tomoko takano, md, phd, hanan elimam, msc, and andrey v....

Complement-Mediated Cellular Injury

Tomoko Takano, MD, PhD, Hanan Elimam, MSc, and Andrey V. Cybulsky, MD

Summary: Complement activation and recruitment of inflammatory leukocytes is an important defense

0270-9295/ -& 2013 Elsevhttp://dx.doi.o

Financial supdian InstituMOP-53335C. and T.TMcLaughlin

Conflict of in

Department oUniversity,

Address reprNephrology,treal, Quebe

586

mechanism against bacterial infection. However, complement also can mediate cellular injury and contribute tothe pathogenesis of various diseases. With the appreciation that the C5b-9 membrane attack complex can injurecells in the absence of leukocytes, a role for the terminal complement pathway in inducing cell injury and kidneydisease was shown in several experimental models, including the rat passive Heymann nephritis model of humanmembranous nephropathy. In podocytes, sublytic C5b-9 activates a variety of downstream pathways includingprotein kinases, lipid metabolism, reactive oxygen species, growth factors/gene transcription, endoplasmicreticulum stress, and the ubiquitin-proteasome system, and it impacts the integrity of the cytoskeleton and slitdiaphragm proteins. C5b-9 also injures other kidney cells, including mesangial, glomerular endothelial, andtubular epithelial cells, and it contributes to the pathogenesis of mesangial-proliferative glomerulonephritis,thrombotic microangiopathy, and acute kidney injury. Conversely, certain C5b-9 signals limit complement-inducedinjury, or promote recovery of cells. In addition to C5b-9, complement cleavage products, such as C5a and C1q,can injure kidney cells. Thus, the complement system contributes to various kidney pathologies by causingcellular damage in both an inflammation-dependent and inflammation-independent manner.Semin Nephrol 33:586-601 C 2013 Elsevier Inc. All rights reserved.Keywords: C5b-9 membrane attack complex, cytoskeleton, glomerulonephritis, phospholipases, proteinkinases

The complement system is composed of threepathways, namely the classic (antibody-acti-vated), alternative, and mannose-binding lectin

pathways, which contain more than 30 proteinsinvolved in activation and regulation.1–3 The functionsof the complement system are numerous, and althoughthis review focuses on complement-mediated cellularinjury, the complement system acts as a rapid andefficient immune surveillance system, eliminates cel-lular debris and infectious microbes, contributes sub-stantially to homeostasis, and even may be involved inthe repair and regeneration of damaged tissues.2,3

Complement-mediated cellular injury has been impli-cated in various diseases including glomerulonephritis,sepsis, lupus, rheumatoid arthritis, myocardial infarc-tion, multiple sclerosis, myasthenia gravis, organ trans-plant rejection, and, more recently, osteoarthritis andage-related macular degeneration.2–5 In early studiesof complement in experimental glomerulonephritis,the focus was on antibody-dependent complement

see front matterier Inc. All rights reserved.rg/10.1016/j.semnephrol.2013.08.009

port: Supported by research grants from the Cana-tes of Health Research (MOP-53264 to A.V.C., andto T.T.), and the Kidney Foundation of Canada (A.V.

.). Andrey Cybulsky was supported by the CatherineHakim Chair.

terest statement: none.

f Medicine, McGill University Health Centre, McGillMontreal, Quebec, Canada.

int requests to Tomoko Takano, MD, PhD, Division ofMcGill University, 3775 University St, Rm 236, Mon-c, Canada H3A2B4. E-mail: [email protected]

Se

activation as a mechanism for recruitment of inflam-matory leukocytes. With the appreciation that the C5b-9 membrane attack complex can injure cells in theabsence of inflammation, a role for the terminal com-plement pathway in inducing podocyte injury andproteinuria was shown in rat passive Heymann neph-ritis (PHN), an experimental model of human mem-branous nephropathy (see Table 1 for a list of animalmodels).6–8 Consistent with induction of proteinuria,in PHN the C5b-9 complex has been localized in imm-une deposits, and in plasma membranes on the soles ofthe foot processes of visceral glomerular epithelialcells (GECs, commonly known as podocytes). C5b-9also has been shown to induce podocyte injury inexperimental anti–glomerular basement membrane(GBM) nephritis, and is the key mediator of mesangialinjury in the anti-Thy1 model of human mesangial-proliferative glomerulonephritis.4,9 Other humankidney diseases or experimental models in whichcomplement is activated and there is C5b-9 assemblyand injury include acute kidney injury (AKI), tubu-lointerstitial disease, and atypical hemolytic uremicsyndrome (aHUS).2–5,9 This article describes themechanisms by which complement induces cellularinjury and its role in the pathogenesis of renal disease.The main focus is on signaling pathways activatedby the terminal pathway, that is, membrane attackcomplex, C5b-9, in kidney glomerular and tubularcells. Other components of the complement system,such as C5a and C1q, are discussed briefly in thecontext of renal cell injury. The effects of complementon non–kidney cell injury also are presented whereappropriate to provide additional insights into relevantmechanisms.

minars in Nephrology, Vol 33, No 6, November 2013, pp 586–601

dx.doi.org/10.1016/j.semnephrol.2013.08.009



mailto:[email protected]

Table 1. Animal Models of Complement-Induced Injury and Corresponding Human Diseases

Animal Model Corresponding Human Disease Site of Injury Complement Component

PHN Membranous nephropathy Visceral GEC (podocyte) C5b-9Chronic BSA overload Proteinuric glomerulopathy Glomerular and tubular cells C5b-9, C3Anti-Thy1 nephritis Mesangial-proliferative glomerulonephritis Mesangial cells C5b-9, C5a, C1qAnti-GBM nephritis Anti-GBM nephritis Visceral GEC and GBM C5b-9, C5a, C1qThrombotic microangiopathy(anti–glomerular endothelialcell antibody nephritis)

HUS Glomerular endothelial cells C5a and/or C5b-9

Ischemia-reperfusion Acute tubular nephrosis Tubular epithelial cells C5b-9, C5aCisplatin-induced AKI Cisplatin nephrotoxicity Tubular epithelial cells C5a, C5a receptor

Abbreviation: BSA, bovine serum albumin.

Complement-mediated cellular injury 587

C5b-9–INDUCED RESPONSES IN CELLS

The assembly of C5b-9 in the plasma membrane of a cellresults in the formation of transmembrane channels orrearrangement of membrane lipids, with loss of mem-brane integrity. A single C5b-9 complex leads to lysisof an erythrocyte, whereas lysis of nucleated cells requi-res multiple C5b-9 lesions, and lower doses of C5b-9induce sublethal (sublytic) injury in nucleated cells.1,10–12

In vivo, there are relatively few examples of comple-ment lysis,2 and complement-mediated cellular injurygenerally is associated with sublytic amounts of com-plement. In addition to disrupting the plasma mem-brane, sublytic doses of C5b-9 lead to cell injurythrough the activation of specific signaling pathways(Table 2).1,10–12 Such pathways have been character-ized in numerous cell lines, as well as in in vivodisease models. Pathways activated by C5b-9 in GECsinclude protein kinases, phospholipases, reactiveoxygen species (ROS), transcription factors, growthfactors, proteinases, stress pathways, and others. Ulti-mately, these signals may impact on metabolic path-ways, and the structure or function of lipids and thecytoskeleton, filtration slit diaphragm, or other cellularcompartments.

Although various signals of C5b-9 injure nucleatedcells, it should be noted that other signals are activated inparallel to limit/restrict the complement-induced injury,or to promote recovery. Indeed, nucleated cells are equi-pped with several mechanisms that support resistance tocomplement-dependent cytotoxicity.1,12 Thus, multiplepathways may be activated simultaneously by sublyticC5b-9 in nucleated cells, and there is likely an equilibriumamong pathways that lead to cellular injury with thosethat are cytoprotective.

ACTIVATION OF PROTEIN KINASES

The activation of a signaling pathway after assembly ofsublytic C5b-9 in all probability starts at the plasmamembrane, where C5b-9 is assembled. Given that C5b-9forms a membrane channel or a leaky patch, a C5b-9–

induced increase in cytosolic free Ca2þ concentrationresulting from a calcium influx has been reported invarious cell lines, including GECs.1,6,7,12 Sublytic C5b-9also can induce Ca2þ release from intracellular storagesites. Together, such changes in intracellular Ca2þ con-centration can lead to the activation of protein kinases,among which is protein kinase C.1,6,12 At the plasmamembrane of lymphoblastoid B cells, C5b-9 assemblywas shown to couple to heterotrimeric G proteins, whichthen can activate downstream effectors.1 Sublytic C5b-9induced transactivation of receptor tyrosine kinases at theplasma membrane, in cultured GECs and in PHN.6,7 Thisresulted in activation of the Ras–extracellular signal-regulated kinase (ERK) pathway and phospholipase C-γ1. Transactivated receptor tyrosine kinases potentiallyserve as scaffolds for assembly and/or activation ofproteins, which then lead to activation of downstreameffector pathways, either independently or in conjunctionwith the increased cytosolic Ca2þ concentration.Complement-mediated activation of ERK was blockedby expression of a dominant-inhibitory mutant of Ras or aconstitutively active RhoA mutant, or disassembly ofF-actin, indicating that the mechanisms involve the Raspathway and are facilitated by an intact actin cytoskele-ton.6,7 The role of ERK in complement-induced injurymay be cytoprotective (see later).

Besides ERK, sublytic C5b-9 can activate the p38kinase pathway in GECs.6,7,13 The functional role ofp38 appears rather complex. In GEC culture, the p38pathway reduced complement-mediated cytotoxicity,and the cytoprotective effect involved heat shockprotein-27. p38 activity was increased in glomeruli ofrats with PHN, and treatment of these rats with a p38inhibitor exacerbated proteinuria. In keeping with theseexperimental data, enhanced phosphorylation of p38 inpodocytes was reported in biopsy specimens of humanmembranous nephropathy.7,14

Apoptosis signal-regulating kinase-1 (ASK1) is amitogen-activated protein kinase kinase kinase, whichis regulated by ROS, and can stimulate the p38 andc-Jun N-terminal kinase (JNK) pathways. ASK1 wasactivated in glomeruli of rats with PHN, and exposure of

Table 2. Signaling Pathways Activated by the C5b-9 Membrane Attack Complex

Pathways Examples and Effects

Intracellular Ca2þ Ca2þ influx and Ca2þ release from intracellular storage sitesProtein kinases Activation of protein kinase C (PKC), receptor tyrosine kinase (RTK), Ras-ERK, JNK, p38, and ASK1 (PHN)Phospholipases Activation of phospholipase C (PLC)-γ1, cPLA2, and iPLA2-γ (phosphorylation) and AA releaseProstanoids Up-regulation of cyclooxygenase (COX)-2 (cultured GEC and PHN glomeruli) and COX-1 (PHN

glomeruli), production of prostanoidsROS Superoxide production via NADPH oxidase and lipid peroxidation (PHN)

ROS production via xanthine oxidase pathway (PHN)Generation of hydrogen peroxide by cytochrome P450 family of hemeprotein monooxygenases (culturedGEC)

Growth factors Up-regulation of platelet-derived growth factor B-chain, HB-EGF (PHN), and Ret (PHN and culturedmouse podocytes)

Increase of p21 and p27 CDK inhibitors, and decrease of CDK2 activityDecrease of p57, and increase of Cdc2; cyclins B1, B2, and D1; and phosphorylated histone-3

Transcription factors DNAdamage

Activation of NF-κB (cultured GEC and in vivo)Production of interleukin-8 and monocyte chemoattractant protein-1Increase of p21, p53, GADD45, and checkpoint kinase-1 and kinase-2 (cultured GEC and PHN)

Endocytosis andectocytosis

Endocytosis (podocyte)Ectocytosis in membrane vesicles (urinary space)

ER stress Damage of ER membrane and unfolded protein response (UPR) inductionUp-regulation of ER chaperones (Bip and GRP94), PERK stimulation, eukaryotic translation initiationfactor-2α subunit phosphorylation, and reduction of protein synthesis

Ubiquitin-proteasomesystem

Polyubiquitination of glomerular proteins (PHN)Up-regulation of ubiquitin proteasome system (cultured GEC)

Podocyte cytoskeleton Disassembly of F-actin filaments and focal adhesion complexesIncrease of RhoA and decrease of Rac1 and Cdc42 activities (cultured GEC)Foot process effacement by induction of active RhoA in podocytes (in mice)TRPC6 up-regulation (cultured GEC)

Filtration slit diaphragm Decrease of nephrin mRNA and protein (PHN)Dissociation of nephrin from actin cytoskeleton and loss of slit diaphragm integrityAlteration of podocin location and nephrin dissociation from podocin

Extracellular matrix Increase of α3, α4, and α5 type IV collagen mRNAs and de novo synthesis of type I collagen (PHN)Up-regulation of matrix metalloprotease 9, TGF-β2, TGF-β3, and TGF-β type I and II receptors (PHN)Transient increase of collagen IV and fibronectin mRNAs (cultured mesangial cells)

Cell cycle Increased DNA synthesis without cell proliferation (podocyte)Activation of cell cycle and cell proliferation via PI3K (cultured mesangial cells)

Anti-apoptosis PI3K/Akt activation, Bad phosphorylation, and dissociation of the Bad/Bcl-XL complexUp-regulation of caspase-8 inhibitor and cFLIPL, and down-regulation of FasL

Pro-apoptosis DNA damage via apoptosis regulating proteins (podocytes)Up-regulation of ATF3 (mesangial cells)Interferon regulatory factor 1–dependent caspase-8 activation (mesangial cells)

T. Takano, H. Elimam, and A.V. Cybulsky588

cultured GECs to sublytic C5b-9 stimulated ASK1activity, in part, via the reduced form of nicotinamide-adenine dinucleotide phosphate (NADPH) oxidase andROS.7,15 In GECs that overexpress ASK1, complement-induced p38 and JNK activation were amplified. Thelevel of ASK1 expression determined the functionaleffect of p38 activation, that is, when ASK1 was over-expressed, p38 activation was amplified, and C5b-9assembly led to GEC injury via ASK1 and p38.

PHOSPHOLIPASE A2 AND PROSTANOIDS

Assembly of C5b-9 results in phospholipid hydrolysisand the production of arachidonic acid (AA) andprostanoids in various cells, including GECs and

tubular epithelial cells.1,12,16 These effects are mediatedby phospholipase C, phospholipase A2 (PLA2), andothers.6–8 In GEC, cytosolic PLA2 (cPLA2) plays animportant role in AA release, and sublytic C5b-9stimulated cPLA2 activation, which was regulated bychanges in cytosolic Ca2þ concentration and by phos-phorylation. Activation of cPLA2 and release of AAwas compartmentalized to specific organelles, in partic-ular the endoplasmic reticulum (ER). cPLA2 mediatesC5b-9–dependent GEC injury (eg, cPLA2 provides AAfor metabolism to injurious prostanoids), and it mayinduce membrane damage via phospholipid hydrolysis.cPLA2 also modulates stress responses, which confercytoprotection (see later). Sublytic C5b-9–dependentAA release in GECs also is mediated by calcium-


independent PLA2-γ (iPLA2-γ), via an ERK or p38pathway that involves iPLA2-γ phosphorylation.17,18

Activation of iPLA2-γ attenuated complement-inducedcytotoxicity. The mechanism of protection requiresfurther elucidation, although it has been suggested thatiPLA2-γ may facilitate metabolism of damaged/oxidizedphospholipid acyl chains to repair membrane lipids.

The AA released by PLA2 enzymes is metabolizedto prostanoids by cyclooxygenase (COX) enzymes. Inresting GECs in culture and glomeruli of normal rats,there is primarily expression of COX-1. In culture,sublytic C5b-9 up-regulated COX-2 expression,whereas both isoforms were up-regulated in glomeruliof rats with PHN.6,7 Up-regulation of COX-2 wasdependent on activation of protein kinase C and JNK(see later). Both COX isoforms were involved inenhanced prostanoid generation. Nonselective COXinhibition and selective inhibition of COX-2 reducedglomerular prostanoid generation and proteinuria inPHN; however, inhibition of both COX isoforms wasrequired to achieve a maximum antiproteinuriceffect.6,7,19 The antiproteinuric effects of COX inhib-itors were independent of changes in glomerularfiltration. Reduction of proteinuria in PHN also wasinduced by a thromboxane synthase inhibitor.6,7 A fishoil diet, which shifts production of dienoic prostanoidsto inactive trienoic metabolites, decreased the produc-tion of glomerular thromboxane A2 and had both aprotective effect on the development of proteinuria inPHN, as well as a therapeutic effect after the onset ofproteinuria.6,7,20 Thus, a number of distinct experimen-tal approaches support a role for C5b-9–dependentprostanoid production in exacerbating proteinuria inPHN. In regard to the mechanism, prostanoids mayincrease glomerular transcapillary pressure, such thatproteinuria is increased on a hemodynamic basis, and/or production of prostanoids may exacerbatecomplement-induced GEC injury directly.6,8

PRODUCTION OF ROS

Production of ROS after assembly of sublytic C5b-9 hasbeen described in leukocytes, mesangial cells, GECs,and proximal tubular epithelial cells.6–8,12,13,21,22 Amongthe pathways for generating ROS is a multicomponentenzyme complex, the NADPH oxidase, which in thepresence of NADPH reduces molecular oxygen to thesuperoxide anion. Superoxide is metabolized further toother ROS. GECs in culture and in vivo expresscomponents of the NADPH oxidase, and incubation ofcultured GECs with complement stimulated superoxideproduction via the NADPH oxidase.6,7 ROS also areproduced in PHN, and their generation is associated withlipid peroxidation and modification of GEC membraneproteins and GBM components by malondialdehydeadducts.6,7,23 Inhibition of ROS production and lipid

peroxidation reduced proteinuria in PHN, implying thatcomplement-induced ROS exacerbate proteinuria bydamaging the glomerular capillary wall. The xanthineoxidase pathway also has been implicated in ROSgeneration in PHN, and blockade of xanthine oxidasereduced proteinuria.24 A recent study addressed the cyto-chrome P450 family of hemeprotein monooxygenases,which generate superoxide and hydrogen peroxide duringthe course of NADPH-dependent electron transfer.25

Overexpression of cytochrome P450 2B1 (CYP2B1) incultured GECs enhanced the complement-mediated gen-eration of hydrogen peroxide, disruption of actin fila-ments, and cytotoxicity. Conversely, these parameterswere reduced with CYP2B1 gene silencing. PHN wasassociated with generation of glomerular cerium–hydro-gen peroxide reaction products, malondialdehyde add-ucts, proteinuria, and reduced expression of nephrin, thekey component of the filtration slit diaphragm.26 Inhib-ition of CYP2B1 with piperine or cimetidine reducedthe cerium-hydrogen peroxide reaction products, malon-dialdehyde adducts, and proteinuria, and attenuated thereduction in nephrin expression.

Although robust production of ROS induces glomer-ular damage, low concentrations of ROS may function assignaling molecules. For example, in GECs in culture andin vivo, complement-induced superoxide productionactivated JNK.6,7,27 The process was dependent oncPLA2 and AA release; the latter stimulated JNK activityvia the NADPH oxidase and production of ROS. JNKactivation was associated with cytoprotection. Thus, inGECs exposed to C5b-9, ROS may both contribute tocell injury and facilitate transduction of protective signals.

Another example of ROS production resulting fromC5b-9 assembly is the anti-Thy1 model of nephritis.Glomeruli from rats with anti-Thy1 nephritis showedenhanced production of hydrogen peroxide, superoxide,and hydroxyl radicals. In parallel, there were increases inglomerular Nicotinamide adenine dinucleotide (NADH)-dependent and NADPH-dependent oxidative activities,as well as a reduction in antioxidative activities, poten-tially amplifying ROS-mediated glomerular damage.28 Inanother study of anti-Thy1 nephritis, treatment of ratswith the antioxidant α-lipoic acid reduced glomerularROS production. In addition, α-lipoic acid reducedglomerular cell proliferation, reversed an increase inthe expression of glomerular transforming growthfactor-β (TGF-β), and prevented transformation ofmesangial cells into myofibroblasts.29

GROWTH FACTORS, TRANSCRIPTION FACTORS,AND DNA DAMAGE

The actions of complement may include stimulation ofcell proliferation and/or apoptosis, and may be asso-ciated with induction of growth factors, growth factor


receptors, and proteins that regulate the cell cycle.6–8

C5b-9 stimulated production of TGF-β in mesangialcells.30 In PHN, platelet-derived growth factor B-chainwas up-regulated before the onset of proteinuria.Although platelet-derived growth factor C-chainexpression was not shown in PHN, it was prominentin podocytes in human membranous nephropathy.6–8,31

Heparin-binding epidermal growth factor-like factor(HB-EGF) was induced in PHN, before the onset ofproteinuria.6,7 At this early time point, HB-EGF waslocalized in the cytoplasm of GECs; as the diseaseprogressed, HB-EGF showed a nodular pattern withinGECs and along the GBM, consistent with secretionand binding to the GBM. Ret (glial cell-derivedneurotrophic factor receptor tyrosine kinase) was up-regulated in podocytes in PHN, and in cultured mousepodocytes after stimulation with sublytic C5b-9.6,7 Therole of these growth factors in cell injury will requirefurther study.

Podocytes are believed to be terminally differenti-ated cells with low proliferative capacity in vivo bothunder normal circumstances and after injury.32 InPHN, there is no evidence for podocyte proliferationin response to C5b-9–induced injury, although anincrease in proliferating cell nuclear antigen expressionhas been shown.6,7 Changes in cyclins, cyclin-dependent kinases (CDKs), and CDK inhibitors alsohave been shown in PHN.6–8 The p21 and p27 CDKinhibitors were increased, in association with decreasedCDK2 activity. p57 was decreased, whereas Cdc2;cyclins B1, B2, and D1; and phosphorylated histone-3were increased. The low proliferative capacity of GECsin response to injury may be, at least in part, caused byan increase in the expression of specific CDK inhib-itors. Administration of basic fibroblast growth factorto rats with PHN attenuated increases in p21, andstimulated GEC mitosis and ploidy.6–8

Nuclear factor-κB (NF-κB) and NF-κB–dependentgene transcription was activated by complement incultured GECs and in vivo.6,7 NF-κB may, at least inpart, contribute to podocyte injury and proteinuria inPHN. In autologous phase PHN with modest proteinu-ria, administration of the NF-κB inhibitor, pyrrolidonedithiocarbamate, reduced proteinuria; however, thedrug had no effect in PHN with high levels of urineprotein excretion.33 Down-regulation of nephrin expres-sion up-regulated NF-κB in podocytes.7,34 Althoughnephrin down-regulation has been reported in PHN(discussed further later), it remains to be determined ifthis specific nephrin effect on the NF-κB pathway ispresent in membranous nephropathy. In human umbil-ical vein endothelial cells, formation of sublytic C5b-9promoted an increase in NF-κB.35 C5b-9 also stimu-lated production of interleukin-8 and monocyte chemo-attractant protein-1, and preincubation of these cellswith pyrrolidine dithiocarbamate prevented the C5b-9–

induced increases in interleukin-8 and monocyte che-moattractant protein-1. Thus, sublytic C5b-9 can pro-mote proinflammatory endothelial cell activation viaNF-κB.

Conversely, the NF-κB pathway can play a cyto-protective role in complement-mediated injury.36

Enhanced complement-dependent cytotoxicity wasshown in mouse embryonic fibroblasts lacking thep65 subunit of NF-κB or the IκB kinases α or β, andin fibroblasts pretreated with an IκB kinase inhibitor.Reconstitution of p65 into p65-null cells and over-expression of p65 in wild-type cells decreased theirsensitivity to cytotoxicity. This effect of the NF-κBpathway appeared to be mediated via JNK.

In cultured GECs and in PHN, sublytic C5b-9induced DNA damage. These changes occurred inassociation with increases in p21, the tumor-suppressorgene p53, growth-arrest DNA damage-45 gene, andcheckpoint kinase-1 and kinase-2.6,7,37 Inhibition of theERK pathway attenuated the increase in p21 andgrowth-arrest DNA damage-45 gene, and augmentedC5b-9–induced DNA damage. Induction of DNA dam-age by C5b-9 may, in part, block the ability of GECs toproliferate in response to injury.

C5b-9 ENDOCYTOSIS AND ECTOCYTOSIS

As noted earlier, nucleated cells possess variousresistance mechanisms against complement-dependentinjury. Among these is the elimination of C5b-9complexes from the plasma membrane by endocytosisand/or release of membrane vesicles (ectocytosis ofC5b-9).10,12 Elimination of C5b-9 is, in part, linked toCa2þ influx through the C5b-9 pores; chelation ofintracellular Ca2þ inhibits removal of C5b-9 com-plexes and results in increased cell death.38 The effectsof Ca2þ appear to be mediated through various proteinkinases, phospholipases, and cytoskeletal elements. InK562 cells, shedding of C5b-9 from the plasmamembrane began within minutes after C5b-9 assembly.Concomitantly, C5b-9 entered the cells, accumulatedin a perinuclear, late recycling compartment, and co-localized with endocytosed transferrin, whereas inhib-ition of protein kinase C reduced endocytosis of C5b-9.39 In the K562 cells, a key protein found to mediatethe membrane vesiculation of C5b-9 is mortalin/GRP75. This protein promoted shedding of membranevesicles containing C5b-9 and protected cells fromcomplement-mediated lysis.39 Endocytosis and exo-cytosis of C5b-9 also occurs in disease. In PHN, thereis morphologic evidence for C5b-9 endocytosis inpodocytes, as well as exocytosis in membrane vesi-cles in the urinary space, potentially representing amechanism to protect podocytes from complementattack.6,7,40

Figure 1. ER stress in the PHN experimental model of membra-nous nephropathy and the anti-Thy1 nephritis model of mesan-gial-proliferative glomerulonephritis. (A) The unfolded proteinresponse is induced in C5b-9–mediated podocyte injury. Glomer-uli were isolated from normal rats (control) and from rats with PHNon day 14, and lysates were immunoblotted with antibodies to BiP,GRP94, or phospho-eukaryotic translation initiation factor-2αsubunit (eIF2α). (B) Preconditioning of rats with a subnephrito-genic dose of Adriamycin or tunicamycin reduces proteinuria inPHN. Rats were untreated, or were injected with Adriamycin ortunicamycin, to up-regulate ER stress proteins. Four days later,rats were injected with nephritogenic antibody to induce PHN (day0). Urine protein excretion was measured on days 0, 7, 9, and 13.This research was published originally by Cybulsky et al43,104 inthe Journal of Biological Chemistry. © American Society forBiochemistry and Molecular Biology. (C) Preconditioning of ratswith a subnephritogenic dose of thapsigargin or tunicamycinreduced proteinuria in rats with anti-Thy1 nephritis. Figure pro-vided courtesy of Dr. Reiko Inagi (University of Tokyo).


ENDOPLASMIC RETICULUM INJURY AND ERSTRESS

ER stress is a pathologic state that results in accumu-lation of misfolded proteins in the ER.41,42 To rescuesuch misfolded proteins, the ER has quality controlmachinery in place, including the unfolded proteinresponse and ER-associated degradation. Complement-dependent ER stress has been described in GECs andmesangial cells. In GECs, sublytic C5b-9 induceddamage to the membrane of the ER, as well as theunfolded protein response, and these effects were, inpart, enhanced by the activation of cPLA2, presumablyowing to phospholipid hydrolysis at the membraneof the ER.7,41–43 Specifically, complement increasedexpression of ER chaperones (glucose regulated protein(GRP)-78 (BiP) and GRP94), and stimulated activationof PERK (protein kinase R-like ER kinase), phosphor-ylation of eukaryotic translation initiation factor-2αsubunit, and a reduction in protein synthesis. Reductionof BiP expression or PERK deletion exacerbatedcomplement-mediated GEC injury, confirming a func-tionally important, protective role for the unfoldedprotein response. In addition, preconditioning of cellswith other stimuli that can increase expression of ERchaperones (eg, puromycin aminonucleoside) limited theamount of injury during subsequent complement attack.

The unfolded protein response was induced in glo-meruli of proteinuric rats with PHN (Fig. 1A).7,41–44 Arecent study also showed simultaneous induction ofautophagy markers in PHN,44 and although ER stressmay be linked with autophagy, this link has not beenshown conclusively in PHN. Preconditioning of ratswith subnephritogenic doses of Adriamycin, or tunica-mycin, attenuated proteinuria in PHN.7,41–43 Thus,in vivo preconditioning to increase ER stress reducedC5b-9–mediated GEC injury (Fig. 1B). ER stresspreconditioning also reduced C5b-9–mediated mesan-gial cell injury in vivo (Fig. 1C).45 The ER chaperones,BiP and ORP150, activation of PERK, and phosphor-ylation of eukaryotic translation initiation factor-2αsubunit were enhanced in glomeruli of rats with anti-Thy1 nephritis corresponding to the stage of mesangialhypercellularity.45 Tunicamycin or thapsigargin givento rats before induction of anti-Thy1 nephritis did notaffect renal pathology or function, although the com-pounds increased expression of BiP and ORP150.After induction of nephritis, glomeruli displayingmicroaneurysms were reduced markedly in the ratspreconditioned with tunicamycin or thapsigargin com-pared with nonpretreated rats, and mesangial prolifer-ation, as well as proteinuria, were ameliorated. Thesestudies provide a rationale for developing nontoxicpreconditioning methods to induce ER stress in vivo,which eventually may have applications to therapy ofcomplement-mediated glomerular injury.

UBIQUITIN-PROTEASOME SYSTEM

Two recent studies suggested that podocyte proteinmisfolding is a consequence of C5b-9–induced injuryin membranous nephropathy. The studies showedincreased polyubiquitination of glomerular proteins inPHN (Fig. 2A).46,47 In addition, increased ubiquitincontent and aggregation in podocytes was shown in

Figure 2. Ubiquitin is up-regulated in podocytes in experimental and human membranous nephropathy (MN).(A) Confocal micrographs of immunofluorescent staining against ubiquitin (green) and collagen type IV (red) incontrol and PHN rats (day 14). Arrows indicate podocytes. (B) Immunohistochemical staining of ubiquitin (red) inhuman MN renal biopsy specimens. Short scale bars, 200� magnification; long scale bars, 1000� magnification.p, podocyte. (C) Control and PHN rats were treated with vehicle or with the proteasome inhibitor MG132. Urinealbumin concentrations (in milligrams per deciliter) were measured on day 14, and were normalized against urinecreatinine concentrations (in grams per liter). *P o .005 versus control rats. **P o .01 versus PHN rats. Thisresearch was published originally by Meyer-Schwesinger et al46 in the American Journal of Pathology. © AmericanSociety for Investigative Pathology.


human membranous nephropathy (Fig. 2B).48 Admin-istration of a proteasome inhibitor to rats with PHNexacerbated proteinuria, implying that degradation ofmisfolded proteins by the ubiquitin-proteasome systemis a protective response in podocytes undergoing com-plement attack (Fig. 2C).46,47 Ubiquitin C-terminalhydrolase-L1, an enzyme with both ubiquitin ligaseand deubiquitinating activities, was implicated inaltering glomerular ubiquitination and development ofproteinuria in PHN,46 however, the precise mechanismof action of ubiquitin C-terminal hydrolase-L1 willrequire further study.

Global function of the ubiquitin-proteasome systemwas enhanced in cultured GECs undergoing comple-ment attack despite a modest reduction in proteasomal

enzymatic activity.47 This increase in the ubiquitin-proteasome system function may have been owing tocomplement-stimulated conjugation of ubiquitin toproteins, although changes in deubiquitination cannotbe excluded. Exposure to complement, however,impaired ER-associated degradation function in GECs;possibly, misfolded proteins in the ER may have beencompeting for ER resident enzymes or chaperones, orthere was disrupted protein retrotranslocation to theproteasome.47 Another study showed that PHN isassociated with increased expression of genes encodingubiquitin-conjugating enzymes and de-ubiquitinatingenzymes.49 The results of these various studies implythat complement may up-regulate the activity and/orexpression of ubiquitin-proteasome system components


as a cytoprotective mechanism to alleviate proteinmisfolding in podocytes.

ALTERATIONS TO THE PODOCYTE CYTOSKELETON

The structure of the podocyte foot processes is sup-ported by F-actin filaments. In glomerulopathies (eg,membranous nephropathy), there is condensation ofactin filaments at the base of effaced podocyte footprocesses together with disruption of slit diaphragmsand filtration slits. In cultured GECs, it has been shownthat sublytic C5b-9 can disassemble F-actin filamentsand focal adhesion complexes.6–8,13,50 The process wasassociated with adenosine triphosphate depletion anddephosphorylation of paxillin, but was not preventedby tyrosine phosphatase inhibition. Proteolysis of focaladhesion or cytoskeletal proteins did not occur, whichfavors dynamic disassembly and reassembly, perhapsmediated by Rho guanosine triphosphatases (GTPases).Moreover, α3β1 integrins, which are among the mole-cules mediating attachment of GECs to the underlyingGBM, remained intact, in keeping with the normaldistribution of β1 integrins on the basal aspect ofeffaced foot processes in kidney biopsy specimens ofhuman membranous nephropathy.6

Changes in the actin cytoskeleton may be, at least inpart, regulated by Rho GTPases. In cultured GECs,sublytic C5b-9 increased RhoA activity and decreasedactivities of Rac1 and Cdc42.7,51 Analogous changes inthese activities were observed in glomeruli of rats withPHN. Cultured GECs expressing constitutively activeRhoA showed a smaller and round contour, and prom-inent cortical F-actin, whereas those expressing consti-tutively active Rac1 underwent morphologic changesresembling process formation. Expression of constitu-tively active RhoA attenuated complement-mediatedcytotoxicity, whereas cytotoxicity was augmented by adominant-negative RhoA. Thus, complement can alterthe balance of RhoA, Rac1, and Cdc42 activities inGECs, and the activity of Rac1 may contribute toprocess formation, while activation of RhoA in thesetting of complement attack may contribute to footprocess effacement. Indeed, induction of glomerularRhoA activity in mice results in foot process effacementand proteinuria.52 In keeping with the effect of RhoA,expression of a constitutively active Ras protected GECsfrom complement cytotoxicity.7,53 This protective effectof Ras was dependent on remodeling of the actin cyto-skeleton, and was associated with a reduction in Racactivity, thereby altering the equilibrium of Rho GTPaseactivity.

Another potential mediator of the actin cytoskeletonis transient receptor potential channel (TRPC)-6 (anonselective cation channel). The TRPC6 channel isfunctionally connected to the podocyte actin cytoske-leton, and mutations are associated with hereditary

focal segmental glomerulosclerosis. Enhanced expres-sion of wild-type TRPC6 was found in human protei-nuric kidney diseases, including membranousnephropathy, whereas in cultured GECs, complementup-regulated TRPC6 protein.7,54 Thus, the effects ofcomplement on the cytoskeleton could involve TRPC6.

CHANGES TO THE FILTRATION SLIT DIAPHRAGM

As noted earlier, nephrin is a key component of thefiltration slit diaphragm, and is crucial for maintainingglomerular permselectivity.26 A number of studieshave shown effects of complement on nephrin expres-sion. Nephrin messenger RNA (mRNA) and/or proteindecreased in PHN.6,7 In rats with PHN before the onsetof proteinuria, there was complement-dependent dis-ruption or dislocation of slit diaphragms and formationof occluding-type junctions, as well as a reduction innephrin expression, an alteration in the distribution ofnephrin, and a reduction in the fraction of nephrin thatwas bound to actin.6,7 It was concluded that the onsetof proteinuria in PHN is caused by complement-dependent dissociation of nephrin from the actincytoskeleton and loss of slit diaphragm integrity.Nephrin expression was reduced and its distributionwas altered in human membranous nephropathy.6,7

Podocin, a protein localized at the insertion of theslit diaphragm, associates with the nephrin cytoplasmictail and provides a scaffolding function.26 In PHN, thelocation of podocin was altered before the onset ofproteinuria, and nephrin was dissociated from podo-cin.7,55 The actin-associated portion of nephrin wasreduced, and a fraction of nephrin (but not podocin)appeared to be shed into the urine. By analogy, nephrinwas shed from cultured cells injured by complement,and this finding may, at least in part, explain why totalnephrin is reduced in experimental and human mem-branous nephropathy. Shedding of intact podocytesinto the urine, however, also may account for loss ofnephrin in PHN.6,7

The function and localization of nephrin, as wellinteractions of nephrin with various binding partners,appear to be regulated post-translationally (eg, bytyrosine phosphorylation).26 The cytoplasmic domainof nephrin contains six conserved tyrosine residues,and several of these residues can be phosphorylated bythe Src family tyrosine kinases. Tyrosine phosphor-ylation of nephrin appears to augment its interactionwith podocin.7,56 Src activation was reported in PHNalong with an increase in tyrosine phosphorylation ofnephrin, which would imply an enhanced interaction ofnephrin with podocin. In this study, there were nochanges in nephrin expression. An increasing numberof nephrin binding partners are being discovered, andthe effects of C5b-9 on these various interactions andslit diaphragm integrity will require further study.


EXTRACELLULAR MATRIX TURNOVER

Exposure of cultured mesangial cells to sublytic C5b-9caused a transient increase in type IV collagen andfibronectin mRNAs.57 By analogy, glomerular deposi-tion of type IV collagen was increased in anti-Thy1nephritis, and the increase was attenuated inC6-deficient rats.58 Studies in PHN have suggestedthat expansion of the GBM is, at least in part, causedby altered synthesis of extracellular matrix componentsby the complement-injured podocytes.6–8 Changes inGBM components or enzymes involved in extracellularmatrix remodeling have been reported in active HN orPHN. In active HN, there was altered synthesis oflaminin, fibronectin, entactin, and heparan sulfate pro-teoglycans, whereas mRNA levels of α3, α4, and α5type IV collagen chains were increased in glomeruli ofrats with PHN, and there was de novo synthesis of typeI collagen.6–8 Expression of agrin was decreased inactive HN, and this reduction correlated inversely withthe amount of proteinuria.6,7 Matrix metalloproteinase-9 (an enzyme with proteolytic activity against type IVcollagen) was up-regulated in PHN in a complement-dependent manner, and administration of antihepara-nase antibody reduced proteinuria.6–8 TGF-β familymembers may mediate the metabolism of extracellularmatrix components. Expression of TGF-β2 and TGF-β3 (but not TGF-β1), as well as TGF-β type I and typeII receptors, were increased in PHN in a complement-dependent manner.6–8,59 In human membranous nephr-opathy, expansion of the GBM is the result ofincreased deposition of type IV collagen isoforms α3,α4, and α5, as well as laminin, nidogen, and fibronec-tin, with subsequent accumulation of collagen IV iso-forms α1 and α2.6,7 In summary, assembly of C5b-9 inpodocytes may promote formation of an expanded anddisorganized GBM, whereas up-regulated secretion ofproteolytic enzymes may alter breakdown or turnoverof glomerular capillary wall components, and contrib-ute to proteinuria.

C5b-9 also has been implicated in tubulointerstitialextracellular matrix changes.8,60 C5b-9 up-regulatedcollagen gene expression in cultured tubular epithelialcells.61 In rats with Adriamycin nephropathy, C5b-9deposition was present on the apical surface of prox-imal tubular epithelial cells and in the peritubularregion. There was peritubular myofibroblast accumu-lation, associated with increased interstitial extracellularmatrix deposition and tubulointerstitial injury. Thesechanges were reduced substantially in C6-deficient rats,implicating a role for C5b-9. Thus, the study60 sug-gested that glomerular filtration of complement compo-nents and the intratubular formation of C5b-9 is apromoter of peritubular myofibroblast accumulation inexperimental focal segmental glomerulosclerosis. Anal-ogous results have been reported in a model of chronic

bovine serum albumin overload-induced proteinuria.62

Mice with C3 deficiency were protected significantlyagainst the protein overload–induced renal interstitialinflammatory response and fibrosis, and these mice hadless severe podocyte injury and less proteinuria com-pared with wild-type animals. These interstitial changeswere dependent on C3 that was ultrafiltered through theglomerular capillary wall, rather than locally synthe-sized C3. Finally, there was evidence for complementactivation on the luminal surface of the tubules inkidney biopsy specimen from human patients withproteinuric kidney disease.63

ACTIVATION OF THE CELL CYCLE

Assembly of large amounts of C5b-9 complexes on theplasma membrane inevitably leads to cell lysis anddeath. However, sublytic C5b-9 has a range of effectson cell survival (discussed later), and the cell cycle. Forexample, sublytic C5b-9 has been shown to promoteproliferation in various cell types including Schwanncells, aortic smooth muscle cells, and aortic endothelialcells.1 As discussed earlier, in podocytes, a number ofreports showed modulation of various CDKs and CDKinhibitors by complement,7 and that C5b-9 increasedDNA synthesis, but C5b-9 did not induce cell prolifer-ation.64 Similar dissociation between DNA synthesisand cell proliferation after stimulation by C5b-9 wasobserved in oligodendrocytes, terminally differentiatedcells similar to podocytes.65 Low proliferative capacityof the terminally differentiated cells may account forsuch dissociation; it was suggested that the failure tosustain high-level expression of the cdc2/cyclin Bcomplex required for mitosis or a failure to get thiscomplex into the nucleus may account for the failure toprogress to mitosis.64 In contrast, a recent study showedthat sublytic C5b-9 promotes cell proliferation incultured rat mesangial cells (see later).66 C5b-9–medi-ated cell-cycle activation also was reported inaortic endothelial cells, and inactivation of the fork-head transcription factor, FOXO1, downstream of Aktwas implicated.67 Thus, sublytic C5b-9 can activate thecell cycle and promote proliferation in certain cellsincluding mesangial cells, but in terminally differenti-ated cells such as podocytes and oligodendrocytes,cell-cycle activation does not necessarily result inproliferation.

PRO-APOPTOTIC AND ANTI-APOPTOTIC EFFECTSOF C5b-9

The anti-apoptotic effect of C5b-9 has been studiedextensively in oligodendrocytes. In these cells, sublyticC5b-9 induced strong phosphatidylinositol 3-kinase(PI3K)/Akt activation, phosphorylation/inactivation ofthe pro-apoptotic protein, Bad, and dissociation of the


Bad/Bcl-XL complex, and this cascade of eventspromoted cell survival, likely by preventing mitochon-drial insertion of Bad.68 Other mechanisms used byC5b-9 to exert anti-apoptotic effects in these cellsinclude the up-regulation of the caspase-8 inhibitor,c-FLIPL, and down-regulation of the apoptotic ligand,Fas-ligand (FasL).69 On the other hand, the pro-apoptotic effect of C5b-9 is well documented in variouscells including glomerular mesangial cells70 (also seenext section) and glomerular endothelial cells.71 Inpodocytes, C5b-9 induced DNA damage, but character-istic apoptosis was not detected by the terminaldeoxynucleotidyl transferase–mediated deoxyuridinetriphosphate nick-end labeling or the DNA laddertechniques.37 Based on these mixed pro-apoptotic andanti-apoptotic effects of sublytic C5b-9, it was proposedthat even necrotic cell death induced by C5b-9 mayinvolve apoptosis-regulating proteins and may beviewed as programmed necrosis.72 For example, fibro-blasts from mice deficient in the pro-apoptotic proteinBid were shown to resist C5b-9–induced cell lysis.72

Thus, C5b-9 plays a spectrum of roles in necrotic andapoptotic cell death via apoptosis-regulating proteinsand also exerts anti-apoptotic effect in certain cell types.

COMPLEMENT AND MESANGIAL CELL APOPTOSISAND PROLIFERATION

In addition to podocyte injury, an important role forcomplement has been shown in mesangial cell injury.In the rat anti-Thy1 model of mesangial-proliferativenephritis, an acute peak of glomerular cell apoptosiswas observed 1 hour after injection of the antibody(first phase), followed by a much milder increment ofapoptosis between 6 and 24 hours after injection(second phase).73 The first phase of apoptosis wasabsent in C6-deficient rats, implying dependence onC5b-9.73 In contrast, the second phase of apoptosis wassimilar in control and C6-deficient rats, suggesting thatthis phase is C5b-9–independent.73 Thus, the evidencesupports the role of complement in acute mesangialloss in the anti-Thy1 nephritis model both via assemblyof C5b-9 and via other components (eg, C1q, see later).More recently, sublytic C5b-9 was shown to promoteapoptosis in mesangial cells by up-regulating activat-ing transcription factor 3.74,75 The same investigatorsalso reported a role for interferon regulatory factor1–dependent caspase-8 activation in C5b-9–inducedmesangial cell apoptosis.76 Complement-mediatedmesangial cell apoptosis is likely to contribute to thepathogenesis of diseases such as IgA nephropathy andlupus nephritis, in which mesangial cell damage andproliferation is the main feature. In addition, it wassuggested that mesangial cell apoptosis may play animportant role in regulating the recovery phase ofmesangial proliferative glomerulonephritis.73

In contrast to the pro-apoptotic action of C5b-9 inmesangial cells, the effect of C5b-9 on mesangialproliferation appears to be unresolved. By using anti-Thy1 antibody and human complement, it was shownthat sublytic C5b-9 inhibits mesangial proliferation,measured by 3H-thymidine.77 More recently, however,sublytic C5b-9 was shown to promote mesangial pro-liferation via PI3K.66 The latter study also used the anti-Thy1 antibody and human complement, but quantifiedcell proliferation by bromodeoxyuridine incorporationinto nuclear DNA.66 Further studies including cell-count quantification likely will provide a more definitiveanswer as to whether sublytic C5b-9 has pro-proliferativeor anti-proliferative actions in mesangial cells.

PRO-APOPTOTIC AND ANTI-APOPTOTIC EFFECTSOF C5A

C5a is a cleavage fragment of C5 generated by theaction of C5 convertase. C5a is biologically active, andit signals through a specific G-protein–coupled receptor;C5a classically acts as a potent anaphylactic and chemo-tactic factor and mediates inflammation. In addition,C5a has been implicated in apoptosis, with both anti-and pro-apoptotic actions. In a rat model of sepsis, C5ablockage markedly reduced thymocyte apoptosis whileinhibiting caspase-3, -6, and -9 activation, mitochondrialcytochrome C release, and Bcl-XL down-regulation,78

indicating that C5a is pro-apoptotic in this setting. Onthe other hand, anti-apoptotic, cell-protective actions ofC5a have been reported in various neural cells.C5-deficient mice were more susceptible to hippocam-pal excitotoxic lesions compared with C5-sufficientmice.79 In primary murine cortico-hippocampal neurons,C5a reduced glutamate neurotoxicity and this effectoccurred via the C5a receptor and ERK activation.80 Ananti-apoptotic effect of C5a also was reported inneutrophils, and was mediated by the PI3K-Akt path-way.81 Because neutrophils are important mediators ofinjury in certain forms of inflammatory glomeruloneph-ritis, C5a conceivably may exacerbate glomerularinflammation by protecting neutrophils from death. Ina proliferative glomerulonephritis model induced byinjecting mice with horse spleen apoferritin and lip-opolysaccharide, C5a-receptor knockout mice showed alower number of infiltrating interstitial cells and reducedtubular cell apoptosis.82 In this model, the proliferativeglomerular lesions did not, however, differ betweenknockout and control mice.82 Combined with the pre-vious observation that the C5a receptor is expressed inproximal tubular cells in the human kidney,83 C5a islikely to contribute to tubular damage by promotingtubular cell apoptosis via the C5a receptor in this model.

Mice deficient in the complement regulatory proteinfactor H spontaneously develop a chronic membrano-proliferative glomerulonephritis, which resembles human


dense deposit disease or C3 nephropathy.84 The factorH–deficient mice show thickening of the GBM andelectron dense deposits, albuminuria, glomerular neu-trophil infiltration, and accelerated mortality. In thesemice, additional gene deletion of C5 did not affect thechanges in GBM morphology, although C5 deletionpartially attenuated albuminuria, neutrophil infiltration,and mortality.85 Injection of anti-GBM antibody intoyoung factor H–deficient mice induced acute glomer-ular C3 deposition, albuminuria, and neutrophil infil-tration, all of which were absent in injected wild-typecontrol mice. Deletion of C5, but not C6, in the factorH–deficient mice ameliorated glomerular neutrophilinfiltration, suggesting that C5a, but not C5b-9, isresponsible for glomerular inflammation.85 Thus, C5alikely plays multifaceted roles in kidney pathology,acting as a chemoattractant/anti-apoptotic factor forinflammatory cells and as an apoptotic factor forkidney resident cells (eg, tubular cells).

SIGNALING OF APOPTOSIS BY C1q

C1q is a component of C1 together with C1r and C1s.In human beings, C1q nephropathy is a rare cause ofnephrotic syndrome, and histologically features mesan-gial proliferation, accompanied by electron densedeposits observed by electron microscopy and C1qdeposits by immunofluorescence staining. C1q wasshown to enhance apoptosis in cultured rat mesangialcells induced by a C1q-binding anti-Thy1 monoclonalantibody.86 C1q-induced apoptosis also has beenreported in human fibroblasts. p38 and downstreamtranscription factors such as ATF2, ETS domain tran-scription factor 1, and C/EBP homologous protein havebeen implicated in C1q signaling.87 Thus, in additionto C5b-9, C1q is likely to contribute to mesangial cellapoptosis.

ENDOTHELIAL CELL INJURY AND AHUS

A role for C5b-9 in mediating glomerular endothelialcell injury was shown a number of years ago in a ratmodel of thrombotic microangiopathy induced by theadministration of anti–glomerular endothelial cell anti-body.9,88 Injury to glomerular endothelial cells, capil-lary thrombus formation, and impaired renal functiondeveloped in complement-sufficient rats, but allchanges were prevented in complement-depleted andin C6-deficient rats. In addition to damaging endothe-lial cells in this model, it is likely that C5b-9 activatedthese cells and promoted adherence of blood cells tothe capillary wall. C5b-9 has been shown to stimulatesecretion of von Willebrand factor multimers fromendothelial cells, stimulate endothelial prothrombinaseactivity, and induce expression of tissue factor onthe surface of endothelial cells.88 As noted earlier,

such changes may involve NF-κB and inflammatorycytokines.

In human beings, glomerular endothelial cell injuryis a prominent feature of aHUS. One of the moststriking advances in therapeutics in nephrology inrecent years was the recognition that unregulatedcomplement activation contributes to the developmentof aHUS, which has led to the successful therapeuticuse of an inhibitory humanized anti-C5 antibody(eculizumab). It was reported that 60% of patientswith aHUS have inherited or acquired abnormalities inthe regulatory proteins of the alternative pathway ofcomplement.89 Briefly, in patients who have gain-of-function mutations in activating proteins (eg, C3,factor B) or loss-of-function mutations in regulatoryproteins (factor H, I, and membrane cofactor protein),unregulated activation of the alternative pathwayoccurs on glomerular endothelial cells, resulting inthrombotic glomerulopathy.90 It is not known whyglomerular endothelial cells are particularly vulner-able. Furthermore, given that eculizumab inhibits thecleavage of C5 to C5a and C5b by the C5 convertase,C5a and C5b-9 both may be involved in the patho-genesis of aHUS. The experimental models describedearlier provide insights into how C5b-9 leads toendothelial injury in aHUS. Interestingly, Escherichiacoli Shiga toxin, a pathogenic factor in classic HUS,was shown to induce activation of the complementalternative pathway on the surface of cultured endo-thelial cells.91 In an in vivo mouse model of Shigatoxin/lipopolysaccharide-induced HUS, deletion offactor B partially protected against thrombocytopeniaand glomerular damage, implying a pathogenic rolefor the alternative complement pathway.91 In humanbeings, successful use of eculizumab has been report-ed in Shiga-toxin–producing E coli HUS,92 suggestingthat Shiga toxin HUS also may be complement-mediated.

Complement-mediated endothelial injury also playsan important role in cardiovascular disease.93 C5b-9and C5a activate endothelial cells and up-regulateproinflammatory adhesion molecules such as E-selec-tin, P-selectin, vascular cell adhesion molecule-1, andintercellular adhesion molecule-1. Expression of theseadhesion molecules promotes leukocyte infiltrationinto the vessel wall. Within the intima, monocytestransform into foam cells and release inflammatorycytokines, which promote infiltration and proliferationof vascular smooth muscle cells, leading to formationof a complex atherosclerotic plaque covered by a thinfibrous cap. Continuing cytokine production promotesexpression of tissue factor, a potent activator of bloodcoagulation, in various cellular components of theplaque, and when the fibrous cap thins and breaks(plaque rupture), this internal prothrombotic complexis exposed to blood, leading to thrombus formation.


Further support for the role of complement in athero-sclerosis can be found in increased C3 depositionwithin the intima of human atherosclerotic lesionsand deposition of activated complement componentsC3d and C5b-9 in human coronary thrombi.

ISCHEMIA-REPERFUSION AND CISPLATIN-INDUCEDAKI

Complement activation increasingly has been appreci-ated to play a role in the pathogenesis of AKI.Transient cross-clamping of the renal arteries in exper-imental animals to induce ischemia-reperfusion of thekidney has been used extensively as a model systemto mimic AKI/acute tubular necrosis in human beings.A number of reports have documented complementpathway inhibition as protective against ischemia-reperfusion injury.94 Most of these articles attributedthe protective effect to mitigation of the complementeffects on endothelial cells with less leukocyte infiltra-tion, less local coagulation, and less compromise ofsmall-vessel blood flow. However, there is also evi-dence that C5b-9 injures tubular epithelial cellsdirectly. C3-, C5-, and C6-deficient mice were pro-tected against renal ischemia-reperfusion injurywhereas C4-deficient mice were not,95 suggesting thatthe alternative pathway contributes to the injury.Reconstitution of C6-deficient mice with injection ofC6 restored the injury, indicating that C5b-9 is respon-sible.95 In contrast, inhibition of C5a with an antibodydid not impact on tissue injury and myeloperoxidaselevel in C6-deficient mice, implying that C5a-mediatedneutrophil infiltration does not have a significant rolein this model.95 However, in another study using thesame ischemia-reperfusion model, blockage of the C5areceptor was shown to ameliorate renal injury both in aneutrophil-dependent and neutrophil-independent man-ner.96 More recently, the role of complement in renalischemia-reperfusion injury was addressed in C3a- andC5a-receptor–deficient mice. C5a- and C3aþC5a-receptor–deficient mice were protected from injury,and it appeared that receptors expressed on tubularepithelial cells and inflammatory cells contributed toinjury.97 Miwa et al98 reported that gene deletion ofalternative pathway components (ie, C3, factor B,properdin, C3a receptor, or C5a receptor), but not ofC4 or mannose-binding lectin, protected againstischemia-reperfusion injury in decay-accelerating fac-tor and CD59 double-knockout mice.98 In this study,an anti-C5 antibody showed greater protection than thegene deletion of the C5a receptor, suggesting that bothC5a and C5b-9 contribute to the injury.98 Complementalso was shown to have an important role in cytokineproduction by tubular epithelial cells after renalischemia-reperfusion.99 Thus, complement contributes

to ischemia-reperfusion injury of the kidney at multiplelevels, including neutrophil chemotaxis, vascular endo-thelial cell damage, and direct tubular injury, and bothC5a and C5b-9 most likely contribute. A role forcomplement also has been reported in ischemic injuryof other organs (eg, myocardial infarction andstroke).93

Another example of the role of complement in AKIcan be found in the model of cisplatin-induced neph-rotoxicity. Both C5a- and C5a-receptor–deficient micewere protected against cisplatin-induced AKI.100

Administration of C5 or C5a in the C5-deficient micerestored the sensitivity to cisplatin, whereas doubleknockout of CD59a and CD59b (regulatory proteinsfor the assembly of C5b-9) did not affect sensitivity,suggesting that the signaling triggered by C5a and itsreceptor, but not C5b-9, is likely to contribute tocisplatin-induced nephrotoxicity.100 Cisplatin inducedexpression of multiple cytokines and caspases in therenal tubules, and this induction was diminished in C5-deficient mice and was restored by administration ofC5 or C5a.100 The C5a receptor was shown to beexpressed in proximal and distal tubular epithelial cellsin the human kidney.83 More recently, expression ofthe C5a receptor was reported in the thick ascendinglimb of Henle and the first part of the distal convolutedtubule in the human kidney.101 In addition, underinflammatory conditions, the C5a receptor wasexpressed de novo in the proximal tubule.101 Takentogether, it appears that the C5a receptor is up-regulated when tubular cells are injured and the C5a/C5a-receptor pathway may exacerbate tubular damageby further up-regulating inflammatory cytokines and/orpromoting tubular cell apoptosis.

CONCLUSIONS

At present, treatment of human glomerulonephritisoften relies on nonspecific immunosuppressive agents.There are, however, potential opportunities to usetherapies based on disease mechanisms. These couldinclude drugs that block complement activation, aswell as compounds that inhibit cytotoxic signalingpathways or enhance cytoprotective signals. Experi-mental data indicating that C5b-9 plays a key role inthe development of proteinuria in experimental mem-branous nephropathy prompted a study of eculizumabin patients with idiopathic membranous nephropathy.Unfortunately, there was no effect on proteinuria orrenal function after 16 weeks of treatment, probablybecause of inadequate dosing of eculizumab.102 How-ever, patients who received eculizumab in an open-label extension for a period of 12 months showeda significant decrease of proteinuria. Similarly, the roleof complement activation in endothelial cell injuryand thrombotic microangiopathy has been recognized


in an animal model.9,88 This knowledge led to thesuccessful use of eculizumab in the treatment ofaHUS and Shiga-toxin–induced HUS.89,92 Eculizu-mab currently is being tested in the treatment of densedeposit disease and C3 nephropathy, both of whichhave histologic features of membranoproliferativeglomerulonephritis and are caused by abnormal com-plement activation (US Clinical Trial number:NCT01221181). Thus, eculizumab potentially couldbe a treatment for membranous nephropathy and otherforms of glomerulonephritis. Newer agents that mod-ulate the complement system are being developed. Forexample, TT30 is a fusion protein of the C3 fragment-binding domain of complement-receptor type 2 andthe complement alternative pathway inhibitorydomain of factor H. It was designed to inhibit theC3 activation step and, indeed, was shown to protectparoxysmal nocturnal hemoglobinuria erythrocyteseffectively from complement-mediated hemolysis.103

TT30 and other complement-modulating agentsmay prove useful in complement-mediated glomerulo-pathies.

In this article, we have described many signalingpathways that are triggered by C5b-9. As shown insome experimental models, these pathways maypresent additional opportunities for pharmacologicinterventions, such as drugs to inhibit or activateprotein kinases, modulators of eicosanoid metabolism,blockers of prostanoid receptors, anti-oxidants, poten-tiators or correctors of ER stress, potentiators of theubiquitin-proteasome system, and inhibitors of apop-tosis. It is also conceivable that inhibition of tran-scription factors, such as NF-κB, modulation of thecytoskeleton, or stabilizing the behavior of the slitdiaphragm proteins may have beneficial impact. Inpursuing these opportunities, one has to consider thecomplexities of certain pathways. For example, pros-tanoids could be toxic or protective depending on thecontext.6,7 Similarly, although ROS generally are toxicto the cells, a lower amount of ROS may activatecytoprotective signaling pathways such as JNK.24

Thus, precise understanding of the mechanismsinvolved will be required to identify effective thera-peutic interventions.

Finally, future research also should focus on poten-tial therapeutic interventions in nonglomerular diseases.As discussed earlier, a role for complement has beenshown in experimental models of AKI, and the con-tributions of both C5b-9 and anaphylatoxins such asC5a have been shown.94,100 Atherosclerosis potentiallycan contribute to the development of chronic renovas-cular disease, and complement activation has beenimplicated in the pathogenesis of atherosclerosis.93

Thus, it is conceivable that modulation of the comple-ment system could be beneficial in diseases such asAKI and renovascular disease.

REFERENCES1. Tegla CA, Cudrici C, Patel S, et al. Membrane attack by

complement: the assembly and biology of terminal comple-ment complexes. Immunol Res. 2011;51:45-60.

2. Ricklin D, Hajishengallis G, Yang K, Lambris JD. Comple-ment: a key system for immune surveillance and homeostasis.Nat Immunol. 2010;11:785-97.

3. Leslie M. Immunology. The new view of complement.Science. 2012;337:1034-7.

4. Berger SP, Daha MR. Complement in glomerular injury.Semin Immunopathol. 2007;29:375-84.

5. Quigg RJ. Complement and the kidney. J Immunol.2003;171:3319-24.

6. Cybulsky AV, Quigg RJ, Salant DJ. Experimental membra-nous nephropathy redux. Am J Physiol Renal Physiol.2005;289:F660-71.

7. Cybulsky AV. Membranous nephropathy. Contrib Nephrol.2011;169:107-25.

8. Nangaku M, Shankland SJ, Couser WG. Cellular response toinjury in membranous nephropathy. J Am Soc Nephrol.2005;16:1195-204.

9. Couser WG. Complement inhibitors and glomerulonephritis:are we there yet?. J Am Soc Nephrol. 2003;14:815-8.

10. Morgan BP. Effects of the membrane attack complex ofcomplement on nucleated cells. Curr Top Microbiol Immunol.1992;178:115-40.

11. Nicholson-Weller A, Halperin JA. Membrane signaling bycomplement C5b-9, the membrane attack complex. ImmunolRes. 1993;12:244-57.

12. Bohana-Kashtan O, Ziporen L, Donin N, Kraus S, FishelsonZ. Cell signals transduced by complement. Mol Immunol.2004;41:583-97.

13. Chen ZH, Qin WS, Zeng CH, et al. Triptolide reducesproteinuria in experimental membranous nephropathy andprotects against C5b-9-induced podocyte injury in vitro.Kidney Int. 2010;77:974-88.

14. Koshikawa M, Mukoyama M, Mori K, et al. Role of p38mitogen-activated protein kinase activation in podocyte injuryand proteinuria in experimental nephrotic syndrome. J AmSoc Nephrol. 2005;16:2690-701.

15. Ren G, Huynh C, Bijian K, Cybulsky AV. Role of apoptosissignal-regulating kinase 1 in complement-mediated glomeru-lar epithelial cell injury. Mol Immunol. 2008;45:2236-46.

16. David S, Biancone L, Caserta C, Bussolati B, Cambi V,Camussi G. Alternative pathway complement activation indu-ces proinflammatory activity in human proximal tubularepithelial cells. Nephrol Dial Transplant. 1997;12:51-6.

17. Cohen D, Papillon J, Aoudjit L, Li H, Cybulsky AV, TakanoT. Role of calcium-independent phospholipase A2 incomplement-mediated glomerular epithelial cell injury. Am JPhysiol Renal Physiol. 2008;294:F469-79.

18. Elimam H, Papillon J, Takano T, Cybulsky AV. Complement-mediated activation of calcium-independent phospholipaseA2gamma: role of protein kinases and phosphorylation. JBiol Chem. 2013;288:3871-85.

19. Takano T, Cybulsky AV, Cupples WA, Ajikobi DO, Papillon J,Aoudjit L. Inhibition of cyclooxygenases reduces complement-induced glomerular epithelial cell injury and proteinuria in passiveHeymann nephritis. J Pharmacol Exp Ther. 2003;305:240-9.

20. Weise WJ, Natori Y, Levine JS, et al. Fish oil has protectiveand therapeutic effects on proteinuria in passive Heymannnephritis. Kidney Int. 1993;43:359-68.

21. Adler S, Baker PJ, Johnson RJ, Ochi RF, Pritzl P, Couser WG.Complement membrane attack complex stimulates productionof reactive oxygen metabolites by cultured rat mesangial cells.J Clin Invest. 1986;77:762-7.

http://refhub.elsevier.com/S0270-9295(13)00137-X/sbref1


































































22. Biancone L, David S, Della Pietra V, Montrucchio G, CambiV, Camussi G. Alternative pathway activation of complementby cultured human proximal tubular epithelial cells. KidneyInt. 1994;45:451-60.

23. Neale TJ, Ojha PP, Exner M, et al. Proteinuria in passiveHeymann nephritis is associated with lipid peroxidation andformation of adducts on type IV collagen. J Clin Invest.1994;94:1577-84.

24. Gwinner W, Plasger J, Brandes RP, et al. Role of xanthineoxidase in passive Heymann nephritis in rats. J Am SocNephrol. 1999;10:538-44.

25. Liu H, Tian N, Arany I, et al. Cytochrome P450 2B1 mediatescomplement-dependent sublytic injury in a model of mem-branous nephropathy. J Biol Chem. 2010;285:40901-10.

26. Tryggvason K, Patrakka J, Wartiovaara J. Hereditary protei-nuria syndromes and mechanisms of proteinuria. N Engl JMed. 2006;354:1387-401.

27. Peng H, Takano T, Papillon J, Bijian K, Khadir A, CybulskyAV. Complement activates the c-Jun N-terminal kinase/stress-activated protein kinase in glomerular epithelial cells. JImmunol. 2002;169:2594-601.

28. Gaertner SA, Janssen U, Ostendorf T, Koch KM, Floege J,Gwinner W. Glomerular oxidative and antioxidative systemsin experimental mesangioproliferative glomerulonephritis. JAm Soc Nephrol. 2002;13:2930-7.

29. Budisavljevic MN, Hodge L, Barber K, et al. Oxidative stress inthe pathogenesis of experimental mesangial proliferative glomer-ulonephritis. Am J Physiol Renal Physiol. 2003;285:F1138-48.

30. Gao L, Qiu W, Wang Y, Xu W, Xu J, Tong J. Sublyticcomplement C5b-9 complexes induce thrombospondin-1 pro-duction in rat glomerular mesangial cells via PI3-k/Akt:association with activation of latent transforming growthfactor-beta1. Clin Exp Immunol. 2006;144:326-34.

31. Floege J, Eitner F, Alpers CE. A new look at platelet-derivedgrowth factor in renal disease. J Am Soc Nephrol. 2008;19:12-23.

32. Shankland SJ. The podocyte's response to injury: role inproteinuria and glomerulosclerosis. Kidney Int. 2006;69:2131-47.

33. Mudge SJ, Paizis K, Auwardt RB, Thomas RJ, Power DA.Activation of nuclear factor-kappa B by podocytes in theautologous phase of passive Heymann nephritis. Kidney Int.2001;59:923-31.

34. Hussain S, Romio L, Saleem M, et al. Nephrin deficiencyactivates NF-{kappa}B and promotes glomerular injury. J AmSoc Nephrol. 2009;20:1733-43.

35. Kilgore KS, Schmid E, Shanley TP, et al. Sublytic concen-trations of the membrane attack complex of complementinduce endothelial interleukin-8 and monocyte chemoattrac-tant protein-1 through nuclear factor-kappa B activation. Am JPathol. 1997;150:2019-31.

36. Gancz D, Lusthaus M, Fishelson Z. A role for the NF-kappaBpathway in cell protection from complement-dependent cyto-toxicity. J Immunol. 2012;189:860-6.

37. Pippin JW, Durvasula R, Petermann A, Hiromura K, CouserWG, Shankland SJ. DNA damage is a novel response tosublytic complement C5b-9-induced injury in podocytes. JClin Invest. 2003;111:877-85.

38. Kim SH, Carney DF, Hammer CH, Shin ML. Nucleated cellkilling by complement: effects of C5b-9 channel size andextracellular Ca2þ on the lytic process. J Immunol.1987;138:1530-6.

39. Moskovich O, Fishelson Z. Live cell imaging of outward andinward vesiculation induced by the complement c5b-9 com-plex. J Biol Chem. 2007;282:29977-86.

40. Kerjaschki D, Schulze M, Binder S, et al. Transcellulartransport and membrane insertion of the C5b-9 membrane

attack complex of complement by glomerular epithelial cellsin experimental membranous nephropathy. J Immunol. 1989;143:546-52.

41. Cybulsky AV. Endoplasmic reticulum stress in proteinurickidney disease. Kidney Int. 2010;77:187-93.

42. Cybulsky AV. The intersecting roles of endoplasmic reticulumstress, ubiquitin- proteasome system, and autophagy in thepathogenesis of proteinuric kidney disease. Kidney Int.2013;84:25-33.

43. Cybulsky AV, Takano T, Papillon J, Khadir A, Liu J, Peng H.Complement C5b-9 membrane attack complex increasesexpression of endoplasmic reticulum stress proteins in glo-merular epithelial cells. J Biol Chem. 2002;277:41342-51.

44. Wang L, Hong Q, Lv Y, et al. Autophagy can repairendoplasmic reticulum stress damage of the passive Heymannnephritis model as revealed by proteomics analysis. J Proteo-mics. 2012;75:3866-76.

45. Inagi R, Kumagai T, Nishi H, et al. Preconditioning withendoplasmic reticulum stress ameliorates mesangioprolifera-tive glomerulonephritis. J Am Soc Nephrol. 2008;19:915-22.

46. Meyer-Schwesinger C, Meyer TN, Sievert H, et al. UbiquitinC-terminal hydrolase-l1 activity induces polyubiquitin accu-mulation in podocytes and increases proteinuria in rat mem-branous nephropathy. Am J Pathol. 2011;178:2044-57.

47. Kitzler TM, Papillon J, Guillemette J, Wing SS, CybulskyAV. Complement modulates the function of the ubiquitin-proteasome system and endoplasmic reticulum-associateddegradation in glomerular epithelial cells. Biochim BiophysActa. 2012;1823:1007-16.

48. Meyer-Schwesinger C, Meyer TN, Munster S, et al. A newrole for the neuronal ubiquitin C-terminal hydrolase-L1(UCH-L1) in podocyte process formation and podocyte injuryin human glomerulopathies. J Pathol. 2009;217:452-64.

49. Hauser PV, Perco P, Muhlberger I, et al. Microarray andbioinformatics analysis of gene expression in experimentalmembranous nephropathy. Nephron Exp Nephrol. 2009;112:e43-58.

50. Topham PS, Haydar SA, Kuphal R, Lightfoot JD, Salant DJ.Complement-mediated injury reversibly disrupts glomerularepithelial cell actin microfilaments and focal adhesions.Kidney Int. 1999;55:1763-75.

51. Zhang H, Cybulsky AV, Aoudjit L, et al. Role of Rho-GTPases in complement-mediated glomerular epithelial cellinjury. Am J Physiol Renal Physiol. 2007;293:F148-56.

52. Zhu L, Jiang R, Aoudjit L, Jones N, Takano T. Activation ofRhoA in podocytes induces focal segmental glomeruloscle-rosis. J Am Soc Nephrol. 2011;22:1621-30.

53. Huynh C, Ren G, Papillon J, Guillemette J, Takano T,Cybulsky AV. The cytoprotective role of Ras incomplement-mediated glomerular epithelial cell injury. ClinImmunol. 2009;131:343-53.

54. Moller CC, Wei C, Altintas MM, et al. Induction of TRPC6channel in acquired forms of proteinuric kidney disease. J AmSoc Nephrol. 2007;18:29-36.

55. Nakatsue T, Koike H, Han GD, et al. Nephrin and podocindissociate at the onset of proteinuria in experimental mem-branous nephropathy. Kidney Int. 2005;67:2239-53.

56. Li H, Lemay S, Aoudjit L, Kawachi H, Takano T. Src-familykinase Fyn phosphorylates the cytoplasmic domain of nephrinand modulates its interaction with podocin. J Am SocNephrol. 2004;15:3006-15.

57. Wagner C, Braunger M, Beer M, Rother K, Hansch GM.Induction of matrix protein synthesis in human glomerularmesangial cells by the terminal complement complex. ExpNephrol. 1994;2:51-6.

58. Brandt J, Pippin J, Schulze M, et al. Role of the complementmembrane attack complex (C5b-9) in mediating experimental









































































































































mesangioproliferative glomerulonephritis. Kidney Int. 1996;49:335-43.

59. Shankland SJ, Pippin J, Pichler RH, et al. Differential expression oftransforming growth factor-beta isoforms and receptors in exper-imental membranous nephropathy. Kidney Int. 1996;50:116-24.

60. Rangan GK, Pippin JW, Couser WG. C5b-9 regulatesperitubular myofibroblast accumulation in experimental focalsegmental glomerulosclerosis. Kidney Int. 2004;66:1838-48.

61. Abe K, Li K, Sacks SH, Sheerin NS. The membrane attackcomplex, C5b-9, up regulates collagen gene expression in renaltubular epithelial cells. Clin Exp Immunol. 2004;136:60-6.

62. Abbate M, Zoja C, Corna D, et al. Complement-mediateddysfunction of glomerular filtration barrier accelerates pro-gressive renal injury. J Am Soc Nephrol. 2008;19:1158-67.

63. Gaarkeuken H, Siezenga MA, Zuidwijk K, et al. Complementactivation by tubular cells is mediated by properdin binding.Am J Physiol Renal Physiol. 2008;295:F1397-403.

64. Shankland SJ, Pippin JW, Couser WG. Complement (C5b-9)induces glomerular epithelial cell DNA synthesis but notproliferation in vitro. Kidney Int. 1999;56:538-48.

65. Rus HG, Niculescu FI, Shin ML. Role of the C5b-9 comple-ment complex in cell cycle and apoptosis. Immunol Rev.2001;180:49-55.

66. Qiu W, Zhang Y, Liu X, et al. Sublytic C5b-9 complexesinduce proliferative changes of glomerular mesangial cells inrat Thy-1 nephritis through TRAF6-mediated PI3K-dependentAkt1 activation. J Pathol. 2011;226:619-32.

67. Fosbrink M, Niculescu F, Rus V, Shin ML, Rus H. C5b-9-induced endothelial cell proliferation and migration aredependent on Akt inactivation of forkhead transcription factorFOXO1. J Biol Chem. 2006;281:19009-18.

68. Soane L, Cho HJ, Niculescu F, Rus H, Shin ML. C5b-9terminal complement complex protects oligodendrocytes fromdeath by regulating Bad through phosphatidylinositol 3-kin-ase/Akt pathway. J Immunol. 2001;167:2305-11.

69. Cudrici C, Niculescu F, Jensen T, et al. C5b-9 terminalcomplex protects oligodendrocytes from apoptotic cell deathby inhibiting caspase-8 processing and up-regulating FLIP. JImmunol. 2006;176:3173-80.

70. Nauta AJ, Daha MR, Tijsma O, van de Water B, Tedesco F,Roos A. The membrane attack complex of complementinduces caspase activation and apoptosis. Eur J Immunol.2002;32:783-92.

71. Hughes J, Nangaku M, Alpers CE, Shankland SJ, Couser WG,Johnson RJ. C5b-9 membrane attack complex mediatesendothelial cell apoptosis in experimental glomerulonephritis.Am J Physiol Renal Physiol. 2000;278:F747-57.

72. Ziporen L, Donin N, Shmushkovich T, Gross A, Fishelson Z.Programmed necrotic cell death induced by complementinvolves a Bid-dependent pathway. J Immunol. 2009;182:515-21.

73. Roos A, Sato T, Maier H, van Kooten C, Daha MR. Inductionof renal cell apoptosis by antibodies and complement. ExpNephrol. 2001;9:65-70.

74. Jiang X, Zhang J, Xia M, et al. Role of activating transcriptionfactor 3 (ATF3) in sublytic C5b-9-induced glomerular mesan-gial cell apoptosis. Cell Mol Immunol. 2010;7:143-51.

75. Xu K, Zhou Y, Qiu W, et al. Activating transcription factor3 (ATF3) promotes sublytic C5b-9-induced glomerularmesangial cells apoptosis through up-regulation ofGadd45Î� and KLF6 gene expression. Immunobiology.2011;216:871-81.

76. Liu L, Qiu W, Wang H, et al. Sublytic C5b-9 complexesinduce apoptosis of glomerular mesangial cells in rats withThy-1 nephritis through role of interferon regulatory factor-1-dependent caspase 8 activation. J Biol Chem. 2012;287:16410-23.

77. Mudge SJ, McRae JL, Auwardt RB, Murphy BF, Chen CG,Power DA. Sublytic complement injury does not activate NF-kappa B, or induce mitogenesis in rat mesangial cells. ExpNephrol. 2000;8:291-8.

78. Guo RF, Huber-Lang M, Wang X, et al. Protective effects ofanti-C5a in sepsis-induced thymocyte apoptosis. J Clin Invest.2000;106:1271-80.

79. Pasinetti GM, Tocco G, Sakhi S, et al. Hereditary deficienciesin complement C5 are associated with intensified neuro-degenerative responses that implicate new roles for theC-system in neuronal and astrocytic functions. NeurobiolDis. 1996;3:197-204.

80. Osaka H, Mukherjee P, Aisen PS, Pasinetti GM. Complement-derived anaphylatoxin C5a protects against glutamate-mediated neurotoxicity. J Cell Biochem. 1999;73:303-11.

81. Perianayagam MC, Balakrishnan VS, King AJ, Pereira BJ,Jaber BL. C5a delays apoptosis of human neutrophils by aphosphatidylinositol 3-kinase-signaling pathway. Kidney Int.2002;61:456-63.

82. Welch TR, Frenzke M, Witte D, Davis AE. C5a is important inthe tubulointerstitial component of experimental immune com-plex glomerulonephritis. Clin Exp Immunol. 2002;130:43-8.

83. Zahedi R, Braun M, Wetsel RA, et al. The C5a receptor isexpressed by human renal proximal tubular epithelial cells.Clin Exp Immunol. 2000;121:226-33.

84. Pickering MC, Cook HT, Warren J, et al. Uncontrolled C3activation causes membranoproliferative glomerulonephritis inmice deficient in complement factor H. Nat Genet. 2002;31:424-8.

85. Pickering MC, Warren J, Rose KL, et al. Prevention of C5activation ameliorates spontaneous and experimental glomer-ulonephritis in factor H-deficient mice. Proc Natl Acad SciU S A. 2006;103:9649-54.

86. Sato T, van Dixhoorn MG, Schroeijers WE, van Es LA, DahaMR. Efficient induction of apoptosis in cultured rat glomerularmesangial cells by dimeric monoclonal IgA anti-Thy-1 anti-bodies. Kidney Int. 1997;51:173-81.

87. Bordin S, Whitfield D. Cutting edge: proliferating fibroblastsrespond to collagenous C1q with phosphorylation of p38mitogen-activated protein kinase and apoptotic features. JImmunol. 2003;170:667-71.

88. Nangaku M, Shankland SJ, Couser WG, Johnson RJ. A newmodel of renal microvascular injury. Curr Opin NephrolHypertens. 1998;7:457-62.

89. Kavanagh D, Goodship TH. Atypical hemolytic uremicsyndrome, genetic basis, and clinical manifestations. Hema-tology Am Soc Hematol Educ Program. 2011;2011:15-20.

90. Vernon KA, Cook HT. Complement in glomerular disease.Adv Chronic Kidney Dis. 2012;19:84-92.

91. Morigi M, Galbusera M, Gastoldi S, et al. Alternative pathwayactivation of complement by Shiga toxin promotes exuberantC3a formation that triggers microvascular thrombosis. JImmunol. 2011;187:172-80.

92. Lapeyraque AL, Malina M, Fremeaux-Bacchi V, et al. Ecu-lizumab in severe Shiga-toxin-associated HUS. N Engl J Med.2011;364:2561-3.

93. Carter AM. Complement activation: an emerging player in thepathogenesis of cardiovascular disease. Scientifica. 2012;2012:14.

94. Bonventre JV. Complement and renal ischemia-reperfusioninjury. Am J Kidney Dis. 2001;38:430-6.

95. Zhou W, Farrar CA, Abe K, et al. Predominant role for C5b-9in renal ischemia/reperfusion injury. J Clin Invest. 2000;105:1363-71.

96. de Vries B, Kohl J, Leclercq WK, et al. Complement factorC5a mediates renal ischemia-reperfusion injury independentfrom neutrophils. J Immunol. 2003;170:3883-9.









































































































































97. Peng Q, Li K, Smyth LA, et al. C3a and C5a promote renalischemia-reperfusion injury. J Am Soc Nephrol. 2012;23:1474-85.

98. Miwa T, Sato S, Gullipalli D, Nangaku M, Song W-C.Blocking properdin, the alternative pathway, and anaphyla-toxin receptors ameliorates renal ischemia-reperfusion injuryin decay-accelerating factor and CD59 double-knockout mice.J Immunol. 2013;190:3552-9.

99. Amura CR, Renner B, Lyubchenko T, Faubel S, Simonian PL,Thurman JM. Complement activation and toll-like receptor-2signaling contribute to cytokine production after renal ische-mia/reperfusion. Mol Immunol. 2012;52:249-57.

100. Pan H, Shen Z, Mukhopadhyay P, et al. Anaphylatoxin C5acontributes to the pathogenesis of cisplatin-induced nephro-toxicity. Am J Physiol Renal Physiol. 2009;296:F496-504.

101. van Werkhoven MB, Damman J, Daha MR, et al. Novelinsights in localization and expression levels of C5aR andC5L2 under native and post-transplant conditions in thekidney. Mol Immunol. 2013;53:237-45.

102. Cunningham PN, Quigg RJ. Contrasting roles of complementactivation and its regulation in membranous nephropathy.J Am Soc Nephrol. 2005;16:1214-22.

103. Risitano AM, Notaro R, Pascariello C, et al. The complementreceptor 2/factor H fusion protein TT30 protects paroxysmalnocturnal hemoglobinuria erythrocytes from complement-mediated hemolysis and C3 fragment. Blood. 2012;119:6307-16.

104. Cybulsky AV, Takano T, Papillon J, Bijian K. Role of theendoplasmic reticulum unfolded protein response in glomer-ular epithelial cell injury. J Biol Chem. 2005;280:24396-403.































tomoko takano, md, phd, hanan elimam, msc, and andrey v....

Documents