C3a and C5a Promote Renal Ischemia-Reperfusion Injury

BASIC RESEARCH www.jasn.org

† ‡ Qi Peng,* Ke Li,* Lesley A. Smyth,* Guolan Xing,* Naiyin Wang,* Lucy Meader,* Bao Lu, Steven H. Sacks,* and Wuding Zhou*

*Medical Research Council Centre for Transplantation, King’s College London, London, United Kingdom; †Core Research Laboratory, The Second Afﬁliated Hospital, School of Medicine, Xi’an Jiaotong University, Xi’an, China; and ‡Division of Pulmonary Medicine and Ina Sue Perlmutter Laboratory, Children’s Hospital, Harvard Medical School, Boston, Massachusetts

ABSTRACT Renal ischemia reperfusion injury triggers complement activation, but whether and how the small proinflammatory fragments C3a and C5a contribute to the pathogenesis of this injury remains to be elucidated. Using C3aR-, C5aR-, or C3aR/C5aR-deficient mice and models of renal ischemia-reperfusion injury, we found that deficiency of either or both of these receptors protected mice from injury, but the C3aR/C5aR- and C5aR-deficient mice were most protected. Protection from injury was associated with less cellular infiltration and lower mRNA levels of kidney injury molecule-1, proinflammatory mediators, and adhesion molecules in postischemic kidneys. Furthermore, chimera studies showed that the absence of C3aR and C5aR on renal tubular epithelial cells or circulating leukocytes attenuated renal ischemia-reperfusion injury. In vitro, C3a and C5a stimulation induced inflammatory mediators from both renal tubular epithelial cells and macrophages after hypoxia/reoxygenation. In conclusion, although both C3a and C5a contribute to renal ischemia-reperfusion injury, the pathogenic role of C5a in this injury predominates. These data also suggest that expression of C3aR and C5aR on both renal and circulating leukocytes contributes to the pathogenesis of renal ischemia-reperfusion injury.

J Am Soc Nephrol 23: 1474–1485, 2012. doi: 10.1681/ASN.2011111072

Ischemia reperfusion (IR) injury occurs upon product C5b-9, also called the membrane attack reperfusion of vascularized tissue after an extended complex (MAC). These effector molecules have di- period of ischemia. It is a common source of illness verse biologic functions for example, C3b mediates and death in a wide variety of conditions, including opsonization of pathogens, C3a and C5a induce local myocardial infarction and stroke. IR injury is also inﬂammation and cell activation, and MAC mediates an unavoidable event in organ transplantation and direct lysis of pathogens and tissue damage.10 has a major effect on short- and long-term graft Previous studies using mice deﬁcient in comple- survival.1–4 The pathogenesis of IR injury is com- ment components (e.g., C3, C4, C5, C6, factor B)5,11 plex, and multiple factors (i.e., complement activa- or complement inhibitors (e.g., CD55, CD59)12,13 tion, the coagulation system, leukocytes, cytokines, chemokines, and adhesion molecules) are all – thought to contribute to its development.5 9 Received November 14, 2011. Accepted May 17, 2012.

The complement system consists of a set of dis- Published online ahead of print. Publication date available at tinctplasmaproteinsthatreactinacascademan- www.jasn.org. ner upon triggering by various stimuli, such as Correspondence: Dr. Wuding Zhou, MRC Centre for Trans- infection and tissue injury. Complement activation plantation, King’s College London, 5th Floor Tower Wing, Guy’s generates a set of effector molecules, including the Hospital, Great Maze Pond, London SE1 9RT, United Kingdom. large fragment C3b and its metabolites (iC3b, C3dg); Email: [email protected] the small fragments (C3a, C5a); and the terminal Copyright © 2012 by the American Society of Nephrology

1474 ISSN : 1046-6673/2309-1474 JAmSocNephrol23: 1474–1485, 2012 www.jasn.org BASIC RESEARCH have demonstrated that complement activation contributes to the pathogenesis of renal IR injury. Several studies have also suggested that the formation of MAC (C5b-9) in renal tubular epithelial cells is a critical effector mechanism through which complement mediates renal IR injury.13,14 In addition, the small fragments (C3a, C5a) could be involved in renal IR injury. Renal IR injury is a nonspecificinflammatory process; C3a and C5a, which are released during complement activation, play an important role in the initiation and regulation of inflammatory responses. Both renal cells and circulating leukocytes express C3aR and C5aR.15–20 Elevated C3a levels were detected in the circulation of mice after renal IR.21 Blocking C5aR signaling using a small interfering RNA technique or a receptor-specific Figure 1. Renal expression of C3aR and C5aR is upregulated after antagonist reduced renal functional impairment or decreased renal ischemic insult. Renal IR injury was induced in WT mice. tissue inflammation within the kidneys after renal IR,22–24 Kidney tissues were harvested at 24 hours after reperfusion. suggesting a pathogenic role for C5aR in renal IR injury. Quantitative real-time RT-PCR for analyzing C3aR and C5aR was Several important questions still remain unanswered, performed in naive and postischemic kidneys (n=5 per group). (A) C3aR mRNA expression. (B) C5aR expression. Data were however: Does C3a, like C5a, participate in the development shown as mean 6 SEM and analyzed by t test. **P,0.005 for of the renal IR injury via receptor signaling? Does combined naive versus postischemic kidneys. C3aR and C5aR signaling have a greater effect on the injury than does signaling of each receptor alone (because the ligands for C3aR and C5aR are likely to coexist)? How do C3a and C5a participate in the development of the injury? Which types of Deficiency of C3aR and C5aR Protects Mice from Renal cells are likely to contribute to C3aR- and C5aR-mediated re- IR Injury nal IR injury? To address these questions, we used in vivo and To assess the effect of C3a and C5a receptor deficiency on renal in vitro models of renal IR injury. Using mice with targeted IR injury, we induced renal IR injury in four groups of mice deletions of C3aR and C5aR or of C3aR alone or C5aR alone WT, C3aR2/2,C5aR2/2, and DKO. Renal function was and bone marrow chimeras (wild type [WT]↔ C3aR/C5aR assessed at 24 and 48 hours after reperfusion by measuring double knockout [DKO]), we assessed the effect of single or BUN. All three knockout groups had significantly lower levels double deficiency of C3aR and C5aR on renal functional im- of BUN at both time points compared with WT mice, indi- pairment, renal tissue damage and cellular infiltration, and cating that renal function was protected in all three knockout production of inflammatory mediators in postischemic kidneys. groups (Figure 2A). Toassess renal tubular damage, histologic We also evaluated the contribution of C5aR/C3aR-bearing renal analysis of kidneys collected at 24 and 48 hours after reper- and circulating cells to this injury. fusion was performed. In agreement with renal functional impairment, tubular injury consisting of tubule thinning, di- latation, loss of proximal brush border, and protein casts was RESULTS reduced in all three knockout groups compared with meas- ures in WTmice, the reduction of tubular damage was greater Renal Expression of C3aR and C5aR Is Upregulated in C5aR2/2 or DKO mice (Figure 2, B–D and Supplemental after Renal Ischemic Insult Figure 1). Upregulation of C3aR or C5aR expression has been observed Kidney injury molecule-1 (KIM-1) has been described as in certain pathologic conditions, such as asthma and prolonged a biomarker for acute tubular injury because its expression cold ischemia.14,25 To assess whether expression of C3aR and increases dramatically after injury in proximal tubule epithe- C5aR in kidney tissues is upregulated after renal IR in our lial cells in postischemic rodent kidneys.26 We therefore as- system (a warm ischemia model), we performed quantitative sessed mRNA expression of KIM-1 in injured kidneys of real-time RT-PCR to measure C3aR and C5aR expression in WT, C3aR2/2,C5aR2/2, and DKO mice. In correlation both noninjured and injured kidney tissues (collected 24 with renal function impairment and tubule injury, renal ex- hours after reperfusion). Both C3aR and C5aR were detected pression of KIM-1 was significantly lower in the knockout in noninjured kidney tissues; however, expression of both re- groups of mice than in the WT group at 24 hours after receptors was significantly increased after renal ischemic insult perfusion, with about 25% (C3aR2/2), 50% (C5aR2/2), (Figure 1, A and B). Thus, upregulation of C3aR and C5aR and 60% (DKO) reduction in the WT expression levels expression in ischemic kidneys, together with the observation (Figure 2E). Significantly low renal expression of KIM-1 of elevated C3a in the circulation after renal IR,21 would support was also observed at 48 hours after reperfusion in the the notion that C3aR- and C5aR-ligand interaction participates C5aR2/2 and DKO tissue compared with WT tissue (Sup- in the process of renal IR injury. plemental Figure 2). Taken together, these observations

J Am Soc Nephrol 23: 1474–1485, 2012 C3aR and C5aR in Renal Injury 1475 BASIC RESEARCH www.jasn.org

Figure 2. Deficiency of C3aR and C5aR protects mice from renal IR injury. Renal IR injury was induced in four groups of mice: WT, C3aR2/2,C5aR2/2, and DKO. Serum samples and kidney tissues were collected at 24 and 48 hours after reperfusion. (A) BUN levels at 24 and 48 hours after reperfusion. Number of mice studied per group is included in parentheses. The dotted line indicates a normal BUN level, which was obtained from pooled results of 10 normal mice. (B) Normal kidney histologic features. Light microscopic images of hematoxylin-eosin (H&E) and periodic acid-Schiff (PAS) staining of kidney from naive WT mouse. (C) Histopathologic features of injured kidneys. Representative light microscopic images of H&E and PAS staining of kidneys from the four groups, at 24 hours after reperfusion. (D) Histologic scores of renal tubule injury. Tissue injury was scored in cortical medullary junction areas at magnification of 3200 (n=6 per group three fields of each kidney). (E) KIM-1 gene expression in postischemic kidneys. The dotted line indicates KIM-1 gene expression level in normal kidney tissues. Quantitative real-time RT-PCR was performed in postischemic kidneys (n=4 per group), at 24 hours after reperfusion. Data in parts A, D, and E are shown as mean 6 SEM and were analyzed by Mann-Whitney test *P,0.05, **P,0.005, ***P,0.0001 for WT versus C3aR2/2 or C5aR2/2 or DKO mice.

1476 Journal of the American Society of Nephrology J Am Soc Nephrol 23: 1474–1485, 2012 www.jasn.org BASIC RESEARCH demonstrate that the lack of C3aR or C5aR receptors pro- vides functional and structural protection against renal IR injury; however, protection is better with the lack of C5aR or both C3aR and C5aR.

Deficiency of C3aR and C5aR Reduces Renal Production of Inflammatory Mediators after Renal Ischemic Insult Inflammatory mediators have been suggested to contribute to AKI after renal ischemic insult;5,9,27 we therefore assessed the production of inflammatory mediators in postischemic kidneys from WT, C3aR2/2, C5aR2/2, and DKO mice. Using quantitative real-time RT-PCR, we analyzed the gene expression of proinflammatory cytokines (TNF-a, IL-1b,IFN-g), chemokines (keratinocytes chemoattractant [KC, murine IL-8 homologous], monocyte chemoattractant protein-1 [MCP-1], macrophage inflammatory protein [MIP]-1a, MIP-1b)and adhesion molecules (intercellular adhesion molecule [ICAM]-1, CD146, von Willebrand factor) in kidney tissues collected at 24 and 48 hours after reperfusion. Renal tissue derived from C5aR2/2 or DKO mice exhibited significantly lower levels of proinflammatory cytokines (TNF-a, IL-1b, IFN-g) and chemokines (KC, MCP-1) compared with WT renal tissue at both 24 and 48 hours after reperfusion (Figure 3, A and B and Supplemental Figure 3, A and B). C3aR2/2 mice had less reduction in renal synthesis of these cytokines and chemokines; only TNF-a was significantly reduced at 24 hours after reperfusion (Figure 3, A and B and Supplemental Figure 3, A and B). In addition, we observed reduced levels of chemokines MIP-1a and MIP-1b as well as adhesion molecules ICAM-1, CD146, and von Willebrand factor in postischemic kidneys from C3aR2/2 or C5aR2/2 or DKO mice compared with WT mice however, significant reduction was observed on only some of those molecules (i.e.,MIP-1a,MIP-1b, ICAM-1) in DKO mice (Figure 3, B and C andSupplemental Figure 3, B and C). Thus, our data indicate that the presence of C3aR or C5aR is required for renal production of inflammatory mediators after renal ischemic insult, particularly at 24 hours after Figure 3. Deficiency of C3aR and C5aR reduces renal synthesis of reperfusion. inflammatory mediators after renal ischemic insult. Renal IR injury was induced in WT, C3aR2/2,C5aR2/2,andDKOmice(n=5 Deficiency of C3aR and C5aR Reduces Cellular per group). Kidney tissues were harvested at 24 hours after re- fi perfusion and used for quantitative real-time RT-PCR analysis. (A) In ltration in Postischemic Kidneys fl fi Proin ammatory cytokines (TNF-a,IL-1b,andIFN-g). (B) Chemokines In addition to tubule damage, renal interstitial cellular in l- (KC, MCP-1, MIP-1a,andMIP-1b). (C) Adhesion molecules and tration is a characteristic pathologic change in renal IR injury. endothelial activation/injury markers (ICAM-1, CD146, and von To assess the influence of deficiency of C3aR and C5aR on Willebrand factor). The dotted lines in each part indicate the cellular infiltration, we performed flow cytometric analysis gene expression levels in normal kidney tissues. Data are shown (for CD45, Gr-1, and F4/80) and immunochemical staining as mean 6 SEM and were analyzed by Mann-Whitney test *P,0.05, (for Gr-1 and F4/80) in kidney tissues collected at 24 hours **P,0.005 for WT versus C3aR2/2 or C5aR2/2 or DKO mice. after reperfusion. Immunochemical staining suggested lower levels of Gr-1+ or F4/80+ cells in postischemic kidneys from F4/80+) in single-cell suspension prepared from individual C3aR2/2, C5aR2/2, and DKO mice compared with the levels postischemic kidneys was lower in C3aR2/2,C5aR2/2, in WT kidneys (Figure 4A). Flow cytometry results confirmed and DKO mice than in WT mice (Figure 4B). To investigate + + the reduction of cellular infiltration in all receptor knockout whether the observed lower number of renal CD45 , Gr-1 , mice. The percentage of three types of cells (CD45+,Gr-1+, and F4/80+ cells in the receptor knockout mice after renal IR

J Am Soc Nephrol 23: 1474–1485, 2012 C3aR and C5aR in Renal Injury 1477 BASIC RESEARCH www.jasn.org

Figure 4. Deficiency of C3aR and C5aR reduces cellular infiltration in postischemic kidneys. Renal IR injury was induced in WT, C5aR, C3aR, and DKO mice (n=5 per group). Kidney tissues were harvested at 24 hours after reperfusion. Frozen sections and single cell suspensions were prepared from the kidneys and used for immunohistochemistry and flow cytometry, respectively. (A) Immunohis- tochemical staining of Gr-1 and F4/80. Representative light microscopic images at original magnification of 3 200 from each group are shown. (B) Flow cytometric analysis of CD45+, Gr-1+, and F4/80+ cells in single-cell suspensions prepared from individual kidneys. Left four panels: representative dot plots from each group. Right panel: quantified percentages of CD45+,Gr-1+, and F4/80+ cells from each group (n=5 per group). Data are shown as mean 6 SEM and were analyzed by Mann-Whitney test *P,0.05, **P,0.005, ns not significant, comparing WT and C3aR2/2 or C5aR2/2 or DKO mice.

1478 Journal of the American Society of Nephrology J Am Soc Nephrol 23: 1474–1485, 2012 www.jasn.org BASIC RESEARCH is due to the effect of the intrinsic cell defect, we evaluated renal CD45+, Gr-1+,andF4/80+ cells in naive C3aR2/2, C5aR2/2, DKO, and WT mice. The proportions of the three cell types are similar in all the receptor knockout mice and WT mice, suggesting that the observed differences in cellular infiltration between WT and the receptor-deficient mice depend on ischemic insult (Supplemental Figure 4). Thus, our data demonstrate that single or double deficiency of C3aR and C5aR reduces cellular infiltration of leukocytes (mainly consisting of neutrophils and monocytes or macrophages) in kidneys at 24 hours in response to renal ischemic insult, although the deficiency of C3aR alone has a relatively small effect on this process.

Contribution of C3aR and C5aR on Renal Cells and Circulating Leukocytes to Renal IR Injury Because both renal cells (e.g., tubular epithelial cells) and circulating leukocytes (e.g., neutrophils, monocytes, or macrophages) express C3aR and C5aR, we next determined whether C3aR and C5aR on both groups of cells can contribute to the development of renal IR injury. To do this, we generated murine chimeras between WT and DKO mice by bone marrow transplantation. We generated four groups of mice (1)WT Figure 5. Contribution of C3aR and C5aR on renal cells and mice reconstituted with WT bone marrow (WT/WT, both re- circulating leukocytes to renal IR injury. Renal IR injury was in- nal cells and circulating leukocytes express C3aR and C5aR), duced in four groups of chimeric mice: (1)WTbonemarrowinto (2) DKO mice reconstituted with DKO bone marrow (DKO/ WT recipient (WT/WT), (2) DKO bone marrow into DKO recipi- DKO, both renal cells and circulating leukocytes lack C3aR ents (DKO/DKO), (3) DKO bone marrow into WT recipient (DKO/ WT), and (4) WT bone marrow into DKO recipients (WT/DKO) and C5aR), (3) DKO mice reconstituted with WT bone mar- (n=5 per group). Renal function was assessed at 24 hours after row (DKO/WT, renal cells lack C3aR and C5aR while circu- reperfusion by measuring BUN. The dotted line indicates a nor- lating leukocytes express C3aR and C5aR), and (4)WTmice mal BUN level, which was obtained from pooled results of 10 reconstituted with DKO bone marrow (WT/DKO, renal cells normal mice. Data are shown as mean 6 SEM and were analyzed express C3aR and C5aR while circulating leukocytes lack by Mann-Whitney test *P,0.05, **P,0.005, comparing WT/WT C3aR and C5aR). We induced renal IR injury in these mice and DKO/DKO, or DKO/WT or WT/DKO groups. ns, not signifi- at 8 weeks after transfusion and evaluated renal function cant, comparing DKO/WT and WT/DKO groups. impairment at 24 hours after reperfusion. BUN levels were significantly lower in the DKO/DKO, DKO/WT, and WT/ DKO groups than in WT/WT mice after injury (Figure 5), We prepared macrophages from naive WT mice and exposed indicating that the lack of C3aR and C5aR expression in renal the cells to hypoxia followed by re-oxygenation in the pres- cells or circulating leukocytes protected against renal IR in- ence or absence of C3a or C5a. Under hypoxia-reoxygenation sult and further confirms that expression of C3aR and C5aR conditions, C3a or C5a induced a clear increase of TNF-a, on both renal intrinsic cells and infiltrating leukocytes contrib- IL-6, and KC production by macrophages, in a dose-dependent utes to renal IR injury. manner (C3a: 0–50 nM, C5a: 0–20 nM), compared with medium controls, and a combination of C3a and C5a in- C3a/C5a Stimulation Increases Proinflammatory duced a further increase in both cytokines and chemokines Cytokine/Chemokine Production by Macrophages by macrophages (Figure 6, A–C). These data suggest that under Hypoxia-Reoxygenation Conditions leukocytes such as macrophages during IR can be activated Our data presented so far suggest that C3aR and C5aR signal- by C3a and C5a to produce proinflammatory cytokines ing is a critical pathogenic factor in the development of renal and chemokines. IR injury; this involves both renal cells and infiltrating leukocytes. Macrophages are important sources of proinflam- C3a/C5a Stimulation Increases Proinflammatory matory mediators. To further investigate the mechanisms by Cytokine and Chemokine Production and KIM-1 which C3aR and C5aR on infiltrating cells contribute to local Expression by Proximal Tubular Epithelial Cells under inflammation, we evaluated whether engagement of C3aR Hypoxia-Reoxygenation Conditions and C5aR on macrophages under hypoxia-reoxygenation con- Renal proximal tubular epithelial cells (PTECs) located at the ditions enhances their production of inflammatory mediators. corticomedullary zone are highly susceptible to IR injury. They

J Am Soc Nephrol 23: 1474–1485, 2012 C3aR and C5aR in Renal Injury 1479 BASIC RESEARCH www.jasn.org

are also capable of producing several inflammatory cytokines and chemokines in response to pathogenic stimuli. We therefore evaluated whether engagement of C3aR and C5aR on primary cultured PTECs under hypoxia-reoxygenation conditions enhances their production of inflammatory mediators. Similar to our observations on macrophages, secretion of TNF-a and KC by PTECs was significantly increased by the stimulation of C3a or C5a or C3a/C5a (Figure 7, A and B). In addition to cytokines and chemokines, we also evaluated whether engagement of C3aR and/or C5aR on PTECs under hypoxia-reoxygenation conditions enhances their expression of KIM-1 (Figure 7C). Consistent with the observation in postischemic kidneys that KIM-1 expression was significantly lower in all receptor knockout mice, KIM-1 expression in PTECs was upregulated by the stimulation of C3a or C5a or C3/C5a. Taken together the data suggest a direct activation of renal epithelial cells by C3a and C5a during renal IR.

DISCUSSION

Previous studies have suggested that complement activation is a key mediator of renal IR injury and that the terminal product of complement activation (C5b-9) contributes to the effect of complement on mouse renal IR injury.5,13 More re- cently, several studies have also suggested a pathogenic role for the small complement fragment C5a in renal IR injury.22–25 However, it is unclear whether C3a also contributes to the injury, whether integration of the actions of C3a and C5a is required for the pathogenesis, and how they participate in the development of renal IR injury. These questions are important because they have implications for therapeutic strategies. This study has addressed these issues. The use of receptor knockout mice (C3aR2/2,C5aR2/2, DKO) allowed us to define the role of C3aR and C5aR signaling in renal IR injury in a more specific manner than that used with the pharmaceutical tar- geting of C5 or C5aR described in previous studies.22,23 The use of chimera mice allowed us to evaluate the contribution of C3aR and C5aR on renal cells and circulating leukocytes to renal IR injury. In addition, an in vitro model of hypoxia- Figure 6. C3a and C5a stimulation increases proinflammatory reoxygenation helped us to investigate the potential mecha- cytokine and chemokine production by macrophages under nisms by which expression of C3aR and C5aR on renal tubular hypoxia-reoxygenation conditions. Peritoneal macrophages epithelial cells and macrophages could contribute to renal IR were prepared from naive C57BL/6 mice and exposed to hypoxia injury. The results presented here provide definitive evidence for 24 hours, followed by 24 hours of reoxygenation, in the supporting a predominant role for C5a and a minor role for presence or absence of C5a (10 and 20 nM) and/or C3a (20 and C3a in the pathogenesis of renal IR injury. Our study also 50 nM). The supernatants were harvested and used for analysis of identified several processes of renal IR injury that depend on cytokines and chemokine. Data are shown as mean 6 SEM C3aR and C5aR signaling, such as renal tissue inflammation, of pooled results from four individual experiments and were upregulation of KIM-1, activation of renal tubular epithelial , , analyzed by t test *P 0.05 **P 0.005 for nonstimulation versus cells, cellular infiltration, and endothelial activation. C3a or C5a or C3a/C5a stimulation under hypoxia-reoxygenation An important question arising from our observation that conditions. the lack of C3aR and C5aR protects mice from renal IR injury is how C3a and C5a participate in the development of renal IR injury. The pathologic hallmark of renal IR injury, during the acute phase, is renal tubule destruction and cellular

1480 Journal of the American Society of Nephrology J Am Soc Nephrol 23: 1474–1485, 2012 www.jasn.org BASIC RESEARCH

infiltration of neutrophils,monocytes, andmacrophages. Given that significant expression of C3aR and C5aR is observed on both renal tubular epithelial cells and granulocytes, and the biologic functions of C3a and C5a in cell activation and C5a in chemotaxis, it is therefore possible that C3a and C5a participate in the pathogenesis of renal IR injury through promoting local inflammation by production of cytokines and chemokines by both renal tubular epithelial cells and infiltrating cells after engagement with C3aR and C5aR. The observations made in this study strongly support this hypothesis. First, C3a and C5a stimulation increased TNF-a and KC production by PTECs and macrophages, in response to hypoxia-reoxygenation. Production of TNF-a and KC may mediate epithelial cell injury by autocrine and paracrine signaling, whereas KC promotes leukocyte chemotaxis to the renal tissue. C3a and C5a stimulation also increased IL-6 production by macrophages, which could stimulate innate IL-17–producing cells leading to local inflammation. Second, consistent with the above observations, renal expression of inflammatory mediators (e.g., TNF-a, IFN-g,IL-1b, KC, MCP-1) after renal IR was significantly reduced in C5aR2/2 and DKO mice and slightly reduced in C3aR2/2 mice when compared with WTmice, as was the infiltration of neutrophils, monocytes, and macrophages in postischemic kidneys. Furthermore, renal injury was reduced in mice that lack C3aR and C5aR on renal intrinsic cells or circulating leukocytes, suggesting that C3aR C5aR on both renal cells and circulating leukocytes contribute to renal IR injury. Our in vivo and in vitro studies showed an important effect of C3aR and C5aR signaling on KIM-1 expression (Figure 2, E and C). Although the exact function of this molecule still remains unclear, recent evidence has suggested that KIM-1 is not only a specific marker for acute renal tubule injury but also a functional molecule in the process of renal injury. It has been shown that KIM-1 can act as an endogenous ligand for leukocyte mono-immunoglobulin-like receptor 5 expressed on myeloid cells such as neutrophils, engagement of which leads to activation of cells and increased cell migration and chemokine production in the process of renal IR injury,28 thereby con- tributing to local inflammation. Therefore, our observation that increased KIM-1 expression in response to IR or hypoxia- regeneration depends on C3aR and C5aR signaling suggests Figure 7. C3a and C5a stimulation increases proinflammatory that C3a and C5a could participate in the pathogenesis of renal cytokine and chemokine production and KIM-1 expression by IR injury by modulating KIM-1 expression, consequently pro- PTECs under hypoxia-reoxygenation conditions. Primary PTECs moting local inflammation and renal tissue injury. were prepared from WT kidneys and exposed to hypoxia for With regard to the involvement of endothelial cells in renal 24 hours, followed by 24 hours of reoxygenation, in the presence IR injury, our previous studies have found no apparent signs of or absence of C5a (10 and 20 nM) and/or C3a (20 and 50 nM). vascular abnormalities (e.g., complement deposition, throm- The supernatants were used for analyzing cytokine and chemokine secretion, and the cell pellets were used for analyzing gene bosis) in postischemic kidney by histologic and immuno- expression. (A and B) Secretion of TNF-a and KC. (C) mRNA chemical analysis, suggesting that endothelium injury is not 5 expression of KIM-1. Data are shown as mean 6 SEM of pooled the major feature in murine renal IR injury. In this study, to results from four individual experiments and were analyzed by investigate whether there are some more subtle changes in t test *P,0.05, **P,0.005, ***P,0.0001 for nonstimulation ver- endothelium and whether such changes would depend on sus C3a or C5a or C3a/C5a stimulation under hypoxia/reoxyge- C3aR and C5aR signaling, we analyzed renal expression of nation conditions. several adhesion molecules and endothelial activation associated molecules in postischemic kidneys. Our data suggest that

J Am Soc Nephrol 23: 1474–1485, 2012 C3aR and C5aR in Renal Injury 1481 BASIC RESEARCH www.jasn.org renal endothelial activation occurs in response to renal IR, this infiltration in postischemic kidneys. A question raised from being more notable in the presence of C3aR/C5aR signaling. these studies is how C3a could contribute to leukocyte However, given that expression of C3aR and C5aR is more chemotaxis. Several in vitro studies have shown that C3a abundant in both renal tubular epithelial cells and inflamma- induces the expression of cytokines and chemokines (e.g., tory cells than in renal endothelial cells,15,23,29 and the relative IL-1, IL-6, IL-8, MCP-1, and RANTES [regulated upon acti- lack of endothelium involvement in renal IR injury,5 it re- vation, normal T cell expressed and secreted]) by endothelial mains possible that the activation of renal tubular epithelial cells and astroglioma cells,41,42 whereas selective antagonism cells and the infiltration or activation of inflammatory immune of C3aR significantly attenuates production of MIP-2 and KC cells are the major mechanisms of C3aR- and C5aR-mediated by a renal proximal tubular epithelial cells line.27 These ob- renal IR injury; C3a- and C5a-mediated local inflammation servations, together with our findings that C3a stimulation leads to endothelial activation, enabling leukocyte recruitment. increases proinflammatory cytokine and chemokine produc- C3a and C5a are commonly described as inflammatory me- tion by both renal tubular epithelial cells and macrophages un- diators with similar profiles of activity; however, differen- der hypoxia-reoxygenation conditions, strongly suggest that tial or opposing effects of C3a and C5a have been reported C3a can contribute to leukocyte chemotaxis, possibly through in several other disease models, such as asthma,30,31 systemic inducing the production of other chemoattractants (e.g.,IL-8, lupus erythematosus,32,33 gram-negative bacteremia, and KC,MCP-1,orMIP)bykidneycellsandimmunecellsatthe shock.34 In renal IR injury, previous studies have suggested site of injury. However, whether C3a has direct chemotactic that C5a is a pathogenic factor. Although the contribution of effect on neutrophils and monocytes in renal IR injury remains C3a to renal IR injury is unclear, studies in other organs (heart, to be elucidated. intestine, brain) have demonstrated that the severity of IR in- On the basis of our findings, together with published jury in these tissues was reduced in the absence of C3aR sig- literature about the role of C3a and C5a in local inflammation naling.35–38 We have now clearly shown that both C3aR and and cell activation, we postulate the mechanisms by which C5aR signaling can contribute to renal IR injury; multiple C3a and C5a contribute to the pathogenesis of renal IR injury pathogenic processes (e.g., KIM-1 expression, cellular infiltra- through acting on renal and circulating cells. As illustrated in tion, generation of proinflammatory mediators in ischemic Figure 8, complement activation in response to renal ische- kidneys) depend on signaling by both receptors. Together, mic insults results in the production of C3a and C5a. Binding these studies suggest that in contrast to the opposing effect of both C3a and C5a to specific receptors (C3aR and C5aR) observed in other disease models, in the setting of renal IR, on renal tubular epithelial cells and granulocytes, such as both C3a and C5a contribute to the acute renal injury monocytes and macrophages, leads to cell activation and despite a predominant role for C5a in the pathogenesis. The upregulation of proinflammatory cytokine and chemokine different roles of C3a in different disease models could be ex- production. These inflammatory mediators lead not only to plained by different types of inflammation (e.g., with or with- tubular epithelial cell injury but also to recruitment of more out involvement of infectious agent or antigen, acute or leukocytes into the interstitium. In addition, C3a and C5a chronic) and different types of target cells (e.g., lymphocytes cause renal endothelial activation and thereby enable leuko- or granulocytes and parenchymal cells). It is worth noting that cyte recruitment, which can be mediated by C5a and other although our in vitro experiments indicate that C3a and C5a chemoattractants. Therefore, in addition to C5b-9–medi- have similar effects on activation of renal tubular epithelial ated direct injury of renal tubule cells,5,14 C5a- (and possible cells and macrophages to produce inflammatory mediators, C3a-) induced inflammatory responses also contribute to our in vivo and ex vivo experiments show that renal functional renal tissue destruction. Our findings offer new insight into impairment, renal expression of inflammatory mediators, and the effector molecules of complement activation involved in the infiltration of leukocytes in postischemic kidneys are more the pathogenesis of renal IR injury, which has important impli- dependent on C5aR signaling. This finding suggests that al- cations for therapeutic strategies needed to treat this disease. though C3a contributes to renal IR injury, the pathogenic role of C5a in renal IR injury predominates over that of C3a. With regard to the chemotactic activity of C3a and C5a, CONCISE METHODS although it is well known that C5a is a potent chemoattractant for leukocytes, the role for C3a in leukocyte chemotaxis is less Reagents clear. Several lines of evidence have suggested that C3a can Recombinant mouse C5a was from Hycult (Uden, The Netherlands). contribute to leukocyte chemotaxis. Studies in C3aR2/2 mice Human C3a was purchased from Fitzgerald Industries International have found that absence of C3aR attenuates the spinal cord- (Concord, MA). Hydrocortisone, tri-iodothyronine, collagenase II, infiltrating macrophages and CD4+ T cells in an experimental and deoxyribonuclease I were from Sigma-Aldrich (Poole, United autoimmune encephalomyelitis model and reduces airway Kingdom). The antibodies for CD45, Gr-1, and F4/80 were from eosinophils, neutrophils, and lymphocytes in a model of al- BioLegend (San Diego, CA). Horseradish peroxidase–conjugated lergic bronchopulmonary inflammation.39,40 In the present rabbit antirat polyclonal antibody was from DAKO (Cambridge, study, we observed that absence of C3aR reduces cellular United Kingdom). FcR-blocking antibody (CD16/32) and mouse

1482 Journal of the American Society of Nephrology J Am Soc Nephrol 23: 1474–1485, 2012 www.jasn.org BASIC RESEARCH

arteries and veins were bilaterally occluded for 30 minutes with microaneurysm clamps. After re- moval of the clamps, 0.8 ml of warm saline was placed in the abdomen, the incision sutured, and animals returned to their cages. Blood samples were taken at 24 hours and 48 hours after reperfusion by tail bleeding (or cardiac puncture for terminal mice). In some experiments, mice were euthanized at 24 or 48 hours to collect the kidneys. Renal function was assessed by measuring concentrations of BUN using a Urea Assay Kit from Thermo Fisher (Epsom, United Kingdom). To assess renal pathologic features, the injured kidneys were fixed in a solution of 4% formalin in PBS for 24 hours and embedded in paraffin. Sections (3 mm) were stained with hematoxylin and eosin and periodic acid-Schiff. Histologic score was performed in a blinded fashion by two persons as described previously,5 with some modification. The percentage of tubules in the Figure 8. Proposed mechanism underlying the contribution of C3a and C5a to renal IR corticomedullary junction showing epithelial injury. Complement activation in response to ischemic insults results in the production of C3a and C5a. Binding of C3a and C5a to their specific receptors C3aR and C5aR on injury (i.e., loss of proximal tubule brush bor- (1) renal tubular epithelial cells and (2)inflammatory cells causes the cell activation and der, cell blebbing or vacuolization, and cell ne- fi upregulation of proinflammatory cytokine and chemokine production, (3) on endothelial crosis) was quanti ed using a 7-point scale (0, cells causes the cell activation that enables leukocyte recruitment. In addition to the direct noinjury;1,0%–15%; 2, 15%–30%; 3, 30%– effect of C5a on leukocyte chemotaxis, both C3a and C5a may indirectly mediate leu- 45%, 4, 45%–60%; 5, 60%–75%; 6, .75%) at a kocyte chemotaxis through inducing the production of other chemoattractants by renal magnification of 3100; three fields per kidney cells and monophages and macrophages at the site of injury. Thus, C5a and possibly C3a were assessed. mediated these inflammatory events, and, together with the actions by other complement components, such as C5b-9, contribute to acute renal injury after renal ischemic insult. Quantitative Real-Time RT-PCR Total RNA was extracted from the kidney tissues using an SV total RNA isolation system and used cytometric bead array kit for KC, IL-6, and TNF-a were from BD for cDNA synthesis using an RT kit according to manufacturer’s Biosciences (Cowley, United Kingdom). Collagenase D and Dispase instructions from Promega (Southampton, United Kingdom). II were from Roche (Welwyn, United Kingdom). GoTaq Hot Start Quantitative PCR was performed with a DyNAmo HS SYBR Green Polymerase was from Promega (Southampton, United Kingdom). qPCR kit from Thermo Fisher (Epsom, United Kingdom and an MJ Cell culture medium, FCS, and insulin-transferrin-selenium solution Research PTC-200 Peltier Thermal Cycler from Bio-Rad (Hemel were purchased from Invitrogen (Paisley, United Kingdom). Hempstead, United Kingdom). Amplification was performed according to the manufacturer’s cycling protocol and done in dupli- 2DD (Ct) Mice cate. Gene expression was expressed as 2 , where Ct is cycle Homozygous C3aR2/2 and C5aR2/2 mice were derived by homol- threshold, DD (Ct) is testing samples D (Ct) 2 control samples D ogous recombination in embryonic stem cells43,44 and backcrossed (Ct), and D (Ct) is testing gene (Ct) -18s (Ct). The control samples on to the C57BL/6 (H-2b) parental strain for at least 12 generations. were normal kidney tissues (in the experiments of postischemic kid- C5aR and C3aR DKO mice were generated from cross-breeding of neys) or unstimulated cells (in the experiments of epithelial cells and C3aR2/2 and C5aR2/2 mice. WT C57BL/6 mice were purchased macrophages). The information for primer sequences is given in from Harlan Laboratories (United Kingdom). Male mice (8–12 weeks Supplemental Table 1. old) were used in all experiments. Animal procedures adhered to criteria outlined by the Animals (Scientific Procedures) Act of 1986. Immunochemical Staining Frozen sections (4 mm) of optimal cutting temperature–embedded Induction of Renal IR Injury and Assessment of Renal kidneys were air-dried for a minimum of 20 minutes at room tem- Function and Pathology perature and then acetone fixed. Sections were incubated with rat Renal IR injury was induced as we described previously with a modified antimouse Gr-1 or F4/80 antibody, followed by horseradish perox- ischemic time.5 In brief, mice were anesthetized with isoflurane and the idase–conjugated rabbit antirat antibody. Stained kidney sections body temperature was kept constant by placing a warm pad beneath were visualized under light microscopy and photographed at a mag- the animals during surgery. Using a midline abdominal incision, renal nification of 3200.

J Am Soc Nephrol 23: 1474–1485, 2012 C3aR and C5aR in Renal Injury 1483 BASIC RESEARCH www.jasn.org

Flow Cytometric Analysis Department of Health via the National Institute for Health Re- Single renal cell suspension was prepared using the method described search comprehensive Biomedical Research Centre award to Guy’s elsewhere,45 with some modification. In brief, the whole kidney was and St Thomas’s NHS Foundation Trust in partnership with King’s minced and digested with a cocktail containing collagenase D College London and King’s College Hospital NHS Foundation (0.25 mg/ml), Dispase II (1.5 mg/ml), deoxyribonuclease (0.01 mg/ml), Trust. and FCS (10%) for 40 minutes at 37°C. The digested tissue mixture was passed through a 40-mm nylon sieve, and the cells were used for flow cytometry. Renal suspension cells (33106) were preincubated with FcR-blocking antibody (CD16/32) and then stained with rat antimouse DISCLOSURES allophycocyanin–conjugated CD45 and phycoerythrin-conjugated None. Gr-1 or F4/80 antibodies or the appropriate isotype control antibodies at 4°C for 20 minutes, and then fixed in 400 ml of 1% paraformaldehyde in PBS. All flow cytometric analyses were performed using Calibur Flow REFERENCES Cytometer (BD) and Flowjo software (Tree Star, Ashland, OR). 1. Troppmann C, Gillingham KJ, Benedetti E, Almond PS, Gruessner RW, Generation of Chimeric Mice Najarian JS, Matas AJ: Delayed graft function, acute rejection, and outcome Chimeric mice were generated using a method established in our after cadaver renal transplantation. The multivariate analysis. Trans- – laboratory.46 Mice were first irradiated at a dose of 9 Gy and recon- plantation 59: 962 968, 1995 2. Pagtalunan ME, Olson JL, Tilney NL, Meyer TW: Late consequences of acute stituted with an intravenous administration of 7.53106 donor bone ischemic injury to a solitary kidney. J Am Soc Nephrol 10: 366–373, 1999 marrow cells within 3 hours after irradiation. At 8 weeks after bone 3. Herrero-Fresneda I, Torras J, Cruzado JM, Condom E, Vidal A, Riera M, marrow transplantation, mice were used for induction of renal IR. Lloberas N, Alsina J, Grinyo JM: Do alloreactivity and prolonged cold ischemia cause different elementary lesions in chronic allograft nephropathy? Am J Pathol 162: 127–137, 2003 Induction of Hypoxia-Reoxygenation Injury in 4. Abu Jawdeh BG, Rabb H: Delayed kidney allograft function what does it Naive Macrophages and Primary Renal tell us about acute kidney injury? Contrib Nephrol 174: 173–181, 2011 Tubular Epithelial Cells 5. Zhou W, Farrar CA, Abe K, Pratt JR, Marsh JE, Wang Y, Stahl GL, Sacks Peritoneal macrophages were harvested from naive male C57BL/6 SH: Predominant role for C5b-9 in renal ischemia/reperfusion injury. mice (two or three mice per experiment) and cultured for 1 hour to remove J Clin Invest 105: 1363–1371, 2000 6. Loubele ST, ten Cate H, Spronk HM: Anticoagulant therapy in critical nonadhesive cells. Macrophages were placed in duplicate in a 24-well plate – 3 6 organ ischaemia/reperfusion injury. Thromb Haemost 104: 136 142, (1 10 cells per well). Primary tubular epithelial cell cultures were pre- 2010 pared from kidneys of naive male C57BL/6mice as we described previ- 7. Rabb H, Daniels F, O’Donnell M, Haq M, Saba SR, Keane W, Tang WW: ously.47 In brief, minced cortex was digested with 0.1% collagenase II and Pathophysiological role of T lymphocytes in renal ischemia-reperfusion passed through a 40-mm nylon sieve. The cells and tubules were collected injury in mice. Am J Physiol Renal Physiol 279: F525–F531, 2000 and cultured in a DMEM-12 medium that contained 2% FCS, insulin 8. Awad AS, Rouse M, Huang L, Vergis AL, Reutershan J, Cathro HP, Linden J, Okusa MD: Compartmentalization of neutrophils in the kidney and lung (5 mg/ml), transferrin (5 mg/ml), selenium (5 ng/ml), hydrocortisone following acute ischemic kidney injury. Kidney Int 75: 689–698, 2009 212 (40 ng/ml), and tri-iodothyronine (10 M). Nonpassaged 7-day cultured 9.TakadaM,NadeauKC,ShawGD,MarquetteKA,TilneyNL:The cells were used. Both cells were cultured under hypoxic conditions (5% cytokine-adhesion molecule cascade in ischemia/reperfusion injury of the rat kidney. Inhibition by a soluble P-selectin ligand. J Clin Invest CO2 and 95% N2) in a gas chamber (Billups-Rothenberg, Del Mar, CA) for – 24 hours, followed by reoxygenation (21% O and 5% CO ) for 24 99: 2682 2690, 1997 2 2 10. Ricklin D, Hajishengallis G, Yang K, Lambris JD: Complement: A key hours in the presence of C5a or C3a. The supernatants were collected system for immune surveillance and homeostasis. Nat Immunol 11: and used for measuring cytokines (TNF-a,IL-6)andchemokine(KC)by 785–797, 2010 cytometric bead array assay (according to the manufacturer’s instructions); 11. Thurman JM, Ljubanović D, Royer PA, Kraus DM, Molina H, Barry NP, in some experiments, total RNA was extracted from mouse PTECs and Proctor G, Levi M, Holers VM: Altered renal tubular expression of quantitative RT-PCR was performed to assess KIM-1 expression. the complement inhibitor Crry permits complement activation after ischemia/reperfusion. JClinInvest116: 357–368, 2006 12. Yamada K, Miwa T, Liu J, Nangaku M, Song WC: Critical protection Statistical Analyses from renal ischemia reperfusion injury by CD55 and CD59. JImmunol Statistical analyses between different experimental groups were per- 172: 3869–3875, 2004 formed using the Mann-Whitney test or t test. All the analyses were 13. Turnberg D, Botto M, Lewis M, Zhou W, Sacks SH, Morgan BP, Walport MJ, Cook HT: CD59a deficiency exacerbates ischemia-reperfusion in- performed using Graphpad Prism software. jury in mice. Am J Pathol 165: 825–832, 2004 14. Fregonese L, Swan FJ, van Schadewijk A, Dolhnikoff M, Santos MA, Daha MR, Stolk J, Tschernig T, Sterk PJ, Hiemstra PS, Rabe KF, Mauad T: Expression of the anaphylatoxin receptors C3aR and C5aR is in- ACKNOWLEDGMENTS creased in fatal asthma. J Allergy Clin Immunol 115: 1148–1154, 2005 15. Braun MC, Reins RY, Li TB, Hollmann TJ, Dutta R, Rick WA, Teng BB, Ke B: Renal expression of the C3a receptor and functional responses of This study was supported by Medical Research Council of the United primary human proximal tubular epithelial cells. JImmunol173: 4190– Kingdom. The authors also acknowledge the support from the 4196, 2004

1484 Journal of the American Society of Nephrology J Am Soc Nephrol 23: 1474–1485, 2012 www.jasn.org BASIC RESEARCH

16. Tang Z, Lu B, Hatch E, Sacks SH, Sheerin NS: C3a mediates epithelial- 32. Wenderfer SE, Ke B, Hollmann TJ, Wetsel RA, Lan HY, Braun MC: C5a to-mesenchymal transition in proteinuric nephropathy. JAmSoc receptor deficiency attenuates T cell function and renal disease in Nephrol 20: 593–603, 2009 MRLlpr mice. JAmSocNephrol16: 3572–3582, 2005 17. Abe K, Miyazaki M, Koji T, Furusu A, Nakamura-Kurashige T, Nishino T, 33. Wenderfer SE, Wang H, Ke B, Wetsel RA, Braun MC: C3a receptor Ozono Y, Harada T, Sakai H, Kohno S: Enhanced expression of com- deficiency accelerates the onset of renal injury in the MRL/lpr mouse. plement C5a receptor mRNA in human diseased kidney assessed by in Mol Immunol 46: 1397–1404, 2009 situ hybridization. Kidney Int 60: 137–146, 2001 34. Hollmann TJ, Mueller-Ortiz SL, Braun MC, Wetsel RA: Disruption of the 18. Fayyazi A, Scheel O, Werfel T, Schweyer S, Oppermann M, Götze O, C5a receptor gene increases resistance to acute Gram-negative bac- Radzun HJ, Zwirner J: The C5a receptor is expressed in normal renal teremia and endotoxic shock: Opposing roles of C3a and C5a. Mol proximal tubular but not in normal pulmonary or hepatic epithelial cells. Immunol 45: 1907–1915, 2008 Immunology 99: 38–45, 2000 35. Busche MN, Stahl GL: Role of the complement components C5 and 19. Martin U, Bock D, Arseniev L, Tornetta MA, Ames RS, Bautsch W, Köhl J, C3a in a mouse model of myocardial ischemia and reperfusion injury. Ganser A, Klos A: The human C3areceptoris expressed on neutrophils and Ger Med Sci 8: 8, 2010 monocytes, but not on B or T lymphocytes. JExpMed186: 199–207, 1997 36. Proctor LM, Arumugam TV, Shiels I, Reid RC, Fairlie DP, Taylor SM: 20. Soruri A, Kim S, Kiafard Z, Zwirner J: Characterization of C5aR expres- Comparative anti-inflammatory activities of antagonists to C3a and sion on murine myeloid and lymphoid cells by the use of a novel C5a receptors in a rat model of intestinal ischaemia/reperfusion injury. monoclonal antibody. Immunol Lett 88: 47–52, 2003 Br J Pharmacol 142: 756–764, 2004 21. ThurmanJM,RoyerPA,LjubanovicD,DursunB,LenderinkAM,Edelstein 37. Ducruet AF, Hassid BG, Mack WJ, Sosunov SA, Otten ML, Fusco DJ, CL, Holers VM: Treatment with an inhibitory monoclonal antibody to Hickman ZL, Kim GH, Komotar RJ, Mocco J, Connolly ES: C3a receptor mouse factor B protects mice from induction of apoptosis and renal is- modulation of granulocyte infiltration after murine focal cerebral ischemia/reperfusion injury. J Am Soc Nephrol 17: 707–715, 2006 chemia is reperfusion dependent. J Cereb Blood Flow Metab 28: 22. Arumugam TV, Shiels IA, Strachan AJ, Abbenante G, Fairlie DP, Taylor 1048–1058, 2008 SM: A small molecule C5a receptor antagonist protects kidneys from 38. Mocco J, Mack WJ, Ducruet AF, Sosunov SA, Sughrue ME, Hassid BG, ischemia/reperfusion injury in rats. Kidney Int 63: 134–142, 2003 Nair MN, Laufer I, Komotar RJ, Claire M, Holland H, Pinsky DJ, Connolly 23. de Vries B, Köhl J, Leclercq WK, Wolfs TG, van Bijnen AA, Heeringa P, ES Jr: Complement component C3 mediates inflammatory injury fol- Buurman WA; de VB: Complement factor C5a mediates renal ischemia- lowing focal cerebral ischemia. Circ Res 99: 209–217, 2006 reperfusion injury independent from neutrophils. JImmunol170: 39. Boos L, Campbell IL, Ames R, Wetsel RA, Barnum SR: Deletion of the 3883–3889, 2003 complement anaphylatoxin C3a receptor attenuates, whereas ectopic 24. Zheng X, Zhang X, Feng B, Sun H, Suzuki M, Ichim T, Kubo N, Wong A, expression of C3a in the brain exacerbates, experimental autoimmune Min LR, Budohn ME, Garcia B, Jevnikar AM, Min WP: Gene silencing encephalomyelitis. JImmunol173: 4708–4714, 2004 of complement C5a receptor using siRNA for preventing ischemia/ 40. Drouin SM, Corry DB, Hollman TJ, Kildsgaard J, Wetsel RA: Absence of reperfusion injury. Am J Pathol 173: 973–980, 2008 the complement anaphylatoxin C3a receptor suppresses Th2 effector 25. Lewis AG, Köhl G, Ma Q, Devarajan P, Köhl J: Pharmacological tar- functions in a murine model of pulmonary allergy. JImmunol169: geting of C5a receptors during organ preservation improves kidney 5926–5933, 2002 graft survival. Clin Exp Immunol 153: 117–126, 2008 41. Monsinjon T, Gasque P, Chan P, Ischenko A, Brady JJ, Fontaine MC: 26. Ichimura T, Bonventre JV, Bailly V, Wei H, Hession CA, Cate RL, Regulation by complement C3a and C5a anaphylatoxins of cytokine Sanicola M: Kidney injury molecule-1 (KIM-1), a putative epithelial cell production in human umbilical vein endothelial cells. FASEB J 17: adhesion molecule containing a novel immunoglobulin domain, is 1003–1014, 2003 up-regulated in renal cells after injury. J Biol Chem 273: 4135–4142, 42. Jauneau AC, Ischenko A, Chan P, Fontaine M: Complement component 1998 anaphylatoxins upregulate chemokine expression by human astrocytes. 27. Thurman JM, Lenderink AM, Royer PA, Coleman KE, Zhou J, Lambris FEBS Lett 537: 17–22, 2003 JD, Nemenoff RA, Quigg RJ, Holers VM: C3a is required for the pro- 43. Humbles AA, Lu B, Nilsson CA, Lilly C, Israel E, Fujiwara Y, Gerard NP, duction of CXC chemokines by tubular epithelial cells after renal ishe- Gerard C: A role for the C3a anaphylatoxin receptor in the effector mia/reperfusion. J Immunol 178: 1819–1828, 2007 phase of asthma. Nature 406: 998–1001, 2000 28. Yamanishi Y, Kitaura J, Izawa K, Kaitani A, Komeno Y, Nakamura M, 44. Höpken UE, Lu B, Gerard NP, Gerard C: The C5a chemoattractant re- Yamazaki S, Enomoto Y, Oki T, Akiba H, Abe T, Komori T, Morikawa Y, ceptor mediates mucosal defence to infection. Nature 383: 86–89, 1996 Kiyonari H, Takai T, Okumura K, Kitamura T: TIM1 is an endogenous 45. Dong X, Swaminathan S, Bachman LA, Croatt AJ, Nath KA, GriffinMD: ligand for LMIR5/CD300b: LMIR5 deficiency ameliorates mouse kidney Resident dendritic cells are the predominant TNF-secreting cell in early ischemia/reperfusion injury. JExpMed207: 1501–1511, 2010 renal ischemia-reperfusion injury. Kidney Int 71: 619–628, 2007 29. Zahedi R, Braun M, Wetsel RA, Ault BH, Khan A, Welch TR, Frenzke M, 46. Brown HJ, Sacks SH, Robson MG: Toll-like receptor 2 agonists exacerbate Davis AE: The C5a receptor is expressed by human renal proximal tu- accelerated nephrotoxic nephritis. JAmSocNephrol17: 1931–1939, 2006 bular epithelial cells. Clin Exp Immunol 121: 226–233, 2000 47. Li K, Patel H, Farrar CA, Hargreaves RE, Sacks SH, Zhou W: Comple- 30. Zhang X, Lewkowich IP, Köhl G, Clark JR, Wills-Karp M, Köhl J: A pro- ment activation regulates the capacity of proximal tubular epithelial cell tective role for C5a in the development of allergic asthma associated to stimulate alloreactive T cell response. JAmSocNephrol15: 2414– with altered levels of B7-H1 and B7-DC on plasmacytoid dendritic cells. 2422, 2004 J Immunol 182: 5123 –5130, 2009 31. Lajoie S, Lewkowich IP, Suzuki Y, Clark JR, Sproles AA, Dienger K, Budelsky AL, Wills-Karp M: Complement-mediated regulation of the IL-17A axis is a central genetic determinant of the severity of experi- This article contains supplemental material online at http://jasn.asnjournals. mental allergic asthma. Nat Immunol 11: 928–935, 2010 org/lookup/suppl/doi:10.1681/ASN.2011111072/-/DCSupplemental.

J Am Soc Nephrol 23: 1474–1485, 2012 C3aR and C5aR in Renal Injury 1485 Supplemental figure 1 Histopathology of kidneys after 48h reperfusion

5 * 4

3 *** *** 2

1 Histology scores Histology

0 - - -/ -/ WT R R DKO C3a C5a

Renal IR injury was induced in four groups of mice: WT (n=12), C3aR-/- (n=6), C5aR-/- (n=6) and DKO (n=12 per group). Kidney tissues were collected at 48h after reperfusion. Tissue injury were analyzed in PAS stained sections by histological scores. Data were shown as mean ± SEM (n=4 per group, 3 fields of each kidney) and analyzed by Mann- Whitney test . * P < 0.05 , *** P < 0.0001 , comparing WT and C3aR -/- or C5aR -/- or DKO mice. Supplemental figure 2

Renal expression of KIM-1 is reduced in C3aR-/-, C5aR-/- and DKO mice, compared with WT mice

400

300

200 ** **

100 (Arbitrary unit) (Arbitrary

0 KIM-1 mRNA expression mRNA KIM-1 - - -/ WT R 6 DKO B 3a C C5aR-/

Renal IR injury was induced in four groups of mice: WT, C3aR-/- , C5aR-/- and DKO (n=6 per group). kidney tissues were collected at 48h after reperfusion. KIM-1 expression in postischemic kidneys was analysed by quantitative real time RT-PCR. The dotted line indicates KIM-1 gene expression level in normal kidney tissues. Data were shown as mean ± SEM and analyzed by Mann-Whitney test, ** P < 0.005, comparing WT and C3aR-/- or C5aR-/- or DKO mice. Supplemental figure 3

Deficiency of C3aR/C5aR reduces renal production of inflammatory mediators following renal IR

A 120 WT C3aR-/- 90 * C5aR-/- DKO * 60 *

(Arbitrary (Arbitrary unit) 30 mRNA expression mRNA * * 0 TNF- ααα IL-1 βββ IFN- γγγ

B 40 WT C3aR-/- 30 C5aR-/- DKO * 20 * * *

(Arbitrary (Arbitrary unit) 10 mRNA expression mRNA * * 0 KC MCP-1 MIP-1 ααα MIP-1 βββ C 40 WT C3aR-/- 30 C5aR-/- DKO 20 *

(Arbitrary Units) (Arbitrary 10 mRNA expression mRNA

0 ICAM-1 CD146 vWf

Renal IR injury was induced in WT, C3aR-/-, C5aR-/- and DKO mice (n=5 per group). Kidney tissues were harvested at 48h after reperfusion and used for quantitative real time RT-PCR analysis. A. Pro-inflammatory cytokines (TNF-α, IL-1β and IFN-γ). B. Chemokines (KC, MIP-1α, MIP-1β and MCP-1). C. Adhesion molecules and endothelial activation/injury markers (ICAM-1, CD146 and vWF). The dotted lines in A.B.C indicate the gene expression levels in normal kidney tissue. Data were shown as mean ± SEM and analyzed by Mann-Whitney test, * P < 0.05, comparing WT and C3aR-/- or C5aR-/- or DKO mice. Supplemental figure 4

Proportions of CD45 +, Gr-1+ or F4/80 + cells in kidneys from naïve WT, C3aR-/-, C5aR-/- and DKO mice

3 3 3

2 2 2

1 1 1 CD45+ cells (%) cells CD45+ F4/80+ cells (%) cells F4/80+ Gr-1+ cells (%) Gr-1+ cells

0 0 0 T /- O - W R- /- O -/ DK WT R-/- WT R aR- a DK DKO C3aR-/- C5a 3a C3 C5 C C5aR-/-

Renal single cell suspensions were prepared from naïve WT, C3aR-/-, C5aR-/- and DKO mice (n=6 per group). Flow cytometric analysis of CD45 +, Gr-1+ and F4/80 + cells was performed in the cell suspensions prepared from individual kidneys. Supplemental table

PCR primer sequences and product sizes

Primers Oligonucleotide Sequence Product Size Gene bank (5’ → 3’) (bp) code 18S-1 ATCCCTGAGAAGTTCCAGCA 153 NM_011296.1 18S-2 CCTCTTGGTGAGGTCGATGT C5aR-1 CAGGACATGGACCCCATAGAT 155 NM_007577.2 C5aR-2 ACCAGGAACACCACCGAGTAG C3aR-1 GATTCCATCTCAGTGTGCTTGAC 290 NM_009779.1 C3aR-2 GTGTCCTTGGAGAATCAGGTGAG TNF α-1 TGAGCACAGAAAGCATGATCC 200 ENSMUSG00000 TNF α-2 GCCATTTGGGAACTTCTCATC 024401 IL-1β-1 GCTCTCCACCTCAATGGACA 182 ENSMUSG00000 IL-1β-2 TTGGGATCCACACTCTCCAG 027398 IFN γ-1 ACT GGC AAA AGG ATG GTG AC 237 NM_008337.3 IFN γγγ-2 TGA GCT CAT TGA ATG CTT GG

KC-1 CTT GAA GGT GTT GCC CTC AG 181 ENSMUST00000 KC -2 ACA GGT GCC ATC AGA GCA GT 031327 MCP-1-1 GGCTCAGCCAGATGCAGTTA 219 ENSMUSG00000 MCP-1-2 ATTTGGTTCCGATCCAGGTT 035385 MIP-1α-1 CACTGCCCTTGCTGTTCTTC 262 ENSMUSG00000 MIP-1α-2 GGCATTCAGTTCCAGGTCAG 000982 MIP-1β-1 CCCAGCTCTGTGCAAACCTA 248 ENSMUSG00000 MIP-1β-2 TCTGCCTCTTTTGGTCAGGA 018930 ICAM-1-1 AGAGGTGACTGAGGAGTTCGAC 284 NM_010493.2 ICAM-1-2 AGAAGCTTCGTTTGTGATCCTC CD146-1 AGCCCCAGAGGAACCAACTA 234 ENSMUST00000 CD146-2 AGGCGTGCACTCAGAACACT 098852 vWf-1 CTTTGGGGACGACTTCATCA 228 ENSMUSG00000 vWf-2 TAGGGCATGGAGATGCTTTG 001930 KIM-1-1 AAGCCGCAGAAAAACCCTAC 206 ENSMUSG00000 KIM-1-2 TTGTCTTCAGCTCGGGAATG 040405

* Primer-1 is identical to the coding strand; primer-2 is complementary to the coding strand. All primers were designed such that there are intronic sequences between the primer 1 and primer 2.