Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Sep 1;132(17):e153519.
doi: 10.1172/JCI153519.

The epithelial-specific ER stress sensor ERN2/IRE1β enables host-microbiota crosstalk to affect colon goblet cell development

Affiliations

The epithelial-specific ER stress sensor ERN2/IRE1β enables host-microbiota crosstalk to affect colon goblet cell development

Michael J Grey et al. J Clin Invest. .

Abstract

Epithelial cells lining mucosal surfaces of the gastrointestinal and respiratory tracts uniquely express ERN2/IRE1β, a paralogue of the most evolutionarily conserved endoplasmic reticulum stress sensor, ERN1/IRE1α. How ERN2 functions at the host-environment interface and why a second paralogue evolved remain incompletely understood. Using conventionally raised and germ-free Ern2-/- mice, we found that ERN2 was required for microbiota-induced goblet cell maturation and mucus barrier assembly in the colon. This occurred only after colonization of the alimentary tract with normal gut microflora, which induced Ern2 expression. ERN2 acted by splicing Xbp1 mRNA to expand ER function and prevent ER stress in goblet cells. Although ERN1 can also splice Xbp1 mRNA, it did not act redundantly to ERN2 in this context. By regulating assembly of the colon mucus layer, ERN2 further shaped the composition of the gut microbiota. Mice lacking Ern2 had a dysbiotic microbial community that failed to induce goblet cell development and increased susceptibility to colitis when transferred into germ-free WT mice. These results show that ERN2 evolved at mucosal surfaces to mediate crosstalk between gut microbes and the colonic epithelium required for normal homeostasis and host defense.

Keywords: Cell stress; Gastroenterology; Inflammatory bowel disease.

PubMed Disclaimer

Conflict of interest statement

Conflict of interest: JR Turner is a founder and shareholder of Thelium Therapeutics and has served as a consultant for Entrinsic, Immunic, Johnson & Johnson, Kallyope, and 89Bio. WIL is founder of Transcera.

Figures

Figure 1
Figure 1. Ern2–/– mice have fewer goblet cells in distal colon under CONV conditions.
(A and B) Representative images of (A) AB-stained and (B) anti-MUC2 antibody–stained sections of distal colon from CONV-WT and CONV-Ern2–/– mice. Bar graphs show AB+ cells and MUC2+ cells in the upper and lower half of crypts (normalized by number of crypt epithelial cells), AB-stained theca area, crypt length, and KI67+ cells per crypt in distal colon of WT and Ern2–/– littermates. Symbols represent average values for individual mice (WT, n = 8; Ern2–/–, n = 9) from 2 independent cohorts. Data are represented as mean ± SEM. Mean values were compared by unpaired t test. Scale bars: 50 μm. (C) Violin plot shows the distribution of relative mRNA expression of epithelial cell signature genes (2) in colon crypts. G, goblet; E, enterocyte; S, stem; EE, enteroendocrine; T, tuft. Bar graph shows enrichment of differentially expressed genes. (D) Bar graphs show relative mRNA expression in colon crypts measured by qPCR. Symbols represent individual mice, and data are represented by mean ± SEM. Mean values were compared by unpaired t test. (E) Left panels: Violin plot shows the distribution of relative mRNA expression for genes in goblet cell subpopulations (3), and bar graph shows enrichment of differentially expressed genes. Right panel: volcano plot shows differential expression of conventional crypt goblet cell genes. (F) Bar graph shows the number of AB+ cells in upper half of crypts (normalized by number of crypt epithelial cells) for mice treated with or without the γ-secretase inhibitor DBZ. Symbols represent the average values for individual mice, and data are represented as mean ± SEM. Mean values were compared by 2-way ANOVA (n = 4–8 mice per group). Heatmap shows differential mRNA expression in colon epithelial cells measured by qPCR. *P < 0.05; **P < 0.01; ****P < 0.0001.
Figure 2
Figure 2. Gut microbes induce colon goblet cell development in an ERN2-dependent manner.
(A) Bar graph shows the number of AB+ cells in the upper half of crypts in the distal colon of WT and Ern2–/– under CONV conditions, GF conditions, and GF followed by colonization with gut microbes from CONV-WT donor mice (Colonized). Symbols represent the average value for an individual animal, and data are represented as mean ± SEM. Mean values were compared by 2-way ANOVA. Data for CONV mice is from Figure 1A. (B) Violin plot showing relative mRNA expression for goblet cell signature genes upregulated in CONV-WT mice compared with GF-WT mice. Relative expression for those same genes is plotted for Ern2–/– mice. (C) Bar graphs show relative mRNA expression in colon crypts from WT and Ern2–/– mice for select genes measured by qPCR. Symbols represent individual mice, and data are represented as mean ± SEM. Mean values were compared by 2-way ANOVA. (D) Scatter plot compares differential expression of goblet cell signature genes for CONV-WT versus GF-WT compared with COLONIZED-WT versus GF-WT. (E) Bar graph shows enrichment of differentially expressed goblet cell signature genes for COLONIZED-WT and COLONIZED-Ern2–/– mice compared with GF controls. The heatmap shows the relative expression of differentially expressed genes. The subset of the genes not differentially expressed (Padj > 0.01) in Ern2–/– mice is indicated. (F) Violin plot showing the relative mRNA expression of epithelial cell signature genes from colon crypts of GF-Ern2–/– mice compared with GF-WT mice. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
Figure 3
Figure 3. ERN2-mediated Xbp1 splicing and XBP1 expand ER function and prevent ER stress for goblet cell maturation.
(A) Representative images of AB-stained sections of distal colon from Xbp1fl/fl;Vil-Cre (n = 6) and Xbp1fl/fl;Vil-Cre+ (n = 12) littermates. Bar graphs show the number of AB+ cells in the upper half of crypts and the AB-stained goblet cell theca area. Symbols represent individual mice, and data are represented as mean ± SEM. Mean values were compared by unpaired t test. Scale bar: 50 μm. (B) Bar graph shows relative expression of spliced Xbp1 transcript in colon crypts from (left panel) GF-WT (n = 4) and CONV-WT mice (n = 6) and (right panel) CONV-WT (n = 9) and CONV-Ern2–/– (n = 11) mice. Symbols represent individual mice, and data are represented as mean ± SEM. Mean values were compared by unpaired t test. (C) Left panel: representative images of mouse colonoids treated with the γ-secretase inhibitor DAPT in the presence or absence of the IRE1 inhibitor 4μ8C. Scale bar: 500 μm. Center panel: differentiation status was assayed by scoring spheroid versus nonspheroid morphology. Bar graph shows the change in the percentage of colonoids with spheroid morphology relative to untreated controls for a given colonoid line within an experiment. Symbols represent independent experiments (WT, n = 5; Ern2–/–, n = 7). Mean values within genotypes were compared by 2-way ANOVA. Right panel: heatmap shows relative mRNA expression for select genes measured by qPCR from a single experiment. (D) Bar graph shows the number of AB+ cells in upper half of well-defined crypts following colonization of GF-Ern2–/– mice with a microbiota from CONV-WT donor mice in the presence or absence of TUDCA. Symbols represent the average value for an individual mouse, and data are represented as mean ± SEM (WT: GF/COLONIZED, n = 5/4; Ern2–/–: GF/COLONIZED/COLONIZED+TUDCA, n = 5/5/5). Mean values were compared by 1-way ANOVA. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
Figure 4
Figure 4. Impaired assembly of the colon mucus layer in Ern2–/– mice.
Representative images of the colon mucus layer assessed by AB/PAS staining of Carnoy’s fixed tissue from (A) CONV-WT and CONV-Ern2–/– mice and (B) Xbp1fl/fl;Vil-Cre and Xbp1fl/fl;Vil-Cre+ mice. Bar graphs show the thickness of the inner mucus layer. Symbols represent the average value of the distribution of measures around at least 2 full cross sections for individual mice (WT, n = 4; Ern2–/–, n = 5; Xbp1fl/fl;Vil-Cre, n = 3; Xbp1fl/fl;Vil-Cre+, n = 7). Data are represented as mean ± SEM. Mean values were compared by unpaired t test. **P < 0.01; ****P < 0.0001. (C) Representative images of Carnoy’s fixed colon tissue from CONV-WT and CONV-Ern2–/– mice stained with AB/PAS or the 16S rRNA in situ hybridization probe EUB338 (red) to detect lumenal bacteria. Nuclei were stained with DAPI (blue). (D) Representative images of Carnoy’s fixed tissue from CONV-WT and GF-WT mice stained with AB/PAS. Scale bars: 50 μm.
Figure 5
Figure 5. Ern2–/– mice have increased susceptibility to C. rodentium–and DSS-associated colitis.
(A and B) Time courses represent the onset of infection monitored by (A) the percentage of mice with C. rodentium detected in stool and (B) C. rodentium CFUs in stool. Time courses represent the average measured in 47 WT and 38 Ern2–/– mice from 5 independent experiments. (B) Bar graph shows C. rodentium CFUs in stool of infected mice at 8 days after infection. Symbols represent individual mice. Data are represented as mean ± SEM. Mean values were compared by unpaired t test. (C and D) Bar graphs show C. rodentium CFUs cultured from colon tissue and (D) histology scores of epithelial damage in control mice and infected mice at 8 and 13 days after infection. Symbols represent individual mice. Data are represented as mean ± SEM. Mean values were compared by 2-way ANOVA. (E and F) Time courses shows (E) change in body weight and (F) survival for WT and Ern2–/– mice during administration of DSS for 8 days followed by recovery for 14 days. Symbols represent average values (WT, n = 8; Ern2–/–, n = 8). Survival curves were compared using a log rank test. *P < 0.05.
Figure 6
Figure 6. Ern2–/– microbiota are unable to support goblet cell development when transferred into WT recipient mice.
Bar graphs show number of AB+ cells per crypt in the upper half of crypts in the distal colon of (A) GF-WT mice (n = 3) and GF-WT mice colonized with microbiota from WT (n = 9) or Ern2–/– (n = 8) donor mice and (B) CONV-WT mice (n = 11), antibiotic-treated CONV-WT mice (n = 11), and antibiotic-treated CONV-WT mice that were cohoused with WT (n = 9) or Ern2–/– (n = 7) donor mice. Symbols represent individual mice. Data are represented as mean ± SEM. Mean values were compared by 1-way ANOVA. (C) Bar graphs show relative mRNA expression for indicated genes measured by qPCR in colon epithelial cells from antibiotic-treated WT mice that were cohoused with either WT (n = 5) or Ern2–/– (n = 4) donor mice. Expression levels are shown relative to control WT mice (no antibiotics, no cohousing). (D) Time courses show (left panel) change in body weight and (right panel) survival during administration of DSS for 8 days followed by 14 days of recovery for WT (n = 6), Ern2–/– (n = 7), and antibiotic-treated WT mice cohoused with WT (n = 7) or Ern2–/– donors (n = 6). Survival curves were compared using log rank test. (E) Box plots show α diversity indices for microbiota from WT and Ern2–/– mice. Symbols represent values for individual mice, and error bars represent minimum and maximum. (F) Box plots show relative abundance data for indicated taxa that are significantly different between microbiota from CONV-WT and CONV-Ern2–/– mice. Symbols represent relative abundance for an individual mouse. FDR q values were calculated with MaAsLin2 (63) (WT, n = 14; Ern2–/–, n = 9). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
Figure 7
Figure 7. The SCFA butyrate links gut microbes with Ern2 expression.
(A) Box plots show whole genome sequence reads for enzymes in butyrate metabolism measured in stool from WT (n = 14) and Ern2–/– (n = 9) mice. Symbols represent an individual mouse, and bars represent the range. Mean values were compared using multiple t tests. (B) Bar graphs show mole fraction of SCFAs measured in stool from WT (n = 19) and Ern2–/– (n = 19) mice. Values are shown as the mole fraction of each relative to the total SCFAs measured in each sample. Symbols represent an individual mouse. Data are represented as mean ± SEM. (C) Left panel: bar graph shows the number of AB+ cells in the upper half of crypts in the distal colon of GF-WT and GF-Ern2–/– mice that received regular drinking water or drinking water supplemented with sodium butyrate. Symbols represent the average value for an individual mouse. Data are represented as mean ± SEM. Mean values were compared by 2-way ANOVA. Right panel: bar graph shows relative expression of Ern2 mRNA measured by qPCR in colon crypt epithelial cells from GF-WT mice (control, n = 1; butyrate n = 3). *P < 0.05; **P < 0.01.
Figure 8
Figure 8. ERN2 expression is associated with UC and goblet cell development in primary human colonoids.
(A and B) Left panels: scatter plots show relative mRNA expression for ERN2 in rectal biopsies from individuals with and without UC. Right panels: scatter plots show correlation between mRNA expression levels of ERN2 and goblet cell transcription factors ATOH1 or SPDEF in individuals with UC. (C) Representative images of undifferentiated and differentiated human colonoids. Fixed colonoids were stained with DAPI (blue) and phalloidin (magenta). Differentiated and undifferentiated colonoids were quantitated in cultures treated with DAPT and the IRE1 inhibitor 4μ8C. Symbols represent individual biological replicates from 3 independent experiments. Data are represented as mean ± SEM. Mean values were compared by 2-way ANOVA. Original magnification, ×10. (D) Bar graphs show relative mRNA expression for ATOH1, SPDEF, ERN2, and spliced XBP1 in primary human colonoids treated with DAPT and 4μ8C. Symbols represent individual biological replicates from 3 independent experiments. Data are represented as mean ± SEM. Mean values were compared by 2-way ANOVA. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.
Figure 9
Figure 9. Model for function of ERN2 in microbe-epithelial-mucus feedback loop regulating mucosal homeostasis.

Comment in

  • Goblet cells need some stress.
    Johansson ME, Hansson GC. Johansson ME, et al. J Clin Invest. 2022 Sep 1;132(17):e162030. doi: 10.1172/JCI162030. J Clin Invest. 2022. PMID: 36047499 Free PMC article.

References

    1. Bertolotti A, et al. Increased sensitivity to dextran sodium sulfate colitis in IRE1beta-deficient mice. J Clin Invest. 2001;107(5):585–593. doi: 10.1172/JCI11476. - DOI - PMC - PubMed
    1. Haber AL, et al. A single-cell survey of the small intestinal epithelium. Nature. 2017;551(7680):333–339. doi: 10.1038/nature24489. - DOI - PMC - PubMed
    1. Nyström EEL, et al. An intercrypt subpopulation of goblet cells is essential for colonic mucus barrier function. Science. 2021;372(6539):eabb1590. doi: 10.1126/science.abb1590. - DOI - PMC - PubMed
    1. Vieira Braga FA, et al. A cellular census of human lungs identifies novel cell states in health and in asthma. Nat Med. 2019;25(7):1153–1163. doi: 10.1038/s41591-019-0468-5. - DOI - PubMed
    1. Tschurtschenthaler M, et al. Defective ATG16L1-mediated removal of IRE1α drives Crohn’s disease-like ileitis. J Exp Med. 2017;214(2):401–422. doi: 10.1084/jem.20160791. - DOI - PMC - PubMed

Publication types