Exosome-like Nanoparticles from Intestinal Mucosal Cells Carry Prostaglandin E 2 and Suppress Activation of Liver NKT Cells

This information is current as Zhong-Bin Deng, Xiaoying Zhuang, Songwen Ju, Xiaoyu of September 29, 2021. Xiang, Jingyao Mu, Yuelong Liu, Hong Jiang, Lifeng Zhang, James Mobley, Craig McClain, Wenke Feng, William Grizzle, Jun Yan, Donald Miller, Mitchell Kronenberg and Huang-Ge Zhang

J Immunol 2013; 190:3579-3589; Prepublished online 6 Downloaded from March 2013; doi: 10.4049/jimmunol.1203170 http://www.jimmunol.org/content/190/7/3579 http://www.jimmunol.org/ Supplementary http://www.jimmunol.org/content/suppl/2013/03/07/jimmunol.120317 Material 0.DC1 References This article cites 50 articles, 11 of which you can access for free at: http://www.jimmunol.org/content/190/7/3579.full#ref-list-1

Why The JI? Submit online. by guest on September 29, 2021

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

• No Triage! Every submission reviewed by practicing scientists

• 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 Errata An erratum has been published regarding this article. Please see next page or: /content/196/11/4832.full.pdf

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 © 2013 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Exosome-like Nanoparticles from Intestinal Mucosal Cells Carry Prostaglandin E2 and Suppress Activation of Liver NKT Cells

Zhong-Bin Deng,* Xiaoying Zhuang,* Songwen Ju,* Xiaoyu Xiang,* Jingyao Mu,* Yuelong Liu,† Hong Jiang,* Lifeng Zhang,* James Mobley,‡ Craig McClain,x Wenke Feng,x William Grizzle,† Jun Yan,* Donald Miller,* Mitchell Kronenberg,{ and Huang-Ge Zhang*,‖

Regulation and induction of anergy in NKT cells of the liver can inhibit autoimmune and antitumor responses by mechanisms that are poorly understood. We investigated the effects of PGE2, delivered by intestinal, mucus-derived, exosome-like nanoparticles Downloaded from (IDENs), on NKT cells in mice. In this study, we demonstrate that IDENs migrate to the liver where they induce NKT cell anergy.

These effects were mediated by an IDENs’ PGE2. Blocking PGE2 synthesis attenuated IDENs inhibition of induction of IFN-g and IL-4 by a-galactosylceramide (a-GalCer)–stimulated liver NKT cells in a PGE2 E-type prostanoid 2/E-type prostanoid 4 recep- tor–mediated manner. Proinflammatory conditions enhanced the migration of IDENs to the liver where a-GalCer and PGE2 induced NKT anergy in response to subsequent a-GalCer stimulation. These findings demonstrate that IDENs carrying PGE2 can be transferred from the intestine to the liver, where they act as immune modulators, inducing an anergic-like state of NKT cells. http://www.jimmunol.org/ These reagents might be developed as therapeutics for autoimmune liver diseases. The Journal of Immunology, 2013, 190: 3579– 3589.

he gut-associated is complex and involves thought to play an active role in systemic tumor surveillance and extensive communication with the liver (1, 2). Glycolipids maintenance of tolerance (9). T derived from dietary metabolic products and lipids de- Exosomes are microvesicles released naturally from a variety of rived from some intestinal (3, 4) migrate constantly into cells (10). They carry a cargo of , lipids, mRNAs, and/or the liver, and these lipids can activate liver NKT cells through microRNAs and can transfer their cargo to recipient cells, thus TCR engagement in vitro (5, 6). Liver NKT cells can also be serving as extracellular messengers to mediate cell–cell commu- by guest on September 29, 2021 activated in a TCR-independent manner (7, 8). It is remarkable nication. Numerous recent reports indicate that exosomes have a that liver NKT cells are normally quiescent, even though they are significant effect on physiological and pathological processes in constantly exposed to intestinal-derived products and are also recipient cells (11). Intestinal epithelial cells release exosomes (30–100 nm) (12) and have been shown to be capable of regulating

*Department of Microbiology and Immunology, Brown Cancer Center, University of the differentiation and activity of several different types of immune Louisville, Louisville, KY 40202; †Department of Pathology, University of Alabama, cells (12–14), which suggests the possibility that intestinal-derived Birmingham, AL 35294; ‡Department of Radiology, University of Alabama, Bir- exosome-like nanoparticles could play a key role in communica- mingham, AL 35294; xDivision of Gastroenterology, Department of Medicine, Uni- versity of Louisville, Louisville, KY 40202; {La Jolla Institute for Allergy and tion between the intestine and liver and in gut-related induction of Immunology, La Jolla, CA 92037; and ‖Louisville Veterans Administration Medical immune tolerance or anergic-like state (12, 15). Center, Louisville, KY 40206 In this study, we demonstrate that IDENs migrate into the liver and Received for publication November 20, 2012. Accepted for publication January 23, that inflammation promotes this migration. We also demonstrate that 2013. IDENs can induce an anergic-like state in liver NKT cells. Induction This work was supported by National Institutes of Health Grants RO1CA137037, R01CA107181, RO1AT004294, and RO1CA116092; Louisville Veterans Adminis- of a NKT cell anergic-like state by IDENs occurs through a previ- tration Medical Center merit review grants (to H.-G.Z.); and a grant from the Susan ously undescribed mechanism where the transport of PGE2 by the G. Komen Breast Cancer Foundation. IDENs (16) into the liver creates an environment that promotes the Z.-B.D. and H.-G.Z. designed the research, analyzed and interpreted data, and drafted NKT cell anergic-like state. the manuscript; D.M., M.K., C.M., and W.G. interpreted findings and drafted the manuscript; and X.Z., S.J., X.X., Y.L., J. Mu, H.J., L.Z., J. Mobley, W.F., and J.Y. performed experiments and interpreted data. Materials and Methods Address correspondence and reprint requests to Dr. Huang-Ge Zhang and Dr. Zhong- Isolation and purification of IDENs Bin Deng, Brown Cancer Center, University of Louisville, Clinical and Translational Research Building 309, 505 Hancock Street, Louisville, KY 40202. E-mail addresses: Mice were euthanized, and the small intestines were removed. The luminal [email protected] (H.-G.Z.) and [email protected] (Z.-B.D.) contents of the intestines were removed by gently flushing the intestines with The online version of this article contains supplemental material. 10 ml ice-cold PBS. The intestines were then opened longitudinally, the mucus was collected by mild physical separation using round forceps, and the Abbreviations used in this article: ALT, alanine aminotransferase; AST, aspartate ami- mucus was soaked in PBS before agitating on a rotator at 100 rpm for 15 min. notransferase; BMDC, bone marrow-derived dendritic cell; COX, cyclooxygenase; DC, dendritic cell; DiR, dioctadecyl-tetramethylindotricarbocyanine iodide; DSS, dextran The mucus–PBS mixture was processed by differential centrifugation and sodium sulfate; EP, E-type prostanoid; a-GalCer, a-galactosylceramide; IDEN, sucrose gradient centrifugation to isolate and purify the exosome-like intestinal, mucus-derived, exosome-like nanoparticle; PKA, kinase A. nanoparticles or microparticles (pellets after 10,000 3 g centrifugation) using a previously described protocol (17). The identical protocol was also Copyright Ó 2013 by The American Association of Immunologists, Inc. 0022-1767/13/$16.00 used for isolation and purification of intestinal IDENs from surgi- www.jimmunol.org/cgi/doi/10.4049/jimmunol.1203170 3580 INTESTINAL-DERIVED EXOSOMES INDUCE LIVER NKT ANERGY cally removed tissue. Specimens were prepared for electron microscopy Enzymatic Assay Kit (Thermo Scientific). For histopathology, H&E staining using a conventional procedure (18), stained with 1% uranyl acetate for was performed on paraffin-embedded liver sections. For immunofluores- 15 s, and observed in an FEI Tecnai Spirit T12 electron microscope operated cence analysis, OCT (Sakura Finetek)-embedded tissue cryosections (9- at 200 kV at a magnification of 338,000 and defocus of 2.5 mm. Photo- mm-thick) were fixed at 220˚C in methanol/acetone (3:1). Slides were micrographs were taken using a Gatan Ultrascan 4000 charge-coupled de- hydrated in PBS and blocked for 30 min at 25˚C with Fc block (10 mg/ml) vice camera. Protein concentration was determined using the Bio-Rad and 5% (v/v) normal horse serum in PBS. After blockade, slides were in- Protein Quantitation Assay with BSA as a standard. The levels of PGE2 cubated for 30 min at 25˚C with “mixtures” containing the following pri- in IDENs were determined using an ELISA kit (Cayman Chemical). mary Abs in PBS: anti-A33 (Santa Cruz Biotechnology), anti-CD31 (MEC 13.3; BD Biosciences), and anti–Lyve-1 (R&D Systems). Primary Abs were Mice detected with Alexa Fluor 647–donkey anti-goat, rabbit Ab to FITC–Alexa Fluor 488 (A11090) (S21381; all from Invitrogen Life Sciences), or C57BL/6 mice 6–12 wk of age were obtained from The Jackson Labora- tory. All animal procedures were approved by the University of Louisville streptavidin–allophycocyanin (BD Biosciences). Slides were mounted with Institutional Animal Care and Use Committee. Slow Fade Gold Antifade plus DAPI (S36938; Molecular Probes and Mice were administered by gavage or i.v. injection IDENs resuspended in Invitrogen Life Sciences). Tissues were assessed using a Zeiss LSM 510 PBS as described previously (19). confocal microscope equipped with a digital image analysis system (Pix- Autoimmune hepatitis was induced as previously described (20) using era). Apoptosis was evaluated using the TUNEL assay performed on frozen liver sections using the in situ cell death detection kit (Roche). Con A from Vector Laboratories administered i.v. at 25 or 37.5 mg/kg + + body weight. For survival studies, mice were injected with 37.5 mg/kg Con Cell sorting. For measuring induction, NK1.1 CD5 NKT cells A. Colitis was induced using 3% solution of dextran sodium sulfate (DSS) were sorted from liver leukocytes using a method as described (23). The with a molecular mass of 36,000–50,000 Da (MP Biomedicals) that was purity of the NKT cells collected, typically .90%, was assessed using provided as drinking water to mice (18–25 g, 6 wk old, male) for 5 d. In CD1d tetramers loaded with a-galactosylceramide (a-GalCer; provided by some experiments, mice were administered indomethacin (4 mg/kg/day) the National Institute of Allergy and Infectious Diseases Tetramer Facility, (Sigma-Aldrich, St. Louis, MO) in the drinking water. National Institutes of Health). For proliferation assays, NKT cells were Downloaded from Clinical specimens. Human intestinal specimens were collected at the time also enriched from liver by negative magnetic bead sorting (Miltenyi a of gastric bypass surgery. All samples were collected with the informed Biotec) using anti-CD11b, B220, CD8 , Gr-1, L-, and CD11c Abs consent of the patients, and the experiments were approved by the Institu- using a standard protocol (24). Although CD4 is not depleted in this case, tional Review Board Committee of the University of Alabama (Birmingham, the specificity of NKT cell proliferation was determined by stimulation of a AL). -GalCer, which specifically activates NKT cells but not CD4 T cells. The purity of NKT and CD11c cells was determined by FACS analysis as

Cytokine detection described previously (24). http://www.jimmunol.org/ IFN-g and IL-4 in culture supernatants and liver homogenates were quan- Cell proliferation and CFSE dilution analysis tified using ELISA kits (eBioscience). NKT cells enriched by negative selection (4 3 105) were cultured with liver 3 5 Preparation of mononuclear cells from the liver and flow dendritic cells (DCs) (1 10 ) isolated by CD11c magnetic bead sorting (Miltenyi Biotec). In vitro cultures were incubated for 72 h under the in- cytometry analysis dicated culture conditions, and [3H]thymidine was added for the last 16 h. Mononuclear cells were prepared from perfused livers (21) that were ho- Incorporation of radioactivity was determined using a Betaplate liquid mogenized with a 70-mm pore filter. After washing extensively, the homog- scintillation counter (Packard Bioscience). Results are expressed as mean 6 enates were resuspended in a 33% Percoll gradient at 22˚C (25 ml/liver) and cpm of triplicate wells SEM. For CFSE (Invitrogen) dilution analysis, splenocytes were labeled with 5 mM CFSE for 10 min at 37˚C in PBS centrifuged at 2300 rpm for 20 min. The cell pellet was collected, and RBCs by guest on September 29, 2021 were removed using red cell lysis buffer. containing 0.1% BSA and washed twice with complete RPMI 1640 me- 3 5 For flow cytometry analysis, the cells were labeled using standard pro- dium. Labeled NKT cells (4 10 cells/well) were cultured with liver 3 4 cedures (22). The following mAbs reactive with murine cells were purchased DCs (5 10 ) under the indicated culture conditions for 96 h in complete RPMI 1640 medium and stained with TCR-b and a-GalCer/CD1d. CFSE from BioLegend: CD69 (H1.2F3), CD40L (MR1), TCRab (H57-597), and + IFN-g (XMG1.2); these were used as direct conjugates to FITC, PE, PerCP, dilution was examined on gated a-GalCer/CD1d tetramer NKT cells. PE Cy7, or allophycocyanin. Mouse CD1d tetramer loaded with the a-C- Western blot analysis galactosylceramine analog PBS-57 and conjugated with APC or PE and the unloaded control were provided by the National Institutes of Health Tetramer Western blots were carried out as described previously (19). In brief, cells Core Facility (Atlanta, GA). Cells were analyzed by flow cytometry using or exosomes were lysed in protein lysis buffer containing 1% Triton X- a FACScan (BD Biosciences protocol; BD Pharmingen). Histogram analysis 100, 0.1% SDS, 150 mM NaCl, 50 mM Tris-HCl, 1 mM EDTA, 1 mM was performed using FlowJo software (Tree Star, Ashland, OR). EGTA, 5 mM sodium molybdate, and 20 mM phenylphosphate with protease and phosphatase inhibitors (1 mM PMSF, 10 mg/ml aprotinin, 20 In vivo imaging of IDENs mg/ml leupeptin, 20 mg/ml pepstatin A, 50 mM NaF, and 1 mM sodium Mice were anesthetized with ketamine (100 mg/kg body weight) and xylazine orthovanadate). Proteins of lysed cells were separated on 10% polyacryl- (10 mg/kg body weight) via i.p. injection, and inhaled isoflurane was used as amide gels using SDS-PAGE and transferred to nitrocellulose membranes. necessary. To monitor the trafficking of IDENs administered by gavage, IDENs The membranes were stained with Ponceau Red to validate that all samples contained similar amounts of protein. were first labeled with a near-infrared lipophilic carbocyanine dye, dioctadecyl-tetramethylindotricarbocyanine iodide (DiR; Invitrogen, RNA extraction and PCR Carlsbad, CA), using a previously described method (17). The mice were imaged over a 48-h period using a Carestream Molecular Imaging sys- Total RNA was isolated from livers or NKT hybridomas with TRIzol, tem (Carestream Health, Woodbridge, CT). For controls, mice (five per group) according to the manufacturer’s specifications (Invitrogen), and the RNA received nonlabeled IDENs in PBS at the same concentration as for DiR dye– repurified with an RNeasy mini kit (Qiagen). RNA (1 mg) was reverse labeled exosomes. Images were collected using a high-sensitivity charge- transcribed with Superscript III and random primers (Invitrogen). For coupled device camera with an exposure time for imaging of 2 min. Regions quantitation of of interest, cDNA samples were amplified in a of interest were analyzed for signals and were reported in units of relative CFX96 Real-time System (Bio-Rad Laboratories, Hercules, CA) and photon counts per second. The total photon count per minute was calculated SsoFast Evagreen supermixture (Bio-Rad Laboratories), according to the (five animals) using Living Image software. The effects of DiR dye–labeled manufacturer’s instructions. Fold changes in mRNA expression between versus nonlabeled IDENs on mice was determined by dividing the number of treatments and controls were determined by the dCT method as described photons collected for DiR dye-labeled IDEN treated mice by the number of previously (25). Differences between groups were determined using a two- photons collected for nonlabeled IDEN–treated mice at different imaging time sided Student t test and one-way ANOVA. Error bars on plots represent 6 points. Results were represented as pseudocolor images indicating in- SE, unless otherwise noted. The data were normalized to a GAPDH ref- tensity that were superimposed over grayscale reference photographs. erence. All primers were listed in the Supplemental Table I. Assessment of liver damage Agonists, antagonists, and inhibitors The quantity of alanine aminotransferase (ALT) and aspartate amino- IDEN lipids were extracted according to Bligh and Dyer (26) in chloroform/ transferase (AST) in the sera of the mice was measured using the Infinity methanol/water (2.5:2.5:2.1, v/v/v). Lipids for treatment of cells were The Journal of Immunology 3581 dried under nitrogen and sonicated in PBS at pH 7.4 and 0.05% (v/v) V predominantly detected in microparticles was demonstrated by Tween-20 before use. Polyclonal anti-A33 Ab was purchased from Santa FACS analysis of Annexin V (Fig. 1B, left panel). Electron micro- Cruz Biotechnology (Santa Cruz, CA), and cAMP analogs (dibutyryl graphs (Fig. 1C, left panel) show that IDENs were also present in the cAMP and N6-Bnz-cAMP) and protein kinase A (PKA) inhibitors (H-89, Rp-8-CPT-cAMPS, and Rp-8-Br-cAMPS) were purchased from Sigma- small intestines of human and shared a number of exosomal protein Aldrich. PGE2, Butaprost, misoprostol, sulprostone, and PGE2 ELISA markers identified on mouse IDENs (Fig. 1C, right panel). kits were obtained from Cayman Chemical. E-type prostanoid (EP) 4 Intestinal mucus is a physically cross-linked hydrogel with mesh agonists (ONO-AE1-329), EP1 antagonist (ONO-8713), EP3 antagonist sizes on the order of 10–100 nm (34, 35). Penetration of this barrier (ONO-AE3-240), and EP4 antagonist (ONO-AE3-208) were gifts from Ono Pharmaceutical (Chuo-ku, Osaka, Japan). is largely dependent on the size of the particles (34, 35). The size of the IDENs is ∼30–100 nm, and we used gavage administration Cell cultures of IDENs labeled with an infrared fluorescent membrane dye Bone marrow-derived dendritic cells (BMDCs) were cultured as described (DiR) to track their migration patterns in vivo (17). Imaging previously (27). To study the activity of IDENs or IDEN lipid, total liver revealed an accumulation of fluorescent signals in the gut over the mononuclear (4 3 105/well) or BMDCs cultured with FACS-sorted NKT cells first 24 h (Fig. 1D). At 48 h, signals were detectable in dissected 3 5 3 5 (1 10 BMDCs + 4 10 NKTs/well) or in the presence of IDENs (30 mg/ liver and mesenteric lymph nodes but not the lung, spleen (Fig. ml) for 3 h and then stimulated by a-GalCer (100 ng/ml) for 24–48 h in round bottom 96-well plates containing RPMI 1640 medium supplemented with 1E), or other organs (data not shown). The presence of IDENs in 10% FCS. In some experiments, sorted NKT cells pretreated with IDENs the liver was confirmed by immunohistology and costaining for were stimulated by a-GalCer–tetramer. In some cases, 100 mMPGE2,10mM A33 (Fig. 1F). The colocalization of the DiR and A33 signals with butaprost (EP2 agonist), 40 mM misoprostol (EP4 agonist), 10 mM sulpro- CD31, a marker of endothelial cells (Fig. 1G), and LYVE1, stone (EP1/EP3 agonist), or 100 mM dibutyryl-cAMP,100 mM N6-Bnz- a marker of lymphatic capillaries (Fig. 1H), along the length of the Downloaded from cAMP, 3 mM H-89, 100 mM Rp-8-Br-cAMPS, or 100 mM Rp-8-CPT- cAMPS were added to cocultures. Cells were assessed by flow cytometry, intestinalendothelialandlymphaticvesselswithin6hofad- or supernatants were used for cytokine measurement by ELISA. ministration of IDENs suggests that the IDENs migrate into the liver from the gut primarily through the vascular and lymphatic Liposomes assembled from IDENs vessels. The migration of A33-positive IDENs to the liver (Fig. 1I) Liposomes were assembled from IDENs used the protocol as previously was enhanced in mice treated with either Con A, a lectin that described (28) with minor modification as described below: 1 ml IDENs in induces inflammatory liver injury in mice and is a model of au- http://www.jimmunol.org/ PBS were mixed together with 3.75 ml chloroform/methanol at a 1:2 ratio toimmune liver disease (36), or DSS, which induces colonic in- (v/v) in borosilicate glass test tubes, before adding 1.25 ml chloroform to the mixed sample and vortexing the mixture. Subsequently, 1.25 ml dei- flammation with characteristics of acute ulcerative colitis. Thus, nonized water was added, and the mixture was vortexed again. The sample IDENs can gain access and traffic within the vascular or lymphatic was centrifuged at 1000 rpm in a table-top microcentrifuge for 5 min at system of the liver. Inflammation in the liver or colon enhances 22˚C. The bottom-phased liquid that contained lipids was collected using a this migration. pasteur pipette. Solvent in the collected sample was completely removed via a constant flow of nitrogen gas at 60˚C. Extracted IDEN lipids were IDENs protect against inflammation-mediated liver damage hydrated with 800 ml Milli-Q water prior to bath sonication for 3 min and then sonicated for an additional 2 min after addition of HEPES-buffered To explore the biological effects of the accumulation of IDENs in saline (final contration,10 mM HEPES and 50 mM NaCl). This resulted in the liver, we tested the effects of pretreatment of mice with purified by guest on September 29, 2021 a clear dispersion being produced that was stored at 4˚C for 48 h prior to IDENs on Con A–induced liver injury (36). Mice that were pretreated further experimentation or examination for nanoparticle formation using a conventional procedure (29). with vehicle only succumbed rapidly (Fig. 2A) and exhibited symptoms characteristic of Con A–induced liver injury, including Coupling of IDENs or microparticles to latex beads elevated levels of serum ALT and AST (Fig. 2B). In addition, there and immunofluorescence analysis were histopathological changes in the liver (Fig. 2C) with extensive IDENs or microparticle-coated beads were prepared using a conventional areas of confluent hepatocellular necrosis, marked sinusoidal hy- procedure (29) and washed twice in PBS containing 10% FCS, stained with peremia associated with hemorrhage, and mononuclear cell infiltrates Ab directed against Annexin V (eBioscience), and analyzed using a BD scattered diffusely throughout the viable parenchyma. Pretreatment Biosciences FACSCalibur. A “bead-only” control and an isotype-matched Ab control were prepared and fluorescence intensity was normalized based with IDENs markedly enhanced the survival of the mice (Fig. 2A), on these controls. dampened the Con A–induced elevation in the levels of serum ALT and AST (Fig. 2B), and limited the liver histopathology to minor Statistical analysis inflammatory infiltrates and small necrotic patches (Fig. 2C). Alle- Survival data were analyzed using the log- test. A Student t test was used viation of liver injury appeared to be associated with a reduction in for comparison of two samples with unequal variances. One-way ANOVA inflammatory responses because mice pretreated with IDENs had with Holm’s post hoc test was used for comparing means of three or more serum levels of the proinflammatory IFN-g and TNF-a variables. that were significantly lower than in the mice that were pretreated with vehicle only (Fig. 2D). Furthermore, the administration of Results IDENs reduced apoptosis of liver cells, as assessed by TUNEL IDENs migrate into the liver and inflammation promotes this staining (Fig. 2E). migration Using standard techniques (30), we isolated nanoparticles from the IDENs induce hyporesponsiveness of liver NKT cells to a mucus of the small intestines of naive C57BL/6 mice that were -GalCer identifiable as exosome-like nanoparticles based on electron mi- Con A–induced hepatitis is also dependent on NKT cell activity croscopic examination and their expression of exosomal markers (37). Unlike CD4/CD8 T cells, NKT cells respond to glycolipid (Fig. 1A). Western blot analysis for A33, a marker of intestinal ligands. The glycolipid a-GalCer is a potent stimulator of NKT cell epithelial cells (31), further demonstrated that the preparation con- expansion when first administered but loses its effectiveness on tained a high proportion of intestinal epithelial cell–derived IDENs. repeated administration (9). To determine the effects of accumu- These results were further validated by Western blot analysis (Fig. lation of IDENs in the liver on a-GalCer–driven NKT cell ex- 1A, 1B, right panel) of exosome makers including RAb5 and pansion, mice were injected i.v. with IDENs every 2 d for 2 wk and TSG101 or the microparticle marker, Annexin V (32, 33). Annexin 3 h after the last injection the mice were injected i.v. with a-GalCer 3582 INTESTINAL-DERIVED EXOSOMES INDUCE LIVER NKT ANERGY Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 1. Characterization of IDENs isolated from intestinal mucus. Intestine from B6 mice (A)orhuman(C) was used for isolation of nano- particles or microparticles (B) by differential centrifugation as described previously (27). Sucrose-purified nanoparticles were examined by electron microscopy [(A)and(C), left panel] (bar, 100 nm). FACS analysis was performed to determine expression of Annexin V (Figure legend continues) The Journal of Immunology 3583

FIGURE 2. IDENs protect mice against Con A–induced liver damage. (A–E) C57BL/6 mice were administered IDENs (100 mg/mouse in PBS, injected i.v. every 2 d for 14 d) prior to injection of Con A (37.5 mg/kg of body weight, injected i.v. 4 h after the last injection of exo- somes) and assessed at specified times (n =10 mice/group) for survival of mice after Con A injection (A) and levels of ALT and AST in the sera (B). *p , 0.05, **p , 0.01 as compared with mice injected with PBS (Student t test); (C) H&E-stained sections of the livers; original magnification 320. (D) Levels of IFN-g and TNF-a in the sera. *p , 0.05, **p , 0.01 as compared with mice injected with PBS (Student

t test). (E) Apoptosis in the liver as assessed by Downloaded from TUNEL analysis. Stained with DAPI (blue) and TUNEL (green). Original magnification 320. http://www.jimmunol.org/

(5 mg/mouse). The rationale for choosing a 2-wk treatment schedule suppressed the ex vivo–proliferative response of the NKT cells to was based on data published by others (24). In naive mice, a single a-GalCer (Fig. 4B). We also analyzed the responses of FACS- injection of a-GalCer caused TCR downregulation of NKT cells sorted cells in vitro. BMDCs were cocultured with naive liver within 24 h and a decrease in the percentages of TCR+ NKT cells NKT cells treated with IDENs or PBS for 3 h and then stimulated recovered 3 d after the injection. This result is also consistent with with a-GalCer for an additional 24 h (Fig. 4C), or FACS-sorted what has been reported by others (24). In contrast, pretreatment with liver NKT cells were treated with IDENs or PBS for 3 h and sub- IDENs dampened the recovery of TCR+ NKT cells after a-GalCer sequently stimulated with a-GalCer tetramer (Fig. 4D). Analysis of injection (Fig. 3A) and reduced the a-GalCer–induced expression of the supernatants indicated that the a-GalCer–induced production by guest on September 29, 2021 the markers of early activation of NKT cells (i.e., CD69 and its of IFN-g and IL-4 by the NKT cells was significantly reduced when costimulator CD40L (Fig. 3B). Moreover, the a-GalCer induced IDENs were added to either NKT cells cocultured with DCs pulsed elevation of ALTand AST serum levels (Fig. 3C), and production of with a-GalCer (Fig. 4C) or purified NKT cells stimulated with IFN-g and IL-4 (Fig. 3D) by liver NKT cells was reduced sub- tetramer (Fig. 4D). stantially in mice that had been pretreated with IDENs, indicating a that IDENs can induce an anergic-like state in NKT cells exposed to IDEN-associated PGE2 inhibits -GalCer induction of inflammatory signals in vivo. NKT cell activation To directly test whether IDENs can induce an anergic-like state in We found that liver NKT cells stimulated in vitro with aGalCer in NKT cells to glycolipids, we purified NKT cells from the livers of the presence of lipids extracted from IDENs produced much less mice that had been administered IDENs or vehicle i.v. and cocul- IFN-g and IL-4 (Fig. 5A), suggesting that the lipid content of the tured NKT cells in vitro with DCs from the livers of untreated mice IDENs plays a role in the IDEN-induced effects on NKT cells. in the presence of a-GalCer. The results confirmed that the NKT Liposomes assembled from IDENs also have a similar effect as cells purified from the mice that had been administered IDENs IDENs and IDEN lipids (Supplemental Fig. 1), suggesting that the were hyporesponsive to a-GalCer stimulation (Fig. 4A). Consistent structure of liposomes assembled from IDEN lipids may not be with this, the addition of IDENs to liver NKT cells in culture essential for IDEN-mediated inhibition of activation of NKT cells

on the surface of IDENs or microparticles derived from mouse intestinal mucus (B, left panel). The presence of intestinal epithelial–specific marker (A33 and villin-1) and/or exosome markers (CD63, CD9, CD81, RAB5, RAB11, and HSP70) in the sucrose gradient–purified exosomes from the supernatants of 36- h cultured TS/A mouse breast tumor cells (51), mouse intestinal nanoparticles (A), mircoparticles (B), or human intestinal nanoparticles (C) was determined by Western blot analysis [(A)–(C), right panel]. The presence of Annexin V was determined by FACS analysis [(B), left panels]. The imaging and Western blot are representative of at least three experiments. (D) In vivo imaging of and quantitation of signals in C57BL/6 mice gavage-administered IDENs labeled with the infrared dye DiR (100 mg/mouse in 100 ml PBS) or nonlabeled IDENs in PBS, and (E) images of the washed intestines (cut longitudinally) and other organs obtained 48 h after administration of IDENs. (F) Confocal image analysis of frozen sections of the livers of mice administered IDENs labeled with PKH26 red dye. Sections stained for A33, an intestinal epithelial cell–specific marker (green), and nuclei stained with DAPI. Distribution of IDENs in the intestines of mice 12 h after gavage administration of IDENs labeled with DiR dye (100 mg in 100 ml of PBS) assessed by (G) confocal image analysis of the frozen sections of intestine with immunofluorescent staining for A33 (green) and CD31 (red); original magnification 340. (H) Confocal image analysis of the frozen sections of intestine with immunofluorescent staining for LYVE (green) (original magnification 360 [left panel]) with en- largement of indicated area shown in the right panel.(I) Accumulation of DiR dye-labeled IDENs (100 mg/mouse in 100 ml PBS) administered by gavage 24 h after treatment with Con A, DSS, or PBS representative scan. Data are the mean 6 SEM of three experiments (D) or representative of three experiments (E–I)(n = 10 mice/group). 3584 INTESTINAL-DERIVED EXOSOMES INDUCE LIVER NKT ANERGY

FIGURE 3. IDENs induce an anergic-like state in NKT cells. (A–D) C57BL/6 mice were injected i.v. with IDENs (100 mg/mouse in PBS) every 2 d for 14 d, and 4 h after the last injection of IDENs, the mice were injected with a-GalCer (5 mg/ml, i.v.). (A) Frequency of TCR-b+CD-1d tetramer+ cells (NKT cells) in mononuclear cell preparations from the liver at indicated time points after a-GalCer treatment. (B) Expression of CD40L and CD69 on the TCR-b+CD-1d tetramer+ cells after a-GalCer treatment for 24 h. (C and D) Levels of ALT and AST (C) and IFN-g and IL-4 (D) in the serum. Data are representative of three experiments (A, B) or are the mean 6 SEM of five experiments (C, D). *p , 0.05 and **p , 0.01 (Student t test). Downloaded from

stimulated by a-GalCer. Our data (38) and that of others (39) of the expression of the four known PGE2 receptor subtypes in- http://www.jimmunol.org/ indicate that PGE2 from various sources is enriched on exosomes, dicated selective upregulation of the EP2 and EP4 receptors on and we found that the IDENs carried significant amounts (.80 ng/ a-GalCer–stimulated liver NKT cells (Supplemental Fig. 2). This mg) of PGE2 (Fig. 5B). PGE2 is known to regulate activation suggested the possibility that the IDEN-associated PGE2 could through the modulation of expression of PGE2 receptors on T cells play a role in the induction of the NKT cell anergic-like state. (40). Whether the expression of PGE2 receptors on NKT cells are Next, we tested the effects the amount indomethacin (a cyclo- regulated by a-GalCer stimulation is not known. RT-PCR analysis oxygenase [COX] 2 inhibitor that blocks the generation of PGE2) would have on the association of PGE2 with IDENs. ELISA results indicated that indomethacin treatment reduced the amount by guest on September 29, 2021 of PGE2 associated with the IDENs (Fig. 5B). When comparing the IDENs isolated from the intestines of naive mice pretreated with vehicle only to the IDENs isolated from the intestines of indomethacin-treated mice, the IDENs from indomethacin- treated mice were less potent at inducing the anergic-like state of NKT cells exposed to a-GalCer stimulation (Fig. 5C). Indometh- acin treatment resulted in promoting Con A–induced mouse mor- tality (Fig. 5D), more severe hepatocellular necrosis (Fig. 5E), and higher sera levels of IFN-g and ALT (Fig. 5F). Our data (Fig. 6A) and the data published by others suggest that there was a selective decrease in liver NKT cells in DSS-induced colitis (41). Whether IDENs isolated from DSS-induced colitis have less capacity to inhibit live NKT cell activation is unknown. We found that in comparison with IDENs isolated from the intestines of naive mice, the IDENs isolated from the intestines of mice pretreated with 3% DSS in drinking water are less potent in inducing the IDEN- mediated anergic-like state of NKT cells to a-GalCer stimulation FIGURE 4. Ex vivo analysis of NKT cell responses to IDENs and (Fig. 6B). This reduction appears to be associated with a decreased a-GalCer. (A) Sorted NKT cells from mice treated with PBS or IDENs and amount of IDEN PGE2 (Fig. 6C). PGE2 is also known to modulate + CD11c DCs sorted from livers of naive mice were cocultured for 3 d in CD4 T cell activation, and our data indicated that IDENs or PGE2 the presence of a-GalCer. Cell division was assessed by labeling the treatment leads to a significant reduction of both IFN-g and IL-2 NKT cells with CFSE (5 mM). (B) Proliferation of NKT cells from naive produced by Con A–induced CD4 T cells activation (Fig. 6D), mice cultured with liver DCs in the presence of IDENs or PBS for 3h, and suggesting that IDEN PGE2 may play a role in the inhibition of then stimulated by a-GalCer. The proliferation of the NKT cells was de- 3 CD4 T cell activation, as well as in the Con A–induced mouse liver termined after a 16 h pulse with [ H]thymidine, and the cultured cells were inflammation model. subsequently harvested. (C) Cytokine production by liver NKT cells from Although the immunomodulatory effects of PGE are well naive mice cultured with BMDCs in the presence of IDENs or PBS for 3 h 2 and then stimulated by a-GalCer (100 ng/ml) for 24 h. (D) Cytokine known (42, 43), PGE2 has not been implicated previously in production by NKT cells from naive mice pretreated by IDENs or PBS for modulation of NKT cell activity. We found that addition of ex- 3 h and then stimulated by a-GalCer–tetramer for 24 h. Data are repre- ogenous PGE2 to monocytes isolated from the livers of naive mice sentative of three experiments (A) or are the mean 6 SEM of five (Fig. 7A) or to cocultures of mature BMDCs with NKT cells experiments (B–D). *p , 0.05, **p , 0.01 (Student t test). purified from the livers of naive mice (Fig. 7B) led to inhibition The Journal of Immunology 3585 Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 5. IDEN PGE2 induces an anergic-like state in NKT cells. (A) Levels of IFN-g and IL-4 in supernatants from sorted liver NKT cells cultured with BMDCs in the presence of IDENs or total lipids isolated from IDENs for 3 h and then stimulated by a-GalCer (100 ng/ml) for 24 h. (B) ELISA analysis of PGE2 levels in IDENs from naive or indomethacin-treated C57BL/6 mice. (C) Levels of IFN-g and IL-4 in supernatants of sorted liver NKT cells cultured with BMDCs in the presence of IDENs isolated from naive C57BL/6 mice or indomethacin-treated mice for 3 h and then stimulated by a-GalCer (100 ng/ml) for 24 h. (D–F) Exacerbation of liver injury by indomethacin in mice treated with Con A. C57BL/6 mice (n =5)weread- ministered indomethacin (4 mg/kg/day) orally in the drinking water for 1 wk. Con A was administered i.v. at 37.5 mg/kg body weight. (D) Effects of indomethacin pretreatment on survival in mice treated with Con A. Survival curves are depicted as time after Con A injection. (E) H&E-stained sections of livers from Con A–injected mice pretreated with PBS or indomethacin; original magnification 320. (F) Levels of IFN-g and ALT were assessed in serum at different time points in Con A–injected mice pretreated by PBS (PBS–Con A) or indomethacin (indomethacin–Con A). Data [(A)–(C), (D), (F)] are the mean 6 SEM of 3 experiments (n =7).(E) is representative of at least three independent experiments (n =7).*p , 0.05, **p , 0.01 (one-tailed unpaired Student t test). of a-GalCer–induced production of both IFN-g and IL-4 in The cAMP–PKA pathway plays an essential role in the a PGE2 dose-dependent manner. Similar results were obtained IDEN-associated PGE2-dependent induction of NKT cell anergy using the NKT1.2 NKT hybridoma (Supplemental Fig. 3). Addi- Both EP2 and EP4 are Gas protein–coupled receptors, which upon tion of either EP2- or EP4- selective agonists to cocultures of binding PGE2 act to increase the levels of intracellular cAMP (44). FACS-purified liver NKT cells with BMDCs in the presence of We found that addition of a dibutyryl cAMP analog (dB-cAMP), a-GalCer suppressed the secretion of IFN-g and IL-4 (Fig. 7C), aPKA-specificagonist(N6-Bnz-cAMP), or forskolin, a down- whereas agonists selective for EP1 and EP3 had no effect (data not stream activator of the cAMP-mediated pathway, suppressed the shown). Furthermore, the addition of either the EP4 antagonist, in vitro production of IFN-g and IL-4 by NKT cells (Fig. 8A). ONO-AE3-208 (Fig. 7D), or the EP2 antagonist, AH6809 (Fig. Furthermore, both the IDEN (Fig. 8B)- and PGE2 (Fig. 8C)-me- 7E), partially reversed the PGE2 or IDEN-mediated inhibition of diated suppression of IFN-g and IL-4 production was inhibited by IFN-g and IL-4 production, whereas EP1 and EP3 antagonists had H-89, a PKA inhibitor, or Rp-8-Br-Camps and Rp-8-CPT-cAMPS, no effect (data not shown). These findings lead to the novel type I and type II PKA inhibitors, respectively, confirming that the conclusion that PGE2 can inhibit the production of IFN-g and IL-4 IDEN-mediated inhibition of IFN-g and IL-4 production by in a-GalCer–stimulated NKT cells and that these effects are NKT cells requires PGE2-EP2/EP4 signaling and is dependent on mediated by the EP2 and EP4 receptors. the cAMP/PKA pathway. 3586 INTESTINAL-DERIVED EXOSOMES INDUCE LIVER NKT ANERGY

FIGURE 6. IDENs derived from DSS-in- duced colitis mouse model have significantly less suppression of liver NKT cell activation. (A) Frequency of TCR-b+NK1.1+ cells (NKT cells) in mononuclear cell preparations from the liver at day 8 after mice were treated with 3% DSS in drinking water. (B) Levels of IFN-g and IL-4 in supernatants of sorted liver NKT cells cultured with BMDCs in the presence of IDENs isolated from naive C57BL/6 mice or DSS-treated mice for 3 h, and then stimulated with a-GalCer (100 ng/ml) for 24 h. (C) ELISA analysis of PGE2 levels in IDENs from naive or DSS-treated C57BL/6 mice. Data are representative of three experiments (A) or are the mean 6 SEM of five experiments [(B) and (C)]. **p , 0.01 (Student t test). (D) Levels of IFN-g and IL-2 in super-

natants of sorted naive CD4 T cells in the pres- Downloaded from ence of IDENs or PGE2 for 3 h and then stimulated by Con A (20 mg/ml) for 24 h. Data are the mean 6 SEM of five experiments. **p , 0.01 (Student t test). http://www.jimmunol.org/

Discussion process involving the integration of numerous signals. The current In this study, we demonstrate that IDENs migrate to the liver where studies pinpoint a regulatory mechanism by which IDENs can in- they induce a NKT cell anergy-like state. We also establish for the fluence this balance through mechanisms that involve considerable first time, to our knowledge, that a PGE2-mediated pathway plays cross-talk between intestinal IDENs, PGE2 , and NKT cells. For a role in the induction of the anergic-like state in NKT cells, which example, in the current studies we found that inflammation induced can occur through a cAMP/PKA activity-dependent mechanism. by either Con A injection or drinking DSS water can alter the The rapid enhancement of the activity occurs through a traditional amount of PGE2 carried by the IDENs. Specifically, our data suggest by guest on September 29, 2021 positive feedback mechanism in which the PGE2-induced en- that although inflammation induced by Con A– or DSS-induced hancement of expression of the EP2 and EP4 receptors can amplify colitis resulted in increasing amounts of IDENs accumulating in PGE2 signaling in NKT cells. Notably, the activation of the PGE2 the liver and significant reduction of PGE2 carried by IDENs, which pathway also appears to create in the liver a microenvironment in made them less potent for inhibiting a-GalCer stimulated liver which naive NKT cells are resistant to activation by a-GalCer NKT cell activation. This suggests that different microenviron- stimulation, which is one of the major obstacles in the use of mental conditions may tailor the ability of the IDENs to either a-GalCer for the treatment of certain types of cancer. promote or suppress NKT cell activation. Although two acute in- It is a reasonable assumption that control of the balance between flammatory mouse models were used in this study, IDENs derived the anergic-like state and activation of NKT cells is a complex from chronic inflammatory driven disease models such as colon

FIGURE 7. IDEN PGE2 induces an anergic- like state in NKT cells via the EP2/EP4 signaling pathway. Levels of IFN-g and IL-4 in super- natants of liver leukocytes (A) or sorted hepatic NKT cells (B–E) cocultured with BMDCs that were pretreated for 3 h with the following agents and then stimulated with a-GalCer for 24 h. The

agents were diluted PGE2 (A, B) EP2 (10 mM butaprost), or EP4 (100 nM ONO-AE1-329)

agonist (C) or in combination with PGE2 or IDEM with/without EP4 (1 mM ONO-AE3-208) or EP2 (AH6809) antagonist (D, E). Data (A–E) are the mean 6 SEM of three experiments (n = 5). *p , 0.05, **p , 0.01 (one-tailed unpaired Student t test). The Journal of Immunology 3587

FIGURE 8. IDEN PGE2 induces an anergic-like state in NKT cells via EP2/EP4 mediating the cAMP–PKA pathway. Levels of IFN-g and IL-4 in supernatants of sorted hepatic NKT cells cocultured with BMDCs that were pretreated for 3 h with the following agents and then stimulated with a-GalCer 6 for 24 h. The agents were db-cAMP (100 mM), N -Bnz-cAMP (100 mM), forskolin (10 mM) (A), or in combination with IDENs (B) or PEG2 (C) with/ without H-89 (3 mM), Rp-8-Br-cAMPS (100 mM), or Rp-8-CPT-cAMPS (100 mM). ELISA of IFN-g and IL-4 in the cultured supernatants. Data (A–C) are the mean 6 SEM of four experiments (n = 5). *p , 0.05, **p , 0.01 (one-tailed unpaired Student t test). Downloaded from cancer may have different phenotypes, which we have not ad- A large number of glycolipids thatcan activate NKT cells have been dressed in this study. In addition, although the regulator that con- studied, whereas only a few of the NKT inhibitory lipids have been trols the amounts of PGE2 carried by IDENs has not been studied. Earlier studies demonstrated that neutral glycolipid, Gg3Cer identified, we are interested in determining whether IDENs prefer shed from a tumor cell line can inhibit CD1d1-mediated Ag pre- carrying PGE2 rather than other PGs regulated by Cox1 and Cox2. sentation to NKT cells (47). Di-palmitoyl-phosphatidyl-ethanolamine Furthermore, COX-1 is constitutively expressed in most tissues (48) covalently attached to polyethyleneglycol acts as an antagonist http://www.jimmunol.org/ including intestine. In contrast, COX-2 is induced in inflammation to inhibit the a-GalCer–induced phosphorylation of ERK tyrosine by various stimuli including cytokines and growth factors. Whether kinase in NKT cells, thus blocking NKT TCR signaling and activa- IDENs have a higher capacity in preferentially binding PGs regu- tion of NKT cells. However, under physiological conditions these lated by Cox1 than Cox2 needs to be addressed by comparing lipid complex are not detected in vivo, and mechanism(s) underlying lipidomic profiles of IDENs isolated from Cox1 and Cox2 how the anergic-like state in liver NKT cells is maintained is not knockout mice. known. Data presented in this study suggest that intestinal mucus We also speculate that exosome or exosome-like nanoparticle derived PGE2 carried by exosome-like nanoparticle-mediated can recipient cells may also interact in a coordinated manner with the induce the NKT cell anergy-like state in mouse models. This finding exosome producing cells. For an example, when the host micro- opens up a new avenue to further study whether other intestinal- by guest on September 29, 2021 environment is under “fire” such as in infection or when there is an related Ags carried by intestinal mucus derived exosome-like nano- accumulation of damaged cells, inflammation is required for re- particles play a role in induction of immune tolerance. moving these cells. In this case, immune suppressors such as PGE2 Although the role of IDEN associated PGE2 in the induction of carried by exosomes or exosome-like nanoparticles must be re- the NKT cell anergic-like state are addressed in this study, we duced for the receipt cells to be activated. Coordinated interactions speculate that this finding may also be applied to other types of of exosome or exosome-like nanoparticle producing cells with cells that express PGE2 receptors. A large body of evidence sug- recipient cells may cause a subsequent release of particles pre- gests that PGE2 receptors are expressed on not only NKT, but also dominantly packed with proinflammatory molecules that further CD4 T cells (49) and other types of cells (50) as well. However, if amplify inflammatory responses via paracrine and autocrine path- those cells do not have the in vivo capacity of taking up IDENs, ways until the infected or damaged cells are removed. As a result, PGE2 carried by IDENs may not have an effect on these cells even the newly achieved balance between proinflammatory and anti- if these cells are PGE2 receptor positive. inflammatory factors may be more related to the ratio of proin- In summary, to our knowledge, this is the first study to demon- flammatory and anti-inflammatory factors to each other than ab- strate a PGE2 role in inducing NKT cell anergy. Our data support solute quantities of each. Therefore, it is conceivable that the data a model in which the PGE2 carried by IDENs is taken up by Ag presented in Fig. 2D, where the sera levels of TNF-a and IFN-g are presenting cells or NKT cells of the liver and then initiates a unique higher than those from naive mice, that the overall balance or ratio suppressive mechanism. This mechanism does not require recog- of proinflammatory and anti-inflammatory factors is achieved such nition of lipid Ags. In addition, the IDENs work as a functional that liver damage because of the Con A injection was prevented. bridge between the intestinal and liver immune systems. We pro- Dysregulation of this balance may contribute to the development of pose that this mechanism of suppression is a major pathway that disease. This assumption is supported by recent results reported by counterbalances intestinal lipid-mediated activation of NKT cells another groups. Kornek et al. (45) showed that activated T cells in the liver. released microparticles and that the levels of released micro- particles correlated with the histological severity of patients with Acknowledgments active hepatitis C. Another study (46) indicated that microparticles We thank the National Institutes of Health Tetramer Facility for providing + derived from CD14 monocytes and NKT cells are critical to ad- PBS-57 ligand complexed to CD1d monomers or tetramers; Ono Pharma- ipose liver inflammation of patients with NAFL/NASH. Further ceutical for providing EP4 (ONO-AE1-329), EP1 antagonist ONO-8713, research is required to determine whether IDEN-mediated inhibi- EP3 antagonist ONO-AE3-240, and EP4 antagonist (ONO-AE3-208); tion of liver NKT activation also has an inhibiting effect on the and Dr. Mitchell Kronenberg, who provided the NKT1.2 hybridoma. We releasing of proinflammatory microparticles from NKT. also thank Drs. Fiona Hunter and Jerald Ainsworth for editorial assistance. 3588 INTESTINAL-DERIVED EXOSOMES INDUCE LIVER NKT ANERGY

Disclosures cells requires type I interferon and charged glycosphingolipids. Immunity 27: 597–609. The authors have no financial conflicts of interest. 24. Kim, S., S. Lalani, V. V. Parekh, T. L. Vincent, L. Wu, and L. Van Kaer. 2008. Impact of bacteria on the phenotype, functions, and therapeutic activities of invariant NKT cells in mice. J. Clin. Invest. 118: 2301–2315. References 25. Xiang, X., X. Zhuang, S. Ju, S. Zhang, H. Jiang, J. Mu, L. Zhang, D. Miller, 1. Crispe, I. N. 2009. The liver as a lymphoid organ. Annu. Rev. Immunol. 27: 147–163. W. Grizzle, and H. G. Zhang. 2011. miR-155 promotes macroscopic tumor 2. Adams, D. H., and B. Eksteen. 2006. Aberrant homing of mucosal T cells and formation yet inhibits tumor dissemination from mammary fat pads to the lung extra-intestinal manifestations of inflammatory bowel disease. Nat. Rev. Immu- by preventing EMT. Oncogene 30: 3440–3453. nol. 6: 244–251. 26. Bligh, E. G., and W. J. Dyer. 1959. A rapid method of total lipid extraction and 3. Al-Lahham, S. H., M. P. Peppelenbosch, H. Roelofsen, R. J. Vonk, and purification. Can. J. Biochem. Physiol. 37: 911–917. K. Venema. 2010. Biological effects of propionic acid in ; metabolism, 27. Yu, S., C. Liu, K. Su, J. Wang, Y. Liu, L. Zhang, C. Li, Y. Cong, R. Kimberly, potential applications and underlying mechanisms. Biochim. Biophys. Acta 1801: W. E. Grizzle, et al. 2007. Tumor exosomes inhibit differentiation of bone 1175–1183. marrow dendritic cells. J. Immunol. 178: 6867–6875. 4.Seki,S.,Y.Habu,T.Kawamura,K.Takeda,H.Dobashi,T.Ohkawa,and 28. Laulagnier, K., C. Motta, S. Hamdi, S. Roy, F. Fauvelle, J. F. Pageaux, H. Hiraide. 2000. The liver as a crucial organ in the first line of host defense: T. Kobayashi, J. P. Salles, B. Perret, C. Bonnerot, and M. Record. 2004. Mast the roles of Kupffer cells, natural killer (NK) cells and NK1.1 Ag+ T cells in cell- and dendritic cell-derived exosomes display a specific lipid composition T helper 1 immune responses. Immunol. Rev. 174: 35–46. and an unusual membrane organization. Biochem. J. 380: 161–171. 5. Pellicci, D. G., A. J. Clarke, O. Patel, T. Mallevaey, T. Beddoe, J. Le Nours, 29. Wang, G. J., Y. Liu, A. Qin, S. V. Shah, Z. B. Deng, X. Xiang, Z. Cheng, C. Liu, A. P. Uldrich, J. McCluskey, G. S. Besra, S. A. Porcelli, et al. 2011. Recognition J. Wang, L. Zhang, et al. 2008. Thymus exosomes-like particles induce regu- of b-linked self glycolipids mediated by receptors. latory T cells. J. Immunol. 181: 5242–5248. Nat. Immunol. 12: 827–833. 30. Xiang, X., Y. Liu, X. Zhuang, S. Zhang, S. Michalek, D. D. Taylor, W. Grizzle, 6. Kinjo, Y., D. Wu, G. Kim, G. W. Xing, M. A. Poles, D. D. Ho, M. Tsuji, and H. G. Zhang. 2010. TLR2-mediated expansion of MDSCs is dependent on K. Kawahara, C. H. Wong, and M. Kronenberg. 2005. Recognition of bacterial the source of tumor exosomes. Am. J. Pathol. 177: 1606–1610.

glycosphingolipids by natural killer T cells. Nature 434: 520–525. 31. Johnstone, C. N., N. C. Tebbutt, H. E. Abud, S. J. White, K. L. Stenvers, Downloaded from 7. Subleski, J. J., V. L. Hall, T. B. Wolfe, A. J. Scarzello, J. M. Weiss, T. Chan, N. E. Hall, S. H. Cody, R. H. Whitehead, B. Catimel, E. C. Nice, et al. 2000. D. L. Hodge, T. C. Back, J. R. Ortaldo, and R. H. Wiltrout. 2011. TCR- Characterization of mouse A33 antigen, a definitive marker for basolateral dependent and -independent activation underlie liver-specific regulation of surfaces of intestinal epithelial cells. Am. J. Physiol. Gastrointest. Liver Physiol. NKT cells. J. Immunol. 186: 838–847. 279: G500–G510. 8. Wang, X., K. A. Bishop, S. Hegde, L. A. Rodenkirch, J. W. Pike, and 32. van Genderen, H. O., H. Kenis, L. Hofstra, J. Narula, and C. P. Reutelingsperger. J. E. Gumperz. 2012. Human invariant natural killer T cells acquire transient 2008. Extracellular annexin A5: functions of phosphatidylserine-binding and innate responsiveness via histone H4 acetylation induced by weak TCR stimu- two-dimensional crystallization. Biochim. Biophys. Acta 1783: 953–963. lation. J. Exp. Med. 209: 987–1000. 33. Morel, O., N. Morel, L. Jesel, J. M. Freyssinet, and F. Toti. 2011. Microparticles: http://www.jimmunol.org/ 9. Parekh, V. V., M. T. Wilson, D. Olivares-Villago´mez, A. K. Singh, L. Wu, a critical component in the nexus between inflammation, immunity, and throm- C. R. Wang, S. Joyce, and L. Van Kaer. 2005. Glycolipid antigen induces long- bosis. Semin. Immunopathol. 33: 469–486. term natural killer T cell anergy in mice. J. Clin. Invest. 115: 2572–2583. 34. Wang, Y. Y., S. K. Lai, C. So, C. Schneider, R. Cone, and J. Hanes. 2011. 10. Gyo¨rgy,B.,T.G.Szabo´,M.Pa´szto´i, Z. Pa´l, P. Misja´k, B. Aradi, V. La´szlo´, Mucoadhesive nanoparticles may disrupt the protective human mucus barrier by E. Pa´llinger, E. Pap, A. Kittel, et al. 2011. Membrane vesicles, current altering its microstructure. PLoS ONE 6: e21547. state-of-the-art: emerging role of extracellular vesicles. Cell. Mol. Life Sci. 68: 35. Lai, S. K., Y. Y. Wang, and J. Hanes. 2009. Mucus-penetrating nanoparticles 2667–2688. for drug and delivery to mucosal tissues. Adv. Drug Deliv. Rev. 61: 158– 11. The´ry, C., M. Ostrowski, and E. Segura. 2009. Membrane vesicles as conveyors 171. of immune responses. Nat. Rev. Immunol. 9: 581–593. 36. Arshad, M. I., M. Rauch, A. L’helgoualc’h, V. Julia, M. C. Leite-de-Moraes, 12. Van Niel, G., J. Mallegol, C. Bevilacqua, C. Candalh, S. Brugie`re, E. Tomaskovic- C. Lucas-Clerc, C. Piquet-Pellorce, and M. Samson. 2011. NKT cells are re- Crook, J. K. Heath, N. Cerf-Bensussan, and M. Heyman. 2003. Intestinal epi- quired to induce high IL-33 expression in hepatocytes during ConA-induced

thelial exosomes carry MHC class II/peptides able to inform the immune system acute hepatitis. Eur. J. Immunol. 41: 2341–2348. by guest on September 29, 2021 in mice. Gut 52: 1690–1697. 37. Miyagi, T., T. Takehara, T. Tatsumi, T. Suzuki, M. Jinushi, Y. Kanazawa, 13. Mallegol, J., G. Van Niel, C. Lebreton, Y. Lepelletier, C. Candalh, C. Dugave, N. Hiramatsu, T. Kanto, S. Tsuji, M. Hori, and N. Hayashi. 2004. Concanavalin J. K. Heath, G. Raposo, N. Cerf-Bensussan, and M. Heyman. 2007. T84- a injection activates intrahepatic innate immune cells to provoke an antitumor intestinal epithelial exosomes bear MHC class II/peptide complexes po- effect in murine liver. Hepatology 40: 1190–1196. tentiating antigen presentation by dendritic cells. Gastroenterology 132: 1866– 38. Xiang, X., A. Poliakov, C. Liu, Y. Liu, Z. B. Deng, J. Wang, Z. Cheng, 1876. S. V. Shah, G. J. Wang, L. Zhang, et al. 2009. Induction of myeloid-derived 14. van Niel, G., G. Raposo, C. Candalh, M. Boussac, R. Hershberg, N. Cerf- suppressor cells by tumor exosomes. Int. J. Cancer 124: 2621–2633. Bensussan, and M. Heyman. 2001. Intestinal epithelial cells secrete exosome- 39. Subra, C., D. Grand, K. Laulagnier, A. Stella, G. Lambeau, M. Paillasse, P. De like vesicles. Gastroenterology 121: 337–349. Medina, B. Monsarrat, B. Perret, S. Silvente-Poirot, et al. 2010. Exosomes ac- 15. Chen, X., C. H. Song, B. S. Feng, T. L. Li, P. Li, P. Y. Zheng, X. M. Chen, count for vesicle-mediated transcellular transport of activatable phospholipases Z. Xing, and P. C. Yang. 2011. Intestinal epithelial cell-derived integrin ab6 and prostaglandins. J. Lipid Res. 51: 2105–2120. plays an important role in the induction of regulatory T cells and inhibits an 40. Kabashima, K., T. Saji, T. Murata, M. Nagamachi, T. Matsuoka, E. Segi, antigen-specific Th2 response. J. Leukoc. Biol. 90: 751–759. K. Tsuboi, Y. Sugimoto, T. Kobayashi, Y. Miyachi, et al. 2002. The prosta- 16. Deng, Z. B., X. Zhuang, S. Ju, X. Xiang, J. Mu, Q. Wang, J. Hong, L. Zhang, glandin receptor EP4 suppresses colitis, mucosal damage and CD4 cell activa- M. Kronenberg, J. Yan, et al. Intestinal mucus-derived nanoparticles mediate tion in the gut. J. Clin. Invest. 109: 883–893. activation of Wnt/b-catenin signaling plays a role ininduction of liver NKT cell 41. Yoshioka, K., Y. Ueno, S. Tanaka, K. Nagai, T. Onitake, R. Hanaoka, anergy. Hepatology In press. H. Watanabe, and K. Chayama. 2012. Role of natural killer T cells in the mouse 17. Zhuang, X., X. Xiang, W. Grizzle, D. Sun, S. Zhang, R. C. Axtell, S. Ju, J. Mu, colitis-associated colon cancer model. Scand. J. Immunol. 75: 16–26. L. Zhang, L. Steinman, et al. 2011. Treatment of brain inflammatory diseases by 42. Yao, C., D. Sakata, Y. Esaki, Y. Li, T. Matsuoka, K. Kuroiwa, Y. Sugimoto, and delivering exosome encapsulated anti-inflammatory drugs from the nasal region S. Narumiya. 2009. Prostaglandin E2-EP4 signaling promotes immune inflam- to the brain. Mol. Ther. 19: 1769–1779. mation through Th1 cell differentiation and Th17 cell expansion. Nat. Med. 15: 18. Spilman, M. S., C. Welbon, E. Nelson, and T. Dokland. 2009. Cryo-electron 633–640. tomography of porcine reproductive and respiratory syndrome virus: organiza- 43. Nataraj, C., D. W. Thomas, S. L. Tilley, M. T. Nguyen, R. Mannon, B. H. Koller, tion of the nucleocapsid. J. Gen. Virol. 90: 527–535. and T. M. Coffman. 2001. Receptors for prostaglandin E(2) that regulate cellular 19. Sun, D., X. Zhuang, X. Xiang, Y. Liu, S. Zhang, C. Liu, S. Barnes, W. Grizzle, immune responses in the mouse. J. Clin. Invest. 108: 1229–1235. D. Miller, and H. G. Zhang. 2010. A novel nanoparticle drug delivery system: the 44. Liu, X., A. J. Lazenby, and G. P. Siegal. 2006. Signal transduction cross-talk anti-inflammatory activity of curcumin is enhanced when encapsulated in exo- during colorectal tumorigenesis. Adv. Anat. Pathol. 13: 270–274. somes. Mol. Ther. 18: 1606–1614. 45. Kornek, M., Y. Popov, T. A. Libermann, N. H. Afdhal, and D. Schuppan. 20. Halder, R. C., C. Aguilera, I. Maricic, and V. Kumar. 2007. Type II NKT cell- 2011. Human T cell microparticles circulate in blood of hepatitis patients mediated anergy induction in type I NKT cells prevents inflammatory liver and induce fibrolytic activation of hepatic stellate cells. Hepatology 53: disease. J. Clin. Invest. 117: 2302–2312. 230–242. 21. Deng, Z. B., Y. Liu, C. Liu, X. Xiang, J. Wang, Z. Cheng, S. V. Shah, S. Zhang, 46. Kornek, M., M. Lynch, S. H. Mehta, M. Lai, M. Exley, N. H. Afdhal, and L. Zhang, X. Zhuang, et al. 2009. Immature myeloid cells induced by a high-fat D. Schuppan. 2012. Circulating microparticles as disease-specific biomarkers of diet contribute to liver inflammation. Hepatology 50: 1412–1420. severity of inflammation in patients with hepatitis C or nonalcoholic steatohe- 22. Deng, Z. B., A. Poliakov, R. W. Hardy, R. Clements, C. Liu, Y. Liu, J. Wang, patitis. Gastroenterology 143: 448–458. X. Xiang, S. Zhang, X. Zhuang, et al. 2009. Adipose tissue exosome-like vesi- 47. Sriram, V., S. Cho, P. Li, P. W. O’Donnell, C. Dunn, K. Hayakawa, J. S. Blum, cles mediate activation of macrophage-induced insulin resistance. Diabetes 58: and R. R. Brutkiewicz. 2002. Inhibition of glycolipid shedding rescues recog- 2498–2505. nition of a CD1+ T cell lymphoma by natural killer T (NKT) cells. Proc. Natl. 23. Paget, C., T. Mallevaey, A. O. Speak, D. Torres, J. Fontaine, K. C. Sheehan, Acad. Sci. USA 99: 8197–8202. M. Capron, B. Ryffel, C. Faveeuw, M. Leite de Moraes, et al. 2007. Acti- 48. Lombardi, V., P. Stock, A. K. Singh, J. Kerzerho, W. Yang, B. A. Sullivan, X. Li, vation of invariant NKT cells by toll-like receptor 9-stimulated dendritic T. Shiratsuchi, N. E. Hnatiuk, A. R. Howell, et al. 2010. A CD1d-dependent The Journal of Immunology 3589

antagonist inhibits the activation of invariant NKT cells and prevents develop- 50. Kalinski, P. 2012. Regulation of immune responses by prostaglandin E2. J. ment of allergen-induced airway hyperreactivity. J. Immunol. 184: 2107–2115. Immunol. 188: 21–28. 49. Valdez, P. A., P. J. Vithayathil, B. M. Janelsins, A. L. Shaffer, P. R. Williamson, 51. Liu, C., S. Yu, K. Zinn, J. Wang, L. Zhang, Y. Jia, J. C. Kappes, S. Barnes, and S. K. Datta. 2012. Prostaglandin E2 suppresses antifungal immunity by R. P. Kimberly, W. E. Grizzle, and H. G. Zhang. 2006. Murine mammary car- inhibiting interferon regulatory factor 4 function and interleukin-17 expression cinoma exosomes promote tumor growth by suppression of NK cell function. J. in T cells. Immunity 36: 668–679. Immunol. 176: 1375–1385. Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021 The Journal of Immunology

Corrections

Deng, Z.-B., X. Zhuang, S. Ju, X. Xiang, J. Mu, Y. Liu, H. Jiang, L. Zhang, J. Mobley, C. McClain, W. Feng, W. Grizzle, J. Yan, D. Miller, M. Kronenberg, and H.-G. Zhang. 2013. Exosome-like nanoparticles from intestinal mucosal cells carry prostaglandin E2 and suppress activation of liver NKT cells. J. Immunol. 190: 3579–3589.

A source of funding was omitted in this article. The funding information footnote should read “This work was supported by National Institutes of Health Grants RO1CA137037, R01CA107181, RO1AT004294, and RO1CA116092; Robley Rex VA Medical Center merit review grants (to H.-G.Z.); and a grant from the Susan G. Komen Breast Cancer Foundation. H.-G.Z. is supported by a Veterans Affairs Research Career Scientist Award.” www.jimmunol.org/cgi/doi/10.4049/jimmunol.1600479

Copyright Ó 2016 by The American Association of Immunologists, Inc. 0022-1767/16/$30.00