.81 and deep LFC = 3.73), showing relatively similar expression. The Lcn2 gene exhibited significant upregulation in both layers, with higher expression in the deep layer (LFC = 4.44) compared to the superficial layer (LFC = 2.98). Cxcl2 was significantly upregulated in superficial and deep layers of symptomatic mice (LFC = 2.37 and 1.71, respectively), with greater expression in the superficial layer. The gene expression analysis showed significant downregulation of Cd74, Gstm1, B2m, H3c2, and H4c17 in symptomatic mice compared to asymptomatic. Cd74 is notably downregulated in both the superficial and deep layers (LFC= −3.81 and − 4.04, respectively). Gstm1 was also more downregulated in the deep layer (LFC= −3.23) than in the superficial layer (LFC= −1.08). B2m showed a similar trend, with more significant downregulation in the superficial layer (LFC= −1.76) than in the deep layer (LFC= −1.13). H3c2 was significantly reduced in the superficial layer (LFC= −1.85), with a minor decrease in the deep layer (LFC= −0.19). H4c17, on the other hand, showed significant downregulation in the superficial layer (LFC= −1.53) but a slight increase in the deep layer (LFC = 0.20). These results indicate distinct layer-specific differences in gene expression patterns in the intestines of symptomatic mice and asymptomatic mice.

Differential gene expression between intestinal layers in symptomatic mice

This analysis compared gene expression between the deep and superficial layers in symptomatic mice. Figure 6 shows the differences in gene expression between these layers in symptomatic mice.

Fig. 6
figure 6

Volcano plot analysis identified 14,025 genes when comparing gene expression between the superficial and deep layers of the intestine in symptomatic mice. Of these, 4,012 genes were upregulated with a Log2 fold change (LFC) of ≥ 1, indicating at least a two-fold increase in expression in the superficial layer compared to the deep layer. On the other hand, 739 genes were significantly downregulated, with an LFC of ≤−1, reflecting a two-fold decrease in expression in the superficial layer relative to the deep layer. This analysis highlights the differences in gene expression between the two intestinal layers in symptomatic mice.

Full size image

The analysis of previously identified upregulated and downregulated genes in symptomatic mice compared their expression between the superficial and deep layers of the intestine. For S100a8, the LFC was 0.68; for S100a9, it was − 0.28, indicating no significant changes. Dusp1 had an LFC of −0.63, while Lcn2 was − 1.32, showing significant downregulation. Cd74 (LFC = 0.63), Gstm1 (LFC = 1.52), and B2m (LFC = 0.15) varied, with Gstm1 significantly upregulated. H3c2 (LFC= −4.00) and H4c17 (LFC= −3.37) were significantly downregulated. The analysis revealed that genes, including Gstm1 and H3c2, had more pronounced differences in expression profiles.

Differential gene expression between intestinal layers in asymptomatic mice

Figure 7 Shows the differences in gene expression between deep and superficial layers in asymptomatic mice.

Fig. 7
figure 7

Volcano plot analysis compared gene expression between the superficial and deep layers of the intestine in asymptomatic mice. A total of 12,888 genes were identified in this analysis, 803 genes were significantly upregulated (Log2 fold change [LFC] of ≥ 1), and 203 genes were significantly downregulated (LFC of ≤−1). This analysis highlighted the differences in gene expression between the two layers of the intestine in asymptomatic mice.

Full size image

Gene expression was compared between the superficial and deep layers of the intestine in asymptomatic mice. Out of the genes identified in previous comparisons between symptomatic and asymptomatic conditions, Cd74, Gstm1, H3c2, and H4c17 were significantly downregulated in the superficial layer compared to the deep layer (LFC = −1.10, −1.50, −1.42 and − 1.52, respectively). These differences showcase expression profile nonconformance in gene expression between the intestinal layers in asymptomatic mice.

Quantitative PCR confirms Spatial transcriptomic gene expression patterns using Log2 fold change

Quantitative PCR analysis validated the expression patterns of genes identified in the spatial transcriptomics dataset. Log2 fold changes (–ΔΔCt) demonstrated consistent and statistically significant upregulation of S100a8, S100a9, Lcn2, Cxcl2, and Dusp1, and downregulation of Cd74 in symptomatic mice relative to asymptomatic controls (p < 0.01, unpaired t-test). These results were consistent across three independent experiments using identical protocols (Fig. 8).

Fig. 8
figure 8

Validation of spatial transcriptomic results using q-PCR (log2 fold change). Gene expression levels for selected targets were quantified in colon samples from symptomatic (blue) and asymptomatic (red) mice (n = 6 per group). Expression was normalized to β-actin, and relative expression was calculated using the –ΔΔCt method, then transformed to log2 fold change. Each point represents an individual biological replicate. Asterisks indicate statistically significant differences between groups (p < 0.01, unpaired t-test).

Full size image

Pathway analysis

Fig. 9
figure 9

Volcano plot: symptomatic CDI mice upregulated and downregulated pathways compared with infected asymptomatic mice, showing 793 pathways, 350 pathways upregulated, and 443 pathways downregulated.

Full size image

After individual genes were compared, we were curious to see the difference in pathways between symptomatic and asymptomatic C. difficile-infected mice. Figure 9 shows the volcano plot compared upregulated and downregulated pathways in symptomatic versus asymptomatic mice, mainly highlighting pro-inflammatory pathways. Among the most upregulated pathways were Toll-like receptors, C-type lectin receptors (CLRs), IL-1, and IL-17. We focused on IL-17 because of the significant upregulation of this pathway and genes belonging to this pathway in symptomatic mice. In addition, IL-17 stimulates other proinflammatory pathways, including the MAPK signaling pathway and NF-κB signaling pathway. The IL-17 signaling pathway was significantly upregulated, covering 100% of the related genes, with a positive normalized enrichment score (NES) and a significant adjusted p-value of 0.018. This pathway activated downstream pathways like MAPK and NF-κB signaling. Some of the upregulated genes in the NF-κB and MAPK signaling pathways include Cd14, Jun, Mapk8, and Areg, with LFC of 1.38, 1.37, 1.05, and 1.8, respectively. The MaPK signaling pathway had 80% coverage, a positive NES, and an adjusted p-value of 0.01. The TRAF-mediated NF-κB and MaP kinase activation pathways showed 98.55% coverage, positive NES, and a significant adjusted p-value of 0.003. Conversely, pathways significantly downregulated in symptomatic mice included antigen processing and presentation, metabolic pathways for Lipids, amino acids, and the PPAR signaling pathway. These pathways had negative NES values and p-values below 0.05. Specifically, the MHC class II antigen presentation pathway had a significant downregulation, with a p-value of 0.001. Similarly, the antigen processing-cross presentation pathway showed significant downregulation with a negative NES and a p-value of 0.03.

Discussion

The primary goal of this research was to explore the functional and spatial roles of genes and pathways altered in symptomatic mice to understand C. difficile pathogenesis, particularly the inflammatory pathways’ contribution to the clinical signs and lesions of CDI. This study investigated the differential gene expression in the intestinal mucosa of mice infected with C. difficile. Symptomatic infected mice exhibited altered gene expression compared to the asymptomatic mice. Generally, inflammatory mediators related to epithelial damage and ulceration were found to be upregulated in the symptomatic mice (e.g. Cxcl2, Lcn2, and NF-kB). In addition, pro-inflammatory pathways (e.g. IL-17) were significantly upregulated in symptomatic mice36,37,38.

In this study, approximately 30% of mice consistently remained asymptomatic following exposure to a standardized infectious dose (5 × 10⁵ CFU) of the virulent, toxigenic Clostridioides difficile strain ATCC 43255 (ribotype 087), despite uniform pre-treatment with a five-antibiotic cocktail followed by clindamycin. This observation aligns with previous reports documenting asymptomatic outcomes in 20–40% of infected mice, even when infection parameters such as strain, dose, and antibiotic preconditioning are standardized30. In our model, all animals received identical bacterial and treatment protocols; however, a subset showed no histologic lesions, neutrophilic infiltration, or upregulation of proinflammatory genes. These findings suggest that, even under controlled conditions, intrinsic host factors—such as mucosal immune tolerance, microbiota composition and resilience, and individual inflammatory thresholds—may drive divergent clinical outcomes. This shows the complexity of C. difficile pathogenesis and highlights the biological and translational relevance of asymptomatic carriage in murine CDI models.

The IL-17 pathway is significantly upregulated in symptomatic mice, playing a critical role in host defense against infections39. This pathway is essential for recruiting immune cells, like neutrophils, to the site of infection and regulating inflammation40,41,42,43. The role of IL-17 in CDI has been studied, illustrating its involvement in both host defense and exacerbation of disease. A study on cytokine profiles in CDI patients found that IL-17 levels were significantly elevated in severe cases, with a shift from Th1 to Th17 dominance observed in these patients. This shift suggests that IL-17 and the Th17 response contribute to advanced CDI’s intense inflammation and tissue damage, marking IL-17 as both a severity marker and a potential cause of pathogenic immune responses in severe cases​44. Another study expanded on these findings by examining IL-17 production by γδ T cells, showing that these cells are a primary source of IL-17 during CDI and play a protective role in the immune response. In murine models, IL-17-producing γδ T cells were essential for recruiting neutrophils to infection sites, facilitating pathogen clearance while managing inflammation. Mice deficient in γδ T cells exhibited increased susceptibility to CDI, indicating that IL-17 from γδ T cells is crucial for effective host defense, especially in early infection stages42. A recent study focused on IL-17’s role in recurrent CDI (RCDI), where elevated IL-17 levels were observed in an RCDI mouse model. Treatment with a RORγt inhibitor to block Th17 function significantly improved clinical signs, including reduced inflammation and restored gut barrier integrity. These findings show IL-17’s role in perpetuating inflammation in recurrent cases and suggest that targeting Th17 pathways may offer therapeutic potential for managing RCDI ​43. In this study, chemokines like Cxcl2 were upregulated, reflecting active recruitment of immune cells in symptomatic mice. Molecules such as defensins, occluding (Ocln), and Reg3g, which are involved in the IL-17 signaling pathway, showed significant differences in expression in expression in symptomatic mice versus asymptomatic mice. Defensins were notably downregulated, indicating a weakened innate immune defense, while Reg3g and Ocln were upregulated, reinforcing gut barrier integrity45. Muc1, typically involved in mucosal barrier protection46,47, did not show significant changes in this study, suggesting its role may not be as prominent. The study showed that while some mucosal immunity genes, like Reg3g and Ocln, were more highly expressed in the superficial intestinal layers of symptomatic mice, overall changes across all layers were insignificant. The significant downregulation of defensins in symptomatic mice likely weakens the host’s ability to fight bacterial infections, particularly in CDI48. The study showed that the general decrease in defensin expression compromises the host’s defense mechanisms despite localized upregulation of some immune-related genes49. Since IL-17 is a critical pathway in inflammatory bowel disease (IBD) and ulcerative colitis (UC)50, it is currently targeted via monoclonal antibodies (e.g. secukinomab) for these conditions51,52. These findings suggest that targeting the IL-17 signaling pathway, particularly at the level of receptors or key signaling molecules like Traf6, could offer an effective therapeutic strategy for CDI. It remains unclear whether the upregulation of the IL-17 pathway observed in symptomatic mice directly results from increased neutrophil infiltration or if the pathway itself induces neutrophil recruitment. This distinction is critical for understanding the exact role of IL-17 signaling in CDI pathology and could influence future therapeutic strategies to modulate the inflammatory response. Although IL-17 pathway activation is a known response to C. difficile-induced inflammation, these data suggest its further upregulation is associated with symptomatic disease rather than asymptomatic infection. This supports the idea that modulation of the host immune response, rather than the mere presence of bacteria may be the ultimate determinant of disease severity.

The upregulation of Cxcl2 in symptomatic mice likely contributes to the excessive neutrophil infiltration seen in the inflamed intestinal tissue, exacerbating mucosal damage and contributing to the pathology of the disease53. While neutrophil recruitment is essential for pathogen clearance, it can also lead to tissue injury, as seen in the characteristic signs of pseudomembranous colitis54. Blocking Cxcl2 or its associated pathways could be a potential therapeutic strategy to reduce neutrophil-driven damage while preserving the immune system’s ability to combat C. difficile infection, potentially improving disease outcomes in CDI. Lipocalin-2 (Lcn2) was also upregulated in symptomatic mice. Lcn2 is a glycoprotein involved in the innate immune response and inflammation, particularly through the IL-17 signaling pathway55. The role of Lcn2 in neutrophil recruitment and inflammation56 suggests its relevance to the clinical signs of CDI, where neutrophil-driven tissue damage is a hallmark. The increased expression of Lcn2 in symptomatic mice indicates its contribution to the inflammatory processes that exacerbate CDI. Given that Lcn2 is also involved in inflammation and cancer progression57,58,59, it is plausible that its upregulation enhances neutrophil activity and perpetuates the inflammatory response seen in CDI.

S100a8 and S100a9 were among the most upregulated genes in symptomatic mice, correlating with the clinical signs of CDI. The upregulation of S100a8 and S100a9 supports the inflammatory response in CDI, as these genes are known to play a crucial role in neutrophil recruitment and activation60. Their increased expression in symptomatic mice suggests that they significantly contribute to the neutrophil-driven immune response, leading to tissue damage and exacerbating the clinical manifestations of CDI. The increased activity of these genes amplifies the inflammatory lesions characteristic of the infection, highlighting their roles in the disease’s progression and severity. The S100 protein family, particularly S100a8, is essential for regulating cell cycle progression and differentiation. This protein acts as a danger-associated molecular pattern (DAMP), crucial in alerting the innate immune system61. It triggers immune responses by interacting with pattern recognition receptors such as Toll-like receptor 4 (TLR4) and the receptor for advanced glycation end products (Ager)62. Ager was significantly upregulated when comparing symptomatic and asymptomatic mice. Micro- or macro-molecules that block the function of these molecules might provide an effective approach to mitigate CDI symptom onset.

Some genes with an anti-inflammatory function were also upregulated including Dusp1. Dusp1, known as dual-specificity phosphatase-1 or MAPK phosphatase-1, and is a negative regulator of MAPK activity63. Upregulation of Dusp1 was also seen in our previous study27. It is part of the MAPK phosphatase family and is crucial for anti-inflammatory responses64. The increased expression of Dusp1 in symptomatic mice indicates an attempt by the host to activate anti-inflammatory responses. However, this response seems inadequate as the overall MAPK signaling pathway activity, which typically drives inflammation, remains elevated in symptomatic mice compared to asymptomatic ones. Research involving the silencing of the Dusp1 gene using lentiviral vector-mediated siRNA in mice with acute pancreatitis demonstrated a higher release of pro-inflammatory cytokines65.

The expression of genes involved in the arachidonic acid pathway did not change significantly in this study. A human study found that patients who received NSAIDs had slightly lower mortality rates than those who did not; however, this difference was not statistically significant66. We aimed to explore the effects of NSAIDs in a mouse model. In the current study, Fabp6 and Fads2 were significantly downregulated in symptomatic mice, while Fabp4 was significantly upregulated. These genes, involved in lipid metabolism and transport, are linked to the arachidonic acid pathway by regulating the availability of fatty acids for inflammatory mediators67. The downregulation of Fabp6 and Fads2 suggests a disruption in lipid metabolism, potentially exacerbating the inflammatory environment, while the upregulation of Fabp4 indicates a possible compensatory mechanism in lipid handling in symptomatic mice. The downregulation of these genes suggests a disruption in lipid metabolism, potentially exacerbating the inflammatory environment in symptomatic mice, which may have contributed to the worsened outcomes observed in the prior NSAID treatment study.

Cd74 expression was significantly downregulated in symptomatic mice compared to asymptomatic mice, which may have important implications for the immune response during CDI. Cd74 is essential for proper immune function, particularly in antigen-presenting cells (APCs)68, where it acts as a chaperone for MHC class II molecules and supports the activation of immune pathways like NF-κB and ERK69. This reduction in Cd74 may impair immune surveillance and hinder the host’s ability to present antigens to clear the infection effectively. The decreased expression of Cd74 in symptomatic mice suggests that the immune response is compromised, potentially contributing to the severity of CDI. While reduced Cd74 might prevent excessive inflammation in some contexts, it may also weaken the ability to eliminate pathogens during infection70.

B2m gene was downregulated in symptomatic mice, which likely plays a role in the impaired immune response observed during CDI. B2m encodes beta-2 microglobulin, an essential component of MHC class I molecules for presenting antigens to cytotoxic T cells71. This process is vital for immune surveillance and for efficiently eliminating infected cells72. The downregulation of B2m may reduce the ability to present antigens, weakening the host’s immune defense against infection. This immune dysfunction could exacerbate the severity of the disease. The downregulation of Cd74 and B2m suggests a broad impairment in the immune system’s ability to present antigens and mount an effective immune response. Enhancing Cd74 and B2m expression could be a beneficial therapeutic strategy, improving antigen presentation and pathogen clearance in CDI, thereby mitigating the disease severity.

Gstm1 gene was downregulated in symptomatic mice, likely contributing to the increased oxidative stress observed during CDI. Gstm1 encodes the enzyme glutathione S-transferase M1, a key player in detoxification of free radicals and regulating oxidative stress73. This enzyme neutralizes reactive oxygen species (ROS) and harmful electrophilic compounds that can damage cells and tissues74. Oxidative stress and free radical damage are significant contributors to the pathophysiology of CDI75 and the decreased expression of Gstm1 partially explains the shift toward a pro-oxidative state in symptomatic mice. This shift likely worsens the inflammatory damage to the mucosa, leading to the clinical signs observed in CDI, such as tissue necrosis and ulceration. Restoring Gstm1 expression or targeting oxidative stress pathways could offer therapeutic potential to mitigate the damage caused by ROS and improve recovery in CDI.

H3c2 and H4c17 genes were downregulated in symptomatic mice, impacting the formation of neutrophil extracellular traps (NETs) during CDI. NETs, composed of decondensed chromatin and histones such as H3 and H4, play a crucial role in the immune response by trapping76 and neutralizing pathogens like C. difficile. The downregulation of these histone genes in symptomatic mice could compromise the ability of acute inflammatory response to sequester the infectious agent, allowing C. difficile to evade immune responses and exacerbate inflammation. This reduction in NET formation might explain the heightened infection severity and tissue damage observed in symptomatic mice6. Targeting pathways involved in NET formation or enhancing the expression of H3c2 and H4c17 may represent potential therapeutic strategies to boost immune defenses and improve pathogen clearance in CDI. Based on the findings of this study, significant differences in gene expression were observed between symptomatic and asymptomatic mice. These differences could explain why some mice remained asymptomatic. Several factors might have influenced this, including variability in host responses and immune regulatory mechanisms. Differences in immune responses to the pathogens along with variations in gene expression and pathway activation, could have contributed to the differences in disease presentation, even under identical experimental conditions. Another important factor that could have played a role is the gut microbiome. However, this study did not include the gut microbiome factor as a variable. Given the identical conditions of these inbred mice, this would be surprising. To better understand how the immune system behaves in different layers of the intestine during a C. difficile infection, we compared gene expression between the superficial and deep layers of the intestinal lining. As different layers of the gastrointestinal tract contain distinct populations of immune cells77, analyzing these layers separately can provide more insight into spatial immune responses. This is the first study to examine gene expression with this layer-specific detail in this context. We found that specific inflammation-related genes, including S100a8, S100a9, and Cxcl2, were more active in the superficial layer, suggesting that this region may play a frontline role in activating the body’s innate immune defenses. Meanwhile, the gene Lcn2 was more active in the deeper mucosa, indicating different roles in antimicrobial defense depending on tissue depth. Some genes, like Cd74 and Gstm1, were less active overall but showed variation between layers, highlighting the complexity of immune regulation. Understanding how gene expression differs by layer might help shape better diagnostic tools, targeted treatments, and more personalized approaches to treating intestinal infections.

To validate our spatial transcriptomic findings, we performed qPCR and presented the results as log2 fold change (–ΔΔCt). This approach confirmed that S100a8, S100a9, Lcn2, Cxcl2, and Dusp1 were significantly upregulated, while Cd74 was downregulated in symptomatic mice. The agreement between spatial transcriptomics and qPCR highlights the reproducibility and reliability of our findings. These validated gene expression changes reflect the activation of innate immune pathways and epithelial stress responses that contribute to the pathogenesis of C. difficile infection in symptomatic animals.

In the previous study employing RNA sequencing27, several genes were identified as significantly upregulated in symptomatic mice, including Lcn2, Cxcl2, S100a9, Trem-1, Duox2, OSM, and MMP8. Validation of many of these genes was performed using qPCR analysis. In the present investigation, spatial transcriptomic analysis was conducted on tissue samples collected from the same cohort, and upregulation of Lcn2, Cxcl2, and S100a9 in symptomatic mice was again observed. Additional insights were provided by spatial transcriptomics through the identification of region-specific gene expression patterns not captured by pooled RNA-seq. For example, elevated expression of Lcn2 was detected in the deeper mucosal layers, whereas Cxcl2 and S100a9 were found to be more prominently expressed in the superficial layers. Furthermore, spatial profiling enabled the identification of previously unreported downregulated genes in symptomatic mice, including Cd74, B2m, Gstm1, and histone genes such as H3c2 and H4c17. These findings demonstrate the complementary nature of spatial transcriptomics in enhancing traditional RNA-seq approaches by providing anatomical context to immune and metabolic gene expression during Clostridioides difficile infection.

One of the limitations of our study is that symptomatic mice were sacrificed when they became severely ill with clinical signs. In contrast, asymptomatic mice were observed and sacrificed one-week post-infection. The difference in timing of tissue collection could introduce temporal gene expression variance independent of clinical status. While our goal was to capture gene expression linked with active disease compared to asymptomatic colonization, some of the differences we have observed may be influenced by time-dependent transcriptional change. Experiments in the future using time-matched control groups or longitudinal sampling would delineate the effects of disease severity from those of time post-infection. While differences in timing represent one limitation, another is the relatively small number of mice used for spatial transcriptomic analysis. This constraint was primarily due to the high cost and technical complexity of the NanoString GeoMx DSP platform. However, these samples were obtained from a well-characterized murine model previously published by our group27 supporting our findings’ biological relevance. Although the limited sample size may affect the generalizability of results, we addressed this by independently validating key gene expression changes using qPCR on a larger set of biological replicates. Future studies involving larger cohorts and broader sampling will be critical to confirm and expand on these observations.

Conclusions

Our research elucidates the intricate interactions of various genes and pathways contributing to the pathology attribute to CDI. We identified several genes and pathways linked to inflammation, immune response, and cellular metabolism by comparing gene expression profiles between symptomatic and asymptomatic mice post-CDI. Disruption in these genes impair the host’s capacity to mount effective inflammatory and immune responses, leading to heightened susceptibility to CDI-induced tissue damage and exacerbated clinical signs. This study is the first to evaluate differential gene expression between the superficial and deep layers of the intestinal mucosa. Our findings reveal specific genes are altered in the superficial versus deep intestinal mucosal layers after CDI. This dimensional analysis provides deeper insights into the spatial regulation of gene expression in the context of CDI. Future research will focus on targeted interventions to modulate expression of these genes and pathways or their activity, which could pave the way for innovative therapeutic strategies against CDI. Given that some upregulated pathways, such as the IL-17 signaling pathway, have pro-inflammatory and anti-inflammatory roles, our approach will focus on precisely targeting select upregulated genes rather than broadly suppressing entire pathways. This refined strategy aims to effectively reduce CDI symptoms while preserving the beneficial aspects of these pathways, offering more targeted therapeutics.