Research Paper Mediators of Inflammation, 8, 245–251 (1999)

The aim of the present study was to investigate the interrelationship of the kinin system, nitric oxide and eicosanoids in the acute phase of antigen-induced arthritis (AIA) in rabbits. The arthritis was induced in immunized rabbits and the following parameters were evaluated 24 hours later: leukocyte influx (total and differential white cell count), vascular permeability (Evans's blue method), and synovial PMN cell infiltrate. PGE2 and LTB4 (radioimmunoassay) levels were quantified in the synovial fluid. The animals were pre-treated with 20mg/kg/day during 14 days with L-NAME or D-NAME and/or Enalapril (0.12 mg/kg/day-14 days), and/or the B2 antagonist of Bradykinin HOE 140 (0.9 mg/kg). Our results showed that L-NAME was effective in the prevention of AIA with reduction of all Inflammatory parameters analyzed. Enalapril partially reverted the L-NAME anti-inflammatory effects. The simultaneous treatment with HOE 140 abolished this reversion and returned the inflammatory parameters to the levels observed in L-NAME treated animals. Our results suggest that pressoric alterations induced by L-NAME could not account for all its anti-inflammatory action in this model of experimental arthritis. Additionally the contribution of the kinin system in AIA was characterized as well as its interaction with eicosanoids and nitric oxide.


Introduction
Rhe umatoid arth ritis is an autoimmune disease in w hich chronic inflammation of the synovial lining ce lls produc e pain, sw e lling and progre ssive e rosion of the synovial joints. Dumonde & Glynn 1 first desc ribed antige n-induc ed arth ritis (AIA) in rabbits. AIA e volve s as a progressive immunological arthritis , w hich in his topathologic al te rms closely re sembles rheumatoid arthritis in man. The acute phas e of AIA is a typical Arthus re action w hich oc curs in the joint w ith compleme nt activatio n and polymorphonuc lear (PMN) le ukocyte migration . 2 Nitric ox ide (NO) is a ve ry small and ubiquitous molecule synthe size d from L-arg inine by nitric ox ide synthas es (NOS). NO is produc ed in the joint by chondroc yte s, 3-6 synovioc yte s 6,7 and oste oblasts. 6 It has bee n invo lve d in inflammatory re actions w ith both p ro and anti-inflammatory prope rtie s. Endogenous re le as e of NO c ould act as anti-inflammatory through the re duction of leukoc yte adhesion, 8,9 and inhibition of synthe sis of c yclo-ox ygenase produc ts such as prostaglandin E 2 (PGE 2 ), thrombox ane B 2 (Tx B 2 ) and inte rle ukin-6 (IL-6). 10 Endogenous NO has also be en described as a pro-inflammator y molecule show ing inte rac tions w ith c ytokine s [11][12][13] and inflammator y produc ts of the cyclo-ox ygenase pathw ay. [14][15][16][17][18][19][20] Studie s have show n that NO inc re as es inte rleukin-1b (IL-1) and e icosanoid produc ts w hich may re sult in the ex acerbation of the inflammatory re spons e. 12,13 The p artic ipatio n of NO in intraartic ular inflammatory proce ss has bee n fully de sc ribed in arthritic patie nts. 21,22 NOS inhibitors have bee n w ide ly use d in studie s of NO influenc e in the inflammator y process. The adminis tration of NOS inhibitors in the ex perime ntal arthritis in rats w as follow ed by re duction in the intra-artic ular ac cumulation of leukocyte s, joint e ros ion, and paw sw elling as w ell as his topathologic al abnormalitie s. 23 -27 Inhibition of NO biosynthe sis by L-arginine antago nism re duced PGE 2 produc tion by condroc yte s 3 w ithout effect upon the inc re ase d produc tion of PGE 2 by synoviocyte s stimulate d w ith IL-1. 7 More ove r, e ndogenous re le as e of NO enhanc e s macrophage c yclooxygenase (COX) activ ity and inc re ase s the produc tion of p ro-inflammatory prostaglandins . 13,28 Se veral studie s sugge st that NO partic ipate s in the inflammatory re spons e and joint destruc tion. Slic es of rabbit artic ular cartilage synthe sized large quantitie s of NO follow ing stimulatio n w ith human re c ombinant IL-1b . Tre atme nt of cartilage fragments w ith NOS inhibitor decre as ed NO synthe sis stimulate d by IL-1b and re store d prote oglycan synthe sis. 29 We pre vious ly verifie d that chronic tre atme nt of arth ritic rabbits w ith L-NAME evoke d re duc tion of the cellular influx and prote in le akage to the artic ular c avity. Additionally the synovial fluid of these animals show ed re duction of PGE 2 , IL1 and NO 2 /NO 3 le ve ls. 30 Chronic adminis tration of NOS inhibitors caus es vasoconstric tion w ith pote ntial re duc tion in c ell migratio n. The contrib ution of this pre ssoric alte ration evoked by chronic tre atment w ith L-NAME to its anti-inflammatory action as w e ll as the p ossible links betw e en NO and othe r me diato rs of inflammatory re ac tion have so far bee n poorly studie d in the artic ular environme nt. Thus, in the pre sent study w e analyz ed the leukoc yte migratio n, prote in leakage , synovium polymorphonuclear ce ll infiltrate and e icosanoids produc tion in the synovial fluid of kne e joints of rabbits w ith AIA tre ate d w ith NOS inhibitor, Enalap ril and the B2 antago nist of Bradykinin HOE 140.

Induction of arthritis
The Animal Ethic s Committe e of COBEA (Brazilian College of Ex perime ntal Animals ) has approve d all ex pe rimental proce dure s pe rforme d on animals in accordanc e w ith procedure s set by UFAW (The Unive rs itie s Federation for Animals We lfare ). Male New Zealand White rabbits w ere se nsitiz ed w ith 5 mg of me thylate d bovine se rum albumin (mBSA -Sigma) in 1 ml Fre und's c omple te adjuvant (Gibc o) and 1 ml of ste rile saline through inje c tions at subc utane ous and muscular site s in the suprasc apular and glute al re gions, re spe ctive ly. Seven days afte r the immuniz atio n the animal w as onc e a w ee k booste d w ith intrade rmic al inje ction of 1 mg/ml of mBSA. The n, cutane ous Arthus re action charac te rize d by central ne c rosis w as obse rve d 24 h late r. Simultane ously, the serum antib ody title s agains t mBSA w ere quantifie d by immunodiffusion. The animals used in the ex perime nt w e re those w ho appropriate ly re sponde d to the se cond cutane ous challenge and have had anti-mBSA title s ove r 1/8. Seven days afte r the third booste r, arthritis w as induc ed in the kne e joint by the inje c tion of 0.5 ml of a ste rile solution of mBSA (2 mg/ml) into the artic ular cavity. The contralate ral joint w as inje cte d w ith saline . Tw e nty four hours afte r intra-artic ular challenge the animal w as anae sthetiz ed w ith a mix ture of x ylazine (5 mg/kg) associate d w ith ke tamine (50 mg/kg) by intramuscular inje ction and afte r that it w as kille d by intrave nous inje c tion of 2.0 ml of KCI 20%.

Treatments
A group of animals w as rando mly tre ate d w ith L-NAME (Sigma), a competitive NOS inhibitor, or its inac tive form D-NAME (Sigma). The dose for both tre atments w as of 20 mg/kg/day mix ed w ith drinking w ate r. Drugs w ere adminis te re d for 2 w ee ks prior to the induc tion of arthritis , and starte d simultane ously w ith the sec ond booste r. Anothe r group of animals w as rando mly tre ate d w ith Enalap ril, an angiote nsin conve rting enzyme (ACE), w hich produce s blood pre ssure decre ase . The tre atme nt w ith Enalap ril starte d simultane ously w ith L-NAME and the dose of the drug w as 0.12 mg/kg/day mix ed w ith drinking w ate r. The B2 antag onis t of Bradykinin HOE 140 (D-Arg-[Hyp 3 , Thl 5 , Dtic 5 , Oic 8 ]-bradykinin) w as inje cte d subc utane ously in a dose of 0.3 mg/kg eve ry 8 hours afte r the arthritis induc tion.

Sampling of synovial fluid
Imme diate ly afte r the sacrific e 2 ml of saline containing EDTA (1 mg/ml) w as inje c te d into the kne e joint. Synovial fluid w as as pirate d, the joint w as opene d and the re mainde r of the synovial fluid as w ell as synovial me mbrane w as re c ove re d. Total and diffe re ntial le ukocyte count w as done in a Ne ubaue r chamber unde r light microsc op y. For the diffe re ntial w hite c ell count sme ars w ere pre pare d from a c ell pellet and staine d w ith Gie msa. The synovial fluid w as store d at -70°C for e icosanoids dete rmination .

Assessment of vascular permeability response with Evans's blue dye
Before mBSA inje c tion to induc e arthritis the animals re c eive d iv. 20 mg/kg of Evans's blue dye in a 2.5% saline solution. The dye combine s w ith the prote ins of the p lasma giving rise to a tagged macromole cule that passe s the endothe lial barrie r only to a ne gligible ex te nt unde r normal conditio ns, but it is de posite d in the tissue s unde r circ umstanc es of inc re ased vas cular permeability. The synovial fluid w as ce ntrifuge d and the optic al density colorime tric ally assessed at 630 nm. The c onc entratio n of dye bound to prote in in the synovial fluid w as e stimate d from a standard graph re c ording the optic al density of serial dilutio n of a know n samp le of Evans's blue in NaCl. Results are ex pre sse d as m g of prote in/ml of synovial fluid.

Synovial membrane histology
Afte r sampling the synovial fluid and opening the artic ular cavity, the synovial membrane w as ex cise d and sec tions w ere staine d w ith he matox ylin and eosin for e valuation of PMN c ell infiltrate unde r light mic rosc opy. The synovium PMN ce ll infiltrate w as quantifie d as absent or rare , mild, or se ve re by an observe r blinde d to the ex p erime ntal groups. Assay of eicosanoids PGE 2 and LTB 4 levels w e re as saye d in the synovial fluid using comme rc ial kits (Ame rsham, UK) as pre viously de scribe d. 31,32 The radioim munoassays w ere pe rforme d in polyprop yle ne tube s w ith the re age nts dilute d in phosphate -buffere d saline w ith gelatin and thimerosal. One hundre d microlite rs each of unkno w n sample s, trac ers (I 125 PGE 2 or H 3 LTB 4 ), and rabbit antis e rum to each e icosanoid w ere combine d and inc ubate d ove rnight at 8°C for LTB 4 dete rmination and inc ubate d for 2 hours at 25°C in a w ate r bath for PGE 2 de te rminatio n. Unbound radiolabe led PGE 2 w as re move d by the additio n of 250 m l of donke y anti-rabbit serum coate d onto magne tizable polymer partic le s, and LTB 4 by the additio n of 250 m l of 2% charc oal suspens ion c oate d w ith 0.4% dex tran to the tubes. Afte r ce ntrifug atio n the re sidual bound activ ity w as measure d in each tube by c ounting I 125 PGE 2 in a gamma scintillatio n counte r and H 3 LTB 4 in a beta scintillation counte r. Cross re activ itie s of the antis e rum w ith other eicosanoids w ere below 0.05%. The as say se nsitiv itie s w ere 1.25 to 160 pg/tube (PGE 2 ), and 1.6 to 200 pg/tube (LTB 4 ).

Statistics
Results are ex pre ssed as mean ± s.e .m. The re sults w e re analyz ed by Stude nt's t te st or by re pe ate d me as ure d ANOVA and c ompare d w ith stude nt New man Keuls te st. The chose n level of signific anc e w as 0.05.

Leukocyte influx and vascular permeability in the synovial fluid
To inve stigate the c ontrib ution of pre ssoric alte ration to the anti-inflammatory prope rtie s of chronic L-NAME tre atme nt, le ukoc yte influx and prote in leakage to the artic ular c avity w as quantifie d in the synovial fluid c ollec te d from animals tre ate d w ith L-NAME or D-NAME w ith or w ithout Enalap ril. Synovial fluid c ells and vascular permeability w ere measure d 24 hours afte r artic ular challe nge w ith mBSA. Pane l A (Fig. 1) show s the effe ct of L-NAME compare d w ith c ontrol animals in the re duc tion of cellular migratio n to the inflamed are a. The re duc ed migratio n occ urs mainly in the number of polymorphonuc le ar (PMN) leukocyte s, the c entral c ell in the acute p hase of the inflammatory re ac tion, although mononuc lear (MN) ce ll migratio n w as also affe c te d. The simultaneous adminis tratio n of Enalap ril w ith L-NAME partially re store s this parame te r. The ce llular influx to the artic ular cavity observe d in animals tre ate d w ith Enalap ril alone w as similar to that obse rve d in the animals tre ate d w ith L-NAME plus Enalap ril. The synovial fluid of the contro l joint, inje cte d w ith saline ex hibite d only re side nt cells (dotte d line ). Leukoc yte counts in the artic ular c avity of D-NAME tre ate d animals did not diffe r from the leukoc yte count in control anim als, w hich drank only w ate r 30 .
The vascular pe rmeability show n in pane l B w as quantifie d by Evans's blue method. Our re sults show ed a signific ant dec re ase in prote in leakage to the artic ular cavity in the synovial fluid c ollec te d from L-NAME tre ate d animals . The kne e joint inje c te d w ith saline show e d low le ve ls of dye bound to prote in. Similar finding s in the prote in leakage to the artic ular cavity w e re observe d in both D-NAME tre ate d anim als and those w hich drank only w ate r. The prote in leakage analyzed in the joint fluid of anim als tre ate d w ith Enalap ril or Enalap ril plus L-NAME did not diffe r from the c ontrol anim als. In all groups evaluate d the synovial me mbrane PMN c ell infiltrate show ed a similar patte rn to the PMN infiltrate observe d in the synovial fluid (data not show n). In a pre vious pap er w e show e d that the synovial membrane of anim als re c eiving mBSA intra-artic ularly had inte nse PMN ce ll infiltrate at 24 hours. Tre atme nt w ith the NOS inhibitor L-NAME, prior to the induc tion of arth ritis , gre atly re duce d the ce ll infiltratio n in the synovial me mbrane 24 hours afte r the induc tion of arth ritis . 30 Fig. 2, Pane ls A and B, show s the inflammatory parame te rs analyz ed in anim als simultane ously tre ate d w ith the B2 antago nist of Bradykinin HOE 140. The cellular influx and the prote in le akage analyzed in the synovial fluid of the se animals w e re re markably re duced w he n c ompare d w ith the same parame te rs observe d in animals w ithout HOE 140 tre atme nt. It is w orth noting that the HOE tre atme nt drastic ally re duced the inflammatory paramete rs me asure d in the synovial fluid of L-NAME plus Enalap ril tre ate d animals . The se re sults show the importanc e of the kinin syste m in the de ve lopme nt of AIA in rabbits.

Eicosanoids levels in the synovial fluid
The synovial fluid of animals tre ate d w ith L-NAME, D-NAME, Enalap ril and HOE140 w as c ollec te d, and the eic osanoids PGE 2 and LTB 4 as saye d by radioimmunoas say. The re sults are show n in Figs 3 and 4 re spective ly. The PGE 2 le vel w as signific antly re duce d w ith L-NAME tre atme nt, in ac cordanc e w ith our pre vious re sults. 30 The simultane ous tre atme nt w ith Enalap ril complete ly re ve rte d this inhibitio n. The tre atment of animals w ith HOE 140 drastic ally re duced the level of PGE 2 in all groups e valuate d. It is important to note that the same re sult e ven occ urre d in the group of animals tre ate d only w ith Enalap ril plus HOE 140. The level of LTB 4 show n in Fig. 4 ex hibite d a quite diffe re nt patte rn of alte ratio n. Although re duc ed afte r tre atme nt w ith L-NAME alone or combine d w ith Enalap ril, the level of LTB 4 w as re verte d to that observe d in c ontro l animals w hen the tre atment inc lude d HOE 140. Tre atme nt w ith D-NAME did not affe c t the synovial fluid le vels of e icosanoids Bra dy kin in in vo lvem e n t in in du c ed NO a n d PGE 2 pr od uc tio n Mediators of Inflammation · Vol 8 · 1999 w hen compare d w ith the re sults obtaine d in animals w hich drank only w ate r. 30

Discussion
Our pre vious re sults clearly implic ate nitric ox ide in the IL-1 induc ed PGE 2 p roduc tion in the synovial fluid of acute arthritis in rabbits. 30 In this study w e also demonstrate that tre atme nt w ith the NOS inhibitor L-NAME w as as sociate d w ith a re duction in some signs of ac ute artic ular inflammation in rabbits such as vascular pe rmeability and synovial membrane PMN ce ll infiltrate . In the ac ute phas e of AIA w e notic ed a re duction in the numbe r of leukocyte s c oming to the inflamed are a. This finding is consiste nt w ith the study of Belenky e t a l., 33 w ho found that NOS inhibitors atte nuate d chemotax is of unstimulate d and prime d PMN leukoc yte s. Kaplan e t a l., 34 using pe ripheral human ne utrophils inc ubate d w ith the NOS inhibitor, LNMA, de mons trate d a re duction of leukocyte che motax is to FMLP. This inhibitio n could be ove rc ome if L-arginine or dibutyryl cGMP w e re adde d togethe r w ith the LNMA. These studie s point out the importanc e of NO in the dire c tion of ce ll moveme nt. Our re sults cle arly show that unde r inflammatory challe nge the migrator y re sponse of PMN ce lls and the synovium PMN c ell infiltrate w e re re duce d in animals tre ate d w ith L-NAME. This sugge sts that the endothelial lining of ve sse ls w ithin the rabbit synovium may be an unusual microvasc ular ne tw ork. We hypothe sized that chronic tre atment w ith an NOS inhibitor promote s vasc ular pre ssoric alte ratio n that could be account for its anti-inflammatory action. The simultane ous tre atme nt of the se animals w ith Enalapril, an ACE inhibitor, re store d the ce llular influx , prote in le akage and eicosanoids produc tion almost to the normal range observe d in contro l arth ritic joints . BK induc e s inc re ase in the blood flow of the rabbit F. A. S. Pa lá c ios et al. synovium via activatio n of B2 re c eptors. 35 Then, w e also analyz ed the partic ipation of the kinin syste m through co-adminis tration of a B2 antagon ist of BK HOE 140 to the animals , before and during the arthritis de ve lopme nt. This tre atme nt promote d re markable re duction of c ellular influx , prote in leakage and PGE 2 le ve l in the synovial fluid. It is w orth noting that conc omitant adminis tration of HOE 140 comple te ly abolished the inc re ase of the inflammatory parame te rs promoted by Enalap ril to L-NAME tre ate d animals . NO appears to be gene rate d by the vasc ular endothelial ce lls via constitutive NO synthas e stimulate d by inflammatory mediators like histamine , 5 hydrox ytryptamine and Bradykinin or by the inducible enzyme carrie d to the inflammator y site by PMN ce lls. 36 The magnitude in the re duc tion of the inflammator y parame te rs observe d w ith HOE 140 tre atment suggests that NO gene ratio n in ex perime ntal arth ritic joints follow s BK stimulation of B2 re c eptors. The re sults obtaine d in all animals tre ate d w ith the B2 antago nist of Bradykinin irre sp ective of othe r tre atme nts w e re similar to the value s observe d in anim als tre ate d only w ith L-NAME and inc luded the level of NO 2 /NO 3 (data not show n). The only ex c eption w as in the le ve l of LTB 4 me as ure d in the ex pe rimental joints of HOE tre ate d animals , w hich re sult is compare d w ith those obtaine d w ithout HOE tre atment. This finding suggests a possible inhibitor y action of kinin syste m produc ts in the lipoox yge nase pathw ay. LTB 4 w ould be c onside re d a p ote nt chemotac tic in the ac ute phase of inflammatory re spons e. 37 We have re c ently found that LTB 4 le vels dec re as ed in the synovial fluid of untre ate d anim als at 24 hours of arthritis w hen compare d w ith LTB 4 le vels found in Bra dy kin in in vo lvem e n t in in du c ed NO a n d PGE 2  animals at 4 hours of arthritis (LTB) 4 4th h= 2.48±0.16 ng/ml; LTB 4 24th h = 0.95±0.14 ng/ml). 38 The re duce d levels of LTB 4 in anim als tre ate d only w ith Enalap ril c orroborate pre vious re sults in arth ritic patie nts tre ate d for arte rial hyperte nsion w ith ACE inhibitors . 39,40 The re duce d levels of LTB 4 and the impaire d migration of le ukocyte s to the artic ular cavity in L-NAME tre ate d animals suggest that another me diato r apart, LTB 4 possibly can be invo lve d in the ce ll attrac tion to the artic ular cavity in this phase of AIA development. The inte rac tions be tw een PGE 2 IL-1 and NO synthesis have been re late d in many studie s. 3,4,5,12,20,41 These studie s show that the e ndoge nous re le as e of NO e nhanc es COX ac tivity w ith furthe r inc re ase of PGE 2 . High c once ntrations of PGE 2 inhibite d NOS ex pre ssion and decre ase d NO synthe sis. 14 Whe n NO production w as blocked by NOS inhibitors both NO and PGE 2 simultane ously de cre ase d. 12,[16][17][18][19]25,28,42 NOS inhibitors also block cytokine induc ed re leas e of PGE. 11,13 In the se in v itro ex perime nts NO see ms to be one of the most important signals in the activatio n of c yclo-ox ygenas e to produc e PGE 2 . We pre viously found that inhibitio n of NO produc tion w ith L-arg inine antago nist L-NAME signific antly re duced IL-1b and PGE 2 le ve ls in the synovial fluid, w ith simultane ous re duc tion of total NO 2 /NO 3 and suppre ssion of AIA. 30 In this study w e verifie d that tre atment w ith an ACE inhibitor, Enalap ril, w as effic ie nt in the re version of inflammatory parame te rs in the synovial fluid induc e d by NOS inhibition. This effe ct may not be in conse que nce of arte rial pre ssure normaliz ation but due to Bradykinin accumulation at the site of injury and w as complete ly blocked by adminis tration of BK antagon ist HOE 140.
Toge the r our re sults show that the pre ssoric me chanis m ac ting in chronic adminis tratio n of L-NAME did not ac count for all anti-inflammatory propertie s of NOS inhibitors in the developme nt of AIA in rabbits . Additio nally w e demons trate the partic ipatio n of Bradykinin in the acute phase of antige n-induc e d arthritis in rabbits as w ell as the inte rre lations hip of the kinin syste m w ith NO and eic osanoids produc tion in the arthritic joint.