IL-22 Preserves Gut Epithelial Integrity and Promotes Disease Remission during Chronic Salmonella Infection

The cytokine IL-22 is a member of the IL-10 family and is produced primarily by RORγt⁺ lymphocytes (1). It is recognized by the heterodimeric receptor complex consisting of IL-10R2 and the epithelial cell–restricted IL-22RA1 subunit; consequently, IL-22 acts exclusively on epithelial cells of the pancreas, liver, and kidney as well as those lining the skin, lung, and gut (1). IL-22 is critical for the clearance of many pathogens through the induction of antimicrobial peptides, including regenerating islet-derived protein (Reg) family members, β-defensins, and the S100 proteins (2, 3). Similarly, IL-22–dependent antimicrobial peptides are involved in the sequestration of commensal bacteria in the gut and lymph nodes to establish host-bacterial homeostasis; disruption in these processes results in systemic dissemination of bacteria (4, 5). In the intestine, IL-22 signaling can shape the composition of the microbiome to reduce susceptibility to experimental colitis or promote resistance and tolerance to enteric infections (6–8). Moreover, IL-22 is required for the preservation of the intestinal epithelial barrier and efficient wound repair processes via the maintenance and proliferation of epithelial stem cells (9–11). However, in contrast to its varied protective functions at barrier surfaces, IL-22 may also promote inflammation-driven pathologies by inducing chemokine production and neutrophil recruitment (12). Furthermore, IL-22 may increase disease severity when epithelial cell hyperplasia is the underlying feature of pathology as is the case for colitis-associated cancer or psoriasis (13, 14).

In patients with active inflammatory bowel disease (IBD), IL-22 expression by T lymphocytes is greatly enhanced in the inflamed intestine and can directly induce the production of proinflammatory cytokines and matrix metalloproteinases by subepithelial fibroblasts (15). These observations led to speculation that IL-22 may contribute to the pathological remodeling observed in IBD, including intestinal fibrosis (15). In addition, serum levels of IL-22 in combination with MMP9 and fecal calprotectin levels strongly correlate with the severity of disease activity in Crohn disease (CD) (16). Intestinal innate lymphoid cells (ILCs) from patients with IBD were also found to exhibit enhanced IL-22 production (17). In contrast, IL-22 has been linked to tissue-protective responses because IL-22⁺ ILCs are selectively depleted in CD patients, and this was associated with a reciprocal increase in potentially pathogenic IFN-γ⁺ ILCs (18). The contrasting beneficial versus detrimental effects of IL-22 may be attributed to heterogeneity in patient cohorts and the complexity of IL-22–mediated processes in IBD.

Salmonella enterica serotype Typhimurium (S. Typhimurium) can efficiently colonize the gut of mice following pretreatment with the antibiotic streptomycin and elicit a severe form of gastroenteritis (19). Previous studies showed that although IL-22 expression is highly induced during acute S. Typhimurium infection, it had no impact on cecal pathology while promoting pathogen persistence by suppressing the growth of competing commensals (20, 21). In contrast to the acute infection model, chronic gut infection by the attenuated S. Typhimurium ΔaroA strain results in severe transmural inflammation and fibrosis (22). Although we previously found that collagen deposition in this model is dependent on ILC3s and IL-17A, the role of IL-22 in CD-like fibrotic tissue remodeling is unclear (23). In this paper, we show that Ab-mediated IL-22 suppression following S. Typhimurium ΔaroA infection impedes intestinal epithelial repair leading to exaggerated inflammation and delayed resolution of pathology. Although IL-22 does not affect collagen deposition in the cecum during the peak of disease, we found that in disease remission, IL-22 blockade results in a striking constriction of the cecal lumens, closely resembling the stricturing pathology seen in CD patients with fibrotic disease. These findings highlight a host-protective role for IL-22 during chronic intestinal inflammation and sugges that enhancing IL-22–dependent responses may be of therapeutic benefit in severe and fibrosing CD.

C57BL/6J mice were maintained in a specific pathogen-free environment at the Biomedical Research Centre. Sex- and age-matched mice were pretreated with 20 mg of streptomycin by oral gavage 24 h prior to oral infection with 3 × 10⁶ S. Typhimurium ΔaroA CFU in 100 μl of PBS (22). Beginning at day 7 postinfection (p.i.), mice were injected i.p. with 200 μg of anti–IL-22 Abs (8E11; Genentech) or control Abs to ragweed (10D9.1E11.1F12; Genentech) three times per week (24). Animals were euthanized at days 21 or 42 p.i. for analyses. All experiments performed were approved by the University of British Columbia Animal Care Committee.

Formalin-fixed and paraffin-embedded tissues were cut into 5-μm sections for Masson trichrome or picrosirius red staining. Composite images of whole cecal cross-sections were acquired on a Nikon Brightfield microscope. Morphometric analyses were performed using ImageJ, mucosal depth was determined by measuring distance between the muscularis mucosae and surface epithelium, and the luminal area was determined by measuring area demarcated by the surface epithelial–luminal border. For immunostaining, gut sections underwent Ag retrieval in citrate buffer and were stained using Abs against β-catenin (Cell Signaling) and Ki-67 (Thermo Fisher Scientific); slides were then incubated with Alexa Fluor–conjugated Abs and mounted using ProLong Gold Antifade (Life Technologies). Optical z-stack images were captured on a Leica SP5X confocal microscope.

Total RNA extracted from the terminal end of the cecum using Trizol (Invitrogen) was reverse transcribed with a high-capacity cDNA Reverse Transcription kit (Thermo Fisher Scientific). Quantitative real-time PCR was performed on an AB7900 RT-PCR system using SYBR green chemistry (KAPA) and the following gene-specific primer pairs: Gapdh forward, 5′- ATTGTCAGCAATGCATCCTG-3′ and reverse, 5′-ATGGACTGTGGTCATGAGCC-3′; Lgr5 forward, 5′-GGACCAGATGCGATACCGC-3′ and reverse, 5′-CAGAGGCGATGTAGGAGACTG-3′; Reg3b forward, 5′-AGGACACCTCGTATCTGTGCT-3′ and reverse, 5′-CGTCATTGTTACTCCATTCCCA-3′; Reg3g forward, 5′-CCGTGCCTATGGCTCCTATTG-3′ and reverse, 5′-GCACAGACACAAGATGTCCTG-3′; Col1a2 forward, 5′-TGTTGGCCCATCTGGTAAAGA-3′ and reverse, 5′-CAGGGAATCCGATGTTGCC-3′; Col3a1 forward, 5′-TGGTTCTGGCTTCCAGACAT-3′ and reverse, 5′-GCTTTGTGCAAAGTGGAACC-3′; Tgfb1 forward, 5′-TGACGTCACTGGAGTTGTACGG-3′ and reverse, 5′-GGTTCATGTCATGGATGGTGC-3′; Igf1 forward, 5′-TTCAGTTCGTGTGTGGACCGAG-3′ and reverse 5′-TCCACAATGCCTGTCTGAGGTG-3′; Ctgf forward, 5′-CAGACTGGAGAAGCAGAGCC-3′ and reverse, 5′-GCTTGGCGATTTTAGGTGTC-3′.

Cytokine concentrations in cecal homogenates were quantified using U-PLEX multispot assays (Meso Scale Discoveries). The lower limits of detection (pg/ml) for each cytokine were as follows: IL-12p70 (0.9), IFN-γ (0.07), IL-10 (1.5), IL-6 (2.4), IL-23 (0.18), IL-1β (1.3), IL-22 (0.41), IL-17A (0.09), IL-17F (25.2), GM-CSF (0.09), IL-21 (4.1), and TNF-α (0.54). Cytokine levels were normalized to total protein determined by BCA assay (Pierce).

Cecal and splenic tissues were collected at various time points and homogenized in 1 ml of sterile PBS; serial dilutions of the homogenates were plated on LB agar plates (100 μg/ml streptomycin) for bacterial enumeration.

Fecal pellets were collected from individual mice and stored at −70°C. DNA extraction, PCR, and amplicon sequencing were performed by Microbiome Insights (Vancouver, Canada). DNA was extracted using the PowerSoil for KingFisher kit (MO BIO, Carlsbad, CA) following manufacturer’s instructions. 16S rRNA V4 gene fragments were amplified using barcoded primers as described (25) with the following primer regions (5′ to 3′): forward: 5′-GTGCCAGCMGCCGCGGTAA-3′, reverse: 5′-GGACTACHVGGGTWTCTAAT-3′. Pooled PCR amplicons were diluted to 20 ng/ml and sequenced using MiSeq 2000 bidirectional Illumina sequencing and Cluster Kit v4 (Macrogen). Library preparation was done using TruSeq DNA Sample Prep v2 Kit (Illumina) with 100 ng of DNA per sample. The library was quantified, and quality was checked using Qubit (Thermo Fisher Scientific, Waltham, MA). Sequence data were trimmed, quality filtered, and clustered at 97% identity into operational taxonomic units (OTUs) using a modified MOTHUR standard operating procedure (26). OTUs were taxonomically annotated using the SILVA database (27). Global community structure comparisons were made in an R environment using Phyloseq (28).

Significance was determined using Student t test or Mann–Whitney U test; statistical analyses were performed using Prism 5.0.

Streptomycin-pretreated C57BL/6J mice were infected with the attenuated S. Typhimurium vaccine strain ΔaroA by oral gavage. Administration of neutralizing Abs against IL-22 began 7 d p.i. when peak pathogen burdens in the intestine and spleen were already established (Fig. 1A) (22). This approach was used to delineate IL-22 function during chronic stages of the disease and excludes the effects of cytokine blockade on S. Typhimurium ΔaroA colonization of the intestine over the first week. During the peak of inflammatory responses at day 21 p.i., we found a similar increase in extracellular matrix (ECM) deposition in the cecal submucosa of anti–IL-22–treated and isotype control–treated animals as measured by Masson trichrome staining (Fig. 1B, 1C). Moreover, in both treatment groups, edema and transmural inflammatory cell infiltration were observed (Fig. 1C). Isotype-treated cecal tissue displayed an increase in mucosal depth, indicative of infection-driven intestinal epithelial hyperplasia. Interestingly, IL-22 neutralization led to severe disruption of the intestinal epithelial barrier and ablation of defined epithelial crypt structures. Many regions in the cecum were completely denuded of epithelia (Fig. 1C). By day 42 p.i., the pathology and inflammation was largely resolved in the ceca of isotype-treated animals, although some areas of ECM in the submucosa persisted. In contrast, IL-22 neutralization led to a striking loss of luminal caliber reminiscent of stricture-like pathology (Fig. 1D). Despite exaggerated CD-like features, the residual crypts and epithelial barrier remain intact. Thus, IL-22 protects the integrity of the intestinal epithelium during the course of Salmonella infection.

Morphometric analyses of mucosal depth as measured by the distance between the muscularis mucosae and surface epithelium indicated dramatic crypt hyperplasia 21 d after S. Typhimurium ΔaroA infection. This response, however, was highly attenuated in anti–IL-22–treated animals (Fig. 1E). Luminal area was also reduced during the peak of S. Typhimurium ΔaroA infection at 21 d compared with uninfected controls, but this reverted to a naive-like state by day 42 p.i. in isotype controls. In contrast, the reduction in luminal area persists in IL-22–suppressed animals up to 42 d p.i. (Fig. 1E). These results suggest that IL-22 promotes resolution of chronic S. Typhimurium ΔaroA-induced pathology and that its neutralization delays tissue recovery and leads to enhanced intestinal stricturing.

Next, we evaluated the effects of IL-22 on epithelial cell proliferation following S. Typhimurium ΔaroA infection by costaining for β-catenin and the cellular proliferation marker Ki-67. In cecal cross-sections of isotype-treated animals 21 d p.i., highly proliferative intestinal epithelial stem cells and progenitors were observed at the base of the crypts (Fig. 2A). However, anti–IL-22–treated animals displayed a discontinuous epithelial barrier and a striking reduction in defined crypt units and proliferating intestinal epithelial precursors (Fig. 2A, 2C). Despite the severe depletion of epithelial cells with anti–IL-22 treatment, crypt structures were restored by day 42 p.i. (Fig. 2B). This conclusion is further supported by a significant decrease in the expression levels of Lgr5, a marker of intestinal epithelial stem cells, in response to IL-22 blockade at day 21 p.i. but a recovery of Lgr5 expression in the cecum by day 42 (Fig. 2D). These data suggest the most critical role for IL-22 is in preserving epithelial progenitors and barrier integrity during peak disease of S. Typhimurium ΔaroA infection.

During chronic S. Typhimurium ΔaroA infection, epithelial barrier integrity is altered because of IL-22 neutralization. To evaluate the related inflammatory responses, we quantified cytokine levels of cecal tissue homogenates. At day 21 p.i., IL-22 neutralization resulted in significantly elevated levels of Th1-associated cytokines (IL-12p70 and IFN-γ), Th17 cytokines (IL-1β, IL-23, IL-22, IL-17F, IL-21) and, in addition, the immunosuppressive cytokine IL-10 (Fig. 3). Moreover, the exaggerated inflammatory responses in anti–IL-22–treated animals persisted to day 42 p.i. as significantly elevated levels of GM-CSF, IL-22, IL-21, and IL-10 were further detected (Fig. 3). Therefore, our data suggest that chronic S. Typhimurium ΔaroA infection is a potent inducer of Th1 and Th17 responses and that IL-22 helps restrain the magnitude of such inflammation in the cecum.

A major feature of chronic S. Typhimurium ΔaroA infection in mice is the development of intestinal fibrosis. To evaluate this aspect of disease, we performed picrosirius red staining of cecal sections to visualize and quantify collagen deposition. In comparison with uninfected control tissues, S. Typhimurium ΔaroA infection led to a dramatic increase in the accumulation of collagen by day 21 p.i. primarily in the submucosa as detected by picrosirius red staining in both isotype- and anti–IL-22–treated groups (Fig. 4A, 4B). In addition, mild fibrosis was detected in the subepithelial regions of the mucosa. By day 42 p.i., isotype control–treated animals underwent normal resolution and repair and much of the fibrotic scarring had resolved. In contrast, submucosal collagen staining persisted in anti–IL-22–treated animals (Fig. 4C). Despite severe epithelial defects during peak disease at day 21, quantification of collagen staining suggests that IL-22 blockade does not directly alter collagen levels in the intestine at day 21 but instead delayed the resolution of fibrosis during the latter stages of the disease (Fig. 4D). These results suggest that IL-22 may enhance reparative responses in the epithelia that resolve CD-like intestinal fibrosis.

To further characterize the magnitude of fibrotic remodeling in response to chronic S. Typhimurium ΔaroA infection and IL-22 suppression, we performed quantitative PCR analyses of cecal tissues to determine relative expression of collagen subunits Col1a2 and Col3a1 and the profibrotic factors Tgfb1, Igf1, and Ctgf. Consistent with the degree of tissue pathology observed, transcripts for these collagen components and fibrotic factors were elevated in the ceca of infected animals treated with anti–IL-22 Abs during the course of infection (Fig. 5A, 5B). Although we did not observe an increase in collagen deposition at 21 d p.i., increased relative expression of these fibrotic markers likely reflect the alterations in activation state of fibroblast cells in response to dysregulated barrier integrity and exacerbated inflammation as a consequence of IL-22 suppression.

The gut epithelia serve critical host-protective functions against enteric pathogens through the coordination of appropriate immune responses and the induction of antimicrobial peptides (29). Despite enhanced immunopathology in the ceca of anti–IL-22–treated mice at day 21 p.i., cecal S. Typhimurium ΔaroA burdens were similar to isotype control–treated animals. Interestingly, by day 42 p.i., S. Typhimurium ΔaroA CFUs in the cecal and colonic tissues of anti–IL-22–treated animals were 100-fold lower when compared with isotype controls (Fig. 6A, Supplemental Fig. 1). Moreover, Reg3b and Reg3g expression was highly attenuated in the ceca of anti–IL-22–treated animals (Fig. 6B). This suggests that as a consequence of IL-22 suppression, pathogen clearance is enhanced and is associated with increased inflammation and more pronounced tissue damage. IL-22 and antimicrobial products are also known to influence commensal microbe composition in the intestine and consequently the efficiency of pathogen clearance (30). Therefore, we hypothesized that IL-22–dependent persistence of S. Typhimurium ΔaroA burdens could be associated with an altered microbiome in the intestinal lumen. Through DeepSeq 16S sequencing of fecal pellets, we evaluated the effects of IL-22 blockade on commensal populations during S. Typhimurium ΔaroA infection. At day 7 p.i., we found that Salmonella was the dominant bacterial population in the fecal pellets of animals prior to Ab intervention. With the initiation of isotype or anti–IL-22 Ab treatments, principal component analyses indicate divergence of bacterial communities by days 21 and 42 p.i. (Fig. 7A). At day 21 p.i. (during peak inflammation and fibrosis), Salmonella represented 20% of detected OTUs in the feces (Fig. 7B). Interestingly, Bacteroides acidifaciens OTUs were highly elevated in the feces of animals treated with IL-22–neutralizing Abs (Fig. 7B, 7C). It has previously been reported that S. Typhimurium clearance in Reg3β-deficient animals is enhanced due to the overgrowth of B. acidifaciens, which is susceptible to direct Reg3β-mediated killing. Moreover, this commensal can augment pathogen clearance when fed to Salmonella-infected animals (31). Therefore the reduced intestinal S. Typhimurium ΔaroA burdens after prolonged anti–IL-22 treatment of mice may be attributed to Ab-induced changes in the composition of the gut microbiome.

CD is characterized by chronic transmural inflammation of the bowel. Although the etiology of the disease is unclear, it is thought that disruption of the epithelial barrier and unrestrained inflammation in response to luminal Ags are key contributing factors to the pathogenesis of CD (32). Animal studies of intestinal inflammation have revealed several key roles for IL-22 in maintaining tissue homeostasis through modulating efficient antimicrobial responses and by promoting tissue repair. Although these studies have provided valuable insights into the beneficial functions of IL-22 to the host through pathogen clearance, these models do not address the role of the cytokine in the development of intestinal fibrosis and stricture pathology, which affects nearly a third of CD patients and is arguably the most challenging clinical pathology (33). During fibrotic complications of CD, patients suffer from obstructive bowel symptoms as a result of the accumulation of ECM in the bowel wall and narrowing of the intestinal lumen (34). Chronic S. Typhimurium ΔaroA colonization of the intestine in our model induces robust ECM deposition in the mucosal and submucosal regions and promotes transmural inflammation associated with Th1 and Th17 cytokines, closely mimicking CD-associated fibrosis. Although these pathological features in this model are transient, our current study suggests that prolonged Ab-mediated IL-22 neutralization disrupts the epithelial barrier integrity during the peak of infection and disease. Moreover, this led to enhanced strictures in the cecum characterized by a striking loss in cecal caliber by day 42 p.i. when much of the pathology had been resolved in isotype-treated controls. IL-22 suppression also led to greatly increased mucosal levels of cytokines indicative of exaggerated Th1 and Th17 responses. This is likely due to disruption of the mucosal barrier, resulting in exaggerated inflammatory response to luminal Ags. Thus, anti–IL-22 treatment enhanced pathological hallmarks resembling the degenerative and progressive features observed in CD patients.

The IL-23–associated immune pathway, composed of RORγt⁺ CD4⁺ Th cells, ILCs, and NKTs, has been implicated in the pathogenesis of several chronic immune disorders including IBD (35). Polymorphisms in Il23r are associated with increased susceptibility to CD, whereas Th17 cells and their associated activating and effector cytokines are elevated in the intestinal mucosa of patients with active disease (36–38). Therefore, the suppression of the IL-23/Th17 axis has been investigated extensively in the pursuit of improved therapies for CD. Ab-mediated blockade of the p40 subunit of IL-23/12 has shown promise, whereas trials investigating IL-17A blockade were terminated due to lack of efficacy and increased susceptibility to opportunistic fungal infections in some patients (39, 40). Many of these confounding observations were subsequently clarified in animal studies demonstrating that IL-17A plays a critical role in preserving intestinal epithelial barrier integrity after injury. However, blocking IL-23 suppressed immunopathology in the gut without affecting innate γδ T cell sources of protective IL-17A (41, 42). Importantly, similar functions for IL-22 in epithelial barrier integrity through tight junction regulation have also been reported. However, these animal models did not directly address intestinal fibrosis. Similarly, severe fibrosis was an exclusion criterion for patients in the aforementioned anti–IL-17A trial. In our previous studies of Salmonella-induced intestinal fibrosis in mice, we found that blockade of IL-17A after established intestinal S. Typhimurium ΔaroA infection can attenuate collagen deposition in the cecum (23). In the same study, we found that hematopoietic RORα-deficient animals were protected from Salmonella-induced fibrosis due to defective ILC3 cytokine production; animals displayed reduced levels of mucosal IL-17A and IL-22. In the current study, we find that in contrast to IL-17A neutralization, IL-22 blockade does not directly alter ECM accumulation in the intestine during peak Salmonella-dependent disease. Instead, this cytokine preserves epithelial crypt structures and enhances resolution of chronic Salmonella-induced inflammation and pathology. The effects of IL-22 on the outcome of fibrosis are likely dictated by the context of tissue injury and its requirement for barrier integrity. For instance, IL-22 can directly enhance fibrosis in the liver by modulating hepatic stellate cell responsiveness to TGF-β1 (43). In contrast, IL-22 plays a protective role during kidney injury and fibrotic remodeling by improving tubular epithelial integrity and barrier function (44). Our collective work with the chronic S. Typhimurium ΔaroA infection model suggests that bolstering IL-22 responses while suppressing IL-17A may be beneficial for the treatment of fibrotic complications in CD. In IBD, similar strategies of selectively suppressing type 3 immunity without altering the tissue-protective effects of IL-22 have been proposed (45, 46).

IL-22 is critical for host immunity against infections through the induction of epithelial cell–derived antimicrobial peptides or by inducing fucosylation of surface proteins that can act directly on the invading pathogen or influence pathogen burdens through modulation of the microbiota (2, 6, 7, 30). Previously it was shown that IL-22 neutralization had minimal impact on cecal immunopathology during acute S. Typhimurium–induced colitis but promoted pathogen growth in the inflamed gut by suppressing competing commensals, likely through the induction of antimicrobial peptides (20). Similarly, the IL-22–inducible antimicrobial peptide Reg3β can promote Salmonella persistence and prolong colitis by suppression of commensal B. acidifaciens; supplementation of S. Typhimurium–infected animals with B. acidifaciens can reduce pathogen burdens in the gut through vitamin B6 metabolic responses (31). Moreover, Bacteroides species directly promote colonization resistance against S. Typhimurium infection by inhibiting pathogen growth through the production of proprionate (47). We find that prolonged IL-22 neutralization leads to a 100-fold reduction in S. Typhimurium ΔaroA CFUs in the intestine by day 42 p.i. Microbiota profiling of the feces indicates that sustained treatment with anti–IL-22 Abs results in a dramatic overgrowth of B. acidifaciens. Given that we find that IL-22 blockade represses induction of Reg3 peptides, such alterations in commensal composition may account for the enhanced clearance of S. Typhimurium ΔaroA from the gut. Although we find a discrepancy between the reduced levels of Salmonella recovered by plate CFU counts from the ceca of IL-22–suppressed animals and the higher relative abundance of Salmonella in feces detected by 16S rRNA sequencing in the same animals when compared with isotype controls, a large portion of fecal bacteria is nonviable. A limitation of traditional 16S rRNA sequencing is the inability to distinguish between live and dead bacteria (48); therefore, tissue CFUs are a more accurate reflection of pathogen burdens.

Previous reports and this current study suggest that the severity of intestinal pathology in this model of chronic infection is, to some degree, independent of Salmonella burdens (23, 49, 50). Salmonella-induced fibrosis is self-propagating after initiation of acute colitis and is unaltered by the use of antibiotics to reduce pathogen levels during the latter stages of the disease (50). It has been reported that an avirulent type III secretion system–nonexpressing population of Salmonella becomes overrepresented in the intestine during latter stages of prolonged infection, as they benefit from expression of this virulence trait by other Salmonella without themselves incurring the high fitness cost (51). Moreover, Salmonella pathogenicity island 1 and Salmonella pathogenicity island 2, which encode type III secretion systems, are required for the induction of intestinal fibrosis (22). In our current study, we find that cecal pathology is largely resolved by day 42 p.i. in isotype-treated mice despite sustained S. Typhimurium ΔaroA burdens in the tissue as compared with 21 d p.i. Furthermore, our microbiome analyses revealed that the relative abundance of Salmonella peaks 7 d p.i. but declines to 20% of detectable OTUs in the stool during peak inflammation and fibrosis at day 21. Therefore, these collective observations indicate that S. Typhimurium ΔaroA induces intestinal fibrosis by initiating the exaggerated intestinal inflammation through its disruption of the epithelial barrier disruption and that these pathological features can be further exacerbated by IL-22 suppression.

We thank Biomedical Research Centre core members Ingrid Barta, Wei Yuan, Mike Williams, and Rupi Dhesi for contributions. Neutralizing Abs against IL-22 were kindly provided by Genentech.

The authors have no financial conflicts of interest.