Research Paper Mediators of Inflammation, 7, 339–346 (1998)

We have provided evidence that: (a) lethality of mice to crude Bothrops venom varies according the isogenic strain (A/J > C57Bl/6 > A/Sn > BALB/c > C3H/HePas > DBA/2 > C3H/He); (b)BALB/c mice (LD50=100.0 microg) were injected i.p. with 50 microg of venom produced IL-6, IL-10, INF-gamma, TNF-alpha and NO in the serum. In vitro the cells from the mice injected and challenged with the venom only released IL-10 while peritoneal macrophages released IL-10, INF-gamma and less amounts of IL-6; (c) establishment of local inflammation and necrosis induced by the venom, coincides with the peaks of TNF-alpha, IFN-gamma and NO and the damage was neutralized when the venom was incubated with a monoclonal antibody against a 60 kDa haemorrhagic factor. These results suggest that susceptibility to Bothrops atrox venom is genetically dependent but MHC independent; that IL-6, IL-10, TNF-alpha, IFN-gamma and NO can be involved in the mediation of tissue damage; and that the major venom component inducers of the lesions are haemorrhagins.


Introduction
Bite s by snakes be longing to the genus Bo thro p s sp. can re sult in an immediate loc al lesion. This p rogre sse s from an acute inflammatory re ac tion to an impre ssive tissue de struction. 1 Syste mic effec ts of the venom at various grade s of se ve rity include blood inc oagulability and acute kidne y failure. 2 ,3 Although some venom c ompone nts are e ndow e d w ith the ability to re p roduce in ex pe rimental animals some loc al or syste mic lesions such as hae morrhage (haemorrhagins) and blood incoagulability (thrombin-like factors) re spe ctively, observations in human victims have be e n isolate d and characte rize d, the intimate e ve nts elapsing from the venom pe ne tration to the establishment of the lesions are not ye t know n. Among the se are included the venom compone nts and the host mediators involve d in the re cruitme nt of leukoc yte s from the c irc ulation to the site of the tissue injury, the elicitation of spec ific leukoc yte populations, activation or damage of the e ndothelial ce lls lining the p ost c apillary venules at the site at w hich ex travasation of leukocyte s and/or erythrocytes occ ur. 4 In order to address these issue s, basic information obtain ed in a re pre se ntativ e ex perime ntal animal mode l using adequate ve nom sample s as inflammatory induc er are ne eded. Previous re ports have show n that susce ptibility to Cro ta lu s du ris s u s te rrific u s ve nom varie s ac cording to the strain of mice use d, A/J and DBA/2 have be en show n to be the most susc eptible and BALB/c and C57Bl/6 the most re sis tant strains. 5 It is w ell know n that le thality and tox icity of snake venoms c an vary acc ording to the age , sex , nutritional state and geographic re gions w here the animal w ere capture d. 6 Thus, aiming to minimize the ex pe rimental bias BALB/c mice , a strain of mic e moderately susc eptible to the B. a tro x venom ac tion and a mix ture of venom obtain ed from se ve ral adult snake s from the same geographic re gion w ere used throughout all ex pe riments.
The pre se nt study show s: that the susce ptibility to B. a tro x ve nom varie s ac cording to the strain of mice and that cytokines are re le as ed in vivo during envenoming and in v itro upon stimulation w ith the venom. The inflammatory e ffec ts induce d by this venom are also desc ribed at mic rosp ic al le vel. As haemorrhage is one of most c onspicuous tissue le sions produce d by Bo thro p s sp. venom, the blocking action of a monoclonal antibody against a haemorrhagic fac tor pre se nt in B. a tro x venom w as te ste d.

Venoms
Bo thro p s a tro x, venom, w as provided by the Laboratório de Herpe tologia, Instituto Butantan, São Paulo, Brazil. The venom w as c ollec te d, pooled, lyophilize d and store d at -20°C. Samples w e re p re pare d in 0.15 M phosp hate buffere d saline (PBS) pH 7.2 at use.

Animals
Male mic e of the strains C57Bl/6 (H-2 b ), BALB/c (H-2 d ), C3H/He (H-2 k ), C3H/He Pa (H-2 k ), DBA/2 (H-2 d ), A/J (H-2 a ), A/Sn (H-2 a ) w eighing 18-22 g and age d 5-6 w eeks, w e re obtained from the Biotério de Camundongos Isogênic os, Univers idade de São Paulo, São Paulo, Brazil and use d throughout the ex p eriments. The animals w e re maintain ed and used under strict e thic al conditions ac cording to the animal w elfare international re c ommendation s (Commitee Membe rs, International Socie ty on Tox inology, 1991).

Antibodies
A monoclonal antib ody against a hae morrhagic factor re c ognizing a 60 kDa c omponent w as pre pared and immunochemic ally analysed by the authors (manusc ript in pre paration).

Lethality and LD 50 calculation
The lethal tox icity of B. a tro x venom w as assessed in isogenic mouse strains by intradermal injection of differe nt venom concentration s in 0.1 ml of 0.85% NaCl solution. Five animals w ere used for each ve nom dose. The LD 5 0 w as de te rmined by the Spearman-Karber method 7 and calc ulated ac cording to probit analysis. 8

Venom treatment
Groups of four mic e pe r group w ere inje cte d i.p. w ith 500 m l of saline c ontainin g 50 m g of ve nom.

Sera and cell suspensions
Blood w as collecte d from the re tro-orbital plex us at differe nt times after ve nom injec tion. Ce lls w ere obtain ed either from spleens or from the pe ritone al cavity of mic e pre viously injec te d w ith 50 m g of venom.

Cell culture supernatants
Sple en ce lls w ere obtained from venom or saline injecte d mice 3 h be fore harvesting. Ce lls w ere cultured in supplemented RPMI-1640 medium at 10 5 cells/w ell in 5% CO 2 at 37°C. Ce lls re ceive d in v itro either 250 ng of venom or 1 m g of LPS or 1 m g of Con A.
Adhere nt pe ritone al c ells w e re collecte d from the same donors of spleen ce lls. Ce lls w e re c ultured in RPMI-1640 medium supp le mented w ith 2-ME, L-glutamine , and 5% FCS, and left to adhe re onto 24 mic row ell plates for 24 h at a initial c oncentration of 10 6 cells/w e ll in 5% CO 2 at 37°C be fore in v itro addition of 250 ng of ve nom or 1 m g of LPS. Supernatants w ere c ollec te d at 48 h of culture.
IL-6, IFN-g and IL-10 concentrations in serum or spleen or adherent peritoneal cell cultures The pre se nce of c ytokine s IL-6, IL-10 and IFN-g in mouse sera w as dete rmined by indirect ELISA following the p roc edure s desc ribed elsew here . 9 Brie fly, ELISA plates w e re c oated w ith 100 m l (1 m g/ml) of monoclonal antib odie s anti-IL-6, anti IFN-g or anti-IL-10 in 0.1 M sodium carbonate coating buffer, pH 8.2, and inc ubate d for 6 h at room te mpe rature. The w ells w ere the n w ashe d w ith 0.1% PBS/Tw e en 20 and blocked w ith 100 m l of FCS 10% PBS for 2 h at room te mperature. After w ashing, duplicate mouse serum samples of 50 m l w e re added to each w e ll. Recombinant murine c ytokines w e re used to ge ne rate standard curve s. Afte r 18 h of incubation at 4°C the w ells w ere w as he d and incubated w ith 100 m l (2 m g/ml) of the biotinylat ed monoclonal antib odies anti-IL-6, anti-IFN-g or anti-IL-10 re spe ctive ly as sec ond antib odies for 45 min at room te mperature . After a final w ash, the re action w as de ve lope d by the addition of OPD to each w e ll. Optic al de nsitie s w ere me as ure d at 405 nm in a microplate re ade r. IL-6, IFN-g and IL-10 levels are ex pre sse d as ng/ml. The dete ction limits of the se assays w ere 0.78 ng/ml for IL-6 and IL-10 and 7.8 ng/ ml for IFN-g .

TNF-a concentration in serum or spleen or adherent peritoneal cell cultures
The pre senc e of TNF-a in mic e sera w as e valuated by the me thod of Ruff and Gifford. 10 Briefly, L929 cells maintained in RPMI 1640 medium suppleme nte d w ith 5% FCS w e re plated at 3 3 10 5 c ells/w ell on a 96-multiw e ll tissue culture plate and inc ubated at 37°C for 18 h in a 5% CO 2 atmosp here . Dup lic ate samples of 100 m l of mouse se ra c ollec te d from the saline or venom inje cte d animals w ere se rially diluted in RPMI 1640 me dium c ontainin g 1.0 m g/ml of ac tinomyc in D and added to the L-929 cell cultures. After 18 h of inc ubation, the supernatants w ere re moved and the re maining live c ells asse sse d by fix ing and staining w ith 1% crystal violet. Absorbance [A] w as measured in each w ell at 620 nm in a mic roplate re ader. Cytotoxicity w as calculate d by the formula: Cytotox icity = A c ontrol -A sample 3 100/A control Titres w ere c alc ulated as the re ciproc al of the dilution of the sample in w hich 50% of the c ells in the monolaye rs w ere lyse d. TNF-a activity is ex pre ssed as ng/ml, e stimate d from the ratio of a 50% c ytotox ic dose of the te st to that of the standard mouse re c ombinant TNF-a The de te c tion limit of this assay w as 3.429 pg/ml.

Nitrite concentration in serum, spleen or adherent peritoneal cell cultures
The nitrite levels in mice serum as an indication of NO production w ere dete rmined as de scribed previously. 11 Brie fly, 40 m l of each mouse se rum sample w ere inc ubate d in a 96-w ell, flat-bottom plate w ith 40 m l of the re duction solution (NADPH 1.25 mg/ml; FAD 10.4 mg/ml; KH 2 PO 4 0.125 M) containin g 0.5 U NO 2 re ductase for 2 h at 37°C; after incubation, 80 m l of Griess re age nt (0.1% naphtyle ne diamine hydrocloride, 1% sulphonylamide , 3% H 3 PO 4 ) w ere adde d to each w ell. The optical densities w ere me as ure d at 540 nm in a mic ro plate re ade r. NO 2 conc entrations w ere de te rmined using a standard c urve of NaNO 3 ranging from 1.25 to 270 mM and ex pre ssed as nmol/ 10 5 c ells. Nitrite le vels in the c ell c ulture sup ernatants w ere dete rmined by mix ing e qual volume of Grie ss re age nt and and proce sse d as above.

Determination of LD 50
Groups of five BALB/c, C3H/He, C3H/Pas, C57Bl/6, DBA/2, A/J and A/Sn mic e w ere inje cte d i.p. w ith differe nt doses of B. a tro x crude venom and the death survival ratio de te rmined after 48 h. The LD 50 value w as c alculated by probit analysis at 95% confidenc e. As show n in Table 1 (Fig. 1A). As indicate d by the arrow s the skeletal musc le c ells surrounde d by the infiltratin g le ukoc ytes are disrupte d. In contrast, fragments of tissues c ollec te d from the c orresponding skin site s injecte d w ith the same dose s of c rude venom pre -incubated w ith a monoclonal antibody against a 60 kDa purified haemorrhagic c ompone nt did not show pate nt inflammatory signals ex ce pt for the pre se nce of a disc re te ex udate be tw ee n the skeletal muscle ce lls (Fig. 1B).

In vivo inflammatory cytokines and nitrite release upon venom injection
Groups of BALB/c mic e w ere inje cte d i.p. w ith 50 m g of B. a tro x ve nom and samples of blood w e re collecte d after 2, 4, 12, 18 and 24 h. Cytokines and NO 2 w ere individually titrated in the sera. IL-6, IL-10 and IFN-g attaine d max imal re lease after 4 h ( Fig.  2A,B,C). follow ing venom injec tion, w hile a sec ond peak for IFN-g w as at 12 h TNF-a started to appear after 4 h attainin g a p eak at 6 h de clining the re after (Fig. 2D). NO 2 starts to appear at 2 h attainin g a peak at 4 h and de clining at 6 h (Fig. 2E).

Ex vivo inflammatory cytokines and nitrite release upon venom injection
The cytokines w hich w e re de te c te d in serum are mostly synthe size d by mac rophages. In order to be tter unders tand the mechanism of the c ell activation by venom splee n c ells or adhe re nt peritone al ce lls obtaine d from the same mice pre -stimulated in v ivo w ith B. a tro x ve nom w ere analysed in culture s.

Splenocytes
Groups of BALB/c mice w e re injec te d ip w ith venom 3 h before c ell harvesting. Venom, LPS or Con A w ere adde d to splenoc yte cultures and IL-6, IL-10, IFN-g , TNF-a and NO 2 w ere measured in these supe rnatant s afte r 48 h. Figure 3A show s that in contrast w ith LPS and Con A the addition of venom to ce lls obtained either from mice pre viously inje cte d w ith saline or venom does not induce IL-6 re lease . Re lease of IL-10, how e ve r, w as induce d by venom, LPS and Con A e ithe r in animals inje cte d w ith ve nom or w ith saline ( Fig. 3B). Re leas e of IFNg w as only obtained by Con A. (Fig. 3C). TNF-a and NO 2 w e re not dete cte d in these supernatants (data not show n). Figure 4 show s that re sident peritone al c ells e ithe r obtain ed from animals injec te d w ith ve nom or saline produce significant amounts of IL-6, IL-10 and IFN-g after ve nom or LPS challenge. TNF-a and NO 2 w ere not de te c te d in the se supe rnatant s (data not show n).

Discussion
These data have show n that subc utaneous injection of B. a tro x c rude venom in mice induce d a set of gross inflammatory signals: plasma ex udation, leukoc yte migration, vascular w all damage re sulting in haemorrhage and ske le tal musc le ce ll disruption. Some or all of these effe cts caused by B. a tro x ve nom have be en alre ady desc ribed either in human victims of snake bite or in ex pe rimentally inje cte d animals. 12 ,1 3 The aim of this w ork has be en to provide an ex p erimental mode l to study in v ivo and in vitro the p roduction and re le ase of mediators (p ro-inflammatory substanc es and cytokines) know n to be involve d in leukoc yte migration, vascular w all damage and in ce ll activity or de ath 1 4 and to analyse the le sions induce d by B. a tro x crude ve nom. Such an ex p erimental mode l should involve animals gene tically homoge- ne ous and moderately susce ptible to the ve nom tox ic effects, as w ell as ve nom samples containin g all the pre sumptive tox ic c ompone nts. This w as achie ved by pre -dete rmining the LD 5 0 of the venom in se veral isogenic strains of mice . Among the analysed strains A/Sn, A/J and C57Bl/6 w ere significantly more susc eptible to the venom le thal e ffec ts than C3H/He , C3H/ He Pas and DBA/2 re sistant strains. BALB/c mic e w ere found to be a moderately susc eptible strain. The gene tic influenc e on an ex pe rimental animal susce ptibility to the le thal ve nom ac tion is not novel, sinc e gene tic background also influe nce s susceptibility of mic e to the le thal activity pre sent in C. d. te rrific u s venom. 5   the le thality appears to be de pendent on the inflammatory factors such as thrombin-like, haemorrhagins, p hospholipas es and prote ase s enzymes. Plasma ex udation and oe de ma at the skin sites of venom injection c an be at least partially ex plaine d by loc al re le as e of the w ell-know n me diators of the early e ve nts of the acute inflammatory events such as bradykinin 1 5 and anaphylatox ins. 1 6 In fac t Bo thro ps venom c ontains proteolytic e nzyme s c apable of promoting in vitro cle avage of plasma kallikreinogen and/or the complement c omp one nts C3 and C5, re leasing bradykinin 15 and C3a and C5a, 1 7 re spe ctively. The loc al re le as e of these vasoactive pe ptide s although pre sumptive has not ye t be en ex pe rimentally p roven. Although C3a and C5a are pote nt che moattractants for PMN le ukoc ytes and monocyte s, 1 8 acc umulation of mac rophage-like cells, lymphoc yte scattering among the inflamed tissues, the ce ll w all damage and ske letal muscle c ells disruption cannot be ex plaine d by the loc al formation of those me diators. Published re sults obtained in other inflam- matory models have indic ated that pro-inflammatory substanc es and some cytokines are deep ly involved in the ac tivation of the e ndothelial ce lls and le ukocytes leading the m to ex pre ss on their ce ll surfac e adhesion molecules as E-selec tin and P-selec tin and the corresponding spec ific ligands PSGL-1 and Sialyl Lew is and othe rs. 1 9 Be sides, some cytokine s such as IL-8 are che moattractants for basophils and T lymphocyte s 2 0 -22 and p ote nt activators of these ce lls. 2 3 The c ytokine profile s observed in vivo upon B. a tro x venom injec tion show s that production of IL-6, IL-10 and IFN-g attaine d max imal values at 4 h follow ing ve nom inje ction. While the first tw o cytokines dec ay along the follow ing hours be coming almost unde ctable, IFN-g se rum le vels attain a sec ond higher pe ak 12 h later. TNF-a in c ontrast, appears at a later time in se rum re aching high amounts 18 h after the venom injec tion (Fig. 2). NO 2 se rum le vels, on the othe r hand, starts to appear e ve n be fore 4 h. follow ing venom injec tion inc re as ing slow ly but cons is te ntly re aching max imal values after 24 h. These re sults agree w ith pre vious re p orts show ing that IL-6 is re le as ed in Sw is s mic e injec te d w ith B. a s pe r venom 1 3 and human vic tims of snake bite. 12 Although the first authors w e re not able to de te c t TNF-a in mic e injecte d w ith B. a s pe r venom, others pre se nte d indire ct evide nce indic ating that this c ytokine can be produce d along e nve nomation by snake venom. 2 4 These authors p rovided e vide nce that metallo-prote as es from B. ja ra ra c a and Ech is py ra m idiu m le a ke y ve noms c an cle ave pro-TNF-a into its mature form in vitro . Besides, the y have show n that TNF-a antib ody signific antly re duc ed the re sultant ne c rotic lesions in mice inje cte d w ith E. p. le a ke y venom. IL-10 is c apable of mediating suppre ssion on activate d macrophage s and monocytes leading to striking dow n-re gulation of nume rous inflammatory monokines as TNF-a , IL-1,IL-6, IL-8, GM-CSF and G-CSF. 2 5,2 6 IL-10 is also capable of effective ly prote c ting mic e from endotox in-induce d shock, a le thal inflammatory re action mediated by monokine s as TNF-a and IL-10. 2 7 A similar suppre ssor effe ct of IL-10 w as also found on NO production by activated mac rophages. 2 8 Our data show s that the IL-10 peak at 4 h after venom inje ction coinc ide s w ith TNF-a and NO 2 low serum le vels are compatible w ith the se observations. Some c oinc ide nce s or disc re pancie s w ere also observe d w hen cytokine production w as analysed in v itro using eithe r sple en c ells or adhere nt pe ritoneal ce lls obtained from BALB/c mic e injecte d and challenge d ex v ivo w ith venom: (a) IL-6 w as not produce d by spleen ce lls; (b) IL-10 w as produce d both by sple en or peritone al ce lls; (c ) IFN-g w as produce d by adhere nt pe ritone al cells but not by sple en ce lls.As the in v ivo ve nom stimulate d c ells re sponded in v itro to the classic al cytokine induce rs, these discre panc ie s cannot be impute d to ge ne ral dere gulation of the machinery used to produce cytokines. Production of IL-10 by spleen ce lls and pe ritoneal adhe re nt cells agre es w ith the w ell-established observations showing that this c ytokine is produc ed by a large array of ce lls in w hich T and B lymphocyte s and macrophage s are included. 29 Depe nding on the re lap sing time, the production of the various cytokines w ill be influenc ed by the one s that are first re lease d into the surrounding millieu, inte rfering there fore w ith maturation or se cre tion of the othe rs or w ith the ex pre ssion of their correspondent re ce ptors.
The discre panc ie s dete cte d in the c ytokine profiles produce d in v ivo and ex vivo can be ex plaine d by the inte rplay involving the factors gove rning the production of the conc erne d proteins, the lack of some essential factors for the synthesis or ex pre ssion of a particular c ytokine or by the low sens itivity of the assay methods of dete cting smaller amounts of cytokines. As a monoclonal antib ody sp ecific for a 60 kDa B. a tro x haemorrhagic factor w as able to block le ukocyte infiltration and haemorrhagic ac tivitie s of the ve nom. Such findings sugge st that the haemorrhagic compone nts c ould be the conspic uous inflammatory induc ers pre sent in the venom. The monoclonal certainly re cognizes and blocks the action of the haemorrhagic factor pre sumably by combining w ith a functionally important e pitope.
On the basis of our data, w e can hypothesize that B. a tro x venom can trigger various host inflammatory me chanisms by using the same or some compone nts acting in conc ert: (a) the e arly plasma ex udation can be mediate d by loc al re lease of bradykinin and histamin; (b) polymorphonuclear c ells adhe re nc e and migration may be trigge re d by C3a, C5a and or IL-8; (c) the endothelial cells be longing to the blood vessels crossing the tissue sites injec te d w ith the venom can be ac tivated and ex pre ss P-selec tin and E-se le ctin unde r the p arac rinal action of his tamine, thrombin-like factor and TNF-a ; (d) basophils and lymphoc yte ac cumulation can be mediate d by  and (e ) ske letal muscle c ell disruption c an be induce d either direc tly by enzymes as prote ase s, phospholipase A 2 , hae morrhagins or e ve n othe r cytolitic compone nts pre se nt in the venom, or indire ctly by me diators re leased from invading leukoc ytes.