Maladaptive Role of IL-6 in Ischemic Acute Renal FailureMariusz L. Kielar,* Reji John,* Michael Bennett,† James A. Richardson,†‡ John M. Shelton,‡§

Liying Chen,* D. Rohan Jeyarajah,� Xin J. Zhou,† Hui Zhou,† Brett Chiquett,*Glenn T. Nagami,¶ and Christopher Y. Lu*#

Departments of *Internal Medicine (Nephrology), †Pathology, ‡Molecular Biology, §Internal Medicine (Cardiology), and�Surgery and #Graduate Program in Immunology, University of Texas Southwestern Medical Center, Dallas, Texas;and ¶Nephrology Section, Medical Research Services, Veterans Affairs Greater Los Angeles Healthcare System, LosAngeles, California

The role of IL-6 was investigated in murine ischemic acute renal failure. The renal pedicles were clamped for 17 min, and themice were studied at various times after reperfusion. We found that serum IL-6 increased after murine ischemic renal injury.This increase was associated with increased IL-6 mRNA in the ischemic kidney but not in the contralateral kidney or the liver.Maximal IL-6 production occurred at 4 to 8 h and decreased to baseline by 24 h. Reperfusion of the kidney was required forIL-6 production. In situ hybridization and immunohistochemistry showed that macrophages infiltrated areas adjacent to thevascular bundles in the outer medulla within hours of reperfusion and showed that these macrophages produced IL-6 mRNA.For understanding how macrophages were stimulated to produce IL-6, an in vitro model in which S3 proximal tubular cellswere injured by reactive oxygen species was set up. These injured cells released molecules that activated macrophages toproduce IL-6 in vitro. IL-6 that was produced in response to renal ischemia was maladaptive because transgenic knockout ofIL-6 ameliorated renal injury as measured by serum creatinine and histology. IL-6 transgenic knockout mice were lethallyirradiated, and their bone marrow was reconstituted with wild-type IL-6 cells. Such bone marrow transfers abolished theprotective effects of transgenic IL-6 knockout. It is concluded that macrophages infiltrate the area of the vascular bundles ofthe outer medulla, these macrophages produce IL-6, and this IL-6 exacerbates ischemic murine acute renal failure.

J Am Soc Nephrol 16: ???–???, 2005. doi: 10.1681/ASN.2003090757

A fter ischemia/reperfusion initiates injury to epithelialand vascular cells during ischemic acute renal failure(ARF), maladaptive responses “extend” the injury

(see reviews [1,2]). Inflammation is one maladaptive response(2–5), but the regulation of the inflammatory response to isch-emic renal injury is not well understood.

This report focuses on IL-6 because, as reviewed in the Dis-cussion section, this cytokine is a major regulator of inflamma-tion. Furthermore, IL-6 production may be a common feature ofischemic injury of any organ. IL-6 not only is found afterischemia of the brain (6), gut (7), and heart (8), but also theamount of IL-6 correlates with the amount of ischemic injury(9). In human renal allografts with ischemia-reperfusion injury,IL-6 is detected in urine, and its level correlates with the sever-ity of that injury (10).

The overall goal of this article is to understand the role of IL-6on ischemic acute renal failure (ARF). We make the followingpoints: First, IL-6 protein increases in the serum after ischemicrenal injury. This increase is associated with increased IL-6mRNA in the ischemic kidney. Second, in situ hybridization

and immunohistochemistry localize IL-6 production to macro-phages near the vascular bundles of the outer medulla. Mole-cules that are released by injured S3 proximal tubular cellsactivate macrophages to produce IL-6 in vitro. Third, transgenicknockout of IL-6 ameliorates renal injury as measured by serumcreatinine and histology. Finally, transfer of IL-6–sufficientmacrophages by means of bone marrow transplantation intoIL-6 knockout mice restores the susceptibility of the knockoutmice to ischemic renal injury.

Materials and MethodsAnimals and Surgical Protocols

Male C57Bl/6J [IL-6 (�/�)] and C57BL/6J-Il6tm1Kopf [IL-6 (�/�)]6-wk-old mice were purchased from Jackson Laboratories (Bar Harbor,ME). The genotype was confirmed by genomic PCR of tail snips (http://www.jax.org). The IL-6 (�/�) mice have a 174-bp DNA fragment fromthe wild-type allele, whereas homozygous IL-6 (�/�) mice have asingle 280-bp band as a result of insertion of a neocassette. Mice werehandled in accordance with institutional and National Institutes ofHealth guidelines.

Mice were anesthetized using inhaled isoflurane (Fortec System,Fraser Lake, NY) and maintained at 37°C by a TR100 temperaturecontrolling system with a rectal probe (Fine Science Tools, Foster City,CA). In most mice, first the right kidney was removed and the left renalartery and vein then were clamped for 17 min. In some mice, the leftrenal artery and vein were clamped, but the right (“contralateral”)kidney was not. Other mice had “sham” surgery: Laparotomy anddissection but not clamping of the left renal pedicle. Peripheral serum

Received September 14, 2003. Accepted August 18, 2005.

Published online ahead of print. Publication date available at www.jasn.org.

Address correspondence to: Dr. Christopher Y. Lu, Department of InternalMedicine (Nephrology), University of Texas Southwestern Medical School, 5323Harry Hines Boulevard, Dallas, TX 75390-8856. Phone: 214-648-3959; Fax: 214-648-2071; E-mail: christopher.lu@utsouthwestern.edu

Copyright © 2005 by the American Society of Nephrology ISSN: 1046-6673/1611-0001

. Published on September 28, 2005 as doi: 10.1681/ASN.2003090757JASN Express

was assayed for creatinine by the Refletron automated system (RocheDiagnostics, Indianapolis, IN) and IL-6 by ELISA (Endogen, Woburn,MA). This ELISA measures an active form of IL-6.

In some experiments, the cortex and outer medulla were isolated bydissection using a �3 to �30 operating microscope. These were ana-lyzed by RNase Protection Assay.

RNase Protection AssayTotal RNA was extracted from frozen kidneys using RNA-Easy Midi

Kits (cat. no. 75144; Qiagen, Santa Clara, CA). Total RNA from S3proximal tubular cells (see below) was harvested using the RNA-Easykit (Qiagen). The P32-labeled probes were made using In Vitro Tran-scription Kits, mCK-2b and mCK-3b templates, and Riboquant RPAkits from Pharmingen (Pasadena, CA). The RNase protection gels wereexposed on a phosphor image screen and analyzed with a MolecularDynamics Storm 820 Phosphorimager (Piscataway, NJ). Densitometryanalysis was performed using ImageQuant software (Molecular Dy-namics).

In Situ Hybridization for IL-6 mRNA and Immunohistologyfor F4/80 Macrophage

Ischemic and “contralateral” kidneys were harvested 4 h after reper-fusion and fixed in 4% buffered paraformaldehyde in PBS, embeddedin paraffin, cut into 3-�m sections, and used for in situ hybridization ormacrophage immunostaining. The IL-6 fragment for the in situ probewas amplified from an ischemic kidney cDNA library (11). The detailsof probe preparation and in situ hybridization were published previ-ously by our group (12). To stain with F4/80 antibody, the kidneysections were deparaffinized, blocked with Protein Block Serum (cat.no. 0909; DAKO Labs, Glostrup, Denmark), and incubated overnight at4°C with 1:200 F4/80 antibody (cat. no. RM2900; Caltag Laboratories,San Francisco, CA). Subsequently, sections were incubated with biotin-ylated rabbit anti-rat IgG from DAKO (cat. no. DK-2600) at 1:50, thenwith streptavidin, horseradish peroxidase, and finally with DAB as perthe manufacturer’s instruction (cat. no. K0377; DAKO Labs).

In Vitro Model of Ischemic ARFWe developed a two-stage assay using the S3 cell line. This cell line

was originally dissected from the S3 segment of the proximal tubule ofthe kidney of an SV40 large T antigen transgenic mouse (13). In stage 1,S3 tubules were injured by reactive oxygen species (ROS) for 12 h; theROS were generated by the action of reaction of 0.01 U/ml xanthineoxidase on 5 mM hypoxanthine (both from Sigma Chemical Co., St.Louis, MO), as previously reported by our laboratory (12). In stage 2,the supernatant was cultured with a murine macrophage line (J774),and IL-6 mRNA was measured 4 h later by RNase protection assay(Pharmingen).

In Vivo Injections of Anti–IL-6IL-6–sufficient (C57Bl6/J) mice received 1 dose of monoclonal rat

anti–IL-6 antibody (MP5-20F3 cat. no. 16-7061; eBioscience, www.ebio-science.com) in the amount of 72 �g per mouse. The control groupreceived an equivalent dose of rat IgG1� (cat 16-4301). The antibodieswere suspended in 0.2 ml of 0.1% BSA and injected into the penile veinimmediately before the nephrectomy/clamp procedure. Peripheral se-rum was assayed for creatinine at 24 h after reperfusion. In some mice,peripheral blood was obtained at 4 h after reperfusion and used formeasurement of IL-6 levels by the ELISA method (Endogen).

In addition, we used polyclonal goat neutralizing anti-murine IL-6(AB-406-NA) and control goat IgG (AB-108-C) from R&D systems(www.rndsystems.com). These antibody preparations had �0.01 EU

per 1 �g of antibody. According to the manufacturer’s specifications,3.2 ng of antibody will neutralize 1 ng of IL-6. According to Figure 1,the IL-6 concentration at 4 h after ischemia is 2500 ng/ml. If the mouseweighs 20 g and the volume of distribution of IL-6 is 50%, then the totalIL-6 in the mouse is 25,000 ng and the dose of antibody to neutralize is80 �g. We administered a dose of 125 �g.

Histologic ExaminationThree IL-6 (�/�) and three IL-6 (�/�) mice were killed for mor-

phologic studies at 24 h after reperfusion. The kidneys were fixed in10% buffered formalin, embedded in paraffin, cut into 5-�m sections,and stained with hematoxylin and eosin. The morphologic analysis wascarried out in a blinded manner as detailed previously (14). Briefly, thecortex and outer stripe of the outer medulla were evaluated for epithe-

Figure 1. Ischemic acute renal failure (ARF) increased serumIL-6 protein. Serum IL-6 was measured by ELISA (see Materialsand Methods) in three different groups of mice: (1) ShamC57BL/6J mice (f) had their left renal arteries and veins dis-sected free of surrounding fat, but these vessels were notclamped. (2) Clamp-only mice (u) had their left renal arteriesand veins dissected and also clamped for 17 min; this resultedin an ischemic left kidney and a contralateral nonischemic rightkidney. The contralateral kidney maintained a normal GFR;both the clamp-only and sham mice had serum creatinines of�0.5 mg/dl. These mice had renal ischemia but no uremia. (3)Clamp � nephrectomy mice (�) had their left renal arteryclamped like the above group but, in addition, had their rightkidney removed. These mice had serum creatinines of 1.9mg/dl. They had renal ischemia and uremia. Peripheral bloodwas taken from the sham group at 4 h after surgery and fromthe clamp and the clamp � nephrectomy groups at the indi-cated times of reperfusion. The means and SE are shown. *P �0.01 versus control by t test. There was no statistically significantdifference between the clamp-only and the clamp � nephrec-tomy groups at 4 h of reperfusion.

2 Journal of the American Society of Nephrology J Am Soc Nephrol 16: ???–???, 2005

lial necrosis, loss of brush border, tubular dilation, and cast formation.The kidney sections were scored on the basis of the percentage ofaffected tubules as follows: 0, none; 1, �10%; 2, 11 to 25%; 3, 26 to 50%;4, 51 to 75%; 5, �75%. At least 10 high-power fields (�400) werereviewed for each slide. In addition, leukocyte infiltration in the outerstripe of the outer medulla was counted on hematoxylin and eosin–stained sections. The number of leukocytes were averaged on the basisof at least 5 high-power fields for each slide.

Bone Marrow TransplantBone marrow cells were isolated from femurs and tibias, filtered

through nylon mesh, counted using an electronic particle counter, andwashed, and 8 � 106 cells in 0.5 ml of PBS were injected intravenouslyinto recipients within 6 h of their receiving two doses of 5 Gy separatedby 3 h. We transplanted IL-6 �/� bone marrow into IL-6 �/�, or viceversa. The mice were kept in a sterile environment for 8 wk to allow fullcellular reconstitution. Full chimerism of each mouse was confirmed by

Figure 2. Reperfusion of the ischemic kidney increased IL-6 mRNA in the injured kidney. Ischemic, contralateral, and shamkidneys are defined in the legend to Figure 1. (A) IL-6 mRNA is increased in the ischemic kidney but not in the liver. RNaseprotection assays for IL-6 mRNA were performed on the total renal RNA at the indicated reperfusion times. In addition, IL-6mRNA was assayed from livers that were isolated from a mouse with a sham kidney (a), or from mice with ischemic kidneys atthe following renal reperfusion times: 1 h (b), 2 h (c), 8 h (d), and 24 h (e). L32 and glyceraldehyde-3-phosphate dehydrogenase(GAPDH) are housekeeping genes. (B) IL-6 mRNA is increased in the ischemic but not the contralateral kidney. Total RNA wastaken at the indicated reperfusion times from sham, ischemic, and contralateral kidneys; the indicated cytokines were assayed byRNase protection assays. (C) Reperfusion is required for increased renal IL-6 mRNA. Total RNA was harvested from a sham(control) kidney, from an ischemic kidney without releasing the clamp on the renal artery (no reperfusion), and either 1 or 4 h afterrelease of the clamp (reperfusion). (D) IL-6 mRNA is found in the ischemic outer medulla. Total RNA was isolated from the cortex(C) or outer medulla (M) and was analyzed by RNase protection assay.

J Am Soc Nephrol 16: ???–???, 2005 IL-6 in Ischemic Acute Renal Failure 3

genotyping of DNA from peripheral blood and tails using Jackson Labprotocol (see above). Renal ischemia reperfusion was induced as de-scribed above.

ResultsIschemic ARF Increased Serum IL-6 Protein and IL-6mRNA in the Ischemic Kidney

Figure 1 shows that mice with ischemic kidneys (“clamponly” and “clamp � nephrectomy”) had increased serum IL-6protein at 4 h after reperfusion compared with mice with non-ischemic sham-operated kidneys. The ischemic left kidney hadsimilarly increased serum IL-6, in the presence or absence of afunctioning right kidney. This finding showed that the in-creased IL-6 was not due to decreased renal elimination by theischemic left kidney and suggested hepatic, rather than renal,elimination of IL-6 (15).

Figure 2A shows that the increased serum IL-6 originated inthe ischemic kidney, not the liver. In these experiments, liverand a kidney were taken from the same mouse. Minimalamounts of IL-6 mRNA were present in the sham-operatedkidney and the associated liver. In the ischemic kidney, IL-6mRNA abundance increased until 8 h of reperfusion and de-creased at 24 h of reperfusion; little IL-6 mRNA was present inthe livers of these mice.

Figure 2B compares IL-6 mRNA abundance in sham-oper-ated kidneys and ischemic left versus contralateral right non-ischemic kidneys at various times after reperfusion. IL-6 mRNAincreased in the ischemic kidney at 1 h, peaked at 4 to 8 h, anddecreased by 24 h. The IL-6 mRNA did not increase in thecontralateral or in the sham kidneys. These data confirm pre-vious observations (16). Figure 3 summarizes RNase protectionassays, quantified by densitometry, on four ischemic versus foursham kidneys. The IL-6 mRNA was significantly increased at 1and 4 h and then decreased to baseline at 72 h.

Figure 2C compares IL-6 mRNA abundance in ischemic kid-

neys before release of the renal arterial clamp (“ischemia, noreperfusion”) with ischemic kidneys after 1 or 4 h of reperfu-sion. IL-6 mRNA abundance did not increase unless the kidneywas reperfused.

In addition to IL-6, the Pharmingen RNase protection assaysprovided information about other cytokines. These were notthe focus of our studies, but we comment on them briefly.Figure 2B shows that mRNA for IL-1�, IL-1Ra, and IL-18 in-creased in the ischemic kidney and to a lesser extent in thecontralateral kidney. These molecules are expressed by theischemic kidney (16,17). However, expression by the contralat-eral kidney to our knowledge has not been previously reported.Such expression is consistent with the relatively minor inflam-mation there and may result from the systemic release of cyto-kines from the ischemic kidney (18). In addition, Figure 2Dshows increased leukemia inhibitory factor expression in theischemic kidney and confirms a previous report (16).

Macrophages in the Ischemic Outer Medulla ExpressIL-6 mRNA

To determine which part of the ischemic kidney expressedIL-6 mRNA, we dissected the cortex and the outer medulla.Figure 2D shows that the greatest increase in IL-6 mRNA at 4and 8 h was in the ischemic outer medulla.

To identify the cell population expressing IL-6 mRNA in theischemic kidney, we performed in situ hybridization. We se-lected the 4-h reperfusion time point on the basis of the timecourse for IL-6 mRNA expression shown in Figure 2. Low-power darkfield photomicrographs localize IL-6 expression tothe ischemic outer medulla: Figure 4A shows that the S35-labeled antisense IL-6 mRNA bound to IL-6 mRNA in the outermedulla of the ischemic kidney, Figure 4B shows absent stain-ing of the ischemic kidney by control S35-labeled sense IL-6mRNA, and Figure 4C shows absent staining of the contralat-eral nonischemic kidney by S35-labeled antisense IL-6 mRNA.

Figure 3. Increased IL-6 mRNA abundance in the ischemic kidney after reperfusion: Composite of four experiments. L32 is ahousekeeping gene. The means and SE of four kidneys per group are shown. By t test: *P � 0.04 sham versus ischemic 1-hreperfusion; **P � 0.01 sham versus ischemic 4-h reperfusion.

4 Journal of the American Society of Nephrology J Am Soc Nephrol 16: ???–???, 2005

Higher power views show that the IL-6–expressing cells aremononuclear cells and that the mononuclear cells are locatedadjacent to the vascular bundles. Figure 4D is a medium-power

brightfield photomicrograph of the ischemic outer medulla.Four mononuclear cells that express IL-6 are designated byarrows and outlined in boxes. The neighboring vascular bun-

Figure 4. IL-6 expression by interstitial macrophages near the vascular bundles of the ischemic outer medulla at 4 h of reperfusion.(A through C) These are darkfield, low-magnification photomicrographs. (A) Main panel; shows IL-6 mRNA in the outer medullaof an ischemic kidney at 4 h after reperfusion. The coarse white dots denote silver grains of antisense S35-labeled IL-6 mRNA. (B)Control ischemic kidney with in situ hybridization with a S35-labeled sense IL-6 probe. As expected, this sense probe does nothybridize to the ischemic kidney. (C) Another control: In situ hybridization of a nonischemic contralateral kidney with an antisenseS35-labeled IL-6 probe. There was minimal IL-6 mRNA in the nonischemic kidney. (D) A �40 objective brightfield of the outermedulla of the ischemic kidney shown in A. The dark dots denote S35-labeled antisense IL-6 mRNA. The arrows (both black andred) point at boxes that outline macrophages that express IL-6 mRNA. VB, vascular bundle. (E) High-magnification (�80 objective)brightfield photomicrograph of the two macrophages designated by red arrows in D.

J Am Soc Nephrol 16: ???–???, 2005 IL-6 in Ischemic Acute Renal Failure 5

dles are designated VB. Figure 4E is a high-power view of thetwo mononuclear cells designated by red arrows in Figure 4D.

We immunostained some of the above ischemic kidney sec-tions with F4/80, an mAb that detects murine macrophages(19). There were no detectable F4/80 macrophages in nonisch-emic kidneys (Figure 5, top). After 4 h of reperfusion, F4/80macrophages were found adjacent to the vascular bundles ofthe outer medulla (Figure 5, bottom).

Macrophages Express IL-6 in Response to MoleculesReleased by S3 Proximal Tubular Cells In Vitro

Molecules that are released by injured cells activate macro-phages in vitro (see reviews [20]). To determine whether thisoccurs in the ischemic kidney, we injured S3 renal tubular cellswith ROS in vitro (Figure 6). Such ROS are produced duringischemic ARF (21), and ROS injury of renal tubular cells in vitrohas previously been used to model ischemic ARF in vitro (e.g.,22). The injured renal tubular cells released molecules into thesupernatant that activated macrophages to express IL-6 mRNA.Control experiments established that resting renal tubular cells

did not release activating molecules (Figure 6, Grp C). Grp Dwas ROS and medium incubated overnight in the absence of S3tubules. The ROS of Grp D were so unstable that they degradedduring the overnight incubation and were not a factor in stage 2.

Altogether, these in vitro experiments support the in vivoexperiments of Figures 4 and 5 that show that renal macro-phages express IL-6 in response to ischemic ARF. Further sup-port for macrophages as the source of IL-6 in ischemic kidneyswill be provided when we discuss our bone marrow chimeraexperiments at the end of the Results section.

Removal of IL-6 Ameliorates Ischemic ARFTo determine whether IL-6 exacerbates ischemic ARF, we

injected two different anti–IL-6 antibodies intravenously at thetime of renal ischemia and measured the effect on renal injury.First, we injected monoclonal rat anti–IL-6 (MP5–20F; eBio-science) intravenously at a dose seven times larger than thatpreviously used successfully to neutralize IL-6 and amelioratemurine septic shock (23). Control mice received an equivalentamount of rat IgG. The anti–IL-6 decreased the peripheralblood IL-6 at 4 h of reperfusion from 7000 � 1000 ng/ml in therat IgG–injected group to 1600 � 600 ng/ml in the anti–IL-6–injected group (mean � SE; n � 6; P � 0.01 by t test). Despitethis 77% decrease in peripheral blood IL-6, there was no sig-nificant effect on ischemic ARF; the serum creatinine in the ratIgG–treated group was 0.8 � 0.2 mg/dl and 0.7 � 0.1 mg/dl inthe anti–IL-6–treated group (mean � SE; n � 6; P � 0.38 by ttest). In a second experiment, we injected 125 �g of neutralizingpolyclonal goat anti-murine IL-6 (AB-108-C; R&D Systems).This antibody preparation contained �0.01 EU of endotoxinper 1 �g of antibody. We chose this dose because it exceeds thatnecessary to neutralize the 2500 ng/ml of IL-6 found in blood at4 h after ischemia (see Figure 1 and Materials and Methods).We increased the clamp time to induce greater injury in theseexperiments. The serum creatinine was 1.6 � 0.3 mg/dl in thegoat control IgG-injected mice and 1.5 � 0.4 mg/dl in theanti–IL-6–injected mice. Again, the anti–IL-6 did not ameliorateischemic ARF. The inability of anti–IL-6 to ameliorate ischemicARF most likely represents an inability of sufficient quantitiesof the large antibody molecule (molecular weight � 160 k) toaccess the outer medulla, where the IL-6 is located.

To determine further the role of IL-6 in ischemic ARF, weexamined the effect of transgenic knockout of IL-6 as assessedby both function (Figure 7) and morphology (Figure 8). Weused C57BL/6J-Il6tm1Kopf [IL-6 (�/�)] and C57Bl/6J wild-type[IL-6 (�/�)] mice. These mice are the same except for thenonfunctional IL-6 gene in the former. Figure 7 compares renalfunction during ischemic ARF. The serum creatinine (Scr) wasdetermined at 24 h after reperfusion. The mean serum creati-nine level in the IL-6 (�/�) group was 0.89 � 0.136 mg/dlversus 1.83 � 0.1 mg/dl in the wild-type group (mean and SE ofat least eight mice per group; P � 0.05).

Figure 8, A and B, are low-power photomicrographs thatcompare the pathology of the IL-6 (�/�) versus IL-6 (�/�)kidneys at 24 h of reperfusion. The IL-6 (�/�) kidneys hadsevere injury (Figure 8B). There were many necrotic tubules inthe outer medulla, indicated by *. These tubules were filled

Figure 5. Macrophages are found adjacent to the ischemic vas-cular bundles of the outer medulla. F4/80-stained macrophagesare found near the vascular bundles in the outer medulla ofischemic (bottom) but not contralateral nonischemic (top) kid-neys at 4 h of reperfusion. nt, necrotic tubules; c, casts. Whitearrows indicate a few of many F4/80� (brown) macrophages.

6 Journal of the American Society of Nephrology J Am Soc Nephrol 16: ???–???, 2005

with casts. There was less but still significant injury in thecortex. There were many neutrophils in the interstitial spaces,indicated by arrows. In comparison, the IL-6 (�/�) kidneys(Figure 8A) had much less injury by all of the above criteria.

Figure 8, C and D, are high-power photomicrographs of theouter medulla of the ischemic IL-6 (�/�) versus IL-6 (�/�)kidneys. Necrotic tubules and inflammation are present in theIL-6 (�/�) kidneys. The morphometric analysis (Figure 9)confirmed the decreased medullary and cortical damage andthe decreased inflammation in the IL-6 (�/�) kidneys.

Production of IL-6 by Bone Marrow–Derived Cells IncreasesIschemic ARF

Figure 10 shows that chimeric mice with IL-6 (�/�) renalparenchymal cells and IL-6 (�/�) macrophages have greaterischemic ARF than chimeric mice with IL-6 (�/�) renal paren-chymal cells and IL-6 (�/�) bone marrow–derived cells. Inthese experiments, C57BL/6J-Il6tm1Kopf [IL-6 (�/�)] mice werelethally irradiated and then received bone marrow transplants

from C57BL/6J [IL-6 (�/�)] mice. Similarly, lethally irradiatedIL-6 (�/�) mice received IL-6 (�/�) bone marrow. Becausethese mice differed only in the IL-6 gene, there was no graft-versus-host or host-versus-graft disease. Genomic PCR (see Ma-terials and Methods) of the radioresistant tail and radiosensi-tive peripheral blood confirmed the chimerism.

DiscussionThis article makes several points. IL-6 is produced by the

ischemic kidney. Immunohistology and in situ hybridizationshow that macrophages adjacent to the ischemic vascular bun-dles of the outer medulla produce IL-6. An in vitro model ofischemic ARF suggests that this production is stimulated bymolecules that are produced by ischemic renal tubular cells.Transgenic knockout of IL-6 ameliorates renal injury. In chi-meric mice whose renal parenchymal cells and macrophageswere or were not capable of producing IL-6, maximal ischemicinjury required IL-6–producing macrophages.

Figure 6. In an in vitro model of ischemic ARF, macrophages express IL-6 in response to molecules released by S3 proximal tubularcells. (A) In stage 1, SV40-transformed S3 tubules were injured by reactive oxygen species (ROS); they then released putativemacrophage-activating molecules into the supernatant for 12 h. In stage 2, the supernatant was cultured with a murinemacrophage line (J774), and the IL-6 mRNA was measured 4 h later by RNase protection assay (Pharmingen). The positive control“B” was endotoxin added directly to the macrophages. The negative controls were “C,” supernatant conditioned by S3 tubules inthe absence of ROS, and “D,” ROS and media incubated overnight in the absence of S3 tubules. (Top) The experimental designand the means and SE of four experiments. *P � 0.01 versus media control. The y axis is the densitometry (ratio of IL-6:L32housekeeping gene expression). (B) An RNase protection assay from one of the four experiments.

J Am Soc Nephrol 16: ???–???, 2005 IL-6 in Ischemic Acute Renal Failure 7

Our transgenic mice carry the caveat of all such experiments,i.e., mice with transgenic knockout of IL-6 may produce mole-cules that compensate for the absence of IL-6 from birth. It ispossible that the lesser ischemic ARF of the knockout mice isdue to these compensatory molecules and not the deficiency ofIL-6. One way to overcome this caveat is by inhibiting IL-6 inwild-type mice. Although others ameliorated ischemic ARFwith anti–IL-6 injections (24), we found that exogenous anti–IL-6 did not ameliorate ischemic ARF. Our antibody experi-ments are inconclusive because the antibodies may not accessthe interstitial spaces of the injured kidney. We have overcomethis caveat using bone marrow chimeras. These experimentsdemonstrate a specific effect of wild-type IL-6–producing mac-rophages. Thus, wild-type IL-6–producing macrophages in theIL-6 knockout kidneys produce more ischemic ARF than IL-6knockout macrophages in wild-type kidneys (Figure 10). Inother words, bone marrow transfer of IL-6–producing wild-

Figure 7. Transgenic knockout of IL-6 ameliorated ischemicARF: Renal function. We removed the right kidney andclamped the left renal pedicle for 17 min. The serum creatininewas measured after 24 h of reperfusion. f, homozygous IL-6knockout [IL-6 (�/�)] mice; s, wild-type mice [IL-6 (�/�)].The means and SE are shown.

Figure 8. Transgenic knockout of IL-6 ameliorated ischemic ARF: Pathology. Kidneys were removed after 24 h of reperfusion, fixedin formalin, sectioned, and stained with hematoxylin and eosin. (A) IL-6 (�/�). (B)IL-6 (�/�). (C) IL-6 (�/�) view of the outermedulla. (D) IL-6 (�/�) view of the outer medulla. G, glomeruli; *, necrotic tubules. Arrows show neutrophils. Magnification,�200 in A and B; �400 in C and D.

8 Journal of the American Society of Nephrology J Am Soc Nephrol 16: ???–???, 2005

type macrophages overcame the protective effect of IL-6 knock-out to ischemic renal injury.

In addition to our group, many other laboratories (25–27)found macrophages in the outer medulla early during ischemicARF. Preventing the macrophage infiltration ameliorates renalinjury (26,28,29). However, how macrophages injure the kidneywas not known. We now suggest that one mechanism is theproduction of IL-6.

The maladaptive renal effects of IL-6 remain to be elucidatedin detail. We favor the following hypothesis. Macrophages area major component of the initial inflammatory response to renalinjury, as shown in Figure 5 and previously reported by others(26,27). These macrophages release IL-6, which further in-creases renal inflammation by recruiting more neutrophils intothe injured kidney. This hypothesis is consistent with our ob-servations (Figures 8 and 9), which show decreased inflamma-tion in the kidney at 24 h in the IL-6 (�/�) ischemic kidney.This hypothesis is also consistent with previous reports that thepeak neutrophilic infiltrate after renal ischemia does not occuruntil 12 h (30,31), well after we and others (26) found IL-6–expressing renal macrophages (Figures 4 and 5).

Also consistent with this hypothesis are the known proin-flammatory effects of IL-6. These include stimulation of neu-trophil release from the bone marrow (32), prevention of neu-trophil apoptosis (33), activation of neutrophils to produce toxicenzymes (34), and activation of endothelial cells to expressintercellular adhesion molecule 1 and chemokines (35,36).These effects of IL-6 are confirmed by the decreased inflamma-tory response in transgenic knockout mice after ischemia in thelung (37), gut (7,38), or brain (6); after injections of sterileturpentine (39) or carrageen (35); and after infections. (40,41).Further support of the proinflammatory effects of IL-6 are theincreased isletitis in mice overexpressing IL-6 driven by theinsulin promoter (42) and the anti-inflammatory effects of re-

moving IL-6 in experimental (43) and clinical arthritis (44).Furthermore, IL-6 contributes to neutrophilic influx and in-creased damage in the models of hemorrhagic shock (45) andspinal cord injury (46).

Another issue raised by our experiments is what activates themacrophages after they have entered the kidney. Figure 6shows that injured renal tubules release molecules that activatemacrophages in vitro. Ongoing experiments in our laboratoryaim at identifying these molecules. One possibility is that mac-rophages are stimulated by molecules that ordinarily resideinside renal tubular cells but are released into the extracellularspace after these cells are injured. Heat-shock proteins areexamples of such molecules. Extracellular heat-shock proteinsdo activate macrophages via their Toll-like receptor 4 (TLR4)receptor (20,47–49). Consistent with the idea that TLR4 partic-ipates in ischemic ARF are our data that TLR4-deficient mice

Figure 9. Transgenic knockout of IL-6 ameliorated ischemicARF: Morphometrics. Morphometric comparison of ischemickidneys at 24 h of reperfusion; f, IL-6 (�/�); u, IL-6 (�/�).The means and SE of three kidneys in each group are shown. *P� 0.015 between wild-type and IL-6 knockout groups by t test.Morphometry (see Materials and Methods) was performed ac-cording to the method of Thurman et al. (14).

Figure 10. Production of IL-6 by bone marrow–derived cellsincreases ischemic ARF. Lethally irradiated IL-6 (�/�) micereceived IL-6 (�/�) bone marrow; lethally irradiated IL-6(�/�) mice received IL-6 (�/�) bone marrow. The IL-6 (�/�)and IL-6 (�/�) mice differed only at the IL-6 gene (see Mate-rials and Methods). Top panels are representative genomic PCRthat demonstrate full replacement of hematopoietic cells (in-cluding macrophages) in the chimeric mice. Genomic DNA wasobtained from the tails (recipient radioresistant cells) and blood(donor hematopoietic cells); we used primers specific for theIL-6 gene [IL-6 (�/�)] and for the neocassette in the IL-6knockout [IL-6 (�/�)] mice. All mice underwent right ne-phrectomy and left renal pedicle clamp for 17 min. The serumcreatinine was measured after 24 h of reperfusion. f, IL-6(�/�) bone marrow to IL-6 (�/�) recipient mice; �, IL-6(�/�) bone marrow to IL-6 (�/�) recipient mice. The meansand SE are shown.

J Am Soc Nephrol 16: ???–???, 2005 IL-6 in Ischemic Acute Renal Failure 9

suffer less injury after renal ischemia (R.J. and C.Y.L, unpub-lished observations, 2005).

Although our data and most reports in the literature, re-viewed in the beginning of this article, indicate that IL-6 exac-erbates ischemic injury, no fair discussion would be completewithout acknowledging that, in a few models, IL-6 amelioratesrather than exacerbates injury (37,50). Why IL-6 exacerbatesinjury in most models but ameliorates injury in a few modelsremains to be explained.

In summary, this article makes the following major points:First, IL-6 protein increases in the serum after ischemic renalinjury. This increase is associated with increased IL-6 mRNA inthe ischemic kidney. Second, in situ hybridization and immu-nohistochemistry localize IL-6 production to macrophages nearthe vascular bundles of the outer medulla. Molecules that arereleased by injured S3 proximal tubular cells activate macro-phages to produce IL-6 in vitro. Third, transgenic knockout ofIL-6 ameliorates renal injury as measured by serum creatinineand histology. Finally, transfer of IL-6–sufficient macrophagesby means of bone marrow transplantation into IL-6 knockoutmice increases ischemic renal injury.

AcknowledgmentsM.L.K. is supported by a KO 8 grant from the National Institutes of

Health. C.Y.L. is supported by RO-1 and R21 grants from the NationalInstitutes of Health.

We thank Kathy Trueman for secretarial assistance, as well as NicoleFranz, Vipin Bhagat, and Vanessa Woodward for technical assistance.

References1. Molitoris BA: Transitioning to therapy in ischemic acute

renal failure. J Am Soc Nephrol 14: 265–267, 20032. Bonventre JV, Weinberg JM: Recent advances in the patho-

physiology of ischemic acute renal failure. J Am Soc Nephrol14: 2199–2210, 2003

3. Rabb H, O’Meara YM, Maderna P, Coleman P, Brady HR:Leukocytes, cell adhesion molecules and ischemic acuterenal failure. Kidney Int 51: 1463–1468, 1997

4. Deng J, Hu X, Hewitt S, Miyaji T, Wang Y, Li S, Nibhanu-pudy B, Star RA: Interleukin 10 inhibits acute renal injuryfollowing cisplatin administration and ischemia [Abstract].J Am Soc Nephrol 12: A4183, 2001

5. Chiao H, Kohda Y, McLeroy P, Craig L, Housini I, Star RA:Alpha-melanocyte-stimulating hormone protects againstrenal injury after ischemia in mice and rats. J Clin Invest 99:1165–1172, 1997

6. Clark WM, Rinker LG, Lessov NS, Hazel K, Hill JK, Stenzel-Poore M, Eckenstein F: Lack of interleukin-6 expression isnot protective against focal central nervous system isch-emia. Stroke 31: 1715–1720, 2000

7. Yang R, Han X, Uchiyama T, Watkins SK, Yaguchi A,Delude RL, Fink MP: IL-6 is essential for development ofgut barrier dysfunction after hemorrhagic shock and resus-citation in mice. Am J Physiol Gastrointest Liver Physiol 285:G621–G629, 2003

8. Kukielka GL, Youker KA, Michael LH, Kumar AG, Ballan-tyne CM, Smith CW, Entman ML: Role of early reperfusionin the induction of adhesion molecules and cytokines in

previously ischemic myocardium. Mol Cell Biochem 147:5–12, 1995

9. Vila N, Castillo J, Davalos A, Esteve A, Planas AM,Chamorro A: Levels of anti-inflammatory cytokines andneurological worsening in acute ischemic stroke. Stroke 34:671–675, 2003

10. Kwon O, Molitoris BA, Pescovitz M, Kelly KJ: Urinaryactin, interleukin-6, and interleukin-8 may predict sus-tained ARF after ischemic injury in renal allografts. Am JKidney Dis 41: 1074–1087, 2003

11. Fee D, Grzybicki D, Dobbs M, Ihyer S, Clotfelter J,Macvilay S, Hart MN, Sandor M, Fabry Z: Interleukin 6promotes vasculogenesis of murine brain microvessel en-dothelial cells. Cytokine 12: 655–665, 2000

12. Zhang Y, Woodward VK, Shelton JM, Richardson JA, LinkDC, Zhou XJ, Kielar ML, Jeyarajah DR, Lu CY: Ischemia/reperfusion induces G-CSF gene expression by renal med-ullary thick ascending limb cells in vivo and in vitro. Am JPhysiol Renal Physiol 286: F1193–F1201, 2004

13. Kauntiz JD, Smith Cummins VP, Misler D, Nagami GT:Inhibition of gentamicin uptake into cultured mouse prox-imal tubule epithelial cells by L-lysine. J Clin Pharmacol 33:63–69, 1993

14. Thurman JM, Ljubanovic D, Edelstein CL, Gilkeson GS,Holers VM: Lack of a functional alternative complementpathway ameliorates ischemic acute renal failure in mice.J Immunol 170: 1517–1523, 2003

15. Castell JV, Geiger T, Gross V, Andus T, Walter E, Hirano T,Kishimoto T, Heinrich PC: Plasma clearance, organ distri-bution and target cells of interleukin-6/hepatocyte-stimu-lating factor in the rat. Eur J Biochem 177: 357–361, 1988

16. Lemay S, Rabb H, Postler G, Singh AK: Prominent andsustained up-regulation of gp130-signaling cytokines andof the chemokine MIP 2 in murine renal ischemia-reperfu-sion injury. Transplantation 69: 959–963, 2000

17. Melnikov VY, Ecder T, Fantuzzi G, Siegmund B, Lucia MS,Dinarello CA, Schrier RW, Edelstein CL: Impaired IL-18processing protects caspase-1-deficient mice from ischemicacute renal failure. J Clin Invest 107: 1145–1152, 2001

18. Meldrum KK, Meldrum DR, Meng X, Ao L, Harken AH:TNF-alpha-dependent bilateral renal injury is induced byunilateral renal ischemia-reperfusion. Am J Physiol HeartCirc Physiol 282: H540–H546, 2002

19. Hume DA, Perry VH, Gordon S: The mononuclear phago-cyte system of the mouse defined by immunohistochemicallocalisation of antigen F4/80: Macrophages associated withepithelia. Anat Rec 210: 503–512, 1984

20. Seong SY, Matzinger P: Hydrophobicity: An ancient dam-age-associated molecular pattern that initiates innate im-mune responses. Nat Rev Immunol 4: 469–478, 2004

21. Nath KA, Norby SM: Reactive oxygen species and acuterenal failure. Am J Med 109: 665–678, 2000

22. di Mari JF, Davis R, Safirstein RL: MAPK activation deter-mines renal epithelial cell survival during oxidative injury.Am J Physiol 277: F195–F203, 1999

23. Gennari R, Alexander JW, Pyles T, Hartmann S, Ogle CK:Effects of antimurine interleukin-6 on bacterial transloca-tion during gut-derived sepsis. Arch Surg 129: 1191–1197,1994

24. Patel NS, Chatterjee PK, Di Paola R, Mazzon E, Britti D, DeSarro A, Cuzzocrea S, Thiemermann C: Endogenous inter-leukin-6 enhances the renal injury, dysfunction, and in-

10 Journal of the American Society of Nephrology J Am Soc Nephrol 16: ???–???, 2005

flammation caused by ischemia/reperfusion. J PharmacolExp Ther 312: 1170–1178, 2005

25. Persy VP, Verhulst A, Ysebaert DK, De Greef KE, De BroeME: Reduced postischemic macrophage infiltration andinterstitial fibrosis in osteopontin knockout mice. KidneyInt 63: 543–553, 2003

26. De Greef KE, Ysebaert DK, Dauwe SE, Persy VP, Vercau-teren S, Mey D, De Broe ME: Anti-B7–1 blocks mononu-clear cell adherence in vasa recta after ischemia. Kidney Int60: 1415–1427, 2001

27. Day YJ, Huang L, Ye H, Linden J, Okusa MD: Renal isch-emia-reperfusion injury and adenosine 2A receptor-medi-ated tissue protection: Role of macrophages. Am J PhysiolRenal Physiol 288: F722–F731, 2005

28. Furuichi K, Wada T, Iwata Y, Kitegawa K, Hashimoto H,Ishiwata Y, Asano M, Wang H, Matsushima K, Takeya M,Kuziel WA, Mukaida N, Yokoyama H: CCR2 signalingcontributes to ischemia-reperfusion injury in the kidney.J Am Soc Nephrol 14: 2503–2515, 2003

29. Furuichi K, Wada T, Iwata Y, Kitagawa K, Kobayashi K,Hashimoto H, Ishiwata Y, Tomosugi N, Mukaida N, Mat-sushima K, Egashira K, Yokoyama H: Gene therapy ex-pressing amino-terminal truncated monocyte chemoattrac-tant protein-1 prevents renal ischemia-reperfusion injury.J Am Soc Nephrol 14: 1066–1071, 2003

30. Farrar CA, Wang Y, Sacks SH, Zhou W: Independent path-ways of P-selectin and complement-mediated renal isch-emia/reperfusion injury. Am J Pathol 164: 133–141, 2004

31. Miura M, Fu X, Zhang QW, Remick DG, Fairchild RL:Neutralization of Gro alpha and macrophage inflamma-tory protein-2 attenuates renal ischemia/reperfusion in-jury. Am J Pathol 159: 2137–2145, 2001

32. Suwa T, Hogg JC, Quinlan KB, Van Eeden SF: The effect ofinterleukin-6 on L-selectin levels on polymorphonuclearleukocytes. Am J Physiol Heart Circ Physiol 283: H879–H884,2002

33. Biffl WL, Moore EE, Moore FA, Barnett CC Jr: Interleukin-6delays neutrophil apoptosis via a mechanism involvingplatelet-activating factor. J Trauma 40: 575–578, 1996

34. Johnson JL, Moore EE, Tamura DY, Zallen G, Biffl WL,Silliman CC: Interleukin-6 augments neutrophil cytotoxicpotential via selective enhancement of elastase release.J Surg Res 76: 91–94, 1998

35. Romano M, Sironi M, Toniatti C, Polentarutti N, FruscellaP, Ghezzi P, Faggioni R, Luini W, van Hinsbergh V, Soz-zani S, Bussolino F, Poli V, Ciliberto G, Mantovani A: Roleof IL-6 and its soluble receptor in induction of chemokinesand leukocyte recruitment. Immunity 6: 315–325, 1997

36. Cole N, Krockenberger M, Bao S, Beagley KW, HusbandAJ, Willcox M: Effects of exogenous interleukin-6 duringPseudomonas aeruginosa corneal infection. Infect Immun 69:4116–4119, 2001

37. Meng ZH, Dyer K, Billiar TR, Tweardy DJ: Essential role

for IL-6 in postresuscitation inflammation in hemorrhagicshock. Am J Physiol Cell Physiol 280: C343–C351, 2001

38. Cuzzocrea S, De Sarro G, Costantino G, Ciliberto G, Maz-zon E, De Sarro A, Caputi AP: IL-6 knock-out mice exhibitresistance to splanchnic artery occlusion shock. J LeukocBiol 66: 471–480, 1999

39. Fattori E, Cappelletti M, Costa P, Sellitto C, Cantoni L,Carelli M, Faggioni R, Fantuzzi G, Ghezzi P, Poli V: De-fective inflammatory response in interleukin 6-deficientmice. J Exp Med 180: 1243–1250, 1994

40. Dalrymple SA, Lucian LA, Slattery R, McNeil T, Aud DM,Fuchino S, Lee F, Murray R: Interleukin-6-deficient miceare highly susceptible to Listeria monocytogenes infection:Correlation with inefficient neutrophilia. Infect Immun 63:2262–2268, 1995

41. Kopf M, Baumann H, Freer G, Freudenberg M, Lamers M,Kishimoto T, Zinkernagel R, Bluethmann H, Kohler G:Impaired immune and acute-phase responses in interleu-kin-6-deficient mice. Nature 368: 339–342, 1994

42. Campbell IL, Hobbs MV, Dockter J, Oldstone MB, AllisonJ: Islet inflammation and hyperplasia induced by the pan-creatic islet-specific overexpression of interleukin-6 intransgenic mice. Am J Pathol 145: 157–166, 1994

43. Alonzi T, Fattori E, Lazzaro D, Costa P, Probert L, KolliasG, De Benedetti F, Poli V, Ciliberto G: Interleukin 6 isrequired for the development of collagen-induced arthritis.J Exp Med 187: 461–468, 1998

44. Naka T, Nishimoto N, Kishimoto T: The paradigm of IL-6:From basic science to medicine. Arthritis Res 4[Suppl 3]:S233–S242, 2002

45. Hierholzer C, Kalff JC, Omert L, Tsukada K, Loeffert JE,Watkins SC, Billiar TR, Tweardy DJ: Interleukin-6 produc-tion in hemorrhagic shock is accompanied by neutrophilrecruitment and lung injury. Am J Physiol 275: L611–L621,1998

46. Lacroix S, Chang L, Rose-John S, Tuszynski MH: Deliveryof hyper-interleukin-6 to the injured spinal cord increasesneutrophil and macrophage infiltration and inhibits axonalgrowth. J Comp Neurol 454: 213–228, 2002

47. Dybdahl B, Wahba A, Lien E, Flo TH, Waage A, Qureshi N,Sellevold OF, Espevik T, Sundan A: Inflammatory re-sponse after open heart surgery: Release of heat-shockprotein 70 and signaling through toll-like receptor-4. Cir-culation 105: 685–690, 2002

48. Wallin RP, Lundqvist A, More SH, von Bonin A, KiesslingR, Ljunggren HG: Heat-shock proteins as activators of theinnate immune system. Trends Immunol 23: 130–135, 2002

49. Beg AA: Endogenous ligands of Toll-like receptors: Impli-cations for regulating inflammatory and immune re-sponses. Trends Immunol 23: 509–512, 2002

50. Camargo CA Jr, Madden JF, Gao W, Selvan RS, ClavienPA: Interleukin-6 protects liver against warm ischemia/reperfusion injury and promotes hepatocyte proliferationin the rodent. Hepatology 26: 1513–1520, 1997

J Am Soc Nephrol 16: ???–???, 2005 IL-6 in Ischemic Acute Renal Failure 11