Constitutive Expression of CCL22 Is Mediated by T Cell−Derived GM-CSF Ignazio Piseddu, Natascha Röhrle, Maximilian Martin Ludwig Knott, Stefan Moder, Stephan Eiber, Konstantin This information is current as Schnell, Viola Vetter, Bastian Meyer, Patrick Layritz, of September 24, 2021. Benjamin Kühnemuth, Gabriela Maria Wiedemann, Juliane Gruen, Carolin Perleberg, Moritz Rapp, Stefan Endres and David Anz

J Immunol published online 9 September 2020 Downloaded from http://www.jimmunol.org/content/early/2020/09/08/jimmun ol.2000004

Supplementary http://www.jimmunol.org/content/suppl/2020/09/08/jimmunol.200000 http://www.jimmunol.org/ Material 4.DCSupplemental

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists by guest on September 24, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

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

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2020 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published September 9, 2020, doi:10.4049/jimmunol.2000004 The Journal of Immunology

Constitutive Expression of CCL22 Is Mediated by T Cell–Derived GM-CSF

Ignazio Piseddu,*,1 Natascha Ro¨hrle,*,1 Maximilian Martin Ludwig Knott,* Stefan Moder,* Stephan Eiber,* Konstantin Schnell,* Viola Vetter,* Bastian Meyer,* Patrick Layritz,* Benjamin Ku¨hnemuth,* Gabriela Maria Wiedemann,† Juliane Gruen,* Carolin Perleberg,* Moritz Rapp,* Stefan Endres,* and David Anz*,‡

CCL22 is a key mediator of leukocyte trafficking in inflammatory immune responses, allergy, and cancer. It acts by attracting regulatory T cells and Th2 cells via their receptor CCR type 4 (CCR4). Beyond its role in inflammation, CCL22 is constitutively expressed at high levels in lymphoid organs during homeostasis, where it controls immunity by recruiting regulatory T cells to

dendritic cells (DCs). In this study, we aimed to identify the mechanisms responsible for constitutive CCL22 expression. Downloaded from We confirmed that CD11c+ DCs are the exclusive producers of CCL22 in secondary lymphatic organs during homeostasis. We show that in vitro both murine splenocytes and human PBMCs secrete CCL22 spontaneously without any further stimulation. Interestingly, isolated DCs alone, however, are unable to produce CCL22, but instead require T cell help. In vitro, only the coculture of DCs with T cells or their supernatants resulted in CCL22 secretion, and we identified T cell–derived GM-CSF as the major inducer of DC-derived CCL22 expression. In vivo, Rag12/2 mice, which lack functional T cells, have low CCL22 levels in

lymphoid organs, and this can be restored by adoptive transfer of wild-type T cells or administration of GM-CSF. Taken together, http://www.jimmunol.org/ we uncover T cell–derived GM-CSF as a key inducer of the CCL22 and thus, to our knowledge, identify a novel role for this as a central regulator of immunity in lymphatic organs. This knowledge could contribute to the development of new therapeutic interventions in cancer and autoimmunity. The Journal of Immunology, 2020, 205: 000–000.

he chemokine CCL22 is induced upon inflammation (1) the inflammatory process (6). In cancer, CCL22 is expressed in the and is constitutively expressed under homeostatic condi- tumor tissue of many human malignancies, where it contributes to T tions (2, 3). So far, the only known receptor for CCL22 the recruitment of Tregs and has thus been linked to accelerated is CCR type 4 (CCR4), which is mainly expressed by activated tumor progression and unfavorable outcome (7–9). Further, CCL22 Th2-biased CD4+ T cells (4) and by regulatory T cells (Tregs) (5) is upregulated in contact-sensitized skin, where it promotes the by guest on September 24, 2021 under homeostatic conditions. In the settings of asthma and atopic activation of CD4+ T cells (10). dermatitis, CCL22 levels are elevated in the lungs and the skin, In inflammatory immune responses, various CCL22-inducing respectively, and induce migration of Th2 cells that maintain pathways have already been described. In and den- dritic cells (DCs), CCL22 expression can be induced by IL-4 and IL-13, which plays an important role in the pathogenesis of asthma *Center of Integrated Science Munich, Division of Clinical Pharmacology, Department of Internal Medicine IV, University Hospital of Munich, 80337 Munich, and atopic diseases (3). In cancer, tumor cell–derived IL-1a up- Germany; †Department of Medicine II, University Hospital Rechts der Isar, Technical regulates CCL22 in tumor-infiltrating DCs, which leads to Treg University of Munich, 81675 Munich, Germany; and ‡Department of Internal Med- icine II (Gastroenterology and Hepatology), University Hospital of Munich, 81377 recruitment (11). In addition, CCL22 in macrophages and DCs is Munich, Germany strongly induced by LPS, suggesting a regulatory role in bacterial 1I.P. and N.R. contributed equally. infection (1, 3, 12). ORCIDs: 0000-0001-8331-9778 (I.P.); 0000-0001-5767-3386 (N.R.); 0000-0002- In addition to its inducible secretion during inflammation, CCL22 6995-3702 (M.M.L.K.); 0000-0002-4096-6577 (K.S.); 0000-0001-5104-7683 (M.R.). is expressed constitutively at high levels in the lymph node and the Received for publication January 3, 2020. Accepted for publication August 6, 2020. thymus. We could recently show that DC-derived CCL22 promotes The work was supported by German Research Foundation Grant DFG AN 801/2-1 cellular contacts with Tregs in the lymph node, representing a central (to D.A. and S.E.), Wilhelm Sander-Stiftung Grant 2016.028.1 (to D.A. and B.K.), regulator of T cell immunity (13). Further, constitutive CCL22 Deutsche Krebshilfe Grant 111326 (to D.A.), and the Fo¨rderprogramm fu¨r Forschung und Lehre of the Ludwig-Maximilians-Universita¨t Munich (to I.P., B.M., N.R., V.V., expression may control immune homeostasis by retaining Tregs and M.M.L.K.). in the lymph node and by enabling thymocyte trafficking from Data acquisition: I.P., N.R., M.M.L.K., S.M., S.E., V.V., K.S., B.M., P.L., and J.G. thymic cortex to medulla (14, 15). In the pancreas, CCL22 was Writing, review, and revision of the manuscript: I.P., B.K., C.P., N.R., M.M.L.K., reported to protect from autoimmune b cell destruction through G.M.W., and D.A. Conception and design: I.P., B.K., M.M.L.K., M.R., S.E., and D.A. the recruitment of Tregs in the setting of type-1 diabetes (16). Address correspondence and reprint requests to Dr. David Anz, Division of Further, it was suggested that CCL22 expression in the spleen may Clinical Pharmacology and Department of Internal Medicine II, Klinikum der Universita¨tMu¨nchen, Ziemssenstraße 1, 80336 Mu¨nchen, Germany. E-mail play a role in maintaining tolerance during apoptotic cell clear- address: [email protected] ance (17). Taken together, CCL22 is constitutively expressed at The online version of this article contains supplemental material. high levels in lymphoid organs and serves as an important regu- Abbreviations used in this article: CCR4, CCR type 4; DC, dendritic cell; Non-DC, lator of adaptive immunity. DC-depleted cell; qRT-PCR, quantitative real-time PCR; Treg, regulatory T cell; Whereas the mode of CCL22 induction in the settings of allergy WT, wild-type. and cancer is, in part, understood (e.g., induction by inflammatory Copyright Ó 2020 by The American Association of Immunologists, Inc. 0022-1767/20/$37.50 such as IL-4, IL-13, or IL-1a), the mechanisms leading

www.jimmunol.org/cgi/doi/10.4049/jimmunol.2000004 2 T CELL–DERIVED GM-CSF MEDIATES CONSTITUTIVE CCL22 EXPRESSION to the constitutive secretion of CCL22 in lymphoid organs were so RNA isolation and quantitative real-time PCR far unknown. RNA was isolated from single-cell suspensions using Direct-zol RNA In this study, we demonstrate that both murine splenocytes and MiniPrep Plus Kit (Zymo Research) according to manufacturer’s instruc- human PBMCs secrete CCL22 spontaneously in vitro. Isolated tions. One microgram of RNA was converted into cDNA using the DCs, however, are unable to produce CCL22 and require help from RevertAid First Strand cDNA Synthesis Kit (Fermentas, St. Leon-Rot, T cells. We show that GM-CSF, which is secreted by a subset of Germany). Quantitative real-time PCR (qRT-PCR) was performed using the LightCycler 480 II (Roche Diagnostics, Basel, Switzerland). cDNA T cells (18), is the major driver for CCL22 secretion in DCs. These copies of different mRNAs were quantified and normalized to mRNA of a findings are confirmed using rGM-CSF, GM-CSF–blocking Abs reference (Actb). qRT-PCR primers (IL-3: forward 59-TCTACATGGTC- and T cells derived from GM-CSF–deficient (Csf22/2) mice. The CACCTTAACGA-39, reverse 59-AGAGACGGAGCCAGATGC-39; IL-5: importance of T cell help for constitutive CCL22 expression forward 59-ACATTGACCGCCAAAAAGAG-39, reverse 59-ATCCAGG- 2/2 AACTGCCTCGTC-39; GM-CSF: forward 59-GCATGTAGAGGCCATC- is further demonstrated in vivo using Rag1 mice that were AAAGA-39, reverse 59-CGGGTCTGCACACATGTTA-39; IL-7: forward adoptively transferred with wild-type (WT) T cells. Our findings 59-CGCAGACCATGTTCCATGT-39, reverse 59-TCTTTAATGTGGCAC- uncover a so-far unknown role of GM-CSF and GM-CSF–producing TCAGA-39; CCL22: forward 59-TCTTGCTGTGGCAATTCAGA-39, reverse T cells in the regulation of immunity, presenting new opportunities 59-GAGGGTGACGGATGTAGTCC-39; and Actb: forward 59-CTAAGG- for therapeutic interventions in cancer and autoimmunity. CCAACCGGGAAAAG-39, reverse 59-ACCAGAGGCATACAGGGACA-39) were designed using Roche Universal Probe Library and purchased from Metabion (Munich, Germany). Materials and Methods Mice and cell culture Tissue lysates and ELISA Downloaded from WT C57BL/6 mice were purchased from Janvier Labs. Mice deficient Organs were frozen in liquid nitrogen, mechanically disrupted, and sub- for Rag1 (Rag12/2) were purchased from Charles River Laboratories sequently lysed using Bio-Plex Cell Lysis Buffer (Bio-Rad Laboratories). 2 2 . 3 (Wilmington, MA). GM-CSF–deficient (Csf2 / ) mice were provided by The lysates were incubated on ice for 15 min and centrifuged at 13.000 g Prof. Dr. M. Kopf (ETH Zurich, Switzerland). CCL22-deficient (CCL222/2) for 15 min at 4˚C. Total protein levels were determined using the DC Protein mice were retrieved from the National Institutes of Health–founded Knockout Assay Kit following manufacturer’s instructions. CCL22 protein levels in Mouse Project. All mice used were on C57BL/6 background, around 8 wk old, organ lysates and cell culture supernatants were determined by ELISA and age matched at the onset of the experiments. The animal studies were (R&D Systems) according to manufacturer’s instructions. CCL22 levels approved by the local regulatory agency (Regierung von Oberbayern, were calculated as pictograms of CCL22 per milligram of protein in the http://www.jimmunol.org/ Munich, Germany). Primary murine immune cells were cultured in RPMI respective organ lysate. 1640 medium supplemented with 10% FCS, 100 IU/ml penicillin, 100 mg/ml Murine serum was gained by adding of 500 ml blood to 25 IU heparin 3 streptomycin, 2 mM L-glutamine, 1 mM sodium pyruvate, and 1% MEM– with subsequent centrifugation at 1.000 g for 10 min. CCL22 serum nonessential amino acids solution. For murine T cell medium, 1% MEM– protein levels were determined by ELISA (R&D Systems) according to nonessential amino acids solution was replaced by 1% HEPES. To neutralize manufacturer’s instructions. the effect of cytokines on the production of CCL22, 1 3 106 WT splenocytes were incubated with either 0.2 mg/ml of rat IgG1k isotype control Ab Flow cytometry (BioLegend) or 0.2 mg/ml of cytokine-neutralizing Abs alone or in com- CD34-PE/Cy7 (clone: HM34), B220-APC (clone: RA3-6B2), CD3-FITC bination. Anti–IL-2, anti–IL-3, and anti–GM-CSF Abs were purchased (clone: 17A2), CD11c-PE (clone: N418), and Zombie UV Fixable Via- from R&D Systems; anti–IL-4 and anti–IL-5 Abs were obtained from

bility Kit were all from BioLegend. Counting beads were purchased from by guest on September 24, 2021 BioLegend. rGM-CSF, both murine and human, was purchased from Invitrogen. Events were measured on an FACSCanto II and an FACS Miltenyi Biotec (Munich, Germany). rG-CSF as well as rM-CSF, both LSRFortessa flow cytometer (BD Biosciences) and analyzed with FlowJo murine and human, were purchased from PeproTech (Hamburg, Ger- software (BD Biosciences). many). Recombinant cytokines for splenic DC stimulation (Fig. 3A) were all purchased from PeproTech. Statistical analysis Cell sorting Statistical analysis has been performed using GraphPad Prism Version 5.01. Significance was determined by two-tailed Student t test if not indicated Human DCs and T and B cells were isolated from PBMCs obtained by differently. Error bars show SEM. The p values are as follows: *p , 0.05, density gradient centrifugation using BioColl separating solution (Bio- **p , 0.01, ***p , 0.001, and ****p , 0.0001. chrom). Subsequently, blood myeloid and plasmacytoid DCs and T and B cells were purified by magnetic cell sorting according to manufacturer’s instructions (Miltenyi Biotec, Bergisch Gladbach, Germany): human BDCA1+, Results + + BDCA3 , and BDCA4 cells (Blood Dendritic Cell Isolation Kit II, human), Constitutive CCL22 expression by DCs is mediated by T cells T cells (Pan T Cell Isolation Kit II, mouse), CD3+ cells (CD3 MicroBeads, human), and CD19+ cells (CD19 MicroBeads, human). Human T and In the lymph node and the spleen, CCL22 is known to be expressed B cells were purified using the untouched T cell isolation kit first and, constitutively at high levels exclusively by DCs (13). In analogy, subsequently, a positive selection kit targeting CD3 and CD19, respec- freshly isolated murine immune cells spontaneously secrete CCL22 tively. Human CD4+ and CD8+ T cells were purified by immunomagnetic negative selection according to the manufacturer’s instructions (STEMCELL without any stimulation in vitro (13). In this study, we show that Technologies): human CD4+ T cells (EasySep Human CD4+ T Cell Isolation also human PBMCs exhibit a potent spontaneous CCL22 secretion Kit) and CD8+ T cells (EasySep Human CD8+ T Cell Isolation Kit). over a period of 3 d, similar to murine splenocytes (Fig. 1A). To Murine splenocytes were obtained by mechanical disruption of spleens confirm that DCs are the source of CCL22 secretion in vitro, we using 30-mm cell strainers (Miltenyi Biotec). Distinct immune cell pop- + + + + ulations were isolated using the designated magnetic cell sorting kits from isolated CD11c murine and BDCA1 , BDCA3 , and BDCA4 Miltenyi Biotec following manufacturer’s instructions: murine T cells (Pan human DCs by magnetic cell sorting and incubated purified DCs T Cell Isolation Kit II, mouse), CD3+ (CD3e MicroBeads, mouse), CD4+ and DC-depleted cells (Non-DCs) separately. Surprisingly, no (CD4+ T Cell Isolation Kit, mouse), CD8+ (CD8a MicroBeads, mouse), + + relevant CCL22 secretion was detectable in either isolated DCs or CD45R/B220 (CD45R [B220] MicroBeads, mouse), and CD11c (CD11c Non-DCs even after several days of culture. However, coculture of MicroBeads, mouse). Purity of all cell populations, as determined by flow cytometry, was DCs with Non-DCs completely restored CCL22 secretion to the on average above 90%. For adoptive transfer of T cells into Rag12/2 level known from splenocytes or PBMCs, respectively (Fig. 1B). 2 2 mice, WT or CCL22 / T cells were generated from splenocytes. In We therefore hypothesized that DCs may require help from non- brief, splenocytes were incubated with anti-CD3e (1:1000) and anti- myeloid cells to be able to secrete CCL22. Indeed, cocultures of CD28 (1:10,000) Abs (eBioscience) and IL-2 (Novartis) for 24 h. Prior to T cell expansion, anti-CD3 and anti-CD28 beads (1:100) (Life DCs with purified T cells resulted in strong CCL22 secretion both Technologies) were added. rIL-15 (PeproTech) was used for T cell in mice and in humans, whereas no effect was seen in a DC–B cell expansion. coculture (Fig. 1C). Within the T cell fraction, CD4+ cells showed The Journal of Immunology 3 Downloaded from http://www.jimmunol.org/ by guest on September 24, 2021

FIGURE 1. Constitutive CCL22 expression by DCs is mediated by T cells. (A–F) Cells isolated from murine spleen or human blood samples were cultured as described, and CCL22 protein levels in culture supernatants were determined by ELISA. DCs were magnetically sorted from murine spleen or human blood using CD11c+ and BDCA1+, BDCA3+, and BDCA4+ DCs, respectively. Graphs show representatives of two to four independent experiments (n = 3–5 mice or healthy human blood donors per group, respectively). (A) CCL22 levels of 1 3 106 freshly isolated WT murine splenocytes and human PBMCs were assessed after 24, 48, and 72 h of culture. (B) DCs (1 3 105 murine and 2.5 3 104 human) and Non-DCs (1 3 106 murine and 2.5 3 105 human) were cultured separately or together for 4 d (murine) or for 3 d (human). Unsorted splenocytes (1 3 106) and PBMCs (2.5 3 105) were used as controls. (C) DCs (1 3 105 murine and 2.5 3 104 human) were cultured either alone or with naive CD3+ T cells (Figure legend continues) 4 T CELL–DERIVED GM-CSF MEDIATES CONSTITUTIVE CCL22 EXPRESSION the strongest induction of CCL22 in both mice and humans GM-CSF was able to significantly reduce CCL22 secretion (Fig. 3D). (Fig. 1D). To further determine the relevance of T cells in the The additional neutralization of IL-3 and IL-5, whose receptors share process of CCL22 induction, we used Rag12/2 mice that lack the same b-chain for signaling (b common [bc] receptor) with the functional T and B cells because of defective V(D)J recombina- GM-CSFR (22, 23), did not further diminish CCL22 induction tion. Indeed, splenocytes derived from Rag12/2 mice showed significantly (Fig. 3D). strongly decreased CCL22 levels even after several days of culture The significance of GM-CSF as the inducer of constitutive (Fig. 1E). Addition of WT CD3+ T cells as well as CD4+ T cells, CCL22 secretion in lymphoid organs was further evaluated by but not B cells or CD8+ T cells, to Rag12/2 splenocytes restored quantification of spontaneous CCL22 secretion in freshly isolated CCL22 secretion (Fig. 1F). splenocytes derived from GM-CSF–deficient Csf22/2 mice; as expected, CCL22 secretion was strongly reduced in Csf22/2 T cell–derived soluble factors are responsible for DC-derived compared with WT splenocytes (Fig. 3E). Importantly, CCL22 homeostatic CCL22 expression induction by Csf2 2/2 T cells and their respective supernatants Cross-talk between DCs and T cells comprises Ag-dependent in WT DCs was strongly diminished compared with WT T cells physical interactions as well as indirect communication via cyto- (Fig. 3F). Taken together, these findings uncover an important role kines (19). To discriminate whether the T cell–mediated CCL22 of GM-CSF in CCL22 induction. expression in DCs requires direct DC–T cell contact or paracrine signaling via soluble factors, we harvested supernatants from CCL22 is differentially induced by GM-CSF, G-CSF, and unstimulated T cell and B cell cultures and added them to freshly M-CSF in mice and humans isolated murine and human DCs, respectively (Fig. 2A). Although Next, we aimed to evaluate the differential capacity of GM-CSF Downloaded from B cell supernatants did not stimulate CCL22 expression, T cell and its clinically relevant partner cytokines G-CSF and M-CSF supernatants strongly promoted DC-derived CCL22 secretion. in CCL22 regulation. Corresponding to our findings in mice, we Further, no increase in CCL22 expression was seen when T cells could show that isolated human DCs also secrete high levels of were cultured in supernatants derived from DCs, definitely dem- CCL22 upon GM-CSF stimulation (Fig. 4A). Next, we incubated onstrating that T cell–derived paracrine factors induce CCL22 in DCs with the cytokine G-CSF, which is broadly used clinically to

DCs. A similar CCL22 induction was observed when only the enhance myelopoiesis in patients under chemotherapy. Interestingly, http://www.jimmunol.org/ supernatants from WT T cells were added to Rag12/2 splenocytes in human DCs, G-CSF was also able to induce CCL22 secretion, but (Fig. 2B). Hao et al. (17) have shown that injection of apoptotic the overall levels were much lower compared with GM-CSF stimu- cells induces CCL22 secretion in CD8a+ CD103+ DCs. In our lation (Fig. 4B). Murine DCs, splenocytes, and lymph node–derived experimental setting, engulfment of cell fragments was not re- cells, in contrast, did not respond to G-CSF stimulation (Fig. 4B, sponsible for the induction of CCL22, as filtered T cell–derived Supplemental Fig. 3A, 3B). In vivo, s.c. G-CSF injection resulted in supernatants also induced CCL22 in Rag12/2 splenocytes the typical alterations in WBC, granulocyte, and stem cell counts (Supplemental Fig. 1). Taken together, a soluble T cell–derived in peripheral blood but did not induce CCL22 in lymphatic organs factor is responsible for DC-derived CCL22 expression. (Supplemental Fig. 3C–F). Similar results were seen for M-CSF,

which led to some CCL22 induction in human, but not murine, by guest on September 24, 2021 T cell–derived GM-CSF is the major inducer of DCs (Fig. 4C). In summary, although G-CSF and M-CSF induce constitutive CCL22 little amounts of CCL22 in human DCs, GM-CSF is by far the We cultured isolated DCs with a set of different cytokines to most potent CCL22 inductor in both humans and mice in vitro. identify the T cell–derived factors responsible for constitutive CCL22 secretion. Relevant CCL22 induction was observed upon T cell–derived GM-CSF induces CCL22 expression in vivo stimulation with IL-1, IL-2, IL-3, and IL-4, as described previ- To verify the relevance of our findings in vivo, we quantified ously (3, 11) (Fig. 3A). Surprisingly, a strong CCL22 induction in CCL22 levels using fresh tissue lysates from lymphoid organs. DCs also occurred upon treatment with IL-7 and GM-CSF Indeed, constitutive CCL22 expression was strongly reduced in (Fig. 3A). Interestingly, GM-CSF, but not IL-7, is constitutively the thymus and peripheral as well as mesenteric lymph nodes of expressed on mRNA level by unstimulated murine T cells ex vivo, Rag12/2 mice (Fig. 5A). Importantly, adoptive transfer of WT which is in accordance with previous reports (20, 21) (Fig. 3B). T cells into Rag12/2 mice restored CCL22 expression in pe- Further, also on protein level, we could observe a constant ripheral and mesenteric lymph nodes as well as in the spleen secretion of GM-CSF by freshly isolated, unstimulated murine (Fig. 5B), demonstrating the requirement of T cell help for ade- as well as human T cells (Fig. 3C). Among the known CCL22- quate CCL22 production also in vivo. Injection of CCL222/2 inducing cytokines, IL-3 and IL-5 were further detectable T cells into Rag-deficient mice was also able to restore CCL22 in T cells (Supplemental Fig. 2A). Murine DCs as wells as expression in peripheral and mesenteric lymph nodes, proving that human DCs and monocytes could be excluded as source of the adoptively transferred cells themselves were not the source of GM-CSF (Supplemental Fig. 2B, 2C). To identify the most CCL22 (Supplemental Fig. 4A). Of note, T cell transfer did not relevant T cell–derived, CCL22-inducing factor, we cocultured restore CCL22 expression in the thymus of Rag12/2 mice, dem- splenocytes with blocking Abs targeting potential T cell–derived, onstrating that thymic CCL22 expression may be regulated in a CCL22-inducing cytokines. Interestingly, only neutralization of different way, independently of mature T cells. This is further in

(1.5 3 106 murine and 1 3 105 human) or B cells (1.5 3 106 murine and 1 3 105 human) for 4 d (murine) or 3 d (human). (D) CCL22 protein levels of culture supernatants from 1 3 105 murine DCs cultured alone or with 1 3 106 CD3+CD4+CD25neg T cells or CD3+CD8+ T cells for 72 h were analyzed (Fig. 1D, left). CCL22 protein levels of culture supernatants of 2.5 3 104 human DCs cultured alone or with 1 3 106 CD4+ T cells or CD8+ T cells for 72 h were determined (Fig. 1D, right). (E) A total of 1 3 106 WT and Rag12/2 splenocytes were cultured for 72 h, and supernatants were harvested at the indicated time points. The graph displays a representative of five independent experiments (n = 5 mice per group). (F) A total of 1 3 106 Rag12/2 splenocytes alone (n = 4 mice) or cocultures of 0.9 3 106 Rag12/2 splenocytes with 1 3 105 of either isolated CD3+ pan T cells, CD4+ T cells, CD8+ T cells, or B cells (n = 2–4 mice) were carried out for 3 d, and supernatants were harvested. Data were analyzed using one-tailed Student t test. *p , 0.05. The Journal of Immunology 5

spontaneous CCL22 secretion by DCs in vitro are generally based on bone marrow–derived DCs that were generated by the use of different cytokines, among other GM-CSF (1, 2). Thus, the ob- servation that both human and murine unstimulated DCs do not secrete CCL22 intrinsically is novel. Additionally, we were able to demonstrate that naive unstimulated DCs (human as well as mu- rine) require T cell help to be able to produce CCL22, and we identified T cell–derived GM-CSF as one major driver. This was surprising, as GM-CSF, in contrast to the below-mentioned cyto- kines, has never been associated with CCL22 regulation before (3). Several studies have addressed CCL22 induction under in- flammatory conditions and identified LPS, IL-1, IL-4, or IL-13 as stimulating factors (1, 2, 11). We found that GM-CSF does not only mediate constitutive CCL22 expression but further augments the physiologic CCL22 expression both in vitro and in vivo. Thus, in addition to its role as T cell–derived factor during homeostasis, GM-CSF may also promote CCL22 induction under inflammatory conditions.

It is particularly interesting that CCL22 secretion by DCs is Downloaded from induced by conventional CD4+CD25neg and on a lower level even by CD8+ T cells. In consequence, T cell–mediated CCL22 ex- pression in DCs will enhance the frequency of DC–Treg contacts, leading to immune suppression (13). Thus, GM-CSF–producing T cells also exert an immunoregulatory function by secreting low

amounts of GM-CSF, which presents a novel concept of T cell– http://www.jimmunol.org/ mediated immune control in lymphatic organs. The secretion of GM-CSF by T cells has been studied exten- sively in the setting of inflammation. Th1- and Th17-derived GM-CSF secretion is one key driver of disease in murine and human autoreactive neuronal diseases (26–29). Additionally, even the existence of a distinct ThGM-CSF CD4+ Tcellsubset exclusively secreting GM-CSF in an IL-7/STAT-5–dependent manner is under debate (18, 30). Furthermore, Th cell–secreted GM-CSF occurs in the settings of various other autoimmune dis- by guest on September 24, 2021 orders, such as myocarditis (31) and human rheumatoid arthritis FIGURE 2. T cell–derived soluble factors are responsible for DC- derived homeostatic CCL22 expression. (A)DCs(13 105 murine and (32). The concept that T cells are able to secrete GM-CSF is thus 2.5 3 104 human) or T cells (1 3 105, murine) were cultured for 3 d either well established. To our knowledge, we demonstrate a novel in fresh medium or conditioned supernatants (SN) derived from a previous mechanism of low-dose GM-CSF production by T cells under 24 h (murine) or 72 h (human) T cell, B cell, or DC culture. (B) A total of homeostatic conditions in the lymph nodes, inducing CCL22 and 5 3 105 Rag12/2 splenocytes were cultured either in fresh medium (n =5 thereby representing an immunoinhibitory feedback loop. Inter- mice) or SN conditioned for 48 h by 1 3 106 CD3+ T cells (n = 5 mice) or estingly, GM-CSF secretion by T cells can be controlled by IL-7, a B cells (n = 5 mice) for 72 h. *p , 0.05, **p , 0.01. cytokine that is constitutively present in lymphatic organs and widely known for its capacity to mediate T cell homeostasis and survival (33, 34). Thus, T cell–derived GM-CSF does not only line with previous studies showing that only very low numbers of drive Th1- or Th17-dependent inflammatory processes but may mature peripheral T cells are able to re-enter into and persist in also contribute to the maintenance of immune homeostasis in the thymus (24). To evaluate whether GM-CSF is the inducer of lymph nodes in the absence of inflammation. CCL22 secretion also in vivo, we injected murine rGM-CSF into In our study, we used Rag12/2 mice as a T cell–depleted mouse 2 2 Rag1 / mice. Indeed, GM-CSF treatment restored CCL22 levels model and showed a reduction of CCL22 levels in lymphoid or- 2 2 in the lymph nodes of Rag1 / mice (Fig. 5C) and increased gans. Injection of WT ex vivo expanded T cells (Fig. 5B), but also baseline CCL22 levels in WT mice (Fig. 5D). In summary, these naive T cells (Supplemental Fig. 4B) restored CCL22 expression findings demonstrate a key role of T cell–derived GM-CSF in the in Rag12/2 mice. Whereas GM-CSF expression by T cells has induction of homeostatic CCL22 secretion also in vivo. been mainly described to occur after T cell activation (35), we have shown in vitro that naive T cells secrete GM-CSF without Discussion any further stimulation, leading to CCL22 induction in DCs. In The chemokine CCL22 is expressed constitutively at high levels our in vivo model, we cannot exclude that the adoptively trans- in lymphoid organs and controls the distribution of immune cells ferred T cells experience activation upon injection into Rag12/2 and T cell immunity through the induction of DC–Treg contacts mice as even the transfer of nonactivated T cells into lymphopenic (13, 14, 25). Despite its important role as central immune regulator, hosts is known to induce different types of T cell activation that the mechanisms that maintain constitutive CCL22 expression in are required for their expansion and persistence (36, 37). We an- lymphatic organs have been unknown so far. Under homeostatic alyzed CCL22 levels in Rag12/2 mice 4 wk after injection of conditions in the lymph node, CCL22 is expressed exclusively by donor T cells when a certain state of homeostasis may already be DCs (10, 13). In this study, we show that isolated unstimulated reestablished (38). Thus, although we cannot exclude activation of DCs alone are unable to secrete CCL22. Previous data demonstrating the transferred T cells in the Rag12/2 model, it seems likely that 6 T CELL–DERIVED GM-CSF MEDIATES CONSTITUTIVE CCL22 EXPRESSION Downloaded from http://www.jimmunol.org/ by guest on September 24, 2021

FIGURE 3. T cell–derived GM-CSF is the major inducer of constitutive CCL22. (A) A total of 2 3 105 murine DCs were stimulated with 20 ng/ml IL-1a, IL-1b, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-9, IL-10, IL-12, IL-21, IL-23, or GM-CSF and 1000 U/ml IFN-a or IFN-g for 48 h. CCL22 secretion was analyzed by ELISA, and the graph shows representative of two to four independent experiments (n = 5 mice per group). (B) A total of 1 3 106 CD3+ T cells or B cells were cultured for 48 h. Subsequently, T and B cells were collected, mRNA was isolated, and the gene expression of GM-CSF (n = 8 mice per group) and IL-7 (n = 3 mice per group) was analyzed by qRT-PCR. Graph shows pooled data of 6 (GM-CSF) and 3 (IL-7) independent experiments. (C) A total of 13 106 of either murine or human T cells and B cells were incubated for 48 h in fresh medium. Supernatants were then analyzed for GM-CSF protein expression by ELISA. Graph of murine experiments show pooled data of five independent experiments (n = 5 mice per group), graph of human experiments shows representative of one independent experiments (n = 1 donor per group). (D)Atotalof13 106 WT splenocytes (n = 6 mice per group) were cultured for 3 d with 0.2 mg/ml of either isotype or the indicated neutralizing Abs. CCL22 levels from culture supernatants were determined using ELISA. Graphs show pooled data of three independent experiments. (E) CCL22 levels of 1 3 106 freshly isolated WT or GM-CSF–deficient splenocytes from Csf22/2 mice were determined after 24, 48, and 72 h of culture. CCL22 levels from culture supernatants were determined using ELISA. Graphs show pooled data of three independent experiments. Data were analyzed using one-tailed Student t test. (F) A total of 1 3 105 WT DCs was cultured alone or together with 1 3 106 WT or Csf2 2/2 T cells for 72 h, and CCL22 levels in the supernatant were determined by ELISA (Fig. 3F, left). A total of 1 3 106 WTTcellsorCsf2 2/2 T cells was cultured separately for 48 h; supernatants were harvested and added to 1 3 105 WT DCs for a culture of 72 h. Subsequently, supernatants were harvested, and CCL22 levels were determined by ELISA (Fig. 3E, right). Graph shows representative of three independent experiments. *p , 0.05, **p , 0.01. our in vitro data of GM-CSF secretion by naive T cells that induce expressed by nonimmune cells, such as epithelial cells or fibroblasts CCL22 expression in DCs are also applicable in vivo. (35), and these cells could thus contribute to CCL22 secretion by DCs Although CCL22 levels in Rag12/2 mice are strongly reduced, in Rag12/2 mice. In general, CCL22 induction is a redundant pro- CCL22 is still produced at relevant amounts. GM-CSF can also be cess, with many different factors being able to mediate DC-derived The Journal of Immunology 7

FIGURE 4. CCL22 is differentially in- duced by GM-CSF, G-CSF, and M-CSF in mice and humans. CCL22 levels in culture supernatants were determined by ELISA. Graphs show pooled data of three (murine) and seven (human) independent experiments. DCs (5 3 104 murine and 2.5 3 104 human) were cultured either in normal medium or with different concentrations of murine and human (A) rGM-CSF (0.1, 1, or 10 ng/ml), (B) G-CSF (0.1, 1, or 10 ng/ml), or (C)M-CSF (1, 10, or 100 ng/ml), respectively, for 48 h. Downloaded from *p , 0.05, ***p , 0.001, ****p , 0.0001. http://www.jimmunol.org/

CCL22 expression (Fig. 3A). Besides GM-CSF, especially IL-7, a understood yet, and it would be interesting to dissect the role of cytokine that is mainly secreted by nonhematopoietic cells, such CCL22 in GM-CSF–mediated immune regulation. as stromal cells in lymphoid tissues, represents a potent CCL22 A pleiotropic role of GM-CSF has also been reported in the inducer. Interestingly, it has been shown that IL-7 levels are setting of cancer with several reports on either anti- or protu- by guest on September 24, 2021 elevated in lymphopenic hosts such as Rag12/2 mice because morigenic effects. GM-CSF potently induces maturation of DCs by of diminished scavenging by T cells (39). Thus, besides GM-CSF, promoting Ag-presentation through upregulation of B7 recep- the cytokine IL-7 may be one important inducer of CCL22, par- tors (54). In several murine cancer models, GM-CSF–mediated ticularly in lymphopenic hosts. DC activation led to potent antitumor immune responses (54, 55). CCL22 expression by DCs regulates T cell immunity in the However, it is generally thought that GM-CSF can also inhibit lymph node. As GM-CSF induces CCL22, one may expect that it antitumor immunity and promote tumor immune evasion (56, 57). exerts immunosuppressive functions. Indeed, GM-CSF prevents Mechanistically, GM-CSF–mediated immunosuppression was the development and progress of disease in mouse models of type 1 reported to be due to an expansion of Tregs (58) and myeloid- diabetes, autoimmune thyroiditis, myasthenia gravis, DSS-induced derived suppressor cells, hence contributing to abrogated effector colitis, and chronic graft versus host disease (40–44). Additionally, T cell responses (59, 60). Interestingly, low doses of GM-CSF maintenance of tolerance was described to be driven by GM-CSF– used in clinical trials were associated with an augmented vaccine- mediated induction of tolerogenic DCs that enhance Treg function induced immune response in cancer patients, whereas high doses (40, 41). Thus, it is possible that also the CCL22–CCR4 axis is resulted in a distorted immune response and impaired antitumor mechanistically involved in GM-CSF–driven immunosuppres- immunity (61). In the clinical setting, GM-CSF–based therapeutic sion in these disease models, and this merits further evaluation. In approaches have been shown to be effective and even received U.S. contrast to its immunosuppressive role, GM-CSF is important for Food and Drug Administration approval: Sipuleucel-T (Provenge), survival, differentiation, and maturation of granulocytes, macro- a vaccination based on prostate acid phosphatase and GM-CSF, phages, and DCs, which leads to immune activation (45). Indeed, was approved in 2010 for metastatic prostate cancer. In addition, GM-CSF–deficient mice are resistant to experimental autoimmune talimogene laherparepvec (T-VEC or Imlygic), an oncolytic virus encephalitis and myocarditis as well as rheumatoid arthritis encoding GM-CSF, was approved in 2015 for the treatment of (31, 46–48), showing that GM-CSF is also important at the onset melanoma. Overall, however, GM-CSF–based immunotherapy has or maintenance of autoimmune responses in certain settings. experienced no striking clinical benefit so far. It will be impor- In line with these data, clinical trials with GM-CSFR a-chain tant to investigate the role of CCL22 in the diverging effects of antagonists such as mavrilimumab and GM-CSF cytokine inhib- GM-CSF in cancer. itors like MOR103 showed persuasive therapeutic potential in In addition to the effects of GM-CSF, we also investigated the patients with rheumatoid arthritis (49, 50). In contrast, adminis- role of G-CSF and M-CSF that also belong to the family of colony tration of GM-CSF led to significant reduction of disease activity stimulation factor cytokines. Whereas M-CSF is infrequently used, and corticosteroid use in patients with Crohn disease (51–53). The G-CSF as well as GM-CSF are routinely used in clinical practice factors determining the reported opposing effects of GM-CSF in for prophylaxis and treatment of patients with febrile neutropenia the regulation of inflammation and autoimmunity are not fully under certain myelosuppressive chemotherapy regimens and for 8 T CELL–DERIVED GM-CSF MEDIATES CONSTITUTIVE CCL22 EXPRESSION Downloaded from http://www.jimmunol.org/ by guest on September 24, 2021

FIGURE 5. T cell–derived GM-CSF induces CCL22 expression in vivo CCL22 levels from culture supernatants or organ lysates were determined using ELISA. (A) CCL22 protein levels were analyzed in the peripheral lymph nodes (LN), mesenteric LN, thymus, and spleen of WT (n = 10) and Rag12/2 mice (n = 10 mice). Graphs show pooled data from three independent experiments. (B) CCL22 protein levels in peripheral LN, mesenteric LN, thymus, and spleen of Rag12/2 mice were determined 4 wk after i.v. injection of either PBS (n = 8 mice) or 10 3 106 WT T cells (n = 8 mice). Graphs show pooled data of two independent experiments. (C) CCL22 protein levels were determined in peripheral and mesenteric LN 12 h after s.c. injection of PBS or 0.5 mg murine rGM-CSF in Rag12/2 mice. (D) CCL22 protein levels were determined in spleen, peripheral LN, mesenteric LN, and serum 3, 6, 12, and 24 h after s.c. injections of PBS (unstimulated) or 0.5 mg murine rGM-CSF in WT mice (n = 3 mice per group). Graphs show pooled data of one out of two in- dependent experiments. *p , 0.05, **p , 0.01, ***p , 0.001, ****p , 0.0001. The Journal of Immunology 9 mobilization of peripheral blood stem cells prior to autologous or 15. Hu, Z., J. N. Lancaster, C. Sasiponganan, and L. I. Ehrlich. 2015. CCR4 pro- motes medullary entry and thymocyte-dendritic cell interactions required for allogeneic peripheral stem cell transplantation (62, 63). One pre- central tolerance. J. Exp. Med. 212: 1947–1965. vious study showed that in vivo application of G-CSF can lead to 16. Montane, J., L. Bischoff, G. Soukhatcheva, D. L. Dai, G. Hardenberg, increased CCL22 secretion by ex vivo FACS-sorted CD8a2 DCs M. K. Levings, P. C. Orban, T. J. Kieffer, R. Tan, and C. B. Verchere. 2011. Prevention of murine autoimmune diabetes by CCL22-mediated Treg recruit- derived from peripancreatic lymph nodes of NOD mice (64). Our ment to the pancreatic islets. J. Clin. Invest. 121: 3024–3028. in vitro and in vivo data demonstrate that G-CSF, in contrast to 17. Hao, S., X. Han, D. Wang, Y. Yang, Q. Li, X. Li, and C. H. Qiu. 2016. Critical GM-CSF, does not induce CCL22 in WT mice and induces only role of CCL22/CCR4 axis in the maintenance of immune homeostasis during apoptotic cell clearance by splenic CD8a(+) CD103(+) dendritic cells. Immu- minor amounts in human PBMCs in vitro. Thus, clinical G-CSF nology 148: 174–186. use will probably not interfere with CCL22-mediated immune 18. Herndler-Brandstetter, D., and R. A. Flavell. 2014. Producing GM-CSF: a unique regulation, which may be interesting to verify by determining T helper subset? Cell Res. 24: 1379–1380. 19. Hivroz, C., K. Chemin, M. Tourret, and A. Bohineust. 2012. Crosstalk between CCL22 levels during G-CSF therapy. T lymphocytes and dendritic cells. Crit. Rev. Immunol. 32: 139–155. Taken together, to our knowledge, we identified GM-CSF as a 20. Fry, T. J., and C. L. Mackall. 2002. -7: from bench to clinic. Blood 99: 3892–3904. so-far unknown factor of CCL22 induction in DCs and un- 21. Griffin, J. D., S. A. Cannistra, R. Sullivan, G. D. Demetri, T. J. Ernst, and covered a novel concept of T cell–dependent CCL22 induc- Y. Kanakura. 1990. The biology of GM-CSF: regulation of production and in- tion under homeostatic conditions. This novel GM-CSF–CCL22 teraction with its receptor. Int. J. Cell. Cloning 8(Suppl. 1): 35–45. 22. Kitamura, T., N. Sato, K. Arai, and A. Miyajima. 1991. Expression cloning of the axis may contribute to immune regulation in cancer and auto- human IL-3 receptor cDNA reveals a shared beta subunit for the human IL-3 and immunity and thereby opens a new way of therapeutic immune GM-CSF receptors. Cell 66: 1165–1174. interventions. 23. Tavernier, J., R. Devos, S. Cornelis, T. Tuypens, J. Van der Heyden, W. Fiers, and G. Plaetinck. 1991. A human high affinity interleukin-5 receptor (IL5R) is composed of an IL5-specific alpha chain and a beta chain shared with the Downloaded from Acknowledgments receptor for GM-CSF. Cell 66: 1175–1184. 24. Michie, S. A., E. A. Kirkpatrick, and R. V. Rouse. 1988. Rare peripheral T cells We thank Jana Mandelbaum and Dan Draganovici for assistance in tissue migrate to and persist in normal mouse thymus. J. Exp. Med. 168: 1929–1934. embedding and Prof. Wolfgang Zimmermann for careful reading of the 25. Anders, H. J., P. Romagnani, and A. Mantovani. 2014. Pathomechanisms: ho- manuscript. meostatic in health, tissue regeneration, and progressive diseases. Trends Mol. Med. 20: 154–165. 26. Noster, R., R. Riedel, M. F. Mashreghi, H. Radbruch, L. Harms, C. Haftmann,

Disclosures H. D. Chang, A. Radbruch, and C. E. Zielinski. 2014. IL-17 and GM-CSF ex- http://www.jimmunol.org/ The authors have no financial conflicts of interest. pression are antagonistically regulated by human T helper cells. Sci. Transl. Med. 6: 241ra80. 27. Grifka-Walk, H. M., D. A. Giles, and B. M. Segal. 2015. IL-12-polarized Th1 cells produce GM-CSF and induce EAE independent of IL-23. Eur. J. Immunol. References 45: 2780–2786. 28. El-Behi, M., B. Ciric, H. Dai, Y. Yan, M. Cullimore, F. Safavi, G. X. Zhang, 1. Vulcano, M., C. Albanesi, A. Stoppacciaro, R. Bagnati, G. D’Amico, S. Struyf, P. Transidico, R. Bonecchi, A. Del Prete, P. Allavena, et al. 2001. Dendritic cells B. N. Dittel, and A. Rostami. 2011. The encephalitogenicity of T(H)17 cells as a major source of -derived chemokine/CCL22 in vitro and in vivo. is dependent on IL-1- and IL-23-induced production of the cytokine GM-CSF. Eur. J. Immunol. 31: 812–822. Nat. Immunol. 12: 568–575. ´ 2. Godiska, R., D. Chantry, C. J. Raport, S. Sozzani, P. Allavena, D. Leviten, 29. Codarri, L., G. Gyu¨lveszi, V. Tosevski, L. Hesske, A. Fontana, L. Magnenat, A. Mantovani, and P. W. Gray. 1997. Human macrophage-derived chemokine T. Suter, and B. Becher. 2011. RORgt drives production of the cytokine GM-CSF by guest on September 24, 2021 (MDC), a novel chemoattractant for monocytes, monocyte-derived dendritic in helper T cells, which is essential for the effector phase of autoimmune neu- cells, and natural killer cells. J. Exp. Med. 185: 1595–1604. roinflammation. Nat. Immunol. 12: 560–567. 3. Mantovani, A., P. A. Gray, J. Van Damme, and S. Sozzani. 2000. Macrophage- 30. Sheng, W., F. Yang, Y. Zhou, H. Yang, P. Y. Low, D. M. Kemeny, P. Tan, A. Moh, derived chemokine (MDC). J. Leukoc. Biol. 68: 400–404. M. H. Kaplan, Y. Zhang, and X. Y. Fu. 2014. STAT5 programs a distinct subset 4. D’Ambrosio, D., A. Iellem, R. Bonecchi, D. Mazzeo, S. Sozzani, A. Mantovani, of GM-CSF-producing T helper cells that is essential for autoimmune neuro- and F. Sinigaglia. 1998. Selective up-regulation of chemokine receptors CCR4 inflammation. Cell Res. 24: 1387–1402. and CCR8 upon activation of polarized human type 2 Th cells. J. Immunol. 161: 31. Sonderegger, I., G. Iezzi, R. Maier, N. Schmitz, M. Kurrer, and M. Kopf. 2008. 5111–5115. GM-CSF mediates autoimmunity by enhancing IL-6-dependent Th17 cell de- 5. Iellem, A., M. Mariani, R. Lang, H. Recalde, P. Panina-Bordignon, F. Sinigaglia, velopment and survival. J. Exp. Med. 205: 2281–2294. and D. D’Ambrosio. 2001. Unique chemotactic response profile and specific 32. Reynolds, G., J. R. Gibbon, A. G. Pratt, M. J. Wood, D. Coady, G. Raftery, expression of chemokine receptors CCR4 and CCR8 by CD4(+)CD25(+) reg- A. R. Lorenzi, A. Gray, A. Filer, C. D. Buckley, et al. 2016. Synovial CD4+ ulatory T cells. J. Exp. Med. 194: 847–853. T-cell-derived GM-CSF supports the differentiation of an inflammatory dendritic 6. Romagnani, S. 2002. Cytokines and chemoattractants in allergic inflammation. cell population in rheumatoid arthritis. Ann. Rheum. Dis. 75: 899–907. Mol. Immunol. 38: 881–885. 33. Tan, J. T., E. Dudl, E. LeRoy, R. Murray, J. Sprent, K. I. Weinberg, and 7. Klarquist, J., K. Tobin, P. Farhangi Oskuei, S. W. Henning, M. F. Fernandez, C. D. Surh. 2001. IL-7 is critical for homeostatic proliferation and survival of E. R. Dellacecca, F. C. Navarro, J. M. Eby, S. Chatterjee, S. Mehrotra, et al. naive T cells. Proc. Natl. Acad. Sci. USA 98: 8732–8737. 2016. Ccl22 diverts T regulatory cells and controls the growth of melanoma. 34. Bradley, L. M., L. Haynes, and S. L. Swain. 2005. IL-7: maintaining T-cell Cancer Res. 76: 6230–6240. memory and achieving homeostasis. Trends Immunol. 26: 172–176. 8. Wu, S., H. He, H. Liu, Y. Cao, R. Li, H. Zhang, H. Li, Z. Shen, J. Qin, and J. Xu. 35. Shi, Y., C. H. Liu, A. I. Roberts, J. Das, G. Xu, G. Ren, Y. Zhang, L. Zhang, 2018. C-C motif chemokine 22 predicts postoperative prognosis and adjuvant Z. R. Yuan, H. S. W. Tan, et al. 2006. Granulocyte-macrophage colony- chemotherapeutic benefits in patients with stage II/III gastric cancer. OncoIm- stimulating factor (GM-CSF) and T-cell responses: what we do and don’t munology 7: e1433517. know. Cell Res. 16: 126–133. 9. Curiel, T. J., G. Coukos, L. Zou, X. Alvarez, P. Cheng, P. Mottram, M. Evdemon- 36. Min, B., H. Yamane, J. Hu-Li, and W. E. Paul. 2005. Spontaneous and ho- Hogan, J. R. Conejo-Garcia, L. Zhang, M. Burow, et al. 2004. Specific recruitment meostatic proliferation of CD4 T cells are regulated by different mechanisms. of regulatory T cells in ovarian carcinoma fosters immune privilege and predicts J. Immunol. 174: 6039–6044. reduced survival. Nat. Med. 10: 942–949. 37. Kim, J., J. Y. Lee, K. Cho, S.-W. Hong, K. S. Kim, J. Sprent, S. H. Im, + 10. Tang, H. L., and J. G. Cyster. 1999. Chemokine Up-regulation and activated C. D. Surh, and J. H. Cho. 2018. Spontaneous proliferation of CD4 T cells in T cell attraction by maturing dendritic cells. Science 284: 819–822. RAG-deficient hosts promotes antigen-independent but IL-2-dependent strong 11. Wiedemann, G. M., M. M. Knott, V. K. Vetter, M. Rapp, S. Haubner, J. Fesseler, proliferative response of naı¨ve CD8 + T cells. Front. Immunol. 9: 1907. B. Ku¨hnemuth, P. Layritz, R. Thaler, S. Kruger, et al. 2016. Cancer cell-derived 38. Pennock, N. D., J. T. White, E. W. Cross, E. E. Cheney, B. A. Tamburini, and IL-1a induces CCL22 and the recruitment of regulatory T cells. OncoImmu- R. M. Kedl. 2013. T cell responses: naive to memory and everything in between. nology 5: e1175794. Adv. Physiol. Educ. 37: 273–283. 12. Yamashita, U., and E. Kuroda. 2002. Regulation of macrophage-derived che- 39. Guimond, M., R. G. Veenstra, D. J. Grindler, H. Zhang, Y. Cui, R. D. Murphy, mokine (MDC, CCL22) production. Crit. Rev. Immunol. 22: 105–114. S. Y. Kim, R. Na, L. Hennighausen, S. Kurtulus, et al. 2009. sig- 13. Rapp, M., M. W. M. Wintergerst, W. G. Kunz, V. K. Vetter, M. M. L. Knott, naling in dendritic cells regulates the homeostatic proliferation and niche size of D. Lisowski, S. Haubner, S. Moder, R. Thaler, S. Eiber, et al. 2019. CCL22 CD4+ T cells. Nat. Immunol. 10: 149–157. controls immunity by promoting regulatory T cell communication with dendritic 40. Gaudreau, S., C. Guindi, M. Me´nard, G. Besin, G. Dupuis, and A. Amrani. 2007. cells in lymph nodes. J. Exp. Med. 216: 1170–1181. Granulocyte-macrophage colony-stimulating factor prevents diabetes development in 14. Vitali, C., F. Mingozzi, A. Broggi, S. Barresi, F. Zolezzi, J. Bayry, G. Raimondi, NOD mice by inducing tolerogenic dendritic cells that sustain the suppressive func- I. Zanoni, and F. Granucci. 2012. Migratory, and not lymphoid-resident, dendritic tion of CD4+CD25+ regulatory T cells. J. Immunol. 179: 3638–3647. cells maintain peripheral self-tolerance and prevent autoimmunity via induction of 41. Vasu, C., R. N. Dogan, M. J. Holterman, and B. S. Prabhakar. 2003. Selective iTreg cells. Blood 120: 1237–1245. induction of dendritic cells using granulocyte macrophage-colony stimulating 10 T CELL–DERIVED GM-CSF MEDIATES CONSTITUTIVE CCL22 EXPRESSION

factor, but not fms-like tyrosine kinase receptor 3-ligand, activates thyroglobulin- with corticosteroid-dependent Crohn’s disease: a randomised, double-blind, specific CD4+/CD25+ T cells and suppresses experimental autoimmune thyroiditis. placebo-controlled, phase 2 study. Gut 58: 1354–1362. J. Immunol. 170: 5511–5522. 54. Mach, N., S. Gillessen, S. B. Wilson, C. Sheehan, M. Mihm, and G. Dranoff. 42. Sheng, J. R., L. Li, B. B. Ganesh, C. Vasu, B. S. Prabhakar, and M. N. Meriggioli. 2000. Differences in dendritic cells stimulated in vivo by tumors engineered 2006. Suppression of experimental autoimmune myasthenia gravis by granulocyte- to secrete granulocyte-macrophage colony-stimulating factor or Flt3-ligand. macrophage colony-stimulating factor is associated with an expansion of FoxP3+ Cancer Res. 60: 3239–3246. regulatory T cells. J. Immunol. 177: 5296–5306. 55. Dranoff, G., E. Jaffee, A. Lazenby, P. Golumbek, H. Levitsky, K. Brose, 43. Xu, Y., N. H. Hunt, and S. Bao. 2008. The role of granulocyte macrophage- V. Jackson, H. Hamada, D. Pardoll, and R. C. Mulligan. 1993. Vaccination with colony-stimulating factor in acute intestinal inflammation. Cell Res. 18: 1220– irradiated tumor cells engineered to secrete murine granulocyte-macrophage 1229. colony-stimulating factor stimulates potent, specific, and long-lasting anti- 44. Hotta, M., H. Yoshimura, A. Satake, Y. Tsubokura, T. Ito, and S. Nomura. 2019. tumor immunity. Proc. Natl. Acad. Sci. USA 90: 3539–3543. GM-CSF therapy inhibits chronic graft-versus-host disease via expansion of 56. Slingluff, C. L., Jr., G. R. Petroni, W. C. Olson, M. E. Smolkin, M. I. Ross, regulatory T cells. Eur. J. Immunol. 49: 179–191. N. B. Haas, W. W. Grosh, M. E. Boisvert, J. M. Kirkwood, and K. A. Chianese- 45. Hamilton, J. A. 2008. Colony-stimulating factors in inflammation and autoim- Bullock. 2009. Effect of granulocyte/macrophage colony-stimulating factor on munity. Nat. Rev. Immunol. 8: 533–544. circulating CD8+ and CD4+ T-cell responses to a multipeptide melanoma 46. McQualter, J. L., R. Darwiche, C. Ewing, M. Onuki, T. W. Kay, J. A. Hamilton, vaccine: outcome of a multicenter randomized trial. Clin. Cancer Res. 15: H. H. Reid, and C. C. Bernard. 2001. Granulocyte macrophage colony-stimulating 7036–7044. factor: a new putative therapeutic target in multiple sclerosis. J. Exp. Med. 57. Faries, M. B., E. C. Hsueh, X. Ye, M. Hoban, and D. L. Morton. 2009. Effect of 194: 873–882. granulocyte/macrophage colony-stimulating factor on vaccination with an allo- 47. Ponomarev, E. D., L. P. Shriver, K. Maresz, J. Pedras-Vasconcelos, D. Verthelyi, geneic whole-cell melanoma vaccine. Clin. Cancer Res. 15: 7029–7035. and B. N. Dittel. 2007. GM-CSF production by autoreactive T cells is required 58. Jinushi, M., Y. Nakazaki, M. Dougan, D. R. Carrasco, M. Mihm, and G. Dranoff. 2007. MFG-E8-mediated uptake of apoptotic cells by APCs links the pro- and for the activation of microglial cells and the onset of experimental autoimmune antiinflammatory activities of GM-CSF. J. Clin. Invest. 117: 1902–1913. encephalomyelitis. J. Immunol. 178: 39–48. 59. Morales, J. K., M. Kmieciak, K. L. Knutson, H. D. Bear, and M. H. Manjili. 48. Campbell, I. K., M. J. Rich, R. J. Bischof, A. R. Dunn, D. Grail, and J. A. Hamilton. 2010. GM-CSF is one of the main breast tumor-derived soluble factors involved

1998. Protection from collagen-induced arthritis in granulocyte-macrophage colony- Downloaded from in the differentiation of CD11b-Gr1- bone marrow progenitor cells into myeloid- J. Immunol. stimulating factor-deficient mice. 161: 3639–3644. derived suppressor cells. Breast Cancer Res. Treat. 123: 39–49. 49. Burmester, G. R., I. B. McInnes, J. Kremer, P. Miranda, M. Korkosz, 60. Dufait, I., J. K. Schwarze, T. Liechtenstein, W. Leonard, H. Jiang, D. Escors, J. Vencovsky, A. Rubbert-Roth, E. Mysler, M. A. Sleeman, A. Godwood, et al; M. De Ridder, and K. Breckpot. 2015. Ex vivo generation of myeloid-derived EARTH EXPLORER 1 study investigators. 2017. A randomised phase IIb study suppressor cells that model the tumor immunosuppressive environment in co- of mavrilimumab, a novel GM-CSF receptor alpha monoclonal antibody, in the lorectal cancer. Oncotarget 6: 12369–12382. treatment of rheumatoid arthritis. Ann. Rheum. Dis. 76: 1020–1030. 61. Parmiani, G., C. Castelli, L. Pilla, M. Santinami, M. P. Colombo, and 50. Behrens, F., P. P. Tak, M. Østergaard, R. Stoilov, P. Wiland, T. W. Huizinga, L. Rivoltini. 2007. Opposite immune functions of GM-CSF administered as V. Y. Berenfus, S. Vladeva, J. Rech, A. Rubbert-Roth, et al. 2015. MOR103, a

vaccine adjuvant in cancer patients. Ann. Oncol. 18: 226–232. http://www.jimmunol.org/ human monoclonal antibody to granulocyte-macrophage colony-stimulating 62. Smith, T. J., K. Bohlke, G. H. Lyman, K. R. Carson, J. Crawford, S. J. Cross, factor, in the treatment of patients with moderate rheumatoid arthritis: results J. M. Goldberg, J. L. Khatcheressian, N. B. Leighl, C. L. Perkins, et al; American of a phase Ib/IIa randomised, double-blind, placebo-controlled, dose-escalation Society of Clinical Oncology. 2015. Recommendations for the use of WBC trial. Ann. Rheum. Dis. 74: 1058–1064. growth factors: American Society of Clinical Oncology clinical practice 51. Dieckgraefe, B. K., and J. R. Korzenik. 2002. Treatment of active Crohn’s guideline update. J. Clin. Oncol. 33: 3199–3212. disease with recombinant human granulocyte-macrophage colony-stimulating 63. Hume, D. A., and K. P. A. MacDonald. 2012. Therapeutic applications of factor. Lancet 360: 1478–1480. macrophage colony-stimulating factor-1 (CSF-1) and antagonists of CSF-1 52. Korzenik, J. R., B. K. Dieckgraefe, J. F. Valentine, D. F. Hausman, and receptor (CSF-1R) signaling. Blood 119: 1810–1820. M. J. Gilbert, Sargramostim in Crohn’s Disease Study Group. 2005. Sar- 64. Layseca-Espinosa, E., S. Korniotis, R. Montandon, C. Gras, M. Bouillie´, R. Gonzalez- gramostim for active Crohn’s disease. N. Engl. J. Med. 352: 2193–2201. Amaro, M. Dy, and F. Zavala. 2013. CCL22-producing CD8a- myeloid 53. Valentine, J. F., R. N. Fedorak, B. Feagan, P. Fredlund, R. Schmitt, P. Ni, and dendritic cells mediate regulatory T cell recruitment in response to G-CSF

T. J. Humphries. 2009. Steroid-sparing properties of sargramostim in patients treatment. J. Immunol. 191: 2266–2272. by guest on September 24, 2021