Regulatory Inhibit Alternative Activation and Attenuate Pathology Associated with Fibrosis

This information is current as Prabha Chandrasekaran, Salman Izadjoo, Jessica Stimely, of September 25, 2021. Senthilkumar Palaniyandi, Xiaoping Zhu, Wagner Tafuri and David M. Mosser J Immunol published online 20 September 2019 http://www.jimmunol.org/content/early/2019/09/19/jimmun

ol.1900270 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2019/09/20/jimmunol.190027 Material 0.DCSupplemental http://www.jimmunol.org/

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 25, 2021

*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 © 2019 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published September 20, 2019, doi:10.4049/jimmunol.1900270 The Journal of Immunology

Regulatory Macrophages Inhibit Alternative Macrophage Activation and Attenuate Pathology Associated with Fibrosis

Prabha Chandrasekaran,*,† Salman Izadjoo,*,† Jessica Stimely,*,† Senthilkumar Palaniyandi,‡ Xiaoping Zhu,‡ Wagner Tafuri,x and David M. Mosser*,†

Diversity and plasticity are the hallmarks of macrophages. The two most well-defined macrophage subsets are the classically activated macrophages (CAMfs) and the IL-4–derived alternatively activated macrophages (AAMfs). Through a series of studies, we previously identified and characterized a distinct population of macrophages with immunoregulatory functions, collectively termed regulatory macrophages (RMfs). Although considerable advances have been made in understanding these various macrophage subsets, it is not known whether macrophages of one activation state can influence the other. In this study, we examined whether RMfs capable of inhibiting inflammatory responses of CAMfs could also inhibit AAMfs and their profi-

brotic responses. Our results demonstrated that RMfs significantly dampened the alternate activation phenotype of AAMfs Downloaded from generated in vitro and intrinsically occurring AAMfs from TACI2/2 macrophages. Further, RMfs inhibited AAMf-promoted arginase activity and fibroblast proliferation in vitro. This inhibition occurred regardless of the strength, duration, and mode of alternative activation and was only partially dependent on IL-10. In the chlorhexidine gluconate–induced peritoneal fibrosis model, AAMfs worsened the fibrosis, but RMfs rescued mice from AAMf-mediated pathological conditions. Taken together, our study demonstrates that RMfs are a specialized subset of macrophages with a nonredundant role in limiting overt pro-

regenerative functions of AAMfs, a role distinct from their well-defined role of suppression of inflammatory responses by http://www.jimmunol.org/ CAMfs. The Journal of Immunology, 2019, 203: 000–000.

acrophages (Mfs) are heterogeneous cells with varied exposure of murine Mfs to IL-4, IL-13, or helminth functions and transcriptional signatures (1–3). All Mfs, results in Mfs with a well-defined phenotype (7) that is profibrotic M regardless of their origin, can respond to danger signals in character and includes the expression of reliable biomarkers and assume an inflammatory phenotype. The (so-called M1) Mfs Ym1 and Relma (8). The M2 subset of Mfs has been loosely produce a vast armamentarium of inflammatory , che- expanded to encompass a wide range of Mfs with noninflam- mokines, and mediators. Classically activated Mfs(CAMfs) matory roles, including tissue remodeling, immunoregulation, and arise in response to a combination of danger signals and host (9). The expansion of this M2 designation has led by guest on September 25, 2021 IFNs, which are typically generated during a cell-mediated im- to considerable ambiguity in understanding the phenotype and mune response (4, 5). In 1992, Stein et al. (6) described a pop- functions of the variety of existing and emerging Mf subsets, ulation of “alternatively activated” Mfs (AAMfs), which arose in including tumor-associated Mfs (10), Mox Mfs that are associ- response to the TH2 IL-4. Mfs exposed to IL-4 failed to ated with the oxidized low-density lipoprotein–rich environment induce inflammatory cytokine production as CAMfsdid,but of atherosclerosis (11), and Mfs involved in the resolution of upregulated the expression of the and MHC immune responses (12). class II and were therefore termed alternatively activated (6). We had proposed a scheme to group Mfs based on their ho- Subsequent studies by many other groups have demonstrated that meostatic activities of host defense, , and immune regulation (3). This wider grouping followed our description of a population of Mfs with potent anti-inflammatory and immuno- *Department of Cell Biology and Molecular Genetics, University of Maryland, College † f f Park, MD 20742; Maryland Pathogen Research Institute, University of Maryland, regulatory activity that we termed regulatory M s(RM s) (3, 13–15). College Park, MD 20740; ‡Virginia-Maryland Regional College of Veterinary Med- x We subsequently demonstrated that there were many ways to icine, University of Maryland, College Park, MD 20470; and Departamento de generate Mfs with immunoregulatory and growth-promoting ac- Patologia Geral, Instituto de Cieˆncias Biolo´gicas, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais 31270-901, Brazil tivity. The transcriptional programs of Mfs stimulated with three ORCIDs: 0000-0002-2916-280X (J.S.); 0000-0002-1672-9504 (X.Z.); 0000-0002- different “reprogramming” signals, namely, high-density immune 9503-4187 (D.M.M.). complexes (IC), PGE2, and Adenosine (Ade), were compared Received for publication March 7, 2019. Accepted for publication August 14, 2019. (16). All three of these populations of Mfs secreted lower levels This work was supported by grants from the National Institutes of Health (R01 of inflammatory mediators but higher levels of anti-inflammatory GM102589-01 and U01 AI088650). cytokines and growth and angiogenic factors. All three populations Address correspondence and reprint requests to Dr. David M. Mosser, University of of Mfs could confer some level of protection against lethal Maryland, Department of Cell Biology and Molecular Genetics, 3102 Bioscience endotoxemia (7, 16, 17). We, therefore, grouped them together as Research Building, College Park, MD 20742. E-mail address: [email protected] RMfs and conceded that each of these RMfs should be named The online version of this article contains supplemental material. by the stimulus that was used to generate it (1). Importantly, all Abbreviations used in this article: AAMf, alternatively activated Mf; Ade, adeno- f sine; BMDM, bone marrow–derived Mf; CAMf, classically activated Mf; CG, three of these RM s were transcriptionally quite distinct from chlorhexidine gluconate; IC, immune complex; LAde, LPS + Ade; LIC, LPS + IC; IL-4–treated AAMfs. f f f LPGE, LPS + PGE2;M , macrophage; PF, peritoneal fibrosis; RM , regulatory M ; Although there are several studies pertaining to Mf polariza- WT, wild-type. tion, only a very few have attempted to explore how Mfs of one Copyright Ó 2019 by The American Association of Immunologists, Inc. 0022-1767/19/$37.50 polarized state interact and influence the other under basal or

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1900270 2 INHIBITION OF ALTERNATE ACTIVATION BY REGULATORY MACROPHAGES pathological conditions. In the current study, we investigated Fig. 1A), and the expression of these housekeeping genes did not differ how RMfs affect the activation status, phenotype, and function under stimulation conditions (Supplemental Fig. 1B). of AAMfs. The results demonstrate that RMfs are capable of ELISA and Western blot inhibiting the alternate activation state of AAMfsandcanres- cuemicefromAAMf-mediated pathologic fibrosis. Mouse IL-10 and IL-12/23p40 were measured by Duoset ELISA kits (R&D Systems). Mfs were lysed with RIPA buffer containing protease and phosphatase inhibitors (Cell Signaling, Danvers, MA). Anti-OVA Abs Materials and Methods added as IC during stimulation were cleared from protein lysates by in- cubating with Protein A Agarose beads (Santa Cruz Biotechnologies, Mice Dallas, TX) for 2 h at 4˚C. HRP-conjugated actin Ab, anti-Arg1, and rabbit In vitro experiments were performed using cells isolated from 6- to 8-wk- polyclonal Ab against mouse Ym1 were purchased from Santa Cruz old C57BL/6 female mice, and the fibrosis experiments were conducted Biotechnologies, Abcam, and Stemcell Technologies (Vancouver, Canada), using 8-wk-old C57BL/6 female mice, purchased from Taconic Biosciences respectively. 2/2 (Germantown, MD). TACI knockout mice (TACI )onaC57BL/6back- Arginase and urea assays ground were obtained from Dr. Mustafa Akkoyunlu (U.S. Federal Food and Drug Administration, Silver Spring, MD), as previously described (18). Arginase activity and urea levels were measured in Mf lysates using kits Mice were housed at the University of Maryland animal facility in College purchased from Abcam. AAMfs were treated with conditioned media for Park, MD, in a pathogen-free environment. All procedures were ap- 12 h in the DMEM/F12 medium supplemented with 2% FBS. Two million proved by the University of Maryland Institutional Animal Care and Use Mfs were lysed in 100 ml of assay buffer, and 20 ml of lysate was added to Committee. each well for assaying arginase activity. Mouse bone marrow–derived Mf and peritoneal Mf isolation Scratch assay Downloaded from Bone marrow cells were collected from femurs and tibiae and cultured in migration was evaluated with the scratch assay using L-929 DMEM/F12 medium supplemented with 10% FBS, 10 mM L-glutamine, mouse fibroblasts cell lines purchased from the American Type Culture 100 IU/ml penicillin, 100 mg/ml streptomycin, and 15% L929-conditioned Collection (Manassas, VA). The cells were grown to 90% confluency, and medium. On day 7, matured bone marrow–derived Mfs (BMDMs) were the scratch assay was performed in DMEM/F12 media containing 5% FBS. dislodged from petri dishes using Cellstripper solution (Corning) and The monolayer was scratched using a P200 pipette tip, and the cells were f plated for stimulation. Peritoneal M s were obtained by lavaging the allowed to migrate for the indicated times in the presence of media alone or peritoneal cavity once with PBS containing 1% penicillin and streptomy- conditioned media derived from AAMf/RMf/AAMf-RMf admixture. http://www.jimmunol.org/ cin. All stimulations were carried out in 10% FBS–supplemented DMEM/ To observe the direct effects of polarized Mfs on fibroblast migration, F12 medium without L929-conditioned medium. 0.5 million polarized Mfs were seeded in 5-mm upper transwell cham- Mf culture and stimulation bers (Corning) placed on 12-well plates containing L-929 fibroblasts. The scratch closure was imaged using an inverted microscope (Axio Observer; Mfs were cultured and stimulated as described previously (16). Mfs were Zeiss). The cells were counted, and the distance was measured in four treated with 30 ng/ml IL-4 or IL-13 (R&D Systems, Minneapolis, MN) to different view fields of the scratch, using Fiji software (20). generate AAMfs and 20 ng/ml Ultrapure LPS (InvivoGen, San Diego, CA) to generate inflammatory CAMfs. Three different RMf populations Peritoneal fibrosis were obtained by adding 20 ng/ml LPS in combination with one of the PF was induced in mice following a protocol described elsewhere (21, 22). following reprogramming stimuli: 1) OVA-anti-OVA IC, 2) 200 nM PGE2 Eight-week-old C57BL/6 mice weighing ∼20 g were administered 200 ml by guest on September 25, 2021 m (Cayman Chemical, Ann Arbor, MI), or 3) 200 M Ade (Sigma-Aldrich, (∼0.015 mg/g body weight) of 0.1% chlorhexidine gluconate (CG) and St. Louis, MO), which are henceforth referred to as LPS + IC (LIC), LPS + PGE2 15% ethanol (Sigma-Aldrich) dissolved in 13 PBS i.p. every other day for (LPGE), and LPS + Ade (LAde), respectively. All stimulations were car- 3 wk (a total of 10 injections). To avoid any artifacts arising from damage ried out for 8 h. IC was generated by mixing 1 mg/ml chicken OVA to the peritoneum due to repeated injections, CG was injected at the lower (Sigma-Aldrich) and 10-fold molar excess of rabbit anti-OVA (Bio-Techne, part of the peritoneum, and the upper portion of the parietal peritoneum Minneapolis, MN) for 30 min at room temperature. IL-10 stimulations of was used for the histological examination. Mice were divided into six Mfs were carried out at a concentration of 20 ng/ml recombinant mIL-10 f groups for this study: 1) control group treated with 15% ethanol dissolved (R&D). For inhibition assays, AAM s were treated with supernatants in PBS (n = 8); 2) CG group treated with CG in 15% ethanol/PBS (n =8);3) derived from RMfs or cultured in the presence of RMfsin5-mm upper f f f CG + AAM group treated with CG and 1 million AAM s(n = 9); 4) transwell chambers (Corning). For experiments in which RM s were used CG + 2:1 group treated with CG, 0.7 million AAMfs, and 0.3 million for the peritoneal fibrosis (PF) study or added to the transwell chambers, RMfs(n = 5); 5) CG + 1:1 group treated with CG and 0.5 million each of the Mfs were cultured and stimulated in 12-well ultra-low attachment f f f f AAM s and RM s(n = 5); and 6) CG + RM group treated with CG plates (Corning). Following 8 h of stimulation, the M s were dislodged by and 1 million RMfs(n = 9). For groups 3–6, mice were initiated with CG gentle pipetting, washed three times with plain DMEM/F12 media, f m alone for the first three injections, then given in combination with the M s counted, resuspended into a volume of 200 l of complete media, and for the subsequent seven injections. RMfs generated with LIC were used seeded onto transwell chambers. For injection into mice with PF, the for the injections. Two individual experiments were performed, each 5 mo stimulated Mfs were washed three times and rested for 1 h in plain apart, with new sets of reagents and mice. DMEM/F12 medium, washed and resuspended in PBS, and injected. To neutralize the effects of IL-10, BMDMs were treated with rat mAb (clone Pathological scoring and histology 1B1.3a) against IL-10R (CD210; Abcam, Cambridge, MA) 4 h prior to the addition of conditioned media. All animals were sacrificed, and median laparotomy was performed after 21 d. The macroscopic i.p. changes and the extent of fibrosis were graded Real-time PCR semiquantitatively on a 0–3 scale, by scoring in a blinded manner based on 13 different criteria, as described elsewhere (21). Following peritoneal RNA was isolated using the TRIzol-chloroform method (Thermo Fisher flushingtoobtainperitonealMfs, parietal peritoneum, intestines, Scientific, New York, NY), and cDNA was synthesized using SuperScript liver, spleen, and kidney were harvested from mice. The tissues were VILO cDNA Synthesis Kit (Invitrogen, Carlsbad, CA). Real-time PCR was immediately fixed with 4% paraformaldehyde and embedded in par- set up using our published primer pairs (16) and 23 Maxima SYBR Green/ affin blocks. For morphological examination, 4-mm thick, paraffin- ROX qPCR Master Mix (Thermo Fisher Scientific, Waltham, MA). The embedded tissues were deparaffinized and stained with Masson trichrome qPCR was carried out in a LightCycler 480 system (Roche, Reinach, staining. The mounting and staining were performed at Histoserv Switzerland), and the fold induction was calculated by 2^(2DDCT) (Germantown, MD). method (19), using the expression of genes in naive Mfs as control. The geometric mean expression of gapdh and eif3a (eukaryotic translation PCR array initiation factor 3 subunit A) was used for the normalization of the RNA expression profile. The forward and reverse sequences of eif3a are 59-G- Mouse Fibrosis RT2 Profiler PCR Array (PAMM-120Z; Qiagen, Germany) CCCACCAGAAAAGGAGTGG-39 and 59-CTGCATTACGCCAGGAAG- was performed on peritoneal cells collected at the time of sacrifice from GA-39, respectively. No difference was observed in gapdh or eif3a our fibrosis study. The RNA samples were transcribed to cDNA using RT2 normalized expression (representative data shown for ym1 in Supplemental First Strand Kit, mixed with the RT2 SYBR Green Mastermix, and 25 mlof The Journal of Immunology 3 the mixture was added to each well of the PCR array. The 384-well plates (Fig.3A).TotestwhetherRMfs can inhibit the intrinsically oc- (4 3 96 format) were run on a LightCycler 480 system (Roche). curring M2 phenotype, we chose to study Mfs from mice knocked Data and statistical analysis out for TACI (transmembrane activator, calcium modulator, and cyclophilin ligand interactor), which are intrinsically AAMf skewed All data were analyzed with GraphPad Prism version 7 software (San Diego, CA). The heatmap for PCR array was generated using R studio and have higher steady-state expression levels of M2 markers, such version 1.1.442 (23). PCR array data were analyzed in GeneGlobe as CD206, arg1, and ym1 (18, 25). In our study, peritoneal Mfs 2 2 Data Analysis Center (Qiagen). The significance of the observed dif- from TACI / mice expressed higher intrinsic levels of retnla and ferences was evaluated by nonparametric one-way ANOVA. Data are significantly higher expression of ym1 and retnla upon IL-4 stimu- 6 , presented as mean SEM, and p values 0.05 are considered to be lation, compared with wild-type (WT) peritoneal Mfs(Supplemental statistically significant. Fig. 2A). After being stimulated with LPS + reprogramming stimuli, TACI2/2 Mfs were capable of polarizing to regulatory pheno- Results type, as characterized by high IL-10 production and inhibition of RMfs inhibit alternate activation of Mfs LPS-induced TNF-a and IL-12 secretion (Supplemental Fig. 2B). RMfs are defined by their ability to inhibit the proinflammatory When we tested whether RMfs can inhibit the stronger alternate responses of stimulated Mfs (7). Although RMfs are typically activation phenotype in TACI2/2 mice, the supernatants derived confused with AAMfs, we previously demonstrated that RMfs from all three RMf populations of WT or TACI2/2 mice inhibited bore different protein and transcriptional signatures than AAMfs the ym1 and retnla markers expression in TACI2/2 AAMfs (Fig. 3B). and were functionally distinct (16). Hence, we hypothesized that Together, the results indicate that RMfs or their secreted products RMfs would have the ability to modulate alternate activation of can inhibit alternative activation irrespective of the strength, du- Downloaded from Mfs. To test this, cell culture supernatants derived from inflam- ration, or mode of activation. matory (CAMf)Mfs (BMDMs treated with LPS) and three RMfs inhibit AAMf-promoted arginase activity and fibroblast different populations of RMfs (LIC, LPGE, LAde; for details, see proliferation in vitro the Materials and Methods section) were added to the (IL-4– treated) AAMfs, and the expression levels of markers of two well- IL-4 treatment of Mfs increases Arginase 1 (Arg1) enzyme levels

described markers of alternate activation, ym1 and retnla,were through a STAT6-dependent alternative activation pathway, which http://www.jimmunol.org/ assessed. The differentiation of naive MfstoCA-Mf,AAMf,and plays an important role in fibroblast proliferation and fibrosis (26). RMf was ascertained for each experiment by performing real-time Arginase is the key enzyme in the terminal step of the urea cycle PCR for il-10 or ELISA for TNF-a and IL-12. Representative in which arginine is converted to urea and ornithine. From our graphs are shown in Supplemental Fig. 1C. The addition of as little published RNA sequencing data (16), we knew that the arg1 RNA as one-fourth volume of RMf supernatants to AAMfs significantly levels were upregulated by IL-4 treatment of murine Mfs and reduced the mRNA expression of ym1 and retnla (Fig. 1A). The were downregulated in LPS-treated CAMfs and in all three RMf reduction in transcript expression occurred as early as 4 h and populations (Fig. 4A). To assess if the inhibition of markers of continued for 8 and 24 h. Increasing the volume of supernatants to alternate activation by RMfs is associated with reduced down- 500 ml (1:2) or 1 ml (complete) to AAMfs resulted in further in- stream functions of alternative activation, arginase activity and by guest on September 25, 2021 hibition of AAMf transcript expression (Fig. 1B, 1C). To observe a urea levels were assessed in AAMfs treated with or without su- direct cell-dependent effect, similar experiments were carried out by pernatants derived from RMfs. The results demonstrated that both incubating 1 3 106 RMfs that were washed out of their superna- the protein levels and the activity of the arginase enzyme were tants, to transwell chambers containing AAMfs. The inhibition by significantly inhibited by RMfs (Fig. 4B1). This inhibition in RMfs was comparable to the inhibition by supernatants (Fig. 1D, arginase activity was reflected by the significant reduction of urea 1E). The inhibitory effect of RMfs on alternate activation was also levels (Fig. 4B2). observed for primary tissue-resident peritoneal Mfs (Fig. 1F). The AAMfs have long been demonstrated to have a role in tissue observations were further corroborated at the protein level by repair by critically regulating fibroblasts, the key effector cells Western blot analyses of YM1 in AAMfs incubated overnight with involved in creating and collagen structures 1:4 volume of RMf supernatants (Fig. 1G). (27). To study the physiological effects of RMfsonAAMf We next tested if RMfs could inhibit AAMf markers regardless functions, we performed an in vitro scratch assay (28) and of the strength or duration of alternative activation by IL-4. First, assessed the mobility of fibroblasts. The addition of supernatants RMf supernatants (1:4 volume) were added to AAMfs exposed derived from AAMfs accelerated the scratch closure by pro- to increasing concentrations of IL-4 (20, 50, and 100 ng). Irre- moting fibroblast proliferation (Fig. 5A, 5B, top). The addition of spective of the concentration of IL-4, the markers of alternate AAMfs (washed off the IL-4) in the transwell coculture with the activation were inhibited by supernatants derived from all three fibroblasts produced similar results, with the complete closure of subsets of RMfs (Fig. 2A). Second, the supernatants were added the scratch occurring within 24 h (Fig. 5B, bottom). Conversely, after Mfs were stimulated with 30 ng/ml of IL-4 for 4, 8, 12, or the addition of supernatants derived from AAMfs exposed to 1:4 24 h. Again, RMf supernatants inhibited alternative Mf activa- volume of RMfs supernatants significantly inhibited fibroblast tion regardless of the duration of IL-4 priming (Fig. 2B). To ad- proliferation and scratch closure (Fig. 5A, 5B, top). Transwell dress if the level of inhibition is the same throughout different coculture of AAMf/RMf admixture (1:4 cells; washed before kinetics, we reanalyzed the data presented in Fig. 2 and calculated addition) produced a similar reduction in fibroblast migration the percentage of downregulation for different signal strengths. (Fig. 5B, bottom). These results demonstrate that RMfs can The results show that in the presence of 1:4 volume of RMf- inhibit alternative activation of AAMfs and their downstream derived supernatants, ∼50–75% of alternate activation is inhibited, effector functions. regardless of the strength and duration of alternate activation Inhibition of alternate activation by RMfs is not entirely (Supplemental Fig. 1D). IL-13 is another TH2-type cytokine that signals through STAT6 and is capable of inducing alternative dependent on IL-10 activation of Mfs (24). The addition of RMf supernatants to One of the distinguishing features of Mfs with regulatory functions IL-13–primed Mfs also inhibited markers of alternate activation is their ability to secrete high levels of IL-10 (15, 16). To determine 4 INHIBITION OF ALTERNATE ACTIVATION BY REGULATORY MACROPHAGES Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021

FIGURE 1. RMfs inhibit alternative activation of Mfs. BMDMs were treated with 10 ng/ml LPS alone (LPS) or LPS and 25 ml of OVA-anti-OVA IC

(LIC)/200 nM PGE2 (LPGE)/200 mM Ade (LAde) for 8 h to generate three different populations of RMfs. For (A)–(E) and (G), BMDMs were treated with 30 ng/ml of IL-4 for 8 h to generate AAMfs. The resulting AAMfs were incubated with (A) 1:4 volume of RMf-derived supernatant for 4, 8, or 24 h; (B) 1:2 volume or (C) complete volume of RMf-derived supernatants for 6 h; or (D) 1:4 count or (E) 1:2 count of RMfs seeded in a 0.5-mm transwell coculture for 6 h, and the RNA expression levels of alternate activation markers Ym1 or Retnla were measured by quantitative real-time PCR. (F) AAMfs incubated with 1:4 volume of RMf-derived supernatants for 6 h with both AAMfs and RMfs generated from peritoneal Mfs. (G) Representative Western blot and a bar graph summarizing blot analyses of four independent experiments assessing for YM1 protein levels in AAMfs treated with 1:4 volume of RMfs for 12 h. Bars indicate mean 6 SEM of at least four independent experiments. Naive Mfs with no stimulation were used as a control group (fold change = 1; not shown in the figure), and the data were normalized against gapdh and eif3a expression. Data from at least three independent experiments are presented. Significant p values are represented as follows: **p , 0.01, ***p , 0.0001, ****p , 0.00001. if there is a role for IL-10 in inhibiting alternative activation, neutralizing Ab (Fig. 6A). We also tested whether AAMfs from we tested the ability of RMf-derived supernatants to inhibit the the IL-10R–deficient mice would respond to a 1:4 ratio of the expression of ym1 and retnla in AAMfs pretreated with anti–IL-10R RMf-derived supernatants (Fig. 6B). Under both of these conditions, The Journal of Immunology 5 Downloaded from http://www.jimmunol.org/

FIGURE 2. RMfs inhibit AAMfs regardless of the strength and duration of alternate activation. (A) BMDMs were treated with 20, 50, or 100 ng/ml of IL-4 for 8 h to generate AAMfs, followed by treatment with 1:4 volume of RMf-derived supernatants for 6 h. (B) Alternatively, AAMfs generated with 30 ng/ml of IL-4 were treated with 1:4 volume of RMf-derived supernatants for increasing times of 4, 8, 12, or 24 h. The AAMfs were harvested in TRIzol reagent, and the RNA expression levels of alternate activation markers Ym1 or Retnla were measured by quantitative real-time PCR. Mfs from at least four mice per experiment were analyzed. Naive Mfs with no stimulation were used as a control group (fold change = 1; not shown in the figure), and the data were normalized against gapdh and eif3a expression. *p , 0.05, **p , 0.01, ***p , 0.0001. by guest on September 25, 2021 RMfs inhibited the expression of AAMf markers, although the (21) (Fig. 7A). To explore the contribution of different polarized inhibition was somewhat less than that observed with the IL-10– states of Mfs, the mice were administered CG injections for a responsive AAMfs. Finally, both ym1 and retnla continued to be week (i.e., three injections) to initiate the fibrotic process, and then expressed in AAMfs that were exposed to 30 ng/ml of rIL-10 Mfs were coadministered with CG for the subsequent 2 wk (six (Supplemental Fig. 3). Because IL-10 plays a critical role in injections). The mean macroscopic histopathological fibrotic score wound healing by limiting the early inflammatory response and was the highest for CG + AAMf group of mice, and the severity fibroblast proliferation (29), we performed the in vitro scratch was significantly higher than that observed for CG alone, dem- assay on L929 fibroblasts pretreated with the IL-10R neutralizing onstrating that AAMfs worsened CG-induced PF in mice. Ad- Ab. Making fibroblasts unresponsive to IL-10 resulted in better mixture of RMfs (LIC-generated) to CG + AAMf significantly fibroblast proliferation, but itdidnotresultinthecomplete ameliorated the PF in a dose-dependent manner (Fig. 7A). When reversal of inhibition of L929 proliferation by RMf-derived RMfs alone were injected in combination with CG (CG + RMf), supernatants (Fig. 6C, 6D). These results demonstrate that it protected the mice from the CG-induced PF (Fig. 7). The RMf-driven inhibition of alternative activation is not entirely macroscopic observations were further corroborated with the mi- dependent on IL-10. croscopic observations of peritoneal membrane and intestine sections stained with Masson trichrome stain (Fig. 7B). At day 22, RMfs ameliorate AAMf-promoted PF the peritoneal membrane showed loss of mesothelial cells and AAMfs play an important role in wound healing in which RELMa increased thickness of the submesothelial zone, with massive in- produced by AAMfs instructs fibroblasts to produce lysyl hy- filtration and collagen deposition (blue layer) in mice from CG droxylase 2 (LH2), an enzyme that facilitates collagen cross-linking and CG + AAMf groups. Concurrent administration of RMfs (27, 30). However, when the healing process continues unregulated, with CG or CG + AAMf was effective in preventing and re- it results in progressive fibrosis (31, 32). Following our observations versing PF as visualized by reduced membrane thickness. Ad- on the suppression of fibroblast proliferation in the in vitro scratch ministration of CG has been previously demonstrated to affect the assay, we inquired if RMfs could mitigate the profibrotic effects visceral organs, such as colon and liver (33). The macroscopic of AAMfs. observations showed stiffness and rubbery appearance of the in- Long-term peritoneal dialysis often results in PF, which is testines in the CG-administered mice. Masson trichrome staining phenocopied in mice by i.p. injections of CG. Three weeks of i.p. of the peritoneal membrane and intestines showed a significant administration of CG solution induced significant macroscopic i.p. thickening with CG alone or CG + AAMf administration, which changes in mice, including poor adhesion to diaphragm, dullness of was ameliorated in a dose-dependent manner by RMfs (Fig. 7C). liver, “rubber tube”–like appearance of intestines, thickening of Similarly, stained liver tissues exhibited increased collagen intestinal wall, and the encapsulation of i.p. organs like a cocoon deposition and inflammatory foci in CG or AAMf +CGmice, 6 INHIBITION OF ALTERNATE ACTIVATION BY REGULATORY MACROPHAGES Downloaded from http://www.jimmunol.org/

FIGURE 3. RMfs are capable of inhibiting an extrinsically or intrinsically generated AAMf phenotype. (A) To extrinsically generate AAMf phenotype, BMDMs were treated with 30 ng/ml of IL-13 for 8 h. These AAMfs were then incubated with 1:4 or 1:2 volume of RMf-derived supernatants for 6 h. (B) Thioglycollate-elicited peritoneal Mfs obtained from TACI2/2 mice were treated with 30 ng/ml of IL-4 to amplify the intrinsic AAMf response and were then treated with 1:4 volume of supernatants from WT-RMfs(B1 and B3)orTACI2/2-RMfs(B2) for 6 h for the RNA estimation of by guest on September 25, 2021 Ym1 and Retnla or 12 h to obtain lysates for Western blot (B3). The graphs represent mean 6 SEM of independent experiments from at least four mice. Naive Mfs with no stimulation were used as a control group (fold change = 1; not shown in the figure), and the data were normalized against gapdh and eif3a expression. *p , 0.01, **p , 0.001. which was observed to be reduced by RMfs (Supplemental upregulated in the CG + AAMf group compared with the CG- Fig. 4A). alone group (Fig. 8, Supplemental Fig. 4B). Genes associated with RNA expression analysis of 84 key genes involved in tissue proinflammatory functions (tnf, il1a, il1b, stat1) and in TGF-b remodeling and fibrosis was measured in cells of the peritoneal signaling were also upregulated in these mice. Genes associated cavity obtained from CG-injected mice either alone or in combi- with antifibrotic properties (il-10, bmp-7) and tissue remodel- nation with AAMfs/RMfs. Profibrotic genes such as acta-2, ing, such as integrins, were upregulated in CG + RMf injected ccl11, ccl12, ccl3, il-4, and il-113 were highly and significantly mice (Supplemental Fig. 4B). Together, the data show that the

FIGURE 4. RMfs inhibit arginase levels in AAMfs. (A) RNA expression levels of arginase-1 in different RMf populations. (B1) Colorimetric es- timation of arginase activity and estimation of arginase-1 protein levels by Western blot and (B2) colorimetric estimation of urea levels in cell lysates of AAMfs alone or AAMfs cultured in the presence of supernatants derived from three different RMf populations at the indicated volume ratios. Each bar represents the average of at least three independent experiments. The Journal of Immunology 7 Downloaded from http://www.jimmunol.org/

FIGURE 5. RMfs inhibit AAMf-promoted fibroblast proliferation in vitro. Confluent monolayers of L929 mouse fibroblasts were subjected to scratch assay, and the cell migration was monitored by bright field microscopy at original magnification 310 for 6, 12, and 24 h in the presence of DMEM/F12 media alone (medium), supernatant derived from IL-4–treated AAMfs (AAM sup) alone, or with the 1:4 admixture of AAMf supernatant + RMf supernatants derived from three different RMf populations (LIC, LPGE, and LAde). (A) Representative images for each condition and time point are shown. (B)Graphical representation of rate of migration in the presence of different supernatants. (C) Alternatively, the migration was monitored in the presence of 1 3 106 AAMfs alone or mixed with a 1:4 ratio of various RMfs in the transwell inserts. by guest on September 25, 2021 administration of AAMfs with CG worsened fibrosis, and the Despite several attempts to arrive at a consensus nomenclature addition of RMfs ameliorated CG-induced PF and rescued mice (1, 34), there remains some confusion about the different Mf from AAMf-mediated pathological conditions. phenotypes and their biological significance. One of the persistent inaccuracies is to classify together all Mfs that are not inflam- Discussion matory into a “catch-all” M2 category. This has impeded Mf Inflammatory Mfs arise in response to pathogen- or damage- research by creating confusion in defining markers and biological associated molecular patterns at the site of tissue injury. If the functions for the various Mf subsets and sometimes even im- primary role of Mfs is to maintain homeostasis, then it is logical plying in the extreme that Mfs exist only as two subsets. Through to assume that Mfs should be capable of not only producing in- a series of studies, including this one, we established strong evi- flammatory mediators and growth factors in response to tissue dence for subsets of Mfs with immunoregulatory and angiogenic injury but also anti-inflammatory and regulatory mediators to re- activity, which we collectively refer to as RMfs. In the pre- store homeostasis. In a series of articles (7, 13–16) we identified sent work, we show that these RMfs are not only distinct from and characterized a population of Mfs termed RMfs that could both the traditional inflammatory CAMfs and the IL-4–derived inhibit inflammatory responses by CAMfs, capable of promoting AAMfs but are actually capable of regulating both of these Mf an anti-inflammatory, growth-promoting transcriptional program subsets. Now, in addition to these cells having distinct transcrip- in them. In the present work, we sought to examine whether the tional, metabolic, and protein signatures (16), we can ascribe a RMf that inhibited inflammatory responses could also regulate broader regulatory role for these Mfs. We emphasize that we are the overt proregenerative and profibrotic response of AAMfs. We describing three different populations of Mfs generated with three demonstrate that RMfs were capable of inhibiting the alterna- distinct reprogramming signals. However, all three of these RMf tive activation phenotype of AAMfs generated in vitro and the populations, RMf-IC, RMf-Ade, and RMf-PGE2, were able to intrinsically occurring AAMfs derived from TACI2/2 mice. rescue mice from lethal endotoxemia (16), and all three of these This inhibition occurred regardless of the strength, duration, and populations inhibited the expression of Relma (retnla) and Ym1 mode of alternative activation and was only partially dependent (chi3l3) on AAMf. on IL-10. In the CG-induced PF model, RMfs rescued mice Our standard approach to generate RMfs was to stimulate naive from AAMf-mediated pathological conditions. Taken together, Mfs with a TLR ligand, such as LPS, along with a secondary our study demonstrates that the inhibition of the alternative ac- signal of IC/PGE2/Ade (16). We speculate that there may be many tivation phenotype by RMfs might play an important role in other possible secondary stimuli, which we have yet to explore, in- restricting pathologies associated with AAMfs. cluding resolvins, histamines, sphingosine 1-phosphate, melanocortin, 8 INHIBITION OF ALTERNATE ACTIVATION BY REGULATORY MACROPHAGES Downloaded from http://www.jimmunol.org/

FIGURE 6. Inhibition of alternate activation by RMfs is not entirely IL-10 dependent. (A) Alternatively activated BMDMs (AAMfs) were generated by adding IL-4 for 6 h. To block IL-10 signaling, the AAMfs were treated or not with anti–IL-10R (CD210) Ab for 4 h prior to the addition of conditioned media. (B) BMDMs derived from IL-10R2/2 mice were polarized to AAMfs by adding IL-4 for 6 h. (C and D) The scratch closure was measured in L929 by guest on September 25, 2021 mouse fibroblasts, whose IL-10 signaling was blocked by addition of CD210. (C) A representative image of cell migration at 24 h after addition of conditioned media is shown (original magnification 310). (D) Graphical representation of cell migration in the presence of different supernatants is shown. Cells were treated with 1:4 volume of supernatants derived from Mfs treated with LPS, each of the RMf populations (LIC, LPGE, or LAde). After 8 h, the RNA expression levels of ym1 and retnla were measured. The figures represent mean values from at least three independent experiments. Naive Mfs with no stimulation were used as a control group (fold change = 1; not shown in the figure), and the data were normalized against gapdh and eif3a expression. *p , 0.05, **p , 0.01, ***p , 0.0001. vasoactive intestinal peptides, adiponectin, or even a toxin from reported excessive human skin fibrosis (lipodermatosclerosis) Bacillus anthracis, which is a GPCR ligand (3). The secondary under unrestrained retnlb-positive AAMf activity. Similarly, un- stimuli reprograms Mfs to turn down the production of inflam- checked AAMfshavebeendemonstratedtounderliehuman matory mediators, such as IL-1, IL-6, and IL-12, and inflamma- idiopathic pulmonary fibrosis (35), pancreatic fibrosis in chronic tory chemokines while inducing increased production of IL-10 pancreatitis (36), PF (37), and several other fibrotic patholo- and a myriad of growth and angiogenic factors. The requirement gies (32, 38). These studies indicate that a dysregulated state of for a secondary stimuli to reprogram Mfs to a regulatory phe- alternative activation can lead to pathological conditions. Our notype suggests that RMfs may naturally arise after the primary in vitro observations that RMfs are capable of inhibiting alter- Mf stimulation to dampen Mf effector functions. The results native activation in Mfs (Fig. 1) led us to hypothesize a protective from the in vitro coculture experiments demonstrated that when role for RMfs in pathological fibrosis. The simple in vitro scratch AAMfs came in contact with the RMfs or their supernatants, the experiments with mouse L929 fibroblast (Fig. 5) confirmed the markers of alternate activation and related functional molecules profibrotic role for AAMfs, in which the closure of the scratch (arginase and urea) were inhibited. Surprisingly, Mfs stimulated was faster in the presence of AAMf-conditioned media or AAMfs with LPS alone also had a low level of AAMf inhibitory activity. in a transwell coculture. The addition of RMfsortheirderived This inhibition often required a higher volume or longer duration media suppressed the migration of fibroblasts and scratch closure. of incubation with AAMfs to achieve a comparable level of in- Treatment with LPS has been reported to accelerate or decelerate hibition by RMfs. We speculate that this inhibition may be a migration, depending on the concentration, cell type, and FBS used reflection of the natural progression of inflammatory Mfs toward by various studies (39–41). Addition of LPS with AAMf did not a regulatory phenotype to maintain homeostasis. have any noticeable effect on the closure in our experiments. This The contribution of AAMfs to wound healing and tissue repair could be due to the low concentration of LPS (10 ng/ml) we use in has been previously demonstrated by several groups (30). In skin our studies, which might not be enough to induce significant pro- repair studies, IL-4Ra–dependent Mf activation controlled col- liferative or migratory changes in the L929 cells. Further, we have lagen fibril assembly and repair (27). However, the same study used 5% FBS in our culture, which could have contributed to some The Journal of Immunology 9 Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021

FIGURE 7. RMfs reduce AAMf-promoted CG-induced PF. PF was induced by administration of CG. Mfs were coadministered along with CG after the first three CG-alone injections for 2 wk. (A) Anatomic display of fibrosis developed in CG alone or CG + AAMf group demonstrating thickening of peritoneum membrane, reduced adhesion (top panel, arrows), dullness of liver, intestines with thick rubber tube–like appearance, and contracted mesenterium (bottom panel, arrows). The fibrosis effect is ameliorated in the group receiving CG + RMf (B). Fibrosis was morphometrically scored using the system described by Suga et al. (21). Average fibrosis score for each condition 6 SEM is represented in the bar graph. ***p , 0.001 compared with control, ##p , 0.01 compared with CG group. (C) Photomicrographs illustrate Masson trichrome staining of peritoneal membrane and intestines collected at the time of sacrifice of mice administered with CG alone, CG + AAMf,CG+RMf, or CG + AAMf/RMf admixture. The pictures are representative sections from three different mice samples, captured with original magnification 320. Enhanced submesothelial zone thickness (vertical arrows) of peritoneum and extracellular matrix deposition in intestines (blue area, arrows) can be observed for mice from the CG or CG + AAMf group. RMfs generated with LIC were used in the injections. The increase in the thickening of layers and the collagen deposition are shown by the arrows. basal level of proliferation in our system. Nevertheless, the effect, if inhibitory effects of RMfs, suggesting that RMfsexertboth any, presumably will be consistent across culture conditions. IL-10–dependent and –independent mechanisms of fibropro- Many of the known functions of RMfs have been attributed to liferative inhibition. the IL-10 that is secreted by these cells (16, 42). We therefore The mouse model of CG-induced PF resembles human disease selected this cytokine for further investigation. To our surprise, characterized by increased thickness of the submesothelial tissues IL-10 appeared to make a modest contribution to the inhibition of of the peritoneum and prominent Mf infiltration. Studies focusing IL-4–driven alternative activation. RMfs deficient in IL-10 were on the role of Mfs in PF identified a profibrotic role for AAMfs, capable of partially inhibiting alternative activation, and AAMfs demonstrating high expression of ym1, yetnla, and arg-1 by Mfs treated with anti–IL-10R Ab remained sensitive to RMfs. IL-10 isolated from the peritoneum (44, 45). These studies showed has been linked with the mitigation of fibrosis based on its ability that administration of liposome-encapsulated clodronates, to to inhibit the inflammation that drives fibroproliferation (32). In a deplete Mfs, mitigated PF in animals. The administration randomized trial, administration of rIL-10 to 24 chronic hepatitis of liposome-encapsulated clodronates targeting Mfs has also C patients who had not previously responded to IFN-based ther- proven to be effective in reducing fibrosis of other organs, such apy decreased hepatic inflammation and reduced liver fibrosis as lung, renal, and biliary fibrosis (35, 38). The intriguing as- (43). Similar results have been reported for patients receiving pect of our study is that the RMfs ameliorated CG-induced renal transplants and patients with Crohn disease (29). We ob- fibrosis and rescued mice from the AAMf-mediated patho- served that treatment of L929 fibroblasts with anti–IL-10R Ab logical fibrosis, providing direct evidence for the antifibrotic facilitated the scratch closure but did not completely relieve the role for RMfs. The historical approach of Mf depletion might 10 INHIBITION OF ALTERNATE ACTIVATION BY REGULATORY MACROPHAGES

FIGURE 8. RMfs modulate fibrotic gene expres- sion in peritoneal cells of CG-induced fibrosis in mice. Differential expression of genes identified by mouse fibrosis array in cells from the peritoneal cavity of mice administered with CG alone (CG) or in combination with IL-induced AAMfs (AAM + CG)/LIC-induced Downloaded from RMfs (RMf + CG). Heatmap represents genes with significant upregulation (green) or downregulation (red) compared with the cells from PBS-administered control mice. Fold differences in gene regulation are shown for relevant antifibrotic, profibrotic, and in- flammatory genes in the color-coded tables. Five http://www.jimmunol.org/ housekeeping genes present in the array were used for the normalization of mRNA expression analysis. Each value represents the mean of two independent arrays run on RNA samples pooled from three independent experiments each. by guest on September 25, 2021

have overlooked the possibility of depleting not only patho- The results from this study indicate that Mfs not only exist in logical Mfs but also a subset of Mfs with a potentially pro- different polarized states but can also influence the functions of tective role. Although we have not addressed the mechanism one another through direct or indirect interactions. The ability of of fibrosis inhibition in this study, we are currently working RMfs to regulate AAMfs highlights another important role for toward identifying the most relevant molecules using high- these cells and may provide a scientific rationale for designing throughput approaches. new Mf-targeted therapies. In this study, we describe the inhibition of fibrosis and al- ternative activation by RMfs. We hypothesize that in normal Disclosures instances of wound healing, Mfs become reprogrammed to D.M.M. has a financial interest in a company (LeukoSight) that seeks to assume a regulatory phenotype to restore homeostasis. The manipulate immune responses at the level of Mfs. The other authors have mechanism(s) of RMf-mediated inhibition of fibrosis have not no financial conflicts of interest. been resolved by this study. However, the transcriptional pro- filing suggests that RMfs drive higher expression of anti- References fibrotic genes, such as Bmp7 and Il-10, with the concomitant 1. Murray, P. J., J. E. Allen, S. K. Biswas, E. A. Fisher, D. W. Gilroy, S. Goerdt, reduction in the expression of profibrotic and inflammatory S. Gordon, J. A. Hamilton, L. B. Ivashkiv, T. Lawrence, et al. 2014. Macrophage genes in the peritoneal cells of CG-administered mice. Insuf- activation and polarization: nomenclature and experimental guidelines. Immunity 41: 14–20. ficient RMf responses might lead to an extended repair pro- 2. Ginhoux, F., and M. Guilliams. 2016. Tissue-resident macrophage ontogeny and cess resulting in pathological conditions. This hypothesis needs homeostasis. Immunity 44: 439–449. 3. Mosser, D. M., and J. P. Edwards. 2008. Exploring the full spectrum of macrophage to be confirmed, and subsequent studies will be geared to- activation. [Published erratum appears in 2010 Nat. Rev. Immunol. 10: 460.] Nat. ward understanding the late events during repair and fibrosis. Rev. Immunol. 8: 958–969. The Journal of Immunology 11

4. Nau, G. J., J. F. Richmond, A. Schlesinger, E. G. Jennings, E. S. Lander, and 27. Knipper, J. A., S. Willenborg, J. Brinckmann, W. Bloch, T. Maaß, R. Wagener, R. A. Young. 2002. Human macrophage activation programs induced by bac- T. Krieg, T. Sutherland, A. Munitz, M. E. Rothenberg, et al. 2015. Interleukin-4 terial pathogens. Proc. Natl. Acad. Sci. USA 99: 1503–1508. receptor a signaling in myeloid cells controls collagen fibril assembly in skin 5. Martinez, F. O., S. Gordon, M. Locati, and A. Mantovani. 2006. Transcriptional repair. Immunity 43: 803–816. profiling of the human -to-macrophage differentiation and polarization: 28. Liang, C.-C., A. Y. Park, and J.-L. Guan. 2007. In vitro scratch assay: a con- new molecules and patterns of gene expression. J. Immunol. 177: 7303–7311. venient and inexpensive method for analysis of cell migration in vitro. Nat. 6. Stein, M., S. Keshav, N. Harris, and S. Gordon. 1992. potently Protoc. 2: 329–333. enhances murine macrophage mannose receptor activity: a marker of alternative 29. Sziksz, E., D. Pap, R. Lippai, N. J. Be´res, A. Fekete, A. J. Szabo´,andA´ . Vannay. immunologic macrophage activation. J. Exp. Med. 176: 287–292. 2015. Fibrosis related inflammatory mediators: role of the IL-10 cytokine family. 7. Edwards, J. P., X. Zhang, K. A. Frauwirth, and D. M. Mosser. 2006. Biochemical Mediators Inflamm. 2015: 764641. and functional characterization of three activated macrophage populations. 30. Chen, F., Z. Liu, W. Wu, C. Rozo, S. Bowdridge, A. Millman, N. Van Rooijen, J. Leukoc. Biol. 80: 1298–1307. J. F. Urban, Jr., T. A. Wynn, and W. C. Gause. 2012. An essential role for TH2- 8. Mantovani, A., A. Sica, S. Sozzani, P. Allavena, A. Vecchi, and M. Locati. 2004. type responses in limiting acute tissue damage during experimental helminth The chemokine system in diverse forms of macrophage activation and polari- . Nat. Med. 18: 260–266. zation. Trends Immunol. 25: 677–686. 31. Gordon, S., and F. O. Martinez. 2010. Alternative activation of macrophages: 9. Martinez, F. O., and S. Gordon. 2014. The M1 and M2 paradigm of macrophage mechanism and functions. Immunity 32: 593–604. activation: time for reassessment. F1000Prime Rep. 6: 13. 32. Wilson, M. S., and T. A. Wynn. 2009. Pulmonary fibrosis: pathogenesis, etiology 10. Qian, B. Z., and J. W. Pollard. 2010. Macrophage diversity enhances tumor and regulation. Mucosal Immunol. 2: 103–121. progression and metastasis. Cell 141: 39–51. 33. Lee, C.-J., Y.-M. Subeq, R.-P. Lee, H.-H. Liou, and B.-G. Hsu. 2014. Cal- 11. Kadl, A., A. K. Meher, P. R. Sharma, M. Y. Lee, A. C. Doran, S. R. Johnstone, citriol decreases TGF-b1 and angiotensin II production and protects against M. R. Elliott, F. Gruber, J. Han, W. Chen, et al. 2010. Identification of a novel chlorhexide digluconate-induced liver peritoneal fibrosis in rats. Cytokine macrophage phenotype that develops in response to atherogenic phospholipids 65: 105–118. via Nrf2. Circ. Res. 107: 737–746. 34. Guilliams, M., F. Ginhoux, C. Jakubzick, S. H. Naik, N. Onai, B. U. Schraml, 12. Gilroy, D., and R. De Maeyer. 2015. New insights into the resolution of in- E. Segura, R. Tussiwand, and S. Yona. 2014. Dendritic cells, and flammation. Semin. Immunol. 27: 161–168. macrophages: a unified nomenclature based on ontogeny. Nat. Rev. Immunol. 14: 13. Sutterwala, F. S., G. J. Noel, P. Salgame, and D. M. Mosser. 1998. Reversal of 571–578. Downloaded from proinflammatory responses by ligating the macrophage Fcgamma receptor type 35. Murray, L. A., Q. Chen, M. S. Kramer, D. P. Hesson, R. L. Argentieri, X. Peng, I. J. Exp. Med. 188: 217–222. M. Gulati, R. J. Homer, T. Russell, N. van Rooijen, et al. 2011. TGF-beta driven 14. Sutterwala, F. S., G. J. Noel, R. Clynes, and D. M. Mosser. 1997. Selective lung fibrosis is macrophage dependent and blocked by serum amyloid P. Int. J. suppression of interleukin-12 induction after macrophage receptor ligation. Biochem. Cell Biol. 43: 154–162. J. Exp. Med. 185: 1977–1985. 36. Xue, J., V. Sharma, M. H. Hsieh, A. Chawla, R. Murali, S. J. Pandol, and 15. Lucas, M., X. Zhang, V. Prasanna, and D. M. Mosser. 2005. ERK activation A. Habtezion. 2015. Alternatively activated macrophages promote pancreatic following macrophage FcgammaR ligation leads to chromatin modifications at fibrosis in chronic pancreatitis. Nat. Commun. 6: 7158.

the IL-10 locus. J. Immunol. 175: 469–477. 37. Bello´n,T.,V.Martı´nez, B. Lucendo, G. del Peso, M. J. Castro, L. S. Aroeira, http://www.jimmunol.org/ 16.Fleming,B.D.,P.Chandrasekaran,L.A.Dillon,E.Dalby,R.Suresh,A.Sarkar, A. Rodrı´guez-Sanz, M. Ossorio, R. Sa´nchez-Villanueva, R. Selgas, and N. M. El-Sayed, and D. M. Mosser. 2015. The generation of macrophages with anti- M. A. Bajo. 2011. Alternative activation of macrophages in human perito- inflammatory activity in the absence of STAT6 signaling. J. Leukoc. Biol. 98: 395–407. neum: implications for peritoneal fibrosis. Nephrol. Dial. Transplant. 26: 17. Hamidzadeh, K., S. M. Christensen, E. Dalby, P. Chandrasekaran, and 2995–3005. D. M. Mosser. 2017. Macrophages and the recovery from acute and chronic 38. Kitamoto, K., Y. Machida, J. Uchida, Y. Izumi, M. Shiota, T. Nakao, H. Iwao, inflammation. Annu. Rev. Physiol. 79: 567–592. T. Yukimura, T. Nakatani, and K. Miura. 2009. Effects of liposome clodronate on 18. Allman, W. R., R. Dey, L. Liu, S. Siddiqui, A. S. Coleman, P. Bhattacharya, renal leukocyte populations and renal fibrosis in murine obstructive nephropathy. M. Yano, K. Uslu, K. Takeda, H. L. Nakhasi, and M. Akkoyunlu. 2015. TACI J. Pharmacol. Sci. 111: 285–292. deficiency leads to alternatively activated macrophage phenotype and suscepti- 39. Zheng, X., W. Zhang, and X. Hu. 2018. Different concentrations of bility to Leishmania infection. Proc. Natl. Acad. Sci. USA 112: E4094–E4103. regulate barrier function through the PI3K/Akt sig- 19. Livak, K. J., and T. D. Schmittgen. 2001. Analysis of relative gene expression nalling pathway in human pulmonary microvascular endothelial cells. Sci.

data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Rep. 8: 9963. by guest on September 25, 2021 Methods 25: 402–408. 40. Li, X. J., F. Z. Huang, Y. Wan, Y. S. Li, W. K. Zhang, Y. Xi, G. H. Tian, and 20. Schindelin, J., I. Arganda-Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, H. B. Tang. 2018. Lipopolysaccharide stimulated the migration of NIH3T3 cells S. Preibisch, C. Rueden, S. Saalfeld, B. Schmid, et al. 2012. Fiji: an open-source through a positive feedback between b-Catenin and COX-2. Front. Pharmacol. platform for biological-image analysis. Nat. Methods 9: 676–682. 9: 1487. 21. Suga, H., S. Teraoka, K. Ota, S. Komemushi, S. Furutani, S. Yamauchi, and 41. Vuong, T. T., T. M. Reine, A. Sudworth, T. G. Jenssen, and S. O. Kolset. 2015. S. Margolin. 1995. Preventive effect of pirfenidone against experimental sclerosing Syndecan-4 is a major syndecan in primary human endothelial cells in vitro, peritonitis in rats. Exp. Toxicol. Pathol. 47: 287–291. modulated by inflammatory stimuli and involved in wound healing. J. Histochem. 22. Kurata, K., S. Maruyama, S. Kato, W. Sato, J. Yamamoto, T. Ozaki, A. Nitta, Cytochem. 63: 280–292. T. Nabeshima, Y. Morita, M. Mizuno, et al. 2009. Tissue-type plasminogen 42. Verreck, F. A., T. de Boer, D. M. Langenberg, M. A. Hoeve, M. Kramer, activator deficiency attenuates peritoneal fibrosis in mice. Am. J. Physiol. Renal E. Vaisberg, R. Kastelein, A. Kolk, R. de Waal-Malefyt, and T. H. Ottenhoff. Physiol. 297: F1510–F1517. 2004. Human IL-23-producing type 1 macrophages promote but IL-10-producing 23. RStudio Team. 2015. RStudio: integrated development for R. RStudio, Inc., Boston, type 2 macrophages subvert immunity to (myco)bacteria. Proc. Natl. Acad. Sci. MA. Available at: https://www.rstudio.com/. Accessed: August 28, 2018. USA 101: 4560–4565. 24. Van Dyken, S. J., and R. M. Locksley. 2013. Interleukin-4- and interleukin-13- 43. Nelson, D. R., G. Y. Lauwers, J. Y. Lau, and G. L. Davis. 2000. mediated alternatively activated macrophages: roles in homeostasis and disease. treatment reduces fibrosis in patients with chronic hepatitis C: a pilot trial of Annu. Rev. Immunol. 31: 317–343. interferon nonresponders. Gastroenterology 118: 655–660. 25. Liu, L., W. R. Allman, A. S. Coleman, K. Takeda, T. L. Lin, and M. Akkoyunlu. 44. Kushiyama, T., T. Oda, M. Yamada, K. Higashi, K. Yamamoto, N. Oshima, 2018. Delayed onset of autoreactive production and M2-skewed mac- Y. Sakurai, S. Miura, and H. Kumagai. 2011. Effects of liposome-encapsulated rophages contribute to improved survival of TACI deficient MRL-Fas/Lpr clodronate on chlorhexidine gluconate-induced peritoneal fibrosis in rats. mouse. Sci. Rep. 8: 1308. Nephrol. Dial. Transplant. 26: 3143–3154. 26. Murthy, S., J. L. Larson-Casey, A. J. Ryan, C. He, L. Kobzik, and A. B. Carter. 45.Wang,J.,Z.-P.Jiang,N.Su,J.-J.Fan,Y.-P.Ruan,W.-X.Peng,Y.-F.Li,and 2015. Alternative activation of macrophages and pulmonary fibrosis are modu- X.-Q. Yu. 2013. The role of peritoneal alternatively activated macrophages in lated by scavenger receptor, macrophage receptor with collagenous structure. the process of peritoneal fibrosis related to peritoneal dialysis. Int. J. Mol. Sci. FASEB J. 29: 3527–3536. 14: 10369–10382. Supplementary Figure 1.

Supplementary Figure 1: Induction of regulatory macrophages. (A) BMDMs were treated with 20ng/ml of IL-4 and the expression of ym1 were assessed by real time PCR. Both gapdh and eif3a normalized data are presented. (B) RNA expression levels of gapdh and eif3a in regulatory macrophage populations (top panel) and AAMs treated with supernatants of CAMs or RMs (bottom panel). (C) BMDMs were treated with 10 ng/ml LPS alone or a combination of LPS and 25 µl Ova/anti-Ova ICs (LIC, 200 nM PGE2 (LPGE), 200 μM Ado (LAde) for 8 h. The mRNA expression of Il-10 was measure by qRTPCR and IL-12/23p40 and TNF- protein levels were measured in the supernatants by ELISA. Error bars indicate means ± SEM of 6 independent experiments. **p < 0.01, and *p< 0.05. (D) Reanalyzed data of Main Figure 2. The percentage of alternate activation inhibition by RM was calculated for ym1 and retnla with respective to their expression levels in AAM. Supplementary Figure 2.

Supplementary Figure 2: Macrophages derived from TACI-/- are intrinsically biased toward an alternatively activated phenotype, but can be induced to a regulatory phenotype. (A) Peritoneal macrophages from TACI-/- mice expressed higher RNA levels of markers of alternate activation when stimulated with IL-4 for 6 hours. (B). Peritoneal macrophages (PMs) from TACI-/- mice were treated with 10 ng/ml LPS alone or a combination of LPS and 25 µl Ova/anti-Ova ICs (L+I), 200 nM PGE2 (L+P), 200 μM Ado (L+A) for 8 h, followed by estimation of IL-10, IL-12/23p40 and TNF- levels in the supernatants by ELISA. Error bars indicate means ± SEM of 3 independent experiments. ***p<0.001, **p < 0.01, and *p< 0.05. Supplementary Figure 3.

Supplementary Figure 3: IL-10 alone does not reduce the expression of markers of alternative activation in AAMs. BMDMs were induced to alternative activation by treating with IL-4 or IL-13 for 8 hours. Mouse recombinant IL-10 was added at the indicated concentrations and cells were incubated overnight before the estimation of RNA expression of ym1 and retnla by qRTPCR. Data represents mean±SEM from three independent experiments. Supplementary Figure 4.

Supplementary Figure 4: Regulatory macrophages modulate fibrosis-related gene expression in CG-induced fibrosis in mice. (A) Masson trichrome staining of liver from mice with CG-induced peritoneal fibrosis. Mice were administered with PBS, CG-alone (CG), CG with alternatively activated macrophages (AAM+CG) or CG with regulatory macrophages (RMac+CG). The collagen deposition (blue) is indicated with the arrowheads. (B) Differential RNA expression from cells from the peritoneal cavity of mice administered with CG-alone (CG) or in combination with IL- induced AAMs (AAM+CG) /LPS + immune complex-induced RMs (RMac+CG). Fold differences in expression is shown for genes with indicated functions. Each value represents mean of two independent PCR arrays run on RNA samples pooled from three independent experiments each.