Local haemorrhage induced by Bothrops jararaca venom: relationship to neurogenic inflammation.

We investigated morphological alterations induced by s.c. injection of 2.5 microg of Bothrops jararaca venom in rats. Intense disorganisation of collagen fibres was observed 1 min after the venom injection, particularly at regions near vessels and nerves. Mast cells were degranulated, and erythrocytes were seen leaving venules throughout the endothelial junctions. At this time, damaged endothelial cells were not observed. In rats envenomed as above, but immediately after cardiorespiratory failure induced by deep ether anaesthesia, alterations in the connective tissue structures, as previously described, were not observed. The mediation of this haemorrhage was investigated by injecting the venom into the foot pad of mice and compared to the mediation of oedema. Local haemorrhage was significantly reduced in mice pre-treated with capsaicin or guanethidine or submitted to a surgical section of sciatic and saphenous nerves. In these animals, oedema was not affected. Groups treated with methysergide or morphine showed both haemorrhage and oedema significantly reduced. Indomethacin or dexamethasone pre-treatments significantly reduced the oedema, but not the haemorrhage. Moreover, in animals treated with promethazine or mepyramine, oedema and haemorrhage were not affected. These data suggest that local haemorrhage induced by Bothrops jararaca venom is partially controlled by serotonin and neurohumoral mediators. Furthermore, results indicate that haemorrhage and oedema are mediated by different pharmacological systems.


Introduction
Haemorrhage and oe de ma, besides blood c oagulation disturbance s, are the main signs of e nve noming by Bo thro p s snake venoms. 1,2 De spite serum therapy being an effective tre atment for the control of systemic symptoms in this type of e nve noming, it doe s not control the oedema and local haemorrhagic and necrotic lesions w hich de ve lop at the site of venom inoculation. 1 ,3 -5 Among the tox ins re sponsible for these local effects, hae morrhagic factors w ere isolated from various viperid venoms. 6 Eve n though some of the se haemorrhagic fac tors are devoid of p rote olytic ac tivitie s, 7-11 the y are charac te rise d as me talloproteinases. 6 Some of the se factors digest prote ins from the ex trace llular matrix , and this prop erty has be en related to the pathogeny of haemorrhagic lesions. 6,1 2 -17 On the other hand, e fforts have also bee n made to charac te rise the pharmac ological me diators involve d in oedema formation, 18 -23 pain 2 4 and other para-me te rs of the inflammatory response 2 5 -29 induce d by vipe rid venoms. Ne ve rthe less, all this information is not suffic ie nt to draw a compre hensive me chanism to ex plain the local le sions observed in this type of envenoming.
In the pre sent study it is show n that oedema and haemorrhage induc ed by B. ja ra ra ca venom have distinc t pharmacologic al me diation and our re sults suggest that hae morrhage is partially c ontrolled by se rotonin and neuroge nic me diators.

Venoms
A pool of lyophilised venom obtained from adult spe cime ns of B. ja ra ra ca snake s, supplied by the Laboratory of Herpe tology, Instituto Butantan, w as used throughout this inve stigation. The ve nom w as ke pt at -20°C, and venom solutions w ere p re pare d (w /v ) w ith sterile saline just before use .

Animals
Male Wistar rats (220-250 g ) or male Sw iss mice (18-22 g ) w ere used. The animals w ere maintaine d w ith w ater and food a d libitu m in app ropriate environment conditions and used under e thic al c onditions, acc ording to guide lines of the International Socie ty of Tox inology. 30

Morphology of venom-induced lesions
Rats, maintaine d unde r e ther anaesthe sia, w ere inje cte d into the subcutaneous tissue of the sc rotal bag w ith 2.5 m g of venom (50 m l) in saline . Control rats rec eived the same volume of saline. One minute after the inje ction, animals w ere killed and samples of tissue from inje cte d are as w ere obtained. In orde r to e valuate the role p laye d by blood circulation in the e volution of the venom-induced lesions, the same proce dure as above w as performed in rats w hose he artbeats had just stoppe d due to prolonged e ther anae sthesia. Tissue samples from the sites of venom injection w e re c ollec te d 1 min afte r the injec tion and fix ed in 2.5% buffered glutaraldehyde . Routine double fix ation, e mbedding in Araldite and staining w ith uranyl ace tate and le ad citrate w ere employe d. The proce sse d samples w ere analysed in a Phillips (EM-201 ) ele ctron microsc ope.

Evaluation of haemorrhage and oedema
Loc al hae morrhage and oedema w ere simultane ously e valuate d using a modification of the me thods de scribed by Ow nby e t a l. 3 1 and by Yamakaw a e t a l. 3 2 Mice w ere injected s.c. in the right foot pad w ith 50 m l of the venom solution. The c ontra-lateral foot pad rec eived the same volume of ste rile saline. Three hours later (time determine d by a time-course ex perime nt for hae morrhage and oedema ), the animals w ere sacrificed in an ether chambe r. The paw s w ere remove d at the le ve l of the tibio-tarsic junction, w eighted, fragme nte d and put in tube s containing 4 ml of von Kampen -Zijlstra re agent modified by Mats ubara e t a l. 33 6 ], 50 mg KCN, 120 mg KH 2 PO 4 , 50 mg NaCl and 1 ml non-ionic dete rgent per litre ). The tissue w as homoge nise d and the tube s w e re ce ntrifuge d at 3000 g for 30 min. The supe rnatant w as filtered in Millip ore me mbranes (0.8 m pore s ), and the absorbanc e at 540 nm w as dete rmined in a spec trophotomete r (Micronal B382, Braz il).
The haemoglobin conc entration w as dete rmine d as follow s: In order to e liminate the influe nce of oedema, the conc entration of hae moglobin obtaine d in the paw injected w ith ve nom w as divided by the w eight of the paw injec te d w ith saline . The difference betw e en haemoglobin c ontent of the paw injected w ith venom and of that injected w ith saline w as the e stimated haemorrhage, ex pressed as mg Hb /g of tissue. The minimal hae morrhagic dose (MHD) w as defined as the minimal c once ntration of venom able to induc e an inc rease of thre e times the haemoglobin conc entration in re lation to that of a c ontrol tissue. Finally, oedema, ex pre sse d in mg, w as evaluate d by the difference in w eight be tw ee n one paw injec te d w ith venom and the othe r inje cte d w ith saline. 3 2

Statistics
Results w ere analysed by Student's t-test and considered significantly different w he n p< 0.05.

Morphology of the envenomed connective tissue
Dis organisation of collage n fibres, particularly near blood vessels and nerve bundles ( Fig. 1A and 1B, respective ly ), mast ce ll degranulation (Fig. 1C) and inte nse vasc ular congestion (Fig. 1D) w ere observe d 1 min afte r venom inje ction into the subcutane ous tissue of the sc rotal bag of rats. Morphologic al alterations in venular endothe lial ce lls w e re not clearly detected, although red blood cells w e re se eing squee zing out through open e ndothelial cell junctions, characte rising a pe r dia pe de s is type of haemorrhage (Fig. 2 ). Alte rations such as disorganisation of collagen fibres and mast ce ll de granulation w ere not observe d w hen the venom w as injected into the s.c. tissue of clinically dead animals.

Time-course of haemorrhage and oedema induced by the venom
Haemorrhage induc ed by 1.25 m g of venom show ed a peak at the 3rd hour after injec tion, persisting up to 6 h. Oe de ma w as max imal from 30 min up to 3 h, decre asing afte r 6 h of ve nom inje ction. As both haemorrhage and oedema w ere max imal 3 h after venom inje ction, this time w as sele cte d to e valuate some parame te rs of the pharmacologic al mediation of these proc esses.
The hae morrhagic activity w as dose-depe nde nt, and the MHD obtained w as 0.95 m g. A dose of 5 m g of venom w as use d for the study of the pharmacologic al me diation of the le sions.

Mediation of venom-induced haemorrhage and oedema
With the dose of venom used, only local e ve nts w ere e valuate d, as it w as considered that envenome d mic e do not prese nt such syste mic effe cts as blood inc oagulability or thrombocytopenia (data not show n ). Dex amethasone and indome thac in treat- me nts did not alte r the hae morrhagic p attern induce d by the venom, but signific antly inhibite d oedema (Fig.  3 ). Pre-treatment w ith morphine or methyse rgide inhibited both hae morrhage and oedema (Fig. 4 ), and guane thidine , surgic al denervation and capsaicin tre atments significantly inhibited haemorrhage, but not oede ma induce d by the venom (Fig. 5 ). In promethazine or mep yramine treated animals, oedema and haemorrhage w e re not affe cte d.

Discussion
Loc al hae morrhage induce d by snake venoms is cre dited to metalloprote inases prese nt in the se venoms. 6 This effe ct has been associated w ith the ac tivity of the se tox ins on prote ins of the ex trac ellular matrix . 6,12 In the pre sent study, the injection of a low dose of B. ja ra ra c a venom into the conne ctive tissue of the sc rotal bag of rats induced prompt and significant alterations to the morphology of this tissue. The premature manifestation of these alterations contrast w ith the long time (hours ) needed for the in v itro degradation of prote ins of ex tracellular matrix by haemorrhagic factors isolate d from vip erid ven- oms. 1 3 -1 7 Moreove r, the absenc e of alterations such as disorganisation of collage n fibre s and degranulation of mast c ells, w he n the venom w as inje cte d into the conne ctive tissue of rats imme diately after ceasing he artbeat, sugge st that such an e ffec t de pends on the me diation of some fac tor(s ) libe rate d from inje cte d tissue, and possibly from the ac tive blood circ ulation.
The p ossible particip ation of me diators from the damage d tissue in the pathogene sis of local haemorrhage induce d by the ve nom w as tested in paw of mic e submitted to pharmac ological or surgical treatme nts and compare d w ith the me diation of the oedema induc ed by the ve nom in the same paw. As show n in the Results sec tion, the major me diators of oedema induce d by this ve nom w ere derivatives of arachidonic ac id, sinc e treatme nts w ith dex amethasone (phospholipase A 2 inhibitor ) or indomethacin (cycloox igenase inhibitor ) w ere e ffec tive in inhibiting this effect. These results are in agre eme nt w ith data from Tre bien and Calix to 1 9 and Pe rales e t a l. 21 The participation of the se mediators in the proce ss w as also demonstrate d for othe r viperid venom-induce d oedema. 18 ,23 While histamine se ems to play a minor role in oe de ma induc ed by B. ja ra ra c a ve nom in rats, 19 in mic e this vasoactive amine doe s not participate at all in the proc ess, confirming previous observations made by Pe rales e t a l. 21 By contrast to the observations of the se authors, w e found that se rotonin partially me diate s venom-induce d oe de ma, considering that pre-treatment of the animals w ith me thysergide dec reased the intensity of the process.
The oe de ma induce d by B. ja ra ra c a ve nom see ms to be also mediate d by opioid re ce ptors, considering that treatme nt of the animals w ith morphine inhibits the proce ss. This inhibitory effect of morphine w as also demonstrated for oe de ma caused by other vipe rid venom (Trim e re s u ru s fla vo v iridis ) 20 , as w e ll as that c ause d by other flogistic agents such as carrageenan. 3 4 Neve rthe le ss, the me chanisms w hich underlie the anti-oe de matoge nic effect of morphine are not yet understood.
Our re sults indic ate that the mediation of oedema formation differs from that involved in the induction of haemorrhage . Based on results obtaine d in mic e pre-treated w ith dex ame thasone or indomethacin, arachidonic acid de rivatives see ms not to participate in the loc al haemorrhage induce d by B. ja ra ra ca venom, confirming previous observations by Pe rale s e t a l. 21 Converse ly, in mice pre-treate d w ith guane thi- dine, capsaicin or surgic al de nervation, loc al haemorrhage but not oe de ma w as attenuate d. On the othe r hand, methyse rgide or morphine significantly diminished both oe de ma and hae morrhage, indic ating that, at least in mic e, serotonin and opioid rec eptors are partially implic ated in both effects. The fact that partial inhibition w as obtained w ith these treatments suggests that haemorrhage and oe de ma are indee d multimediated phenome na.
The p artic ipation of opioid rece ptors in inflammatory eve nts, as w e ll as the re lationship betw e en antidromic stimulation of sensitive ne rve fibre s and vascular perme ability disturbance , have bee n demonstrate d. 2 0, 3 4 -43 It is know n that neurope ptides, particularly substance P, can induc e a direc t vasc ular disturbance or may act indirec tly by induc ing libe ration of histamine from mast c ells. 3 9 Added to this hypothesis is the fact that the blockade of opioid rec eptors (e .g. tre atment w ith morphine ) can pre ve nt the liberation of substance P. 44 ,45 These results, and the inhib ition of haemorrhage obse rved in animals tre ate d w ith c apsaic in or w hich are chronic ally dene rvated, strongly indic ate that hae morrhage induc ed by the venom could be classified as a ne urogenic type of hae morrhagic inflammatory response. Re inforcing this interpretation is the fac t that ne uroge nic vascular disturbance s can be inhibite d by sectioning nerve routes or by chemic al depletion of neurop eptides after capsaicin tre atment. 3 6 Taken togethe r, the se data might ex p lain the ex plosive alterations in the connec tive tissue of rats 1 min afte r ve nom injec tion. Maluc elli and Mariano 4 0 also observe d a similar p henome non of instantane ous haemorrhagic re sponse on the diaphragm of guine a pigs after e le ctric or chemic al stimulation of the phre nic ne rve. These authors described a pe r dia p ede s is haemorrhage in diap hragmatic vessels. Furthermore, this type of haemorrhage w as also observe d after T. fla vo viridis venom inje ction. 4 6 Conversely, afte r intramusc ular injection of various vipe rid c rude ve noms or isolated haemorrhagic factors, the usual feature is a pe r rh ex is hae morrhage observe d in c apillarie s. 4 7-4 9 Whe the r these differences are due to the dose of venom injected, the route of inje ction or other pe culiar charac te ristics of the microvasculature in these tissue s remains to be investigate d. It must be considered that pe r dia pe des is hae morrhage is mainly observed in venules w hile pe r rh ex is hae morrhage is reported to occ ur mainly in capillaries. Nevertheless, only vessels greater than 20 m m in diame te r, like venules, are innervate d. 5 0 This fact and results he re de sc ribe d indic ate that pe r dia pe d e s is hae morrhage induce d by B. ja ra ra ca venom can be c ontrolled by neuroge nic fac tors. Furthe r studie s should be made in orde r to identify the spec ific neuroge nic fac tor(s ) involved in this proce ss.
In c onclusion, our re sults sugge st that loc al hae morrhage induce d by B. ja ra ra c a ve nom is partially me diated by se rotonin and by neuroge nic fac tors. Furthe rmore , the y show that oedema and loc al haemorrhage induce d by this venom have distinc t che mical me diation.