of July 10, 2018.This information is current as

Synthase Exacerbates Erosive Joint DiseaseSelective Inhibition of Inducible Nitric Oxide

and Sharon M. WahlNancy L. McCartney-Francis, Xiao-yu Song, Diane E. Mizel

http://www.jimmunol.org/content/166/4/2734doi: 10.4049/jimmunol.166.4.2734

2001; 166:2734-2740; ;J Immunol 

Referenceshttp://www.jimmunol.org/content/166/4/2734.full#ref-list-1

, 10 of which you can access for free at: cites 30 articlesThis article

        average*  

4 weeks from acceptance to publicationFast Publication! •    

Every submission reviewed by practicing scientistsNo Triage! •    

from submission to initial decisionRapid Reviews! 30 days* •    

Submit online. ?The JIWhy

Subscriptionhttp://jimmunol.org/subscription

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/About/Publications/JI/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/alertsReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists All rights reserved.Copyright © 2001 by The American Association of1451 Rockville Pike, Suite 650, Rockville, MD 20852The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

by guest on July 10, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

by guest on July 10, 2018

http://ww

w.jim

munol.org/

Dow

nloaded from

Selective Inhibition of Inducible Nitric Oxide SynthaseExacerbates Erosive Joint Disease

Nancy L. McCartney-Francis,1 Xiao-yu Song, Diane E. Mizel, and Sharon M. Wahl

NO is an essential cytotoxic agent in host defense, yet can be autotoxic if overproduced, as evidenced in inflammatory lesions andtissue destruction in experimental arthritis models. Treatment of streptococcal cell wal1-induced arthritis in rats withNG-mono-methyl-L-arginine (L-NMMA), a competitive nonspecific inhibitor of both constitutive and inducible isoforms of NO synthase(NOS), prevents intraarticular accumulation of leukocytes, joint swelling, and bone erosion. Because increased inducible NOS(iNOS) expression and NO generation are associated with pathogenesis of chronic inflammation, we investigated whether aselective inhibitor of iNOS, N-iminoethyl-L-lysine (L-NIL), would have more directed anti-arthritic properties. Whereas bothL-NMMA and L-NIL inhibited nitrite production by streptococcal cell wall-stimulated rat mononuclear cells in vitro and systemictreatment of arthritic rats with L-NMMA ablated synovitis, surprisingly L-NIL did not mediate resolution of inflammatory jointlesions. On the contrary, daily administration of L-NIL failed to reduce the acute response and exacerbated the chronic inflam-matory response, as reflected by profound tissue destruction and loss of bone and cartilage. Although the number of iNOS-positivecells within the synovium decreased after treatment withL-NIL, immunohistochemical analyses revealed a distinct pattern ofendothelial and neuronal NOS expression in the arthritic synovium that was unaffected by the isoform-specificL-NIL treatment.These studies uncover a contribution of the constitutive isoforms of NOS to the evolution of acute and chronic inflammationpathology which may be important in the design of therapeutic agents. The Journal of Immunology,2001, 166: 2734–2740.

Overproduction of NO, a short-lived free radical, has beenassociated with persistent inflammation and tissue de-struction, and modulation of the biosynthesis or activity

of NO results in amelioration of pathogenesis in experimental ar-thritis models (1). NO is synthesized by the enzyme NO synthase(NOS)2 through the five-electron oxidation of one of the chemi-cally equivalent guanidine nitrogens associated withL-arginine (2).Of the three NOS isoforms, two are constitutively expressed, en-dothelial NOS (eNOS, NOS3) and neuronal NOS (nNOS, NOS1),and one, inducible NOS (iNOS, NOS2), is regulated at the genelevel by a variety of inflammatory mediators (2, 3). Because iNOSis present in phagocytic leukocytes and is responsible for the pro-duction of sustained high levels of NO, it is often considered theprimary perpetrator of autotoxicity associated with chronic inflam-matory lesions (2, 4). In rodents, inhibition of NOS usingNG-monomethyl-L-arginine (L-NMMA), which belongs to a class ofarginine analogues that competitively inhibit all three NOS iso-forms, significantly reduced NO and, concomitantly, the synovialinflammation and tissue damage which occurs in streptococcal cellwall (SCW)-induced arthritis (5) and adjuvant arthritis (6). Theidentification of iNOS in human rheumatoid synovium (7, 8) sug-

gests that regulation of the NO pathway in this chronic debilitatingdisease may also offer a route of therapeutic intervention.

One potential drawback to usingL-NMMA as a systemic inhib-itor of the NO pathway is its apparent inability to discriminatebetween the different isoforms of NOS. Interference with neuronaland endothelial constitutive NOS (cNOS) could affect cell-cellcommunication, vascular tone, and neurotransmission (9).Whereas rats treated with NMMA did not display any overt det-rimental effects (5), inhibitors selective for iNOS could potentiallyprove useful in the long-term treatment of arthritides and otherinflammatory diseases to avoid hypertension and other potentialside effects. In this regard, aminoguanidine (AG), which showssome selectivity for iNOS, andN-iminoethyl-L-lysine (L-NIL), amore specific iNOS inhibitor (10, 11), inhibit adjuvant-inducedarthritis (10) and experimental osteoarthritis (12). WhereasL-NMMA inhibits the constitutive isoforms of NOS (cNOS: eNOSand nNOS) and iNOS,L-NIL reportedly exhibits 33-fold selectiv-ity for the inducible form of the enzyme (IC50, 3 mM) comparedwith cNOS (IC50, 100mM) (10). In this study, we compareL-NIL,a specific iNOS inhibitor, with AG and the nonspecific inhibitor ofNOS, L-NMMA, and demonstrate that targeting of iNOS is actu-ally less, not more, protective. Based on these data, we reveal arole of eNOS and nNOS in the evolution of bacterial cell wall-induced joint pathology.

Materials and MethodsArthritis induction

Pathogen-free Lewis (LEW/N) female rats (;100 g; Charles River Breed-ing Laboratories, Wilmington, MA) were injected i.p. with peptidoglycan-polysaccharide fragments (30mg rhamnose/g body mass) prepared fromgroup A SCW (Lee Laboratories, Grayson, GA). At indicated intervals, thearthritic response was quantified by scoring the four distal joints on a scaleof 0–4 based on swelling, redness, and degree of distortion. The individualjoint scores were summed to determine the articular index (AI), with amaximum possible score of 16. The indices for a group of animals wereaveraged and reported as the mean AI6 SEM. Statistical significance wasdetermined using the nonparametric Mann-WhitneyU test.

Oral Infection and Immunity Branch, National Institute of Dental and CraniofacialResearch, National Institutes of Health, Bethesda, MD 20892

Received for publication April 27, 2000. Accepted for publication November30, 2000.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby markedadvertisementin accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 Address correspondence and reprint requests to Dr. Nancy McCartney-Francis, 30Convent Drive, Building 30, Room 326, National Institute of Dental and CraniofacialResearch, National Institutes of Health, Bethesda, MD 20892-4352.2 Abbreviations used in this paper: NOS, NO synthase;L-NMMA, NG-monomethyl-L-arginine;L-NIL, N-iminoethyl-L-lysine; iNOS, inducible NOS; cNOS, constitutiveisoforms of NOS: eNOS, endothelial NOS; nNOS, neuronal NOS; SCW, streptococ-cal cell wall; AG, aminoguanidine; AI, articular index; PMN, polymorphonuclearcell; PEC, peritoneal exudate cell.

Copyright © 2001 by The American Association of Immunologists 0022-1767/01/$02.00

by guest on July 10, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

Administration of inhibitors

SCW-injected and control LEW/N rats were randomly selected for treat-ment. L-NIL was obtained from Searle (St. Louis, MO).L-NMMA waspurchased from Calbiochem (San Diego, CA). Previous studies have dem-onstrated the effectiveness of oral delivery ofL-NIL (10) and i.v. admin-istration of NMMA (5) to inhibit NO production. In the present study,comparisons between oral, i.v., and i.p. delivery of inhibitors yielded sim-ilar results. When added to the drinking water,L-NIL was given at a doseof 100mg/ml and theL-NIL-containing water was changed every 2–3 days.Alternatively, L-NIL in PBS was injected i.p. daily at a dose of 3 mg/kgbody weight. In our hands, this dosage reduced plasma nitrite1 nitratelevels in SCW-injected rats by 90%.L-NMMA was injected i.v. or i.p.daily at a dose of 30 mg/kg body weight (5). Control animals received anequal volume (1 ml) of PBS. Potential side effects of theL-NIL treatmentwere assessed by multiple parameters, including visual appearance as wellas white blood cell counts, hematocrit levels, and body weights.

Histology and immunohistochemistry

Joint tissues from control and arthritic rats were excised and fixed in 10%buffered Formalin, decalcified in 10% EDTA, embedded in paraffin, sec-tioned (6mm), and stained with hematoxylin and eosin for histopathology.Immunohistochemical staining for NOS was performed by overnight in-cubation at 4°C with either preimmune rabbit sera (1:1000), anti-iNOS(1:200; Upstate Biotechnology, Lake Placid, NY), anti-eNOS (2mg/ml;Transduction Laboratories, Lexington, KY), or anti-nNOS (2mg/ml;Transduction Laboratories). The monospecificity of the NOS isoform Abswas confirmed by Western blot analysis of macrophage, endothelial cell,and pituitary lysates (Transduction Laboratories) (Fig. 1).

Leukocyte isolation

Peripheral blood was diluted in PBS (1:4), layered on Ficoll (Histopaque1083; Sigma, St. Louis, MO), and centrifuged for 30 min. The mononu-clear cells (PBMC) at the interface were collected and resuspended inDMEM containing gentamicin (10mg/ml) and glutamine (2 mM). Poly-morphonuclear cells (PMN) were isolated from peripheral blood by densitygradient centrifugation (Polymorphprep diluted 5:1; Life Technologies,Gaithersburg, MD). Resident peritoneal exudate cells (PEC) were obtainedby lavaging the peritoneal cavities with PBS (BioWhittaker, Walkersville,MD). Mononuclear cells and neutrophils (53 106/ml) were cultured withor without SCW (5mg/ml). Cells were harvested after 6 h of culture forRNA analysis or 16 h for protein analysis. For nitrite determinations, PEC(1 3 106/ml) and PBMC (23 106/ml) were cultured for 36 h and super-natant fluids were collected and stored at220°C.

Nitrite assays

Nitrite content in cell culture supernatant fluids was measured by the col-orimetric microplate method using Griess reagent (13). Plasma nitrite1nitrate concentrations were determined using a fluorometric assay (11, 14).Fluorescence was measured at a wavelength of 365/450 excitation/emis-sion using a fluorescence plate reader (Idexx Laboratories, Westbrook,ME). Data were reported as mean concentration6 SEM. Statistical dif-ferences were determined using the unpaired two-tailedt test.

RT-PCR

Frozen hindleg joints were pulverized using a freezer/mill (Spex CertiPrep,Metuchen, NJ) and total cellular RNA was isolated using Trizol (Life

Technologies, Gaithersburg, MD). RNA was reverse transcribed andcDNA levels were normalized based on an initial PCR amplification withGAPDH primers. iNOS and nNOS cDNA was amplified by semiquantita-tive RT-PCR as described previously (14). For analysis of eNOS, cDNAwas amplified in the presence of a known amount of pretitered competitorDNA (mimic) which had an identical sequence to the eNOS cDNA exceptfor a 50-bp deletion. The competitor DNA was generated by PCR ampli-fication using the 59eNOS-specific primer and a modified 39competitorprimer that contained an additional 20-base sequence located 50 basesupstream of the 39eNOS-specific primer sequence (15, 16). Conditions forPCR were as follows: 94°C for 45 s, 60°C for 45 s, and 72°C for 1 min(GAPDH, 28 cycles; iNOS, 35 cycles; nNOS, 34 cycles; eNOS, 36 cycles).The primers (59to 39) and their product sizes are as follows: GAPDH (555bp): 59-GTGAAGGTCGGTGTCAACGGATTT, 39-CACAGTCTTCT-GAGTGGCAGTGAT; iNOS (496 bp): 59-CCCTTCCGAAGTTTCTGGCAGCAGC, 39-GGGTGTCAGAGTCTTGTGCCTTTGG; nNOS (328 bp): 59-TGGAAGAGCACACGTTTGGGGTCCAGCAGA, 39-GGTGGTCTCCAGGTGTGTAGTAAAGCCCTC; eNOS (189bp): 59-TGAACCCTTCCGGGGATTCTGGCAAGACCG, 39-GGATCCCTGGAAAAGGCGGTGAGGACTTGT; and eNOS mimic (139 bp):39-GAAAAGGCGGTGAGGACTTGTGTCCAGTTGGGAGCATCGGC.

PCR products were electrophoresed in a 1.5% agarose (iNOS, nNOS) or2% NuSieve low-melting point agarose (FMC Bioproducts, Rockland,ME) (eNOS) gel and ethidium bromide-stained bands were scanned andquantitated by a fluorimager (Molecular Dynamics, Sunnyvale, CA) usingImageQuant software (Molecular Dynamics). The densitometric values ofiNOS and nNOS products were normalized to GAPDH, and eNOS wascompared with eNOS mimic. The ratios were then divided by that of theSCW-injected rat sample to assess relative RNA expression.

Western blotting

Cell pellets (153 106) or bone powder (100 mg) were suspended in lysisbuffer (25 mM Tris (pH 7.4), 150 mM NaCl, 1% Nonidet P-40, 1 mMEDTA (pH 8.0), 1 mM EGTA (pH 8.0), 1 mM PMSF, 1 mM Na3VO4, 5mg/ml aprotinin, 5mg/ml leupeptin, and 5mg/ml 4-(2-aminoethyl)-benzenesulfonylfluoride-HCl) and placed on ice for 10 min. The resulting lysateswere centrifuged at 14,000 rpm for 10 min and protein content of thesupernatant was measured (Bio-Rad protein assay; Bio-Rad, Hercules, CA)using BSA as the standard. Lysate samples (100mg) were boiled for 5 minin SDS sample buffer, electrophoresed in 7.5% SDS-acrylamide gels, andtransferred to nitrocellulose membranes. The membranes were blockedwith 5% nonfat dry milk in TBST (100 mM Tris-HCl (pH 7.5), 154 mMNaCl, and 0.1% Tween 20) overnight at 4°C and incubated with isoform-specific NOS Abs (iNOS, 1:1000; Upstate Biotechnology; eNOS, 1:1000;nNOS, 1:500; Alexis, San Diego, CA) or anti-actin (1:250; Sigma) for 3 hat room temperature or overnight at 4°C. After three washes in TBST, themembranes were incubated with secondary Ab conjugated with HRP (don-key anti-rabbit IgG, 1:3000) for 1 h at room temperature, followed by threewashes in TBST. The membranes were then developed with the ECL de-tection system (Renaissance; New England Nuclear, Boston, MA).

ResultsComparison ofL-NMMA andL-NIL on NOS activity in vitro andin vivo

Based on the reported specificity and potency ofL-NMMA, L-NIL,and AG for the NOS isozymes (10), we monitored their ability tosuppress SCW-induced NO production in vitro. All three com-pounds inhibited, in a dose-dependent manner, nitrite productionby SCW-stimulated rat mononuclear cells in vitro (Fig. 2).L-NILwas 10-fold more potent thanL-NMMA (10 mM L-NIL, 72% in-hibition vs 100mM L-NMMA, 66% inhibition) and.50-fold morepotent than AG (500mM AG, 52% inhibition) in inhibiting NOproduction in this in vitro assay. Because bothL-NIL andL-NMMA were more effective in vitro inhibitors of NO synthesisand represented both a nonspecific and an iNOS-specific inhibitor,we compared the in vivo efficacy ofL-NIL with L-NMMA whichwe have previously shown to be effective in ameliorating SCW-induced tissue damage (5).

First, to document a suppressive effect in vivo, PBMC wereobtained from animals receiving arthritogenic doses of SCW pep-tidoglycan-polysaccharide complexes which had been treated witheitherL-NMMA or L-NIL (Fig. 3A). Without further challenge in

FIGURE 1. Monospecificity of NOS isoform Abs. Western blot analy-sis of macrophage (1), endothelial (2), and pituitary (3) cell lysates (5mg)using isoform-specific NOS Abs. The upper band (;150 kDa) in the pi-tuitary sample (3) represents the nNOS protein and the lower minor band(;100 kDa) may be a variant form of nNOS as recently described (30) ora degradation fragment.

2735The Journal of Immunology

by guest on July 10, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

vitro, PBMC from arthritic animals constitutively released sub-stantial levels of NO. Similar to the in vitro data, nitrite productionby PBMC from L-NIL-treated rats was reduced by 80% as com-pared with the untreated arthritic rats (SCW alone,p , 0.01),whereasL-NMMA treatment resulted in 64–70% reduction in ni-trite production (p5 0.05; Ref. 5). The systemic administration ofNO inhibitors also influenced plasma levels of NO metabolitesnitrite and nitrate (Fig. 3B). Plasma nitrite1 nitrate concentrationswere elevated.10-fold during chronic arthritis following SCWtreatment (248.66 40.0 vs 18.036 5.71 for PBS control,p ,0.01). L-NMMA administered i.v. (or i.p.) at 30 mg/kg signifi-cantly reduced plasma nitrite1 nitrate (p , 0.01), whereasL-NILreduced plasma nitrite1 nitrate levels to control levels (p ,0.0001). Inhibition byL-NIL was effective whether administeredorally (at a dose of 100mg/ml in the drinking water) or given i.p.daily (at 3 mg/kg). Thus, circulating levels of NO increase in as-sociation with the development of acute and chronic arthritis andthe NO inhibitors, whether nonspecific or specific for iNOS, ef-fectively diminish these systemic levels. At this level, the twodifferent inhibitors appeared indistinguishable.

Effect of selective inhibitors of iNOS on SCW-induced arthritis

Having established that both nonspecific and selective NOS inhib-itors block peripheral leukocyte and plasma increases in NO me-tabolites in animals challenged with SCW, we next evaluatedwhether these inhibitors had comparable effects on the develop-ment of arthritis. As evident in Fig. 4, and as we previously re-ported (5), the daily administration ofL-NMMA partially sup-presses acute arthritis (days 3–5, 20–30% inhibition) andsignificantly inhibits the clinical manifestations of the chronic ar-thritic lesions (AI, 0.56 0.29 vs 6.56 0.93 for SCW alone,p 50.007). However, surprisingly, the continuous administration ofL-NIL, a specific inhibitor of iNOS, failed to inhibit either acute orchronic synovial lesions. In fact, the daily delivery ofL-NIL,whether orally or by i.p. injection, often exacerbated the arthritis(oral L-NIL: AI 5 10.3 6 0.5 vs 7.16 0.6 for SCW alone,p ,0.012; i.p.L-NIL: AI 5 9.5 6 1.0 vs 6.56 0.9 for SCW alone)

(Fig. 4A). The increased AIs reflected the histopathology of thejoints obtained from these animals.

Histological analysis of joints from rats treated only with SCWexhibited marked infiltration of inflammatory cells into the syno-vial and subsynovial tissue (Fig. 4B). Synovial hyperplasia, a nar-rowing of the synovial space, and severe deterioration of the car-tilage and subchondral bone were also evident. By comparison, thetissues from the arthritic animals receivingL-NIL exhibited ex-treme tissue destruction including extensive loss of bone and car-tilage (Fig. 4C). In contrast, joints fromL-NMMA-treated animalsexhibited sharply reduced inflammatory changes consonant withthe decreased clinical inflammatory index (5).

Systemic effects of specific iNOS inhibition

To assess whether theL-NIL treatment was associated with sys-temic effects, we evaluated several parameters (Table I). Whenadministered orally,L-NIL treatment partially reduced the leuko-cytosis normally observed in SCW-induced arthritis by 25%(24.63 103/mm3 vs 32.93 103/mm3 in untreated SCW-injectedrats,p 5 0.045). In similar fashion, packed RBC volume as mea-sured by hematocrit, which is typically reduced in arthritic ani-mals, was increased followingL-NIL treatment (436 3 vs 36 61.6 for SCW alone as compared with 476 0.6 in PBS control rats),

FIGURE 2. NOS inhibitors block NO synthesis in vitro. PEC (106/ml)from normal female Lewis rats (n5 2) were cultured in duplicate for 36 hwith SCW (5mg/ml) and increasing micromolar amounts of NOS inhibi-tors. Supernatant fluids were assayed in triplicate for nitrite content by theGriess reaction. Values represent mean6 SEM and are representative ofthree similar experiments.p, p # 0.05 compared with SCW alone, unpairedtwo-tailed t test.

FIGURE 3. NOS inhibitors reduce NO production in the periphery ofarthritic rats.A, PBMC were isolated on day 24 from arthritic rats treatedwith NOS inhibitors and cultured (23 106/ml) for 36 h without furtherstimulation. Supernatant fluids were assayed in triplicate for nitrite contentby the Griess reaction. Values represent mean6 SEM and are represen-tative of two similar experiments.B, Nitrite 1 nitrate concentrations inplasma were measured by a fluorescent microplate assay using 2,3-diami-nonaphthalene. Data represent mean plasma nitrite1 nitrate6 SEM fromchronic phase rats treated with SCW alone (n5 7), SCW1 L-NIL (n 510), or SCW1 L-NMMA (n 5 4). Control groups (PBS,L-NIL, or L-NMMA alone) consisted of three to four rats.p, p # 0.05 compared withSCW 1 PBS, unpaired two-tailedt test.

2736 FAILURE OF iNOS-SPECIFIC INHIBITOR TO SUPPRESS SCW-ARTHRITIS

by guest on July 10, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

although not fully restored to normal levels. Additionally, the char-acteristic weight loss in arthritic rats (148.36 5.9 g for SCW-injected rats vs 2066 3.9 g for PBS rats,p , 0.05) was not

reversed byL-NIL treatment (132.26 4.2 g,p , 0.04 as comparedwith SCW alone). By comparison, the lack of toxicity ofL-NMMAin the rat arthritis model was reflected in the normal hematocrit andbody weights ofL-NMMA-treated arthritic rats as reported previ-ously (5). With regard to all of these parameters as well as physicalappearance, rats treated withL-NIL alone (no SCW) were indis-tinguishable from rats treated with PBS alone.

Identification of NOS isoenzymes in arthritic lesions

Based on the unanticipated effect ofL-NIL on synovial pathology,we next focused on whether theL-NIL had any effect on localizedsynovial production of NO and whether it had any impact on tissueiNOS and/or the other NOS isoforms. Although no iNOS stainingwas detected in synovial tissue obtained from untreated controlanimals (data not shown), iNOS immunoreactivity was clearly ev-ident within the hyperplastic synovial tissue from SCW-arthriticrats (Fig. 5A). The primary cells exhibiting specific iNOS immu-noreactivity were morphologically consistent with macrophages.Synovial lining cells and some chondrocytes were also iNOS pos-itive. After treatment with the iNOS-specific inhibitorL-NIL, thenumber of positive cells decreased, although some staining per-sisted proximal to the bone (Fig. 5B). Parallel staining with theAbs for eNOS and nNOS revealed an unanticipated distribution ofthese additional isoforms in the arthritic synovium, but not in nor-mal synovium (PBS,L-NIL alone, data not shown). In addition toeNOS staining in the endothelial cells, multinucleated osteoclast-like cells residing on the borders of degrading bone were clearlyeNOS positive (Fig. 5,C andD), but negative for iNOS and nNOS.Scattered cells, some of which were multinucleated, within thesynovium and bone marrow stroma also were eNOS positive.nNOS staining, previously identified in bone (17), but not synovialtissues, was seen in the infiltrating leukocytes including macro-phages and polymorphonuclear leukocytes (Fig. 5,E andF). Thesecells were negative for eNOS and did not stain in the absence ofthe primary Ab, indicating that staining was not due to a nonspe-cific peroxidase effect. More importantly, after treatment withL-NIL, even when the iNOS staining was reduced (Fig. 5B), neithereNOS (Fig. 5D) nor nNOS (Fig. 5F) staining was altered, docu-menting the specificity of the inhibitor, but suggesting that addi-tional sources of NO were sustained even in the presence ofL-NIL.In contrast, treatment withL-NMMA reduced both eNOS- andnNOS-positive cells as well as iNOS (data not shown), coincidentwith reduced inflammatory pathology (5).

Inflammatory cells as a source of iNOS and nNOS

To correlate the cellular distribution of the NOS isoforms identi-fied by immunohistochemical analysis with cell-specific proteinlevels, mononuclear cells and neutrophils were isolated from rat

FIGURE 4. Differential effects of NOS inhibitors on SCW-induced ar-thritis. A, Female Lewis rats received a single i.p. injection of SCW at day0 and daily i.p. injections of NOS inhibitorsL-NMMA or L-NIL for 20–24days. Data represent mean AI6 SEM at day 3 (acute) and days 20–24(chronic) and are representative of two experiments (totaln for L-NILtreatment, 19). Chronic arthritic joints from SCW (B) and SCW1 L-NIL(C) were embedded in paraffin, sectioned, and stained with hematoxylinand eosin. A single injection of SCW induces the accumulation of inflam-matory cells within the synovial tissue and bone erosion and cartilage deg-radation (B). AfterL-NIL treatment (C), a pronounced increase in the cel-lularity and joint erosion was evident. Syn, synovium; JS, joint spacecavity; C, cartilage; B, bone; BM, bone marrow. Original magnification,320.

Table I. Effects ofL-NIL treatment on circulating leukocytes in arthriticratsa

Treatment WBC (3103/mm3) Body Weight (g)

PBS 6.06 0.8 206.36 3.9L-NIL 6.36 1.1 191.86 3.3SCW 32.96 1.6b 148.36 5.9b

SCW 1 L-NIL 24.6 6 1.8b,c 132.26 4.2b,c

a Values are the mean6 SEM. Rats were injected i.p. with PBS or SCW on day1. L-NIL was administered in the drinking water (100mg/ml) and rats were killed onday 37 (n5 2–8 rats for each group).

b p # 0.05 in comparison withL-NIL control group.c p # 0.05 in comparison with SCW treatment group using the Mann-WhitneyU

test.

2737The Journal of Immunology

by guest on July 10, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

peripheral blood, stimulated with SCW in vitro, and the cell ex-tracts were monitored for iNOS, eNOS, and nNOS proteins usingisoform-specific polyclonal Abs. As shown in Fig. 6, PBMC didnot constitutively express iNOS, but produced the 130-kDa iNOSprotein in response to SCW. Moreover, constitutive nNOS expres-sion was evident which was elevated following exposure to SCW.

eNOS levels were barely seen and minimally augmented by SCW.Neutrophils, on the other hand, were strongly positive for iNOSfollowing exposure to SCW, but did not synthesize detectable lev-els of eNOS or nNOS protein before or after stimulation withSCW. In additional experiments, peritoneal macrophages stimu-lated by SCW were found to express low levels of both eNOS andnNOS proteins as well as substantially enhanced levels of iNOS.Thus, inflammatory cells triggered by SCW have the capacity tocontribute to the overproduction of NO, not only through the ex-pression of iNOS but also constitutive isoforms of NOS.

Gene expression of NOS isoforms in arthritic joints is altered byNOS inhibitors

Based on this surprising NOS isoform expression, we further as-sessed NOS gene expression in inflamed and treated joints at theRNA level. RNA was isolated from arthritic joint tissue and rel-ative RNA levels were determined by RT-PCR. Increased expres-sion of iNOS mRNA was observed in arthritic joints (Fig. 7,A andC), consistent with immunohistochemical data for iNOS proteinand our previous data (5, 14). Interestingly, nNOS mRNA expres-sion in control animals (PBS,L-NIL alone) was comparable toarthritic rats, even though immunohistochemical staining for

FIGURE 5. Reduced iNOS but persistent cNOS expression inL-NIL-treated arthritic rats. Chronic arthritic joints for SCW (A,C, andE) andSCW 1 L-NIL (B, D, andF) were processed and sectioned for immuno-histochemical analysis with NOS isoform-specific Abs. NOS-positive cellswere visualized by the reddish brown diaminobenzidine stain. Arrows de-pict multinucleated cells (A,C, andD). A reduction in iNOS staining wasobserved in joint tissue fromL-NIL-treated rats (A, SCW alone;B, SCW1L-NIL). However, eNOS (CandD) and nNOS (EandF) staining remainedprominent in the same joints. Original magnification,340.

FIGURE 6. SCW induction of NOS enzymes in rat peripheral bloodleukocytes. Western blot analysis of cell lysates from normal rat mononu-clear cells (Ficoll purified, PBMC), polymorphonuclear cells (polyprep pu-rified, PMN), and PEC using isoform-specific NOS Abs. iNOS protein wasdetected in each cell population after stimulation with SCW for 16 h. Low-level constitutive expression of eNOS in PBMC and PEC populations wasincreased after SCW stimulation but not detected in the PMNs. nNOSexpression was observed in the mononuclear cells only (PBMC, PEC). Thedata are representative of three experiments.

FIGURE 7. Differential expression of NOS mRNA expression in jointtissue. RNA was isolated from pulverized joints of treated rats (day 24 afterSCW injection), reverse transcribed, and amplified by PCR in the absence(A) or presence of a competitor (B). PCR products were separated by gelelectrophoresis and stained with ethidium bromide.C, The stained bandswere scanned and quantitated by a fluorimager. iNOS and nNOS densito-metric values were normalized to GAPDH (A; iNOS, nNOS) or and eNOSwas compared with eNOS mimic (B). Data are reported as ratios relative tothe SCW-injected rat sample and are representative of two experiments(n 5 2–4 rats/group).

2738 FAILURE OF iNOS-SPECIFIC INHIBITOR TO SUPPRESS SCW-ARTHRITIS

by guest on July 10, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

nNOS protein in control tissues was undetectable. In contrast to thehigh constitutive expression of nNOS mRNA, eNOS mRNA ex-pression in joint tissue was not detectable by conventional RT-PCR and could only be observed in the presence of a mimic cDNA(competitive RT-PCR)(Fig. 7,B and C). After L-NIL treatment,iNOS mRNA expression, although somewhat variable, was con-sistently reduced (23–60% reduction as compared with SCWalone,p , 0.05), whereas nNOS mRNA expression was relativelyunchanged. On the other hand, eNOS mRNA expression, albeitlow, was increased inL-NIL-treated arthritic rats.

Joint tissues were examined in parallel by Western blot analysesfor the presence of the NOS proteins (Fig. 8). Consistent with theimmunohistochemistry and RNA data, immunoreactive iNOS pro-tein was elevated in arthritic joints and markedly reduced byL-NILtreatment. Both eNOS and nNOS proteins were expressed in jointsof SCW-injected rats but not control rats and the levels of expres-sion were unaffected byL-NIL treatment. In contrast, the nonspe-cific NOS inhibitor NMMA not only suppressed expression ofiNOS protein but also the constitutive NOS isoforms eNOS andnNOS, likely reflecting, at least in part, the marked reduction ininflammatory infiltrate. These data are consistent with differentialregulation of the three NOS isoforms during joint inflammationand following selective inhibition of one isoform.

DiscussionInducible NOS has long been considered the villain responsible forexcessive and autotoxic levels of NO in chronic pathogenic in-flammatory lesions (3, 4). For these reasons, we considered thatselective blockade of the iNOS pathway, without impinging oneNOS- or nNOS-dependent events, would provide an ideal strat-egy to disrupt only the presumptive harmful NO being produced inthe synovium, and be a more effective anti-arthritic agent thanL-NMMA, which acts as a competitive inhibitor of both constitu-tive and inducible isoforms of NOS. Unanticipated was the exac-erbation of disease which occurred when the animals were treatedwith the iNOS-specific inhibitorL-NIL. These observations areconsistent with our recent inability to suppress SCW-induced ar-thritis with AG which is also a preferential iNOS inhibitor (un-published observations). Based on our previous experience with

L-NMMA (5), it was clear that excess NO contributed to the tissuepathology triggered by SCW deposition in the synovium. This di-chotomy suggested that either iNOS-dependent NO was protectiveand/or that eNOS/nNOS were, in fact, also perpetrators of tissueinjury. Although we cannot totally rule out a protective role foriNOS-derived NO (18), in these studies we have documented acontribution of the constitutive isoforms to the emerging arthrop-athy in the SCW model and that failure to concomitantly regulatethe products of these pathways is detrimental. First, by immuno-histochemical analysis, we demonstrate that all three NOS iso-forms are found in inflamed synovium. eNOS is present not onlyin synovial endothelial cells, but also in bone-degrading cells, theosteoclasts, which express high levels of this isoenzyme in arthriticjoints, as recently reported (17, 19). As expected, iNOS was ex-pressed by inflammatory cells and synovial lining cells. Furtheranalysis revealed that nNOS protein was also identified within in-filtrated leukocytes in the inflamed synovium. Thus, all threeisoenzymes were clearly represented but in a unique distribution inthe arthritic synovium.

Second, to confirm the cellular distribution of the NOS seen byimmunohistochemistry, isolated inflammatory cell populationswere found to constitutively or inducibly express the isoenzymeproteins by Western blot analysis. Both mononuclear cells andpolymorphonuclear cells expressed substantial iNOS upon stimu-lation in vitro with SCW. PBMC and peritoneal lavage macro-phages also produced increased levels of nNOS protein and limitedeNOS as detected in the cell extracts following challenge with theSCW. Although we were unable to detect nNOS protein by West-ern blot in PMNs, perhaps due to proteolytic degradation, recentstudies have demonstrated constitutive expression of nNOSmRNA by rat circulating neutrophils which could be inhibited by thenNOS inhibitor 7-nitroindazole but was not inhibited byL-NIL (20).

Third, iNOS mRNA was up-regulated in joint tissue isolatedfrom SCW-induced arthritic rats. nNOS mRNA was constitutivelyexpressed without pronounced enhancement in inflamed tissues,whereas eNOS mRNA, although very low, was somewhat elevatedin arthritic joints. Although we have not yet isolated sufficientosteoclasts from these inflamed joints to monitor mRNA expres-sion for eNOS, the specific staining pattern of these cells with theeNOS Ab and recent identification of eNOS protein in culturedosteoclasts (17) predicts the presence of the eNOS gene.

Based on our new findings, the failure of specific iNOS inhib-itors to prevent SCW-induced synovial pathology may be due tothe inability of these inhibitors to influence the NO generated bythe eNOS and nNOS pathways, which at least by immunohisto-chemical staining as well as RNA and protein levels are relativelyuntouched byL-NIL therapy. Clearly, theL-NIL is active, since thecirculating levels of nitrite and nitrate, the stable decompositionproducts of NO, are definitely reduced following treatment. Al-though the maximal levels of plasma nitrite1 nitrate correspondwith the peak inflammatory response, it is now less clear how theseplasma NO metabolites relate to the pathology since bringing theselevels to normal or near normal by inhibiting iNOS is not neces-sarily associated with a corresponding decrease in synovial dis-ease. In either case,L-NIL or L-NMMA both suppress iNOS, re-sulting in decreased NO, but it may be the localized production ofeNOS and nNOS along with iNOS which collectively drive theNO-mediated autotoxicity. However, the data are consistent withthe idea that the enhanced levels of plasma nitrite1 nitrate whichoccur during the development of arthritis result from the up-reg-ulation of iNOS, with a lesser contribution of cNOS tosystemic NO.

Whether the contribution of eNOS and nNOS to inflammatorypathology is unique to the synovium remains to be resolved. In this

FIGURE 8. Western blot analysis of arthritic joint tissue. Proteins frompulverized joints of treated (day 24 after SCW injection) and control (L-NIL alone) rats were separated on 7.5% SDS-acrylamide gels and immu-noblotted with NOS isoform-specific Abs. Immunoreactive bands were de-tected with HRP-conjugated donkey anti-rabbit secondary Ab followed byan ECL detection system. Equal loading of protein per lane was verifiedwith anti-actin Ab.

2739The Journal of Immunology

by guest on July 10, 2018http://w

ww

.jimm

unol.org/D

ownloaded from

regard, the granulomatous response to SCW in the liver is reversedby L-NIL, demonstrating not only the bioactivity of theL-NIL but,for the first time, differential tissue responses (N. McCartney-Fran-cis and S. Wahl, manuscript in preparation). Thus, suppression ofiNOS alone might be sufficient in soft tissues such as gut and liver(21) to block NO toxicity. AlthoughL-NIL has been shown tosuppress adjuvant-induced arthritis if administered prophylacti-cally (10), it is not effective therapeutically (22). Similarly, pro-phylactic treatment withL-NIL reduced progression of experimen-tal osteoarthritis in dogs by;50% (12). Contrasted with the failureof L-NIL to prevent or reverse bacterial cell wall (SCW)-inducedarthritis, these differential responses toL-NIL may reflect criticaldifferences in clinical and immunological sequelae in these inflam-matory models. In the arthritic joints of SCW-injected rats, bothreactive nitrogen species and reactive oxygen species play pivotalroles in the pathologic process and therapeutic targeting of NO byL-NMMA (5) or superoxide and hydrogen peroxide by superoxidedismutase or catalase (23) is effective in blocking the inflammationand tissue destruction. However, the ineffectiveness of the iNOSspecific inhibitorL-NIL in preventing inflammatory damage in thearthritic joint highlights the complex roles these reactive speciesplay in the pathologic process. Deletion of the iNOS gene by ho-mologous recombination in mutant mice renders these animalsmore susceptible to adjuvant (24) and septic (25) arthritis, consis-tent with our findings. Significantly, administration of a selectiveneuronal NOS inhibitor 7-nitroindazole reduced adjuvant-inducedjoint inflammation in rats (26), suggesting an inexplicable role fornNOS in arthritic pathology. These studies highlight the complex-ity of the pro- and anti-inflammatory activities of iNOS (27–29)and define a previously unrecognized role for the constitutivelyexpressed NOS in the evolution of acute and chronic inflammatorysynovial pathology. These isoenzymes must now also be consid-ered in the design of NO-related therapeutic targets.

AcknowledgmentsWe thank Dr. Mark Currie and Dr. Pam Manning for supplying theL-NILused for this project, Dr. Jane Connor for plasma nitrite determinations, andDr. Sue Silverton for pathologic assessments of joint tissue. We also thankWenwen Jin for technical assistance and Dr. Wanjun Chen and Dr. LarryWahl for critical review of this manuscript.

References1. Zamora, R., Y. Vodovotz, and T. R. Billiar. 2000. Inducible nitric oxide synthase

and inflammatory diseases.Mol. Med. 6:347.2. MacMicking, J., Q. W. Xie, and C. Nathan. 1997. Nitric oxide and macrophage

function.Annu. Rev. Immunol. 15:323.3. Clancy, R. M., A. R. Amin, and S. B. Abramson. 1998. The role of nitric oxide

in inflammation and immunity.Arthritis Rheum. 41:1141.4. Bogdan, C. 1998. The multiplex function of nitric oxide in (auto)immunity.

J. Exp. Med. 187:1361.5. McCartney-Francis, N., J. B. Allen, D. E. Mizel, J. E. Albina, Q. W. Xie,

C. F. Nathan, and S. M. Wahl. 1993. Suppression of arthritis by an inhibitor ofnitric oxide synthase.J. Exp. Med. 178:749.

6. Stefanovic-Racic, M., K. Meyers, C. Meschter, J. W. Coffey, R. A. Hoffman, andC. H. Evans. 1994.N-monomethyl arginine, an inhibitor of nitric oxide synthase,suppresses the development of adjuvant arthritis in rats.Arthritis Rheum. 37:1062.

7. Sakurai, H., H. Kohsaka, M. F. Liu, H. Higashiyama, Y. Hirata, K. Kanno,I. Saito, and N. Miyasaka. 1995. Nitric oxide production and inducible nitricoxide synthase expression in inflammatory arthritides.J. Clin. Invest. 96:2357.

8. McInnes, I. B., B. P. Leung, M. Field, X. Q. Wei, F. P. Huang, R. D. Sturrock,A. Kinninmonth, J. Weidner, R. Mumford, and F. Y. Liew. 1996. Production ofnitric oxide in the synovial membrane of rheumatoid and osteoarthritis patients.J. Exp. Med. 184:1519.

9. Moncada, S., and A. Higgs. 1993. The L-arginine-nitric oxide pathway.N. Engl.J. Med. 329:2002.

10. Connor, J. R., P. T. Manning, S. L. Settle, W. M. Moore, G. M. Jerome,R. K. Webber, F. S. Tjoeng, and M. G. Currie. 1995. Suppression of adjuvant-induced arthritis by selective inhibition of inducible nitric oxide synthase.Eur.J. Pharmacol. 273:15.

11. Misko, T. P., R. J. Schilling, D. Salvemini, W. M. Moore, and M. G. Currie.1993. A fluorometric assay for the measurement of nitrite in biological samples.Anal. Biochem. 214:11.

12. Pelletier, J. P., D. Jovanovic, J. C. Fernandes, P. Manning, J. R. Connor,M. G. Currie, J. A. Di Battista, and J. Martel-Pelletier. 1998. Reduced progres-sion of experimental osteoarthritis in vivo by selective inhibition of induciblenitric oxide synthase.Arthritis Rheum. 41:1275.

13. Ding, A. H., C. F. Nathan, and D. J. Stuehr. 1988. Release of reactive nitrogenintermediates and reactive oxygen intermediates from mouse peritoneal macro-phages: comparison of activating cytokines and evidence for independent pro-duction.J. Immunol. 141:2407.

14. McCartney-Francis, N. L., X. Y. Song, D. E. Mizel, C. L. Wahl, and S. M. Wahl.1999. Hemoglobin protects from streptococcal cell wall-induced arthritis.Arthri-tis Rheum. 42:1119.

15. Celi, F. S., M. E. Zenilman, and A. R. Shuldiner. 1993. A rapid and versatilemethod to synthesize internal standards for competitive PCR.Nucleic Acids Res.21:1047.

16. Sun, B., J. Wells, E. Goldmuntz, P. Silver, E. F. Remmers, R. L. Wilder, andR. R. Caspi. 1996. A simplified, competitive RT-PCR method for measuring ratIFN-g mRNA expression.J. Immunol. Methods 195:139.

17. Ralston, S. H. 1997. The Michael Mason Prize Essay 1997. Nitric oxide andbone: what a gas!Br. J. Rheumatol. 36:831.

18. Grisham, M. B., D. Jourd’Heuil, and D. A. Wink. 1999. Nitric oxide. I. Physi-ological chemistry of nitric oxide and its metabolites: implications in inflamma-tion. Am. J. Physiol. 276:G315.

19. Amin, A. R., P. E. Di Cesare, P. Vyas, M. Attur, E. Tzeng, T. R. Billiar,S. A. Stuchin, and S. B. Abramson. 1995. The expression and regulation of nitricoxide synthase in human osteoarthritis-affected chondrocytes: evidence for up-regulated neuronal nitric oxide synthase.J. Exp. Med. 182:2097.

20. Greenberg, S. S., J. Ouyang, X. Zhao, and T. D. Giles. 1998. Human and ratneutrophils constitutively express neural nitric oxide synthase mRNA.Nitric Ox-ide 2:203.

21. Grisham, M. B., R. D. Specian, and T. E. Zimmerman. 1994. Effects of nitricoxide synthase inhibition on the pathophysiology observed in a model of chronicgranulomatous colitis.J. Pharmacol. Exp. Ther. 271:1114.

22. Fletcher, D. S., W. R. Widmer, S. Luell, A. Christen, C. Orevillo, S. Shah, andD. Visco. 1998. Therapeutic administration of a selective inhibitor of nitric oxidesynthase does not ameliorate the chronic inflammation and tissue damage asso-ciated with adjuvant-induced arthritis in rats.J. Pharmacol. Exp. Ther. 284:714.

23. Skaleric, U., J. B. Allen, P. D. Smith, S. E. Mergenhagen, and S. M. Wahl. 1991.Inhibitors of reactive oxygen intermediates suppress bacterial cell wall-inducedarthritis.J. Immunol. 147:2559.

24. Gilkeson, G. S., J. S. Mudgett, M. F. Seldin, P. Ruiz, A. A. Alexander,M. A. Misukonis, D. S. Pisetsky, and J. B. Weinberg. 1997. Clinical and serologicmanifestations of autoimmune disease in MRL-lpr/lprmice lacking nitric oxidesynthase type 2.J. Exp. Med. 186:365.

25. McInnes, I. B., B. Leung, X. Q. Wei, C. C. Gemmell, and F. Y. Liew. 1998.Septic arthritis followingStaphylococcus aureusinfection in mice lacking induc-ible nitric oxide synthase.J. Immunol. 160:308.

26. Pozza, M., C. Bettelli, F. Magnani, M. T. Mascia, E. Manzini, and L. Calza. 1998.Is neuronal nitric oxide involved in adjuvant-induced joint inflammation?Eur.J. Pharmacol. 359:87.

27. Kolb, H., and V. Kolb-Bachofen. 1998. Nitric oxide in autoimmune disease:cytotoxic or regulatory mediator?Immunol. Today 19:556.

28. St. Clair, E. W. 1998. Nitric oxide: friend or foe in arthritis?J. Rheumatol.25:1451.

29. Willenborg, D. O., M. A. Staykova, and W. B. Cowden. 1999. Our shiftingunderstanding of the role of nitric oxide in autoimmune encephalomyelitis: areview.J. Neuroimmunol. 100:21.

30. Park, C. S., G. Krishna, M. S. Ahn, J. H. Kang, W. G. Chung, D. J. Kim,H. K. Hwang, J. N. Lee, S. G. Paik, and Y. N. Cha. 2000. Differential andconstitutive expression of neuronal, inducible, and endothelial nitric oxide syn-thase mRNAs and proteins in pathologically normal human tissues.Nitric Oxide4:459.

2740 FAILURE OF iNOS-SPECIFIC INHIBITOR TO SUPPRESS SCW-ARTHRITIS

by guest on July 10, 2018http://w

ww

.jimm

unol.org/D

ownloaded from