Clinical and Pharmacological Investigation of Myotoxicity in Sri Lankan Russell’s Viper (Daboia russelii) Envenoming

Background Sri Lankan Russell’s viper (Daboia russelii) envenoming is reported to cause myotoxicity and neurotoxicity, which are different to the effects of envenoming by most other populations of Russell’s vipers. This study aimed to investigate evidence of myotoxicity in Russell’s viper envenoming, response to antivenom and the toxins responsible for myotoxicity. Methodology and Findings Clinical features of myotoxicity were assessed in authenticated Russell’s viper bite patients admitted to a Sri Lankan teaching hospital. Toxins were isolated using high-performance liquid chromatography. In-vitro myotoxicity of the venom and toxins was investigated in chick biventer nerve-muscle preparations. Of 245 enrolled patients, 177 (72.2%) had local myalgia and 173 (70.6%) had local muscle tenderness. Generalized myalgia and muscle tenderness were present in 35 (14.2%) and 29 (11.8%) patients, respectively. Thirty-seven patients had high (>300 U/l) serum creatine kinase (CK) concentrations in samples 24h post-bite (median: 666 U/l; maximum: 1066 U/l). Peak venom and 24h CK concentrations were not associated (Spearman’s correlation; p = 0.48). The 24h CK concentrations differed in patients without myotoxicity (median 58 U/l), compared to those with local (137 U/l) and generalised signs/symptoms of myotoxicity (107 U/l; p = 0.049). Venom caused concentration-dependent inhibition of direct twitches in the chick biventer cervicis nerve-muscle preparation, without completely abolishing direct twitches after 3 h even at 80 μg/ml. Indian polyvalent antivenom did not prevent in-vitro myotoxicity at recommended concentrations. Two phospholipase A2 toxins with molecular weights of 13kDa, U1-viperitoxin-Dr1a (19.2% of venom) and U1-viperitoxin-Dr1b (22.7% of venom), concentration dependently inhibited direct twitches in the chick biventer cervicis nerve-muscle preparation. At 3 μM, U1-viperitoxin-Dr1a abolished twitches, while U1-viperitoxin-Dr1b caused 70% inhibition of twitch force after 3h. Removal of both toxins from whole venom resulted in no in-vitro myotoxicity. Conclusion The study shows that myotoxicity in Sri Lankan Russell’s viper envenoming is mild and non-life threatening, and due to two PLA2 toxins with weak myotoxic properties.


Introduction
Snake bite is a significant public health issue in the tropics [1]. Coagulopathy, neuromuscular paralysis, acute kidney injury and local effects are the most important clinical syndromes of snake envenoming [2]. Russell's viper bites cause a large number of envenomings across Asia, and are more medically important than any other snake in the region [3,4]. Both species of Russell's vipers, i.e. D. russelii (found in Sri Lanka, India, Pakistan, Nepal, Bangladesh) and D. siamensis (found in some parts of southeast and east Asia such as Cambodia, Myanmar, Thailand, Taiwan, South China and, some parts of Indonesia including East Java), commonly cause venom-induced consumption coagulopathy, acute kidney injury and mild local effects throughout their distribution [3,5]. However, there is geographical variation in the clinical effects of Russell's viper envenoming with neuromuscular paralysis and rhabdomyolysis only reported from Sri Lanka and South India [3,[6][7][8][9]. In a recent clinical and neurophysiological investigation of neuromuscular paralysis in Sri Lankan Russell's viper envenoming, we showed that the paralysis is mild, non-life threatening with no long term effects [6]. Clinical evidence of myotoxicity, including local and generalized myalgia, muscle tenderness, and dark red or black coloured urine suggestive of myoglobinuria, has been reported in cases of Russell's viper envenoming in Sri Lanka [7][8][9][10]. Elevation of plasma and urinary myoglobin concentrations were reported in 19 Russell's viper bite patients from Sri Lanka further suggesting the existence of myotoxicity in Sri Lankan Russell's viper patients [10]. A recent study of Sri Lankan Russell's viper venom injected into mice reported an elevation in creatine kinase, also suggesting that the venom is myotoxic in-vivo [11].
Myotoxicity is an important effect of snake envenoming and can manifest locally and systemically [12]. Local muscle necrosis is a component of the local necrotic effects [12,13]. Systemic myotoxicity has been reported following envenoming by some vipers [14,15], sea snakes [16][17][18], Australasian elapids [19,20] and some Asian kraits [21,22]. Systemic myotoxicity ranges in severity. Loss of functioning muscle cells due to myotoxicity can aggravate co-existing weakness due to neuromuscular block caused by neurotoxins. More importantly rhabdomyolysis can cause secondary acute kidney injury [23,24] and itself can result in lifethreatening hyperkalaemia due to extensive muscle cell damage [15].
Although many cytotoxic components in snake venoms may contribute to the development of myotoxicity, the most important snake venom myotoxins are phospholipase A 2 (PLA 2 ) toxins [12,25]. Three types of PLA 2 myotoxins, commonly referred as Asp49, Ser49 and Lys49 PLA 2 , have been characterised in viperid venoms. Despite structural similarity, the latter two types lack catalytic activity, and are referred to as 'PLA 2 like' peptides [26][27][28]. It is important to note that the enzymatic activity and the myotoxic activity of PLA 2 myotoxins are independent [29]. Myotoxic PLA 2 s cause muscle damage primarily by destruction of the sarcolemma [12,30]. Several PLA 2 toxins have been isolated from Sri Lankan Russell's viper venom and biochemically characterised, [31,32] Some of these PLA 2 toxins contain a unique Serine residue at the N-terminus (S-type), while others have an Asparagine residue at the N-terminus (N-type) [32]. We have recently shown that the pre-synaptic neurotoxicity of the Sri Lankan Russell's viper venom is primarily due to one of the S-type PLA 2 toxins, which we named U1-viperitoxin-Dr1a [33].
There are several gaps in our understanding of the myotoxicity associated with Sri Lankan Russell's viper envenoming. It is unclear if Sri Lankan Russell's viper envenoming causes severe myotoxicity and whether any resultant myotoxicity can be treated with Indian Polyvalent antivenom. Further, the in vitro myotoxicity of Sri Lankan Russell's viper venom has not been pharmacologically studied. This requires isolation and pharmacological characterisation of the myotoxins from the venom. The present study aims to investigate the clinical severity of myotoxicity from Russell's viper envenoming and isolate the toxins responsible for this activity.

Clinical study
Ethics, study setting and patients. We carried out a prospective cohort study of confirmed Russell's viper bite patients aged over 16 years admitted to the Teaching Hospital Anuradhapura, Sri Lanka for a 14 month-period. The hospital is a tertiary care hospital which admits over 1000 suspected snake bite cases annually. Ethics clearance for this study was granted by the Human Research Ethics Committees of the University of Peradeniya (Approval No: 2012/EC/63), Sri Lanka, Rajarata University of Sri Lanka (Approval No: ERC 2013/019) and Monash University, Australia (Approval No: CF14/970-2014000404). Written informed consent was obtained prior to the recruitment of all patients. For patients 16 and 17 years of age, consent was also obtained from the patient's parent or guardian. The recruitment of patients to this cohort study has been previously described. [6] Case authentication. Cases were authenticated by either positive identification of the snake specimen involved as a Russell's viper by AS, who is a herpetologist, or by detection of the specific Russell's viper venom antigens in blood samples of the patients using enzymeimmunoassay (EIA). Clinical examination of all patients was  undertaken on admission to hospital, 1 h, 4 h, 8 h, 12 h and 24 h post-bite, and then every 24 h  until discharge. In particular, muscle pain and tenderness, in the bitten limb (local myotoxicity) and the other limbs (generalized myotoxicity) were assessed as part of the clinical examinations. Blood samples were collected from patients on admission and then 1 h, 4 h, 8 h, 12 h  and 24 h post-bite, and daily thereafter. All blood samples were immediately centrifuged, serum aliquoted and frozen initially at -20˚C before transfer to -80˚C storage until they were analysed.

Clinical data collection and blood collection.
Antivenom. Indian polyvalent antivenoms from VINS Bioproducts, India (batch numbers: 01AS11118, 1119, 1121, 1123, 3100, 4001, 4025, 4026, 4031) and BHARAT Serums and Vaccines Ltd, India (batch number: A5311018) were used throughout the study period. In the present cohort, the majority of patients received antivenom and a small proportion of these received their first dose of antivenom at a primary care hospital, before being transferred to the study hospital. The decision to administer antivenom was made solely by the treating physician, based on clinical and laboratory evidence of systemic envenoming. Patients were administered between 10 to 20 vials of antivenom which is the standard dose. Antivenom was administered in normal saline and infused over 1h. If an acute adverse reaction to antivenom occurred, the infusion was stopped for 5 to 10 min and the patient treated with adrenaline, antihistamines and corticosteroids, as per the treating clinicians.
Russell's viper venom enzyme immunoassays. Measurement of venom in the samples was carried out using a Russell's viper venom specific Enzyme-Immunoassay (EIA) which has been previously described [6,34,35]. In brief, rabbit IgG was raised against Russell's viper venom from Sri Lanka. Antibodies were bound to microplates as well as conjugated to biotin as the detecting antibodies with streptavidin-horseradish peroxidase. The lower limit of detection for the assay was 2.5 ng/ml. In cases where no venom was detected in pre-antivenom samples or pre-antivenom samples were not available, the post-antivenom samples were subjected to heat dissociation treatment and then tested for venom using the same EIA [35].
Serum creatine kinase (CK) assay. Creatine kinase (CK) was measured using the Thermo Scientific Ltd CK-NAC reagent kit on stored frozen serum samples. Patient serum samples were thawed and 15 μL added to 300 μL of reagent. After 120 s, the samples were read in a plate reader (340 nm, 37˚C) for 3 min. An abnormal CK was defined as a concentration greater than 300 U/l. A screening assay was done on all 24 h post-bite samples, or as close to this time point as possible. If the screening assay was abnormal (> 300 U/L), then serial CK assays were carried out for all samples available for that patient.
Data analysis. Continuous variables were reported as medians, range and interquartile range (IQR). Associations between venom concentration and CK measurements were assessed by Spearman's correlation test. The CK concentrations between groups of patients was compared with the Kruskal-Wallis test. Analysis of clinical data was done using PRISM, version 6.05 (GraphPad Software, Inc.).

Pharmacological study
Venom and antivenom. Freeze-dried Russell's viper (D. russelii) venom from Sri Lanka (pooled from both sexes, including juveniles) donated by Professor Ariaranee Gnanadasan (Faculty of Medicine, University of Colombo, Sri Lanka) was used for this study. Venom was dissolved in MilliQ water and stored at -20˚C until required. Protein quantification of the venom, fractions and toxins was carried out using a BCA Protein Assay Kit (Thermo Fisher Scientific, Rockford, IL, USA), as per the manufacturer's instructions. The relative abundance of the two isolated toxins was determined by the area under the chromatogram of the whole venom. Isolation of the two toxins was performed as described, from whole venom, using the two chromatographic techniques. Here, the two toxins separately and the venom without the two toxins were collected by multiple HPLC runs and pooled separately. Then, the three samples (i.e. venom without both toxins, and the two toxins) were dissolved separately. Based on the protein assay, the protein/toxin amounts in the samples were determined. Then the different combinations of the venom with and without individual toxins were made by mixing the venom without the two toxins with appropriate toxin according to the proportion present in the venom.
Indian polyvalent antivenom (VINS Bioproducts, Thimmapur, India; batch No: 01AS14035) was used for the study. The antivenom was reconstituted in 2 mL of sterile injection water. According to the manufacturer, 1/10 of the antivenom in a single vial neutralises 0.6 mg of Russell's viper venom. For in-vitro studies examining the neutralisation of myotoxins, the required amount of antivenom was calculated based on the relative abundance of the toxin in the venom.
In-vivo myotoxicity studies. Male Sprague-Dawley rats (300-360g) were anaesthetised with sodium pentobarbital (60-100 mg/kg, i.p.) with additional anaesthetic being administered during the experiment as required. An incision was made in the cervical region, the carotid artery was cannulated for blood sample collection and the trachea intubated for artificial respiration, if required. The rats were kept on small animal surgery tables with their body temperature maintained at 37˚C. Venom (250 μg/kg in 50 μl 0.9% saline, n = 4) was injected into the gastrocnemius muscle of the left hind-limb. For control rats (n = 4), 50 μl 0.9% saline was injected in the same location. Blood samples (500 μl) were collected into MiniCollect LH/ Gel separation tubes (Greiner bio-one), through the carotid cannula immediately before the venom/saline injection and at 1.5, 3, 4.5 and 6 h post-injection. At 6 h, rats were humanely killed with an overdose of sodium pentobarbitone. For the baseline measurements of CK, blood samples of an additional seven 'control' (i.e. not envenomed) rats were used. Measurement of CK was done using the same CK assay method used for human samples.
Size-exclusion chromatography: Crude venom (500 μg) was reconstituted in buffer A (ammonium acetate; 0.1 M, pH 6.8). Venom was then separated using a Superdex G-75 column (13 μm; 10 mm × 300 mm; GE Healthcare, Uppsala, Sweden) equilibrated with buffer A. Elution was performed at a flow rate of 0.5 ml/min and monitored at 280 nm. All fractions were freeze-dried immediately and later screened using the chick biventer cervicis nerve-muscle preparation (see 2.2.6 below) to identify those with myotoxicity.

Matrix-Assisted Laser Desorption/Ionization (MALDI-TOF) Mass
Spectrometry. MALDI-TOF Mass Spectrometry analysis was performed in order to determine the intact mass of the toxins, with an Applied Biosystems (Forster City, CA, USA) 4700 TOF TOF Proteomics Analyser, as described previously (Chaisakul et.al.). [36] Amino acid sequence determination by Liquid Chromatogarphy-Mass Spectrometry (LC-MS). Freeze-dried U1-viperitoxin-Dr1b dissolved in 100 mM ammonium bicarbonate buffer, 2 mM DL-Dithiothreitol, incubated for 20 min at 65˚C and alkylated with 5 mM chloroacetamide. The protein mixture was digested with trypsin at 1:100 (w/w) ratio for 18 h at 37˚C. The resulting peptides were desalted with Pierce C18 Spin columns (ThermoFisher Scientific), dried, resuspended in 2% Acetronitrile, 0.1% TFA and analysed by LC-MS/MS analysis as previously described [37]. The raw data acquired by the mass spectrometer was converted into a centroided peaklist file using ProteoWizard (version 3.0.9967) [38]. The MS/ MS results were analysed with PMI-Byonic software (Protein Metrics, ver 2.8.14) against Uniprot serpents (taxID:8570) protein sequences (update 07/2016) using the following search parameters: trypsin digestion with up to 2 missed cleavages, fixed cysteine alkylation, variable methionine oxidation and asparagine and glutamine deamidation, precursor mass tolerance 10 ppm and fragment mass tolerance 20ppm. PLA 2 activity. PLA 2 activity of the venom and U1-viperitoxin-Dr1b was determined using a secretory PLA 2 assay kit (Cayman Chemical, Ann Arbor, MI, USA) as described previously [33] and according to the manufacturer's instructions.
Chick biventer cervicis nerve-muscle preparation. Male chickens (aged 4-10 days) were humanely killed by exsanguination following CO 2 inhalation. Biventer cervicis nerve-muscle preparations were dissected and then mounted on wire tissue holders under 1 g resting tension in 5 ml organ baths. Tissues were maintained at 34˚C, bubbled with 95% O 2 and 5% CO 2 , in physiological salt solution of the following composition (mM); 118.4 NaCl, 4.7 KCl, 1.2 MgSO 4 , 1.2 KH 2 PO 4 , 2.5 CaCl 2 , 25 NaHCO 3 and 11.1 glucose. Direct twitches were evoked by stimulating the muscle belly (rate: 0.1 Hz; pulse duration: 2 ms) at supramaximal voltage (20-30 V) with the two electrode loops directly encircling the muscle belly, using a Grass S88 stimulator (Grass Instruments, Quincy, MA). Contractile responses of the tissues to KCl (40 mM for 30 s) was obtained in the absence of muscle stimulation. Following that, direct stimulation of the muscle was resumed. Selective stimulation of the muscle was ensured by abolishment of any indirect twitches by the continual presence of adding d-tubocurarine (10 μM) to the organ bath. The preparations were then stimulated for 30min, before the addition of venom or toxins. For antivenom experiments, the tissues were equilibrated with antivenom for 20 min before the venom or toxin was added. The preparation was exposed to the treatment/control until abolishment of direct twitches or for a period of 3 h. At the conclusion of the experiment, KCl was re-added as above.
Data analysis and statistics. The quantity of the toxins in the venom was determined by measuring the area under the curve of the size-exclusion and RP HPLC venom elution profiles. Direct twitch responses and responses to KCl were measured via a Grass FTO3 force displacement transducer and recorded on a PowerLab system (ADInstruments Pty Ltd., Australia). Responses were expressed as percentages of their pre-venom/toxin values. T 90 values (i.e. time taken for 90% inhibition of the maximum twitch response to occur) were determined for the two toxins. To compare the responses to exogenous agonists following the administration of venom a one-way ANOVA was used. All ANOVAs were followed by Bonferroni's multiple comparison post-tests. Data are presented in the form of mean ± standard error of the mean (S.E.M.) of n experiments. All statistical analyses and presentation of data were generated using GraphPad Prism 6.05 software. For all statistical tests p < 0.05 was considered statistically significant.
Animal ethics. All animal experiments used in this study were approved by the Monash University Animal Ethics Committee (Approval No: MARP/2014/097).

Myotoxicity in authenticated cases
Clinical features of myotoxicity. Of the 245 authenticated Russell's viper bite patients recruited to the study, 177 (72.2%) had myalgia and 173 (70.6%) tenderness in the bitten limb (local myotoxicity). Thirty-five patients (14.2%) and 29 patients (11.8%) had generalized myalgia and generalized muscle tenderness, respectively. Changes in local and generalized myalgia and muscle tenderness over time among the study participants is shown in Fig 1A and 1B. Myalgia and muscle tenderness resolved in over 80% of the patients by day 4. Six patients (2.4%) had dark coloured urine observed between 12 and 48 h after the bite. The clinico-epidemiological data related this cohort of patients is described in the Table 1.  ; Fig 2A). There was no statistically significant association between 24 h serum CK concentration and peak venom concentrations (Spearman's correlation; p = 0.48; Fig 2B). Similarly, there was no relationship between the time delay from the snakebite to first dose of antivenom and 24 h serum CK (Fig 2C). Of   Of these, 135 patients were given antivenom within 6 h and 31 of these developed a CK > 300 U/l at 24 h, compared to 31 patients who received antivenom after 6 h with 6 developing a CK > 300 U/l at 24 h. Only three of the six patients with dark coloured urine had CK concentrations above 300 U/l (404, 633 and 856 U/l,). However, red blood cells were reported in three of these six patients (S1 Table). The dark coloured urine in Russell's viper envenoming could be due to myoglobinuria, haemoglobinuria, haematuria or a combination of all these.
The serial CK concentrations of the 37 patients with high CK had a median peak value of 666 U/l (maximum: 1066). In most patients, CK reached a peak in 24 h after the bite (Fig 2D). By the time CK started to rise, the venom concentrations in these patients had already decreased, and in most cases was lower than the level of detection (Fig 2E).

In-vivo myotoxic effects in anaesthetised rats
The plasma CK concentrations of rats injected with Sri Lankan Russell's viper venom 250 μg/ kg (i.m.) were not different tocontrol rats and were within the normal range until 6 h of the venom injection (S1 Fig).

Isolation and biochemical characterisation of U1-viperitoxin-Dr1a and U1-viperitoxin-Dr1b
Liquid Chromatography. Eight major fractions were identified with size-exclusion chromatography of the crude venom of Russell's viper (Fig 3A). Peak E, eluting around 33 min showed marked myotoxicity in the chick biventer cervicis nerve-muscle preparation. Peak E was further analysed and fractionated using RP-HPLC (Fig 4B). In vitro myotoxicity was largely confined to two peaks with retention times of 31 min and 38 min, respectively (Fig 3B). The former venom component eluting at 31 min was subsequently isolated as U1-viperitoxin-Dr1b, and constituted 22.7% (SD: 1.2; n = 10) of the crude venom, based on the area under the curve of the size exclusion chromatography and RP-HPLC profiles of the venom. The component eluting at 38 min was named U1-viperitoxin-Dr1a, a PLA 2 toxin which has been previously characterised biochemically and found to be the major neurotoxin (with a relative abundance of 19.2%) in Sri Lankan Russell's viper venom [33].
Intact protein analysis of U1-viperitoxin-Dr1b by MALDI-TOF. Analysis of U1-viperitoxin-Dr1b by MALDI-TOF mass spectrometry indicated a single species with a mass/charge ratio of 13564 Da (Fig 4). A doubly charged ion with a mass/charge ratio of 6779.8 Da was also observed.
PLA 2 activity of U1-viperitoxin-Dr1b. PLA 2 activity of U1-viperitoxin-Dr1b was 1139 (+/-19) μmol/min/mg. Pharmacological investigation of the myotoxicity. Sri Lankan Russell's viper venom caused concentration-dependent inhibition of direct twitches in the chick biventer nerve muscle preparation (Fig 5A) and inhibited the response to 40 mM KCl (Fig 5B). However, high venom concentration such as 80μg/ml did not completely abolish either the direct twitches or the response to KCl, even after 3 h. U1-viperitoxin-Dr1b ( Fig 5C) and U1-viperitoxin-Dr1a (Fig 5D) concentration-dependently inhibited the direct twitches and also inhibited the response to KCl (Fig 5E). When the response to KCl was compared between the crude venom 50μg/ml, each of the two toxins of the same amount of venom individually, a combination of the two toxins, and the whole venom without individual or both toxins, removal of the two toxins led to a loss of all myotoxicity from the venom (Fig 5F). Antivenom, at the recommended concentrations, did not neutralise the myotoxic effect of the venom (Fig 5B)

Discussion
The present study demonstrates that myotoxicity seen in humans following Sri Lankan Russell's viper envenoming is mild, based on the relatively low frequency of generalized myotoxic features and only a very mild elevation in CK. There were small but statistically significant association between clinical features of myotoxicity and the 24 h CK measurements. The rats injected intramuscularly with Sri Lankan Russell's viper venom did not show an elevation of CK within the first 6 h. Although Sri Lankan Russell's viper venom displays concentrationdependent myotoxic effects in-vitro, very high venom concentrations (i.e. >50 μg/ml) are required to completely inhibit direct twitches in the chick biventer nerve-muscle preparation and the response to KCl. Antivenom at recommended concentrations was unable to neutralise the myotoxic properties of the venom in-vitro. The myotoxicity of the whole venom appears to be due to two abundant PLA 2 toxins in the venom, U1-viper1toxin-Dr1a and U1-viperitoxin-Dr1b (named herein), which both have in vitro myotoxic activity, but only at very high concentrations (3 μM). Features of systemic myotoxicity such as generalized muscle tenderness and myalgia, and dark coloured urine were observed in less than 15% of the patients. In addition, three of the six patients with dark coloured urine did not have 24 h CK concentrations outside the normal limit, so the dark urine was more likely due to macroscopic haematuria (due to haemorrhage) which occurs with Russell's viper envenoming. Peak CK concentrations were abnormally high in less than 17% of patients, and even in these cases there were only modest elevations with the median and maximum being 666 and 1066 U/l, respectively. In comparison, the reported median peak CK values following envenoming by mulga snakes and tiger snakes in Australia were 3100 U/l [20] and 4749 U/l [19], with severe myotoxicity being associated with CK values over 100,000 U/L. This means that myotoxicity in Russell's viper envenoming is uncommon and mild when it occurs compared to other myotoxic snakes. The median 24 h CK concentrations were higher in symptomatic patients demonstrating that the clinical features were consistent with biochemical evidence of mild muscle injury.
Anaesthetised rats injected with venom showed no significant increase in serum CK concentrations compared to saline injected rats, even 6 h after the venom injection. Assuming that the blood volume of a rat with a body weight of 330g is approximately 25 ml, the maximum serum venom concentration that this venom dose (i.e. 250 μg/kg), with 100% systemic absorbance, could give rise to is 3300 ng/ml. This venom concentration is greater than the maximum venom concentration observed in the Russell's viper envenomed patients in this study, hence the venom dose used for the envenoming is clinically relevant. The absence of a significant CK rise in the envenomed rats during the 6 h observation period may be due to a combined effect of the weak myotoxicity of the venom and the period of observation being too short to observe any CK elevation. These results are not in agreement with the previously reported high CK concentrations (mean: 16,000 U/l) in mice, 3 h after the intramuscular injection of 5 μg (250 μg/kg dose in 20 g mouse) of Sri Lankan Russell's viper venom [11], despite similar venom doses in the two studies. The reason for the discrepancy is unclear but such a large early increase in CK concentrations is unusual and the mouse model has not previously been validated.
The clinical findings were consistent with the in vitro studies in which two low potency myotoxins were found to be relatively abundant in the venom. Very high venom concentrations up to 80 μg/ml (80 times higher than the maximum peak venom concentration observed in envenomed patients) were required to decrease the direct twitch force by two thirds within 3 h. The low potency of both myotoxins was evident from the fact that high concentrations (3 μM) were required for significant in-vitro myotoxic activity. Venom with these two toxins removed did not cause myotoxicity confirming that these toxins are the only major myotoxins in Sri Lankan Russell's viper venom.
doi:10.1371/journal.pntd.0005172.g005 symptoms in 77% of the cases but no biochemical confirmation of myotoxicity. This study included only severe envenoming with 19/22 with acute kidney injury and 6 deaths. In 1988, Phillips et. al. [10] reported myotoxic symptoms in 32% of the cases, but again this only included more severe envenoming (4/23 died). The mean venom concentration of 375 ng /ml (range: 16.5-702 ng/ml) [10] was much higher compared to the median peak venom concentration of 25 ng/ml (range:2.5-2316 ng/ml) in our study [6]. Phillips et al. reported serum and urine myoglobin, which are difficult to interpret because no studies have correlated these measurements with outcomes. They detected myoglobin in the plasma of all 19 patients tested (range: 100->8000; median: 2745, normal value:<50 ng/ml) and in the urine in 14 of 18 patients (110 to >16 000; median: 4000; normal: <21 ng/ml) [10]. Although these appear to be high, much higher values are reported in Australian mulga snake envenoming, where urinary myogolobin values in three patients were 4129, 28200 and 127000 ng/ml [20]. In a larger study of 336 patients in 2000 by Kularatne et al [7] there were 47 patients (14%) with myotoxicity based on generalised muscle tenderness. This is consistent with our study in terms of clinical effects but they did not provide any biochemical confirmation. Our 245 patients are typical of current patients. They had high circulating peak venom concentrations with a median of 25ng/ml, with 24 patients having venom concentrations >1000 ng/ml [6]. Further, nearly all had local envenoming, 68% had coagulopathy (half with bleeding manifestations), and 53% had neuromuscular paralysis [6]. Therefore, the mild myotoxic effects in Sri Lankan Russell's viper envenoming are most likely due to the weak myotoxicity of the venom.
In patients with Australian Mulga snake (Pseudechis australis) envenoming, a delay in antivenom treatment has led to an increase in the severity of the myotoxicity [20]. Since there was no control group of patients who did not receive antivenom in this study, it could be argued that the weak myotoxicity observed in Sri Lankan Russell's viper envenoming is due to an effect of antivenom. However, there was no association between the delay in antivenom treatment and the number of patients having CK >300 U/l, indicating that the weak myotoxicity observed in the patients is not due to an effect of the antivenom.
Indian Polyvalent antivenom was used at concentrations equivalent to the tested venom as recommended by the manufacturer. At these concentrations antivenom failed to prevent the myotoxic effects of Sri Lankan Russell's viper venom in-vitro. Although this may be due to the low efficacy of the antivenom against the myotoxins [39], testing higher antivenom concentrations was not possible due to the practical limitations of increasing antivenom concentration in the tissue organ bath environment without affecting the osmolarity of the physiological salt solution. It was therefore not possible to determine the efficacy of the Indian polyvalent antivenom to neutralise the myotoxic effects of Sri Lankan Russell's viper venom, because such a large amount of venom was required to cause myotoxicity requiring very high concentrations of antivenom.
Of the several 's' type PLA 2 toxins isolated from Sri Lankan Russell's viper venom, VRV-PL-VIIIa [40] has 100% match with the aligned trypsin digested peptide fragments of U1-viperitoxin-Dr1b. The N-terminal sequence of the first 50 amino acids of the toxin 'P1' [31] and the N-terminal sequence of first 21 amino acids of the toxin PLA 2 4 [32] are a 100% match for the sequence of U1-viperitoxin-Dr1b. Following the suggested rational nomenclature for toxins [41], and given that the 94% sequence homology of the toxin with U1-viperitoxin-Dr1a [33], we have named the above toxin as U1-viperitoxin-Dr1b (Fig 6).
The recent study on the venom proteome of the Sri Lankan Russell's viper [11] reported that five PLA 2 toxins make up a relative abundance of 35% of the whole venom, with VRV-PL-VIIIa (U1-viperitoxin-Dr1b) making 13.9% (as opposed to 22.2% in our study) of the whole venom. However, the same study did not match a PLA 2 with VRV-PL-V (U1-viperitoxin-Dr1a), which makes 19.2% of the venom in our analysis as previously published [33].
In conclusion, myotoxicity in Sri Lankan Russell's viper envenoming is mild and non-life threatening, as evident from the low frequency of generalized myotoxicity and low concentrations of serum creatine kinase in envenomed patients. The whole venom of Sri Lankan Russell's viper has weak myotoxic properties in-vitro. Two PLA 2 toxins, U1-viperitoxin-Dr1a and U1-viperitoxin-Dr1b that make 42% of the whole venom, are the major myotoxins in the venom, but display weak myotoxicity. There is a small possibility that myotoxicity may occur in patients with severe Sri Lankan Russell's viper envenoming in which antivenom is delayed.