Myeloid ATG16L1 Facilitates Host–Bacteria Interactions in Maintaining Intestinal Homeostasis

Myeloid ATG16L1 Facilitates Host−Bacteria Interactions in Maintaining Intestinal Homeostasis

This information is current as Hong Zhang, Libo Zheng, Dermot P. B. McGovern, Ariel of October 2, 2021. M. Hamill, Ryan Ichikawa, Yoshitake Kanazawa, Justin Luu, Kotaro Kumagai, Marianne Cilluffo, Masayuki Fukata, Stephan R. Targan, David M. Underhill, Xiaolan Zhang and David Q. Shih J Immunol published online 27 January 2017 http://www.jimmunol.org/content/early/2017/01/27/jimmun Downloaded from ol.1601293

Supplementary http://www.jimmunol.org/content/suppl/2017/01/27/jimmunol.160129 http://www.jimmunol.org/ Material 3.DCSupplemental

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on October 2, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2017 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published January 27, 2017, doi:10.4049/jimmunol.1601293 The Journal of Immunology

Myeloid ATG16L1 Facilitates Host–Bacteria Interactions in Maintaining Intestinal Homeostasis

Hong Zhang,*,† Libo Zheng,† Dermot P. B. McGovern,* Ariel M. Hamill,* Ryan Ichikawa,* Yoshitake Kanazawa,* Justin Luu,* Kotaro Kumagai,* Marianne Cilluffo,‡ Masayuki Fukata,* Stephan R. Targan,* David M. Underhill,* Xiaolan Zhang,† and David Q. Shih*

Intact ATG16L1 plays an essential role in Paneth cell function and intestinal homeostasis. However, the functional consequences of ATG16L1 deficiency in myeloid cells, particularly macrophages, are not fully characterized. We generated mice with Atg16l1 deficiency in myeloid and dendritic cells and showed that mice with myeloid Atg16l1 deficiency had exacerbated colitis in two acute and one chronic model of colitis with increased proinflammatory to anti-inflammatory macrophage ratios, production of

proinflammatory cytokines, and numbers of IgA-coated intestinal microbes. Mechanistic analyses using primary murine macro- Downloaded from phages showed that Atg16l1 deficiency led to increased reactive oxygen species production, impaired mitophagy, reduced microbial killing, impaired processing of MHC class II Ags, and altered intracellular trafficking to the lysosomal compartments. Increased production of reactive oxygen species and reduced microbial killing may be general features of the myeloid compartment, as they were also observed in Atg16l1-deficient primary murine neutrophils. A missense polymorphism (Thr300Ala) in the essential autophagy gene ATG16L1 is associated with Crohn disease (CD). Previous studies showed that this polymorphism leads

to enhanced cleavage of ATG16L1 T300A protein and thus reduced autophagy. Similar ﬁndings were shown in primary human http://www.jimmunol.org/ macrophages from controls and a population of CD patients carrying the Atg16l1 T300A risk variant and who were controlled for NOD2 CD-associated variants. This study revealed that ATG16L1 deﬁciency led to alterations in macrophage function that contribute to the severity of CD. The Journal of Immunology, 2017, 198: 000–000.

nflammatory bowel diseases (IBD) are chronic debilitating The IBD-associated single nucleotide polymorphism (SNP) inflammatory conditions, which are classically differentiated ATG16L1 T300A is a loss-of-function SNP that leads to increased I into Crohn disease (CD) and ulcerative colitis. To date, .200 caspase-mediated cleavage of ATG16L1 protein (10–12). The susceptibility loci have been identified (1–3). One of the genetic functional consequences of ATG16L1 T300A were assessed in by guest on October 2, 2021 variants at position 300 in the autophagy gene ATG16L1 resulted mice with knock-in of the human ATG16L1 T300A gene (10). The in a threonine to alanine substitution (T300A) in the C-terminal mice in this study developed defective and abnormal appearing domain (2, 4). Macroautophagy (herein referred to as autophagy) Paneth cells and goblet cells, reduced bacteria clearance, in- described a bulk degradation system where cytosolic constituents creased IL-1b production, and worsened cecal inflammation with were engulfed in a double-membrane vesicle and targeted for Salmonella infection compared with wild-type (WT) mice. An- degradation by lysosomal fusion (reviewed in Ref. 5). Selective other group independently showed reduced bacteria clearance and autophagy, such as mitophagy to remove damaged mitochondria, increased expression of IL-1b in primary human ATG16L1 T300A also removed cellular danger signals and is thought to act as a macrophages and mouse ATG16L1 T316A knock-in (correspond- break for the NLRP3 inflammasome (6–9). ing to human ATG16L1 T300A) macrophages (11).

*F. Widjaja Foundation, Inflammatory Bowel and Immunobiology Research Institute, mitochondria luminol studies; J.L., K.K., and M.F. performed the neutrophil assays; Cedars-Sinai Medical Center, Los Angeles, CA 90048; †Department of Gastroenter- Y.K. assisted in animal sacrifice and flow cytometry analysis; M.C. was involved in ology, Second Hospital of Hebei Medical University, Shijiazhuang, Hebei 050000, design and analysis of TEM studies; M.F. was involved in design and analysis of China; and ‡Electron Microscopy Core Facility, Brain Research Institute, University Salmonella infection model and Salmonella uptake studies; D.Q.S. wrote the paper; of California Los Angeles, Los Angeles, CA 90095 and all authors discussed the results and commented on the manuscript. ORCIDs: 0000-0002-8400-8821 (A.M.H.); 0000-0002-2989-658X (D.M.U.); 0000- Address correspondence and reprint requests to Dr. David Q. Shih or Dr. Xiaolan 0002-7047-3634 (X.Z.); 0000-0003-1335-7044 (D.Q.S.). Zhang, Cedars-Sinai Inflammatory Bowel and Immunobiology Research Institute, 8700 Beverly Boulevard, Suite 4066, Los Angeles, CA 90048 (D.Q.S.) or Depart- Received for publication July 26, 2016. Accepted for publication December 28, 2016. ment of Gastroenterology, Second Hospital of Hebei Medical University, 215 West This work was supported by U.S. Public Health Service Grant DK056328 and Na- Heping Road, Shijiazhuang, Hebei 050000, China (X.Z.). E-mail addresses: david. tional Institutes of Health Grant P01 DK046763 (to S.R.T.), National Institutes of [email protected] (D.Q.S.) or [email protected] (X.Z.) Health K08 Career Development Award DK093578 (to D.Q.S.), Crohn’s and Colitis The online version of this article contains supplemental material. Foundation of America Career Development Award 3467 (to D.Q.S.), National Cen- ter for Advancing Translational Sciences Grant UL1TR000124 (to D.Q.S.), and by Abbreviations used in this article: BMM, bone marrow–derived macrophage; CD, funding from the F. Widjaja Foundation Inflammatory Bowel and Immunobiology Crohn disease; DC, dendritic cell; DSS, dextran sodium sulfate; IBD, inflammatory Research Institute. The MIRIAD Biobank is currently supported by the F. Widjaja bowel disease; LPMC, lamina propria mononuclear cell; MFI, mean fluorescence Foundation Inflammatory Bowel and Immunobiology Research Institute, National intensity; MHC-II, MHC class II; MLN, mesenteric lymph node; MOI, multiplicity Institutes of Health Grant P01 DK046763, European Union Grant 305479, National of infection; PBM, peripheral blood monocyte-derived macrophage; ROS, reactive Institute of Diabetes and Digestive and Kidney Diseases Grants DK062413 and U54 oxygen species; SNP, single nucleotide polymorphism; TEM, transmission electron DK102557, and by the Leona M. and Harry B. Helmsley Charitable Trust. microscopy; WT, wild-type.

H.Z., L.Z., D.P.B.M., S.R.T., D.M.U., X.Z., and D.Q.S. designed the research studies Ó and analyzed the data; H.Z. and L.Z. performed most of the experiments; A.M.H. Copyright 2017 by The American Association of Immunologists, Inc. 0022-1767/17/$30.00 performed mitochondrial studies; R.I. performed Salmonella clearance and

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1601293 2 MYELOID ATG16L1 MAINTAINS INTESTINAL HOMEOSTASIS

The cell-specific role of autophagy was reported using condi- (LysM-cre, The Jackson Laboratory stock 004781), and CD11c (CD11c- tional Atg16l1 knockout mice, in which Atg16l1 was deleted cre, The Jackson Laboratory stock 007567). For Ag presentation assay, specifically in intestinal epithelial cells or CD11c+ dendritic cells OTII/RAGII mice (Taconic model 1896) specific for OVA323–339 were used. All mice were in the C57BL/6J genetic background and were main- (DC) (13). Compared to WT, mice with epithelial Atg16l1 defi- tained under specific pathogen-free conditions in the Animal Facility at ciency exhibited Paneth cell abnormalities and were more sus- Cedars-Sinai Medical Center. This study was carried out in strict accordance ceptible to Salmonella typhimurium infection, whereas the phenotype with the Guide for the Care and Use of Laboratory Animals of the National of Atg16l1 deficiency in CD11c+ DC was similar to control WT Institutes of Health. Animal studies were approved by the Cedars-Sinai Medical Center Animal Care and Use Committee (protocol 3723). mice. The investigators found that Atg16l1 in intestinal epithelial cells was important for antibacterial defense and maintenance of gut Induction of colitis, disease activity index, and homeostasis, but that Atg16l1 may be dispensable in CD11c+ DC. histopathological analysis An independent study showed that mice with Atg16l1 deficiency in The protocol for S. typhimurium infection was performed as previously intestinal epithelial cells develop severe spontaneous ileitis in the described with the following modification (16). S. typhimurium 14028 (gift setting of Xbp1 deficiency (but not in single Atg16l1 or Xbp1 defi- from Andreas Baumler, University of California Davis) with nalidixic acid ciency), indicating that Atg16l1 has a compensatory role in inflam- resistance were grown in Luria–Bertani broth supplemented with nalidixic acid (100 mg/ml) to OD 0.2–0.8. Mice pretreated with 20 mg of strepto- mation arising from endoplasmic reticulum stress (14). Additionally, 6 mycin (Sigma-Aldrich) were oral gavaged with 3 3 10 CFU of S. conditional deletion of Atg16l1 in macrophages led to impaired typhimurium and euthanized 5 d postinfection. Acute dextran sodium clearance of Yersinia enterocolitica (11). Given the importance of sulfate (DSS) was performed as previously described except 3% DSS in macrophages and DC in innate immune responses, our goal was to drinking water was used (17). Chronic DSS colitis using four cycles of further elucidate the role of Atg16l1 in myeloid cells. In the present 2.5% DSS in drinking water was performed as described (18). The disease activity index was calculated as described (18). Tissue samples were Downloaded from study, we describe our independent generation of mice with Atg16l1 processed and stained with H&E by the Cedars-Sinai Medical Center deficiency in myeloid cells and DC, and consequential worsening of Histology Core. Histopathological scores were assigned in a blinded manner colonic inflammation in two acute models and one chronic model of by two trained animal pathologists using a previously established scoring colitis in mice with Atg16l1-deficient myeloid cells. We further re- system for S. typhimurium (16), acute DSS (17), and chronic DSS (18). port our findings on non-IBD controls and CD patients who were Human studies either homozygous for WT or the ATG16L1 T300A variant and who did not carry NOD2 CD-associated variants, to dissect the direct Genotyping was at Cedars-Sinai Medical Center using the Illumina Immuno http://www.jimmunol.org/ BeadChip array as previously described (2, 19). All human subjects were functional effect of ATG16L1 versus an indirect effect through in- homozygous non-risk for the three common CD-associated NOD2 variants flammation. Our results showed that the ATG16L1 T300A risk variant (tagged by rs2066842-GG, rs2066844-GG, rs2066845-CC, and rs5743293- resulted in impaired autophagy with reduced conversion of LC3-I to DD) and then stratified by ATG16L1 CD-associated T300A variant status, LC3-II. Loss of ATG16L1 function resulted in increased proin- dividing the study subjects into those who were either homozygous non- risk or homozygous risk (as defined by rs2241880). Written informed flammatory polarization of primary macrophages in mice and hu- consent was received from participants prior to inclusion in the study. All mans. Mice with Atg16l1 deficiency in myeloid cells and humans study subjects were identified by number and not by name and had not who were homozygous for the ATG16L1 T300A risk variant used antibiotics for at least 6 mo. Stool, blood samples, and genetic data exhibited increased colitogenic IgA-coated bacteria (15) and reduced were obtained from Material and Information Resources for Inflammatory by guest on October 2, 2021 mitophagy, which may lead to increased susceptibility to mucosal and Digestive Diseases IBD Biobank and approved under Institutional Review Board nos. 3358 and Pro00027495. inflammation. Mechanistic studies of primary macrophages showed that reduced ATG16L1 function in mice and humans resulted in Cell isolation and culture reduced clearance of S. typhimurium, Ag processing, and altered Isolation and culture of murine lamina propria mononuclear cells (LPMC) intracellular trafficking to the lysosomal compartments. Similarly, and mesenteric lymph node (MLN) cells were carried out as previously Atg16l1-deficient primary murine neutrophils also exhibited in- reported (18). MLN cells and LPMC were cultured in 96-well round- 3 6 creased reactive oxygen species production and impaired clearance bottom plates at 1.5 10 cells/ml of RPMI 1640 containing 2 mmol/l glutamine and 25 mmol/l HEPES (Corning), 10% FBS (Omega Scientific), of S. typhimurium. Taken together, these results indicate the im- 100 mM nonessential amino acids (Life Technologies), 1 mM sodium portance of autophagy in myeloid cells for the maintenance of in- pyruvate (Fisher Scientific), 10 ng/ml LPS (InvivoGen), 50 mM 2-ME testinal homeostasis. (Sigma-Aldrich), and 50 mg/ml gentamicin (Gemini Bio-Products). Media were harvested 48 h after culture and assessed for IL-1b and TNF-a production via an ELISA kit (eBioscience) per the manufacturer’s proto- Materials and Methods col. Mouse bone marrow–derived macrophages (BMM) and subsequent f/f Mice and generation of Atg16l1 mice differentiation to M1 and M2 macrophages were performed as described (20). Murine neutrophils were isolated from mouse bone marrow using a Cloning of Atg16l1 targeting vector and generation of Atg16l1f/f mice were neutrophil isolation kit (Stemcell Technologies) per the manufacturer’s performed in collaboration with genOway. Briefly, Atg16l1 endogenous protocol. Human peripheral blood monocyte-derived macrophages (PBM) locus containing 5.6 kb upstream and 2.1 kb downstream of exon 3 were were obtained by first isolating monocytes as previously described (21). generated by PCR amplification using proprietary C57BL/6J library de- Isolated monocytes then underwent M1 and M2 differentiation condition veloped at genOway. Subsequently, two loxP sites were inserted flanking as previously described (20). For human studies, paired samples (WT and Atg16l1 exon 3 (Fig. 1A). Positive selection neomycin gene flanked by ATG16L1 T300A) were obtained for each PBM isolation and differenti- FRT sites was inserted to the intron between exons 3 and 4 to generate the ation at the same day. targeting vector (Fig. 1A). Every step of the cloning process was validated through restriction enzyme analysis and sequencing. The Atg16l1 gene- Expression analysis targeting construct was linearized and electroporated into genOway proprietary embryonic stem cells with C57BL/6J background. Homologous Total RNA was isolated using an RNeasy mini kit (Qiagen) per the man- recombinants were selected by G418 and confirmed by PCR and Southern ufacturer’s protocol. All RNA samples were reverse transcribed using an blot analysis. Embryonic stem cell clones with correct 59 and 39 recom- Omniscript reverse transcription kit (Qiagen). The following validated bination were microinjected into C57BL/6J blastocysts and introduced into qPCR assays (Integrated DNA Technologies) were used to perform ex- pseudopregnant C57BL/6J mice. Male chimeric offspring were bred to pression analysis: mouse b-actin, Mm.PT.39a.22214843; mouse Atg16l1, obtain germline mutant mice that were then bred to flpe delete mouse Mm.PT.58.19068201; mouse Ptgs2, Mm.PT.58.9154407; mouse Retnla, strain to remove the neomycin cassette and were confirmed by Southern Mm.PT.58.43062398; mouse Nos2, Mm.PT.58.43705194; mouse Arg1, blot. To generate mice with conditional targeting of Atg16l1, Atg16l1f/f Mm.PT.58.8651372; human b-ACTIN, Hs.PT.56a.40703009.g; human mice were bred with mice expressing cre recombinase under the control ARG1, Hs.PT.58.40654839; human RETNLB, Hs.PT.58.1296566; human of villin (Villin-cre, The Jackson Laboratory stock 004586), LysM NOS2, Hs.PT.58.14740388; human PTGS2, Hs.PT.58.77266. The Journal of Immunology 3

FIGURE 1. Atg16l1 deficiency in murine myeloid cells exacerbated acute and chronic colitis. (A) Schematic of Atg16l1 gene targeting. (B) RT-PCR of Atg16l1 mRNA in mouse BMM and bone marrow–derived DC (BMDC). Each filled circle represents an independent experiment, and data are expressed as percentage of b-actin expression. (C) Representative immunoblot of Atg16l1, LC3, and b-actin from three independent experiments. (D) Representative H&E-stained midcolon sections from 6-mo- old mice at original magnification 3100 are shown, and data from multiple mice are quantitated (n = 7–8 per group). Scale bars, 100 mm. (E) Disease activity index of Sal- monella infection model (n = 12–15 per Downloaded from group), acute DSS (n = 10 per group), and chronic DSS (n = 10–19 per group) were quantitated and are shown. *p , 0.05, **p , 0.01, ***p , 0.001. http://www.jimmunol.org/

Western blotting S. typhimurium 14028 (gift from A. Baumler, University of California Davis) was added to all wells (multiplicity of infection [MOI] of 20) and Cells were plated on 24-well plates overnight at 250,000 cells per well and incubated for 1 h. Cells were incubated for 1- and 5-h periods with lysed using sample buffer (Novex) with protease inhibitor (Calbiochem), gentamicin to kill extracellular bacteria. Coverslips were washed, fixed boiled, and loaded onto premade SDS-polyacrylamide gels (Invitrogen), with 4% paraformaldehyde (Electron Microscopy Sciences), blocked blocked for 60 min with 1% milk (Bio-Rad Laboratories), and stained with 5% BSA (Fisher Scientific) and 0.01% Triton X-100 (Fisher Sci- overnight with the indicated primary Abs at 4˚C. Blots were washed and entific), and primary Ab was added and incubated at 4˚C overnight. by guest on October 2, 2021 stained with HRP-conjugated secondary Ab, and binding was detected by Coverslips were washed and secondary Ab was added and incubated for chemiluminescence (Thermo Scientific). Abs used included: Atg16l1, D6D5, 2 h in the dark at room temperature. Coverslips were washed and 1:1000; LC3, D11, 1:1000; b-actin, D6A8, 1:1000; anti–phospho-p40phox; mounted to slides using DAPI (SouthernBiotech). Abs used included: and HRP-linked secondary Ab (all from Cell Signaling Technology). anti–Beclin-1 1:100 (Novus Biologicals), anti-ATG16L1 1:100 (Cell Signaling Technology), anti-Lamp1 1:150 (Novus Biologicals), anti- Flow cytometry and IgA measurement Rab5 1:100 (Abcam), anti-Rab7 1:100 (Abcam), anti–S. typhimurium 0-4 Ab (1E6) 1:1000 (Abcam), goat polyclonal anti-LC3 1:100 (Biorbyt), For flow cytometry, cells were acquired on an LSR II flow cytometer (BD goat anti-mouse IgG H&L 1:200 (Abcam), donkey polyclonal secondary Biosciences, San Jose, CA) and analyzed using FlowJo analysis software. Ab to rabbit IgG H&L 1:200 (Abcam), and donkey F(ab9)2 polyclonal Live cells were selected using Live/Dead stain (Life Technologies), and secondary Ab to goat IgG H&L 1:500 (Abcam). Images were acquired on CD16/CD32 (clone 2.4G2) was used to block nonspecific FcR binding a Leica confocal microscope with an HCX PL APO CS 63.03/1.30 (eBioscience). For determination of intracellular cytokine production by GLYC (21˚C) UV objective. leukocytes, cells were incubated for 5 h at 37˚C with BD Leukocyte Ac- tivation Cocktail with BD GolgiPlug (BD Biosciences). Abs used were: mouse anti–S. typhimurium 0-4 Ab (Abcam), goat anti-mouse IgG H&L Reactive oxygen determination (Abcam), CD206 (clone C068C2; BioLegend), MHC class II (MHC-II; Reactive oxygen production was measured using luminol-ECL (Sigma- clone M5/114.15.2; BioLegend), CD86 (clone GL-1; BioLegend), NOS2 Aldrich) using macrophages and neutrophils isolated from mouse bone (clone CXNFT; eBioscience), CD11b (clone M170; BioLegend), Ly6C marrow and human macrophages differentiated from human monocytes as (clone HK1.4; BioLegend), CD45 (30-f11; clone BioLegend), CD4 (clone previously described (22). Mitochondrial reactive oxygen species (ROS) GK1.5; BioLegend), CD25 (clone PC61.5; BioLegend), CD103 (clone were determined with the addition of S. typhimurium 14028 (MOI of 100) 2E7; eBioscience), CD64 (clone X54-5/7.1; BioLegend), TNF-a and analyzed using flow cytometry after staining with MitoTracker Green (clone MP6-XT22; BioLegend), and IL-1b (clone NJTEN3; eBioscience). (Life Technologies), MitoTracker Red (Life Technologies), or MitoSOX Fecal IgA flow cytometry was performed on two fecal pellets that were (Life Technologies) for 15 min at 37˚C. Cells were washed and analyzed collected directly from four to five cohoused mice of the same genotype at using flow cytometry. 4–5 mo of age or human fecal material shipped overnight with an ice pack and then flash frozen in liquid nitrogen. The fecal samples were homog- Electron microscopy analysis enized in 1 ml of PBS per 100 mg fecal material and then centrifuged at 50 3 g for 15 min to remove large particles. Supernatants were stained Human PBM and mouse BMM were infected with S. typhimurium 14028 for IgA by flow cytometry as described (15). Total murine IgA was mea- (MOI of 20) for 1 h. Cells were fixed in 2% glutaraldehyde/2% para- sured using a mouse IgA ELISA kit the per manufacturer’s protocol formaldehyde in PBS and incubated at 4˚C overnight. Samples were (eBioscience). transferred to the University of California Los Angeles Electron Mi- croscopy Core, where they were embedded, sectioned, and placed on Immunofluorescent stain copper grids and stained with uranyl acetate and lead citrate. Images of sections were obtained using a JEOL 100CX transmission electron mi- For immunofluorescent staining, 50,000 cells were plated on 12-mm- croscope. Images were analyzed with the assistance of the Electron diameter coverslips in 24-well plates overnight. The next day, cells under Microscopy Core by counting vesicles, organelles, and Salmonella taken starvation conditions were incubated with Earle’s balanced salt solution, and up by cells. 4 MYELOID ATG16L1 MAINTAINS INTESTINAL HOMEOSTASIS

FIGURE 2. Splenomegaly and increased cellular infiltrate in mice with Atg16l1 deficiency in myeloid cells. (A) Representative H&E-stained midcolon sections at original magnification 3100 are shown, and histologic Downloaded from inflammation scores are quantitated. Scale bars, 100 mm. (B) Spleen weight in grams and splenocyte number were quantitated from chronic DSS-induced colitis in WT (Flox) mice or from mice with Atg16l1 deficiency in myeloid cells (ΔMye) and DC (ΔDC). (C) Cell http://www.jimmunol.org/ recovery from MLN and LPMC were quantitated and are shown. Each filled circle represents data from an individual mouse. *p , 0.05, **p , 0.01, ***p , 0.001. by guest on October 2, 2021

Gentamicin protection assay Ag presentation proliferation assay A gentamicin protection assay was performed by infecting human PBM, CD4+ T cells were isolated using the EasySep mouse CD4+ T cell isolation mouse BMM, and neutrophils with S. typhimurium as previously de- kit (Stemcell Technologies) from spleen from OTII/RAGII mice and scribed (23) with the following modifications: intracellular killing was stained with CellTrace CSFE (Thermo Fisher Scientific) according to the of S. typhimurium that was infected at MOI of 20. Infected cells were manufacturer’s recommendations. CFSE stained cells were cocultured (one then incubated with gentamicin at 50 mg/ml for 1 h (macrophages) and BMM to four CD4+ T cells) with either WT or Atg16l1-deficient BMM that 30 min (neutrophils) and plated for quantification. were exposed to whole OVA or OVA323–339 for 6 h. Cells from 72-h cocultures were collected and stained with CD4 GK1.5 (BioLegend) and Live/Dead stain + Neutrophil migration/chemotaxis assay (Fisher Scientific). After gating on live CD4 T cells, CFSE staining was analyzed by flow cytometry to determine the CD4+ T cell proliferation. Murine neutrophils were isolated from mouse bone marrow using a neutrophil isolation kit (Stemcell Technologies) per the manufacturer’s Statistical analysis protocol. After cell viability was checked by trypan blue, neutrophils were suspended in DMEM/1% endotoxin-free BSA/10 mM HEPES, and Data are presented as the means 6 SD. Comparison between two groups 200,000 cells were placed in the upper insert of a 12-well Transwell was performed by a two-tailed Fisher exact test for categorical variables chamber (5-mm pore size, 6.5-mm diameter; Costar Corning). The su- and a Student t test for continuous variables. Parametric and nonparametric pernatant from bone marrow–derived WT macrophages stimulated with tests were used depending on the fulfillment of the test assumptions. sonicated S. typhimurium or MIP-2 at 5 ng/ml was applied to the bottom Comparison between three groups was done using ANOVA, followed by well. After 60 min incubation at 37˚C in 5% CO2, the cells in the pairwise post hoc analysis with a Tukey honest significant difference and bottom well were harvested by adding 50 mlof70mMEDTAand Behrens–Fisher test correction for the multiple comparisons. A p value counted with a hemocytometer. ,0.05 was considered significant. The Journal of Immunology 5

FIGURE 3. Increased production of IL-1b and TNF-a with Atg16l1 deficiency in myeloid cells. MHC-II+CD11b+F4/80+ macrophages (A) and MHC-II+CD11b+CD11c+ DC (B) were stained for intracellular IL-1b and TNF-a expression and quantitated for Salmonella infection (right panel), acute DSS (middle panel), and chronic DSS (right panel). Isolated mononuclear cells from MLN and LPMC from the three murine colitis models were cultured for 3 d, and the levels of secreted IL-1b in MLN (C) and LPMC (D) as well as TNF-a in MLN (E) and LPMC (F) were assessed by ELISA. Downloaded from Each filled circle for MLN represents the value obtained from a single mouse. Each filled circle for LPMC represents the value obtained from an independent experiment using pooled samples from two mice. *p , 0.05, **p , 0.01, ***p , 0.001. http://www.jimmunol.org/

Results and acute colitic conditions, we observed significant splenomegaly by guest on October 2, 2021 ΔMye ΔDC Loss of Atg16l1 in myeloid cells exacerbated murine models of with increased splenocytes in the Atg16l1 and Atg16l1 mice colitis as compared with Atg16l1flox mice under chronic DSS-induced colitis (Fig. 2B). Except for MLN cells in the acute DSS model, we re- Because global deficiency of Atg16l1 is lethal to mice (7, 24, 25), Δ covered more MLN cells and LPMC in Atg16l1 Mye than in we generated Atg16l1 floxed mice (hereafter called Atg16l1flox)to Atg16l1flox mice (Fig. 2C). In the chronic DSS-induced colitis model, evaluate its role in gut mucosal homeostasis in colitis models Δ Atg16l1 DC mice had more MLN cells than did Atg16l1flox mice. (Fig. 1A). Using LysM-Cre and CD11c-Cre, we generated mice These results indicated that Atg16l1 deficiency in myeloid cells (and with Atg16l1 deficiency in myeloid cells (hereafter called Δ Δ to a lesser extent in DC) exacerbated colitis in murine models. Atg16l1 Mye) and in DC (hereafter called Atg16l1 DC), respec- tively. Deletion of Atg16l1 in myeloid cells and DC was confirmed by the lack of Atg16l1 mRNA and protein (Fig. 1B, 1C). Mac- Myeloid Atg16l1 deficiency resulted in an increase in rophages and DC with Atg16l1 deficiency also exhibited func- proinflammatory macrophages tional autophagy deficiency by the lack of conversion of LC3-I Flow cytometry analysis did not reveal any differences in the to LC3-II (Fig. 1C). frequencies of macrophages (MHC-II+CD11b+F4/80+), DC (MHC- Under basal conditions, mice did not spontaneously develop colitis, II+CD11b+CD11c+), or expression of the activation marker CD86 and there were no histologic differences between the colons of control on macrophages and DC in the MLN and lamina propria of the Atg16l1flox mice and mice with Atg16l1 deficiency in myeloid cells or three colitic models used (data not shown). We observed an in- in DC up to 10 mo of age (Fig. 1D). Next, Atg16l1flox, Atg16l1ΔMye, creased percentage of macrophages and DC producing IL-1b (but Δ Δ and Atg16l1ΔDC mice underwent two acute models and one chronic not TNF-a)inAtg16l1 Mye mice and in DC in Atg16l1 DC mice model of murine colitis to assess the role of Atg16l1 on mucosal compared with control Atg16l1flox mice (Fig. 3A, 3B, data not inflammation and to exclude colitis model–specific effects of Atg16l1. shown). Upon examining isolated cells from the MLN and LPMC, In all models, disease activity index scores were significantly higher in we observed increased IL-1b and TNF-a, but not IL-6, production Atg16l1ΔMye ($50% of measured time points) and in Atg16l1ΔDC in most (if not all) colitis models used (Fig. 3C–F). Δ Δ (latter part of colitic model) than in control Atg16l1flox mice (Fig. 1E). Because Atg16l1 Mye (but not Atg16l1 DC) mice exhibited increased Histologic examination of the colon revealed worsened inflammation colonic inflammation in all three models of colitis (Fig. 2A), we fo- Δ characterized by increased cellular infiltrate, mucin depletion, crypt cused on Atg16l1 Mye mice for the remainder of the murine experi- abscesses, and architectural changes in Atg16l1ΔMye as compared with ments covered in the present study. Because appropriate balance and Atg16l1flox mice in all three models (Fig. 2A). Histological analysis function of the proinflammatory P2 stage (MHC-II+Ly6C+)andanti- 2 showed that Atg16l1ΔDC mice exhibited worsened colitis only in the inflammatoryP3/4stages(MHC-II+Ly6C ) of macrophages 2 chronic DSS-induced colitis model (Fig. 2A). In contrast to the basal (CD45+CD11b+CD64+CD103 ) have been implicated in maintenance 6 MYELOID ATG16L1 MAINTAINS INTESTINAL HOMEOSTASIS of mucosal immune homeostasis (20, 26), we next sought to de- Increased IgA-coated bacteria in the stool of Atg16l1ΔMye mice termine whether Atg16l1 deficiency in myeloid cells altered these IgA coating has been shown to identify IBD-driving microbes in intestinal macrophage subsets. Prior to initiation of colitis, we mice and humans (15). To assess whether Atg16l1 deficiency in observed an increase in the percentage of P2 cells and a reduction in myeloid cells could led to increased IgA-coated bacteria that may the percentage of P3/4 cells in the LPMC of Atg16l1ΔMye mice com- flox be colitogenic, we examined IgA staining in stool collected from pared with control Atg16l1 mice (Fig. 4A, 4B). Similarly, expression noninflamed 2- to 3-mo-old Atg16l1flox and Atg16l1ΔMye mice. analysis showed a consistent increase in the proinflammatory macro- ΔMye ΔMye Flow cytometry showed that Atg16l1 mice had increased IgA phage markers Nos2 and Ptgs2 in LPMC isolated from Atg16l1 flox flox coating of bacteria compared with control Atg16l1 mice compared with control Atg16l1 mice prior to the initiation of colitis (Supplemental Fig. 1A). Induction of colitis led to a further sig- (Fig. 4C). Similar levels of the regulatory macrophage markers Retnla nificant increase in IgA-coated fecal bacteria in all models and Arg1 were measured in the isolated LPMC between Atg16l1ΔMye flox (Supplemental Fig. 1A). Total IgA in the stools was quantitated by compared with control Atg16l1 mice prior to the initiation of colitis IgA ELISA to determine whether increased IgA coating of fecal (Fig. 4C). Once colitis was established, flow cytometry analysis dem- bacteria was due to increased IgA secretion into the intestinal onstrated an increase in proinflammatory P2 cells in all colitis models, lumen. Total IgA in the stool was less in mice with Atg16l1 de- and a reduction in anti-inflammatory P3/4 macrophage populations was ficiency than in mice without Atg16l1 deficiency prior to induction found in the chronic DSS conditions in Atg16l1ΔMye as compared with flox of colitis (Supplemental Fig. 1B), suggesting that the increased control Atg16l1 mice (Fig. 4A, 4B). IgA-coated bacteria were not driven by increased luminal IgA. To test whether autophagy was involved in the development of Total stool IgA was increased in the Salmonella infection model proinflammatory and regulatory macrophages, we assessed

but not in the acute or chronic DSS model (Supplemental Fig. 1B), Downloaded from whether there are differences in the differentiation of bone marrow– which is likely due to impaired clearance of Salmonella by derived proinflammatory M1 and regulatory M2 macrophages ΔMye flox ΔMye macrophages (Fig. 5A). The worsened colitis in Atg16l1 mice in vitro. Littermate Atg16l1 and Atg16l1 BMM were cul- as compared with control Atg16l1flox mice may in part be due to tured in either M1 (LPS and IFN-g) or M2 conditions (TGF-b, increased colitogenic IgA-coated bacteria in Atg16l1ΔMye mice. IL-10). Compared to Atg16l1flox, there was increased expression of the M1 markers Nos2 and Ptgs2 in Atg16l1ΔMye BMM cultured in Impaired clearance and altered trafficking of bacteria in M1 conditions (Fig. 4D). No differences in the M2 markers Retnla Atg16l1-deficient murine macrophages http://www.jimmunol.org/ and Arg1 were observed between Atg16l1flox and Atg16l1ΔMye Owing to the importance of autophagy in bacterial clearance, we BMM cultured in M2 conditions (Fig. 4D). Taken together, the hypothesized that Atg16l1 deficiency could lead to impaired data showed that inhibition of macrophage autophagy promoted macrophage clearance of bacteria. Infection of Atg16l1-deficient M1 macrophage polarization. BMM with S. typhimurium led to reduced bacteria clearance by guest on October 2, 2021

FIGURE 4. Increased number of proinflammatory macrophages with Atg16l1 deficiency. Macrophage subsets in LPMC were assessed as proinflammatory (P2) and anti-inflammatory (P3/4) macrophages in 2-mo-old noncolitic mice and in the three colitis models. Representative flow cytometry plots of gated CD45+CD11b+CD64+CD1032 cells are shown (A) and quantitated (B). (C) RT-PCR of mRNA for proinflammatory macrophage markers (Nos2, Ptgs2) and regulatory macrophage markers (Retnla, Arg1) were measured from LPMC macrophages isolated from 2-mo-old noncolitic mice and are represented as percentage of b-actin reference gene expression. (D) RT-PCR of mRNA for Nos2 and Ptgs2 were determined in BMM cultured in M1 polarizing conditions, and mRNA for Retnla and Arg1 were determined in BMM cultured in M2 polarizing conditions. Each filled circle represents an independent experiment using pooled LPMC from two mice of the same genotype for (B) and from a single mouse in (C)and(D). *p , 0.05, **p , 0.01, ***p , 0.001. The Journal of Immunology 7

FIGURE 5. Impaired clearance and altered cellular trafficking of S. typhimurium in Atg16l1-deficient BMM. WT and Atg16l1ΔMye BMM were infected with S. typhimurium, and surviving intracellular bacteria were quantified (A), representative TEM images of WT (n = 17) and Atg16l1-deficient (n = 25) BMM infected with S. typhimurium are shown (B), and vesicles containing multiple Salmonella were quantitated (C). White arrowheads indicate vesicles containing S. typhimurium. Scale bars, 2 mm. Each filled circle represents an independent experiment (A) or data from a single macrophage (C). (D) Representative flow cytometry plot of intracellular Salmonella (left panel), quantitated as percentage (middle panel), and MFI (right panel) are shown (n =5 Downloaded from independent experiments). Representative confocal image of WT or Atg16l1-deficient BMM stained with anti-Salmonella (green stain, top panels) and anti-Lamp1, anti-Rab5, and anti-Rab7 (violet stain, middle panels) are shown (E) and quantitated (F). Colocalized images where green stain overlaps with purple http://www.jimmunol.org/ stain are marked by white arrowheads (merged white stain, bottom panels). Scale bars, 20 mm. Each filled circle represents data acquired from an individual BMM from three independent experiments. (G) Quantification of S. typhimurium in either double-membrane versus single-membrane vesicles obtained from TEM in each BMM are shown as mean 6 SD. , , , *p 0.05, **p 0.01, ***p 0.001. by guest on October 2, 2021

(Fig. 5A). When sections of S. typhimurium–infected macro- Salmonella in double-membrane vesicles but increased Salmonella in phages were visualized with transmission electron microscopy single-membrane vesicles (Fig. 5G). Taken together, the data showed (TEM), we observed more vacuoles containing multiple Salmo- that Atg16l1 deficiency in macrophages led to a reduction of bacteria nella in Atg16l1ΔMye compared with Atg16l1flox BMM (Fig. 5B, clearance, due in part to decreased intracellular trafficking of bacteria 5C). We determined the percentage of bacteria in macrophages to to Lamp1-containing vacuoles for bacterial degradation. control for potential differences in bacteria uptake between Atg16l1ΔMye BMM. Although there were no differences in the Atg16l1 deficiency resulted in enhanced reactive oxygen percentage of macrophages containing S. typhimurium (Fig. 5D, production in primary murine macrophages left and middle panels), the mean fluorescence intensity (MFI) of ROS are cytotoxic for a variety of microorganisms, including the Salmonella stain was significantly increased at 60 min after S. Salmonella (27–29). We therefore hypothesized that ROS pro- typhimurium infection (Fig. 5D, right panel), indicating that there duction might be defective in Atg16l1-deficient cells. Unexpect- were more S. typhimurium per macrophage. edly, we observed that Atg16l1-deficient macrophages had higher We hypothesize that the impaired S. typhimurium clearance that total cellular generation of ROS when infected with S. typhimu- leads to its increase in macrophages in Atg16l1-deficient macro- rium (Fig. 6A, left panel) or exposed to zymosan (Fig. 6A, right phages may be due to reduced intracellular trafficking of S. typhi- panel) than did WT macrophages. As an independent approach to murium to intracellular compartments for bacteria killing. assess NADPH oxidase activation, we measured phosphorylation Consistent with our hypothesis, colocalization of S. typhimurium of the cytosolic p40 subunit. When macrophages were infected with the lysosome marker Lamp1 was reduced (Fig. 5E, 5F). with S. typhimurium, p40phox was more phosphorylated in Colocalization of S. typhimurium with early autophagy markers Atg16l1-deficient cells than in control cells (Fig. 6B). These data (Beclin 1 and LC3) was similar between WT and Atg16l1-deficient suggested that Atg16l1 plays a role in limiting ROS production by macrophages (Fig. 5E, 5F). Interestingly, colocalization of Salmo- macrophages and that impaired bacteria killing is not due to re- nella with the endosomal markers Rab5 and Rab7 was increased duced ROS production. (Fig. 5E, 5F). Quantification of double-membrane (autophagy) and Selective autophagy such as mitophagy that removes dysfunc- single-membrane (nonautophagy) vesicles revealed reduction of tional mitochondria has been shown to limit IL-1b production and 8 MYELOID ATG16L1 MAINTAINS INTESTINAL HOMEOSTASIS

FIGURE 6. Atg16l1 deficiency increased ROS production and altered mitochondria homeostasis. (A) Relative ROS production as measured by luminol- dependent chemiluminescence was determined during 60 min for WT (Atg16l1flox) and Atg16l1-deficient (Atg16l1ΔMye) BMM treated with S. typhimurium or zymosan (filled diamond and solid line) and at baseline (filled diamond). Data of eight independent experiments with similar results are shown. (B) Phospho-p40phox and b-actin levels in WT and Atg16l1-deficient BMM at baseline and with S. typhimurium infection were determined by immunoblotting. (C) Representative TEM Downloaded from images of WT and Atg16l1-deficient BMM are shown. Scale bars, 2 mm. White arrowheads indicate mitochondria, and mitochondria per BMM are quantitated. (D) Representative flow cytometry plot of WT and Atg16l1-deficient BMM left unstained (filled) or labeled with MitoSOX (open) are shown and MFI is quantitated. (E) Representative flow cytometry plot of WT or Atg16l1- http://www.jimmunol.org/ deficient BMM with and without S. typhimurium infection and stained with MitoTracker Deep Red and MitoTracker Green are shown. Each filled circle represents an independent experiment from a single mouse. *p , 0.05. by guest on October 2, 2021

inflammation (6, 7, 9). Therefore, we hypothesized that Atg16l1- specific OT-II CD4+ T cells isolated from OTII/RAGII2/2 mice. deficient BMMs may contain more mitochondria and produce These genetically modified T cells recognize MHC-II–restricted increased ROS than do control BMM. Consistent with our hy- presentation of OVA323–339 (30). Proliferation of OT-II cells was pothesis, Atg16l1-deficient BMM had increased macrophages per reduced when stimulated with Atg16l1-deficient macrophages cell (Fig. 6C) and produced more ROS than did controls (Fig. 6D). compared with WT macrophages (Fig. 7A). To dissect the steps of We next hypothesized that Atg16l1-deficient BMM contained Atg16l1 deficiency impairment of MHC-II–restricted Ag presen- more damaged mitochondria with reduced membrane potential. tation to T cells, we pulsed WT and Atg16l1-deficient macro- Thus, we measured mitochondria membrane potential using phages with the OVA323–339 synthetic peptide that did not need MitoTracker Deep Red and counterstained with MitoTracker further intracellular processing to be loaded onto MHC-II. WT and Green, a probe that stains mitochondrial membrane lipids inde- Atg16l1-deficient macrophages induced a similar degree of pro- pendent of membrane potential. Consistent with our hypothesis, liferation to CellTrace-labeled OVA-specific OT-II CD4+ T cells we observed a reduction of membrane potential in mitochondria in (Fig. 7B). These data were consistent with the idea that the im- Atg16l1-deficient cells compared with WT (Fig. 6E). Taken to- paired macrophage presentation of Ag was due to altered pro- gether, our data suggest that the increased overall ROS through cessing of Ag and not due to deficient loading of Ag to MHC-II direct cytotoxic damage and impaired mitophagy with Atg16l1 protein. deficiency could contribute to worsened colitis in our models. Impaired S. typhimurium clearance and increased ROS Atg16l1 is required for optimal processing of MHC-II Ags by production in Atg16l1-deficient murine neutrophils murine macrophages Because utilization of the LysM-Cre system leads to deletion of the To assess whether cellular defects seen with Atg16l1 deficiency floxed allele in the myeloid compartment, deletion efficiency of (e.g., altered Ag trafficking) affected presentation of MHC-II Ags, nearly 100% in granulocytes including neutrophils was also ob- we exposed WT or Atg16l1-deficient macrophages to full-length served (31). We hypothesized that Atg16l1 deficiency in neutro- OVA protein and cocultured them with CellTrace-labeled OVA- phils will also lead to impaired S. typhimurium clearance and The Journal of Immunology 9

FIGURE 7. Atg16l1 is required for optimal Ag processing by macrophages for MHC-II Ag presentation. Representative flow cytometry plots of proliferating OT-II CD4+ T cells labeled with CellTrace stimulated with BMM treated with whole OVA protein (A)orOVA323–339 (B) are shown. Decreased CellTrace fluorescence intensity indicated proliferation. Each filled circle represents an independent experiment. *p , 0.05. Downloaded from

increased ROS production that were observed in Atg16l1-deficient allele (Supplemental Fig. 1C). However, the effect of IgA coating murine neutrophils. To assess neutrophil function, we isolated pri- was more profound in CD patients compared with control subjects http://www.jimmunol.org/ mary murine neutrophils from bone marrow of WT and Atg16l1ΔMye (Supplemental Fig. 1C). These data showed that CD-associated mice. Atg16l1-deficient neutrophils had similar chemotaxis/ ATG16L1 T300A variants are associated with reduced autophagy recruitment in response to supernatant from WT neutrophils cocul- and other proinflammatory changes (increased M1 polarization and tured with S. typhimurium (Supplemental Fig. 2A). Consistent with IgA-coated stool) that may explain the increased susceptibility to the altered Atg16l1-deficient murine macrophages, we also observed CD in individuals carrying ATG16L1 T300A risk alleles. that murine neutrophils with Atg16l1 deficiency exhibit increased ATG16L1 T300A is associated with impaired clearance and ROS production and impaired clearance of S. typhimurium (Supple- altered trafficking of bacteria within human macrophages

mental Fig. 2B, 2C). Our data indicate that increased production of by guest on October 2, 2021 ROS and reduced microbial killing may be general features of We sought to determine whether the ATG16L1 T300A impaired Atg16l1 deficiency in the myeloid compartment. bacteria clearance in human PBM similar to murine Atg16l1- deficient BMM. Infection of PBM with S. typhimurium led to Characterization of the ATG16L1 T300A polymorphism in reduced bacteria clearance in both non-IBD and CD subjects human primary macrophages homozygous for the ATG16L1 T300A CD-associated allele Several studies have indicated that the IBD-associated ATG16L1 (Fig. 8E). The reduction in bacterial clearance was not secondary T300A variant is a loss-of-function SNP that leads to increased to reduced uptake because S. typhimurium uptake did not vary caspase-mediated cleavage of ATG16L1 protein, resulting in reduced with ATG16L1 T300A status (Fig. 9A). In contrast to our findings autophagy (10, 11). Our murine data on Atg16l1 deficiency suggested in the mouse, Salmonella-infected non-IBD human PBM from that macrophages from people carrying the T300A SNP might be individuals with the ATG16L1 T300A risk alleles did not exhibit more M1-polarized, more inflammatory, and less able to kill Salmo- increased numbers of vacuoles containing multiple Salmonella nella. We therefore isolated primary PBM from subjects who were (Fig. 9B, 9C). By TEM, we observed reduced numbers of S. homozygous for either WT or ATG16L1 T300A SNP for functional typhimurium in double-membrane vacuoles in the risk ATG16L1 human studies. Because ATG16L1 has been reported to interact with T300A homozygotes (Fig. 9D), indicating that the risk variant is NOD2, we controlled for the IBD-associated NOD2 polymorphism associated with reduced autophagy. We next examined whether by using only cells that lacked the NOD2 risk allele (12, 23, 32–34). our finding in mice of defective intracellular trafficking of S. Additionally, we compared PBM from non-IBD and CD subjects to typhimurium to the LAMP1+ lysosomal compartments also oc- exclude the possibility that changes seen were due to CD. curred in humans. Consistent with the murine data, there was Macrophages from subjects homozygous for the risk ATG16L1 reduced colocalization of S. typhimurium with the lysosome T300A allele exhibited significantly reduced autophagy (reduced marker LAMP1 in ATG16L1 T300A PBM (Fig. 9E). Colocaliza- LC3-I to LC3-II conversion) upon induction with S. typhimurium tion of S. typhimurium with early autophagy markers (Beclin1 and but not under basal conditions in both non-IBD subjects (Fig. 8A) LC3) and with the endosomal markers (Rab5 and Rab7) were and CD patients (Fig. 8B). Similar to murine Atg16l1-deficint similar (Fig. 9E). Taken together, these data suggested that re- macrophages, PBM from ATG16L1 T300A risk variant homozy- duced bacterial clearance could be in part due to altered intra- gotes from both non-IBD subjects (Fig. 8C) and CD patients cellular trafficking of bacteria to lysosomes for degradation. (Fig. 8D) exhibited increased polarization to M1 proinflammatory macrophages with an increase in the M1 markers NOS2 and ROS production is enhanced in primary human macrophages PTGS2. Consistent with murine findings, individuals who were from subjects with the ATG16L1 T300 variant homozygous for the CD risk ATG16L1 T300A allele had more IgA- Because Atg16l1 deficiency increased ROS production, we hy- coated bacteria in the stool than did individuals not carrying the risk pothesized that PBM with ATG16L1 T300A SNP will also exhibit 10 MYELOID ATG16L1 MAINTAINS INTESTINAL HOMEOSTASIS

FIGURE 8. Reduced autophagy and proinflammatory macrophage polarization with the humanCDriskANPATG16L1 T300A.Rep- resentative immunoblot of LC3 and b-actin from non-IBD control (A) and CD patients (B) who were homozygous for risk or non-risk ATG16L1 T300A alleles are shown and quantitated as percentage of b-actin reference expression. Nos2 and Ptgs2 mRNA levels were Downloaded from measured by RT-PCR in PBM cultured in M1-polarizing conditions, and mRNA for Retnla and Arg1 were measured in PBM cultured in M2-polarizing conditions. mRNA expression is expressed as percentage of b-actin reference expression for non-IBD controls (C) D E and CD patients ( ). ( )Representativeflow http://www.jimmunol.org/ cytometry plots of fecal bacteria stained with anti-IgA Ab from non-IBD controls and from CD patients are shown and quantitated. Each experiment was done in pairs with an individual of each genotype for ATG16L1.(F)PBM from risk or non-risk ATG16L1 T300A variant subjects were infected with S. typhimurium, and surviving intracellular bacteria were quantified. Each filled circle is representative of an independent experiment from a human subject. by guest on October 2, 2021 *p , 0.05, **p , 0.01, ***p , 0.001.

increased ROS production. Homozygote risk PBMs with ATG16L1 panels) and for CD patients (Fig. 11B, middle and right panels). T300A SNP had increased total cellular generation of ROS when Taken together, the data showed that macrophages from subjects exposed to S. typhimurium or zymosan compared with non-risk with the ATG16L1 CD-associated risk variants had defects in macrophages (Fig. 10A). We then examined the effect of the autophagy, inflammatory signaling, ROS production, and mito- ATG16L1 risk variant on human PBM mitochondria and mitochondrial function that are similar to those observed in murine chondrial ROS production. Similar to the murine Atg16l1 defi- macrophages deficient in Atg16l1. ciency, mitochondria numbers were increased in macrophages from ATG16L1 risk homozygotes compared with non-risk ho- Discussion mozygotes (Fig. 10B). MitoSOX fluorescent signal was enhanced Previous studies showed that ATG16L1 in epithelial cells is im- in cells from the risk homozygotes for both non-IBD subjects portant for intestinal homeostasis (13, 14). In the present study, we (Fig. 11A, left panel) and CD patients (Fig. 11B, left panel), in- were interested in the role of ATG16lL1 in myeloid cells and used dicating that there was increased ROS accumulation in the mito- LysM-Cre, which can mediate deletion of loxP-flanked genes to chondria of ATG16L1 risk homozygotes. Using MitoTracker Deep nearly 100% efficiency in macrophages and granulocytes (31). We Red and MitoTracker Green, we observed a reduction of mem- showed that Atg16l1ΔMye mice exacerbate intestinal inflammation in brane potential in the mitochondria of cells from the risk homo- three different murine colitis models, indicating that functional zygotes for both non-IBD subjects (Fig. 11A, middle and right Atg16l1 in macrophages and granulocytes is important for intestinal The Journal of Immunology 11

FIGURE 9. Altered intracellular trafficking in PBM with CD risk ATG16L1 T300A variant. (A) Representative flow cytometry plots (n = 4 independent experiments from four subjects per group) of PBM from risk and non-risk ATG16L1 T300A homozygotes that were uninfected (dotted line) or infected with S. typhimurium and stained with anti- Salmonella Ab (solid line) are shown. Rep- resentative TEM image of PBM from homozygotes risk and non-risk for CD- associated ATG16L1 T300A variant infected with S. typhimurium is shown (B) and quantitated for number of intracellular vesicles that contain more than one Salmonella (C) and for the number of single- and double- membrane vesicles that contain bacteria (D). Each filled circle represents an independent experiment from at least six subjects per group. Scale bars in TEM, 2 mm. (E) Rep- resentative confocal microscopy image of Downloaded from PBM from subjects with risk or non-risk ATG16L1 T300A variants infected with S. typhimurium and stained with anti-Salmonella (green stain) and anti-Lamp1 (violet stain) at original magnification 3630 are shown. Colocalized regions where green stain overlaps with purple stain are marked by white http://www.jimmunol.org/ arrowheads (merged white stain) and represented by percentage of colocalization. Scale bars in confocal images, 20 mm. Each filled circle in (C)–(E) represents a value obtained from a PBM from at least six subjects per group. homeostasis. We next focused on characterizing the cellular and in myeloid cells did not exhibit resistance to C. rodentium, and in by guest on October 2, 2021 molecular function of Atg16l1 in mouse primary macrophages and a bone marrow chimera experiment Atg16l1HM mice exhibited correlated the murine Atg16l1-deficient defects to human primary resistance to C. rodentium infection regardless of the bone marrow macrophages carrying the risk CD ATG16L1 T300A SNP. We source (WT or Atg16l1HM) (41). observed that impaired ATG16L1 function led to increased numbers Our study revealed additional innate and adaptive immune of proinflammatory intestinal macrophages in vivo (Fig. 4A, 4B), mechanisms by which ATG16L1 affected the removal of infectious which was likely in part due to enhanced differentiation of proin- organisms in the gut. We identified an intracellular trafficking flammatory macrophages. A recent study showed that another gene defect associated with ATG16L1 deficiency leading to reduced in the autophagy pathway, ATG5, also promotes proinflammatory fusion of S. typhimurium–containing vesicles with Lamp1+ lyso- macrophage polarization in BMM and Kupffer cells, resulting in somes in mouse and human macrophages. This diminished fusion increased hepatic inflammatory response in high-fat diet/LPS– led to reduced autophagosome maturation in ATG16L1 macro- treated ATG52/2 mice (35). The important role played by macrophages and contributed to the impaired bacterial clearance that we phage subsets in intestinal homeostasis is evidenced by a growing and others have observed with ATG16L1 deficiency (10, 11, 13). body of literature showing that either an increase in proin- The impaired macromolecule trafficking with ATG16L1 defi- flammatory macrophages or a reduction in regulatory macrophages ciency could lead to defective presentation of Ag onto MHC-II to could worsen mucosal inflammation (20, 26, 36, 37; reviewed in activate T cells (Fig. 7). Insufficient activation of the adaptive Ref. 38). immune response with ATG16L1 deficiency could be one of the Our result showing impaired clearance of Salmonella in the mechanisms that contributed to ineffective elimination of patho- ATG16L1-deficient macrophages is consistent with other reports genic bacteria and the persistent activation of intestinal inflam- on the antibacterial function attributed to autophagy (10, 11, 13, mation in the chronic DSS-induced colitis model that were seen in 39, 40). In contrast, there are also reports showing that ATG16L1 our study. The impaired bacteria clearance due to defective innate deficiency could lead to increased clearance of microbes, includ- and adaptive immunity with ATG16L1 deficiency could lead to ing uropathogenic Escherichia coli and Citrobacter rodentium the increased number of colitogenic bacteria that contributed to (41, 42). The differences in antibacterial function of ATG16L1 both murine colitis and human IBD. Recently, IgA-coated bacteria may be dependent on several factors such as the specific experi- were implicated as IBD-promoting microbes (15). Consistently, mental bacteria being tested, affected by the composition of we observed that increased amounts of IgA-coated bacteria were commensal bacterial in the host gut, or cell type–specific function found in mice with Atg16l1 deficiency and in humans with the of autophagy. The cell type–specific function of ATG16L1 was ATG16L1 T300A CD-associated variant, which was further in- illustrated by the fact that the resistance to C. rodentium infection creased in the context of intestinal inflammation. These findings in Atg16L1 hypomorph (Atg16l1HM) mice was due to non- suggest an intriguing rationale for studying whether carriage of the hematopoietic cells because cell type–specific deletion of Atg16l1 ATG16L1 T300A SNP results in a unique microbial composition 12 MYELOID ATG16L1 MAINTAINS INTESTINAL HOMEOSTASIS

FIGURE 10. CD risk ATG16L1 T300A increased ROS production and mitochondria number. (A) Relative ROS production, as measured by luminol-dependent chemiluminescence, was determined for risk or non-risk for ATG16L1 T300A variant PBM treated with S. typhimurium or zymosan (filled diamonds and solid line) and at baseline (filled diamonds). Representative data from at least six independent experiments (from six human subjects in each group) with similar results are shown. (B) Representative TEM image of non-risk (n = 12 PBM) or risk ATG16L1 T300A variant (n = 12 PBM examined) PBM. At least five subjects per group were used. White arrowheads point to mitochondria, and mitochondria per BMM are quantitated. White scale bars in TEM images, 2 mm. **p , 0.01. Downloaded from http://www.jimmunol.org/ that could be a potential target for precision therapeutics for IBD (which may contribute to intestinal inflammation) (7, 10). We patients carrying the ATG16L1 T300A SNP. showed that the increased ROS with ATG16L1 deficiency was In this study, we added novel mechanisms to previous findings mediated by its increased production through NADPH oxidase and that ATG16L1 deficiency increased cellular production of ROS mitochondria. Our study showed that with ATG16L1 deficiency, by guest on October 2, 2021

FIGURE 11. CD risk ATG16L1 T300A alters mitochondria phenotype. Representative flow cytometry plots of risk or non-risk ATG16L1 T300A PBM that are either unstained (filled) or stained with MitoSOX (open) are shown for non-IBD control (A, left panels) and CD patients (B, left panels). Rep- resentative flow cytometry plots of non-risk and risk ATG16L1 T300A variant PBM with and without S. typhimurium infection and stained with MitoTracker Deep Red and MitoTracker Green are shown and quantitated for non-IBD control (A, middle and right panels) and CD patients (B, middle and right panels). Each filled circle represents an independent experiment from a human subject. *p , 0.05, **p , 0.01. The Journal of Immunology 13 there was increased phosphorylated p40phox protein due to re- 7. Saitoh, T., N. Fujita, M. H. Jang, S. Uematsu, B. G. Yang, T. Satoh, H. Omori, duced maturation of p40phox+ phagosomes, which contributed to T. Noda, N. Yamamoto, M. Komatsu, et al. 2008. Loss of the autophagy protein + Atg16L1 enhances endotoxin-induced IL-1b production. Nature 456: 264–268. increased ROS production via increased p40phox phagosomes 8. Zhou, R., A. S. Yazdi, P. Menu, and J. Tschopp. 2011. A role for mitochondria in that had been shown to increase ROS production (22). Addition- NLRP3 inflammasome activation. [Published erratum appears in 2011 Nature 475: 122.] Nature 469: 221–225. ally, our study showed that the mitochondria in both mouse and 9. Zhong, Z., A. Umemura, E. Sanchez-Lopez, S. Liang, S. Shalapour, J. Wong, human ATG16L1-deficient macrophages produced more ROS at F. He, D. Boassa, G. Perkins, S. R. Ali, et al. 2016. NF-kB restricts inflamma- baseline and with S. typhimurium stimulation, owing to both in- some activation via elimination of damaged mitochondria. Cell 164: 896–910. 10. Lassen, K. G., P. Kuballa, K. L. Conway, K. K. Patel, C. E. Becker, creased numbers of mitochondria and increased accumulation of J. M. Peloquin, E. J. Villablanca, J. M. Norman, T. C. Liu, R. J. Heath, et al. abnormal mitochondria that were more prone to produce ROS. 2014. Atg16L1 T300A variant decreases selective autophagy resulting in altered Our data from both non-IBD and CD patients who did not carry cytokine signaling and decreased antibacterial defense. Proc. Natl. Acad. Sci. USA 111: 7741–7746. any of the NOD2 CD-associated variants and were homozygous for 11. Murthy, A., Y. Li, I. Peng, M. Reichelt, A. K. Katakam, R. Noubade, M. Roose- either the risk or non-risk ATG16L1 CD-associated T300A variant Girma, J. DeVoss, L. Diehl, R. R. Graham, and M. van Lookeren Campagne. further validated and added greater relevance to our murine 2014. A Crohn’s disease variant in Atg16l1 enhances its degradation by caspase 3. Nature 506: 456–462. findings. Additionally, potential gene-dependent (ATG16L1) ver- 12. VanDussen, K. L., T. C. Liu, D. Li, F. Towfic, N. Modiano, R. Winter, sus disease (CD)-dependent effects were controlled by studying T. Haritunians, K. D. Taylor, D. Dhall, S. R. Targan, et al. 2014. Genetic variants non-IBD and CD patients with and without the ATG16L1 variant. synthesize to produce paneth cell phenotypes that define subtypes of Crohn’s disease. Gastroenterology 146: 200–209. We observed that polarization of monocytes to M1 proinflammatory 13. Conway, K. L., P. Kuballa, J. H. Song, K. K. Patel, A. B. Castoreno, macrophages was likely an ATG16L1 T300A-dependent effect be- O. H. Yilmaz, H. B. Jijon, M. Zhang, L. N. Aldrich, E. J. Villablanca, et al. 2013. cause it was found in both non-IBD and CD patients to a similar Atg16l1 is required for autophagy in intestinal epithelial cells and protection of

mice from Salmonella infection. Gastroenterology 145: 1347–1357. Downloaded from magnitude and unlikely due to its known interaction with NOD2 14. Adolph, T. E., M. F. Tomczak, L. Niederreiter, H. J. Ko, J. Bo¨ck, E. Martinez- (Fig. 8C, 8D). In contrast, although ATG16L1 itself increased levels Naves, J. N. Glickman, M. Tschurtschenthaler, J. Hartwig, S. Hosomi, et al. of IgA-coated bacteria in the stool, having CD may likely play a 2013. Paneth cells as a site of origin for intestinal inflammation. Nature 503: 272–276. bigger role because the magnitude of IgA-coated bacteria was 15. Palm, N. W., M. R. de Zoete, T. W. Cullen, N. A. Barry, J. Stefanowski, L. Hao, greater (Supplemental Fig. 1C). P. H. Degnan, J. Hu, I. Peter, W. Zhang, et al. 2014. Immunoglobulin A coating In conclusion, this study showed that colitis development in mice identifies colitogenic bacteria in inflammatory bowel disease. Cell 158: 1000– 1010. http://www.jimmunol.org/ and humans could not be attributed solely to ATG16L1-dependent 16. Barthel, M., S. Hapfelmeier, L. Quintanilla-Martıńez, M. Kremer, M. Rohde, epithelial/Paneth cell defects because lack of ATG16L1 in mac- M. Hogardt, K. Pfeffer, H. Russmann,€ and W. D. Hardt. 2003. Pretreatment of mice with streptomycin provides a Salmonella enterica serovar Typhimurium rophages and other myeloid cells similarly exacerbates murine colitis model that allows analysis of both pathogen and host. Infect. Immun. 71: colitis. We showed that ATG16L1 functions in macrophages to 2839–2858. regulate inflammatory versus regulatory macrophage subsets, 17. Ungaro, R., M. Fukata, D. Hsu, Y. Hernandez, K. Breglio, A. Chen, R. Xu, J. Sotolongo, C. Espana, J. Zaias, et al. 2009. A novel Toll-like receptor 4 an- bacteria clearance, Ag processing and trafficking to the lysosomal tagonist antibody ameliorates inflammation but impairs mucosal healing in compartment, mitophagy, and appropriate ROS production. Par- murine colitis. Am. J. Physiol. Gastrointest. Liver Physiol. 296: G1167–G1179. ticularly, bacteria clearance and appropriate ROS production may 18. Barrett, R., X. Zhang, H. W. Koon, M. Vu, J. Y. Chang, N. Yeager, be a general function of ATG16L1 in the myeloid compartment, as M. A. Nguyen, K. S. Michelsen, D. Berel, C. Pothoulakis, et al. 2012. Consti- tutive TL1A expression under colitogenic conditions modulates the severity and by guest on October 2, 2021 similar alterations were also seen in both Atg16l1-deficient mac- location of gut mucosal inflammation and induces fibrostenosis. Am. J. Pathol. rophages and neutrophils. 180: 636–649. 19. Cleynen, I., G. Boucher, L. Jostins, L. P. Schumm, S. Zeissig, T. Ahmad, V. Andersen, J. M. Andrews, V. Annese, S. Brand, et al; International Inflam- Acknowledgments matory Bowel Disease Genetics Consortium. 2016. Inherited determinants of Crohn’s disease and ulcerative colitis phenotypes: a genetic association study. We thank Jeremy Chen for assistance with Western blots, Guo Ying for Lancet 387: 156–167. setting up the mitochondria assay, Wei Zhu for TEM studies, and Loren 20. Shouval, D. S., A. Biswas, J. A. Goettel, K. McCann, E. Conaway, N. S. Redhu, Karp for critical reading of the manuscript. Human specimens were pro- I. D. Mascanfroni, Z. Al Adham, S. Lavoie, M. Ibourk, et al. 2014. Interleukin- vided by the MIRIAD Biobank. 10 receptor signaling in innate immune cells regulates mucosal immune toler- ance and anti-inflammatory macrophage function. Immunity 40: 706–719. 21. Shih, D. Q., L. Y. Kwan, V. Chavez, O. Cohavy, R. Gonsky, E. Y. Chang, Disclosures C. Chang, C. O. Elson, and S. R. Targan. 2009. Microbial induction of inflam- The authors have no financial conflicts of interest. matory bowel disease associated gene TL1A (TNFSF15) in antigen presenting cells. Eur. J. Immunol. 39: 3239–3250. 22. Ma, J., C. Becker, C. Reyes, and D. M. Underhill. 2014. Cutting edge: FYCO1 recruitment to dectin-1 phagosomes is accelerated by light chain 3 protein and References regulates phagosome maturation and reactive oxygen production. J. Immunol. 1. Franke, A., D. P. McGovern, J. C. Barrett, K. Wang, G. L. Radford-Smith, 192: 1356–1360. T. Ahmad, C. W. Lees, T. Balschun, J. Lee, R. Roberts, et al. 2010. Genome- 23. Homer, C. R., A. L. Richmond, N. A. Rebert, J. P. Achkar, and C. McDonald. wide meta-analysis increases to 71 the number of confirmed Crohn’s disease 2010. ATG16L1 and NOD2 interact in an autophagy-dependent antibacterial susceptibility loci. Nat. Genet. 42: 1118–1125. pathway implicated in Crohn’s disease pathogenesis. Gastroenterology 139: 2. Jostins, L., S. Ripke, R. K. Weersma, R. H. Duerr, D. P. McGovern, K. Y. Hui, 1630–1641.e2. J. C. Lee, L. P. Schumm, Y. Sharma, C. A. Anderson, et al; International IBD 24. Kuma, A., M. Hatano, M. Matsui, A. Yamamoto, H. Nakaya, T. Yoshimori, Y. Ohsumi, T. Tokuhisa, and N. Mizushima. 2004. The role of autophagy during Genetics Consortium (IIBDGC). 2012. Host-microbe interactions have shaped the early neonatal starvation period. Nature 432: 1032–1036. the genetic architecture of inflammatory bowel disease. Nature 491: 119–124. 25. Mizushima, N., A. Kuma, Y. Kobayashi, A. Yamamoto, M. Matsubae, T. Takao, 3. Liu, J. Z., S. van Sommeren, H. Huang, S. C. Ng, R. Alberts, A. Takahashi, T. Natsume, Y. Ohsumi, and T. Yoshimori. 2003. Mouse Apg16L, a novel WD- S. Ripke, J. C. Lee, L. Jostins, T. Shah, et al; International Multiple Sclerosis repeat protein, targets to the autophagic isolation membrane with the Apg12- Genetics Consortium; International IBD Genetics Consortium. 2015. Associa- Apg5 conjugate. J. Cell Sci. 116: 1679–1688. tion analyses identify 38 susceptibility loci for inflammatory bowel disease and 26. Zigmond, E., C. Varol, J. Farache, E. Elmaliah, A. T. Satpathy, G. Friedlander, highlight shared genetic risk across populations. Nat. Genet. 47: 979–986. M. Mack, N. Shpigel, I. G. Boneca, K. M. Murphy, et al. 2012. Ly6Chi mono- 4. Hampe, J., A. Franke, P. Rosenstiel, A. Till, M. Teuber, K. Huse, M. Albrecht, cytes in the inflamed colon give rise to proinflammatory effector cells and mi- G. Mayr, F. M. De La Vega, J. Briggs, et al. 2007. A genome-wide association gratory antigen-presenting cells. Immunity 37: 1076–1090. scan of nonsynonymous SNPs identifies a susceptibility variant for Crohn dis- 27. Shiloh, M. U., J. D. MacMicking, S. Nicholson, J. E. Brause, S. Potter, ease in ATG16L1. Nat. Genet. 39: 207–211. M. Marino, F. Fang, M. Dinauer, and C. Nathan. 1999. Phenotype of mice and 5. He, C., and D. J. Klionsky. 2009. Regulation mechanisms and signaling path- macrophages deficient in both phagocyte oxidase and inducible nitric oxide ways of autophagy. Annu. Rev. Genet. 43: 67–93. synthase. Immunity 10: 29–38. 6. Elliott, E. I., and F. S. Sutterwala. 2015. Initiation and perpetuation of NLRP3 28. Mastroeni, P., A. Vazquez-Torres, F. C. Fang, Y. Xu, S. Khan, C. E. Hormaeche, inflammasome activation and assembly. Immunol. Rev. 265: 35–52. and G. Dougan. 2000. Antimicrobial actions of the NADPH phagocyte oxidase 14 MYELOID ATG16L1 MAINTAINS INTESTINAL HOMEOSTASIS

and inducible nitric oxide synthase in experimental salmonellosis. II. Effects on sponse in obese mice by promoting proinflammatory macrophage polarization. microbial proliferation and host survival in vivo. J. Exp. Med. 192: 237–248. Autophagy 11: 271–284. 29. Vazquez-Torres, A., J. Jones-Carson, P. Mastroeni, H. Ischiropoulos, and 36. Tamoutounour, S., S. Henri, H. Lelouard, B. de Bovis, C. de Haar, C. J. van der F. C. Fang. 2000. Antimicrobial actions of the NADPH phagocyte oxidase and Woude, A. M. Woltman, Y. Reyal, D. Bonnet, D. Sichien, et al. 2012. CD64 inducible nitric oxide synthase in experimental salmonellosis. I. Effects on mi- distinguishes macrophages from dendritic cells in the gut and reveals the Th1- crobial killing by activated peritoneal macrophages in vitro. J. Exp. Med. 192: inducing role of mesenteric lymph node macrophages during colitis. Eur. J. 227–236. Immunol. 42: 3150–3166. 30. Barnden, M. J., J. Allison, W. R. Heath, and F. R. Carbone. 1998. Defective TCR 37. Platt, A. M., C. C. Bain, Y. Bordon, D. P. Sester, and A. M. Mowat. 2010. An expression in transgenic mice constructed using cDNA-based a- and b-chain independent subset of TLR expressing CCR2-dependent macrophages promotes genes under the control of heterologous regulatory elements. Immunol. Cell Biol. colonic inflammation. J. Immunol. 184: 6843–6854. 76: 34–40. 38. Bain, C. C., and A. M. Mowat. 2014. Macrophages in intestinal homeostasis and 31. Clausen, B. E., C. Burkhardt, W. Reith, R. Renkawitz, and I. Fo¨rster. 1999. inflammation. Immunol. Rev. 260: 102–117. Conditional gene targeting in macrophages and granulocytes using LysMcre 39. Castillo, E. F., A. Dekonenko, J. Arko-Mensah, M. A. Mandell, N. Dupont, mice. Transgenic Res. 8: 265–277. S. Jiang, M. Delgado-Vargas, G. S. Timmins, D. Bhattacharya, H. Yang, et al. 32. Sorbara, M. T., L. K. Ellison, M. Ramjeet, L. H. Travassos, N. L. Jones, 2012. Autophagy protects against active tuberculosis by suppressing bacterial S. E. Girardin, and D. J. Philpott. 2013. The protein ATG16L1 suppresses in- burden and inflammation. Proc. Natl. Acad. Sci. USA 109: E3168–E3176. flammatory cytokines induced by the intracellular sensors Nod1 and Nod2 in an 40. Lapaquette, P., M. A. Bringer, and A. Darfeuille-Michaud. 2012. Defects in autophagy-independent manner. [Published erratum appears in 2014 Immunity 40: 303.] Immunity 39: 858–873. autophagy favour adherent-invasive Escherichia coli persistence within macro- 33. Travassos, L. H., L. A. Carneiro, M. Ramjeet, S. Hussey, Y. G. Kim, phages leading to increased pro-inflammatory response. Cell. Microbiol. 14: J. G. Magalha˜es, L. Yuan, F. Soares, E. Chea, L. Le Bourhis, et al. 2010. Nod1 791–807. and Nod2 direct autophagy by recruiting ATG16L1 to the plasma membrane at 41. Marchiando, A. M., D. Ramanan, Y. Ding, L. E. Gomez, V. M. Hubbard-Lucey, the site of bacterial entry. Nat. Immunol. 11: 55–62. K. Maurer, C. Wang, J. W. Ziel, N. van Rooijen, G. Nun˜ez, et al. 2013. A de- 34. Cooney, R., J. Baker, O. Brain, B. Danis, T. Pichulik, P. Allan, D. J. Ferguson, ficiency in the autophagy gene Atg16L1 enhances resistance to enteric bacterial B. J. Campbell, D. Jewell, and A. Simmons. 2010. NOD2 stimulation induces infection. Cell Host Microbe 14: 216–224. autophagy in dendritic cells influencing bacterial handling and antigen presen- 42. Wang, C., G. R. Mendonsa, J. W. Symington, Q. Zhang, K. Cadwell, tation. Nat. Med. 16: 90–97. H. W. Virgin, and I. U. Mysorekar. 2012. Atg16L1 deficiency confers protection Downloaded from 35. Liu, K., E. Zhao, G. Ilyas, G. Lalazar, Y. Lin, M. Haseeb, K. E. Tanaka, and from uropathogenic Escherichia coli infection in vivo. Proc. Natl. Acad. Sci. M. J. Czaja. 2015. Impaired macrophage autophagy increases the immune re- USA 109: 11008–11013. http://www.jimmunol.org/ by guest on October 2, 2021 A Precolitic Salmonella Acute DSS Chronic DSS Flox 2.0 % Flox 25.4 % Flox 18.3 % Flox 24.9 %

∆Mye 9.8% ∆Mye 56.5% ∆Mye 30.1% ∆Mye 40.2% Events

IgA *** *** *** 100 *** *** 80 ** * * * * 60

20 % IgA Bacteria IgA % 0 Flox ∆Mye Flox ∆Mye Flox ∆Mye Flox ∆Mye Precolitic Salmonella Acute DSS Chronic DSS

B 50 ** 6000 ** C Control Pair 1 Control Pair 2 20 * 5000

40 WT 5.3% g/g)

g/g) WT 3.2% 16 μ μ 4000 T300A 10.1% T300A 14.1% 30 3000 12 20 2000 8 10 1000

Stool IgA ( IgA Stool 4

Stool IgA ( IgA Stool % IgA Bacteria IgA % 0 0 Flox ∆Mye Flox ∆Mye 0 IgA+ bacteria IgA+ bacteria WT T300A Precolitic Salmonella CD Pair 1 CD Pair 2 100 100 100 WT 30.5% WT 46.8% ***

80 80 T300A 48.2% T300A 68.4% 80

g/g) g/g) μ 60 μ 60 60 40 40 40 20

20 20 Bacteria IgA %

Stool IgA ( IgA Stool Stool IgA ( IgA Stool 0 0 IgA+ bacteria IgA+ bacteria 0 Flox ∆Mye Flox ∆Mye WT T300A Acute DSS Chronic DSS Supplemental Figure 1. Increased IgA-coated bacteria in the stool of mice with Atg16l1 deficiency in myeloid cells. (A) Representative flow cytometry plots of fecal bacteria stained with anti-IgA antibody from 2 month old pre-colitic mice and from mice undergoing the 3 colitis models were shown (top panels) and quantitated (bottom panel). (B) Total IgA in the stools was quantitated by IgA ELISA and represented as microgram (μg) of IgA per gram (g) of stool. Each filled circle represents an independent experiment from each mouse. *P<0.05, **P<0.01.

A B

5 5 5 5 5000 *** 4 4 4000

3 3 3000 Neutrophils 2 2 5 2000 1 1 1000

0 0 CFU/1x10 0 Migrated Neutrophils x 10 x Neutrophils Migrated Migrated Neutrophils x 10 x Neutrophils Migrated Flox ∆Mye Flox ∆Mye Flox ∆Mye MIP-2 Supernatant C Salmonella Zymosan 25000 250000 Atg16l1∆Mye 20000 Atg16l1∆Mye 200000 15000 150000

10000 100000 flox

Atg16l1flox Atg16l1 Relative ROS Units ROS Relative 5000 Units ROS Relative 50000

0 0 0 10 20 30 40 50 60 0 10 20 30 40 50 60 Minutes Minutes

Supplemental Figure 2. Alterations of neutrophil function with Atg16l1 deficiency. (A) MIP-2 or supernatant derived from WT macrophages stimulated with S. typhimurium were added to the bottom well with either WT or Atg16l1-deficient neutrophils added to the upper well of a transwell. The number of neutrophils migrating to the bottom well was counted. Data showed that neutrophils from WT and Atg16l1∆Mye mice has similar migratory responses. (B) WT and Atg16l1∆Mye neutrophils were infected with S. typhimurium and surviving intracellular bacteria were quantified. Each filled circle in (A) and (B) represents an independent experiment. (C) Relative ROS production as measured by luminol-dependent chemiluminescence was determined over 60 minutes for WT and Atg16l1 deficient neutrophils treated with S. typhimurium or zymosan (filled diamond and solid line) and at baseline (filled diamond). Representative data of 8 independent experiments with all experiments showing Atg16l1-deficient neutrophils producing higher ROS are shown. *P<0.05, **P<0.01.