Kidney International, Vol. 58 (2000), pp. 1580–1587

Complement is activated in kidney by endotoxin but does notcause the ensuing acute renal failure

PATRICK N. CUNNINGHAM, V. MICHAEL HOLERS, JESSY J. ALEXANDER, JOEL M. GUTHRIDGE,MICHAEL C. CARROLL, and RICHARD J. QUIGG

Department of Medicine, Section of Nephrology, The University of Chicago, Chicago, Illinois; Department of Medicine,Division of Rheumatology, University of Colorado Health Sciences Center, Denver, Colorado; and Department of Pathology,Harvard Medical School, Boston, Massachusetts, USA

Complement is activated in kidney by endotoxin but does not tion rate is associated with the pathological lesion ofcause the ensuing acute renal failure. acute tubular necrosis. ARF in the setting of sepsis often

Background. Acute renal failure (ARF) in sepsis occurs occurs despite relatively modest levels of hypotension,when the release of multiple inflammatory mediators is inducedimplying that some factor(s) other than ischemia aloneby bacterial endotoxins. C3 mRNA is markedly up-regulatedis involved. In sepsis, bacterial products such as lipopoly-in mouse kidney after exposure to lipopolysaccharide (LPS).

We hypothesized that LPS could induce tubular synthesis and saccharide (LPS) trigger the release of multiple inflam-secretion of C3, leading to activation of the complement cas- matory mediators, including the complement cascade [1].cade and direct renal tubular injury.

Injection of animals with LPS can serve as a useful modelMethods. ARF was induced in mice by intravenous injectionof endotoxemia, leading to widespread cytokine releaseof LPS and was confirmed by an acute rise in blood urea nitrogen

(BUN) and histologically by acute tubular necrosis. Three sepa- and resulting in ARF [2].rate strategies were used to investigate the role of the comple- Several complement regulatory proteins serve to in-ment system in this model of ARF: (1) Crry-Ig, a recombinant hibit and contain the complement cascade. One key regu-protein containing the potent murine complement C3 activation

lator in the mouse is Crry (complement receptor 1-relatedinhibitor Crry was injected at the same time as LPS (N 5 8).gene/protein y) [3–6], a membrane-bound C3 convertase(2) LPS was injected into transgenic mice overexpressing Crry

in glomeruli and tubules (N 5 8), and (3) LPS was injected inhibitor present in a variety of tissues, including the kid-into C3-deficient mice (N 5 5). ney [7, 8]. Various strategies have been used to show that

Results. Compared with unmanipulated mice, C3 staining increasing the degree of complement inhibition decreasesby immunofluorescence (IF) microscopy in mice injected withinjury in rodent models of glomerulonephritis [9–11].LPS was greater in renal cortical tubular cells (IF score of 2.1 6

0.1 vs. 1.4 6 0.2 in controls, P 5 0.013), most prominently at Although the liver is the major source of serum com-the basolateral surface. LPS injection led to a 16- to 42-fold plement, other tissues, including kidney, have been shownincrease in urinary C3 excretion. Despite reduction or complete to synthesize components of the complement system.elimination of renal C3 with maneuvers suppressing comple-

Renal cortical C3 mRNA levels are acutely increasedment activation, BUN values were not statistically differentin vivo in mice after LPS injection [12]. Additionally,across all groups. In no experiment did BUN values correlate

with the extent of C3 staining. cultured human proximal tubular cells immediately syn-Conclusion. Although LPS up-regulates renal C3 synthesis, thesize and secrete functionally active C3 after exposure

resulting in basolateral tubular C3 deposition, this is not re- to various cytokines, including tumor necrosis factor-a,sponsible for LPS-induced ARF in mice.interleukin-1a, interferon-g, or interleukin-2 [13, 14].There is evidence that under immunologic activation,such as occurs with transplant rejection, tubular cells canSepsis is a common cause of acute renal failure (ARF),produce C3 [15] and contribute to the circulating C3an entity in which the abrupt decline in glomerular filtra-pool [16].

Basolateral deposition of complement around renaltubules has been demonstrated in chronic renal paren-Key words: acute tubular necrosis, C3, lipopolysaccharide, recombi-

nant proteins, transgenic mice, knockout mice, sepsis. chymal disease, which has been attributed to comple-ment activation by the high interstitial ammonia concen-Received for publication October 18, 1999tration [17]. Human proximal tubular cells in culture areand in revised form March 23, 2000

Accepted for publication May 4, 2000 capable of activating the alternative pathway of comple-ment [18], terminating in the binding of the C5b-9 mem- 2000 by the International Society of Nephrology

1580

Cunningham et al: Complement-independent RF from endotoxin 1581

brane attack complex to the cellular surface, followed gous recombination (C32/2) and their wild-type con-by release of inflammatory mediators such as arachidonic trols (C31/1), on a 129SvJ and C57BL/6 mixed back-acid metabolites, interleukin-6, and tumor necrosis fac- ground, were generated as previously described [28, 29].tor-a [19]. The luminal brush border may be more vul- Normal CD-1 mice were purchased from Harlan Spraguenerable to complement-mediated injury because it lacks Dawley (Indianapolis, IN, USA).the complement regulatory proteins present on the baso-

Crry-Iglateral cell surface [20]. This has led to the hypothesisthat the chronic tubular damage seen with long-standing A Crry-Ig chimera was used in these studies [11]. Theproteinuria is due to luminal deposition of filtered com- construct used to produce this uses the CMV promoterponents of complement [21]. In support of this, luminal and encodes a signal peptide followed by the five shortcomplement deposition has been detected in several ani- consensus repeats of Crry and the hinge, CH2 and CH3mal models of nonimmunologic proteinuria [22–24]. domains of mouse IgG1, a noncomplement fixing isotype.

Ischemic damage itself, independent of endotoxin, The circulating half-life of Crry-Ig is approximately 40may impair the ability of renal tubular cells to regulate hours, and thus, a single intravenous dose of 3 mg confersthe complement cascade and therefore increase comple- lasting complement inhibition for acute studies [11]. Toment activation on their surface [25]. In addition to the serve as an appropriate control for Crry-Ig, MOPC-21,previously mentioned mechanisms, all of which would an irrelevant monoclonal murine IgG1 antibody, was pu-involve the alternative pathway, ischemic damage to cells rified from ascites (Sigma, St. Louis, MO, USA) via ion-can cause exposure of previously hidden cell membrane exchange chromatography (Amersham-Pharmacia-Bio-antigens and binding of “natural antibodies,” of the tech, Uppsala, Sweden).strongly complement-activating IgM isotype, leading topropagation of the classical pathway [26]. Beyond their Basic experimental designlocal inflammatory effects, complement cleavage prod- A model for ARF in mice was developed in whichucts C3a and C5a have vascular effects [27] that could animals were injected intravenously with LPS (E. colicontribute to changes in renal hemodynamics in ARF. serotype O111:B4 LPS; Sigma). As shown in Figure 1,Thus, by a variety of mechanisms, the complement sys- there was a dose–response relationship between thetem may be a key effector of ARF in sepsis. amount of LPS injected and the resultant ARF. Based

Given these interactions, we hypothesized that theon this dose–response relationship, 0.15 mg per mouse

complement system could potentially play a key role inwas selected as the optimal dose. Three experiments

endotoxemic ARF. LPS could trigger increased localwere conducted to determine the influence of the com-synthesis, activation, and deposition of C3 either at theplement system in endotoxin-induced ARF. In the firstbasolateral or luminal surface of the tubules, resulting inexperiment, eight normal CD-1 mice were injected intra-propagation of the alternative and terminal complementvenously with 0.15 mg of LPS concurrently with 3 mg ofpathways leading to tubular cell damage and acute tubu-Crry-Ig. These were compared with a control grouplar necrosis. This study examines the effect of three dif-of eight normal CD-1 mice injected with the same doseferent strategies to inhibit complement on LPS-inducedof LPS together with 3 mg of MOPC-21 IgG1. For therenal injury.second experiment, eight Crry transgenic mice and eightlittermate animals lacking the transgene were injected

METHODS with 0.15 mg LPS. In the third experiment, five C32/2mice were injected with 0.15 mg LPS and compared withMicefive C31/1 mice treated identically.Transgenic mice expressing recombinant soluble Crry

Mice were anesthetized with methoxyflurane inhala-directed by the metallothionein-I promoter were usedtion. After initial injection, animals were placed in meta-in this study. These mice have circulating levels of solublebolic cages for urine collection overnight. In all mice,Crry that are complement inhibitory, and high intrinsicserum was obtained via retro-orbital bleeding for base-renal production of Crry, including in glomeruli and tu-line blood urea nitrogen (BUN) immediately before in-bules [10]. The Crry transgenic animals used in this studyjection and 20 hours after injection. Following this, ani-were derived from a single founder line and are on amals were sacrificed and renal tissue harvested for lightCD-1 background. In all animals, the presence of theand immunofluorescence (IF) microscopy.Crry transgene was documented by polymerase chain

For comparison of urinary and tissue C3, three normalreaction (PCR), and soluble Crry in sera was identifiedCD-1 mice undergoing no experimental manipulationby enzyme-linked immunosorbent assay (ELISA) [10].were placed in metabolic cages, and urine was collectedAs controls, littermate animals that lacked the Crryovernight. These mice were then sacrificed, and kidneytransgene were used.

Mice in which the C3 gene was inactivated by homolo- tissue was taken for IF microscopy.

Cunningham et al: Complement-independent RF from endotoxin1582

Fig. 1. Lipopolysaccharide (LPS) leads toacute renal failure (ARF) in a dose-dependentfashion. Blood urea nitrogen (BUN) measure-ments taken 24 hours after intravenous injec-tion of various doses of LPS are shown asmean values 6 SEM. Baseline BUN valueswere 21.6 6 0.9 mg/dL.

Analysis of serum and urine pel, Organon Teknika Corp., Durham, NC, USA). IFstaining scores in glomeruli and tubules were compiledBlood urea nitrogen concentration was determinedby an observer blinded to the origin of the samples.by the urease-glutamate dehydrogenase assay using a

Beckman CX5CE autoanalyzer.Statistical analysisSerum and urinary levels of mouse C3 and albumin

All data are reported as the mean 6 SEM. A compari-were measured by previously described ELISA protocolsson between variables at baseline and following LPS[30, 31]. The C3 ELISA will react with C3 in its nativeadministration was made by paired t-testing. Comparisonor activated forms [30] and was sensitive to 1.5 ng/mL.between an experimental group and its respective con-When an animal had undetectable urinary C3 by thistrol group was done using an unpaired t-test. A correla-technique, it was considered to be zero. As expected, C3tion between two variables within groups was done bywas not detected in C32/2 mice, and this group was notunivariate linear regression. Statistical calculations wereentered in analysis of sera C3 levels. Creatinine concen-done using Minitab software (State College, PA, USA).tration was measured by the modified rate Jaffe reaction

using a Beckman CX5CE autoanalyzer. The fractionalexcretion of C3 (FEC3) was determined by dividing CC3 RESULTSby CCr, leading to the following calculation: [plasma creati-

ARF occurs after endotoxin administrationnine][urine C3]/[urine creatinine][plasma C3]. FEAlbuminAs shown in Figure 1, mice given LPS intravenouslyrelative to creatinine was determined in an analogous

developed ARF. Excessive mortality was observed atfashion. For these calculations, we used normal values indoses above 0.15 mg. As all animals receiving 0.15 mgthe mouse for serum creatinine (3 mg/mL), C3 (3 mg/mL),developed renal failure, this dose was chosen for subse-and albumin (37 mg/mL) derived in our laboratory.quent studies in which we determined the effects of in-

Renal tissue processing hibiting complement activation on the resultant ARF.Three different experimental strategies (Crry-Ig in nor-A sagittal section of renal tissue was fixed in bufferedmal mice, Crry transgenic mice, and C32/2 mice), eachformalin. Five micrometer sections were stained with peri-with its relative advantage in this model, were used toodic acid-Schiff and were processed for light microscopicdetermine conclusively the effects of complement in thisevaluation. A renal tubular injury score was adapted fromLPS-mediated ARF.that described by Nomura et al [32], incorporating tubu-

In all groups of animals, baseline BUN values werelar cast formation, dilation, and degeneration, each scoredno different prior to LPS injection (22.8 6 0.7 mg/dL inby a blinded observer from 0 to 3; the composite scorethe 6 groups of animals). Following LPS injection, BUNreported adds all three values together.rose acutely in every animal in each experiment. In theFor IF microscopy, tissue was snap frozen in isopen-three experiments, only one animal, in the C31/1 group,tane on dry ice. Four micrometer cryostat sections weredied. Individual data points are shown in Figure 2. How-processed for direct IF microscopy as described pre-

viously [33] using FITC-conjugated anti-mouse C3 (Cap- ever, no statistically significant differences were found

Cunningham et al: Complement-independent RF from endotoxin 1583

Fig. 2. Effect of complement inhibition onLPS-induced ARF. BUN values obtained 20hours after LPS injection are shown for eachgroup. Each data point corresponds to an indi-vidual animal. Symbols are: (d) groups inwhich a complement-reducing strategy wasused; (s) respective control groups. No sig-nificant difference in BUN was found betweenany experimental group and its particular con-trol group. Baseline BUN values were 22.8 60.7 mg/dL in the six groups of animals.

between any of the experimental and respective control nificantly less tubular C3 staining in the Crry-transgenicgroup compared with its control group (P 5 0.012; Fig. 6),groups. These results show that ARF occurs with endo-

toxin administration, but suggest that complement is un- and as expected, no C3 staining in C32/2 mice (P #0.001 vs. C31/1 animals). Glomerular C3 staining wasimportant in the pathogenesis of this condition.minimal in all experimental and control groups (data not

Light microscopy shown). C3 staining did not correlate with BUN values.Thus, complement activation and deposition occur withTissue from all animals in both experimental and con-

trol groups showed evidence for tubular injury with tubu- LPS activation, but appear to not be responsible forthe ARF.lar dilation, vacuolization, luminal debris, and brush bor-

der attenuation (Fig. 3B). Quantitatively, however, thisSerum and urinary C3 and albuminwas of moderate severity and indistinguishable in degree

between experimental and control groups receiving LPS. Serum C3 levels increased following LPS administra-tion, although this was not statistically significant (3.10 6In confirmation that an LPS dose of 0.15 mg led to

adequate renal injury, the tubular injury score at a dose 0.22 and 3.60 6 0.22 mg/mL before and after LPS admin-istration, respectively, P 5 0.11). With the use of sensi-of 0.15 mg was comparable to that seen with 0.5 mg in

normal CD-1 mice (respective injury scores of 3.0 6 0.2 tive ELISA techniques, there was detectable urinary C3(range, 0 to 198 ng/mL) and albumin (range, 4.8 to 15.3and 3.7 6 0.8, N 5 4, P 5 NS). Glomeruli in all groups

of animals were normal (Fig. 4). mg/mL) in normal animals. FEC3 and FEAlbumin are shownin Table 1. Upon injection of LPS, urinary C3 excretion

Tissue C3 was elevated between 16- and 42-fold relative to normalanimals (P , 0.001 vs. unmanipulated mice), with theNormal unmanipulated mice had focal segmental tubu-

lar staining for C3, as well as discontinuous linear staining obvious exception of C32/2 mice (Table 1). Urinaryalbumin was also elevated in all mice receiving LPS com-of Bowman’s capsule. This tubular staining was at the

basolateral aspect of the tubules and present in cortical pared with unmanipulated normals (P , 0.001), but wasstatistically similar between each experimental groupbut not medullary areas (data not shown). This is not

artifactual, as this staining pattern is absent in C32/2 and its control group. The absolute and relative excretionof albumin was consistently elevated compared with uri-mice. In normal mice given LPS, there was increased

C3 deposition in the basolateral aspects of tubules (IF nary C3 in all groups (Table 1).There was a positive correlation of urinary C3 withstaining scores in normal and LPS-injected mice, 1.4 6

0.2 and 2.1 6 0.1, respectively, P 5 0.013). urinary albumin (r 5 0.66, P , 0.001).There was no correlation in pooled data or individualThe tubular IF scores in the various groups are shown

graphically in Figure 5. There was a trend for less tubular experiments between BUN and either urinary excretionof C3 or albumin as absolute values or as fractionalC3 staining in the experimental group receiving Crry-Ig

when compared with its corresponding control group, excretions relative to urinary creatinine. In addition,there was no correlation in pooled data or individualbut these were not statistically different. There was sig-

Cunningham et al: Complement-independent RF from endotoxin1584

Fig. 3. LPS-induced ARF is associated with tubular damage. Compared with a normal control kidney (A), animals injected with LPS showedchanges of tubular dilation (asterisks), vacuolization (arrows) and brush border attenuation and sloughing (arrowhead; B). This was indistinguishablein severity between experimental and control groups given LPS.

Fig. 4. Glomeruli from mice with LPS-inducedARF are normal. Representative glomerulifrom a normal control (A) and LPS-injectedanimal (B) are shown.

Fig. 5. Immunofluorescence (IF) scores for tubular C3 staining in micereceiving LPS. Data shown are mean 6 SEM. *P 5 0.012; **P # 0.001vs. corresponding control groups.

Fig. 6. Tubular C3 staining in mice with LPS-induced ARF. After LPS administration, tubules stained for C3 by IF microscopy, in a predominantlybasolateral and cortical distribution. Animals positive for the Crry transgene showed significantly less tubular C3 staining (A) than transgenenegative controls (B).

Cunningham et al: Complement-independent RF from endotoxin 1585

Table 1. Urinary excretion of C3 and albumin in the different groups of animals

Group LPS CC3/CCr (3106) CAlbumin/CCr (3106) CAlbumin/CC3

CD-1 — 0.260.2 3.160.8 6.3a

CD-1 1 MOPC-21 1 4.261.4 143652.3 46.7617.3CD-1 1 Crry-Ig 1 10.063.8 125656.0 11.261.3Crry transgene negative 1 5.361.7 44.868.0 10.362.4Crry transgene positive 1 3.860.8 38.867.2 13.663.6C31/1 1 3.060.9 33.4619.1 8.562.5C32/2 1 0.060.0 15.461.2 —a

Abbreviations are: C, clearance; Cr, creatinine; LPS, lipopolysaccharide.a Values of CC3 5 0 were excluded from analysis

experiments between tubular staining for C3 and urinary plex. Although C3 staining was indeed absent on IF,C3 or albumin. these mice uniformly developed renal failure.

By using these three unique experimental approaches,we conclusively established the role of complement inDISCUSSIONLPS-induced renal injury. In this regard, it is necessary

In this study, we developed a model of sepsis-induced to emphasize that there are differences between eventsacute tubular necrosis in which animals were injected occurring in complement-deficient animals and in com-intravenously with LPS. Although LPS triggers basolat- plement-inhibited animals. The former investigates theeral deposition of complement in the renal cortical tu- effects of complete deficiencies, while the latter deter-bules, inhibition of this in three unique experimental mines how inhibiting complement activation will affectmethods did not prevent ARF. outcome measures. The inhibition of complement is not

The first experiment sought to inhibit the complement absolute with inhibitors, but is more relevant to clinicalcascade at the critical C3 convertase stage via injection utility, as such inhibitors are available clinically [35]. Theof the recombinant Crry-Ig concurrently with LPS. We analysis of disease models using knockouts has beenbelieve these studies have particular relevance given the widely used, but there are limitations for mechanisticcurrent availability of recombinant protein inhibitors of

analyses and assessment of therapeutic strategies. Ancomplement such as soluble CR1 and humanized anti-

example of this has recently been reviewed by SteinmanC5 monoclonal antibodies [9, 34, 35]. Although thisregarding the role of the cytokines tumor necrosis factor-group had a trend toward less C3 deposition by IF, ita and lymphotoxin-a in multiple sclerosis and its mo-did not reach statistical significance. This relative lackdel, experimental autoimmune encephalomyelitis (EAE)of effect could be due to a variety of reasons. Perhaps[36]. Specifically, although a double knockout of tumorthe synthesis of complement occurs too rapidly in com-necrosis factor-a and lymphotoxin-a can still developparison with the diffusion of the 160 kD Crry-Ig into theEAE with normal kinetics and manifestations, it is notrenal interstitium. Additionally, poor renal perfusion,correct to assume that there is no role for these cytokineswhich may in fact be the primary mechanism of ARF inin EAE and multiple sclerosis. That is because dozensthis setting, would limit delivery of Crry-Ig to the kidney.of other mechanistic studies in animals and humans sup-Consistent with this, staining for C3 in all groups had aport an important role for these cytokines. There aregeneral patchy quality.several possible explanations for these findings. TheseThe transgenic mice overexpressing the murine C3include the limitations of current knockout strategiesinhibitor Crry as a soluble protein did show notably lessbecause they eliminate expression of the particular genetubular C3 staining than controls. This shows that theproduct throughout the body beginning with conception,increased C3 deposition seen after LPS injection requireswith likely untoward and unknown effects, and the possi-the action of the C3 convertase. The inhibition of com-bility of disease-inducing parallel pathways that can by-plement deposition in the kidney in the Crry-transgenepass one defect [36]. Whatever the explanation, however,group was more effective than in the Crry-Ig group,one cannot eliminate a role for a particular proinflam-perhaps because the smaller size of the 45 kD Crry allowsmatory pathway in disease using knockout strategiesit to diffuse better into the interstitial compartment. De-alone. For this reason, we used two other strategies inspite the decrease in tubular deposition, the transgenicaddition to C32/2 mice.animals still developed ARF.

Normal mice not exposed to LPS do have interstitialFinally, because C3 occupies a pivotal role early instaining for C3 at the basolateral aspects of cortical tu-both the classic and alternative pathways, the C32/2bules. It seems likely this can be attributed to C3 activa-mice have no complement cascade and therefore cannot

produce C3a, C5a, or the C5b-9 membrane attack com- tion as postulated by Nath, Hostetter, and Hostetter [17],

Cunningham et al: Complement-independent RF from endotoxin1586

as it is absent in C32/2 mice [31] and factor B and interspecies differences exist in rodents’ susceptibility toendotoxin, implying that interaction of endotoxin with aC2-deficient mice [37]. However, this degree of staining

increases significantly after LPS injection, consistent with variety of host factors is important. Thus, generalizationof these findings to humans should be made with caution.the hypothesis that increased transcription of C3 in the

renal cortex leads to increased local production and de- The phenomenon of ARF and the role of endotoxinin causing it remain poorly understood. It may be thatposition of C3. The cortical and tubular distribution of

IF staining for C3 protein matched that seen by in situ hemodynamic factors play a more prominent role thandirect tubular injury. Furthermore, tubular cell apoptosishybridization for C3 mRNA described in previous stud-

ies [12]. In addition, there was a trend toward increased also occurs in endotoxemia [42], which is consistent withthe modest degree of tubular necrosis that we have seencirculating C3 levels following LPS administration in all

animals as a group, as has previously been described in these studies [43]. Although LPS triggers local corticalsynthesis and deposition of complement, here we have[38]. Increased synthesis of C3 following LPS administra-

tion is not restricted to proximal tubular cells, but is shown conclusively that complement activation does notlead to the resultant ARF.instead a more general phenomenon, as shown by in-

creased C3 mRNA and protein levels in jejunal mucosaand liver [38]. ACKNOWLEDGMENTS

The appearance of C3 in the urine was also markedly This work was supported by National Institutes of Health grantsR01AI39246 to Dr. Carroll, R01AI31105 to Dr. Holers, and R01DK41873elevated in LPS-stimulated mice, but only in proportionand R01DK55357 to Dr. Quigg; a chapter grant from the Arthritisto urinary albumin. Additionally, FEC3 was consistentlyFoundation, Greater Chicago Chapter; and Biomedical Sciences

less in magnitude than FEAlbumin. This could be explained Grants from the National Arthritis Foundation to Drs. Holers andQuigg. Drs. Cunningham and Alexander were supported by Nationalby the fact that C3 is three times the size of albumin.Institutes of Health training grant T32DK07510. We thank Dr. Y.Whether urinary C3 is derived from glomerular filtrationNakagawa (Section of Nephrology, The University of Chicago, Chi-

and/or tubular synthesis cannot be determined from cago, IL, USA) for performing all BUN and creatinine determinations.these studies. The cause of albuminuria in this model of

Reprint requests to Richard J. Quigg, M.D., Section of Nephrology,ARF, not normally thought of as a proteinuric lesion,The University of Chicago, 5841 South Maryland Avenue, MC5100,

most likely reflects damage of tubular cells with failure Chicago, Illinois 60637, USA.E-mail: rquigg@medicine.bsd.uchicago.eduof normal albumin catabolism [39].

In the setting of endotoxemia, natural antibodies ap-pear to bind endotoxin and activate complement, facili- REFERENCEStating its clearance from the circulation via the reticulo- 1. Camussi G, Ronco C, Montrucchio G, Piccoli G: Role of soluble

mediators in sepsis and renal failure. Kidney Int 66(Suppl):S38–S42,endothelial system [40, 41]. Thus, C32/2 animals given1998high doses of LPS (approximately 10 times the dose used

2. Moll S, Schifferli JA, Huarte J, Lemoine R, Vassalli JD, Sap-in these studies) had a much greater mortality than wild- pino AP: LPS induces major changes in the extracellular proteolytic

balance in the murine kidney. Kidney Int 45:500–508, 1994type mice, which appeared to be due to unrestricted con-3. Wong W, Fearon DT: p65: A C3b-binding protein on murine cellssumption of C1 inhibitor [40]. A similar susceptibility oc-

that shares antigenic determinants with the human C3b receptorcurred in C42/2 animals, consistent with classic pathway (CR1) and is distinct from murine C3b receptor. J Immunol

134:4048–4056, 1985activation by natural antibodies bound to LPS. Thus, in4. Paul MS, Aegerter M, O’Brien SE, Kurtz CB, Weis JH: Thethese studies, it is conceivable that C32/2 mice had pro-

murine complement receptor gene family. I. Analysis of mCRYlonged exposure to LPS due to impaired clearance. How- gene products and their homology to human CR1. J Immunol

142:582–589, 1989ever, Crry transgenic mice do not have the marked abnor-5. Molina H, Wong W, Kinoshita T, Brenner C, Foley S, Holersmality in immune complex processing characteristic of

VM: Distinct receptor and regulatory properties of recombinantC32/2 and C42/2 mice [31] (and unpublished observa- mouse complement receptor 1 (CR1) and Crry, the two genetic

homologues of human CR1. J Exp Med 175:121–129, 1992tions), and thus are likely to clear LPS effectively.6. Kim Y-U, Kinoshita T, Molina H, Hourcade D, Seya T, WagnerThis disease model serves as a useful starting point

LM, Holers VM: Mouse complement regulatory protein Crry/p65for better understanding the pathogenesis of sepsis-asso- uses the specific mechanisms of both human decay-accelerating

factor and membrane cofactor protein. J Exp Med 181:151–159,ciated ARF in humans. As in human sepsis, the animals1995look systemically ill several hours after LPS administra-

7. Li B, Sallee C, Dehoff M, Foley S, Molina H, Holers VM:tion, with increasing mortality at higher doses. Neverthe- Mouse Crry/p65: Characterization of monoclonal antibodies and

the tissue distribution of a functional homologue of human MCPless, clinical septic shock in humans comprises a complexand DAF. J Immunol 151:4295–4305, 1993spectrum of illness, which is difficult to simulate in animal

8. Funabashi K, Okada N, Matsuo S, Yamamoto T, Morgan BP,models in every detail. True gram-negative septic shock Okada H: Tissue distribution of complement regulatory membrane

proteins in rats. Immunology 81:444–451, 1994in humans likely involves other inflammatory bacterial9. Couser WG, Johnson RJ, Young BA, Yeh CG, Toth CA, Ru-products besides LPS, released into the circulation in a

dolph AR: The effects of soluble recombinant complement recep-constant or intermittent pattern, often superimposed on tor 1 on complement-mediated experimental glomerulonephritis.

J Am Soc Nephrol 5:1888–1894, 1995other comorbid conditions. Considerable interstrain and

Cunningham et al: Complement-independent RF from endotoxin 1587

10. Quigg RJ, Lim A, He C, Berthiaume D, Alexander JJ, Kraus 28. Fischer MB, Ma M, Goerg S, Zhou X, Xia J, Finco O, Han S,Kelsoe G, Howard RG, Rothstein TL, Kremmer E, Rosen FS,D, Holers VM: Transgenic mice overexpressing the complement

inhibitor Crry as a soluble protein are protected from antibody- Carroll MC: Regulation of the B cell response to T-dependentantigens by classical pathway complement. J Immunol 157:549–556,induced glomerular injury. J Exp Med 188:1321–1331, 1998

11. Quigg RJ, Kozono Y, Berthiaume D, Lim A, Salant DJ, Wein- 199629. Wessels MR, Butko P, Ma M, Warren HB, Lage AL, Carrollfeld A, Griffin P, Kremmer E, Holers VM: Blockade of antibody-

induced glomerulonephritis with Crry-Ig, a soluble murine comple- MC: Studies of group B streptococcal infection in mice deficientin complement component C3 or C4 demonstrate an essential rolement inhibitor. J Immunol 160:4553–4560, 1998

12. Ault BH, Colten HR: Cellular specificity of murine renal C3 for complement in both innate and acquired immunity. Proc NatlAcad Sci USA 92:11490–11494, 1995expression in two models of inflammation. Immunology 81:655–

660, 1994 30. Quigg RJ, Alexander JJ, Lo CF, Lim A, He C, Holers VM:Characterization of C3-binding proteins on mouse neutrophils and13. Brooimans RA, Stegmann APA, Van Dorp WT, Van Der Ark

AAJ, Van Der Woude FJ, van Es LA, Daha MR: Interleukin 2 platelets. J Immunol 159:2438–2444, 199731. Quigg RJ, Lim A, Haas M, Alexander JJ, He C, Carroll MC:mediates stimulation of complement C3 biosynthesis in human

proximal tubular epithelial cells. J Clin Invest 88:379–384, 1991 Immune complex glomerulonephritis in C4- and C3-deficient mice.Kidney Int 53:320–330, 199814. Gerritsma JS, Van Kooten C, Gerritsen AF, Mommaas AM,

van Es LA, Daha MR: Production of inflammatory mediators and 32. Nomura A, Nishikawa K, Yuzawa Y, Okada H, Okada N, Mor-gan BP, Piddlesden SJ, Nadai M, Hasegawa T, Matsuo S: Tubu-cytokine responsiveness of an SV40-transformed human proximal

tubular epithelial cell line. Exp Nephrol 6:208–216, 1998 lointerstitial injury induced in rats by a monoclonal antibody whichinhibits function of a membrane inhibitor of complement. J Clin15. Andrews PA, Finn JE, Lloyd CM, Zhou W, Mathieson PW,

Sacks SH: Expression and tissue localization of donor-specific Invest 96:2348–2356, 199533. Quigg RJ, Abrahamson DR, Cybulsky AV, Badalamenti J,complement C3 synthesized in human renal allografts. Eur J Immu-

nol 25:1087–1093, 1995 Minto AWM, Salant DJ: Studies with antibodies to cultured ratglomerular epithelial cells: Subepithelial immune deposit forma-16. Tang S, Zhou W, Sheerin NS, Vaughan RW, Sacks SH: Contribu-

tion of renal secreted complement C3 to the circulating pool in tion after in vivo injection. Am J Pathol 134:1125–1133, 198934. Wang Y, Hu Q, Madri JA, Rollins SA, Chodera A, Matis LA:humans. J Immunol 162:4336–4341, 1999

17. Nath KA, Hostetter MK, Hostetter TH: Pathophysiology of Amelioration of lupus-like autoimmune disease in NZB/W F1 miceafter treatment with a blocking monoclonal antibody specific forchronic tubulo-interstitial disease in rats: Interactions of dietary

acid load, ammonia, and complement component C3. J Clin Invest complement component C5. Proc Natl Acad Sci USA 93:8563–8568,199676:667–675, 1985

18. Biancone L, David S, Della Pietra V, Montrucchio G, Cambi 35. Quigg RJ: We should inhibit complement in glomerular protein-uria. Kidney Int 56:2314–2315, 1999V, Camussi G: Alternative pathway activation of complement by

cultured human proximal tubular epithelial cells. Kidney Int 36. Steinman L: Some misconceptions about understanding autoim-munity through experiments with knockouts. J Exp Med 185:2039–45:451–460, 1994

19. David S, Biancone L, Caserta C, Bussolati B, Cambi V, Camussi 2041, 199737. Mitchell DA, Taylor PR, Cook HT, Moss J, Bygrave AE, Wal-G: Alternative pathway complement activation induces proin-

flammatory activity in human proximal tubular epithelial cells. port MJ, Botto M: Cutting edge: C1q protects against the develop-ment of glomerulonephritis independently of C3 activation. J Im-Nephrol Dial Transplant 12:51–56, 1997

20. Ichida S, Yuzawa Y, Okada H, Yoshioka K, Matsuo S: Localiza- munol 162:5676–5679, 199938. Wang Q, Meyer TA, Boyce ST, Wang JJ, Sun X, Tiao G, Fischertion of the complement regulatory proteins in the normal human

kidney. Kidney Int 46:89–96, 1994 JE, Hasselgren PO: Endotoxemia in mice stimulates productionof complement C3 and serum amyloid A in mucosa of small intes-21. Ong AC, Fine LG: Loss of glomerular function and tubulointersti-

tial fibrosis: Cause or effect? Kidney Int 45:345–351, 1994 tine. Am J Physiol 275:R1584–R1592, 199839. Maunsbach AB: Absorption of I-125-labeled homologous albu-22. Eddy AA, Michael AF: Acute tubulointerstitial nephritis associ-

ated with aminonucleoside nephrosis. Kidney Int 33:14–23, 1988 min by rat kidney proximal tubule cells: A study of microperfusedsingle proximal tubules by electron microscopic autoradiography23. Nomura A, Morita Y, Maruyama S, Hotta N, Nadai M, Wang

L, Hasegawa Matsuo S: Role of complement in acute tubulointer- and histochemistry. J Ultrastruct Res 15:197–241, 196640. Fischer MB, Prodeus AP, Nicholson-Weller A, Ma M, Murrowstitial injury of rats with aminonucleoside nephrosis. Am J Pathol

151:539–547, 1997 J, Reid RR, Warren HB, Lage AL, Moore FDJ, Rosen FS,Carroll MC: Increased susceptibility to endotoxin shock in com-24. Hori Y, Yamada K, Hanafusa N, Okuda T, Okada N, Miyata

T, Couser WG, Kurokawa K, Fujita T, Nangaku M: Crry, a plement C3- and C4-deficient mice is corrected by C1 inhibitorreplacement. J Immunol 159:976–982, 1997complement regulatory protein, modulates renal interstitial disease

induced by proteinuria. Kidney Int 56:2096–2106, 1999 41. Reid RR, Prodeus AP, Khan W, Hsu T, Rosen FS, CarrollMC: Endotoxin shock in antibody-deficient mice: Unraveling the25. Vakeva A, Meri S, Lehto T, Laurila P: Activation of the terminal

complement cascade in renal infarction. Kidney Int 47:918–926, role of natural antibody and complement in the clearance of lipo-polysaccharide. J Immunol 159:970–975, 19971995

26. Weiser MR, Williams JP, Moore FD Jr, Kobzik L, Ma M, Hecht- 42. Ortiz-Arduan A, Danoff TM, Kalluri R, Gonzalez C, KarpSL, Elkon K, Egido J, Neilson EG: Regulation of Fas and Fasman HB, Carroll MC: Reperfusion injury of ischemic skeletal

muscle is mediated by natural antibody and complement. J Exp ligand expression in cultured murine renal cells and in the kidneyduring endotoxemia. Am J Physiol 271:F1193–F1201, 1996Med 183:2343–2348, 1996

27. Smedegard G, Cui LX, Hugli TE: Endotoxin-induced shock in 43. Lieberthal W, Koh JS, Levine JS: Necrosis and apoptosis in acuterenal failure. Semin Nephrol 18:505–518, 1998the rat: A role for C5a. Am J Pathol 135:489–497, 1989