Inflammation in the Rat Knee Joint

Annals ofthe Rheumatic Diseases 1991; 50: 747-751 747

SCIENTIFIC PAPERS Ann Rheum Dis: first published as 10.1136/ard.50.11.747 on 1 November 1991. Downloaded from

Neurogenic component of different models of acute inflammation in the rat knee joint

Francis Y Lam, William R Ferrell

Abstract eliciting an inflammatory response.9 There is This study was performed to investigate increasing evidence indicating that substance P whether different models of acute joint in- contained in sensory C fibres plays an important flammation showed a neurogenic component part in what has come to be termed 'neurogenic and to establish whether this is mediated inflammation' for both the skin2 10-12 and the through sensory afferent or sympathetic joint.7 13 efferent nerve fibres. Intra-articular injection There is another body of thought which of 2% carrageenan, 20 pg substance P, 1% suggests that the sympathetic nervous system formalin, and 2% urate all produced an in- may contribute to neurogenic inflammation. flammatory response. Prior surgical dener- Evidence supporting this idea arises from the vation of the joint significantly inhibited this finding that vascular permeability is increased response in the carrageenan and formalin after sympathetic nerve stimulation'4 and that models, but not the others. Pretreatment of the baseline plasma extravasation in the knee the joint with 1% capsaicin (about one week joint of cats was substantially reduced after previously) significantly reduced the inflam- lumbosacral sympathectomy.'5 16 Chemical matory response in all models except forma- sympathectomy with 6-hydroxydopamine,17 lin. In animals pretreated long term with reserpine, 8 or a-methyl-p-tyrosine or guan- reserpine (to deplete sympathetic nerve end- ethidine'9 has been shown to inhibit potently ings of their neurotransmitters) significant plasma extravasation evoked by intradermal reductions occurred in the inflammatory injection of acetic acid, compound 48/80, poly- responses to substance P and urate. Intra- mixin B, histamine, and 5-hydroxytryptamine, articularinjection ofcompound48/80produced as well as that produced by thermal injury. a marked inflammatory response, which was Levine et al have also shown that chemical only significantly reduced by capsaicin pre- sympathectoniy reduces swelling and joint treatment. These results suggest that both the injury in rats with adjuvant arthritis.20 formalin and carrageenan models of inflam- The question of whether sensory nerve fibres mation depend to some extent on the integrity contribute to the inflammatory process in acute http://ard.bmj.com/ of the sensory innervation of the joint, and models of joint inflammation has only recently thus have a neurogenicaily mediated com- been considered. Carrageenan induced inflam- ponent to the inflammatory process they mation in the rat knee has been shown to have a generate. In these models there seems to be significant neurogenic component.2' Substance little contribution from sympathetic efferent P has also been shown to be effective in evoking fibres. Each model of inflammation showed a inflammatory processes in the rat knee joint.22 different pattern of response to the pre- It is not known, however, whether the neuro- on October 2, 2021 by guest. Protected copyright. treatments, suggesting that the mediators of genic component of these inflammatory the inflammatory process may differ in each responses is mediated by the unmyelinated C case. fibres or the sympathetic nerves supplying the rat knee, or both. This study is an attempt to identify the types of nerve fibres and the Involvement of the nervous system in inflam- importance of their contributions to the acute matory responses has been recognised for models of inflammation produced by car- almost a century and is based on the observation rageenan, substance P, formalin, and uric acid. that antidromic electrical stimulation of dorsal In addition, the effects of the mast cell degranu- roots or peripheral sensory nerves results in lator, compound 48/80, on plasma extravasation vasodilatation' and increased vascular per- in the rat knee was also investigated. meability2 in the territory innervated. Lewis,s5 in his classic experiments, first implicated the release of substances from the peripheral term- Materials and methods inals of nociceptive afferents in the production Experiments were performed on male Wistar of cutaneous wheal and flare responses. Since rats (-300 g) anaesthetised with a mixture of evidence has accumulated in support of urethane (- 13 g/kg) and diazepam (2-5 mg/kg). Institute of Physiology, then, University of Glasgow, the notion that the 'axon reflex' is mediated by Evans blue (Sigma; 75 mg/kg) was injected into G12 8QQ, UK unmyelinated afferent fibres. Neurogenic vaso- the external jugular vein. One knee was injected F Y Lam dilatation and plasma extravasation have been with 0-2 ml of 0-9% saline or vehicle to serve as W R Ferrell shown to occur during antidromic nerve stimu- a control while 0-2 ml of either 20 pg substance Correspondence to: Dr Lam. lation at C fibre strength in both the skin2 6 and P, or 1% formalin, or 2 mg compound 48/80 was Accepted for publication the joint.7 8 Direct activation of C polymodal injected into the other knee. At the end of a four 6 October 1990 nociceptive afferents in skin is also effective in hour period the animal was injected with 748 Lam, Ferrell

Euthatal (133 mg/kg) and exsanguinated. The effectiveness of reserpine treatment in depleting

joint capsule from the anterior and posterior sympathetic neurotransmitters of the articular Ann Rheum Dis: first published as 10.1136/ard.50.11.747 on 1 November 1991. Downloaded from regions of the knee was dissected free and nerves can be deduced by comparing the analysed for Evans blue content using a modi- frequency-response curve of reserpine treated fied technique,23 details of which have been rats with that of normal rats. reported previously.2' As the amount of tissue In a final group of rats the effects of the obtained from one rat is small, in any one substance P antagonist [D-Pro4,D-Trp7'9"l0] experiment the joint capsules from the knees of SP411 on the compound 48/80 inflammatory five- rats were used and Evans blue extracted response were investigated by intra-articular from this pooled sample. The Evans blue values injection of 100 [ig (in 0-1 ml volume) of the presented throughout represent the difference antagonist 15 minutes before injection of com- between the values obtained from knees pound 48/80 into the rat knee. injected with inflammatory agent and those For each experimental procedure (with the obtained from saline or vehicle injected (con- exception of the blood flow experiments, which trol) knees. The latter samples were necessary were performed on individual rats) three to six to control for the effect of fluid within the syno- groups offive rats were used. Data are presented vial cavity and the trauma associated with as the mean difference (SEM) in Evans blue penetration of the joint capsule. content between the control and test knee in The procedures for investigating the effects each group. Differences were considered of 2% carrageenan and 2% uric acid were significant if the p values were 5% or less essentially the same as above, except that the (unpaired t test). drugs were allowed to remain in the knee joint for 24 hours and Evans blue injected into the animal four hours before the end of the experi- Results ments. Carrageenan and uric acid were found to In the present experiments it was found that require a longer period in the synovial cavity (24 intra-articular injection of substance P (four hours) to evoke comparable inflammatory hours), carrageenan (24 hours), formalin (four responses. hours), and uric acid (twenty four hours) were In one group of rats 0-2 ml of 1% capsaicin all effective in producing marked plasma extra- (0-02 g capsaicin (Sigma) dissolved in 0-1 ml vasation into the joint capsule. Intra-articular absolute alcohol, 0-1 ml Cremophor (Sigma), and 1-8 ml 0 9% saline) was injected into the 200 200 synovial cavity of one knee under general i anaesthesia with Hypnorm (0 1 mg/kg) and diazepam (0 5 mg/kg). The rats were allowed to ° 150- c 150i

recover and after 9-14 days their response to cn intra-articular injection of the inflammatory cn 100- Z agents into the capsaicin pretreated knee was a) B00

tested. http://ard.bmj.com/ co 0tn In another group of rats the nerves supplying >& 50- the knee joint were transected unilaterally I~~~~~~C under general anaesthesia (Hypnorm 0 1 mg/kg; Ow diazepam 0-5 mg/kg) and the animals then 2% C DEN RES allowed to recover. After 11-14 days these 'AP 2[]tg CAP Car SP 20010 animals were assessed for their response to the TO.

inflammatory agents injected into the denervated on October 2, 2021 by guest. Protected copyright. knee. e /.+1 In a further group of rats reserpine (1 mg/kg) a1 150 ' 150' (5 mg reserpine dissolved in 1 ml absolute cm alcohol and 4 ml (20%) ascorbic acid) was given .a1001 intraperitoneally for three successive days. On a) 0 the fourth day the response of these animals to enoo the inflammatory agents was assessed. In > 50 I 50- addition, the articular blood flows of six reserpine treated rats in response to electrical stimu-

lation were measured individually. These 17o UAP DLEN R1ES Zo CAP DNRtES:a experiments were performed by laser Doppler Form Urate Instruments on rats flowmetry (Moor MBF3) Figure 1 Effect ofintra-articular injection ofdifferent under general anaesthesia with a mixture of inflammatory agents on Evans blue content ofthejoint sodium pentobarbitone (14 mg/kg), urethane capsule. Effect of(A) 2% carrageenan (Car); (B) 20 lAg (142 mg/kg), and diazepam (0-32 mg/kg). The substance P (SP); (C) 1%formalin (Form); (D) 2% urate. In each histogram thefirst value gives the control response to procedure essentially involved electrical stimu- tnjection ofthese agents alone. The effect ofvarious lation of the nerve supply to the anteromedial pretreatments on these responses is given by the adjacent aspect of the knee while monitoring articular values. CAP=intra-articular injection of1% capsaicin blood flow a about one week previously; DEN=surgical transection ofthe by laser probe placed on the nerve supply to the kneejoint 10 dayspreviously; surface, but not in contact with the anterior RES=reserpine pretreatment. *p<0*05, **p<0*01, portion of the knee joint. A frequency-response * * *p<0-001: differs significantlyfrom control. curve was obtained by monitoring the changes *p<0 05, * -p<0 01, *--p<0 001: capsaicin treatment differs significantlyfrom denervation treatment. fp<0 05, in the articular blood flow with different ffp<0o01: reserpine treatment differs signifcantlyfrom frequencies of stimulation (1-50 Hz). The denervation treatment. Neurogenic component ofdifferent models ofacute inflammation 749

3 min injection of 1% capsaicin 9-14 days earlier Normal m

resulted in a significant reduction (36%) of the 1.0- Ann Rheum Dis: first published as 10.1136/ard.50.11.747 on 1 November 1991. Downloaded from carrageenan induced inflammatory response. Surgical denervation also produced a significant ] 25% reduction in this model, but chemical 2 I1Hz 5Hz 10Hz 20Hz 50Hz sympathectomy failed to produce a significant reduction of the inflammatory response (fig ii Reserpine treated IA). .0C 1.0-i Substance P induced inflammation was almost completely abolished (97% reduction) 4 0.5- after capsaicin pretreatment, whereas denervation of the rat knee did not alter the response to O 1OHz 30Hz 50Hz 70Hz substance P. Chemical sympathectomy with Figure 3 Responses to electrical stimulation (10 V; I ms reserpine, on the other hand, produced a width) ofthe nerve supply to the knee in a normal rat (upper significant 50% reduction ofsubstance P induced trace) and in a reserpine pretreated rat (lower trace) at inflammation (fig 1B). differentfrequencies. Black bars denote periods of Neither capsaicin pretreatment nor chemical stimulation. sympathectomy by reserpine produced any significant effects on the formalin induced inflammation, but denervation of the rat knee In the normal rat electrical stimulation of the markedly reduced (57%) the formalin response nerve supply to the knee (10 V, 1 ms width) was (fig IC). With uric acid induced inflammation, effective in reducing articular blood flow with a capsaicin pretreatment produced a significant threshold frequency of 2 Hz and maximal 46% reduction, whereas knee joint denervation reduction at 30 Hz, as illustrated for one animal failed to attenuate the uric acid response in fig 3 (upper trace). In five out of six reserpine significantly (fig ID). Reserpine pretreatment treated rats no vasoconstrictor response to did produce a reduction in the inflammatory articular nerve stimulation was seen (fig 3, response, but this was not significant compared lower trace). The only one of this group of rats with the response to uric acid alone, but just which responded to nerve stimulation showed a reached significance compared with the effect of much smaller maximum vasoconstriction and a uric acid in the denervated knee. higher stimulation threshold (10 Hz) than Compound 48/80 when injected into the rat shown by normal animals. knee elicited a very potent inflammatory action, which was not affected by substance P antagonist pretreatment, reserpine pretreatment, or Discussion denervation of the rat knee. Inflammation Capsaicin is known to produce lasting depletion induced by compound 48/80, however, was of unmyelinated afferent fibres when applied significantly reduced (26%) after capsaicin topically to peripheral nerves.24 The action of a

treatment (fig 2). drug is expected to be attenuated after capsaicin http://ard.bmj.com/ treatment if its action is dependent upon the 350- integrity of the unmyelinated afferent fibres supplying the tissue. Surgical removal of the nerves innervating that particular tissue should 300- also result in a similar attenuation on the response. In these studies carrageenan induced inflammation of the rat knee was found to joint on October 2, 2021 by guest. Protected copyright. 250- follow this expected pattern of action, which is also in accord with our previous findings using carrageenan over a shorter period.2' 200 . , Parratt and West have shown that reserpine treatment is a simple and effective method for 10, depleting sympathetic nerve endings of their neurotransmitters.25 Our studies confirmed this 0 0 view as five of the six reserpine treated rats tested with laser Doppler flowmetry showed no vasoconstrictor response to articular nerve stimulation. The lack of effect of reserpine pretreatment in inhibiting some of the inflammatory responses therefore cannot be ascribed to failure of depletion of neurotransmitters from sympathetic nerve endings. Carrageenan

0 induced inflammation of the rat knee was not ;;Pvdnr% WvOUAQ10n CAPP-AD2 SPAQOTA ncoi Uhflnc affected by reserpine treatment. This suggests Figure 2 Evans blue content ofthe rat kneejoint capsule that the sympathetic nervous system may not be (test-control response)four hours after intra-articular involved in the carrageenan model of inflam- administration of2 mg compound (CPD) 48180. Although pretreatment by surgical denervation (DEN), reserpine mation. (RES), and!prior intra-articular injection of100 ,Ig of[D- The dose of substance P injected into the Pro4,D-Trp 9"'0]SP4,1, a substance P antagonist (SPA), synovial cavity (20 ,ug) was chosen to give an all resulted in a reduction in the inflammatory response, these with that were not significant exceptfor capsaicin pretreatment (CAP). inflammatory response comparable *p<00S: differs significantly from control response. elicited by the other agents, and cannot be 750 Lam, Ferrell

considered physiological. This was dispersed contents or their numbers in the knee joint as

around the joint cavity, however, and the actual the response to compound 48/80 was found to Ann Rheum Dis: first published as 10.1136/ard.50.11.747 on 1 November 1991. Downloaded from concentration at the effector sites cannot be be similar in intact and denervated knees. determined. Assessments as to the physiological Inhibition of formalin induced inflammation dosage of administered agents are difficult to after denervation is therefore more appropriately make as the effectiveness of the agent depends explained by the lack ofnerve innervation in the on the tissue concentration achieved at the knee joint. The failure of capsaicin to attenuate target tissue and this may be dependent on the the inflammatory response highlights the fact point of delivery of the agent. In the case of that this treatment is probably not exertmg a substance P its release from C fibres may occur neurotoxic effect on articular nerve fibres with in close proximity to its target cells (mast cells the intra-articular mode of administration but and blood vessels), thereby achieving high may instead act on subsets of cells mediating the concentrations locally. Under these circum- inflammatory responses. Hence although stances, to obtain comparable concentrations at formalin induced inflammation is dependent these sites by intra-articular administration on the nerve supply to the knee, it is not would require a higher concentration of significantly affected by capsaicin pretreatment. administered substance P. The sympathetic nervous system does not seem It was of interest to find that inflammation to play a part in the formalin induced inflam- induced by substance P and uric acid were both mation as no significant alteration ofthe formalin reduced after capsaicin treatment but were not response was noted after reserpine treatment. affected by denervation of the knee. The period Compound 48/80 induced inflammation was of denervation was checked to be sufficient for reduced after capsaicin treatment. Direct substantial degeneration of the nerves to occur, inhibition of the action of compound 48/80 on as judged by their electron microscopic appear- mast cells by capsaicin is unlikely as capsaicin ance. The inhibitory effects of capsaicin on was given at least nine days before compound substance P and uric acid induced inflammation 48/80 injection. It is more likely that the mast cannot be attributed to its neurotoxic effects as cell contents or numbers were reduced as a complete denervation of the knee was ineffec- result of capsaicin administration. The effect of tive. Possibly capsaicin may have direct in- compound 48/80 is also independent of the hibitory actions on the target cells which sub- substance P content in the rat knee as the stance P acts upon.22 The present findmgs response to compound 48/80 was the same in suggest that capsaicin may also be inhibiting the either the absence or presence of the substance target cells that uric acid acts upon. P antagonist. Chemical sympathectomy by reserpine pre- We initially concluded that capsaicin treat- treatment showed a significant reduction of the ment and surgical denervation were equivalent substance P induced inflammation. This is procedures and thus gave rise to similar unlikely to be due simply to the depletion of inhibition of the carrageenan induced inflam- sympathetic neurotransmitters by reserpine, matory response.2' Clearly, however, although

however, as complete surgical denervation of this may be true in some models of inflam- http://ard.bmj.com/ the knee failed to influence the substance P mation, it does not hold in others. Although induced response. One of the actions of sub- capsaicin may exert neurotoxic effects, this is stance P is to promote mast cell degranulation,26 not always the sole explanation of its inhibitory but the possibility of reserpine interfering with actions in the different models of acute inflam- this process is also unlikely, as the present mation. The substance P and uric acid inflam- studies have shown that the actions of a specific matory models do not seem to have a neurogenic

mast cell degranulator, compound 48/80, was component as they were unaffected by nerve on October 2, 2021 by guest. Protected copyright. not affected by reserpine pretreatment. It is transection, whereas the presence of intact therefore more likely that reserpine is acting innervation is important for carrageenan and through inhibitory actions on other mediators of formalin to exert their full inflammatory actions. the substance P response and perhaps on the Of those inflammatory models which have a target cells that substance P affects. In addition neurogenic component, our findings suggest to promoting mast cell degranulation, substance that it is the unmyelinated C fibres rather than P also causes inflammatory cell chemotaxis,27 the sympathetic efferent fibres that mediate, at neutrophil activation,28 and fibroblast prolifer- least in part, these inflammatory processes. ation,29 all of which are recognised components of the arthritic inflammatory process. Substance Financial support from the Arthritis and Rheumatism Council P has also been shown to activate synoviocytes and ICI Pharmaceuticals is gratefully acknowledged. to secrete prostaglandin E2 and collagenase,30 as well as to stimulate secretion of interleukin-l- 1 Bayliss W M. On the origin from the spinal cord of the like activity from macrophages.31 It is therefore vasodilator fibres ofthe hindlimb and on the nature ofthese fibres. J Physiol (Lond) 1901; 26: 173-209. possible that reserpine pretreatment inhibits 2 Jancso N, Jancso-Gabor A, Szolcsanj J. Direct evidence for one or more of these processes and thus reduces neurogenic inflammation and its prevention by denervation and pretreatment with capsaicin. Brj Phannacol 1%7; 31: the substance P mediated inflammatory 138-51. response. 3 Lewis T. The blood vessels of human skin and their responses. London: Shaw, 1927. In an inverse manner to that of the substance 4 Lewis T. Experiments relating to cutaneous hyperalgesia and P and uric acid induced inflammation, formalin its spread through somatic nerves. Clin Sci 1936; 2: 373-417. induced inflammation was not affected by 5 Lewis T. Pain. New York: Macmillan, 1941. capsaicin pretreatment, but was significantly 6 Gasser H. Unmedullated fibres originating in dorsal root ganglia. Y Gen Physiol 1950; 33: 651-90. reduced by surgical denervation. Denervation 7 Ferrell W R, Russell N J W. Extravasation in the knee of the knee could not have altered the mast cell induced by antidromic stimulation of articular C fibre Neurogenic component ofdifferent models ofacute inflammation 751

afferents ot the anaesthetised cat. J Physiol (Lond) 1986; inflammatory activity of a-adrenoceptor blocking agents. 379: 407-16. BrJ Pharmacol 1974; 51: 45-53. 8 Ferrell W R, Cant R. Vasodilation of articular blood vessels 20 Levine J D, Moskowitz M A, Basbaum Al. The contribution Ann Rheum Dis: first published as 10.1136/ard.50.11.747 on 1 November 1991. Downloaded from induced by antidromic electrical stimulation of joint C of neurogenic inflammation in experimental arthritis. J fibres. In: Schmidt R F, Schaible H G, Vahle-Hinz C, eds. Imnmol 1985; 135: 843s-7s. Fine afferent nerve fibres and pain. Weinheim, FRG: VCH 21 Ferrell W R, Lam F Y. Inhibition of carrageenan induced Verlagsgesellschaft, 1987: 187-92. inflammation in the rat knee joint by substance P antagonist. 9 Kenins P. Identification of the unmyelinated sensory nerves Ann Rheun Dis 1989; 48: 928-32. which evoke plasma extravasation in response to antidromic 22 Ferrell W R, Lam F Y. Capsaicin suppresses substance P- stimulation. Nerosci Leu 1981; 25: 137-41. induced joint inflammation in the rat. Neurosci Lett 1989; 10 Lembeck F, Holzer P. Substance P as neurogenic mediator of 105: 155-8. antidromic vasodilatation and neurogenic plasma extra- 23 Harada M, Takeuchi M, Fukao T, Katagiri K. A simple vasation. Naunyn Schmiedebergs Arch Pharmacol 1979; 310: method for the quantitative extraction of dye from skin. 7 175-83. Pharm Pharmacol 1971; 23: 218-9. 11 Gamse R, Holzer P, Lembeck F. Decrease of substance P in 24 Lynn B, Pini A. Long-term block of afferent C-fibres primary afferent neurones and impairment of neurogenic following capsaicin treatment in the rat. J7 Physiol (Lond) plasma extravasation by capsaicin. BrJ Phmacol 1980; 1985; 362: 19P. 68: 207-13. 25 Parratt J R, West G B. Release of 5-hydroxytryptamine and 12 Foreman J C. Peptides and neurogenic inflammation. BrMed histamine from tissues of the rat. 7 Phsiol (Lond) 1957; BuU 1987; 43: 386-400. 137: 179-92. 13 Yaksh T L, Baily J, Roddy D R, Harty G J. Peripheral 26 Mazurek N, Pecht I, Teichburg V I, Blumberg S. The role of release of substance P from primary afferents. In: Dubner the N-terminal tetrapeptide in the histamine-releasing R, Gebhart G F, Bond M R, eds. Proceedings of Vth World action ofsubstance P. NeuropharmacoLogy 1981; 20: 1025-7. Congress on Pain. 1988: 51-4. 27 Marasco W A, Showeil H J, Becker E L. Substance P binds 14 Linde B, Chisolm G, Rosell S. The influence of sympathetic to the formyl peptide chemotaxis receptor on the rabbit activity and histamine on the blood-tissue exchange of neutrophil. Biochem Biophys Res Commun 1981; 99: solutes in canine adipose tssue. Acta Physiol Scand 1974; 1065-72. 92: 145-55. 28 Bar-Shavit Z, Goldman R, Stabinsky Y, et al. Enhancement 15 Engel D. The influence of sympathetic nervous system on of phagocytosis-a newly found activity of substance P capillary permeability. J Physiol (Lond) 1941; ": 161-81. residing in its N-terminal tetrapeptide sequence. Biochem 16 Engel D. The influence of the sympathetic nervous system on Biophys Res Commun 1980; 94: 1445-51. capillary permeability. Res Exp Med (Berl) 1978; 173: 1-8. 29 Nilsson J, von Euler A M, Dalgard C J. Stimulation of 17 Helme R D, Andrews P V. The effects ofnerve lesions on the connective tissue cell growth by substance P and substance inflammatory response to injury. J Neuosci Res 1985; 13: K. Nature 1985; 315: 61-3. 453-9. 30 Lotz M, Carson D A, Vaughan J H. Substance P activation of 18 Gozsy B, Kato L. Role of norepinephrine and 5-hydroxy- rheumatoid synoviocytes: neural pathway in pathogenesis tryptamne in the delayed phase of the inflammatory of arthritis. Science 1987; 235: 893-6. reaction in rats. Int Arch AUergy Apply Immunol 1966; 30: 31 Kimball E S, Perisco F J, Vaughan J H. Substance P, 553-60. neurokinin A and neurokinin B induce generation of IL-1- 19 Green K L. Role of endogenous catecholamines in the anti- like activity in P388Dl cells. IImmunol 1988; 141: 3564-9. http://ard.bmj.com/ on October 2, 2021 by guest. Protected copyright.