TMED10-mediated unconventional secretion of IL-33 regulates intestinal epithelium differentiation and homeostasis

Dear Editor,

Unconventional protein secretion (UcPS) refers to the release of non-signal peptide-containing secretory proteins independent of the ER–Golgi trafficking.¹ Although two major pathways including vesicle-dependent and -independent UcPS have been revealed,¹ the physiological role of each pathway is largely unclear due to the lack of molecular details. Previously, we identified a translocation pathway termed transmembrane p24 trafficking protein 10 (TMED10)-channeled UcPS (THU), which regulates vesicle-dependent UcPS pathway.^2,3 In THU, TMED10 is a protein translocator key for mediating secretory cargo translocation into the ER–Golgi intermediate compartment to initiate vesicle-dependent UcPS. The identification of TMED10 as a central player for UcPS allows us to employ genetic approaches to understand physiological function of UcPS.

The intestinal lumen is layered with differentiated epithelial cells essential for nutrient absorption and homeostasis. Differentiation of epithelial cells into the enterocyte and secretory lineages requires the action of sophisticated secretory factors and related signaling pathways.^4,5 We hypothesize that the UcPS pathway THU may play a role in intestinal cell differentiation and homeostasis. To explore the role of THU in the function of intestinal epithelium, we generated tissue-specific knockout (KO) mice by crossing Tmed10^flox/flox with Villin-cre mice (Supplementary information, Fig. S1a), through which the protein translocator of THU TMED10 was depleted in the intestinal epithelium (T10^fl/fl;Vil, T10-KO). Secretory lineage (goblet (MUC2⁺), Paneth (LYZ⁺), and endocrine cells (CHGA⁺)) differentiation was reduced in T10-KO mice (Fig. 1a, b and Supplementary information, Fig. S1b). Consistently, the mucus layer thickness was decreased in T10-KO mice (likely due to the decrease of goblet cells, Fig. 1c). Accordingly, expression of several signal or transcription factors promoting (Egf, Egfr, and Atoh1) or inhibiting (Notch1 and Dll4) secretory lineage differentiation was decreased and increased respectively (Supplementary information, Fig. S1c). There is no obvious defect in intestinal morphology or enterocytes (Supplementary information, Fig. S1d). Acute administration of dextran sodium sulfate (DSS) has long been established as an effective model of epithelial damage that results in a highly reproducible acute colitis with weight loss, bloody diarrhea, and mucosal ulceration.⁶ In a DSS-induced colitis model, the T10-KO mice developed more severe colitis and compromised colon recovery compared to the control (T10^fl/fl) mice, including decreased body weight and colon length, while increased disease score, colon inflammation, and death rate (Fig. 1d and Supplementary information, Fig. S2). Therefore, these data suggest that TMED10 regulates the differentiation of secretory lineage in the intestine and protects against colitis.

Fig. 1. TMED10 regulates the differentiation of secretory cell lineage and protects against colitis by regulating IL-33 protein secretion.

a Hematoxylin & eosin staining, immunofluorescence staining and quantification of MUC2⁺ cells in small intestine (SI) and large intestine (LI) from control (Tmed10^fl/fl) and T10-KO (Tmed10^fl/fl; Vil) mice (n = 3 mice/group). b Immunofluorescence staining and quantification of CHGA⁺, LYZ⁺ cells in crypt-villus axis of SI and CHGA⁺ cells in LI under 40× objective area from control and T10-KO mice (n = 3 mice/group). c Alcian blue staining and quantification of mucus layer thickness of LI from control and T10-KO mice (n = 3 mice/group). d DSS-induced colitis model and index analyses of control and T10-KO mice including body weight change, Disease Activity Index (DAI) score, colon length and survival rate (n = 5 mice/group). e Analysis of scRNA-seq data (GSE186917) showing expression of Il-33 in mice epithelial cells of SI. f Diagram showing isolation of tissue pieces and collection of secretion supernatant (left). Secretion of IL-33 in the SI and LI from control and T10-KO mice determined by ELISA (middle). Secretion of mIL33-FLAG in HEK293T-TMED10-KO re-expressing the indicated TMED10 variants determined by immunoblot (right). The TMED10 variants were amplified from TMED10-V5 plasmid (DNASU). The full-length TMED10 and its truncation TMED10△CT (207–219 residue deletion) fused with a V5 tag at the C terminus were constructed into the FUGW vector and the truncation TMED10△GD (32–132 residue deletion) was inserted into the pLX304 vector. g Immunofluorescence staining and quantification of MUC2⁺ cells in SI and LI, LYZ⁺ cells in SI from control and T10-KO mice, after IP with mutIL-33 protein (10 μg/mouse/day) or IL-33 protein (10 μg/mouse/day) for 10 days (n = 3 mice/group). h Body weight change, DAI score and colon length of control and T10-KO mice at the indicated days, after IP with 10 μg/mouse/day mutIL-33 protein or IL-33 protein for consecutive 10 days, and administrated with water or 3% DSS for consecutive 8 days (n = 5 mice/group). i Immunofluorescence staining and quantification of MUC2⁺, LYZ⁺ cells in SI and MUC2⁺ cells in LI under 40× objective area from control (Il33^fl/fl) and Il33-KO (Il33^fl/fl;Vil) mice, after IP with mutIL-33 protein or IL-33 protein (n = 5 mice/group). j Body weight change, DAI score and colon length of control and Il33-KO mice at the indicated times, after IP with 10 μg/mouse/day mutIL-33 protein or IL-33 protein for 10 consecutive days, then administrated with 3% DSS for consecutive 6 days (n = 5 mice/group). k Immunofluorescence staining and quantification of MUC2⁺, LYZ⁺ cells in villus of SI and MUC2⁺ cells in LI under 40× objective area from control (T10^fl/fl;Il33^fl/fl) and T10-Il33-DKO (T10^fl/fl;Il33^fl/fl;Vil) mice, after IP with mutIL-33 protein or IL-33 protein (n = 5 mice/group). l Body weight change, DAI score and colon length of control and T10-Il33-DKO mice at the indicated times, after IP with 10 μg/mouse/day mutIL-33 protein or IL-33 protein for consecutive 10 days, then administrated with 3% DSS for consecutive 6 days (n = 5 mice/group). m A model for TMED10-mediated secretion of IL-33 regulates intestinal epithelium differentiation and homeostasis. The secreted IL-33 may act on ILC2 cells which further release IL-13 to regulate the differentiation of goblet cells and possibly other secretory cells.¹⁰ Data represent means ± SD. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05, unpaired two-tailed Student’s t-test (a, c), one-way ANOVA (d, g, h, i, j, k, l), two-way ANOVA (b, d, f, g, h, i, j, k, l), Log-rank test (d). Scale bars, 25 μm.

We next sought to pinpoint the UcPS factors accounting for the phenotype of THU deficiency. IL-33 was recently identified as a cytokine positively regulating intestinal secretory lineage differentiation.⁷ The release of IL-33 was suggested to be dependent on THU in our previous work.² We next focused on IL-33 as a potential UcPS cargo accounting for THU deficiency because: (1) Il-33 is expressed in different small intestinal epithelial cells as shown by single-cell RNA sequencing (scRNA-seq) data analysis (notably enterocytes are major sources of Il-33; however, some other secretory cells in the epithelium also express Il-33 and therefore we cannot exclude the contribution of IL-33 secretion from epithelial cells other than enterocytes) (Fig. 1e). IL-33 expression was also detected in human colon epithelial cells by scRNA-seq.⁸ (2) IL-33 release was decreased in T10-KO intestinal tissue culture in vitro (Fig. 1f, left). (3) TMED10 regulates the release of mature IL-33, dependent on TMED10 functional domains (GOLD and CT)² in cultured cells (Fig. 1f, right). (4) IL-33 downstream factors Il-13 and Areg in ILC2 cells^9,10 were decreased suggesting a deficiency of IL-33 release in the intestine (Supplementary information, Fig. S3a). Indeed, intraperitoneal injection (IP) of IL-33 protein fully or partially restored the amount of goblet and Paneth cells, and mucus layer thickness in T10-KO mice (Fig. 1g and Supplementary information, Fig. S3b, c). Consequently, IL-33 partially alleviated the severe colitis of T10-KO mice induced by DSS (Fig. 1h and Supplementary information, Fig. S3d). As a control, a cleavage form of IL-33 (mutIL-33) unable to bind to the receptor failed to rescue the phenotype. Therefore, these data collectively indicate that IL-33 secretion is regulated by THU and the lack of IL-33 release in T10-KO mice largely accounts for the decrease of goblet cells and is partially responsible for the reduced number of Paneth cells and severe colitis induced by DSS.

It is noteworthy that two additional cytokines from the IL-1 family (IL-1β and IL-18, the secretion of which is influenced by THU in specific circumstances²) affect secretory lineage differentiation and DSS-induced colitis.^11,12 However, it’s essential to note that the secretion of these cytokines is associated with a contrary effect — it reduces secretory lineage differentiation and simultaneously exacerbates the severity of DSS-induced colitis. This contrasting behavior suggests that IL-1β and IL-18 are unlikely to be the primary drivers behind the observed phenotype in T10-KO.

Previous work found that pericryptal fibroblasts are a major source of IL-33 production.⁷ Since Villin is not expressed in pericryptal fibroblasts, our data indicate that epithelium-derived IL-33 may also play a similar role in regulating secretory lineage differentiation. To confirm this notion, we generated epithelium-specific Il-33 KO mice (Il33^fl/fl;Vil, Il33-KO). Like T10-KO, the secretory lineage (goblet and Paneth cells) as well as mucus thickness was decreased in Il33-KO, which was fully restored by IL-33 injection but not the mutant (Fig. 1i and Supplementary information, Fig. S4a, b). Again, in the DSS-induced colitis model, Il33-KO developed more severe colitis (measured by body weight, disease score, colon length, and inflammation) and IL-33, instead of the mutant, completely reversed the detrimental effect (Fig. 1j and Supplementary information, Fig. S4c). The data confirmed the role of epithelial IL-33 in regulating secretory lineage differentiation and conferring protection against colitis.

The above data favor the possibility that THU regulates IL-33 release in intestinal epithelial cells which determines secretory lineage differentiation and maintains intestinal homeostasis. However, TMED10 and IL-33 may also act in parallel to independently contribute to secretory lineage differentiation and protection against colitis. To further confirm that TMED10 and IL-33 act in the same regulatory axis, we generated Tmed10 and Il-33 double KO mice (T10^fl/fl;Il33^fl/fl;Vil, T10-Il33-DKO). The T10-Il33-DKO mice showed a similar phenotype of decreased secretory lineage differentiation and colitis induced by DSS (Supplementary information, Fig. S5a–g). Importantly, the T10-Il33-DKO mice did not show obviously increased extent of secretory lineage deficiency or colitis induced by DSS compared to that of the T10-KO mice (comparing Fig. 1a–d with Supplementary information, Fig. S5a–g), which suggests that TMED10 and IL-33 may act in the same pathway. To further validate this possibility, we performed IL-33 replenishment experiments using DKO mice. Again, IL-33 largely restored the number of goblet cells and Paneth cells (Fig. 1k), and ameliorated colitis induced by DSS (Fig. 1l and Supplementary information, Fig. S6a). The extent of secretory lineage restoration and colitis protection is similar to those observed in T10-KO mice. Together these data support the notion that in the case of secretory lineage differentiation, TMED10 and IL33 largely function in the same regulatory axis instead of acting in parallel.

In summary, our study for the first time reveals a physiological role of THU by showing that TMED10-regulated IL-33 secretion (together with other sources of IL-33) in the intestinal epithelium governs secretory lineage differentiation, which is essential for protection against colitis (Fig. 1m). Therefore, it deepens our insights into multiple layers of conventional and unconventional secretory regulation in controlling intestinal epithelium differentiation and homeostasis. Conflicting effects of IL-33 on secretory lineage differentiation and DSS-induced colitis have been reported, which are likely caused by the differential sources of IL-33 and different experimental conditions (see discussion in Supplementary information, Data S1). Nonetheless, our study supports a positive effect of IL-33 (at least IL-33 derived from the epithelium) on secretory lineage differentiation and protective role against DSS-induced colitis. Although the pathway of THU-controlled IL-33 secretion was revealed in this work, we also noticed that IL-33 alone cannot completely restore the effect of Paneth cell deficiency or fully reverse the severe colitis phenotype in T10-KO mice, indicating that IL-33 secretion deficiency only partially accounts for the effect of T10-KO. One possibility is that other secretory factors in addition to IL-33 are also regulated by THU and together they contribute to Paneth cell differentiation and intestinal homeostasis. Another possibility may be that extra roles of TMED10 besides unconventional secretion (e.g., regulating cargo trafficking or Golgi morphology) also contribute to the phenotypes we observed.¹³ In addition to its impact on secretory lineages and colitis, it’s worth noting that IL-33 plays a crucial role in regulating the immune response, which is essential for maintaining the integrity of the barrier in the intestine. This relationship between IL-33-regulated immune responses and barrier integrity has a significant influence on the severity of colitis.^14,15 Therefore, it becomes imperative for future research to delve into the potential effects of THU on immune modulation within the intestinal context.

Author contributions

Y.W., M.H., Y.L., M.Z., L.G. and Y.-G.C. conceived the experiments and wrote the manuscript. Y.W., M.H., X.M., W.S. and Q.G. performed experiments and collected and analyzed data.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41422-023-00891-3.