ilable at ScienceDirect

Toxicon 56 (2010) 1007–1017

Contents lists ava

Toxicon

journal homepage: www.elsevier .com/locate/ toxicon

Acute physiopathological effects of honeybee (Apis mellifera) envenomingby subcutaneous route in a mouse model

Mónica Prado a,b, Gabriela Solano-Trejos b, Bruno Lomonte a,*

a Instituto Clodomiro Picado, Facultad de Microbiología, Universidad de Costa Rica, San José, Costa RicabDepartamento de Parasitología, Facultad de Microbiología, Universidad de Costa Rica, San José, Costa Rica

a r t i c l e i n f o

Article history:Received 19 June 2010Received in revised form 8 July 2010Accepted 8 July 2010Available online 16 July 2010

Keywords:Apis melliferaHoneybeeVenomToxicityMouseAfricanized

* Corresponding author. Tel.: þ506 2229 0344.E-mail address: bruno.lomonte@ucr.ac.cr (B. Lom

0041-0101/$ – see front matter � 2010 Elsevier Ltddoi:10.1016/j.toxicon.2010.07.005

a b s t r a c t

Bee stings are a health concern in the Americas, where fatal envenomings due to massiveattacks by Africanized honeybees have been documented in the last decades. Most studieson the toxic effects of honeybee venom in experimental animals have been performedusing the intravenous or intraperitoneal injection routes. The aim of this study was todevelop a mouse model that would better resemble a massive honeybee attack by usingthe subcutaneous (s.c.) route to induce a severe, sublethal systemic envenoming. An arrayof acute venom effects were characterized, including biochemical, hematological, histo-logical, and inflammatory alterations, after the s.c. injection of 0.5 median lethal dose ofvenom. Rapid increases in serum alanine (ALT) and aspartate (AST) transaminases, creat-inine, urea nitrogen, uric acid, sodium and chloride electrolytes, and creatine kinase (CK)were recorded, indicating damage to liver, kidneys, and skeletal muscle. Also, coagulationdisturbances (fibrinogen decrease, and moderate delay in prothrombin and partialthromboplastin times) were demonstrated. Circulating platelet and leukocyte numbersremained unaltered, but a hemoconcentration effect (hematocrit and hemoglobinincrease) was observed. This effect might be related to the marked edema induced by thevenom. In addition, this inflammatory response included a systemic increase in cytokines(IL-1b, IL-6, TNF-a), together with an elevation of serum malondialdehyde and nitric oxide.The myotoxic effects of venom, melittin, and phospholipase A2 were demonstrated afterinjection by s.c. route. No synergistic myotoxicity between melittin and PLA2 was observed.Moreover, these two components, when injected at equivalent concentrations to thosepresent in venom, induced a lower increase in serum CK than venom, suggesting that othercomponents also contribute to its strong systemic toxicity towards skeletal muscle. Themodel here presented may be useful in preclinical studies to assess therapeutic anti-venoms developed to cope with the problem of massive bee attacks.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

The accidental escape of African honeybees (Apis melli-fera scutellata) imported into Brazil in 1956, followed bytheir interbreeding with European species (A. melliferamellifera and A. mellifera ligustica) gave rise to a new hybrid,commonly referred to as the "Africanized" honeybee

onte).

. All rights reserved.

(Spivak, 1992; de Medeiros, 2003). After more than fivedecades of migration and proliferation, this hybrid iscurrently widespread throughout tropical and subtropicalareas of the American continent, spanning from Argentinaand Brazil to the southern regions of the United States ofAmerica (Rinderer et al., 1993).

Bee stings may cause medical emergencies by twodistinct types of mechanisms: type I hypersensitivity(allergic) reactions, and envenomings. In the first, an indi-vidual becomes sensitized to bee venom components by

M. Prado et al. / Toxicon 56 (2010) 1007–10171008

producing IgE antibodies to them. If re-exposed to venom,an acute reaction may followwithin minutes, whereby IgE-sensitized mast cells degranulate and release their potentvasoactive mediators. Such reactions may range froma mild, local inflammation, to a life-threatening systemicanaphylaxis involving dyspnoea, asphyxia, hypotension,and cardiovascular shock (Harvey et al., 1984; Johanssonet al., 1991; Riches et al., 2002). Importantly, an allergicreaction may be triggered in a susceptible individual afterexposure to minute amounts of venom, even after a singlebee sting. This is in sharp contrast with the case ofenvenomings due to massive bee attacks, which mayinvolve hundreds or even thousands of stings, leading tosystemic toxicity.

Massive bee attacks have been related to the emergenceand successful expansion of "Africanized" bees in theAmericas, known to display a more vigorous and persistentcolony defensive behavior than their parent species (Winston,1994; Rinderer et al., 1993). Africanized honeybees producelower amounts of venom than their European counterparts(Schumacher et al., 1990), although the general biochemicalcomposition does not differ (Schumacher et al., 1990; Nelsonet al., 1990). Nevertheless, due to the large number of stingsthat may occur during an attack by Africanized honeybees,a systemic envenoming may ensue, with a potentially fataloutcome.

Bee stings have become a public health concern in theAmerican continent, where severe envenomings in humansand animals have been documented (Taylor, 1986; Françaet al., 1994; Schumacher and Egen, 1995; Kolecki, 1999;Oliveira et al., 2000, 2007; de Medeiros, 2003; Betten et al.,2006; Mitchell, 2006), in spite of the scarce and fragmen-tary epidemiological information available. In the state of SãoPaulo, Brazil, 2462 accidents involving bee stings, includingseven fatalities, were reported from 1993 to 1998(de Medeiros and França, 2003). In Mexico, over 17,000 beesting accidents were recorded during 1998, and in Argentinaseven fatalities occurred between 1997 and 1998 (de Roodtet al., 2005). In the U.S.A., 533 deaths were attributed toHymenoptera (hornets, wasps, and bees) during the decade1991–2001 (Langley, 2005). In Costa Rica, where arrival ofAfricanized honeybees was first detected in 1983 (Spivak,1991, 1992), epidemiological records of deaths attributed toHymenoptera (thevastmajoritycorresponding tobees) stingsduring 1985–2006 revealed 52 fatalities (Prado et al., 2009).

The major toxic effects of honeybee venom in humansystemic envenomings have been described in a number ofclinical case reports and in someautopsystudies (França et al.,1994; Ariue, 1994; Díaz-Sánchez et al., 1998; Kolecki, 1999;Oliveira et al., 2000; Bresolin et al., 2002; Daher et al., 2003,2009; Gabriel et al., 2004; Mitchell, 2006; Betten et al.,2006; Huertas-Franco and Bucknor-Masís, 2008), but arestill only partially characterized. Studies in experimentalanimals have complemented this information by focusing onparticular effects of this venom, or some of its components(Habermann, 1972; Couch and Benton, 1972; Vick andShipman, 1972; Vick et al., 1972; Ishay et al., 1975; Kaplinskyet al., 1977; Marsh and Whaler, 1980; Schumacher et al.,1989; Azevedo-Marques et al., 1992; Schmidt, 1995; Ownbyet al., 1997; Grisotto et al., 2006; Oliveira et al., 2007). Never-theless, animal studies have utilized different species, doses,

and routesof injection todetermineparticular venomactions,resulting in a somewhat fragmentary, and not alwayscomparable information. Most of these studies have utilizedthe intravenous and intraperitoneal injection routes. The aimof the present study was to develop a mouse model thatwould better resemble a massive bee attack, with a slowerabsorption from the skin, by using the subcutaneous (s.c.)route to induce a severe, sublethal systemic envenoming.With this model, an array of acute venom effects were char-acterized, including biochemical, hematological, histological,and inflammatory alterations.

2. Materials and methods

2.1. Bee venom, melittin and phospholipase A2

Dried venom from A. mellifera was obtained from twocommercial sources (Sigma–Aldrich, USA, or New Tech-niques Laboratory Ltd., Georgia) and stored at �20 �C. Forcomparative purposes, a small sample of pooled venomwas also obtained from honeybees collected in Costa Rica,by manual extraction of venom sacs after killing the insectsovernight at 2–4 �C. All venom solutions were freshlyprepared for each assay. Melittin and phospholipase A2(PLA2) were purified from the crude venom of New Tech-niques Laboratory by reverse phase-high performanceliquid chromatography (RP-HPLC) on a C8 column (Vydac,250 � 10 mm, 5 mm particle size), monitored at 280 nmusing an Agilent 1100 chromatograph. Venom aliquots of10–20 mg were dissolved in 0.5–1.0 ml of solution A (0.1%trifluoroacetic acid [TFA], 5% acetonitrile, 95% water),centrifuged, injected, and eluted at a flow rate of 2 ml/minwith a linear gradient from 0 to 70% of solution B (0.1% TFA,95% acetonitrile, 5% water) over 40 min. Fractions werecollected manually, dried by vacuum centrifugation, andstored at�20 �C. The PLA2 peak was identified by detectingits enzymatic activity on micellar phosphatidylcholineusing a phenol red-based colorimetric method (Lobo deAraújo and Radvanyi, 1987). Melittin was identified bycomparing its HPLC retention time with that of a syntheticstandard (Sigma–Aldrich), as well as by its direct hemolyticactivity in vitro, incubating it with a 5% suspension of rabbiterythrocytes in phosphate-buffered saline (PBS; 0.12 MNaCl, 40 mM sodium phosphate, pH 7.2) for 15min at 37 �C,and recording the release of hemoglobin to the supernatantspectrophotometrically at 540 nm.

2.2. Gel electrophoresis

Venom, melittin, and PLA2 were analyzed by sodiumdodecylsulphate-polyacrylamide gel electrophoresis (SDS-PAGE) using 5–15% gradient gels (Laemmli, 1970) or a tri-phasic gel system (Schagger and von Jagow, 1987), afterreduction with 5% 2-mercaptoethanol at 95 �C. Proteinswere stained with Coomassie blue R-250.

2.3. Animals

CD1 mice of either sex were utilized, following guide-lines of the Institutional Committee for the Use and Care ofResearch Animals (CICUA), University of Costa Rica.

M. Prado et al. / Toxicon 56 (2010) 1007–1017 1009

2.4. Median lethal dose (LD50) determinations

LD50 of honeybee venom was determined by injectingvariable amounts, dissolved in PBS, in groups of five mice(16–18 g body weight) by either intravenous (i.v.), intra-peritoneal (i.p.), or subcutaneous (s.c.) routes. Deaths wererecorded after 48 h, and the corresponding LD50 valueswere calculated by the Spearman–Karber method (WorldHealth Organization, 1981). For all subsequent toxicityexperiments, groups of mice (n ¼ 5) were treated with 0.5LD50 of bee venom by the s.c. route (corresponding to20.8 mg/g), in a volume of 0.2 ml. As controls, other groupsof mice were treated identically, but with 0.2 ml of PBSinstead of venom.

2.5. Hematologic parameters

Blood samples were collected from the retroorbitalplexus of mice at 3 and 6 h after the s.c. injection of 0.5 LD50of either venom or PBS, using 3.8% sodium citrate as anti-coagulant, under ether anesthesia, followed by euthanasia.This bleeding procedure, providing 0.5–1.0 ml of blood, wasutilized in all subsequent experiments. Hematologicprofiles (hemoglobin, hematocrit, and erythrocyte, leuko-cyte, and platelet counts) were determined on an auto-mated Coulter Act-8 instrument, as well as by microscopicobservation of Wright-stained blood smears for leukocytedifferential counts.

2.6. Coagulation parameters

Groups of mice received 0.5 LD50 of bee venom, orpurified PLA2 (3.1 mg/g), by the s.c. route. This PLA2 dosewas selected to match its relative content in 0.5 LD50 ofcrude venom (as determined by the HPLC analysisdescribed in Section 2.1). A control group received only PBS.Blood samples were obtained 3 h after injection, using 3.8%sodium citrate as anticoagulant. Prothrombin time wasdetermined on a Sysmex CA500 instrument, whereaspartial thromboplastin time and fibrinogen levels weredetermined on Stago STA-Compact equipment, followingmanufacturer’s instructions.

2.7. Serum biochemical and enzymatic parameters

Blood samples were obtained after 1, 4, 8, 12, 24, 48 or72 h after s.c. venom or PBS injection, and allowed to clot.Serum biochemical (blood urea nitrogen, creatinine, uricacid, sodium, and chloride) and enzymatic (creatine kinase,aspartate transaminase [AST], alanine transaminase [ALT])parameters were determined spectrophotometrically ona Synchron CX9 automated equipment.

2.8. Histological analyses

Groups of mice were euthanized by sodium pentobar-bital i.p. injection (0.02mg/g), at 4 and 24 h after s.c. venomor PBS injection, and dissected. Tissue samples of kidneys,lungs, liver, skeletal muscle, heart, brain, and spleen werefixed in 10% formalin, embedded in paraffin, and processedfor hematoxilyn-eosin staining.

2.9. Myotoxic activity

Groups of mice received 0.5 LD50 of either bee venom(20.8 mg/g), melittin (10.4 mg/g), or PLA2 (3.1 mg/g) by the s.c.route. These melittin and PLA2 doses were selected tomatch their relative content in 0.5 LD50 of crude venom (asdetermined by the HPLC analysis described in Section 2.1).Blood samples were obtained after 1, 2, 4, 8, 12, 24, and 48 hfor determination of plasma creatine kinase (CK) activity,using a kinetic UV assay (CK-Nac, Biocon). Control micereceiving an s.c. PBS injection were utilized to establishbasal reference values.

2.10. Edematogenic activity

Due to the inherent limitations in quantitating localedema at the s.c. site of injection used in all other in vivoexperiments (i.e. the dorsal skinofmice), a footpadassaywasutilized instead. Groups of five mice received either beevenom (5, 10 or 20 mg), melittin (2.5 or 5 mg), or PLA2 (1.5 or3 mg), in a total volume of 40 ml, by the s.c. route in their leftfootpad. Footpad thicknesswasmeasured before and 0.5,1, 2,3, 4, 6,12, 24, and48h after injection, byusing a low-pressurespring caliper (Oditest�) as described (Lomonte et al., 1993).Edema was expressed as a percentage increase, consideringthe footpad thickness before injection as 100%.

2.11. Serum pro-inflammatory cytokines

Blood samples were obtained at 1, 4, 8, 12, 24 or 48 h afters.c. venom injection, and allowed to clot. Serum levels of IL-1b, IL-6, and TNF-a were determined by enzyme-immuno-assay using the corresponding ELISA Ready-Set-Go� kits(eBioScience). Serum cytokine levels were expressed in pg/mlon the basis of calibration curves obtained using recombinantcytokine standards provided by the manufacturer. Controlmice receiving an s.c. PBS injection were utilized to establishbasal reference values.

2.12. Serum nitric oxide and lipid peroxidation

Blood samples were obtained at 1, 4, 12, 24, 48, and 72 hafter s.c. venom injection, and allowed to clot. As an indi-cator of nitric oxide production, serum nitrite was deter-mined indirectly by the Griess reaction (Green et al., 1982),and quantified photometrically on a Labsystems Multiskanmicroplate reader at 540 nm NaNO3 was utilized asa standard for the calibration curve, and results wereexpressed as mmol/l. Lipid peroxidation was determined insimilar serum samples, based on the reaction betweenmalondialdehide (MDA) and thiobarbituric acid (Contiet al., 1991). 1,1,3,3-tetramethoxypropane was utilized asa standard for the calibration curve, and results wereexpressed as mmol/l. Control mice receiving an s.c. PBSinjection were utilized to establish basal reference valuesfor both assays.

2.13. Statistical analyses

Results are expressed as mean � SEM. The significanceof differences between means of two experimental groups

M. Prado et al. / Toxicon 56 (2010) 1007–10171010

was evaluated by Student’s t-test, while differencesbetween more than two groups were evaluated by ANOVA.A p value <0.05 was considered significant.

3. Results

3.1. Bee venom and main fractions

The three samples of whole bee venom (Sigma, NewTechniques Laboratories, and a locally prepared sample)presented a similar electrophoretic pattern by SDS-PAGE, asexpected from the literature (Schumacher et al., 1990;Nelson et al., 1990). The venom from New TechniquesLaboratories, shown in Fig. 1A, was selected for all subse-quent experiments, as well as for the purification of PLA2

Fig. 1. Electrophoretic analysis of bee venom and purified melittin and phosphoconditions. Molecular mass markers (M) are indicated in kDa, at the left. The mdependent "smear" due to its high propensity to aggregation. The second most abuthe PLA2 and melittin (Mel) fractions, in comparison to whole venom (V). (C) RP-HPLThe peaks corresponding to PLA2 and melittin are indicated by arrows.

and melittin (Fig. 1B) by means of RP-HPLC (Fig. 1C). Inte-gration of the chromatogram signal (area under the curve)indicated that PLA2 and melittin constituted approximately15% and 50% of the proteins of this venom, respectively. Onthe basis of these values, doses of PLA2 or melittin wereselected to match their relative content in 0.5 LD50 ofvenom.

3.2. Median lethal dose (LD50) estimations

The lethal potency of bee venomwas very similar whenadministered to mice by the i.v. (LD50 4.3 � 1.5 mg/g) or i.p.(LD50 3.8 � 0.8 mg/g) routes. In contrast, when venomlethality was estimated by the s.c. route, the LD50 valueincreased to 41.6� 11.2 mg/g. This route was selected for the

lipase A2 (PLA2) (A) Triphasic SDS-PAGE of crude venom under reducingost prominent band ate3 kDa, melittin, commonly forms a concentration-ndant band ate17 kDa corresponds to PLA2 (B) Gradient SDS-PAGE (5–15%) ofC fractionation of crude venom on C8, as described in Materials and Methods.

Table 2Acute effects of bee (Apis mellifera) venoma or phospholipase A2

b subcu-taneous injection upon coagulation parameters in mice.

Parameter Control group Venom, 3 h PLA2, 3 h

Prothrombin time (sec) 9.2 � 0.5 10.4 � 0.5* 9.7 � 0.7Partial thromboplastin

time (sec)20.7 � 1.6 27.9 � 3.3* 24.4 � 1.4*

Fibrinogen (mg/dl) 226.2 � 72.8 155.6 � 45.7* 155.2 � 16.0*

*p < 0.05; results are expressed as mean � SD of five animals per group.a 0.5 LD , corresponding to 20.8 mg/g body weight.

M. Prado et al. / Toxicon 56 (2010) 1007–1017 1011

subsequent characterization of bee venom effects in mice,as it would better resemble the route of injection ina massive bee attack (with a slower absorption from theskin). In order to induce a severe, but sublethal envenom-ing of the animals, 0.5 LD50 of bee venom was injected bythe s.c. route in all experiments, corresponding to 20.8 mg/gbody weight. Within few minutes after this injection, miceappeared to be severely affected, but survived throughoutthe sampling times of the experiments in all cases.

50b 3.1 mg/g body weight, corresponding to the PLA2 content in 0.5 LD50 of

venom.

)l

2500A*

3.3. Hematologic and coagulation parameters

Evaluation of hematological parameters after the s.c.injection of 0.5 LD50 of bee venom is summarized in Table 1.The venom caused a statistically significant increase inhematocrit, erythrocyte counts, and hemoglobin by 6 h,whereas no significant changes in total leukocyte andplatelet counts occurred. The differential leukocyte countsdid not vary significantly between envenomed and controlmice (data not shown). On the other hand, the s.c. injectionof venom lead to a slight, but statistically significantprolongation in both prothrombin and partial thrombo-plastin times, as well as to a decrease in fibrinogenconcentration (Table 2). Since some toxic PLA2s are knownto affect coagulation, the purified bee venom PLA2 was alsoevaluated in this model. Injection of PLA2 by s.c. route, ata dose equivalent to its relative content in 0.5 LD50 ofvenom, prolonged the partial thromboplastin time andreduced the plasma fibrinogen concentration (Table 2).

0 12 24 36 48 72

m/U( ytivitca TSA

mureS

0

500

1000

1500

2000

)lm/

U( ytivitca TL

150

200

250 B

*

*

*

*

*

*

3.4. Serum biochemical and enzymatic parameters

Bee venom s.c. injection caused a marked increase ofserum AST since the first hour after injection, peakingbetween 6 and 12 h (Fig. 2A). Although less markedly thanAST, serum ALTalso increased rapidly, reaching amaximumat 12 h (Fig. 2B). The serum levels of both enzymes werestill above control values 24 h after venom injection,returning to normal by 48 h. Assays for renal functionrevealed several alterations induced by the venom, assummarized in Fig. 3. Serum creatinine was significantlyincreased at 3, 12, and 48 h, whereas urea nitrogen wasincreased 6 h after venom injection, thereafter returning tonormal. Uric acid also increased significantly in the sera ofmice receiving bee venom, at 3, 6 and 24 h (Fig. 3). Sodiumand chloride serum electrolytes showed a similar behavior,with an initial decrease 6 h after venom injection, followedby an increment that lasted up to 48 h, and normalized by

Table 1Acute effects of bee (Apis mellifera) venoma subcutaneous injection uponhematologic parameters in mice.

Parameter Control group Venom, 3 h Venom, 6 h

Hematocrit (%) 37.8 � 1.6 41.4 � 3.3 42.8 � 2.8*Hemoglobin (g/dl) 12.1 � 0.4 13.2 � 1.3 14.0 � 0.6*Erythrocytes (106/ml) 6.7 � 0.1 7.5 � 0.4* 7.4 � 0.3*Leukocytes (103/ml) 3.2 � 1.2 4.4 � 1.2 4.6 � 1.1Platelets (103/ml) 973.7 � 88.3 888.0 � 211.7 933.2 � 76.7

*p<0.05; results are expressed as mean � SD of four animals per group.a 0.5 LD50, corresponding to 20.8 mg/g body weight.

72 h (Fig. 4). The s.c. injection of venom also induceda prominent and rapid increase in serum CK levels (Fig. 5A),which returned to normal by 24 h (not shown). PurifiedPLA2 and melittin also induced an increment in serum CKwhen injected s.c., either individually or in combination(Fig. 5B–D), although this increase did not reach values ashigh as those recorded for whole venom (Fig. 5A).

3.5. Histological evaluation

Microscopical evaluation of tissue samples from lungs(Fig. 6D), skeletal muscle (Fig. 6F), spleen, liver, heart, andbrain (not shown) of mice receiving s.c. venom did not

Time (hr)0 12 24 36 48 72

A mureS

0

50

100

*

Fig. 2. Effect of the subcutaneous injection of 0.5 median lethal dose of beevenom (20.8 mg/g) upon serum activity of (A) aspartate transaminase (AST) and(B) alanine transaminase (ALT), in mice. Each point corresponds to themean�SEMoffiveanimals.Asterisks indicateastatisticallysignificant (p<0.05)difference from control mice that received a PBS injection (dashed line).

0 12 24 36 48 60 72

)ld/gm( eninitaer

C 0.2

0.3

0.4

0.5

0 12 24 36 48 60 72

)ld/gm( negortin aer

U 20

25

30

35

40

A

B

* * *

*

Time (hr)0 12 24 36 48 60 72

)ld/gm( dica cir

U

1.2

1.6

2.0

2.4

2.8C* * *

Fig. 3. Effect of the subcutaneous injection of 0.5 median lethal dose of beevenom (20.8 mg/g) upon serum creatinine (A), urea nitrogen (B), and uricacid (C), in mice. Each point corresponds to the mean � SEM of five animals.Asterisks indicate a statistically significant (p < 0.05) difference from controlmice that received a PBS injection (dashed line).

0 12 24 36 48 60 72

)M

m( muidoS

135

140

145

150

155

Time (hr)0 12 24 36 48 60 72

)M

m( edirolhC

105

110

115

120

125

A

B

*

* * *

*

*

*

Fig. 4. Effect of the subcutaneous injection of 0.5 median lethal dose of beevenom (20.8 mg/g) upon serum concentrations of sodium (A), and chloride(B), in mice. Each point corresponds to the mean � SEM of five animals.Asterisks indicate a statistically significant (p < 0.05) difference from controlmice that received a PBS injection (dashed line).

M. Prado et al. / Toxicon 56 (2010) 1007–10171012

reveal alterations, appearing identical to control tissuesamples. However, a hyaline eosinophilic material wasevident in the kidney samples, mainly within proximaltubules (Fig. 6B).

3.6. Inflammatory parameters

S.c. venom injection induced a macroscopically visiblelocal edema, which was nevertheless not possible toquantitate in the mouse dorsal skin. This effect was there-fore characterized using a footpad assay, where a rapidedema induced by crude venom (Fig. 7A), PLA2 (Fig. 7B), ormelittin (Fig. 7C) was demonstrated. Edema induced by

either of the two isolated components was more transientthan that induced by the whole venom, which appeared todisplay a biphasic profile instead. While footpad edemareturned to basal by 6–12 h after injection of PLA2 ormelittin, the venom caused a more sustained, dose-dependent effect (Fig. 7). PLA2 appeared more potent thanmelittin in inducing local edema.

The acute inflammatory reaction induced by s.c. injec-tion of venomwas also characterized by an elevation of thecytokines IL-6, TNF-a, and IL-1b in serum (Fig. 8). Thehighest increase was observed in the case of IL-6, witha nearly tenfold elevation at 2 h after venom injection.Serum TNF-a and IL-1b levels increased only marginally,albeit with statistically significant peaks at 4 h and 2–4 h,respectively. In addition to the release of proinflammatorycytokines, the acute reaction to venom included a rapid andtransient increase in serum malondialdehyde at 4 h afterinjection (Fig. 9A), and in serum nitrite at 1 and 2 h (Fig. 9B).

4. Discussion

Most experimental studies on the toxic actions of beevenom or its isolated components have been performedusing i.v. or i.p. injection routes (Schumacher et al., 1989,

0 4 8 120

500

1000

1500

2000

0 4 8 120

200

400

600

800

1000

0 4 8 12

)lm/

U( ytivitca KC

mureS

0

2000

4000

6000

8000

10000

Time (hr)0 4 8 12

0

100

200

300

400

A B

*

*

*

*

***

*

C D

*

*

* *

* *

*

*

Fig. 5. Effect of the subcutaneous injection of 0.5 median lethal dose of bee venom (20.8 mg/g) upon serum activity of creatine kinase (CK), in mice (A). Anequivalent dose of phospholipase A2 (3.1 mg/g) or melittin (10.4 mg/g) was injected in (B) and (C), respectively. In (D), a combined phospholipase A2 (3.1 mg/g) andmelittin (10.4 mg/g)injection was performed. Each point corresponds to the mean � SEM of five animals. Asterisks indicate a statistically significant (p < 0.05)difference from control mice that received a PBS injection (dashed line).

M. Prado et al. / Toxicon 56 (2010) 1007–1017 1013

1990; Schmidt, 1995), which do not mimick the route ofentrance occurring by stinging through the skin. In thiswork, we utilized the mouse as a model to characterize themain alterations induced by bee venom when injected bys.c. route to induce a sublethal, severe envenoming. The s.c.LD50 of this venom was nearly one order of magnitudeweaker in comparison to LD50 values obtained by i.v. or i.p.routes. To the best of our efforts, no information was foundin the literature on the s.c. LD50 of bee venom for mice. Thislower lethal potency of bee venom by the s.c. route is inagreement with data obtained with a variety of othervenoms (Minton andWeinstein, 1986; Consroe et al., 1992),and should relate to bioavailability, i.e. the proportion ofvenom components reaching systemic circulation afterabsorption from the site of injection. Thus, the presentresults suggest that a substantial fraction of the lethalcomponents of bee venom, when injected by s.c. route inmice, might be prevented to enter the systemic circulation.Venom inactivation by host factors and/or binding tolocally available tissue target sites would offer a reasonableexplanation for this finding.

The s.c. injection of 0.5 LD50 of bee venom induceda series of physiopathological alterations in mice, whichwithin minutes developed general signs of toxicity such aslocal swelling, slower breathing, piloerection, and localmuscle spasms. A number of acute effects were explored inthis experimental setting. Hematological changes werebasically confined to a hemoconcentration effect, withoutsignificant alterations in leukocyte and platelet counts.Hemoconcentration is likely to be a consequence of theability of venom to induce edema. This effect wasconfirmed using the footpad assay, where both PLA2 andmelittin were shown to individually contribute to theimmediate phase of edema, albeit with a lower potency

than the whole venom. Edema at the stung areas isa common sign in patients suffering massive honeybeeattacks, being especially evident in the facial region (Françaet al., 1994; Ariue, 1994; Betten et al., 2006). Phospholipidhydrolysis by PLA2, together with degranulation of mastcells by PLA2 and basic peptides of bee venom, have beenidentified as mechanisms involved in edema formation(Cirino et al., 1989; Hartman et al., 1991; Calixto et al.,2003).

Circulating platelet numbers remained unaltered afters.c. bee venom injection in mice, in similarity with severalclinical envenoming reports (França et al., 1994; Daheret al., 2003; Betten et al., 2006), although thrombocyto-penia may occur in some cases (Díaz-Sánchez et al., 1998;Bresolin et al., 2002; Mitchell, 2006). On the other hand,coagulation was moderately affected in this mouse model,where a delay in prothrombin and partial thromboplastintimes, together with a reduction of fibrinogen levels wererecorded. Likewise, coagulation abnormalities have beendocumented in some human envenomings (França et al.,1994; Daher et al., 2003; Mitchell, 2006). Bee venomPLA2 is known to alter coagulation parameters of humanplasma in vitro (Petroianu et al., 2000), in similarity withthe effect of whole venom (Ouyang et al., 1979), as well asto induce and to inhibit platelet aggregation in vitro(Ouyang and Huang, 1984). In the present study, s.c.injection of PLA2 alone reproduced the moderate coagu-lation abnormalities induced by the whole venom (atequivalent doses), further implicating this component inthe pathophysiology of such effect in vivo.

Serum biochemical markers evidenced that the s.c. injec-tion of bee venom caused damage to several organs andtissues of mice. Within few hours of injection, serum ALTlevels were significantly elevated, indicating an acute liver

Fig. 6. Histological evaluation of the effects of a subcutaneous injection of 0.5 median lethal dose of bee venom (20.8 mg/g) on kidney (B), lung (D), and skeletalmuscle (F) tissues of mice, after 24 h. Tissues of control animals that received a PBS injection (A, C, E). Staining corresponds to hematoxylin-eosin. Arrow in (B)points to the eosinophilic hyaline material in the kidney tubules. Magnification: 100�. A hyaline eosinophilic material was evident in the kidney samples, mainlywithin proximal tubules.

M. Prado et al. / Toxicon 56 (2010) 1007–10171014

injury, whereas the strong increase in serum AST observedcouldoriginatenotonly from liver, but also fromother sourcessuch as red blood cells or skeletal muscle. In addition,increases of serum creatinine, urea nitrogen, and uric acid,together with alterations in chloride and sodium electrolytelevels, are all indicative of significant kidney toxicity. More-over, skeletal muscle necrosis was reflected by the prominentincrease in serum CK induced by the venom. All these effectsare consistent with clinical observations, which report asmajor complications of human envenomings a generalizedrhabdomyolysis (with markedly elevated CK levels and myo-globinuria), the development of acute renal failure, and

multiorgan dysfunction with hepatic involvement andrespiratorydistress (França et al., 1994; Díaz-Sánchez et al.,1998; Kolecki, 1999; Bresolin et al., 2002; Daher et al., 2003,2009; Gabriel et al., 2004; Mitchell, 2006; Betten et al.,2006; Huertas-Franco and Bucknor-Masís, 2008).

Despite the clear alterations demonstrated in mice usinga variety of biochemical markers, histological evaluations didnot reveal widespread damage to major organs, with theexceptionofkidneys.Thesefindings indicate that theextentofacute tissue injury in thismodel of sublethal envenomingwassubtle and disperse, not leading to gross histological alter-ations. In the case of kidneys, however, the pigment deposits

Time (hr)0 3 6 9 12 24

)%( a

medE

0

20

40

605.0 μ g2.5 μ g

0 3 6 9 12 24

)%( a

medE

0

20

40

603.0 g1.5 μ g

0 3 6 9 24 48

)%( a

medE

0

20

40

6010 μ g20 μ g

A

B

C

μ

Fig. 7. Footpad edema induced by the subcutaneous injection of bee venom(A), phospholipase A2 (B), or melittin (C) at the indicated doses, in mice.Each point corresponds to the mean � SEM of five animals.

0 4 8 12 16 20 24

)lm/gp( 6-LI

0

1000

2000

3000

0 4 8 12 16 20 24-F

NT α

)lm/gp(

180

200

220

240

260

A

B

*

*

*

*

Time (hr)0 4 8 12 16 20 24

1-LIβ

)lm/g

m(

120

130

140

150

160

170 C

*

* *

**

*

Fig. 8. Effect of the subcutaneous injection of 0.5 median lethal dose of beevenom (20.8 mg/g) upon serum levels of IL-6 (A), TNF-a (B), and IL-1b (C), inmice. Each point corresponds to the mean � SEM of five animals. Asterisksindicate a statistically significant (p < 0.05) difference from control mice thatreceived a PBS injection (dashed line).

M. Prado et al. / Toxicon 56 (2010) 1007–1017 1015

clearly observed in tubules would be compatible with thepresence of myoglobin and/or hemoglobin, which have beenimplicated in the pathogenesis of acute renal failure (Françaet al., 1994; Daher et al., 2003; Grisotto et al., 2006). Usinga rat model of kidney damage after the i.v. injection of beevenom, Grisotto et al. (2006) provided evidence that vaso-constriction, direct nephrotoxicity and rhabdomyolyis, evenin the absence of hemolysis and hypotension, are importantmechanisms for the development of acute renal failure.

Rhabdomyolysis is a conspicuous feature of severehoneybee envenomings, which present strong elevations ofserum CK, as well as myoglobinuria (França et al., 1994;Betten et al., 2006). Previous studies demonstrated thatboth PLA2 and melittin from bee venom display potentmyotoxicity when injected by the i.m. route in mice(Ownby et al., 1997). In the present work, the myotoxiceffect of honeybee venom, PLA2, and melittin, when

administered by s.c. route is demonstrated for the firsttime. Interestingly, the CK increases induced by either PLA2or melittin alone were not as prominent as that induced bythe whole venom (at equivalent doses). Since a synergismbetween PLA2 and melittin actions has been described(Mollay and Kreil, 1974; Cajal and Jain, 1997), the combinedinjection of these two toxins was investigated. Neverthe-less, the CK increase recorded in such case was basicallyadditive, rather than synergistic, and did not reach the veryhigh levels induced by the whole venom. These findingssuggest that other venom components, which remain to beidentified, would contribute to its strong myotoxic activityby s.c. route. One possibility might be that hyaluronidase,known to promote the spreading of venom from theinjection site (Habermann, 1972), may play an importantsynergistic role by enhancing the systemic myotoxic effectof melittin and PLA2.

Time (hr)0 12 24 48 72

( edyhedlaidnolaM

μ )l/lo

m

0

10

20

30 *

Time (hr)0 12 24 48 72

(μ etirtiN

)l/lo

m

0

100

200

300

**

A

BLPS

Fig. 9. Effect of the subcutaneous injection of 0.5 median lethal dose of beevenom (20.8 mg/g) upon serum levels of malondialdehyde (A), or nitrite (B),in mice. Each point corresponds to the mean � SEM of five animals. Aster-isks indicate a statistically significant (p < 0.05) difference from control micethat received a PBS injection (dashed line). In (B), the nitrite serum levelsinduced by an s.c. injection of bacterial lipopolysaccharide (LPS; 5 mg/g in200 ml) after 8 h, as a positive control, is indicated.

M. Prado et al. / Toxicon 56 (2010) 1007–10171016

The toxic actions of bee venom on diverse tissue targetstriggered an acute inflammatory response. In addition toedema, a rapid release of proinflammatory cytokines (IL-1b,TNF-a, and especially IL-6)was recorded.Moreover, transientpeaks of serum malondialdehyde, an indicator of lipid per-oxidation by free radicals, and of serumnitrite, an indicator ofnitric oxide formation, appeared at 6 h, and 1–2 h, respec-tively. The systemic release of such potent inflammatorymediators by the host may add to the direct toxic action ofvenom, potentially contributing to the physiopathologicalalterations observed in this model. However, in similarity toother envenomings, it is not an easy task to determine ifinflammatory mediators appear mainly as a consequence oftissue damage, or conversely, if mediators actively contributeto such process (Lomonte et al., 1993; Texeira et al., 2009).Further experimental work along these lines, by using cyto-kine knock-out mice, leukocyte-deficient mice, or specificinhibitors of inflammatory pathways, for example, mightprovide highly relevant information to better understand thephysiopathology of bee envenomings.

In summary, honeybee venom was shown to inducea series of acute toxic effects in mice when administered by

the s.c. route, at a sublethal dose. These include at leastliver, skeletal muscle, and kidney damage, together withalterations in the coagulation system and a hemoconcen-tration effect, without marked changes in circulatingplatelet and leukocyte counts. Concomitantly, a notoriousinflammatory response ensues, including a significantedema, the involvement of lipid peroxidation and nitricoxide production, and systemic release of cytokines.Although these findings in mice cannot be directlyextrapolated to human envenomings, they seem to be inconcordance with the effects reported clinically. Someimportant clinical effects such as respiratory distressshould be focused with more detailed analyses in thisexperimental model, since histological evaluation did notreveal any major morphological changes. Finally, as theneed to develop a specific therapeutic antivenom to copewith massive honeybee attacks has been pointed outrepeatedly (França et al., 1994; Schumacher and Egen,1995;Jones et al., 1999), the mouse model here described maybecome useful to evaluate pre-clinically the efficacy ofantivenoms or other inhibitors, not only to preventlethality, but also to prevent or to minimize the importantphysiopathological effects here demonstrated.

Acknowledgements

We acknowledge the financial support to these studiesby CONARE and Vicerrectoría de Investigación, Universityof Costa Rica, as well as the collaboration of Drs. AndreaBadilla, Natalia Castillo, and Clas Une. This work was per-formed in partial fulfillment of the requirements for theM.Sc. degree of M. Prado at the University of Costa Rica.

Conflicts of interest statementThe authors have no conflicts of interest related to this

study.

References

Ariue, B.K., 1994. Multiple Africanized bee stings in a child. Pediatrics 94,115–117.

Azevedo-Marques, M., Ferreira, D.B., Costa, R.S., 1992. Rhabdomyonecrosisexperimentally induced in Wistar rats by Africanized bee venom.Toxicon 30, 344–348.

Betten, D.P., Richardson, W.H., Tong, T.C., Clark, R.F., 2006. Massive honeybee envenomation-induced rhabdomyolysis in an adolescent. Pedi-atrics 117, 231–235.

Bresolin, N.L., Carvalho, F.L.C., Goes, J.E.C., Fernandes, V.R., Barotto, A.M.,2002. Acute renal failure following massive attack by Africanized beestings. Pediatr. Nephrol. 17, 625–627.

Cajal, Y., Jain, M.K., 1997. Synergism between mellitin and phospholipaseA2 from bee venom: apparent activation of intervesicle exchange ofphospholipids. Biochemistry 36, 3882–3893.

Calixto, M.C., Triches, K.M., Calixto, J.B., 2003. Analysis of the inflamma-tory response in the rat paw caused by the venom of Apis melliferabee. Inflamm. Res. 52, 132–139.

Cirino, G., Peers, S.H., Wallace, J.L., Flower, R.J., 1989. A study of phospholi-pase A2-induced oedema in rat paw. Eur. J. Pharmacol. 166, 505–510.

Consroe, P., Gerrish, K., Egen, N., Russell, F.E.,1992. Intravenous dose-lethalitystudy of Americanpit viper venoms inmice using standardizedmethods.J. Wild. Med. 3, 162–167.

Conti, M.,Morand, P., Levillain, P., Lemonnier, A.,1991. Improved fluorometricdetermination of malonaldehyde. Clin. Chem. 37, 1273–1275.

Couch, T.L., Benton, A.W., 1972. The effect of the venom of the honey bee,Apis mellifera L., on the adrenocortical response of the adult male rat.Toxicon 10, 55–62.

M. Prado et al. / Toxicon 56 (2010) 1007–1017 1017

Daher, E.F., Junior, G.B.S., Bezerra, G.P., Pontes, L.B., Martins, A.M.C.,Guimarães, J.A., 2003. Acute renal failure after massive honeybeesting. Rev. Inst. Med. Trop. S. Paulo 45, 45–50.

Daher, E.F., Oliveira, R.A., da Silva, L.S.V., Silva, E.M.B., de Morais, T.P., 2009.Insuficiência renal aguda por picada de abelhas: relato de casos. Rev.Soc. Bras. Med. Trop. 42, 209–212.

Díaz-Sánchez, C.L., Lifshitz-Guinzberg, A., Ignacio-Ibarra, G., Halabe-Cherem, J., Quinones-Galvan, A., 1998. Survival after massive (>2000)Africanized honeybee stings. Arch. Intern. Med. 158, 925–927.

França, F.O., Benvenuti, L.A., Fan, H.W., Dos Santos, D.R., Hain, S.H., Picchi-Martins, F.R., Cardoso, J.L., Kamiguti, A.S., Theakston, R.D., Warrell, D.A., 1994. Severe and fatal mass attacks by "killer" bees (Africanizedhoney bees – Apis mellifera scutellata) in Brazil: clinicopathologicalstudies with measurement of serum venom concentrations. Q. J. Med.87, 269–282.

Gabriel, D.P., Rodrigues, A.G., Barsante, R.C., Silva, V.S., Caramori, J.T.,Martim, L.C., Barretti, P., Balbi, A.L., 2004. Severe acute renal failureafter massive attack of Africanized bees. Nephrol. Dial. Transpl. 19,2680.

Green, L., Warner, D., Glogowski, J., Skipper, P., Wishnock, J.,Tannenbaum, S., 1982. Analysis of nitrate, nitrite and [15N] in bio-logical fluids. Ann. Biochem. 126, 131–138.

Grisotto, L.S.D., Mendes, G.E., Castro, I., Baptista, M.A.S., Alves, V.A., Yu, L.,Burdmann, E.A., 2006. Mechanisms of bee venom-induced acuterenal failure. Toxicon 48, 44–54.

Habermann, E., 1972. Bee and wasp venoms. Science 177, 314–322.Hartman, D.A., Tomcheck, L.A., Lugay, J.R., Lewin, A.C., Chau, T.T.,

Carlson, R.P., 1991. Comparison of antiinflammatory and antiallergicdrugs in the melittin- and D49 PLA2-induced mouse paw edemamodels. Agents Actions 34, 84–88.

Harvey, P., Sperber, S., Kette, F., Heddle, R.J., Roberts-Thomson, P.J., 1984.Bee-sting mortality in Australia. Med. J. Austr. 140, 209–211.

Huertas-Franco, V., Bucknor-Masís, J., 2008. Insuficiencia renal agudaasociada a picadura de abeja africanizada. Acta Méd. Costarricense 50,57–60.

Ishay, J., Ben-Shachar, D., Elazar, Z., Kaplinsky, E., 1975. Effect of mellitinon the central nervous system. Toxicon 13, 277–283.

Johansson, B., Ericksson, A., Ornehult, L., 1991. Human fatalities caused bywasp and bee stings in Sweden. Int. J. Leg. Med. 104, 99–103.

Jones, R.G.A., Corteling, R.L., Bhogai, G., Landon, J., 1999. A novel Fab-basedantivenom for the treatment of mass bee attacks. Am. J. Trop. Med.Hyg. 61, 361–366.

Kaplinsky, E., Ishay, J., Ben-Shachar, D., Gitter, S., 1977. Effects of bee (Apismellifera) venom on the electrocardiogram and blood pressure. Toxicon15, 251–256.

Kolecki, P., 1999. Delayed toxic reaction following massive bee enveno-mation. Ann. Emerg. Med. 33, 114–116.

Laemmli, U.K., 1970. Cleavage of structural proteins during the assemblyof the head of bacteriophage T4. Nature 227, 680–685.

Langley, R.L., 2005. Animal related fatalities in the United States –

an update. Wildern. Environ. Med. 16, 67–74.Lobo de Araújo, A., Radvanyi, F., 1987. Determination of phospholipase A2

activity by a colorimetric assay using a pH indicator. Toxicon 25,1181–1188.

Lomonte, B., Tarkowski, A., Hanson, L.Å., 1993. Host response to Bothropsasper snake venom: analysis of edema formation, inflammatory cells,and cytokine release in a mouse model. Inflammation 17, 93–105.

Marsh, N.A., Whaler, B.C., 1980. The effects of honey bee (Apis mellifera L.)venom and two of its constituents, melittin and phospholipase A2, onthe cardiovascular system of the rat. Toxicon 18, 427–435.

de Medeiros, C., 2003. Himenópteros de importância médica. In: CostaCardoso, J.L., França, F.O., Wen, F.H., Sant’Ana Málaque, C.M.,Haddad, V. (Eds.), Animais Peçonhentos no Brasil: biologia, clínica eterapêutica dos acidentes. Sarvier Editora, São Paulo, Brasil, pp. 237–242.

deMedeiros, C.R., França, F.O., 2003. Acidentes por abelhas e vespas. In: CostaCardoso, J.L., França, F.O., Wen, F.H., Sant’Ana Málaque, C.M., Haddad, V.(Eds.), Animais Peçonhentos no Brasil: biologia, clínica e terapêutica dosacidentes. Sarvier Editora, São Paulo, Brasil, pp. 243–251.

Minton, S.A., Weinstein, S.A., 1986. Geographic and ontogenic variation invenom of the Western diamondback rattlesnake (Crotalus atrox).Toxicon 24, 71–80.

Mitchell, A., 2006. Africanized killer bees, a case study. Crit. Care Nurse26, 23–31.

Mollay, C., Kreil, G., 1974. Enhancement of bee venom phospholipase A2activity by melittin, direct lytic factor from cobra venom and poly-myxin B. FEBS Lett. 46, 141–144.

Nelson, D.R., Collins, A.M., Hellmich, R.L., Jones, R.T., Helm, R.M.,Squillace, D.L., Yunginger, J.W., 1990. Biochemical and immuno-chemical comparison of Africanized and European honeybee venoms.J. Allergy Clin. Immunol. 85, 80–85.

Oliveira, E.C., Pedroso, P.M., Meirelles, A.E., Pescador, C.A., Gouve, A.S.,Driemeier, D., 2007. Pathological findings in dogs after multipleAfricanized bee stings. Toxicon 49, 1214–1218.

Oliveira, F.A., Guimaraes, J.V., dos Reis, M.A., Teixeira, V.P.A., 2000. Acid-ente humano por picadas de abelhas africanizadas. Rev. Soc. Bras.Med. Trop 33, 403–405.

Ouyang, C., Lin, S.C., Teng, C.M., 1979. Anticoagulant properties of Apismellifera (honey bee) venom. Toxicon 17, 197–201.

Ouyang, C., Huang, T.F., 1984. Effect of the purified phospholipases A2 fromsnake andbee venomson rabbit platelet function. Toxicon 22, 705–718.

Ownby, C.L., Powell, J.R., Jiang, M.S., Fletcher, J.E., 1997. Melittin andphospholipase A2 from bee (Apis mellifera) venom cause necrosis ofmurin skeletal muscle in vivo. Toxicon 35, 67–80.

Petroianu, G., Liu, J., Helfrich, U., Maleck, W., Rufer, R., 2000. PhospholipaseA2-induced coagulation abnormalities after bee sting. Am. J. Emerg.Med. 18, 22–27.

Prado, M., Quirós, D., Lomonte, B., 2009. Mortality due to Hymenopterastings in Costa Rica,1985–2006. Pan-Amer. J. Publ. Health 25, 389–393.

Riches, K.J., Gillis, D., James, R.A., 2002. An autopsy approach to bee sting-related deaths. Pathology 34, 257–262.

Rinderer, T.E., Oldroyd, B.P., Sheppard, W.S., 1993. Africanized bees in theU.S. Sci. Am 269, 84–90.

de Roodt, A.R., Salomón, O.D., Orduna, T.A., Robles, L.E., Paniagua, J.F.,Alagón, A., 2005. Envenenamiento por picaduras de abeja. Gac. MédMéx 141, 215–222.

Schagger, H., von Jagow, G., 1987. Tricine-sodium dodecyl sulfate-poly-acrylamide gel electrophoresis for the separation of proteins in therange from 1 to 100 kDa. Anal. Biochem. 166, 368–379.

Schmidt, J.O., 1995. Toxinology of venoms from the honeybee genus Apis.Toxicon 33, 917–927.

Schumacher, M.J., Egen, N.B., 1995. Significance of Africanized bees forpublic health. Archs. Intern. Med. 155, 2038–2043.

Schumacher, M.J., Schmidt, J.O., Egen, N.B., 1989. Lethality of "killer" beestings. Nature 337, 413.

Schumacher, M.J., Schmidt, J.O., Egen, N.B., Lowry, J.E., 1990. Quantity,analysis, and lethality of European and Africanized honey beevenoms. Am. J. Trop. Med. Hyg. 43, 79–86.

Spivak, M., 1991. The Africanization process in Costa Rica. In: Spivak, M.,Fletcher, D.J.C., Breed, M.D. (Eds.), The "African" Honey Bee. WestviewPress, Colorado, pp. 130–148.

Spivak, M., 1992. The relative success of Africanized and European honey-bees over a range of life-zones in Costa Rica. J. Appl. Ecol. 29, 150–162.

Taylor, O.R., 1986. Health problems associated with African bees. Ann.Intern. Med. 104, 267–268.

Texeira, C., Cury, Y., Moreira, V., Picolo, G., Chaves, F., 2009. Inflammationinduced by Bothrops asper venom. Toxicon 54, 988–997.

Vick, J.A., Mehlman, B., Brooks, R., Phillips, S.J., Shipman, W., 1972. Effect ofbee venom and melittin on plasma cortisol in the unanesthetizedmonkey. Toxicon 10, 581–586.

Vick, J.A., Shipman, W.H., 1972. Effects of whole bee venom and its frac-tions (apamin and melittin) on plasma cortisol levels in the dog.Toxicon 10, 377–380.

Winston,M.L.,1994. The Africanized "killer" bee: biology andpublic health.Q.J. Med. 87, 263–267.

World Health Organization, 1981. Progress in the Characterization ofVenoms and Standarization of Antivenoms. W.H.O. offset publicationN� 58, Geneva.