INNATE IMMUNITY

THE JOURNAL OF EXPERIMENTAL MEDICINE SELECTED ARTICLES APRIL 2015 www.jem.org

World-Class Quality | Superior Customer Support | Outstanding Value

Toll-Free Tel: (US & Canada): 1.877.BIOLEGEND (246.5343)Tel: 858.768.5800

biolegend.com

08-0050-01

BioLegend is ISO 9001:2008 and ISO 13485:2003 Certifi ed

BioLegend has a large variety of recombinant cytokines, chemokines, growth factors, and proteinases (e.g., MMPs, Granzymes A and B). The high quality of our recombinant proteins anchor your research, and the exceptional pricing allows you to focus on reaching those lofty research goals.

To learn more, visit: biolegend.com/recombinant_proteins

Recombinant Mouse FLT3LMouse FLT3L induces IL-6 production in murine myeloid leukemia cell line M1.

RFU

/sec

hMMP-9 (ng)

Recombinant Human MMP-9The activity of monomeric hMMP-9 was measured with 10 M of fl uorogenic MMP substrate, Mca-PLGL-Dpa-AR-NH2 , in the presence of 2.5, 5.0, 10, 20, 40, and 80 ng of activated hMMP-9. The activity of activated hMMP-9 is >800 pmole/min/g.

The Pinnacle of DiscoveryRecombinant Proteins for Bioassay

Diverse, with over 400 human, mouse, and rat proteins.

Exceptionally priced, to meet even the tightest budget.

Functionally tested and validated for bioassay.

Available in carrier-free and animal-free formats to suit your assay needs.

BioLegend Recombinant Proteins are:

BioLegendCompetitor

OD

(450

- 57

0 nm

)

ng/mL

WelcomeInnate Immunity

The Journal of Experimental Medicine now prints topic-specific mini collections to showcase a handful of our recent publications. In this installment, we highlight papers focusing on two innate cell subsets, dendritic cells and macrophages in health and disease.

Immunotherapy has proven successful in many types of cancer treatment, but has also been associated with dangerous inflammatory responses in some patients. Our collection begins with an insight by Merghoub and Wolchok discussing the findings of Mirsoian et al. who show that lethal inflammation in response to anti-CD40 and IL-2 immunotherapy is triggered by increased inflammatory macrophages in the accumulated visceral fat of aged mice. Similarly, young obese mice succumbed to treatment while older mice on diets were protected. The study highlights factors to be considered with use of immunotherapeutics.

In the gut, the immune system must balance recognition of infectious pathogens with minimal responses to commensal bacteria. An Insight by Giorgio Trinchieri describes these immune challenges and the findings from Longman et al. showing that in both mice following Citrobacter rodentium infection and in patients with colitis, CX3CR1+ mononuclear phagocytes (MNPs) are potent producers of IL-23 and IL-1b. The production of these cytokines by MNPs efficiently induces IL-22 production by group 3 innate lymphoid cells to promote barrier integrity and mucosal protection.

Lambrecht and Guilliams examine genetic fate mapping of myeloid cells that distinguishes macrophages derived from embryonic precursors vs. circulating monocytes. They highlight the contribution to macrophage biology from Molowi et al. reporting that cardiac macrophages are initially of embryonic origin, but are progressively replaced by monocyte-derived cells as mice age. They speculate about how manipulation of different macrophage populations could be used in therapeutics.

Myelin destruction in multiple sclerosis (MS) is mediated by inflammatory macrophages, but the origin of these cells has been unclear. An Insight from Michael Heneka discusses findings from Yamasaki et al., who use a mouse model of MS to distinguish tissue-resident microglial cells from infiltrating monocytes. Using double chemokine reporter mice, the authors find that monocyte-derived macrophages initiate myelin destruction, mainly in the nodes of Ranvier, while microglia-derived macrophages are involved with clearing debris.

Two papers from the Liu and Nussenzweig groups track the differentiation of human progenitor cells into dendritic cells (DCs). They show that a granulocyte/monocyte/DC progenitor gives rise to a monocyte-DC progenitor that in turn generates both monocytes and a common DC progenitor. The common DC progenitor produces the three major subsets of human DCs, such as CD1c+ cDCs, CD141+ cDCs and pDCs. They go on to identify the immediate precursor of CD1c+ and CD141+ DCs in the circulation of healthy donors. These precursor cells (hpre-cDC) were detectable in cord blood, bone marrow, blood, and peripheral lymphoid organs. An Insight by Frederic Geissmann describes the novel in vitro culture system used in the studies to parse out the development of human DCs as well as a discussion of the developmental similarities between mouse and human DCs.

Together these studies characterize the many roles and functions of innate immune cells. We hope you enjoy this complimentary copy of our Innate Immunity collection. We invite you to explore additional collections at www.jem.org and to follow JEM on Facebook, Google+, and Twitter.

Immunotherapy and the belly of the beastTaha Merghoub and Jedd D. Wolchok

Adiposity induces lethal cytokine storm after systemic administration of stimulatory immunotherapy regimens in aged miceAnnie Mirsoian, Myriam N. Bouchlaka, Gail D. Sckisel, Mingyi Chen, Chien-Chun Steven Pai, Emanuel Maverakis, Richard G. Spencer, Kenneth W. Fishbein, Sana Siddiqui, Arta M. Monjazeb, Bronwen Martin, Stuart Maudsley, Charles Hesdorffer, Luigi Ferrucci, Dan L. Longo, Bruce R. Blazar, Robert H. Wiltrout, Dennis D. Taub, and William J. Murphy

Selected Articles April 2015

Critical role for CX3CR1+ mononuclear phagocytes in intestinal homeostasisGiorgio Trinchieri

CX3CR1+ mononuclear phagocytes support colitis-associated innate lymphoid cell production of IL-22Randy S. Longman, Gretchen E. Diehl,1 Daniel A. Victorio, Jun R. Huh,1 Carolina Galan, Emily R. Miraldi, Arun Swaminath, Richard Bonneau, Ellen J. Scherl, and Dan R. Littman

Monocytes find a new place to dwell in the niche of heartbreak hotelBart Lambrecht and Martin Guilliams

Progressive replacement of embryo-derived cardiac macrophages with ageKaaweh Molawi, Yochai Wolf, Prashanth K. Kandalla, Jeremy Favret, Nora Hagemeyer, Kathrin Frenzel, Alexander R. Pinto, Kay Klapproth, Sandrine Henri, Bernard Malissen, Hans-Reimer Rodewald, Nadia A. Rosenthal, Marc Bajenoff, Marco Prinz, Steffen Jung, and Michael H. Sieweke

Macrophages derived from infiltrating monocytes mediate autoimmune myelin destructionMichael T. Heneka

Differential roles of microglia and monocytes in the inflamed central nervous systemRyo Yamasaki, Haiyan Lu, Oleg Butovsky, Nobuhiko Ohno, Anna M. Rietsch, Ron Cialic, Pauline M. Wu, Camille E. Doykan, Jessica Lin, Anne C. Cotleur, Grahame Kidd, Musab M. Zorlu, Nathan Sun, Weiwei Hu, LiPing Liu, Jar-Chi Lee, Sarah E. Taylor, Lindsey Uehlein, Debra Dixon, Jinyu Gu, Crina M. Floruta, Min Zhu, Israel F. Charo, Howard L. Weiner, and Richard M. Ransohoff

The real thing: How to make human DC subsetsFrederic Geissmann

Restricted dendritic cell and monocyte progenitors in human cord blood and bone marrowJaeyop Lee, Galle Breton, Thiago Yukio Kikuchi Oliveira, Yu Jerry Zhou, Arafat Aljoufi, Sarah Puhr, Mark J. Cameron, Rafick-Pierre Skaly, Michel C. Nussenzweig, and Kang Liu

Circulating precursors of human CD1c+ and CD141+ dendritic cellsGalle Breton, Jaeyop Lee, Yu Jerry Zhou, Joseph J. Schreiber, Tibor Keler, Sarah Puhr, Niroshana Anandasabapathy, Sarah Schlesinger, Marina Caskey, Kang Liu, and Michel C. Nussenzweig

THE JOURNAL OF EXPERIMENTAL MEDICINEExecutive Editor Marlowe S. Tessmerphone (212) 327-8575 email: jem@rockefeller.edu

Senior Editors Teodoro PulvirentiHeather L. Van Epps

Scientific EditorCatarina Sacristn

Editors Jean-Laurent CasanovaDouglas T. FearonDavid HoltzmanLewis L. LanierWilliam A. MullerCarl NathanMichel NussenzweigAnne OGarraAlexander RudenskyAlan SherSasha TarakhovskyAndreas TrumppDavid Tuveson

Editor Emeritus Alan N. Houghton

Manuscript Coordinator Sylvia F. Cuadradophone (212) 327-8575 email: jem@rockefeller.edu

Preflight Editor Rochelle Ritacco

Assistant Production Editor Shauna OGarro

Production Editor Brianna Caszatt

Production Manager Camille Clowery

Production Designer Erinn A. Grady

Advisory EditorsShizuo AkiraKari AlitaloFrederick W. AltK. Frank AustenAlbert BendelacMichael J. BevanChristine A. BironChristian BogdanHal E. BroxmeyerMeinrad BusslingerArturo CasadevallAjay ChawlaYongwon ChoiRobert L. CoffmanDaniel J. CuaMyron I. CybulskyRiccardo Dalla FaveraGlenn DranoffMichael DustinVincent A. FischettiRichard A. FlavellAdolfo Garcia-SastrePatricia GearhartRonald N. GermainChristopher GoodnowSiamon GordonOr GozaniSergio GrinsteinPhilippe GrosKristian HelinChyi HsiehChristopher A. HunterKayo InabaGerard KarsentyJay KollsPaul KubesVijay K. KuchrooRalf Kuppers

Tomohiro KurosakiBart N. LambrechtKlaus F. LeyYong-Jun LiuClare LloydTak MakBernard MalissenJames S. MalterPhilippa MarrackDiane MathisIra MellmanMatthias MerkenschlagerSean J. MorrisonMuriel MoserChristian MnzCornelis MurreBenjamin G. NeelMichael NeubergerVictor NussenzweigJohn J. OSheaPaul H. PattersonFiona PowrieLluis Quintana-MurciKlaus RajewskyGwendalyn J. RandolphJeffrey RavetchSergio RomagnaniNikolaus RomaniDavid L. SacksShimon SakaguchiMatthew D. ScharffOlaf SchneewindStephen P. SchoenbergerHans SchreiberGerold SchulerRobert A. SederRafick-P. SkalyCharles N. SerhanNilabh Shastri

Ethan M. ShevachRoy L. SilversteinJonathan SprentJanet StavnezerAndreas StrasserStuart TangyeSteven L. TeitelbaumThomas J. TempletonKevin J. TraceyGiorgio TrinchieriShannon TurleyMarcel R.M. van den BrinkUlrich von AndrianHarald von BoehmerChristopher M. WalkerRaymond M. WelshE. John WherryLinda S. WickerIan WilsonThomas Wynn

Monitoring EditorsMarco ColonnaJason CysterStephen HedrickKristin A. HogquistAndrew McMichaelLuigi NotarangeloAnjana RaoFederica SallustoLouis M. StaudtToshio Suda

ConsultingBiostatistics EditorsGlenn HellerMadhu Mazumdar

Print ISSN 0022-1007 Online ISSN 1540-9538

Copyright to articles published in this journal is held by the authors. Articles are published by The Rockefeller University Press under license from the authors. Conditions for reuse of the articles by third parties are listed at http://www.rupress.org/terms

INSIGHTS

2 INSIGHTS | The Journal of Experimental Medicine

Immunotherapy has emerged as an effective means to restore immune recognition of cancer. Numerous forms of immunotherapy are currently being explored clinically. The most common are vaccines (dendritic cell, viral, and whole tumor cell based), adoptive T cell therapy and immune checkpoint blockade. The recent successes in melanoma, renal cancer, and lung cancer have gener-ated renewed optimism for treatment of multiple cancer types that were believed not amenable to immune-based therapies. However, immune-related events such as colitis, dermatitis, and, less frequently, endocrinopathy and pneumonitis have been reported and can be a challenge in the clinical use of these approaches. Cytokine release syndrome has been reported in the case of CAR (chimeric antigen receptor) T cell therapy and treatment with agonist antibodies. In this issue, Mirsoian et al. provide evidence that adiposity in aged mice induces a lethal cytokine storm follow-ing systemic administration of stimulatory immunotherapy consisting of anti-CD40 agonist antibody with IL-2.

This is an elegant follow up to a previous study by the same authors in which they showed that systemic immunotherapy adminis-tration in aged mice resulted in the induction of a rapid and lethal cytokine storm. In the current paper, they find that it is the fat that accumulates during aging, and mainly the visceral fat, that is associated with toxicity. This results in increased levels of proinflammatory

M1 macrophages within the peritoneal cavity and visceral adi-pose tissues, leading to heightened production of TNF. In or-der to confirm that it is indeed the fat that is associated with lethal effects, the authors repeated their studies in young mice with genetic (ob/ob) or diet-induced obesity (DIO). While young obese mice also displayed severe toxicity to the therapy, it was much more severe in older mice suggesting that age is also a significant factor. Reciprocally, calorie-restricted aged mice had lower visceral body fat content and reduced cytokine levels, with increased survival following immunotherapy. Obese mice were also protected by macrophage depletion or TNF blockade. These data demonstrate the need to consider age and body fat content as variables in preclinical assessment of therapeutics and when modeling diseases such as cancer.

Since many cancer patients fall into the elderly category, this study brings up a critical point that aged mice respond dif-ferently to immunotherapy than younger mice. This also highlights an important issue when considering potential tox-icities associated with any therapy, as most animal studies are performed in young, healthy mice. Another interesting point is the potential impact of body mass index and adipose accu-mulation in predicting side effects when investigating and uti-lizing immunotherapies. This is a very important and impactful finding for the field, and further investigations in tumor bear-ing mice with commonly used therapies are warranted. If the findings are confirmed in other therapeutic settings, the inclu-sion of supportive measures, such as TNF blockade, should continue to be considered for early management of immune-related toxicities to improve clinical outcomes.Mirsoian, A., et al. 2014. J. Exp. Med. http://dx.doi.org/10.1084/jem.20140116.

Immunotherapy and the belly of the beast

Taha Merghoub and Jedd D. Wolchok, Memorial Sloan Kettering Cancer Center: merghout@mskcc.org and wolchokj@mskcc.org

Insight from Taha Merghoub (left) and Jedd Wolchok (right)

The Murphy group previously showed that immunotherapy with anti-CD40 and IL-2 leads to a productive immune response in young mice (A) but lethal cytokine storm in older mice (B). They now find that the visceral fat that accumulates in aged mice (or young obese mice) is the primary trigger of inflammation after immunotherapy, resulting in increased inflammatory M1 macrophages and toxic cytokine storm (B). Immunotherapy-induced lethality was prevented by blocking TNF or depleting macrophages (B).Whether the success of immunotherapy in the treatment of tumor-bearing mice will also be impaired by obesity remains to be determined.

Article

The Rockefeller University Press $30.00J. Exp. Med. 2014 Vol. 211 No. 12 23732383www.jem.org/cgi/doi/10.1084/jem.20140116

2373

The use of immunotherapy (IT) in cancer has recently resulted in impressive responses. Yet, the usages of IT regimens, especially those involving cytokine therapies, have also resulted in the induction of severe systemic toxicities often not previously characterized in their preclinical animal studies. Because such toxicities often require intensive care management of patients, the causes for

the discrepancies in observations between clinical and preclinical toxicity merit further study.

Cancer has been considered a disease of aging, with persons over 65 accounting for over 60% of newly diagnosed malignancies (Balducci and Ershler, 2005). Consequences of aging include gradual decreases in immunological function, thymic involution leading to decreased naive T cell output, restriction within T cell repertoire

CORRESPONDENCE William J. Murphy: wmjmurphy@ucdavis.edu

Abbreviations used: AL, ad libitum; ALT, alanine aminotransferase; CR, calorie restricted; DIO, dietinduced obese; HD, high dose; IT, immunotherapy; LD, low dose; MRI, magnetic resonance imaging.

*A. Mirsoian and M.N. Bouchlaka contributed equally to this paper.**D.D. Taub and W.J. Murphy contributed equally to this paper.

Adiposity induces lethal cytokine storm after systemic administration of stimulatory immunotherapy regimens in aged mice

Annie Mirsoian,1* Myriam N. Bouchlaka,1* Gail D. Sckisel,1 Mingyi Chen,2 Chien-Chun Steven Pai,1 Emanuel Maverakis,1 Richard G. Spencer,5 Kenneth W. Fishbein,5 Sana Siddiqui,5 Arta M. Monjazeb,3 Bronwen Martin,5 Stuart Maudsley,5 Charles Hesdorffer,5 Luigi Ferrucci,5 Dan L. Longo,5 Bruce R. Blazar,6 Robert H. Wiltrout,7 Dennis D. Taub,4,8** and William J. Murphy4**

1Department of Dermatology, 2Department of Pathology and Laboratory Medicine, 3Department of Radiation Oncology, and 4Department of Dermatology and Internal Medicine, University of California, Davis, Sacramento, CA 95817

5National Institute on Aging-Intramural Research Program, National Institutes of Health, Biomedical Research Center, Baltimore, MD 21224

6Division of Blood and Marrow Transplantation, Department of Pediatrics, University of Minnesota, Minneapolis, MN 554557National Cancer Institute, Frederick, MD 217028Hematology and Immunology Translational Research Center, VA Medical Center, Washington, DC 20422

Aging is a contributing factor in cancer occurrence. We recently demonstrated that sys-temic immunotherapy (IT) administration in aged, but not young, mice resulted in induction of rapid and lethal cytokine storm. We found that aging was accompanied by increases in visceral fat similar to that seen in young obese (ob/ob or diet-induced obese [DIO]) mice. Yet, the effects of aging and obesity on inflammatory responses to immunotherapeutics are not well defined. We determine the effects of adiposity on systemic IT tolerance in aged compared with young obese mice. Both young ob/ob- and DIO-generated proinflammatory cytokine levels and organ pathologies are comparable to those in aged ad libitum mice after IT, culminating in lethality. Young obese mice exhibited greater ratios of M1/M2 macrophages within the peritoneal and visceral adipose tissues and higher percentages of TNF+ macrophages in response to CD40/IL-2 as compared with young lean mice. Macro-phage depletion or TNF blockade in conjunction with CD40/IL-2 prevented cytokine storms in young obese mice and protected from lethality. Calorie-restricted aged mice contain less visceral fat and displayed reduced cytokine levels, protection from organ pathology, and protection from lethality upon CD40/IL-2 administration. Our data dem-onstrate that adiposity is a critical factor in the age-associated pathological responses to systemic anti-cancer IT.

2014 Mirsoian et al. This article is distributed under the terms of an Attribution NoncommercialShare AlikeNo Mirror Sites license for the first six months after the publication date (see http://www.rupress.org/terms). After six months it is available under a Creative Commons License (AttributionNoncommercialShare Alike 3.0 Unported license, as described at http://creativecommons.org/licenses/ by-nc-sa/3.0/).

2374 Obesity induces immunostimulatory therapy toxicity | Mirsoian et al.

study seeks to establish the consequences of increased adipose tissue accumulation throughout aging upon ITinduced toxicities. Here, we find that adiposity results in a skewing toward increased levels of proinflammatory M1 macrophages within the peritoneal cavity and visceral adipose tissues that ultimately result in heightened production of proinflammatory cytokines, namely TNF, during strong immune stimulation with IT. Importantly, aged ad libitum (AL)fed mice and young obese mice succumb to TNFmediated pathological responses culminating in rapid lethality, which are prevented through caloric restriction in the aged or through either macrophage depletion or TNF blockade.

RESULTSIT administration in the aged results in rapid lethal cytokine storm and is associated with increased accumulation of visceral adipose tissueThe combination IT regimen consisting of CD40 and IL2 (CD40/IL2) has previously shown synergistic antitumor effects resulting in the regression of metastatic tumors in young (

JEM Vol. 211, No. 12

Article

2375

were similar to young mice (Fig. 1, GI). MRI scans of young (4 mo), middleaged (9 mo), and aged (15 mo) mice that were either ALfed or CR throughout life showed significant differences in adipose accumulation over time (Fig. 1 I). We found that with increasing age, ALfed mice have a proportional increase in body fat content, whereas mice placed on a CR diet lack increased fat accumulation regardless of age. Aged CR mice show no significant difference in fat volume ratios to their middleage and young counterparts, and remain similar to young AL mice (Fig. 1 I).

Caloric restriction in aged mice results in decreased accumulation of adiposity and protection from IT-induced toxicityTo examine the effect of adiposity and aging on the development of systemic toxicities after IT, aged AL, aged CR, and young AL mice were administered IT or a control therapy

mice (Fig. 1, G and I). These results indicate that aged mice are not only heavier and larger in size but that these size differences are attributed to increases in adipose tissue accumulation.

Adipose tissue is widely published to be a highly active metabolic organ with immune modulatory capabilities. To address whether in our aged ALfed mice, who were allowed to freely access food, the increased adiposity noted was contributory toward the development of ITinduced toxicities in the aged, we aimed to determine the effects of IT administration into an aged calorierestricted (CR) cohort. CR mice are placed on a food restriction diet beginning at 14 wk of age and maintained throughout life. To ensure that agematched CR aged mice contained a decreased accumulation of body fat in comparison with aged AL mice, we quantified total body weight and visceral fat through MRI (Fig. 1, H and I). CR mice weighed significantly less than AL aged mice and exhibited a decreased presence of visceral adiposity to levels that

Figure 1. Aged AL mice show increased proinflammatory cytokines, rapid lethality to LD IT, and display increased fat volume ratios in comparison to aged CR mice. (A) Ad libitumfed C57BL/6 mice were treated with LD CD40/IL-2 or rIgG/PBS and survival was monitored, n = 58 mice/group. (B) Serum cytokines TNF, IL-6, and IFN- were measured day 2 by CBA after the start of CD40/IL-2 or rIgG/PBS therapy in aged AL mice, n = 35 mice/group. (C) Young AL (3 mo) and aged AL (18 mo) mice were treated with CpG/IL-2 IT and survival was monitored, n = 5 mice/group. (D and E) Serum cytokine levels of TNF (D) and IL-6 (E) were measured by CBA on day 2 after the start of therapy, n = 5 mice/group. (F) Representative image of aged and young C57BL/6 mice showing size and waist circum-ference. (G) Total body weight (left graph) and visceral fat weight (right graph), was mea-sured for young (2 mo) and aged AL mice (20 mo), n = 5 mice/group. (H) Total body weight (left graph) or visceral fat (right graph) was measured in naive 15-mo-old C57BL/6 mice on an AL or CR diet, n = 34 mice/group. (I) Visceral fat in young (4 mo), middle-aged (9 mo), and aged (15 mo) mice on either an AL (left) or CR (right) diet was evaluated by MRI. WS is indicative of water suppression, where fat is bright. FS is indicative of fat suppression, fat is dark/black. Data in all panels are repre-sentative of one of at least five experiments with similar results. MRI scans are representa-tive of one of at least three mice imaged per group with similar results. Survival analysis was plotted according to the Kaplan-Meier method, and statistical differences were deter-mined with the log-rank test. Bar graph (mean value SEM) statistics were performed using either one-way or two-way ANOVA with Bonferroni post-hoc tests. ***, P < 0.001; **, P < 0.01; *, P < 0.05.

2376 Obesity induces immunostimulatory therapy toxicity | Mirsoian et al.

succumb to mortality on day 2 of treatment (Fig. 2 G). However, 100% of aged CR mice placed on the HD amount of IT demonstrated complete protection from CD40/IL2mediated lethality (Fig. 2 G). Collectively, these data suggest that increased body fat plays a critical role in the induction of systemic toxicities by IT. Increased adiposity played a critical role in the ageassociated cytokine storm and subsequent organ pathologies after IT and calorie restriction diminished the development of such toxicities conferring a protective effect that allowed for increased survival in the aged.

Increased adipose tissue accumulation through obesity, in the absence of age, results in IT-induced lethal cytokine stormTo further dissect the role of adipose tissue in the ITassociated inflammatory responses seen in aged mice, we next sought to determine the impact of adiposity, in the absence of age, as a possible cofactor. Young ob/ob mice (B6.CgLepob/ob) are leptindeficient and therefore start exhibiting obesity by 3 wk of age. In comparison to agematched young WT (2 mo) control mice placed on an AL diet (young AL), ob/ob mice have significantly higher body masses and weigh, on average, 2.5 greater (Fig. 3 A). MRI analysis of body fat content demonstrated

(rIgG/PBS). IT resulted in increased lymphocytic infiltrates to the liver of aged CR and aged AL mice, in contrast to agematched controltreated mice (Fig. 2, AC). To assess whether IT resulted in organ damage, livers and intestines were collected on day 2 of treatment, the day of mortality in aged AL mice, and assessed for pathological changes. Importantly, livers and intestines from aged CR mice demonstrated significantly less inflammation and less necrosis in comparison with aged AL mice (Fig. 2, A and C). Serum analysis of alanine aminotransferase (ALT) enzyme levels, a correlate of liver damage, showed a significant reduction in ALT levels in aged CR mice in comparison with aged AL mice (Fig. 2 B). Additionally, serum proinflammatory cytokine analysis of TNF (Fig. 2 D), IL6 (Fig. 2 E), and IFN (Fig. 2 F) resulted in all cytokines being significantly decreased in aged CR mice in comparison with the aged AL mice. These decreases in cytokine values were to levels that were either lowerthan or aslowas those in young AL mice that received IT (Fig. 2, DF). To determine if these lower cytokine values during IT therapy would confer protection against ITinduced mortality, aged AL mice and aged CR mice were administered either LD IT or HD IT and monitored for mortality. Independent of treatment dose, aged AL mice all

Figure 2. Calorie restriction in the aged confers protection by decreasing cytokine levels and IT-associated organ pathology, thereby allowing increased survival during IT. (AC) Aged AL and age-matched aged CR mice (18 mo) were treated with HD CD40/IL-2 or rIgG/PBS; at day 2 of IT, mice were sacrificed for histological analysis of livers (A) and intestines (C). Serum was collected on day 2 and assessed for ALT (B) levels, n = 3 mice/group. Representative H&E images of liver and intestines (gut) for aged AL and aged CR groups. Bars, 200 m. Asterisks represent steatosis and arrows denote areas of severe immune infiltration and/or necrosis. (DF) Serum was analyzed day 2 after treatment for TNF (D), IL-6 (E), and IFN- (F). (G) Aged (18 mo) C57BL/6 mice on AL or CR diet were treated with either LD or HD of CD40/IL-2 and survival was monitored. Control groups received rIgG/PBS. All data panels are representative of one of four independent experiments with similar results for all panels. Survival analysis was plotted according to the Kaplan-Meier method and statistical differences were determined by using the log-rank test. Bar graph (mean value SEM) statistics were performed using one-way ANOVA with Bonferroni post-hoc test. ****, P < 0.0001; ***, P < 0.001; **, P < 0.01.

JEM Vol. 211, No. 12

Article

2377

was comparable to the aged AL mice and significantly greater than the young AL (Fig. 3, C and D).

Because HD IT resulted in significant changes within the obese ob/ob mice, we next administered the lower dose (LD) of IT to young ob/ob, young agematched AL mice, aged AL, and aged CR mice. Administration of LD IT still resulted in significant increases in TNF and IL6 proinflammatory cytokine levels within the young ob/ob that were not significantly different to the aged AL mice (Fig. 3, E and F). Both ob/ob and aged AL were significantly greater than both the young AL and aged CR mice, where CR in the aged led to significant downregulation in both TNF and IL6 cytokine production (Fig. 3, E and F). Although the young ob/ob and the aged AL mice displayed increased levels of IFN, these levels were not significant among the groups (Fig. 3 G). Consistent with the increase in proinflammatory cytokines at day 2 of IT treatment, 100% of young ob/ob mice succumbed to ITinduced lethality within 4 days of treatment in contrast to 100% survival in the

that young ob/ob mice have higher total body fat content that is similar in accumulation to the aged mice, being deposited largely in the intraabdominal cavity, in comparison with young AL mice (Fig. 3 B).

To determine if adiposity could solely induce toxic consequences in aged AL, young ob/ob and agematched young AL mice were treated with HD IT and at day 2 of treatment analyzed for the presence of immunemediated organ pathological changes within the liver and intestines (Fig. 3, C and D). IT resulted in mild inflammatory infiltrates in young AL mice but resulted in moderate to severe inflammation in the ob/ob mice, which was comparable to the aged within the intestines (Fig. 3 D). Yet, analysis of ITinduced liver pathology was hindered by the presence of extensive steatosis in ob/ob mice, resulting in a paradoxical interference in determining the intensity of immunemediated inflammatory liver damage (Fig. 3 C). IT resulted in both the intestines and liver displaying the most immunemediated inflammatory damage within the ob/ob that

Figure 3. Increased adiposity indepen-dently of age results in systemic release of proinflammatory cytokines and organ pa-thology after IT in young ob/ob (obese) and aged AL mice. (A) Body weights of naive young (2 mo) WT and young ob/ob C57BL/6 mice, n = 6 mice/group. Representative pic-tures of young-AL WT mice and obese ob/ob naive mice are shown. (B) Total fat content in naive young WT (2 mo), young ob/ob (2 mo), and aged (18 mo) C57BL/6 mice was visualized by MRI. Fat is bright. (C and D) Young AL (2 mo), young ob/ob (2 mo), and aged AL (18 mo) mice were treated with HD CD40/IL-2 and liver (C) and gut (D) were stained with H&E and scored for histopathology on day 2. Bars, 200 m. Arrows indicate areas of steatosis and/or immune cell infiltration and asterisks indicate patchy necrosis and/or fibrosis. (EG) Young AL (2 mo), young ob/ob (2 mo), aged AL (18 mo), and aged CR (18 mo) were treated with LD CD40/IL-2 and serum was collected day 2 of treatment and analyzed by CBA for levels of: TNF (E), IL-6 (F), and IFN- (G), n = 3 mice/group. (H and I) Young AL (2 mo), young ob/ob (2 mo), aged AL (18 mo), and aged CR (18 mo) were treated with LD CD40/IL-2 (H) or CpG/IL-2 (I) or rIgG/PBS control and survival was monitored. (J) Serum levels of TNF were measured on day 2 of CpG/IL-2 therapy by CBA. Data in all panels are representative of one of at least three experiments with similar results. MRI scans are representative of at least three mice imaged per group with similar results. Survival data were plotted using the Kaplan-Meier method and statistical differ-ences were determined with the log-rank test. Bar graph (mean value SEM) statistics per-formed using Students t test or one-way ANOVA with Bonferroni post-hoc tests. ***, P < 0.001; **, P < 0.01; *, P < 0.05.

2378 Obesity induces immunostimulatory therapy toxicity | Mirsoian et al.

This data further supports that increased adiposity plays a crucial role in the induction of systemic toxicities after IT. These data are suggestive that not only does an aging environment result in lethal toxicities but toxicity may also be dependent on the increase in fat accumulation as the young ob/ob mice succumb to cytokine storm induction and mortality after IT, similar to that observed in the aged mice.

IT toxicity is TNF- and M1 macrophagedependent within the peritoneum and visceral adipose tissuesA hallmark of obesity is the increasing infiltration of macrophages into adipose tissues (ATMs) through the secretion of MCP1 by adipocytes. Phenotypically, these macrophages are

young WT cohort (Fig. 3 H). 100% of aged AL mice yielded to lethality from the IT by day 2 of treatment, whereas 100% of aged CR mice survived the entire regimen (Fig. 3 H).

The pathological and lethal response observed within the young ob/ob mice was not CD40/IL2 IT administration specific. Young AL and agematched young ob/ob mice were administered CpG in combination with HD IL2 and, likewise, resulted in lethality of 75% of young ob/ob mice by day 5 of treatment but resulted in complete tolerance by young AL mice (Fig. 3 I). CpG/IL2 therapy led to significantly increased levels of serum proinflammatory cytokine TNF in an analogous trend as CD40/IL2 administration on day 2 of therapy (Fig. 3 J).

Figure 4. IT toxicity is TNF-dependent through M1 macrophage accumulation. (AD) Young (2 mo) WT and young (2 mo) ob/ob mice were treated with either control rIgG/PBS or LD CD40/IL-2. Visceral adipose tissue and peritoneal lavages were collected on day 2 and assessed for the pres-ence of macrophages identified as CD45+CD19F4/80+CD11b+. Ratio of M1/M2 macrophages (A), and total numbers (B) and percentages (C) of TNF- producing macrophages. Macrophages identified as M1 macrophages were characterized as CD206; M2 macrophages were gated as CD206+. (EH) Young AL (2 mo) and young ob/ob (2 mo) were treated with either control or macrophage depleting clodronate liposomes before and during LD CD40/IL-2 and survival was monitored; n = 5 mice/group. (F and G) On day 2, visceral adipose tissues and peritoneal lavages were collected and assessed for the percentages TNF+ macrophages (F and G) as characterized in experiments in AD and serum TNF levels (H), n = 4 mice/group. (IL) Young AL (2 mo) and young ob/ob (2 mo) mice were treated with either control hIgG/PBS or subcutaneous injection of 1.5 mg/0.1 ml etanercept before and during LD CD40/IL-2 and survival was monitored (I), n = 5 mice/group. (J and K) On day 2, visceral adipose tissue and peritoneal lavages were collected and assessed by flow cytometry for the percentages of TNF+ macrophages (J and K) and serum TNF levels (L); n = 4 mice/group. Data in all panels are representative of one of at least three experiments with similar results. Survival data were plotted using the Kaplan-Meier method and statistical differences were determined with the log-rank test. Bar graph (mean value SEM) statistics performed using one-way ANOVA with Bonferroni post-hoc tests. **, P < 0.01; *, P < 0.05.

JEM Vol. 211, No. 12

Article

2379

(Fig. 4, IL). Both macrophage depletion and cotreatment with etanercept during LD CD40/IL2 resulted in complete protection of young ob/ob mice from lethality (Fig. 4, E and I). As expected, treatment with liposomal clodronate led to significantly decreased levels of TNFproducing macrophages within both the visceral fat and peritoneal lavages (Fig. 4, F and G), which significantly reduced systemic circulating serum levels of proinflammatory cytokine TNF (Fig. 4 H). Additionally, coadministration of LD CD40/IL2 with etanercept did not reduce the percentages of TNFexpressing macrophages within the visceral fat or peritoneal cavity (Fig. 4, J and K) but mediated complete protection through an overall reduction in systemic TNF levels (Fig. 4 L).

Together, these data indicate that proinflammatory M1 macrophages are the cellular mediator in the induction of IT toxicity, particularly through increased presence within the peritoneal and visceral adipose tissues. Importantly, these effects are TNFdependent, as blocking TNF levels either through etanercept administration or macrophage depletion resulted in complete protection of young ob/ob mice from lethality.

Diet-induced obese (DIO) mice succumb to cytokine storm, organ pathology, and lethality similar to ob/ob and aged AL mice after ITob/ob mice are congenic for the spontaneous mutation in the OB gene, resulting in leptin hormone deficiency. To extend our findings, we generated young DIO mice. Mice were placed on a purified 60% fat diet to induce obesity beginning at 3 wk of age and were determined to be obese when weighing >40 g (DIO). Agematched counterparts were placed on a matching

predominately classically activated M1 cells that also express proinflammatory mediators IL1, TNF, and IL6 (Lumeng et al., 2007; Fujisaka et al., 2009). As such, we next sought to determine if increases in M1 macrophages within the peritoneal cavity and visceral adipose tissues could account for the pathological increases in systemic TNF during IT administration. Young AL and agematched young ob/ob mice were administered LD CD40/IL2. At day 2 of treatment, visceral adipose tissues and peritoneal lavages were phenotypically assessed by flow cytometry for the presence of M1 (CD206) macrophages, M2 (CD206+) macrophages, and percentages of TNFexpressing macrophages. Consistently, young ob/ob mice demonstrated significant increases in M1 macrophage proportions during IT administration, whereas young AL mice were able to maintain consistent M1toM2 ratios (Fig. 4 A). Additionally, both the visceral adipose tissues and peritoneal lavages of ob/ob mice experienced significant increases in percentages of TNFexpressing macrophages as opposed to young AL mice (Fig. 4, BD). Significant changes were not observed in the spleen of either young ob/ob or AL mice (unpublished data). This data, consistent with obesity literature, suggests that increases in adipose tissue accumulation within the visceral cavity may impact regulation of M1 and M2 macrophage ratios that ultimately may affect TNF responses to immune stimulation with immunotherapeutic regimens.

Given the increases in M1 macrophages within the obese mice and dysregulation observed with IT, young ob/ob and young AL mice were treated with LD CD40/IL2 in conjunction with either liposomal clodronate for macrophage depletion (Fig. 4, EH) or etanercept (Enbrel) for TNF blockade

Figure 5. DIO succumb to cytokine storm, organ pathology, and lethality similar to aged AL mice after IT. Young C57BL/6 mice were placed on a 60% fat diet starting at 3 wk of age to induce obesity (young DIO). Lean age-matched controls (young lean) were fed a 10% fat diet beginning at 3 wk of age. Young mice were ana-lyzed at 5 mo of age, and aged AL mice were analyzed at 18 mo of age, n = 6 mice/group. Av-erage body weight (A), representative pictures (B), and total fat content as measured by MRI (C) are shown for each group. (DF) Young lean (5 mo), young DIO (5 mo), and aged AL (18 mo) mice were treated with HD CD40/IL-2 or rIgG/PBS. Serum was collected day 2 after the start of treatment and analyzed by CBA for levels of TNF (C), IL-6 (D), and IFN- (E) and survival was ana-lyzed (F). Data in AC are representative of one of at least three experiments with similar results and data in DG are representative of one of at least four experiments with similar results. Survival data were plotted using the Kaplan-Meier method and statistical differences were determined with the log-rank test. Bar graph (mean value SEM) statistics performed using one-way ANOVA with Bonferroni post-hoc tests. ****, P < 0.0001; ***, P < 0.001; **, P < 0.01; *, P < 0.05.

2380 Obesity induces immunostimulatory therapy toxicity | Mirsoian et al.

were observed between young obese and aged AL mice, where aged mice die by day 2 but young obese mice die at day 4 of treatment. Therefore, although our data strongly demonstrates that adiposity plays a central role in the induction of toxicities, aging is also a likely key factor. Therefore, our findings highlight the importance in considering both the status of aging and obesity of patients receiving systemic administration of stimulatory IT regimens. It will be important to conduct studies dissociating the consequences of aging upon increased cytokine responsiveness, possibly through either investigating epigenetic modifications induced by the metainflammatory and/or inflammaging states associated with increased adipose accumulation or whether aging is accompanied by changes in expression levels of various receptors that allow for a quicker response.

The development of IT regimens continues on a positive trajectory with increasing successes within a variety of cancers. Active IT has taken a greater comprehensive form focusing on activating humoral and cellmediated immunity, and these strategies have included antigenspecific vaccines (i.e., MART1 vaccines), DCbased vaccines (i.e., SipuleucelT), inhibitors of checkpoint blockade (i.e., CTLA4 and PD1), and systemic stimulatory regimens through cytokine based therapies and application of agonistic antibodies (i.e., CD40, IFN, and IL2). Our data presented herein demonstrates that both young obese and aged AL mice succumb to rapid lethality during systemic administration of stimulatory therapies. Our data strongly suggests that excess adiposity may be a prognostic factor in determining patient outcome and treatment tolerance. However, additional studies need to be conducted examining the effects of adiposity on the development of adverse reactions within other classes of IT regimens, such as checkpoint inhibitors.

Clinically, we hypothesize that taking into account patient BMI, adipose accumulation, and age may better modulate treatment type choices as well as the inclusion of supportive therapeutics such as TNF blockade in conjunction with anticancer therapies that, overall, may increase patient survival and lead to greater anticancer responses. Followup investigations in tumorbearing models merit further study.

MATERIALS AND METHODSAnimals and diets. Female young (25 mo old), middleaged (912 mo old), and aged (1522 mo old) C57BL/6 mice were purchased from Charles River or from the animal production area at the National Cancer Institute. Young female B6.CgLepob/ob and young agematched female WT control C57BL/6 mice were purchased from The Jackson Laboratory at 812 wk of age.

DIO obese and control lean mice were generated from WT C57BL/6 mice purchased at 3 wk of age from Charles River or from the animal production area at the National Cancer Institute. Mice were placed on an opensource purified diet consisting of either 60% highfat (D12492) or 10% lowfat (D12450J) constitution (Research Diets, Inc.). Agematched mice were placed on either a highfat or lowfat diet beginning at 3 wk of age for 1620 wk and weighed weekly. Mice were deemed obese upon reaching a minimum of 40 g of weight.

For studies comparing aged AL and CR dietary effects, both female aged CR and agematched female ALlittermates (1522 mo old) were purchased from the CR rodent colony at the National Institutes on Aging, National Institutes of Health (NIH). In brief, mice were either raised through free access, AL, feeding on the NIH31 regular chow, or placed on a limited daily feeding

purified diet that was 10% fat and served as lean controls. Increases in body mass were validated in DIO mice before IT administration (Fig. 5, AC). DIO mice displayed increases in overall body weight and size (Fig. 5, A and B). MRI analysis of body fat content showed increased total body fat accumulation, largely as increased visceral fat, in young DIO mice that was similar to the aged AL mice (Fig. 5 C).

Confirming our previous findings, DIO mice treated with HD CD40/IL2 displayed increased serum TNF, IL6, and IFN levels compared with agematched young mice on a 10% fat diet after IT administration (Fig. 5, DF). The increases in proinflammatory cytokines in DIO mice were lower, yet not significantly so, than aged mice after IT (Fig. 4, CE). Young DIO mice started succumbing to IT lethality at day 3 of treatment (>60%), and

JEM Vol. 211, No. 12

Article

2381

Murine macrophage studies. On day 2 of treatment, murine macrophages were isolated from either the peritoneal cavity or visceral adipose tissue. For peritoneal lavages, 4 ml of cold PBS was injected into the peritoneal cavity of each mouse and then aspirated. Lavages were spun down at 1,200 rpm for 5 min and the resulting pellet was resuspended in RF10c media and plated for flow cytometry staining.

The stromal vascular fraction of visceral adipose tissues was collected day 2 of treatment. Adipose tissues from the visceral cavity (gonadal) were collected and then incubated for 60 min in a digestion solution consisting of 2 mg/ml type IV collagenase (SigmaAldrich) dissolved fresh in a 2% BSA in PBS solution. After incubation samples were centrifuged at 1,200 rpm for 5 min for layer separation of the stromal vascular fraction from adipocytes. The resulting stromal vascular fraction was then removed, resuspended into a single cell suspension, counted, and plated for flow cytometry staining.

Flow cytometry analysis of macrophages. Single cell suspensions from the peritoneal lavages and visceral adipose tissues were plated in RF10c media with the addition of GolgiPlug (containing Brefeldin A) and GolgiStop (containing Monensin) Protein Transport Inhibitors (BD) according to manufacturers protocols. Plates were incubated for 9 h at 37C and 5% CO2.

After incubation, cells were washed with PBS and stained for flow analysis. Single cell suspensions were labeled with Fc block (purified antimouse CD16/32; BD) for 10 min, followed by labeling with antibodies for 20 min at 4C. Cells were washed twice with a staining buffer consisting of 5% FBS (Gemini Bioproducts) in DPBS (Corning). For intracellular staining, the Cytofix/Cytoperm kit (BD) was used per manufacturers protocols. After staining, cells were analyzed using a custom configured Fortessa cytometer, and by using FACSDiva software (BD). Data were analyzed using FlowJo software vX (Tree Star). Antibodies used to distinguish macrophages were: PBconjugated antimouse CD45, BV711conjugated antimouse CD11b, BV605 and PEconjugated antimouse CD206, BV785conjugated antimouse CD19, APCconjugated antimouse F4/80, and PECy7conjugated antimouse TNF (BioLegend).

Serum cytokine bead array. Serum levels of TNF, IFN, and IL6 were quantified by multiplex measurement using the Cytometric Bead Array (CBA; BD) kit as described previously (Bouchlaka et al., 2013). In brief, serum samples and standards were incubated for 1 h at 25C with a bead mixture containing bound antibodies to TNF, IFN, and IL6 according to manufacturers instructions. Samples and standards were resuspended in wash buffer and analyzed by flow cytometry. Data were acquired on a customconfigured Fortessa cell analyzer using FACSDiva software (BD) and analyzed using FlowJo software (Tree Star). Each sample was analyzed in triplicate. Upon analysis of raw data, the mean fluorescent intensities (MFIs) of each bead cluster were quantified. Where indicated in the figures, protein concentrations were extrapolated relative to a standard curve created by serial dilution of the mouse positive control cytokine.

Colorimetric liver enzyme assay. Serum ALT was quantified using the ID Labs ALT Enzymatic Assay kit (ID Labs Biotechnology Inc.) as previously described (Bouchlaka et al., 2013). Colorimetric determination of ALT levels was performed by reading the absorbance of each well at 340 nm on a plate reader (VERSAmax turntable plate reader). The concentration of ALT (U/liter)

schedule with the NIH31fortified chow. CR is initiated at 14 wk of age at 10% restriction, increased to 25% restriction at 15 wk, and to 40% restriction at 16 wk of age where it is maintained throughout the life of the animal. Information on survival curves, food intake curves, and body weights of the NIA strains has been previously published (Turturro et al., 1999).

All experimental mice were housed in the Animal Facilities at the University of California, Davis under specific pathogenfree (SPF) conditions. For the comparison in fat volume ratios, young, middleaged, and aged AL and CR mice were ordered and sent to the National Institute on Aging for MRI analysis and also housed under SPF conditions. All of the animal protocols were approved by the Institutional Animal Care and Use Committees of both institutions. Body weights were also measured for several mice at different ages and on dietary regimens. Mice were euthanized by CO2 asphyxiation and the visceral fat pads (mesenteric, gonadal, and renal) were collected and weighed together for each mouse to determine the visceral fat weight.

Reagents. The agonistic antimouse CD40 antibody (clone FGK115B3) was generated via ascites production in our laboratory as previously described (Murphy et al., 2003; Bouchlaka et al., 2013). The endotoxin level of the CD40 was

2382 Obesity induces immunostimulatory therapy toxicity | Mirsoian et al.

normal body temperature using warm air. Mice were inserted into the scanner in a headfirst prone orientation. Spinecho T1weighted images were obtained through the mouse body (necktobase of the tail) using a 72mm linear volume coil. Scan sequence parameters were the following: TR 1,000 ms, TE 15 ms, and 2 averages. The field of view was 7.7 3.85 2.0 cm, with a matrix size of 256 128 40. The corresponding voxel size was 0.3 0.3 0.5 mm. Images were acquired using ParaVision 5.0 software. After acquisition, images were transferred into Invenon Research Workplace 4.0 software (Siemens Preclinical) allowing fat to be displayed with high intensity (white).

Statistical analysis. Statistical analyses were performed using Prism software (GraphPad Software Inc.). Data were expressed as mean SEM. For analysis of three or more groups, a nonparametric ANOVA test was performed with a Bonferroni posthoc test. Analysis of differences between two normally distributed test groups was performed using the Students t test. Nonparametric groups were analyzed with the MannWhitney test. Welchs correction was applied to datasets with significant differences in variance before Students t test.

We thank Monja Metcalf and Weihong Ma for technical help.This work has been supported by NIH grants CA0905572 and AG034874 and

by the Intramural Research Program of the National Institute on Aging, NIH.The authors declare no competing financial interests.

Author contributions: A. Mirsoian, M.N. Bouchlaka, D.D. Taub, and W.J. Murphy designed the research; A. Mirsoian, M.N. Bouchlaka, G.D. Sckisel, M. Chen, C.C.S. Pai, A.M. Monjazeb, and D.D. Taub performed the research; A. Mirsoian, M.N. Bouchlaka, D.D. Taub, and W.J. Murphy analyzed the data; M. Chen analyzed and scored all histology data; E. Maverakis provided etanercept (Enbrel) for TNF blockade studies; R.G. Spencer, K.W. Fishbein, and S. Siddiqui performed and analyzed MRI studies; B. Martin, C. Hesdorffer, L. Ferrucci, and S. Maudsley provided aged mice and assisted with data interpretation; A. Mirsoian, M.N. Bouchlaka, D.D. Taub, and W.J. Murphy wrote the paper; and G.D. Sckisel, D.L. Longo, B.R. Blazar, and R.H. Wiltrout helped with the editing of the paper. All authors read and approved the manuscript.

Submitted: 18 January 2014Accepted: 6 October 2014

REFERENCESAilhaud, G. 2000. Adipose tissue as an endocrine organ. Int. J. Obes. Relat.

Metab. Disord. 24:S1S3. http://dx.doi.org/10.1038/sj.ijo.0801267Balducci, L., and W.B. Ershler. 2005. Cancer and ageing: a nexus at several lev

els. Nat. Rev. Cancer. 5:655662. http://dx.doi.org/10.1038/nrc1675Bouchlaka, M.N., G.D. Sckisel, M. Chen, A. Mirsoian, A.E. Zamora, E.

Maverakis, D.E. Wilkins, K.L. Alderson, H.H. Hsiao, J.M. Weiss, et al. 2013. Aging predisposes to acute inflammatory induced pathology after tumor immunotherapy. J. Exp. Med. 210:22232237. http://dx.doi.org/ 10.1084/jem.20131219

Campisi, J. 2013. Aging, cellular senescence, and cancer. Annu. Rev. Physiol. 75:685705. http://dx.doi.org/10.1146/annurevphysiol030212183653

Franceschi, C. 2007. Inflammaging as a major characteristic of old people: can it be prevented or cured? Nutr. Rev. 65:S173S176. http://dx.doi.org/ 10.1111/j.17534887.2007.tb00358.x

in each sample was then directly determined from the change in absorbance within 5 min time. Dilutions of the Pyruvate Control, included in the kit, were used to construct a standard curve to calibrate the assay. Every serum sample was assayed in triplicate.

Histopathology and grading/scoring. Liver and whole intestines were collected on day 2 of IT, flushed, and fixed in 10% paraformaldehyde, embedded in paraffin, cut into 5m sections, and stained with hematoxylin and eosin (H&E). All tissues were prepared and stained at Histology Consultation Services, Inc. (Everson, WA). Images were captured with a microscope (BX4; Olympus) equipped with a Qcolor3 camera and 10 numerical aperture objective lens. Magnification for each captured image is specified in the figure legends. Grading of histopathological inflammation was performed using a scale from 0 to 4 in a blind fashion by a boardcertified pathologist (M. Chen) at the UC Davis Medical Centers Department of Pathology and Laboratory Medicine. Specifically, the grading score for liver and gastrointestinal tract followed our previously published grading method (Bouchlaka et al., 2013), summarized in Tables 1 and 2.

MRI. To quantify abdominal fat, MR images were acquired in collaboration with the National Institute on Aging. Images were acquired using a Biospec 7 Tesla 30cm MRI scanner (Bruker Biospin) with a 72mm diameter transmitreceive birdcage coil. In each experiment, two mice were inserted into the scanner, side by side, in feetfirst prone orientation immediately after sacrifice. During scanning, the mice were cooled with a stream of cold air from a vortex tube (Exair, Inc.) to minimize decomposition. Two different pulse sequences were used: a heavily T1weighted fast spin echo (RARE) sequence yielding brightfat (watersuppressed [WS]) images and a fatsuppressed proton density weighted fast gradient echo (FLASH) sequence yielding fatsuppressed (FS) images. 3D datasets were acquired in axial orientation with a field of view of 72 36 80 mm (leftright anteriorposterior headfoot) and matrix size 256 128 128. The voxel size (spatial resolution) was 281 281 625 m. The spectral bandwidth in both sequences was 100 kHz (391 Hz/pixel) and two averages were acquired for improved signaltonoise ratio. For the RARE sequence, the RARE factor (i.e., number of spin echoes per shot or number of kspace lines per segment) was 8, the repetition time (TR) was 125 ms, the actual echo time (TE) was 11.4 ms, and the effective echo time (TEeff) was 46.3 ms. Each RARE scan took 8 min and 32 s to acquire. For the FLASH sequence, the repetition time was 25 ms and the echo time was 3.2 ms. The excitation pulse had a flip angle of 30. Each FLASH scan took 13 min 39 s to acquire. Image processing was performed in ImageJ 1.45 (NIH). The statistical analyses were performed using oneway ANOVA and the NewmanKeuls posthoc test. P < 0.05 was considered to be significant.

To visualize fat distribution and content imaging of young ob/ob, DIO, and lean agematched controls were conducted in collaboration with the Center for Molecular and Genomic Imaging (CMGI) Facility at the University of California, Davis. Mice were anesthetized before imaging using 2% isoflurane in an induction chamber and then maintained on 12% isoflurane throughout imaging via a nose cone fitted for anesthetic inhalation. Mice were imaged using a BioSpec 70/30 7T (Bruker Biospin) horizontal bore system designed specifically for small animal imaging. Animals were kept at

Table 2. Grading of acute intestinal damageGlobal assessment Criteria

0: None No inflammationI: Minimal/indeterminate Scattered infrequent inflammatory infiltrate in a minority of crypts, which is minimalII: Mild Inflammatory infiltrate in 50% of the crypts, which is generally mild, and confined within the mucosa epithelium with

mild architectural distortionIII: Moderate Inflammatory infiltrate involving most of the crypts, and associated with architectural distortion (villous blunting and

focal mucosal atrophy/erosion) and luminal inflammatory exudateIV: Severe Extensive inflammatory infiltrate (large lymphoid aggregates) with expanding into the lamina/muscular propria, and

associated with prominent mucosal ulceration/denuding, muscular wall necrosis, and/or perforation

JEM Vol. 211, No. 12

Article

2383

Franceschi, C., M. Bonaf, S. Valensin, F. Olivieri, M. De Luca, E. Ottaviani, and G. De Benedictis. 2000. Inflammaging. An evolutionary perspective on immunosenescence. Ann. N. Y. Acad. Sci. 908:244254. http://dx.doi .org/10.1111/j.17496632.2000.tb06651.x

Fujisaka, S., I. Usui, A. Bukhari, M. Ikutani, T. Oya, Y. Kanatani, K. Tsuneyama, Y. Nagai, K. Takatsu, M. Urakaze, et al. 2009. Regulatory mechanisms for adipose tissue M1 and M2 macrophages in dietinduced obese mice. Diabetes. 58:25742582. http://dx.doi.org/10.2337/db081475

Harris, T.B. 2002. Invited commentary: body composition in studies of aging: new opportunities to better understand health risks associated with weight. Am. J. Epidemiol. 156:122124, discussion :125126. http://dx.doi.org/ 10.1093/aje/kwf024

Lee, S., and K. Margolin. 2011. Cytokines in cancer immunotherapy. Cancers (Basel). 3:38563893. http://dx.doi.org/10.3390/cancers3043856

Lumeng, C.N., S.M. Deyoung, J.L. Bodzin, and A.R. Saltiel. 2007. Increased inflammatory properties of adipose tissue macrophages recruited during dietinduced obesity. Diabetes. 56:1623. http://dx.doi.org/10.2337/db061076

Murphy, W.J., L. Welniak, T. Back, J. Hixon, J. Subleski, N. Seki, J.M. Wigginton, S.E. Wilson, B.R. Blazar, A.M. Malyguine, et al. 2003. Synergistic

antitumor responses after administration of agonistic antibodies to CD40 and IL2: coordination of dendritic and CD8+ cell responses. J. Immunol. 170:27272733. http://dx.doi.org/10.4049/jimmunol.170.5.2727

Ronti, T., G. Lupattelli, and E. Mannarino. 2006. The endocrine function of adipose tissue: an update. Clin. Endocrinol. (Oxf.). 64:355365.

Spaulding, C.C., R.L. Walford, and R.B. Effros. 1997. Calorie restriction inhibits the agerelated dysregulation of the cytokines TNFalpha and IL6 in C3B10RF1 mice. Mech. Ageing Dev. 93:8794. http://dx.doi.org/10.1016/ S00476374(96)018246

Turturro, A., W.W. Witt, S. Lewis, B.S. Hass, R.D. Lipman, and R.W. Hart. 1999. Growth curves and survival characteristics of the animals used in the Biomarkers of Aging Program. J. Gerontol. A Biol. Sci. Med. Sci. 54:B492B501. http://dx.doi.org/10.1093/gerona/54.11.B492

VzquezVela, M.E., N. Torres, and A.R. Tovar. 2008. White adipose tissue as endocrine organ and its role in obesity. Arch. Med. Res. 39:715728. http://dx.doi.org/10.1016/j.arcmed.2008.09.005

Yang, H., Y.H. Youm, B. Vandanmagsar, J. Rood, K.G. Kumar, A.A. Butler, and V.D. Dixit. 2009. Obesity accelerates thymic aging. Blood. 114:38033812. http://dx.doi.org/10.1182/blood200903213595

INSIGHTS

2 INSIGHTS | The Journal of Experimental Medicine

A key challenge at the intestinal barrier is to minimize responses to commensal bacteria, which can lead to inflammatory bowel disease (IBD) in genetically predisposed individuals, but retain the ability to recognize and control the growth of infectious pathogens. Group 3 innate lymphoid cells (ILC3) help maintain intes-tinal homeostasis by producing the cytokine IL-22, which promotes mucosal healing and maintains barrier integrity. Microbial signals trigger production of IL-23 and IL-1b, which stimulate ILC3s to produce IL-22, leading to the induction of antibacterial peptides and epithelial cell regeneration.

But the identity of the cell type producing IL-23 in response to microbial signals is unclear and has been the subject of much debate; resident mononuclear phagocytes, inflammatory monocytes, and conventional dendritic cells have all been implicated. In this issue, Longman et al. provide compelling evidence, both in mouse following Clostridium rodentium infection and in patients with colitis, that

CX3CR1+ mononuclear phagocytes (MNPs) are the most potent producers of IL-23 and IL-1b and are very efficient in

inducing IL-22 production by ILC3. The authors demonstrated the importance of microbial stimulation in IL-22 induction

in patients with surgical diversion of the fecal streamIL-22 production by ILC3 was lower in the sites unexposed to the gut microbiota compared with exposed sites. Microbial TLR4 and TLR9 agonists were particularly efficient at inducing IL-23 and IL-1b production by CX

3CR1+

MNPs. The authors also discovered that a gene significantly associated with both ulcerative colitis and Crohns disease in GWAS, TNF-like ligand 1A (TL1A or TNFSF15), is overexpressed in mouse CX

3CR1+ MNPs and synergizes with IL-23 and IL-1b to induce IL-22 production

in both human and mouse ILC3.The study does not completely exclude the role of other cell types in the production of IL-23 and

induction of IL-22 production but in the conditions studied (mouse infection with C. rodentium and human colitis), CX

3CR1+ MNPs were crucial in mediating this mucosal protective loop. In basal

conditions or in other types of inflammatory or preneoplastic conditions, and under stimulation by different microbial TLR agonists (e.g., flagellin), it is quite possible that other cell types, including conventional dendritic cells, may play an important role, as suggested by published studies. However, the study by Longman et al. greatly contributes to our understanding of the mechanisms of human IBD and reassures us that when mouse data are carefully combined with experimental human studies and GWAS, they are powerful in providing mechanistic evidence that explains human pathology.Longman, R.S., et al. 2014. J. Exp. Med. http://dx.doi.org/10.1084/jem.20140678.

Giorgio Trinchieri, National Cancer Institute, National Institutes of Health: trinchig@mail.nih.gov

Critical role for CX3CR1+ mononuclear phagocytes in

intestinal homeostasis

Confocal immunofluorescence image of mouse colon shows juxtaposition (white arrows) of CX3CR1

+ MNPs (green) and RORgt+ ILC3 (red).

Insight from Giorgio Trinchieri

The Rockefeller University Press $30.00J. Exp. Med. 2014 Vol. 211 No. 8 1571-1583www.jem.org/cgi/doi/10.1084/jem.20140678

1571

Article

Inflammatory bowel disease (IBD) has been de-fined as a dysregulated cellular immune response to environmental triggers in genetically predis-posed individuals. Although the initial discovery linking single-nucleotide polymorphisms in the IL23R locus with susceptibility to IBD (Duerr et al., 2006) was consistent with a role for IL-23responsive T cells, more recent evidence supports the importance of IL-23responsive innate lym-phoid cells (ILC) in maintaining epithelial ho-meostasis (Sonnenberg and Artis, 2012). These RORt-dependent ILCs (now named group 3 ILCs, or ILC3 (Spits et al., 2013)) were initially

characterized in mouse models of colitis as pre-dominant producers of IL-22 (Satoh-Takayama et al., 2008), an IL-10 family member that sig-nals via STAT3 to regulate mucosal healing, a critical clinical endpoint in IBD (Pickert et al., 2009; Hanash et al., 2012). In light of their ro-bust production of IL-22 and close proximity to the intestinal epithelial layer (Cella et al., 2009), ILC3 have been proposed to play an important role in mucosal healing and maintenance of barrier integrity, and understanding how they are induced to produce IL-22 has great poten-tial for therapeutic benefit.

CORRESPONDENCE Randy S. Longman: ral2006@med.cornell.edu OR Dan R. Littman: dan.littman@med.nyu.edu

Abbreviations used: CD, Crohns disease; DT, diphtheria toxin; DTR, DT receptor; IBD, inflammatory bowel disease; ILC3, group 3 innate lymphoid cell; MAMP, microbe-associated molecular pattern; MNP, mononuclear phagocyte; PAMP, pathogen-associated molecular pattern; UC, ulcer-ative colitis.

R.S. Longman and G.E. Diehl contributed equally to this paper.J.R. Huhs present address is University of Massachusetts Medical School, Worcester, MA 01605.A. Swaminaths present address is Lennox Hill Hospital, New York, NY 10075.

CX3CR1+ mononuclear phagocytes support colitis-associated innate lymphoid cell production of IL-22

Randy S. Longman,1,3 Gretchen E. Diehl,1 Daniel A. Victorio,1,3 Jun R. Huh,1 Carolina Galan,1 Emily R. Miraldi,1,5,6 Arun Swaminath,4 Richard Bonneau,5,6 Ellen J. Scherl,3,4 and Dan R. Littman1,2

1The Kimmel Center for Biology and Medicine of the Skirball Institute and 2Howard Hughes Medical Institute, New York University School of Medicine, New York, NY 10016

3The Jill Roberts Center for IBD, Department of Medicine, Weill-Cornell Medical College, New York, NY 100214Division of Digestive and Liver Diseases, Department of Medicine, Columbia University Medical Center, New York, NY 100325Center for Genomics and Systems Biology, Department of Biology; and 6Courant Institute of Mathematical Sciences, Computer Science Department, New York University, New York, NY10003

Interleukin (IL)-22producing group 3 innate lymphoid cells (ILC3) promote mucosal heal-ing and maintain barrier integrity, but how microbial signals are integrated to regulate mucosal protection offered by these cells remains unclear. Here, we show that in vivo depletion of CX3CR1+ mononuclear phagocytes (MNPs) resulted in more severe colitis and death after infection with Citrobacter rodentium. This phenotype was rescued by exogenous IL-22, which was endogenously produced by ILC3 in close spatial proximity to CX3CR1+ MNPs that were dependent on MyD88 signaling. CX3CR1+ MNPs from both mouse and human tissue produced more IL-23 and IL-1 than conventional CD103+ dendritic cells (cDCs) and were more efficient than cDCs in supporting IL-22 production in ILC3 in vitro and in vivo. Further, colonic ILC3 from patients with mild to moderate ulcerative colitis or Crohns disease had increased IL-22 production. IBD-associated SNP gene set analysis revealed enrichment for genes selectively expressed in human intestinal MNPs. The product of one of these, TL1A, potently enhanced IL-23 and IL-1-induced production of IL-22 and GM-CSF by ILC3. Collectively, these results reveal a critical role for CX3CR1+ mono-nuclear phagocytes in integrating microbial signals to regulate colonic ILC3 function in IBD.

2014 Longman et al. This article is distributed under the terms of an Attribution NoncommercialShare AlikeNo Mirror Sites license for the first six months after the publication date (see http://www.rupress.org/terms). After six months it is available under a Creative Commons License (AttributionNoncommercial Share Alike 3.0 Unported license, as described at http://creativecommons.org/ licenses/by-nc-sa/3.0/).

1572 CX3CR1+ MNPs regulate ILC3 | Longman et al.

a mouse with the diphtheria toxin receptor (DTR) cDNA in-serted into the Cx3cr1 locus (Diehl et al., 2013). Analysis of colonic lamina propria mononuclear cells (LPMCs) after infec-tion of DT-treated mice revealed a reduction in the percentage of CD11c+ MHCII+ LPMCs (Fig. 1 A), which reflected a pref-erential loss of the CX3CR1+ CD11b+ CD14+ fraction of MNPs (Fig. 1 B; Tamoutounour et al., 2012), as well as CX3CR1+ monocytes in Cx3cr1DTR/+ mice compared with control mice. CD103+ CD11b+ cDCs were not depleted (Fig. 1 C). To induce colitis, mice were infected with C. rodentium, a mouse model for infectious colitis (Zheng et al., 2008; Sonnenberg et al., 2011; Qiu et al., 2012). DT-treated infected Cx3cr1DTR/+ mice, but not uninfected or control infected mice, lost more weight (Fig. 1 D), displayed more severe intestinal pathology (Fig. 1 E), and ultimately succumbed to infection (Fig. 1 F). Infected Cx3cr1DTR/+ mice also had increased bacterial burden in the spleen, consistent with the loss of barrier integrity (Fig. 1 G).

To examine potential involvement of signaling pathways for receptors of pathogen- or microbe-associated molecular patterns (PAMPs or MAMPs) in mediating this phenotype, mice with a conditional deletion of MyD88 in CD11c-expressing MNPs (CD11c-Cre/Myd88fl/fl) were infected with C. roden-tium. Infection of CD11c-Cre/Myd88fl/fl mice, but not litter-mate controls, was lethal by 15 d after infection (Fig. 2 A), implicating PAMP/MAMP signaling as having a critical role in barrier protection mediated by CD11c-expressing MNPs. The C. rodentium colitis model depends on IL-22 for protec-tion (Zheng et al., 2008). Thus, to test if exogenous IL-22 could rescue the susceptibility phenotype described above, CD11c-Cre/Myd88fl/fl (Fig. 2 A) and DT-treated CX3CR1-DTR (Fig. 2 B) mice were hydrodynamically injected with a plasmid encoding IL-22 (Qiu et al., 2012). The exogenous IL-22 rescued both lines of mice from colitis-induced death.

Colonic CX3CR1+ MNPs regulate ILC3 production of IL-22High-dose infection with C. rodentium is controlled by ILC3, which represents the large majority of LPMCs producing IL-22 (Sonnenberg et al., 2011). At day 7 after infection, both the percentage and absolute number of IL-22+ colonic, lineage, CD90hi, and RORt+ ILCs (Fig. S1 shows gating strategy) from mice depleted for CX3CR1+ cells were reduced in com-parison to ILCs from mice with intact CX3CR1+ cells (Fig. 2, CE). Depletion of CX3CR1+ cells did not affect the abso-lute number of ILC3 (Fig. 2 F). Although T cells can also con-tribute to IL-22 production in low-dose C. rodentium infection (Basu et al., 2012), no statistically significant difference in the total IL-22+ (or IL-17+) T cells was noted in mice depleted for CX3CR1+ cells (Fig. 2 G). To assess the ability of colonic CX3CR1+ MNPs to interact with ILC3 within the colonic tis-sue, Cx3cr1GFP/+ mice were used to visualize CX3CR1+ MNPs in situ ( Jung et al., 2000). Consistent with the ability of these cells to regulate ILC3 function, confocal microscopy revealed the spatial proximity of RORt+ ILC3 cells with CX3CR1+ MNPs in the colonic lamina propria (Fig. 2 H).

To evaluate the ability of intestinal CX3CR1+ cells and cDCs to support ILC3 activation, CX3CR1+ (GFP+) cells and

Mononuclear phagocytes (MNPs) are sentinels of the intes-tinal lamina propria, capable of responding to microbial prod-ucts, and play a crucial role in orchestrating intestinal lymphocyte homeostasis. MNPs can be subdivided based on their expression of CD103 or CX3CR1, and each group has been ascribed criti-cal functions in maintaining intestinal homeostasis (Bogunovic et al., 2009; Merad et al., 2013). CD103+ cells, which can be further subdivided based on the expression of CD11b, differen-tiate from a common DC precursor and are thought to be the conventional, migratory myeloid DCs (Varol et al., 2010). CD103+ CD11b DCs require Irf8, Id2, and Batf3 for their de-velopment and are thought to play a critical role in cross-priming virus- and tumor-specific CTLs (Hildner et al., 2008; Merad et al., 2013). Loss of these cells, however, does not alter intestinal T cell homeostasis or lead to spontaneous inflammation (Edelson et al., 2010). CD103+CD11b+ DCs, in contrast, require Notch2 signaling, produce IL-23 in response to flagellin-induced TLR5 activation, resulting in IL-22 production by ILC3, and have ad-ditionally been proposed to support Th17 polarization (Lewis et al., 2011; Kinnebrew et al., 2012). These cells can produce retinoic acid, which promotes the expression of the gut-homing receptor CCR9 and synergizes with TGF to induce regulatory T cells (Sun et al., 2007). One recent study suggests that Notch2-dependent CD103+ CD11b+ DCs regulate protection from C. rodentiuminduced colitis (Satpathy et al., 2013). However, specific depletion of CD103+ CD11b+ intestinal DCs revealed that these cells are not the MNP subset required for protection against C. rodentium or IL-22 production (Welty et al., 2013).

In contrast to CD103+ cDCs, CX3CR1+ MNPs differenti-ate from monocyte precursors (Varol et al., 2010). Although these cells were previously thought to be tissue-resident and to promote local Treg differentiation (Hadis et al., 2011), recent data from our group showed that they can up-regulate CCR7 and migrate to secondary lymphoid organs, suggesting a broader role in orchestrating immunity (Diehl et al., 2013). Notably, we ob-served that interaction with microbiota limits the migration of these cells to mesenteric LNs (MLNs; Diehl et al., 2013), and an increase in CX3CR1+ cells has been described in the lamina propria during mouse (Zigmond et al., 2012) and human colitis (Kamada et al., 2008). A recent study reported that fractalkine receptor (CX3CR1) expression supports innate celldependent clearance of C. rodentium infection (Manta et al., 2013), but a functional role for CX3CR1+ MNPs in regulating colitis-associated ILC3 remains unclear. To evaluate this question, we employed novel mouse models to enable selective depletion of CX3CR1+ MNPs in vivo. Our results reveal a critical role for CX3CR1+ MNPs from both mouse and human tissue in sup-porting IL-22 induction in ILC3 in vitro and in vivo. Moreover, we identify the ability of TL1A produced by MNPs to potently enhance IL-23 and IL-1induced production of IL-22 and GM-CSF by ILC3.

RESULTSCX3CR1+ cells protect against C. rodentiuminduced colitisTo investigate the role of the expanded population of CX3CR1+ cells in the intestinal lamina propria during colitis, we generated

JEM Vol. 211, No. 8 1573

Article

express DTR only upon cre-mediated deletion of a LoxP-Stop cassette inserted into the Cx3cr1 locus (Diehl et al., 2013). In-jection of DT resulted in a selective loss of Ly6Clo MHCIIhi MNPs, which express both intermediate and high levels of CX3CR1-GFP, and spared the Ly6Chi monocytes, which express intermediate levels of CX3CR1 (Fig. 4 A; Diehl et al., 2013). Similar to results with CX3CR1DTR/+ mice, in which both monocytes and MNPs were ablated, depletion of CX3CR1+ CD11c+ cells led to reduction in colitis-induced ILC3 production of IL-22 (Fig. 4 B).

Because both IL-23 and IL-1 regulate ILC3, we wished to determine the contribution of these cytokines to the ob-served regulation of IL-22 production by MNPs. LPS and CpG stimulation induced markedly increased IL-23 and IL-1 pro-duction by CX3CR1+ MNPs compared with CD103+ CD11b+

CD103+ CD11b+ DCs were sorted from the lamina propria of Cx3cr1GFP/+ mice (Fig. 3 A) and co-cultured with intestinal ILCs. TLR-stimulated CX3CR1+ cells were markedly more efficient than CD103+ CD11b+ DCs in supporting IL-22 pro-duction (Fig. 3, B and C). CX3CR1+ cells in the LP include Ly6Chi monocytes and Ly6Clo MHCIIhi MNPs (Fig. 3 D; Zigmond et al., 2012; Diehl et al., 2013). To evaluate the role of these distinct CX3CR1+ populations in supporting ILC3 func-tion, we sorted CX3CR1+ monocytes and CX3CR1+ MNPs and co-cultured them with ILCs in the presence of TLR stim-uli. TLR-stimulated MNPs were much more potent inducers of IL-22 than monocytes (Fig. 3, E and F). In an effort to confirm the functional potential of MNPs compared with monocytes in vivo, we selectively ablated CD11c-expressing CX3CR1+ MNPs. CD11c-cre mice were bred to mice engineered to

Figure 1. Intestinal CX3CR1+ cells pro-tect mice from C. rodentium-induced colitis. (A and B) Depletion efficiency in the Cx3cr1DTR/GFP mice was assessed by flow cytometry of small intestinal lamina propria cells Cx3crDTR/GFP or littermate control mice after administration of DT to both groups daily for 2 d. (A) Surface staining for CD11b versus CX3CR1-GFP. (B) MHCII+ CD11c+ cells were assessed for expression of CX3CR1, CD103, and CD14. Results are representative of five independent experiments with a mini-mum of three animals per group. (C) Total number of MHCII+ CD11c+CD103+ (left) and MHCII+ CD11c+CX3CR1+ cells (right) per intes-tine as determined by flow cytometry analy-sis. n.s., P > 0.05; **, P 0.01. Two-tailed Students t test. Error bars represent the SEM. Results are representative of five independent experiments with a minimum of 3 animals per group. (D) Weight of DT-treated littermate WT (control) mice or Cx3cr1DTR/+ mice following infection with C. rodentium (n = 7 mice/group). DT was administered at days 2, 1, and 0 and every other day after infection. Data are representative of two independent experiments. (E) Representative colonic histol-ogy from littermate control mice or Cx3cr1DTR/+ mice (analyzed in D) infected with C. rodentium at day 7 after infection.

1574 CX3CR1+ MNPs regulate ILC3 | Longman et al.

(CD, n = 8) or ulcerative colitis (UC, n = 6; Table S1) as well as age-matched non-IBD control patients undergoing routine screening colonoscopy (n = 8). Analysis of intracellular cytokine production revealed significantly increased IL-22 production in the CD3 fraction of colonic LPMCs from sites of colonic inflammation in both CD and UC compared with the non-IBD controls (Fig. 5 A and Fig. S2). In contrast, IL-22 production by T cells was not significantly different between the groups. Further characterization of the nonT cells producing IL-22 revealed that a large fraction of these cells expressed c-Kit and CD56, markers of ILCs (Cella et al., 2009; Fig. 5 B). Consis-tent with their being ILC3, these cells were lineage negative (Lin; Fig. S3 A); expressed RORt (Fig. 5 C); were CD45int CD127+ (Fig. 5 D); expressed CD161, NKp44, and CCR6 (Fig. 5 E), which are phenotypic surface markers of ILC3; and produced IL-22 in response to IL-23 stimulation (Fig. 5 F). As

DCs in vitro (Fig. 4, C and D). To evaluate the role of MNP-derived IL-23 and IL-1, co-cultures were performed with in-testinal CD11c+ cells derived from WT or Il23p19/ mice and ILCs from WT or Il1r/ mice. IL-23deficient MNPs and IL1R-deficient ILCs yielded significantly reduced production of IL-22 (Fig. 4 E), consistent with the importance of MNP-derived IL-1 and IL-23 in supporting IL-22 production. These data reveal a mechanistic role for IL-23 and IL-1 pro-duced by colonic MNPs in supporting colitis-associated ILC3 secretion of IL-22.

Human intestinal ILC3 production of IL-22 is regulated by microbial stimulation of MNPsTo evaluate the regulation of intestinal ILC3 in humans with IBD, we prepared LPMCs from descending colon biopsies of patients with endoscopically mild to moderate Crohns disease

Figure 2. CX3CR1+ cells support colonic ILC3 pro-duction of IL-22. (A) Survival curves of C. rodentiuminfected Myd88f/f littermate controls (n = 10, open circle) as compared with CD11c-cre/Myd88f/f mice without (n = 13, filled circle) or with (n = 5, open triangle) exog-enous hydrodynamic delivery of an IL-22producing plasmid. Results are a composite of two independent experiments. (B) Survival curves of DT-treated Cx3cr1DTR/+ mice infected with C. rodentium after hydrodynamic delivery of a plasmid expressing IL-22 (n = 8) or control vector (n = 9). DT was administered at days 2, 1, and 0 and every other day after infection. Results are a com-posite of three independent experiments. (CE) Percent-age (C and D) and total number (E) of colonic Lin CD90.2+ ILCs producing IL-22 from DT-treated Cx-3cr1DTR/+ (n = 10) and littermate control mice (n = 9) 7 d after C. rodentium infection and from uninfected mice (NT; n = 3). Results are a composite of two independent experiments. DT was administered at days 2, 1, and 0 and every other day after infection. Intracellular IL-22 was assayed by flow cytometry after 4-h culture. A represen-tative flow cytometry plot from each group is shown in C. **, P 0.01. One way ANOVA with Bonferronis correction. Error bars represent the SEM. (F) Total number of ILC3 per colon in Cx3cr1DTR/+ (n = 9) or control (n = 8) mice administered DT. Error bars represent the SEM. Results are one of three representative experiments. (G) Total number of IL-17+ or IL-22+ colonic CD4+ T cells from DT-treated Cx3cr1DTR/+ (n = 10) and littermate con-trol mice (n = 9) 7 d after C. rodentium infection and from uninfected mice (n = 3). Intracellular IL-22 and IL-17 was assayed by flow cytometry after 4-h culture. *, P 0.05. One way ANOVA with Bonferronis correction. Error bars represent the SEM. Results are a composite of two independent experiments. (H) Confocal immuno-fluorescence of colonic samples from Cx3Cr1GFP/+ mice stained for CD3 and RORt. Bar, 10 m. White arrows indicate sites of MNP and ILC3 juxtaposition.

JEM Vol. 211, No. 8 1575

Article

Parallel subpopulations of CD11c+ MNPs present in the mouse intestine similarly exist in the human intestine (Fig. 6 A; Merad et al., 2013). To evaluate whether these distinct sub-populations of MNPs from human intestinal tissue functioned similarly, we examined the phenotypic properties of CD103+ DCs and CD14+ MNPs (which express CX3CR1; Kamada et al., 2008) within the CD11c+ MHCII+ fraction of LPMCs (Fig. 6 B). In contrast to CD103+ DCs, CD14+ MNPs ex-pressed CD64 as well as higher levels of CD86. Consistent with the phenotypic characterization of these subsets, transcriptional analysis of these populations by RNA-seq revealed higher levels of CLEC9A, XCR1, and CD207 expression in the CD103+ cells, whereas MERTK, STAB1, and CX3CR1 were higher in the CD14+ cells (Fig. 6 C). We tested the potential

Myd88 deficiency abrogated ILC3 production of IL-22, we hypothesized that signals from the microbiota could induce ILC3 to produce IL-22. To investigate this in human tissue, we evaluated three patients who had a surgical diversion of the fecal stream (i.e., a diverting ostomy) as part of their therapy for IBD. Endoscopic biopsies were taken from a site proximal to the diversion (afferent limb), where the mucosa was exposed to intestinal microbiota in the fecal stream, and from mucosa distal to the diversion (efferent limb), that was unexposed to the fecal stream. ILCs were present at both mucosal locations, but not in PBMCs from the same donor (Fig. S3 B). In all three donors, ILCs from tissue exposed to bacteria in the fecal stream produced substantially more IL-22 compared with ILCs from unexposed tissue (postdiversion; Fig. 5 G).

Figure 3. TLR-stimulated CX3CR1+ MNPs are stronger inducers of ILC3 production of IL-22 than CD103+ CD11b+ DCs and monocytes. (AC) CD103 or CX3CR1+ MHCII+ CD11c+ CD11b+ cells were isolated from the lamina propria of CX3CR1GFP/+ mice (sort strat-egy shown in A and co-cultured with Lin RORt-GFP+ ILCs with or without the indi-cated bacterial TLR ligands or IL-23. IL-22 was assessed by intracellular staining of CD90.2+ ILCs after 18 h. A representative flow cytom-etry plot is shown in B. (C) Percent IL-22+ ILCs is shown from seven independent experi-ments. **, P 0.01; *, P 0.05. One way ANOVA with Bonferronis correction. Error bars represent SEM. (DF) Ly6C+ MHCIIlo (monocytes) and Ly6C MHCIIhi (MNPs) were isolated from CX3CR1+ CD11b+ lamina propria cells (sort strategy is shown in D and co- cultured with intestinal ILCs with LPS or IL-23 as indicated. Intracellular cytokine staining for IL-22 is shown after 18 h (E). Supernatants were harvested after 18 h and IL-22 produc-tion quantitated by ELISA (F). Results are rep-resentative of two independent experiments performed in triplicate. ***, P 0.001. One-way ANOVA with Bonferroni correction. Error bars represent the SEM.

1576 CX3CR1+ MNPs regulate ILC3 | Longman et al.

the expanded population of CD14+ MNPs (Kamada et al., 2008) in supporting colitis-associated IL-22 production by ILC3, through their secretion of IL-23 and IL-1.

CX3CR1+ MNP-derived TL1A synergizes with IL-23 and IL-1 to induce IL-22We hypothesized that MNP-derived factors in addition to IL-23 and IL-1 would contribute to the regulation of ILC3 function. The RNA-seq data from CD14+ human MNPs re-vealed a significant enrichment for genes associated with IBD in GWAS studies, suggesting that IBD-associated pathways are im-portant in these cells (Fig. 7 A and Table S2). Notably, we iden-tified TNF-like ligand 1A (TL1A, also designated TNFSF15) as a significant contributor to the IBD GWAS-derived gene set enrichment. Evaluation of Tnfsf15 transcript in sorted mouse colonic APC subsets by qPCR confirmed higher ex-pression in CX3CR1+ MNPs (Fig. 7 B). DR3/TNFRSF25,

of these subsets to induce IL-22 production by co-culturing TLR-stimulated CD14+ MNPs and CD103+ DCs from human intestinal resections with intestinal ILCs. Intracellular cyto-kine staining at 18 h revealed that the CD14+ MNP were more effective than the CD103+ DCs at stimulating IL-22 production by ILCs (Fig. 6 D). Neither cell population induced signifi-cant IL-17 or IFN- production by ILCs (Fig. 6, D and E).

Consistent with the importance of IL-23 and IL-1 in the mouse co-culture experiments, a higher level of IL23A expression was observed by RNA-seq in CD14+ MNPs com-pared with CD103+ DCs (Fig. 6 C). Stimulation of these human MNP subsets with LPS or flagellin revealed increased IL-23p19 mRNA and IL-1 protein produced by the CD14+ cells com-pared with the CD103+ cells (Fig. 6 F). In accord with this finding, in co-cultures of human intestinal ILC3 and CD14+ MNPs, IL-23 and IL-1 antibody blockade blunted IL-22 production (Fig. 6 G). These data reveal a mechanistic role for

Figure 4. CX3CR1+ MNP-derived IL-23 and IL-1 activate ILC3 to produce IL-22. (A) Phenotype analysis of colonic LPMCs from Cx3cr1STOP-DTR/GFP mice with or without CD11c-cre after DT injection for 2 d. (top) Selective depletion of CX3CR1hi MNPs. (bottom) Expression of Ly6C and MHCII on CX3CR1hi and CX3CR1int populations. (B) Expression of IL-22 in Lin CD90.2+ colonic ILCs from Cx3cr1STOP-DTR/+ (Stop-DTR) or CD11c-Cre x Cx3cr1STOP-DTR/+ (Cre-DTR) mice at 7 d after C. rodentium infection. DT was administered at days 2, 1, and 0 and every other day postinfection. One representative intracellular cytokine flow cytometry plot is shown on the left and a composite graph (n = 6/group) on the right. *, P 0.05, two-tailed Students t test. Error bars represent the SEM. Results are a composite of two independent experiments. (C) Supernatants from APC-ILC co-cultures (Fig. 3, B and C) were harvested after 18 h and assayed for IL-23 by ELISA. Results are averages of three independent experiments and the SEM is shown. (D) CX3CR1+ MNPs or CD103+ CD11b+ DCs were sorted and incubated with media or CpG for 18 h and supernatants were assayed for IL-1 by ELISA. Results are the mean of two independent experiments performed in duplicate and the SEM is shown. *, P 0.05; **, P 0.01. (E) Lin CD90.2hi ILCs from WT or Il1r/ mice were co-cultured with sorted intestinal MNPs from WT or Il23p19/ mice, with or without CpG, as indicated. IL-22 production by the ILCs was assessed after 18 h by ELISA. Data are combined from three independent experiments performed in duplicate. *, P 0.05; ***, P 0.001. One-way ANOVA with Bonferroni correction. Error bars represent the SEM.

JEM Vol. 211, No. 8 1577

Article

with LPS-stimulated CX3CR1+ MNPs. These data reveal a potent ability of TL1A to enhance IL-22 production via DR3/TNFRSF25 on ILC3 in both mouse and human.

DISCUSSIONMononuclear phagocytes are spatially and functionally poised to integrate microbial signals from the luminal microbiota (Niess et al., 2005; Varol et al., 2010). Although MNPs were previously thought to remain in the tissue, we recently showed that CX3CR1+ MNPs can migrate to draining lymph nodes and initiate immune responses under conditions of dysbiosis (Diehl et al., 2013). In the context of inflammation, however, these CX3CR1+ MNPs expand within the lamina propria dur-ing chemical (Zigmond et al., 2012) or infectious colitis and in IBD patients (Kamada et al., 2008), and their function has remained obscure. Conventional DCs (cDCs), rather than the MNPs, have been postulated to regulate intestinal Th17 cell differentiation in response to microbiota (Denning et al., 2011;

the receptor for TL1A, is expressed both on T cells and ILC subsets, with the highest expression on ILC2 and ILC3 (Meylan et al., 2014). Ex vivo stimulation with mouse or human re-combinant TL1A significantly enhanced the ability of IL-23 and IL-1 to induce IL-22 by both mouse (Fig. 7 C and Fig. S4) and human (Fig. 7 D) intestinal ILC3, respectivel