Inducible, but not constitutive, pancreatic REG/Reg isoforms are regulated by intestinal microbiota and pancreatic diseases

Abstract

The REG/Reg gene locus encodes a conserved family of potent antimicrobial but also pancreatitis-associated proteins. Here we investigated whether REG/Reg family members differ in their baseline expression levels and abilities to be regulated in the pancreas and gut upon perturbations. We found, in humans and mice, the pancreas and gut differed in REG/Reg isoform levels and preferences, with the duodenum most resembling the pancreas. Pancreatic acinar cells and intestinal enterocytes were the dominant REG producers. Intestinal symbiotic microbes regulated the expression of the same, select Reg members in gut and pancreas. These Reg members had the most STAT3-binding sites close to the transcription start sites and were partially IL-22 dependent. We thus categorized them as "inducible" and others as "constitutive". Indeed, in pancreatic ductal adenocarcinoma and pancreatitis models, only inducible Reg members were upregulated in the pancreas. While intestinal Reg expression remained unchanged upon pancreatic perturbation, pancreatitis altered the microbial composition of the duodenum and feces shortly after disease onset. Our study reveals differential usage and regulation of REG/Reg isoforms as a mechanism for tissue-specific innate immunity, highlights the intimate connection of pancreas and duodenum, and implies a gut-to-pancreas communication axis resulting in a coordinated Reg response.

Keywords: Antimicrobial peptides; Gene family; Gut-to-pancreas communication; Innate immunity; Pancreatic diseases; REG proteins.

Fig. 1.. Mouse and human Reg/REG gene family members are structurally highly conserved but differ in their expression patterns between pancreas and gut.

A. Sequence similarity (upper right corner) and structural similarity (lower left corner) of all mouse and human REG proteins excluding pro-segment regions. B-E. Expression levels of indicated Reg genes in the pancreas (pan), duodenum (duo), jejunum (jej), ileum (ile) and colon (col) of C57Bl6 mice by Q-PCR (n = 9 except colon, n = 6). F-H. Expression levels of indicated REG genes in human pancreas (pan, n = 31), duodenum (duo, n = 15), distal small intestine (distal SI, n = 14), and colon (col, n = 11) by Q-PCR. I-K. Relative contribution of Reg/REG isoforms to total Reg/REG pool in C57Bl6 mice (I), NOD mice (J) and humans (K) based on data in Fig. 1B-H and S2A-J, M and N. The numbers on right indicate total expression of Reg/REG genes. * p < 0.05, ** p < 0.01, *** p < 0.001 by ANOVA.

Fig. 2.. Gut and pancreas are enriched for Reg transcripts and the upstream signaling receptors.

A, B. Heatmap of Reg gene (A) and IL-6 or IL-22 receptor signaling gene (B) expression in indicated tissues from 8-week-old C57Bl6 mice measured by Q-PCR (n = 2–5, see as indicated by box graph). Maximal expression level per row is indicated at the right of each row. C. Expression of Il22ra1 and Il22ra2 in pancreas (pan), duodenum (duo), jejunum (jej), ileum (ile) and colon (col) of C57Bl6 mice by Q-PCR (n = 4). D-G. Expression levels of indicated IL6 or IL22 receptor signaling genes in human pancreas (pan, n = 31), duodenum (duo, n = 15), distal small intestine (distal SI, n = 14), and colon (col, n = 11) by Q-PCR. H. Annotated scRNAseq feature plot of mouse pancreas. I. Feature plots of mouse Reg2, Reg3b, Reg3d and Amy2b. J. Violin plots of IL-6 or IL-22 receptor subunit genes across mouse pancreatic cell types. K. Annotated scRNAseq feature plot of human pancreas. L. Feature plots of human REG1A, REG1B, REG3A and AMY2B. M. Violin plot of IL-6 or IL-22 receptor subunit genes across human pancreatic cell types. * p < 0.05, ** p < 0.01, *** p < 0.001 by ANOVA.

Fig. 3.. Pancreatic and intestinal Reg gene family transcripts correlate and are regulated by microbial colonization.

A-B. Expression levels of indicated Reg genes in pancreas (A) and duodenum (B) of Il22^+/+, Il22^+/− and Il22^−/− mice (n = 9, 9 and 10, respectively) by Q-PCR. Data pooled from two independent experiments. C-F. Pancreatic and ileal expressions of indicated Reg genes in broad-spectrum antibiotics (Abx)- treated mice or its vehicle control (C, D)(pancreas n = 7, vehicle ileum n = 9, Abx ileum n = 10), or in B6 mice from JAX, Taconic (Tac), or JAX-mice co-housed with Taconic mice (E, F) (n = 4) for 4 weeks. G-J. Expression levels of indicated Reg genes in pancreas and ileum of Il22^−/− mice treated with antibiotics or vehicle (G, H) (n = 3), or Il22^+/+ (n = 4) versus Il22^−/− (n = 3) littermate mice co-housed with Taconic mice and non-co-housed Il22^+/+ littermate controls (n = 4) (I, J). Data on graphs C-J represent samples run on the same Q-PCR plate to permit direct comparison of expression levels between the two experiments. * p < 0.05, ** p < 0.01, *** p < 0.001 by 2-tailed t-test.

Fig. 4.. Differential IL-22 producers in the pancreas and gut upon antibiotics and SFB-colonization suggests context-dependent production.

A-F. Frequencies of T_H17 cells, IL-22⁺RORγt⁺ CD4⁺ T cells, ILC3s, IL-22⁺ ILC3s, γδ T cells, IL-22⁺ _γδ T cells in pancreas (PAN) or small intestine lamina propria (SI) of B6 mice after 4 weeks of broad spectrum antibiotics (Abx) or vehicle treatment. (n = 7, pooled from two cohorts). G-L. Frequencies of T_H17 cells, IL-22⁺RORγt⁺ CD4⁺ T cells, ILC3s, IL-22⁺ ILC3s, γδ T cells, IL-22⁺ γδ T cells in pancreas (PAN) or lamina propria of the duodenum (DUO) or ileum (ILE) of B6 mice from Jax, co-housed Jax B6 mice or B6 mice from Taconic (n = 4). * p < 0.05, ** p < 0.01, *** p < 0.001 by 2-tailed t-test.

Fig. 5.. A mouse model of primary pancreatic ductal adenocarcinoma triggers Reg upregulation in pancreas but not gut.

A, B. Pancreatic expressions of Reg3g (A) and Reg3b (B) in KPC mice or their Cre^- littermates at 4, 8 or 12 weeks post tamoxifen-induced PDAC onset measured by Q-PCR (WT n = 7, 9,10, PDAC n = 9, 8, 6 for 4, 8, 12 weeks, respectively). Data pooled from 3 cohorts per time point. C, D. Intestinal expressions of Reg3g (C) and Reg3b (D) in different gut segments of KPC mice or their Cre^- littermates at 12 weeks post tamoxifen-induced PDAC onset measured by Q-PCR (WT n = 6 (Jej, Ile, Col) or 11 (Duo), PDAC n = 4 (Jej, Ile, Col) or 8 (Duo)). Data pooled from 2 to 3 independent cohorts. E. Pancreatic sections from KPC mice at 12 weeks post tamoxifen treatment or Cre^- littermates stained for ductal marker CK19, acinar marker CPA1, REG3β and DAPI (nuclei). Bar = 200 µ m, inset bar = 40 µ m. F. Bar graph of Log2 fold change of indicated genes in acinar versus metaplastic pancreatic cells in mouse scRNAseq. G. Bar graph of Log2 fold change of indicated genes in PDAC cell lines versus isolated healthy human acinar cells analyzed by Q-PCR. H. Violin plots of REG1A, REG1B, REG3A, AMY2B, KRT19, IL6R, IL6ST, IL22RA1 and IL10RB expressions in human acinar versus ductal cells in PDAC or adjacent tissues based on scRNAseq. I. Violin plots of Il6 and Il22 in all cell types recovered from human PDAC or healthy adjacent tissues analyzed by scRNAseq. * p < 0.05, ** p < 0.01, *** p < 0.001 2-tailed t-test.

Fig. 6.. Cerulein induced pancreatitis triggers upregulation of Reg genes in pancreas and modifies the gut microbiome.

A, B. Pancreatic expression of Reg3g (A) and Reg3b (B) in mice treated with Cerulein for 1, 4 or 8 weeks and age-matched vehicle control mice housed in parallel (n = 4 (1 week) or 8 (4 and 8 week treatment)). Data representative of one experiment per time point. C. Violin plots of Reg3b, Reg3g, Reg3d, Reg2, Amy2b and Krt19 expressions in mouse acinar versus ductal cells from vehicle or Cerulein treated mice based on scRNAseq. D, E. Intestinal expressions of Reg3g and Reg3b in all gut segments from mice treated with Cerulein or vehicle for one week (D) or 4 and 8 weeks (E) measured by Q-PCR (n = 6 (1 week) or 8(4 and 8 weeks)). Data representative of one experiment per time point. F-I. PCA of duodenal luminal bacteria (F) and mucosa-associated bacteria (G) or beta-diversity of duodenal luminal bacteria (H) and mucosa-associated bacteria (I) of mice with indicated treatments (n = 7–8). J-M. Volcano plots of bacteria enriched or de-enriched in the mucosa of mice treated with Cerulein for 4 (J) or 8 (K) weeks versus vehicle control mice, or in the lumen of mice treated with Cerulein for 4 (L) or 8 (M) weeks versus control mice (n = 8). * p < 0.05, ** p < 0.01, *** p < 0.001 by 2-tailed t-test.