Colonic spatial single-cell proteomics and murine models link mitochondrial dysfunction to dimeric IgA-secreting plasma cell deficiency in Crohn's disease

Abstract

Secretory IgA (SIgA) is critical for maintaining the intestinal barrier. A dysregulated B-cell compartment and altered Ig secretion have been well documented in Crohn's disease (CD) patients, although their origin is unknown. To unravel the role of mucosal humoral immunity in CD pathogenesis, we in-depth phenotype colonic plasma cell (PC) differentiation in CD at the single-cell level, linked to ex vivo functional characterization and experimental mouse models with a congenital mitochondrial defect or under glucose-free high-protein dietary intervention. Here, we demonstrate that despite expanded colonic B cells, CD patients in remission present significantly diminished mucosal dimeric IgA and fecal SIgA. Colonic plasmablasts and immature CD19⁺CD45⁺ PCs are increased at the expense of the mature CD19^-CD45^- phenotype. Accordingly, CD-derived ex vivo differentiated PCs display impaired maturation into dimeric IgA-secreting PCs. In this study, patient-derived data from colonic RNA-seq, spatial single-cell proteomics, and plasma metabolomics are combined with data from both mouse models and highlight the crucial role of mitochondrial oxidative phosphorylation in colonic IgA⁺-PC differentiation, suggesting promising directions for future therapeutic strategies.

Fig. 1. Colonic B cells and PBs but not CD138⁺ PCs are expanded in CD^rem.

a RNA-seq, SSCP, and quantification of B-cell marker mRNA expression in isolated LPMNCs were performed utilizing colonic biopsies from age- and sex-matched CD^rem patients and non-IBD controls. b Volcano plot of differentially expressed genes (DEGs; pink: increased, blue: decreased) identified by RNA-seq (CD^remn = 7, non-IBD n = 12) displaying unadjusted P values. c Top 40 upregulated REACTOME pathways according to gene set enrichment analysis (GSEA). The color of the dots corresponds to the adjusted P value. d Significantly upregulated DEGs linked to B cells (pink), MHC class II complex (orange), or immunoglobulin (Ig) chains (turquoise). e Percentage of cells positive for different immune cell markers in the colonic LP of CD^rem patients (n = 6) versus non-IBD controls (n = 6) was determined using SSCP. f Representative images of SSCP of colonic mucosa from CD^rem patients (n = 6, right) versus non-IBD controls (n = 6, left) for CD19 (magenta), CD27 (purple), and CD38 (greenish yellow), and DAPI (gray). Scale bar: 10 µm. g LPMNCs were isolated from colon biopsies of CD^rem patients (n = 6) and non-IBD controls (n = 5) by enzymatic digestion. h Quantification of B-cell markers in CD^rem-derived (n = 6) colonic LPMNCs by RT-qPCR is displayed as median relative to non-IBD (n = 5) (left). CD19 and CD38 mRNA expression was increased >2-fold in colonic LPMNCs isolated from CD^rem patients (right). i Scheme of a simplified differentiation pathway of the B-cell lineage, including crucial surface markers. e Two-tailed multiple Mann–Whitney U-test. h Two-way ANOVA with Šídák’s multiple comparisons test. Box plots depict the median and IQR with whiskers indicating the range. *P ≤ 0.05. IQR interquartile range, SSCP spatial single-cell proteomics. Parts of (a, g, i) were created in BioRender. Jäschke, S. (2026) https://BioRender.com/0mfac2x. Some elements in (i) were provided by Servier Medical Art (https://smart.servier.com).

Fig. 2. CD^rem patients display a higher abundance of colonic CD19⁺CD27⁺CD38⁺CD138^- PBs but not CD138⁺ PCs and reduced fecal IgA levels.

a Immunoglobulin levels were quantified in blood plasma (CD^remn = 10, non-IBD n = 21) and colonic tissue (CD^remn = 7, non-IBD n = 12) using LC-MS proteomics or RNA-seq, respectively. Circle color reflects the P value of the Mann–Whitney U-test comparing CD^rem versus non-IBD, with larger circles indicating significant differences (P ≤ 0.05). b Colonic RNA-seq data of Ig heavy chains of CD^rem patients (n = 7) and non-IBD controls (n = 12) are displayed as transcripts per million (TPM). c Quantification of colonic IgA, IgM, and BCMA in CD^rem patients (n = 9) versus non-IBD controls (n = 13) via Western blotting. d Representative images of IHC staining of colonic IgA (CD^remn = 8, non-IBD n = 13) (left). Total IgA staining in the LP was normalized to nucleic hematoxylin staining (right). e Fecal SIgA levels were quantified by IgA-specific ELISA experiments (CD^remn = 44, non-IBD n = 85). f Schematic representation of epithelial IgA transcytosis and SIgA formation. g Representative images of SSCP of a plasmablast (PB) and a PC showing CD19 (magenta), CD27 (purple), CD38 (greenish yellow), CD138 (cyan), and DAPI (gray) (left). Relative abundance of CD19⁺CD27⁺CD38⁺CD138^- PBs, and CD138⁺ PCs in the colonic LP of CD^rem patients (n = 6) compared to non-IBD (n = 6; median set to 1) (right). a–e Two-tailed (Multiple) Mann–Whitney U-test. g Two-way ANOVA with Šídák’s multiple comparisons test. Box plots depict the median and IQR with whiskers indicating the range. *P ≤ 0.05, ***P ≤ 0.001. IQR interquartile range, SSCP spatial single-cell proteomics. Parts of (a, f) were created in BioRender. Jäschke, S. (2026) https://BioRender.com/0mfac2x.

Fig. 3. Colonic low dIgA-secreting CD19⁺CD45⁺CD138⁺ PCs dominate over well-differentiated CD19^-CD45^-CD138⁺ PCs in CD^rem patients.

a Colonic PCs of CD^rem patients (n = 6) and non-IBD controls (n = 6) were characterized using the 64-plex Immuno-Oncology protein panel on the CosMx^TM spatial molecular imager (SMI) in two independent experiments. b A total of 28,895 CD138⁺ PCs (blue) were identified in a UMAP of all colonic lamina propria (LP) cells (Vim⁺) of the first spatial single-cell proteomics experiment and c reclustered in a UMAP with coloring for non-IBD controls (n = 3) and CD^rem patients (n = 3). d Scheme comparing characteristics of intestinal IgA-secreting CD138⁺ PC subpopulations^,,. Representative images of mucosal CD19⁺CD45⁺ (left), CD19^-CD45⁺ (middle), and CD19^-CD45^- PCs (right) (CD138: cyan, CD19: magenta, CD45: green, DAPI: gray). Scale bar: 2 µm. e Relative levels of CD138, NF-κB p65, pan-RAS, and relative percentage of Ki-67⁺ cells in mucosal PC subpopulations of six non-IBD controls, related to the CD19⁺CD45⁺ PC population per patient. f UMAP displaying the CD138⁺ PC subtype distribution in colon biopsies from non-IBD controls (n = 3) and CD^rem patients (n = 3). g Fractions of CD19⁺CD45⁺ (magenta), CD19^-CD45⁺ (dark green), CD19^-CD45^- (turquoise) and other (gray) Vim⁺CD138⁺ PCs in colon biopsies of non-IBD controls (n = 6; top) and CD^rem patients (n = 6; bottom). h Percentages of colonic PC subpopulations in CD^rem patients (n = 6) compared to non-IBD controls (n = 6). i Quantification of mIgA/total IgA, dIgA/total IgA, and dIgA/mIgA ratios in colonic biopsies of CD^rem patients (n = 9) and non-IBD controls (n = 13) via Western blotting. j The ratio of dIgA/BCMA in the colonic mucosa of CD^rem (n = 9) versus non-IBD (n = 13) was quantified using Western blotting. k Upregulated (left; P ≤ 0.05) and nonregulated (right) cytokines and chemokines crucial for B-cell/PC development were quantified using RNA-seq and are displayed as scaled transcripts per million (TPM) in the colonic mucosa of CD^rem patients (n = 7) versus non-IBD controls (n = 12). e Two-way ANOVA with Tukey’s multiple comparisons test (left) and Friedman test with Dunn’s multiple comparisons test (right), h–k Two-tailed (multiple) Mann–Whitney U-test. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. Box plots depict the median and IQR with whiskers indicating the range. AU arbitrary units, IQR interquartile range, UMAP uniform manifold approximation and projection. Parts of (a) were created in BioRender. Jäschke, S. (2026) https://BioRender.com/0mfac2x.

Fig. 4. Ex vivo IgA-secreting PC differentiation is impaired for CD^rem-derived colonic switched memory B cells (sMBCs).

a Representative images from six patients each in two independent SSCP experiments, showing a PC (red arrow) and a memory B cell (white arrow) making contact to a CD4⁺ T cell (blue arrow) in the colonic mucosa displaying CD19 (magenta), CD38 (greenish yellow), CD27 (purple), CD138 (cyan), CD4 (blue), CD20 (yellow) and DAPI (gray). Scale bar: 5 µm. b SMBCs were enriched from patient-derived colonic LPMNCs (see Fig. 1g; CD^rem patients n = 6, non-IBD controls n = 9) using negative selection magnetic-activated cell sorting (MACS) and were ex vivo differentiated into mature PCs for 14 days. c Secretion of dIgA, mIgA, and IgG by ex vivo differentiating PCs during Day 0–4, Day 4–7, Day 7–10, and Day 10–14 of cultivation (lanes 5-8) was demonstrated by Western blot experiments with respective IMDM samples (+cytokines) (lanes 1–4) as controls. d Expression of PC differentiation markers in patient-derived ex vivo differentiated colonic PCs on Day 14 was quantified using RT-qPCR and depicted in a heatmap as median values for CD^rem (SDC1n = 2, PRDM1/TNFRSF17/XBP1 unspl./XBP1 spl./CD27/CD38/CD19/PTPRCn = 5, MZB1n = 6, TNFRSF13Bn = 4) and non-IBD (SDC1/CD27/TNFRSF13B/PTPRCn = 7, PRDM1/TNFRSF17/XBP1 unspl./XBP1 spl./CD38/CD19n = 8, MZB1n = 9). e Expression of IGHA1, IGHA2, and JCHAIN for CD^rem- (IGHA1n = 5, IGHA2/JCHAINn = 6) versus non-IBD-derived (n = 9) ex vivo differentiated colonic PCs. f Ratio (left) and correlation (right) of IGHA1 and IGHA2 expression in CD^rem- (n = 5) and non-IBD-derived (n = 9) ex vivo differentiated colonic PCs on Day 14. g Normalized quantification (left) and correlation (right) of total and dimeric IgA secretion by ex vivo differentiated colonic PCs from CD^rem patients (n = 6) and non-IBD controls (n = 9; median set to 1), measured between Day 10–14 using ELISA experiments. h Normalized IgG secretion by CD^rem-derived ex vivo differentiated colonic PCs (n = 6) between Day 10–14 compared to non-IBD-derived PCs (n = 9; median set to 1) was quantified by Western blotting. d–f, h Two-tailed (multiple) Mann–Whitney U-test. f, g Spearman correlation and linear regression with 95% confidence interval (dashed lines). g Two-tailed Mann–Whitney U-test (total IgA) and t-test with Welch’s correction (dIgA). Box plots depict the median and IQR with whiskers indicating the range. ^#P ≤ 0.1, *P ≤ 0.05. IQR interquartile range, SSCP spatial single-cell proteomics. Parts of (b) were created in BioRender. Jäschke, S. (2026) https://BioRender.com/0mfac2x.

Fig. 5. Colonic PCs of CD^rem patients display increased surface levels of the NADase CD38.

a SSCP data of colonic CD19⁺CD45⁺, CD19^-CD45⁺, and CD19^-CD45^- PCs (CD^rem patients n = 6, non-IBD controls n = 6). Results are presented in a dotplot as median protein levels per cell (circle color) and percentage of positive cells (circle size). b Representative images of CD38 staining (greenish yellow) in a colonic CD138⁺ PC of CD^rem (n = 6) versus non-IBD (n = 6). Scale bar: 5 µm. c UMAP displaying the minimum (blue) and maximum (magenta) CD38 levels within all colonic PCs of the first SSCP experiment (CD^rem patients n = 3, non-IBD controls n = 3). d Quantification of CD38 levels in CD19⁺CD45⁺, CD19^-CD45⁺, and CD19^-CD45^- colonic PCs of CD^rem patients (n = 6) compared to non-IBD controls (n = 6) using SSCP. Data were presented as median CD38 level per cell. e Schematic representation of mitochondrial NAD⁺ metabolism, showing how increased NAD⁺-hydrolyzing activity of surface CD38 contributes to mitochondrial dysfunction. f Correlations of CD38 with NLRP3, CASP1, and PYCARD expression in the colon of CD^rem patients (n = 7) and non-IBD controls (n = 12) displaying Spearman correlation coefficients. d Two-way ANOVA with Tukey’s multiple comparisons test, f Spearman correlation and linear regression with 95% confidence interval (dashed lines). Box plots depict the median and IQR with whiskers indicating the range. *P ≤ 0.05, **P ≤ 0.01. CAC citric acid cycle, (c)ADPR (cyclic) ADP-ribose, ETC electron transport chain, IQR interquartile range, MFI median fluorescence intensity, NAD⁺/NADH nicotinamide adenine dinucleotide, NAM nicotinamide (a form of vitamin B3), NAMPT Nicotinamide phosphoribosyltransferase, NMN nicotinamide mononucleotide, NMNAT Nicotinamide mononucleotide adenylyltransferase, R5P ribose 5-phosphate, ROS reactive oxygen species, SSCP spatial single-cell proteomics, UMAP uniform manifold approximation and projection. Parts of (e) were created in BioRender. Jäschke, S. (2026) https://BioRender.com/0mfac2x.

Fig. 6. Mitochondrial OXPHOS is impaired in CD patient-derived colonic PCs.

a Representative image from three biological replicates showing immunofluorescence co-staining of a colonic PC for IgA and OXPHOS-promoting mitochondrial P32. Scale bar: 5 µm. b PCA of metabolites measured in plasma samples of CD^rem (red, n = 9) and non-IBD (blue, n = 22) individuals using in vitro diagnostic research (IVDr) proton nuclear magnetic resonance (¹H NMR) spectroscopy. c Significantly regulated plasma metabolites (top left) and lipoproteins (bottom left) in CD^rem (n = 9; bars) compared to the reference average of non-IBD controls (n = 22; dotted center line), with upregulated ketone bodies and CAC intermediates highlighted in a blue box. Plasma L-lactate/D-glucose ratio was determined for CD^rem (n = 9) versus non-IBD (n = 22) (right). d Gene set enrichment analysis (GSEA) of colonic RNA-seq data for the MSigDB “HALLMARK_OXIDATIVE_PHOSPHORYLATION” gene set in CD^rem (n = 7) versus non-IBD (n = 12). NES normalized enrichment score. e Downregulated REACTOME pathways in the colon of CD^rem patients (n = 7) versus non-IBD controls (n = 12) according to GSEA. The color of the dots corresponds to the adjusted P value. f Significantly downregulated colonic gene transcripts linked to mitochondrial function in CD^rem (n = 7) compared to non-IBD (n = 12) are displayed as scaled transcripts per million (TPM). g Schematic overview of all significantly downregulated colonic transcripts encoding subunits of the mitochondrial electron transport chain (ETC) complexes I, III, IV, and V in CD^rem. h Normalized D-glucose consumption and L-lactate release were quantified in the supernatant of CD^rem- (n = 3) and non-IBD-derived (n = 7) ex vivo differentiated colonic PCs between Day 10–14. i Expression of metabolic enzyme(s) (subunits) and nutrient transporters in non-IBD- (n = 9) and CD^rem-derived (n = 5) ex vivo differentiated colonic PCs on Day 14 was quantified using RT-qPCR and depicted as median relative to non-IBD, with mitochondrial genes highlighted in a green box. c, f Two-tailed (multiple) Mann–Whitney U-test, h, i Two-way ANOVA with Šídák’s multiple comparisons test (h) or Fisher’s least significant difference (LSD) test (i). Box plots depict the median and IQR with whiskers indicating the range. *P ≤ 0.05. IQR interquartile range. Parts of (d, g) were created in BioRender. Jäschke, S. (2026) https://BioRender.com/0mfac2x.

Fig. 7. Mitochondrial dysfunction impairs colonic PC differentiation and lowers IgA in mice.

a Mutation in subunit 8 of the ATP synthase and published metabolic imbalance^,. b Schematic of flow cytometry of lymphocytes from mesenteric lymph nodes (MLNs) and Peyer’s patches (PPs) and LPS + IL4-stimulated splenic B cells of male and female Atp8-mutant and B6-WT mice. c Representative contour plots of mitochondrial mass (MitoTracker Green) and membrane potential (TMRE, tetramethylrhodamine ethyl ester) (left). Percentage of TMRE^high cells in LPS + IL4-stimulated or unstimulated B220⁺ cells from Atp8-mutant (n = 3) versus B6-WT mice (n = 3) (right) from two independent experiments. d Proliferation of Atp8-mutant (n = 3) and B6-WT (n = 3) splenic B cells under LPS + IL4 stimulation. e B220 level of ex vivo differentiated PBs from Atp8-mutant mice (n = 3) versus B6-WT-derived PBs (n = 3). f Median percentages of B220⁺ (dark blue) and B220^- (turquoise) CD138⁺ PCs isolated from MLNs and PPs of Atp8-mutant (n = 3) and B6-WT mice (n = 3) in two experiments. g Colonic expression of B-cell/PC markers was quantified for male Atp8-mutant (Cd19n = 9, Ighan = 12, Tnfrsf17n = 11, Sdc1n = 10), and B6-WT mice (n = 12) via RT-qPCR. Data were displayed relative to B6-WT mice for three sampling rounds, with box plots representing median and IQR and whiskers indicating the range. h Pairwise correlation matrix of colonic C1qbp and B-cell/PC marker expression for Atp8-mutant (n = 9–12) and B6-WT mice (n = 12) displaying Spearman correlation coefficients. i Correlation of colonic Igha and C1qbp expression in Atp8-mutant (n = 12) and B6-WT mice (n = 12). j Representative colonic IgA IHC staining of male Atp8-mutant (n = 3) and B6-WT mice (n = 3) (left) and quantification of IgA⁺ cells (right), with mice sampled on two days. k Hypothesized metabolic switch upon glucose-free high-protein (GFHP) dietary intervention. l Representative colonic IgA IHC staining of female GFHP diet (n = 5) versus isocaloric chow diet-fed mice (n = 6) (left) from two independent experiments and quantification of IgA⁺ cells (right). c–e Two-way ANOVA with Tukey’s multiple comparisons test (c, d) or Fisher’s least significant difference (LSD) test (e). (g, j, l) Two-tailed Mann–Whitney U-test, h Spearman correlation, i Spearman correlation and linear regression with 95% confidence interval (dashed lines). Bar graphs display median with range. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. MFI mean fluorescence intensity. Parts of (a, b, and k were created in BioRender. Jäschke, S. (2026) https://BioRender.com/0mfac2x.