Tosirnn, Vol . 32, No . 12_pp_ . 1689-1695, 1994

~)(-0lUl(94)00097-2

Rioted in ßreat ~Hri4in. All ria6 served0041-0101 /94 57.00 + .00

EFFECT OF VARIOUS VIPERIDAE AND CROTALIDAE SNAKEVENOMS ON ENDOTHELIAL CELLS IN VITRO

GADI BORKOWI , BRUNO LoMONTE,2 JOSÉ MARIA GuTIÉRREZ2 and MICHAEL OVADIAI*(Department ofZoology, George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel Aviv 69978, Israel;

and zInatituto Clodomiro Pitxdo, Facultad de Microbiologia Universidad de Costa Rica, Costa Rica

(Received l5 April 1994; accepted 30 June 1994)

G. Boxxow, B. LoMONTE, J . M. GuTt~RREZ and M. OVADIA . Effect of variousViperidae and Crotalidae snake venoms on endothelial cells in vitro . Toxicon32, 1689-1695, 1994.-The effect of various crotalid and viperid venoms at 10,50 and 100 ~g/ml wasexamined on bovine and marine endothelial cells in vitro .The venoms caused the cells to lose their processes, leading to the appearanceof spaces which were gradually enlarged between clusters of cells . The cellsbecame round and finally detached from the substrate . This effect was morepronounced on bovine normal cells than on marine transformed cells . Most ofthe venoms did not affect the viability of the cells even after 24 hr of incubation,as determined by the trypan blue dye exclusion procedure. Moreover, after thecells were washed from the venoms and transferred into fresh medium, theyregained their original morphology after spreading on the substrate and theythen proliferated normally . This reversible effect shows that most of the crotalidand viperid venoms examined were not directly cytotoxic to the endothelial cellsat the concentrations tested .

IT Is somewhat unclear as to whether there is only one or there are several mechanismsby which venoms induce hemorrhage (for review see B~ARNASON and Fox, 19889). Bloodcapillaries appear to be the main target of the hemorrhagic toxins, with a resultingalteration of the vessel permeability and extravasation. OHSAKA et al. (1973) and TSUCHIYAet al. (1974) attributed the hemorrhagc effect of the venom of Trimereserus flavoviridis toenzymatic disruption of the basement membrane of the capillary wall, followed by redblood cells oozing out through inter-endothelial gaps . OWNBY et al. (1978) investigated thehemorrhagic toxins isolated from the venom of Crotales atrox and concluded that theprimary effect of those hemorrhagc toxins was on the endothelial cells themselves and thatthe basement membrane around the capillaries was disrupted afterwards. Similar obser-vations were reported for the hemorrhaggc toxins isolated from venoms of Crotaleshorridus (OWNBY and GEREN, 1987), Agkistrodon bilineates (OWNBY et al., 1990) andCerastes cerastes (RAHMY et al., 1992).

Author to whom correspondence should be addressed.

1689

1690

Short Communications

Our in vivo studies with Bothrops riper crude venom and with the main hemorrhaginisolated from this venom (BaHI, Boi:xow et al., 1993) indicated that the basementmembrane was affected early in the course of envenomation, and that the initial effect onthe endothelial cells included detachment from their surrounding basement membrane andpresence of blebs and cytoplasmic projections (Mota;etxn et al., 1992, 1994). As conse-quence, endothelial cells became thinner until spaces or breaks appeared between the cellsthrough which erythrocytes escapéd, leading to the conclusion that endothelial cellalteration was secondary to degradation of basement membrane wmponents.The controversy over the mechanism of action of snake venom on the capillaries may

be due to the fact that previous studies were carried out in the complex in vivo system and,therefore, suffered from at least two disadvantages: (1) the inability to distinguish betweenenzymatic and cytotoxic activities of the venom because in vivo experiments do not allowfollow-up of the sequence of events continuously under the microscope ; and (2) theinability to isolate the endothelial cells from all surrounding tissues.The significance of this work is, therefore, bound up with the focus on the activity of

many viperid and crotalid venoms on isolated endothelial cells cultured in vitro . Althougha venom is a mixture of various compounds, it is important to examine the overall activityof the whole mixture on the isolated cells before the investigation of each purifiedcomponent .Most viperid venoms were obtained from snakes that were kept and milked at the

Serpentarium of the Canadian Center for Ecological Zoology, Tel Aviv University.Venoms of Bothrops species were prepared at the Instituto Clodomiro Picado (Costa Rica).After the venoms were pooled, the protein content was estimated by measuring theabsorbance at 280 nm in a Gilford UV spectrophotometer; for a 1% solution A,~ = 14.The venoms were lyophilized and stored at -20°C until use. All other Echis and crotalidvenoms were purchased in lyophilized form from Latoxan Laboratories (Rosans, France).Two different types of mammalian endothelial cells were examined : bovine and marine.

Primary bovine aortic endothelial cells were isolated as described by Sexw~+xTZ (1978).Transformed marine endothelial cells (tEnd) of capillary origin were kindly provided byDrs A. M~iv'rovnrn and E. DEJANA (Instituto Mario Negri, Milano, Italy) . For experimen-tation, the cells were grown on uncoated plastic plates with 96 wells of 0.2 ml incubationvolume (Cel-Cult, Sterilin Limited, Feltham, U.K.) . The effect of the various venoms onthe cells was monitored by adding 10 ~g of phosphate buffered saline 0.01 M, pH 7 (PBS),containing 2~g of each venom (final concentration 10 pglml) into wells containing 10'bovine endothelial cells or marine endothelial cells . After various time intervals 20~e1aliquots of each well were diluted in 80 pl fresh medium and the number of floating cellswas determined by a cytometer. Twenty hours after the addition of venom or when9100% of the cells in each well were detached from the bottom, the medium with thefloating cells was removed, centrifuged at 100 x g and the cells were washed twice withfresh medium before they were recultivated with fresh medium in new wells. The totalnumber of cells per well was calculated after trypsinization of the cells in three untreatedwells and by counting as above.

Table 1 shows that most examined viperid and crotalid venoms have a demonstrableeffect on the endothelial cells . Some venoms induce detachment of the cells from thesubstrate within 2-4 hr, whereas other venoms need longer periods to reach this result . Thisactivity of the venoms is summarized in several steps, as shown in Fig. 1 : there is an initialloss of cell processes followed by aggregation of the cells, which form spaces. The size ofthe spaces gradually increases as the incubation period is prolonged; the cells then become

Short Communications

1691

TABLE I . DETACFU~NT OF BOVINE AND MURINE ENDO'l'F~LIAL CELLS 7REATFF~ WITFI VWERID AND CROTALID VENOMS

Two micrograms ofeach venom waa added to endothelial cells cultured in 0.2 ml medium. Theabove valuesarc the clean percentage of detached cells (triplicates).

round and finally detach from the substrate and float in the medium. This effect was morepronounced in bovine normal cells than in marine transformed cells . Similar distinctsusceptibility to cytolytic effects of snake venoms has also been shown for marine B-cellsand Tells at distinct stages of cellular development (LEwls er al., 1990). In all venomstested, when detached floating cells were washed with fresh medium to remove the venomas described above, the cells recuperated after they were transferred into new wellscontaining fresh medium, and spread on the substrate. They then regained their originalmorphology and proliferated, indicating that the effect of the crotalid and viperid venomson the endothelial cells is reversible . The influence of the venoms on the viability of theendothelial cells was also examined by the trypan blue exclusion procedure in a separateexperiment in which the cells were incubated with 10, 50 or 100 ~tg/ml ofeach venom. After24 hr of incubation with the venom, 5 ~l of 8% Trypan blue solution (w/v) was added toeach well and the number ofstained cells counted. Table 2shows that most of the venoms,even at the high concentration of 100 ~g/ml, did not induce staining of the cells by the

I hr 2hr

Bovine cellsDetached cells

4 hr 6hr 20 hr l hr

Marine alla% Detached cells2 hr 4 hr 6 hr 20 hr

ViperidaeV.lebetina 80 !00 0 0 0 IO 20V. ammodytes 60 80 l00 0 0 5 10 !00V, palaestinae 0 0 0 0 15 0 0 0 0 0V. bornnmelleri 0 0 0 0 5 0 0 l0 20 50V. raddel 30 60 90 100 0 0 0 0 0V. russelli 10 30 90 100 0 0 5 20 50V. eristocophis 0 0 10 ( 5 25 0 0 0 5 30Pseudoccrastes fttldi 0 0 0 0 10 0 0 0 0 0E. coloratus 5 35 90 100 0 0 IO 20 SOE.leucogaster 30 60 100 20 40 60 l00E. pyramidewn 30 70 100 0 0 20 50 100E. sochurekl 20 60 100 30 60 100E. multisquamatus 80 100 0 0 30 60 100E. ocelatus 75 100 0 10 40 80 100Bitis arietans 0 0 10 15 25 0 0 0 0 0Bitis gabortica 25 50 95 100 not examined.lspis eerastes 60 95 100 not examined

CrotalidaeBothrops atrox 50 80 100 0 0 0 0 0Botlvops asper 20 SO 95 100 0 10 30 50 80Bothrops picadaei 2l1 35 45 60 90 0 0 0 0 0Bothrops reanmijer 0 10 15 20 25 0 0 0 0 0Botlvops lateralts 0 5 20 40 80 0 0 0 0 0Bothrops sc6legeli 0 5 25 45 70 0 0 0 0 5Bothrops opkyromegas 0 10 20 40 70 0 0 0 0 5Borktrops godrrlarri 0 5 10 20 30 0 0 0 0 0Bothrops nasutus 0 0 5 10 25 0 0 0 0 0Botlvops nigrooiridis 0 0 10 15 25 0 0 0 0 0Crotales atrox 0 0 0 0 10 0 0 0 0 0Crotalw scutulatus 0 0 0 5 15 0 0 0 0 0Crotales diaisus drairus 0 0 0 0 10 0 0 0 0 0Lachesis muta 0 0 0 10 20 0 0 0 0 0Agkistrodon piacioorus SO 100 not examined

1692

Short Communications

dye, indicating that the cells remained viable . Moreover, the viability of the cells wasfurther demonstrated by removing the venom and recultivating the washed cells in freshmedium as described above (Table 2) . Similar findings of a reversible effect were observed

Fta. 1 . Ermec,-r ofL~~exn vrreero vexo~sv ox eovnvE sxno~t.w, ceu.s.(A) Endothelial cells (103) were incubated for 4 hr at 37°C with PBS only: the cells spread well onthe substrate; (B) cell aggregation and spaces caused by the venom of Bothropa asptr (10 pg/ml)after 1 hr, (C) activity of Vipers kbetirta venom (10 pg/ml) after 2 hr. the cells became round and

detached from the substrate.

Short Communications

1693

TAHIE 2. VIASH .rrroFTHE ENDDTIIELUI. CEr.LS AFIER 24 HROFINCUBATION WR'H VAAI(H1SCONCENTRATIONSOF EACHOF THE CROTALID AND VI1FR1D VENOItS

Two, ten or twenty micrograms of each venom was added to endothelial cells cultured in 0.2 ml medium(quadruplicates). After 24 hr 5 itl of 8% Trypan blue solution (w/v) was added to two wells ofeach venom. Thepercentage of the stained cells is given in the table. The medium from the rrmain+ng two wells of each venomwas pooled and centrifuged in 5000 x d for 10 min. The supernatant was washed and fresh medium was addedto the cells, which were recultivated . Their proliferation after 24 hr was examined.TB, Peroentage of cells stained by Trypan blue.V, Viability of the cells as determined by their proliferation.+, Cells proliferated normally ; -, No cells proliferated.

in our previous work investigating the influence of various venoms on melanoma B16F10cells (Gwwn~t-MATYAS and OvwDlw et al., 1987). These results also corroborate the previousultrastructural studies carried out in vivo with Bothrops riper crude venom and with themain hemorrhagic toxin BaHl, purified from this venomby BORKOW et al. (1993) . In thosestudies Moxlmew et al. (1992, 1994) indicated that the basement membrane is affected earlyin the course of envenomation. Moreover, in a subsequent investigation LwMONTE et al.(1994) showed that BaHI was not cytotoxic to the transformed murine endothelial cells,even at a toxin concentration as high as 65 ~g/ml.

It is well known that morphological alterations of the cells may occur as a consequenceof the degradation of the basement membrane ('IhYCCV~+soN et al., 1987; At.BER~ts et al.,1989). A current study on normal bovine endothelial cells indicates that the initial effectof the purified hemorrhagic toxin BaH1 is not directed toward the cells but rather towardextracellular components (BORKOW et al ., 1994). The present findings are also supportedby previous studies that demonstrated the degradation of isolated extracellular matrixproteins by hemonhagic toxins isolated from the venom of Crotales atrox (BARAMOVwet al., 1989; Sxwxxox et al., 1989). The order of proteolytic activity of these snake toxinsi'OX J3113-H

10TB V

Bovine cells(PB/~)

50TB V

100TH V

10TB V

Murine cells(h8/m1)

50TB V

100TH V

ViperidaeV. klxtina 0 + 100 - 100 - 5 + 100 - 100 -V. palatstirrae 0 + 0 - 40 - 0 + 0 + 0 +V. bornmuelkrl 0 + 0 + 30 + 0 + 0 + 5 +V. raddrt 0 + 0 + 40 + 0 + 0 + 5 +E. coloratus 0 + 0 + 20 + 0 + 0 + 5 +E. kucogaster 0 + 0 + 5 + 0 + 0 + 0 +E. pyramideurn 0 + 0 + 0 + 0 + 0 + 0 +E. sochweki 0 + 0 + 0 + 0 + 0 + 0 +E. rtatJtisquamatus 0 + 0 + 30 + 0 + 0 + 5 +BJtis arietatrs 15 + 100 - 100 - 5 + 100 - 100 -Bttis gabontca 20 + I00 - 100 - 20 + 100 - 100 -

CrotalidaeBothrops aspcr 0 + 10 + 100 - 0 + 0 + 0 +Bothrops picadet 0 + 0 + 0 + 0 + 0 + 0 +Bothrops rwrntreije 0 + 5 + 10 + 0 + 0 + 0 +Bothrops lateralis 0 + 10 + 60 + 0 + 0 + 0 +Bothrops schlegelt 0 + 50 + 90 - 10 + 90 - 100 -Bothrops ophyromegas 20 + 40 + 50 + 0 + 0 + 5 +Bothrops godnwni 0 + 0 + 10 + 0 + 0 + 0 +Bothrops nasutus 5 + 30 + 10 + 0 + 0 + 5 +Bothrops ntgroviridis 0 + IS + 30 + 0 + 0 + 5 +Crotahts durisus durJsus 0 + 25 + 70 + 5 + 60 + 90 +Lachesis nruta 0 + 0 + 0 + 0 + 0 + 0 +

1694

Short Communications

on the basement membrane components was the same as their hemorrhagic potenciesin vivo (13JARNASON et al., 1988).However, it could not be excluded that some venoms are also cytotoxic to endothelial

cells, in addition to their disruption of the basement membrane, because endothelial celllysis by venoms was found by McKwY et al. (1970), OwxsY and GSRHN (1987), OWNBYet al. (1978, 1990), and RAHMY et al. (1992) . The ultrastructural studies on capillary vesselalterations after injection of hemorrhagic toxins have demonstrated that endothelial cellsare drastically affxted. However, this effect may be secondary to the detachment of thecells from the substrate, or it may be caused indirectly in the complex i» vivo system.Furthermore, in some cases more than vne endothelial all degenerative pattern mayoccur(MORBIRA et x1.,1992). If there are several populations of endothelial cells in the capillaries,it could be proposed that both mxhanisms ofinitial disruption ofthe basement membraneor initial cytotoxicity to the cells may work in vivo in parallel. A partial demonstration ofthis can be found in Table 2 of this study, which shows that some venoms lysed aproportion of the cell population, whereas the other portion ofcells was able to recuperateafter the venoms were washed .

In vivo experiments are indispensable in the study of venom activity in the real system.The process described by this work may not accurately reflect the events occurring in vivo .For example, the time course of events monitored (hr) does not match the time course ofhemorrhagic activity of venoms (min). However, this in vitro investigation enabled manymore tests to be carried out than is possible with the whole animal study, and addedanother dimension . It enabled us to follow up the sequence of events continuously underthe microscope and to differentiate between the enzymatic activity of each venom and itspossible cytotoxicity. It also made possible additional experiments on the endothelial cellsduring the development ofthe activity ofeach venomthrough changing the medium, whichallowed us to show that most of the viperid and crotalid venoms examined are notcytotoxic to the endothelial cells in the examined concentrations, even after longincubation (up to 24 hr). Instead, most of the examined venoms caused the detachmentof viable cells from the substratum, which was shown to be a reversible effect .

Acknowledgtnrtnt-This study was supported by AID Grant No . DHR-5544-6-00-1064-00.

REFERENCES

At.emus, B., BIeAY, D., Lswu, J., Rte, M., Rosea~ts, K. and We'tsota, J. D. (1989) Alolecular Biology ojtlre Cell,pp. 802-524. New York: Garland.

B~ssüve, E., SFUivivox, J. D., Buxxeaox, J. H. and Fox, J. W. (1989) Degradation of extracellular matrixproteins by hemorrhaic metalloproteinases. .lrch. Bloclrcm. Blophya. 275, 63-71.

B»xesotv, J. B. and Fox,J. W. (1988-89) Hemorrhagictoxins ofsnake venoms . !. Toxtc.-Toxin Reo. 7, 121-209 .BtNeruaotv, J. B., Htiai.rox, D. and Fox, J. W. (1988) Studies on the mechanism of the hemorrhage production

by 8ve proteolytic hemorrhagc toxins from Crotales atrox venom. Blot. Chem. Aoppe-Seykr 369, 121-129 .Ho~ow,G., Gulvéatu~z, J. M. and Ov~ou, M. (1993) Isolation and characteri7stion of synergistic hemorrhaginsfrom the venom of the snake Botlunpa riper. Toxlcott 31, 1137-1150.

HORICOW, G., Gtrrnéanez,J. M. and oveuu, M. (1994) Activity ofBaH1, the main hemorrhagc toxin ofBothropasaper snake venom, on bovine endothelial alla and extraallular matrix proteins. (Submitted) .

Cté-M~Trea, A. and Ovenu, M. (1987) Cytotoxic activity of various snake venoms of melanoma, B16F10and chondraarooma. L(fe Scl. 40, 1601-1607.

Lsvw, G. M., MtrrroN, S. A. and SLAt:1C, J. A. (1990) HALB/cAn B alla andT alla have distinct susceptibilitiesto cytotoxic effects of snake venom. Toxicon 2B, 351-358.

Loxorrre, H., Gtrr~aamz, J. M., Hoaxow,G., twenu, M., T~texowaet, A. and H~t~sort, L. (1994) Activity ofhemorrhagc metalloproteinase BaHI and myotoxin II from Bothropa aaptr snake venom on capillaryendothelial cells br nltro. Toxltb» 32, 505-510.

Short Communications

1695

McI{w, D. G., Moteoz, C., De Vx>ps, A., Csevosar, I. and CRUSe, V. (1970) The action of hemorrhagin andphoapholipese dtrived from Vipers pakstinat venom on the microcirculation. Lab. ImrGSt Z2, 387-399.

Moa~u, L., ßurt~ue~, J. M., Howcow,G.andOv~uu, M. (1992) Ultrastructutal alterations in mouse capillaryblood vessels atkr experimental injection of venom from the snake Bothrops aspen (tercipelo) . Exp. moi. Path.S7, 124-133.

Moastae, L., Hoaxow,G., Ovenu, M. and Gvt~esez,M. J. (1994) Pathological changes induced by BaHI . Ahemorrhagc proteinese isolated from Bothrops aspen (Terciopelo) snake venom, on mouse capillary bloodvessels . Toxicon 32, 977J987.

Otisecw, A., Just, M. andHw~xx, E. (1973) Action of snake venom hemorrhagc principles on isolatedglomerular basement membrane, Biochirn . biophys . Acts 323, 415-428.

Owivsv, C. L. and Ga~v,C. R. (1987) Pathogenesis of hemorrhage induced by hemorrhagc proteinase IV fromtimber rattlesnake (Crotahv horridus horridus) venom. Toxicon 23, 517-526.

OwivaY, C. L., Brwtuv~ox, J. and Tu,A. T. (1978) Hemorrhagic toxins from rattlesnake (Crotahv atrox) venom.Am. J. Path . 93, 201-210.

Owwev, C. L., Ntawt, T., hw,K. and Suatx~u, H. (1990) Pathogenesis of hemorrhage induced by bilitoxin,a hemorrhagc toxin isolated form the venom oftye common cantil (Agkistrodon blümatus bilfneatus) . Toxfcon28, 83746.

R~cY,T., Tu,A. T., Ec.-B~xxews~r, M. A., F~.-Ase~a, M. F. and Hex, F. M. (1992) Cytopathologic effectof C~rastes cerastes (Egyptian sand viper) venom and isolated hemorrhagic toxin on liver and kidney: anelectron microscopic study . J. Nat. Toxins 1, 45-58.

Scxw~erz, S.M. (1978) Selection and characterization of bovine aortic endothelial cells . In Yitro 14, 966-970.St~xxox, J. D., B~wsove, E. N., Btexiv~oN, J. B. and Fox, J.W. (1989) Amino acid sequence of a Crotahis

arrox venom metalloproteinase which cleaves type IV collagen and gelatin . J. bioi. Chem . 264, 11,575-11,583.TRYOavesox, K., HoYxne, M, and Seto, T. (1987) Proteolytic degradation of extracellular matrix in tumor

invasion . Biochim. Btophys. Acts 907, 191-217.Tauctmrw, M., Ottsmo, C., Ot~+atn, M., Owe, A., Suzuct, K. and Fut~meo, Y. (1974) Cinematographic and

electron microscopic analyses of the hemorrhage induced by the main hemorrhagc principle, HR-1, isolatedfrom the venom of Tr6neresurus,Jlauovbidts . In : Plattkts, Thrombosis, and Inhibitors, pp . 43946 (Dm~t,O. Swaexoro,T. andY~x~~ "er, H., Eds). Stuttgart Schafteuer.