TLR2 Signaling in Tubular Epithelial Cells Regulates NK Cell Recruitment in −

This information is current as Hye J. Kim, Jong S. Lee, Ahra Kim, Sumi Koo, Hee J. Cha, of September 29, 2021. Jae-A Han, Yoonkyung Do, Kyung M. Kim, Byoung S. Kwon, Robert S. Mittler, Hong R. Cho and Byungsuk Kwon J Immunol 2013; 191:2657-2664; Prepublished online 31 July 2013;

doi: 10.4049/jimmunol.1300358 Downloaded from http://www.jimmunol.org/content/191/5/2657

Supplementary http://www.jimmunol.org/content/suppl/2013/07/31/jimmunol.130035 Material 8v1.DC1 http://www.jimmunol.org/ References This article cites 97 articles, 15 of which you can access for free at: http://www.jimmunol.org/content/191/5/2657.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 29, 2021

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

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

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2013 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

TLR2 Signaling in Tubular Epithelial Cells Regulates NK Cell Recruitment in Kidney Ischemia–Reperfusion Injury

Hye J. Kim,*,1 Jong S. Lee,*,†,1 Ahra Kim,‡ Sumi Koo,* Hee J. Cha,*,x Jae-A Han,{ Yoonkyung Do,{ Kyung M. Kim,†,‖ Byoung S. Kwon,# Robert S. Mittler,‡,** Hong R. Cho,*,†† and Byungsuk Kwon*,‡

Damage-associated molecular patterns released from damaged kidney cells initiate postischemic inflammation, an essential step in the progression of kidney ischemia–reperfusion injury (IRI). However, the mechanism that coordinates this highly specific process in ischemic kidneys remains to be clarified. Previously, we demonstrated that CD137 from NK cells specifically stimulates CD137 ligand (CD137L) on tubular epithelial cells (TECs) such that TECs produced the high CXCR2 chemokine levels required for chemotaxis. We report in the present study that endogenous TLR2 ligands released from ischemic TECs induce CCR5 chemokine expression, which is critical to promoting NK cell recruitment. By implanting CD137L2/2 TECs into the kidney Downloaded from capsule of TLR22/2 mice, we further showed that TLR2-mediated NK cell recruitment is an uncoupled event that can occur independently of CD137L signaling in TECs, which is responsible for recruiting . Therefore, our findings identify TECs as both a target for kidney damage and also as a master regulator that actively modulates stepwise signaling, leading to the initiation and amplification of acute sterile inflammation that inflicts kidney IRI. Being clinically important, the signaling pathway of innate receptors in epithelial cells may therefore be a good target to block acute sterile inflammation resulting from tissue damage, including kidney IRI. The Journal of Immunology, 2013, 191: 2657–2664. http://www.jimmunol.org/

nflammation is a physiological response that manages both recognize a number of endogenous molecules released from in- external and internal danger signals and restores jured cells (3). Evidence of TLR involvement in kidney ischemia– I (1). Sterile inflammation, caused by tissue damage without reperfusion injury (IRI) is accumulating (4–8). Endogenous TLR2 infection, can occur in conditions such as trauma, ischemia, is- and TLR4 ligands are released from the tubular after chemia and reperfusion (IR), or cellular malfunction (1–3). Sen- IR (9–11). Tubular epithelial cells (TECs) seem to play a central sors that recognize molecules released by damaged tissues induce role in initiating kidney IRI, as they express TLR2 and TLR4. inflammation through signaling that results in the production of Additionally, a parenchymal cell deficiency of either gene dramat- by guest on September 29, 2021 inflammatory mediators by tissue-specific immune cells or pa- ically limits kidney injury following IR and diminishes the asso- renchymal cells (1). TLRs are the most well-known sensors and ciated intrarenal inflammation (4, 5). In addition to this autocrine mechanism, endogenous TLR ligands may act on other cell types such as endothelial cells and innate immune cells in a paracrine *Biomedical Research Center, Ulsan University Hospital, School of Medicine, Uni- manner to promote IRI-associated acute inflammation. TLR2 or versity of Ulsan, Ulsan 682-714, Korea; †Department of Internal Medicine, Ulsan University Hospital, School of Medicine, University of Ulsan, Ulsan 682-714, Korea; TLR4 signaling results in proinflammatory cytokine and chemokine x ‡School of Biological Sciences, University of Ulsan, Ulsan 680-749, Korea; Depart- production, but little is known regarding how the TLR inflammatory ment of Pathology, Ulsan University Hospital, School of Medicine, University of Ulsan, Ulsan 682-714, Korea; {School of Nano-Bioscience and Chemical Engineer- cascade is triggered by cell–cell interactions among TECs, stromal ing, Ulsan National Institute of Science and Technology, Ulsan 689-798, Korea; cells, endothelial cells, and infiltrating leukocytes after IR. ‖ Department of Internal Medicine, Dongnam Institute of Radiological and Medical We have previously identified an inflammatory loop between Sciences, Busan 619-953, Korea; #Division of Cell and Immunobiology, Research and Development Center for Cancer Therapeutics, National Cancer Center, Ilsan 410- TECs and inflammatory cells during kidney IRI (12). NK cells 769, Korea; **Department of Surgery, Emory Vaccine Center, Emory University infiltrate the kidney early after IR and approach the tubular epithe- †† School of Medicine, Atlanta, GA 30322; and Department of Surgery, Ulsan Uni- lium across the basement membrane, which is rapidly destroyed versity Hospital, School of Medicine, University of Ulsan, Ulsan 680-749, Korea by IRI (12, 13). Infiltrating NK cells in turn stimulate TECs 1H.J.K. and J.S.L. contributed equally to this work. to produce CXCR2 chemokines by stimulating CD137 ligand Received for publication February 6, 2013. Accepted for publication June 27, 2013. (CD137L) signaling and inducing neutrophil migration (12). Re- This work was supported by grants from the National Research Foundation of Korea cent studies have revealed a positive feedback among dendritic funded by the Ministry of Education, Science and Technology (Grants NRF-2011- 0012596, BRL2009-0087350, and 2009-0094050) and by Chong Keun Dang Phar- cells, NKT cells, and neutrophils to amplify acute inflammation maceutical Corp., Seoul, Korea. leading to kidney IRI (14). Therefore, the events following IR and Address correspondence and reprint requests to Prof. Byungsuk Kwon or Dr. Hong the complex network of stromal cells, endothelial cells, TECs, and R. Cho, School of Biological Sciences, University of Ulsan, San29 Mukeo-dong, immune cells should be defined to better understand IRI-asso- Nam-ku, Ulsan 680-749, Korea (B.K.) or Department of Surgery, Ulsan University Hospital, University of Ulsan, Ulsan 682-714, Korea (H.R.C.). E-mail addresses: bkwon@ ciated inflammation. mail.ulsan.ac.kr (B.K.) or [email protected] (H.R.C.) Currently, the regulation of signals that recruit NK cells fol- The online version of this article contains supplemental material. lowing IRI is not well understood. The aim of this study was to Abbreviations used in this article: BUN, ; CD137L, CD137 determine the mechanism of NK cell recruitment to the kidney af- ligand; HMGB1, high mobility group box 1; IR, ischemia and reperfusion; IRI, ter IR. We found that TECs produced CCR5 chemokines, which ischemia–reperfusion injury; TEC, tubular epithelial cell; WT, wild-type. are required for NK cell chemotaxis, early after IR in a TLR2- Copyright Ó 2013 by The American Association of Immunologists, Inc. 0022-1767/13/$16.00 dependent manner. Moreover, NK cell and neutrophil recruit- www.jimmunol.org/cgi/doi/10.4049/jimmunol.1300358 2658 NK CELL RECRUITMENT BY TLR2 SIGNALING ment by TECs was a sequential and independent event. These sues were passed through a 40-mm cell strainer (BD Falcon) using the results indicate that epithelial cells actively regulate sterile tissue rubber end of a 1-ml syringe plunger, and a cell suspension was obtained inflammation more than was previously recognized. via centrifugation at 1200 rpm for 10 min. Cells were washed in PBS containing 2% BSA and suspended in 36% Percoll (Amersham Pharmacia Biotech). The suspension was gently overlaid onto 72% Percoll and Materials and Methods centrifuged at 2000 rpm at room temperature for 30 min. Cells were re- Mice trieved from the Percoll interface and washed twice in DMEM medium and once with staining buffer (PBS containing 0.2% BSA and 0.1% so- 2 2 2 2 2 2 2 2 CD137 / , CD137L / , CCR5 / , and TLR2 / C57BL/6 mice were dium azide). maintained in a specific pathogen-free facility and used between 7 and 8 wk of age. Wild-type (WT) C57BL/6 mice were purchased from Orient. Flow cytometry Animal studies were approved by the University of Ulsan Animal Care and Prepared cells were preincubated in blocking buffer (PBS containing anti- Use Committee. CD16/CD32 [2.4G2] mAbs, 0.2% BSA, 0.01% sodium azide) for 20 min at Abs and reagents 4˚C. After washing twice with staining buffer, cells were incubated with the relevant mAbs for 30 min at 4˚C. They were then washed twice with Cells or tissue sections were stained with fluorescent- or biotin-conjugated staining buffer and analyzed using FACSCanto II (BD Biosciences). mAbs that specifically recognize CD3, CD11b, CD45, CCL3, CCL4, CCL5, cytokeratin, Gr-1, or NK1.1 (eBioscience). The following chemokine Analysis of renal function antagonists were purchased from Tocris Bioscience: BX513 hydrochloride Kidney function was determined by measuring and blood urea (CCR1), RS504393 (CCR2), and DAPTA (CCR5). Purified rabbit anti-rat nitrogen (BUN) concentrations in sera obtained at various time points after CX3CR1 was purchased from Torrey Pines Biolabs. The TLR2 ligand kidney IRI. Creatinine concentrations were measured enzymatically using Pam3CSK4 and anti-TLR2 mAb were purchased from InvivoGen. Purified either a creatinine colorimetric/fluorometric assay kit (BioVision) or the Downloaded from anti–high mobility group box 1 (HMGB1) mAb was purchased from IDTox creatinine assay kit (ID Labs). BUN concentrations were measured BioLegend. colorimetrically using the QuantiChrom urea assay kit (Bioassay Systems) Induction of kidney IRI according to the manufacturer’s instructions. Kidney IRI was induced as previously described (12). In brief, following Pathological scoring anesthesia, bilateral flank incisions were performed, both kidney arteries Kidneys were fixed in 10% formalin, embedded in paraffin, and sectioned (5 were exposed and cross-clamped for 35 min, and reperfusion was allowed

mm). Paraffin sections were stained with H&E and analyzed. Kidney injury http://www.jimmunol.org/ for 24 h through clamp release. A sham operation was performed using the was scored by a single pathologist (H.J.C.) as the percentage of damaged same surgical procedure without clamping the renal artery. During the tubules in the corticomedullary junction. Criteria for kidney injury in- surgery, mice were placed on a 37˚C heat pad. cluded tubular , cast formation, loss of brush border, tubular di- Preparation of kidney cells latation, and immune cell infiltration. Scoring for each category was as follows: 0, no change; 1, , 10%; 2, 11–25%; 3, 25–45%; 4, 46–75%; 5, Kidneys were perfused, minced, and placed in DMEM containing 1 mg/ml .76% area change. Scores for all the categories were added for the final collagenase IA (Sigma-Aldrich) at 37˚C for 30 min. Digested kidney tis- injury scoring. by guest on September 29, 2021

FIGURE 1. TLR22/2 mice have defects in recruiting NK cells during kidney IRI. WT and TLR22/2 mice were subjected to 35 min of ischemia, followed by 4 or 24 h of reperfusion. (A) Kidneys were harvested and NK cell infil- trates were analyzed using FACS. Percentage of NK cells is presented. (B) Percentage of neu- trophils in kidneys 24 h after kidney IR. (C and D) Serum creatinine and BUN levels 24 h after kidney IR. (E and F) Histology (H&E staining) and injury scores 24 h after kidney IR. Data are the means 6 SEM (n = 3–4/group) and repre- sent two independent experiments. *p , 0.05, **p , 0.01, ***p , 0.001. The Journal of Immunology 2659

Isolation of NK cells Chemotaxis assay Single-cell suspensions in PBS were prepared from the spleen, filtered Purified NK cells were plated (1 3 105/well) in transwell inserts containing through a sterile mesh (BD Falcon), and washed. After erythrocytes were 5-mm pores. Conditioned medium was added to the lower chamber and the lysed in hemolysis buffer (144 mM NH4Cl and 17 mM Tris-HCl [pH 7.2]), plates were incubated at 37˚C for 12 h. DAPTA (50 mg/ml) was used to the remaining cells were resuspended in MACS buffer (13 PBS containing neutralize the NK cell surface CCR5. EDTA and 3% calf serum) and incubated with anti-CD3 microbeads for 20 min on ice. CD3+ cells were depleted using MACS (Miltenyi Biotec). The Immunohistochemistry remaining cells were stained with anti-NK1.1 mAbs. Labeled cells were sorted using the MoFlo XDP cell sorter (Beckman Coulter). The purities of Kidneys were embedded in OCT (Sakura Finetek) and frozen in liquid 2 + nitrogen. Frozen sections (8 mm thick) were permeabilized with 0.3% CD3 NK1.1 cells routinely reached 95%. Purified NK cells were used for Triton X-100, and nonspecific binding was blocked with 10% donkey adoptive transfer or chemotaxis assay. serum and anti-mouse CD16/CD32 mAbs (10 mg/ml). Sections were la- Culture of TECs beled with the relevant Abs for 1 h at room temperature. All specimens were mounted with ProLong Antifade reagent (Molecular Probes). Slides Primary TECs were obtained from kidneys after being digested with col- were examined under a laser-scanning confocal microscope (Olympus). lagenase IA for 30 min. All TECs were grown in complete DMEM/Ham’s F12 (50:50) culture medium (Invitrogen), supplemented with 5% FBS, 100 Statistics mg/ml streptomycin, 100 U/ml penicillin, hormone mix (5 mg/ml insulin, 5 mg/ml transferrin, 1.25 ng/ml PGE , 1.73 ng/ml sodium selenite), and 25 GraphPad Prism 5 (GraphPad Software) was used to analyze and present 1 t ng/ml . Cells were trypsinized before each passage. data. Differences between groups were analyzed using a two-tailed test or Third-passage TECs were used for experiments. To stimulate TEC TLR2, one- or two-way ANOVAwith posthoc analysis, as appropriate. Values are 5 expressed as the means 6 SEM. A p value , 0.05 was considered sta- 1 3 10 TECs/well were transferred to a 12-well flat-bottom plate and treated

5 tistically significant. Downloaded from with 500 ng/ml Pam3CSK4. In some experiments, 1 3 10 TECs/well were cultured under (1% O2 concentration) in the presence or absence of 500 ng/ml anti-TLR2 or 20 mg/ml anti-HMGB1 mAb. Culture medium was Results harvested after 6 h and saved for ELISA and chemotaxis assays. Impaired NK cell recruitment after IR in TLR22/2 mice Implantation of TECs NK cells are rapidly recruited to the kidney after IR (12, 13). Third-passage TECs were harvested and resuspended in methylcellulose Moreover, NK cells are indispensible in recruiting neutrophils, 5 medium (StemCell Technologies). TECs (2 3 10 cells in 10 ml medium) which are major effectors of secondary kidney damage during IRI http://www.jimmunol.org/ were then implanted beneath the kidney capsule. (12). We examined the possibility that TLR2 signaling might re- ELISA cruit NK cells to the kidney during IRI. TLR22/2 mice had sig- The CCL3, CCL4, and CCL5 concentrations in the culture supernatant or nificantly fewer NK cells in the kidney at the peak of NK cell kidney homogenates were measured using a Quantikine ELISA kit (R&D infiltration (4 h after IR) compared with WT mice (Fig. 1A). This 2 2 Systems) according to the manufacturer’s instructions. impaired NK cell recruitment in TLR2 / mice correlated with by guest on September 29, 2021

FIGURE 2. TLR2 signaling in TECs is required to recruit NK cells in kidney IRI. TECs (2 3 105) from WT or TLR22/2 mice were implanted under the kidney capsules of WT or TLR22/2 mice. Immediately thereafter, kidney IRI was induced. (A) Percentage of NK cells in kidneys 4 h after kidney IR. (B) Percentage of neutrophils in kidneys 24 h after kidney IR. (C and D) Serum creatinine and BUN levels 24 h after kidney IR. (E) Photomicrographs of representative sections from the three groups at 3200 (top)or3400 (bottom) magnification (H&E staining). Outer medullary areas are shown in upper panels, and cortex and subcapsular areas containing implanted TECs are shown in lower panels.(F) Tubular injury scores. Data are the means 6 SEM (n = 5–6/group) and represent two independent experiments. *p , 0.05, **p , 0.01. 2660 NK CELL RECRUITMENT BY TLR2 SIGNALING Downloaded from

FIGURE 3. TLR2-mediated NK cell recruitment occurs independently of CD137L-mediated neutrophil recruitment in TECs. TECs (2 3 105) from WT or CD137L2/2 mice were implanted under the kidney capsules of WT or TLR22/2 mice. Immediately thereafter, kidney IRI was induced. (A) Diagram describing a complementation test. (B) Percentage of NK cells in kidneys 4 h after kidney IR. (C) Percentage of neutrophils in kidneys 24 h after kidney IR. http://www.jimmunol.org/ (D and E) Serum creatinine and BUN levels 24 h after kidney IR. (F) Photomicrographs of representative sections from the three groups at 3200 (top)or 3400 (bottom) magnification (H&E staining). Outer medullary areas are shown in upper panels, and cortex and subcapsular areas containing implanted TECs are shown in lower panels.(G) Tubular injury scores. Data are the means 6 SEM (n = 4/group) and represent two independent experiments. *p , 0.05, **p , 0.01. poor CD45+CD11b+Gr-1hi neutrophil infiltration (Fig. 1B). After to NK cell recruitment. To test this possibility, we used an in vivo 24 h of IR, TLR22/2 mice had significant decreases in kidney experiment in which normal TECs implanted into the kidney dysfunction and damage, as reflected by lower plasma creatinine capsule induce kidney IRI (12). We implanted WT TECs under and BUN levels, less tubular necrosis, and less overall acute tu- the kidney capsule of TLR22/2 mice and induced IRI immedi- by guest on September 29, 2021 bular injury compared with WT mice (Fig. 1C–F). ately after implantation. We implanted WT TECs into WT mice Because TLR2 signaling in TECs is involved in kidney IRI (4, 6, as a positive control and TLR22/2 TECs into TLR22/2 mice 11), we asked whether TLR2 expression on TECs could contribute as a negative control. Comparable numbers of NK cells were

FIGURE 4. CCR5 is involved in recruiting NK cells during kidney IRI. CD32NK1.1+ NK cells were purified from WT or CCR52/2 spleens using the MoFlo cell sorter and adoptively transferred into WT and CCR52/2 or CCR52/2 mice, respectively, after labeling with CFSE (A) or without labeling (B–F) 1 h before kidney IR. Samples were obtained 4 (A) or 24 h after IR (B–F). (A) Percentage of CFSE+CD45+CD3+NK1.1+ NK cells. (B) Percentage of neutrophils. (C and D) Levels of serum creatinine and BUN. (E and F) Histology (H&E staining) and injury scores. Data are the means 6 SEM (n = 4–5/ group) and are representative of two independent experiments. *p , 0.05, **p , 0.01, ***p , 0.001. The Journal of Immunology 2661 Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021 2662 NK CELL RECRUITMENT BY TLR2 SIGNALING

FIGURE 6. Model summarizing the role of the TEC/ NK cell/neutrophil axis. Injury to TECs following IRI (step 1) promotes release of HMGB1 (step 2). This molecule stimulates TECs to produce CCR5 chemo- kines through TLR2 activation (step 3) in an autocrine fashion. CCR5 chemokines in turn induce NK cell re- cruitment (step 5). Infiltrated NK cells use their cell- surface molecule CD137 to stimulate CD137L on the surface of TECs (step 6). CD137L signaling results in the production of additional signaling molecules, CXCL1 and CXCL2, in TECs (step 7). Once infiltrated (step 8), neutrophils participate in active tissue destruction (step 9). Downloaded from

observed in kidneys of WT→TLR22/2 and WT→WT mice 4 h (wherein TLR2 signaling is intact) and TLR22/2 TEC cells 2/2 2/2 after IR whereas TLR2 →TLR2 mice recruited fewer NK (wherein CD137L signaling is intact) in the same kidney could http://www.jimmunol.org/ cells (Fig. 2A). Consistent with this finding, counting neutrophils result in kidney IRI, because CD137L2/2 TECs would comple- indicated that TLR22/2→TLR22/2 mice had significantly fewer ment TLR22/2 TEC cells by recruiting NK cells through TLR2 neutrophil infiltrates in the kidney than did WT→TLR22/2 and signaling and TLR22/2 TECs could complement CD137L2/2 WT→WT mice (Fig. 2B). TLR22/2→TLR22/2 mice had signif- TEC cells by recruiting neutrophils through CD137L signaling icantly lower serum creatinine and BUN levels than did (Fig. 3A). Implanting CD137L2/2 TECs into the kidney capsule WT→TLR22/2 and WT→WT mice (Fig. 2C, 2D). Histological of TLR22/2 mice completely restored NK cell and neutrophil analysis showed broader tissue inflammation, ranging from the recruitment to the extent found in WT→WT mice (Fig. 3B, 3C). subcapsular space through the renal outer medulla and cortex, in Analyzing creatinine and BUN levels and inspecting kidney sec- WT→TLR22/2 and WT→WT mice, with tissue damage being tions indicated severe kidney damage and loss of kidney function by guest on September 29, 2021 evident in these areas (Fig. 2E, 2F). In particular, heavy neutrophil in these two groups (Fig. 3D–G). TLR22/2→TLR22/2 mice had infiltration and tissue damage in the outer cortex beneath im- significantly impaired NK cell and neutrophil recruitment and plantation sites were characteristic of these mice compared with milder kidney IRI than did the other two groups (Fig. 3B–G). TLR22/2→TLR22/2 mice (Fig. 2E, lower panels). Implanting Overall, this complementation assay demonstrated that NK cell TLR22/2 TECs into TLR22/2 mice did not induce the neutrophil and neutrophil recruitment can be mediated by two uncoupled migration or tissue damage seen in the other two groups (Fig. 2E, signaling events that occur as a result of TLR2 and CD137L 2F). Taken together, these results clearly demonstrate that TLR2 signaling in TECs, respectively. expression on TECs was required to recruit NK cells during CCR5 is critical in recruiting NK cells kidney IRI. We hypothesized that TLR2 signaling in TECs would produce TECs control sequential NK cell and neutrophil recruitment chemokines that can recruit NK cells to the kidney during kidney through TLR2 and CD137L signaling during kidney IRI, IRI. To test this hypothesis, we screened chemokine receptors that respectively could be involved in NK cell recruitment in kidney IRI. Among The results presented in Figs. 1 and 2 suggest that TLR22/2 mice CCR1, CCR2, CCR5, and CX3CR1 antagonists, a CCR5 antag- might be resistant to kidney IRI owing to defects in NK cell re- onist blocked NK cell recruitment most potently at 4 h after IR cruitment regulated by TLR2 signaling in TECs. To test this hy- (Supplemental Fig. 1). Consistently, the CCR5 antagonist inhibi- pothesis, we assumed that NK cell and neutrophil recruitment are ted neutrophil recruitment, kidney damage, and loss of kidney func- induced by two temporally separate signaling events that are de- tion (Supplemental Fig. 2). To obtain more concrete evidence for livered through TLR2 and CD137L, respectively, in TECs. Based the involvement of NK cells expressing CCR5 in kidney IRI, we on this assumption, we hypothesized that CD137L2/2 TECs adoptively transferred WT NK cells into CCR52/2 mice. We first

FIGURE 5. TLR2 signaling results in CCL3/4/5 production by TECs. (A) Levels of CCL3, CCL4, and CCL5 in 2 h IR kidney homogenates. (B) TECs (1 3 105/well) were cultured in 12-well plates in the presence or absence of Pam3CSK4 for 6 h. The CCL3/4/5 concentration in the culture medium was measured using ELISA. (C) A chemotaxis assay for purified NK cells (5 3 104/well) from WT and CCR52/2 mice was conducted using conditioned medium from TEC cultures. To block CCR5, DAPTA was added. (D) TECs (2 3 105) from WT or TLR22/2 mice were cultured in 12-well plates under normoxia or hypoxia (1% O2,5%CO2, 94% N2) in the presence or absence of anti-HMGB1 or anti-TLR2 mAb for 6 h. The CCL3/4/5 concentrations in the culture medium were measured using ELISA. (E) A chemotaxis assay for purified NK cells (5 3 104/well) from WT and CCR52/2 mice was conducted using conditioned medium from TEC cultures. Data are the means 6 SEM (n = 3–4/group) of two to four independent experiments. *p , 0.05, **, p , 0.01, ***p , 0.001. The Journal of Immunology 2663 confirmed that a significantly higher number of CFSE-labeled induce inflammation to remove damaged cells. Therefore, TECs WT NK cells were recruited to the kidney of CCR52/2 mice, seem to be able to survey their stress status and make a sponta- compared with CCR52/2 NK cells, 4 h after IR (Fig. 4A). neous decision regarding whether damaged cells should be re- WT→CCR52/2 mice had significantly restored neutrophil infil- moved. The amazing ability of TECs to use immune cells for their tration and kidney damage and loss of function compared with own purpose may ultimately allow the kidney to restore its WT→WT (positive control) and CCR52/2→CCR52/2 (negative function and homeostasis (1, 17). control) mice (Fig. 4B–F). Taken together, these data suggest that Renal TECs produce proinflammatory cytokines and chemo- CCR5 plays a critical role in NK cell recruitment during kidney IRI. kines, including IL-6, CCL2, CCL5, CXCL1, and CXCL2, upon exposure to exogenous or endogenous TLR2 ligands (4, 18). These TECs secrete CCR5 chemokines in response to HMGB1, an results indicate that TLR2 signaling can trigger recruitment of endogenous TLR2 ligand other leukocytes, such as neutrophils and monocytes. However, an We examined whether CCR5 chemokines are produced by TECs influx of neutrophils due to TLR2 signaling in TECs is not suf- during kidney IRI. Immunohistochemical analysis showed expres- ficient to induce fulminant kidney IRI. Instead, TECs initiate sion of the CCR5 chemokines, CCL3, CCL4, and CCL5, in WT signaling through CD137L with help from NK cells to induce 2/2 TECs 2 h after IR compared with TLR2 TECs (Supplemental a second wave of neutrophil influx. Therefore, NK cells are re- Fig. 3). ELISA of 2 h IRI mouse kidney homogenates also dem- quired to recruit the full wave of neutrophils needed for clinical 2/2 onstrated that TLR2 mice had significantly lower levels of kidney IRI. Even though NK cells interact with various immune CCL3, CCL4, and CCL5 compared with WT mice (Fig. 5A). system components (19), our results show that NK cell–paren-

These data suggest that TLR2 signaling should signal TECs to chymal cell interactions control acute tissue inflammation. Downloaded from produce CCR5 chemokines in an autocrine manner. We next ex- Epithelial cells are exposed to a variety of microbial and non- amined the ability of isolated TECs to produce CCR5 chemokines microbial challenges. To manage these stressful conditions (17), in response to Pam3CSK5, a potent TLR2 ligand. WT, but not they have numerous sensors, many of which converge on in- 2/2 TLR2 , TECs secreted high levels of CCL3, CCL4, and CCL5 flammation. Well-known examples are the TLR and complement within 4 h after treatment with Pam3CSK5 (Fig. 5B). Conditioned systems (4–8, 20). Individual signaling through TLR2, TLR4, and

medium from cultured TECs stimulated with Pam3CSK4 induced complement receptors results in proinflammatory cytokine and http://www.jimmunol.org/ NK cell migration in a chemotaxis assay compared with condi- chemokine production (20). This result, however, does not mean 2/2 tioned medium from cultured TLR2 TECs (Fig. 5C). Blocking that they have redundant roles in kidney IRI (6, 20). Rather, there CCR5 with CCR5 antagonist or deleting the CCR5 gene com- seems to be a coordinated regulatory system where each signaling pletely abrogated the increased chemotaxis of NK cells in re- pathway crosstalks to control the progression of kidney IRI. For sponse to conditioned medium from cultured WT TECs stimulated example, targeted deletion of TLR2 and factor B or MyD88 ex- with Pam3CSK4 (Fig. 5D). Next, we tested whether an endoge- acerbated kidney IR compared with deletion of factor B or TLR2 nous TLR2 ligand stimulates TECs to produce CCR5 chemokines alone (6, 20). under hypoxic conditions. We detected significantly higher levels Various chemokines play a role in kidney IRI by recruiting of CCL3, CCL4, and CCL5 in culture medium of WT TECS than leukocytes such as neutrophils, macrophages, and Th1 T cells by guest on September 29, 2021 2/2 in that of TLR2 TECs, and neutralization of HMGB1, a well (21–25). In this study, we added CCR5 chemokines to the list of known endogenous ligand of TLR2 released during kidney IRI chemokines involved in kidney IRI. TLR2 signaling resulted in (reviewed in Ref. 15), markedly decreased levels of CCL3, CCL4, CCL3, CCL4, and CCL5 production in TECs, and blocking of and CCL5 (Fig. 5E). Blocking TLR2 had a similar effect in this CCR5 completely inhibited NK cell migration in vitro and in vivo. culture condition (Fig. 5E). Moreover, we found that levels of Although CCL3 and CCL5 are versatile chemokines acting on two CCL3, CCL4, and CCL5 contained in conditioned medium from to three chemokine receptors (26), our results clearly showed that this set of experiments positively correlated with its ability to in- CCR5 functioned as the major receptor for NK cells during kidney duce NK cell migration in a chemotaxis assay (Fig. 5F). In sum, IRI. Moreover, our observation that blocking CCR5 had a similar our results suggest that after ischemic damage, TECs release effect on NK cell recruitment and kidney IRI as depleting NK HMGB1 that induces CCR5 chemokine production through TLR2 cells (12) provides strong evidence that prior NK cell infiltration is signaling. needed for neutrophil migration during kidney IRI. In summary, our results demonstrate that TECs play a central Discussion role in the pathogenic axis consisting of TECs, NK cells, and In kidney IRI, hypoxic and anoxic cell injuries occur mainly in the neutrophils in kidney IRI. TECs have an autoamplification loop microvasculature and tubular epithelium during the ischemic phase where two important signaling events occur through TLR2 and (16). These injured cells undergo necrosis and are thought to CD137L signaling to recruit NK cells and neutrophils, respectively produce endogenous TLR ligands (15). TLR2 signaling in TECs is (Fig. 6). Therefore, pharmacologically inhibiting innate receptor a major inducer of acute inflammation following kidney IRI (4, 6, activation in TECs may be a novel strategy to prevent renal failure 11). However, the elaborate mechanism through which TLR2 in transplanted kidneys. signaling controls inflammation in TECs remains unclear, largely due to technical difficulties in characterizing this class of signal Acknowledgments in vivo. Implanting TECs into the kidney capsule of recipient mice We thank members of Drs. H.R. Cho’s and B. Kwon’s laboratory for help. is a useful tool to dissect the signaling events in TECs that result in kidney IRI (12). In this study, we used this technique to develop Disclosures another in vivo technique that allowed us to analyze two tempo- The authors have no financial conflicts of interest rally separate signaling events in the same axis leading to kidney IRI. Using this new technique, we demonstrated that TECs are a master regulator of kidney IRI by releasing inducers and me- References 1. Medzhitov, R. 2008. Origin and physiological roles of inflammation. Nature 454: diators of inflammation (HMGB1 and CXCR2 chemokines) for 428–435. kidney IRI. Once damage in TECs is beyond repair, they may 2. Nathan, C. 2002. Points of control in inflammation. Nature 420: 846–852. 2664 NK CELL RECRUITMENT BY TLR2 SIGNALING

3. Barton, G. M. 2008. A calculated response: control of inflammation by the innate 14. Li, L., L. Huang, A. L. Vergis, H. Ye, A. Bajwa, V. Narayan, R. M. Strieter, immune system. J. Clin. Invest. 118: 413–420. D. L. Rosin, and M. D. Okusa. 2010. IL-17 produced by neutrophils regulates 4. Leemans, J. C., G. Stokman, N. Claessen, K. M. Rouschop, G. J. Teske, IFN-g-mediated neutrophil migration in mouse kidney ischemia-reperfusion C. J. Kirschning, S. Akira, T. van der Poll, J. J. Weening, and S. Florquin. 2005. injury. J. Clin. Invest. 120: 331–342. Renal-associated TLR2 mediates ischemia/reperfusion injury in the kidney. J. 15. Rosin, D. L., and M. D. Okusa. 2011. Dangers within: DAMP responses to Clin. Invest. 115: 2894–2903. damage and cell death in kidney disease. J. Am. Soc. Nephrol. 22: 416–425. 5. Wu, H., G. Chen, K. R. Wyburn, J. Yin, P. Bertolino, J. M. Eris, S. I. Alexander, 16. Jang, H. R., and H. Rabb. 2009. The innate immune response in ischemic acute A. F. Sharland, and S. J. Chadban. 2007. TLR4 activation mediates kidney kidney injury. Clin. Immunol. 130: 41–50. ischemia/reperfusion injury. J. Clin. Invest. 117: 2847–2859. 17. Swamy, M., C. Jamora, W. Havran, and A. Hayday. 2010. Epithelial decision 6. Shigeoka, A. A., T. D. Holscher, A. J. King, F. W. Hall, W. B. Kiosses, P. S. Tobias, makers: in search of the “epimmunome.” Nat. Immunol. 11: 656–665. N. Mackman, and D. B. McKay. 2007. TLR2 is constitutively expressed within the 18.Tsuboi,N.,Y.Yoshikai,S.Matsuo, T. Kikuchi, K. Iwami, Y. Nagai, kidney and participates in ischemic renal injury through both MyD88-dependent O. Takeuchi, S. Akira, and T. Matsuguchi. 2002. Roles of Toll-like receptors in and -independent pathways. J. Immunol. 178: 6252–6258. C-C chemokine production by renal tubular epithelial cells. J. Immunol. 169: 7. Pulskens, W. P., G. J. Teske, L. M. Butter, J. J. Roelofs, T. van der Poll, 2026–2033. S. Florquin, and J. C. Leemans. 2008. Toll-like receptor-4 coordinates the innate 19. Vivier, E., E. Tomasello, M. Baratin, T. Walzer, and S. Ugolini. 2008. Functions immune response of the kidney to renal ischemia/reperfusion injury. PLoS ONE of natural killer cells. Nat. Immunol. 9: 503–510. 3: e3596. 20. Thurman, J. M., A. M. Lenderink, P. A. Royer, K. E. Coleman, J. Zhou, 8. Chen, J., R. John, J. A. Richardson, J. M. Shelton, X. J. Zhou, Y. Wang, J. D. Lambris, R. A. Nemenoff, R. J. Quigg, and V. M. Holers. 2007. C3a is Q. Q. Wu, J. R. Hartono, P. D. Winterberg, and C. Y. Lu. 2011. Toll-like receptor required for the production of CXC chemokines by tubular epithelial cells after 4 regulates early endothelial activation during ischemic . renal ishemia/reperfusion. J. Immunol. 178: 1819–1828. Kidney Int. 79: 288–299. 21. Miura, M., X. Fu, Q. W. Zhang, D. G. Remick, and R. L. Fairchild. 2001. 9. Kru¨ger, B., S. Krick, N. Dhillon, S. M. Lerner, S. Ames, J. S. Bromberg, M. Lin, Neutralization of Groa and macrophage inflammatory protein-2 attenuates renal L. Walsh, J. Vella, M. Fischereder, et al. 2009. Donor Toll-like receptor 4 con- ischemia/reperfusion injury. Am. J. Pathol. 159: 2137–2145. tributes to ischemia and reperfusion injury following human kidney transplan- 22. Cugini, D., N. Azzollini, E. Gagliardini, P. Cassis, R. Bertini, F. Colotta, tation. Proc. Natl. Acad. Sci. USA 106: 3390–3395. M. Noris, G. Remuzzi, and A. Benigni. 2005. Inhibition of the chemokine re- 10. Wu, H., J. Ma, P. Wang, T. M. Corpuz, U. Panchapakesan, K. R. Wyburn, and ceptor CXCR2 prevents kidney graft function deterioration due to ischemia/ Downloaded from S. J. Chadban. 2010. HMGB1 contributes to kidney ischemia reperfusion injury. reperfusion. Kidney Int. 67: 1753–1761. J. Am. Soc. Nephrol. 21: 1878–1890. 23. Fiorina, P., M. J. Ansari, M. Jurewicz, M. Barry, V. Ricchiuti, R. N. Smith, 11. Allam, R., C. R. Scherbaum, M. N. Darisipudi, S. R. Mulay, H. Ha¨gele, S. Shea, T. K. Means, H. Auchincloss, Jr., A. D. Luster, et al. 2006. Role of CXC J. Lichtnekert, J. H. Hagemann, K. V. Rupanagudi, M. Ryu, C. Schwarzenberger, chemokine receptor 3 pathway in renal ischemic injury. J. Am. Soc. Nephrol. 17: et al. 2012. Histones from dying renal cells aggravate kidney injury via TLR2 716–723. and TLR4. J. Am. Soc. Nephrol. 23: 1375–1388. 24. Furuichi, K., J. L. Gao, and P. M. Murphy. 2006. Chemokine receptor CX3CR1 12. Kim, H. J., J. S. Lee, J. D. Kim, H. J. Cha, A. Kim, S. K. Lee, S. C. Lee, regulates renal interstitial fibrosis after ischemia-reperfusion injury. Am. J.

B. S. Kwon, R. S. Mittler, H. R. Cho, and B. Kwon. 2012. Reverse signaling Pathol. 169: 372–387. http://www.jimmunol.org/ through the costimulatory ligand CD137L in epithelial cells is essential for 25. Oh, D. J., B. Dursun, Z. He, L. Lu, T. S. Hoke, D. Ljubanovic, S. Faubel, and natural killer cell-mediated acute tissue inflammation. Proc. Natl. Acad. Sci. C. L. Edelstein. 2008. Fractalkine receptor (CX3CR1) inhibition is protective USA 109: E13–E22. against ischemic acute renal failure in mice. Am. J. Physiol. Renal Physiol. 294: 13. Zhang, Z. X., S. Wang, X. Huang, W. P. Min, H. Sun, W. Liu, B. Garcia, and F264–F271. A. M. Jevnikar. 2008. NK cells induce in tubular epithelial cells and 26. Zlotnik, A., and O. Yoshie. 2012. The chemokine superfamily revisited. Im- contribute to renal ischemia-reperfusion injury. J. Immunol. 181: 7489–7498. munity 36: 705–716. by guest on September 29, 2021