2.1 Ethics statement

All animal experiments were performed using protocols approved by the Animal Welfare and Ethical Review Boards of the Liverpool School of Tropical Medicine and the University of Liverpool, in accordance with the UK Animal (Scientific Procedures) Act 1986. All protocols were conducted with approval from the UK Home Office under the conditions of Project Licence P24100D38 for LD₅₀ and ED₅₀ experiments, and Project Licence P58464F90 for MND and eMND experiments.

2.2 Venoms and Antivenoms

Venoms were extracted from B. arietans, D. polylepis, H. haemachatus, N. annulifera and N. mossambica snakes maintained in the E-SRIC facility and which were, with government permission (the Big Game Parks authority) wild-caught in Eswatini. Venoms were frozen, lyophilised, stored and transported to Liverpool at 4°C. Venoms were reconstituted in PBS (pH 7.4, Gibco) and stored at -80°C just prior to use. The same batches of venoms were used throughout the study to ensure experimental consistency.

The antivenom products examined in this study are listed in Table 1. SAIMR Polyvalent antivenom (manufactured by South African Vaccine Producers PTY [SAVP], South Africa) was acquired from a pharmacy in Eswatini. Panafrican antivenom (ICP) and PANAF antivenom (PS&V) were kindly donated by the manufacturers. All three antivenoms consist of equine antibodies. Panafrican (ICP) antivenom was manufactured as intact Ig, whereas SAIMR Polyvalent and PANAF (PS&V) were manufactured as F(ab’)₂ fragments of Ig. SAIMR Polyvalent antivenom was used in the liquid format, as supplied at 10 mL per vial. The lyophilised Panafrican (ICP) and PANAF (PS&V) antivenoms were respectively reconstituted in 10 mL 0.9% saline (Severn Biotech) and 10 mL sterile medical-grade water (supplied by the manufacturer). Antivenom protein concentrations listed in Table 1 were determined by measuring the A280 nm using a NanoDrop ONE (ThermoScientific), using two dilutions of antivenom and read in duplicate. To ensure instrument accuracy, equine IgG standards were run in parallel on the NanoDrop, and a commercial BCA assay (Pierce BCA Protein Assay Kit, ThermoFisher Scientific) was used to corroborate the NanoDrop results.

2.3 Animal maintenance

Male CD1 mice weighing 18–20 g were obtained from Charles River (UK) and acclimatised for at least one week prior to experimentation. Mice were randomly allocated into cages of five upon arrival, and these cage allocations formed subsequent treatment groups—no further randomisation was applied. Mice were housed in specific-pathogen free facilities in Techniplast GM500 cages with nesting material (Sizzlenest zigzag fibres, RAJA), Lignocell bedding (JRS, Germany) and environmental enrichment materials. Mice were housed in room conditions of 22°C temperature with 40–50% humidity and 12/12 hour light cycles. Mice had ad lib access to CRM irradiated food (Special Diet Services, UK) and reverse osmosis water in an automatic water system.

2.4 Venom toxicity profiling (LD₅₀ and MND)

The amount of venom that results in 50% lethality (LD₅₀) was determined using WHO-recommended protocols [16]. Doses of venom were prepared with PBS (Gibco) in a total volume of 100 μL and stored on ice until intravenous administration via the tail vein using a 25 gauge needle. We used the WHO-recommended [16] five mice per treatment group for all injections of venoms or antivenom/venom mixtures. Mice were continuously monitored over the six-hour experiment for humane endpoints (seizure, external haemorrhage, loss of righting reflex, loss of grasping reflex, hind limb paralysis and hypothermia) at which point animals were immediately euthanised using rising concentrations of carbon dioxide or cervical dislocation. Observations were performed by unblinded, mixed gender experimenters. Data were collected on time of euthanasia and the number of mice in each dose group that died or survived the experiment. The LD₅₀ and 95% confidence intervals were determined using Probit analysis.

The minimum dose of venom required to induce a lesion of 5 mm diameter (MND) was determined using doses of venom prepared in PBS to a total volume of 50 μL and stored on ice until intradermal injection. Prior to injection animals were briefly anaesthetised with 5% isoflurane to facilitate shaving at the dose site for monitoring of lesion development. Animals were monitored three times per day for 72 hours to observe general health and to monitor lesion size. After 72 hours the animals were euthanised by rising concentrations of carbon dioxide and lesions were excised from skin and measured using callipers.

2.5 Antivenom efficacy profiling (ED₅₀ and eMND)

Antivenoms were compared for venom neutralising efficacy using ED₅₀ and eMND assays. A refined version of the WHO recommended antivenom ED₅₀ was used [3,17–20], in which 3 x or 5 x the LD₅₀ dose was pre-incubated at 37°C for 30 minutes with varying amounts of each antivenom, prepared in a total volume of 200 μL with PBS and the venom/antivenom mixtures injected intravenously. In line with WHO guidelines, our target venom challenge dose was 5 x LD₅₀s but when this proved 100% lethal when mixed with the highest possible volume of antivenom, we reduced the venom challenge dose to 3 x LD₅₀s. Due to an error, 5.3 x LD₅₀ dose was used for D. polylepis venom (see S1 File for details). The animals were monitored and data collected and analysed as described above for the LD₅₀ assays.

For eMND assays, the dose (volume) of SAIMR Polyvalent that reduced lesion size to less than 5 mm diameter was first determined, in which the venom and antivenom (in a volume of 50 μL) were preincubated for 30 minutes at 37°C. Venom/antivenom mixtures were injected intradermally and animals were monitored as described above for the MND experiments. The effective volume of SAIMR Polyvalent was determined and the same volume of the two test antivenoms was used to compare reduction in lesion size. Data was collected and analysed as above for the MND experiments.

These assays require the use of large numbers of experimental mice that experience pain, harm and duress. We have developed experimental protocols and steps to minimise the numbers of mice used in these assays and these are described in full (S1 File).

The following series of in vitro assays were conducted in an effort to explain the differential efficacies of the three antivenoms to neutralise the lethal and tissue-damaging effects of the Eswatini snake venoms.

2.6 Titration ELISA assays to compare the anti-venom immunoglobulin (Ig) titres of each antivenom

Venom was diluted in 100 mM carbonate-bicarbonate coating buffer pH 9.6 and coated at 100 ng per well in 96-well ELISA plates (Nunc MaxiSorp, ThermoScientific) and incubated overnight at 4°C. The following day plates were washed three times with TBS with 1% Tween20 (TBS-T), blocked with 5% non-fat milk in TBS-T and incubated for 2 hours at room temperature. Plates were washed again three times with TBS-T followed by the addition of primary antibodies (neat antivenom and normal horse IgG [BioRad] as negative control) in duplicate, at an initial dilution of 1 in 100 in blocking buffer, which were five-fold serially diluted across the plate and incubated at 4°C overnight. The following day plates were washed three times with TBS-T and incubated for 2 hours at room temperature with secondary antibody horseradish peroxidase-conjugated rabbit anti-horse IgG (Sigma A6917), diluted to 1 in 1,000 in PBS. Plates were washed three times with TBS-T followed by the addition of substrate (3% 2,2’-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid), Sigma) in citrate buffer pH 4.0 containing 0.1% hydrogen peroxide. Plates were incubated at room temperature for 15 minutes to develop, and the signal was measured spectrophotometrically at 405 nm on an LT-4500 plate reader (Labtech).

Relative avidity ELISAs were performed as described for the end-point ELISA assays, with some differences; primary antibodies were added at a single dilution (1 in 10,000), and after washing the primary antibody with TBS-T, the chaotropic agent ammonium thiocyanate was added at a range of concentrations (0–8 M) and incubated for 15 minutes at room temperature. The plates were washed with TBS-T and then continued to follow the same steps as described for the end-point ELISA. The reduction percentages were then calculated; data are shown in S1 Fig.

2.7 SDS-PAGE and immunoblotting to compare the venom protein-binding specificities of each antivenom

For SDS-PAGE, 5 μg venom in PBS was incubated at 100°C for 5 minutes in equal volume 2 X denaturing protein loading buffer (62.5 mM Tris-Cl pH 6.8, 25% v/v glycerol, 2% SDS, 0.75% bromophenol blue) with 30% beta-mercaptoethanol, then loaded into a 4–20% Mini-Protean TGX gel (BioRad). After electrophoresis at 200V the gels were stained with Coomassie blue (50% methanol, 40% water, 10% glacial acetic acid and 0.1% Coomassie Brilliant Blue) overnight at room temperature. Gels were de-stained 2 hours at room temperature in 50% methanol, 40% deionized water, 10% glacial acetic acid and imaged on a GelDoc (Bio-Rad).

For immunoblotting, SDS-PAGE gel electrophoresis of venoms was performed as described above but using hand-cast 15% SDS-PAGE gels. Following electrophoresis, proteins in the gels were transferred to 0.2 μm nitrocellulose membranes using the TransBlot TURBO system (BioRad). Membranes were blocked in 5% non-fat milk in TBS-T on an orbital shaker overnight at 4°C. The following day immunoblots were washed three times in TBS-T for 5 minutes per wash, then incubated with neat antivenom or normal horse IgG (BioRad) diluted 1:5,000 in blocking solution for 2 hours at room temperature. Immunoblots were washed in triplicate with TBS-T as described above, then incubated for two hours at room temperature with horseradish peroxidase-conjugated rabbit anti-horse IgG (Sigma A6917) diluted 1:1800 in PBS. Immunoblots were washed again with TBS-T and developed by the addition of DAB substrate (0.05% w/v DAB with 0.024% hydrogen peroxide) by placing the membrane into the substrate for 40 seconds, before washing with deionised water and immediately photographed.

Semi-quantitative fluorescent immunoblotting was used to compare the binding of antivenoms to venoms. To allow for comparisons using normalisation to total protein content, the appropriate amount of venom to load per lane was determined using serial dilutions to identify amounts of venom that fall within the linear range, and 5 μg was determined to be within the linear range for all 5 venoms. Subsequently, 5 μg venom in 2 X denaturing protein loading buffer was incubated at 100°C for 5 minutes, then loaded into an Any kDa Mini-Protean TGX gel (BioRad). After electrophoresis the gels were transferred to nitrocellulose membranes using the TransBlot TURBO system (BioRad). Membranes were stained and washed with Revert Total Protein Stain 700 (LI-COR Biosciences) as per manufacturers recommendations, then imaged for 2 minutes in the 700 nm channel on the Licor Odyssey Fc (LI-COR Biosciences). Signal values for each lane were exported and normalised for each venom. Membranes were blocked in 5% non-fat milk in TBS-0.1% Tween20 on an orbital shaker for 2 hours at room temperature, then incubated with 1 in 5,000 dilution of antivenom or 1 in 50 dilution of normal horse IgG (diluted in blocking solution) overnight at 4°C. The following day membranes were washed in TBS-0.1% Tween20 for 3 x 5-minute washes, then probed with rabbit anti-horse IgG (H&L) DyLight 800 (Rockland Immunochemicals) diluted at 1 in 15,000 in 5% non-fat milk in TBS-0.1% Tween20, and incubated at room temperature for 2 hours. Membranes were washed as above and finally with 1 x 5-minute wash in TBS prior to scanning on the Odyssey Fc in the 700 nm and 800 nm channels for 2 minutes each. The 800 nm signal for each lane was exported and normalised to total protein content using the normalisation values obtained from the 700 nm signal as described above. All images were obtained using Image Studio software (Version 5.2, LI-COR Biosciences).

2.8 Bovine plasma clotting assay

Two-fold serial dilutions of each antivenom were made in PBS in a volume of 10 μL per well in a 384-well plate. All conditions were tested in triplicate repeats within the plate. Venom (1 mg/mL) was added at 0.5 μL per well into the respective wells, giving a final venom concentration of 500 ng/well. Plates were incubated at 37°C for 30 minutes then cooled to room temperature. Following incubation, a Multidrop Combi reagent dispenser (ThermoFisher) was used to add 20 μL of 20 mM CaCl₂ to each well, followed immediately by 20 μL of citrated bovine plasma (BioWest, VWR). Plates were immediately read on a FLUOstar Omega plate reader (BMG Labtech) at 595 nm absorbance for 113 minutes. The assay was performed independently in technical triplicate and outliers within one run were manually identified and excluded from analyses. The area under the curve (AUC) for each venom was calculated for each replicate and a one-way ANOVA of AUC was performed to identify significant differences from the PBS control indicative of abnormal clotting, using Prism 9 (GraphPad Software). Neutralisation of venom-induced alterations to plasma clotting were determined by comparing the AUC of the test conditions (venom and antivenom) with that of the PBS ‘normal clotting’ control, and a non-significant difference in the AUC indicated restoration of clotting to normal values. Differences in the AUC between SAIMR Polyvalent and the two test antivenoms, at equivalent volume of antivenom, were identified using a one-way ANOVA with multiple comparisons using Prism 9.

2.9 Snake venom metalloproteinase (SVMP) assay

SVMP activity was measured using the previously described fluorogenic peptide assay [18]. Briefly, 1 μg of venom (in a 1 μL volume) or equal volume of PBS was added to each well in a clear, polystyrene 384-well plate (Greiner Bio-One), followed by 10 μL of antivenom (at dilutions of 1 in 4, 1 in 8 and 1 in 16) or equal volume of PBS. Venom only, antivenom only and PBS only controls were included. The SVMP substrate Mca-KPLGL-Dpa-AR-NH2 fluorogenic peptide (ES010, BioTechne) was suspended in reaction buffer (150 mM NaCl, 50 mM Tris-HCl pH 7.5) to 6.2 mM stock. The assay plate was incubated at 37°C for 25 minutes and then cooled to room temperature for 5 minutes before the addition of 90 μL SVMP substrate to each well (7 μM final well concentration). The plate was immediately read at excitation 320–10 nm and emission 420–10 nm with automatic gain for 103 minutes on a CLARIOstar plate reader (BMG Labtech). All conditions were performed in quadruplicate within the plate. The RFUs at 60 minutes were collected and analysed. SVMP activity was calculated for each venom, in which ‘no antivenom’ wells represented 100% activity and the change in SVMP activity in the presence of the test antivenoms was calculated as a percentage of the ‘no antivenom’ wells. Differences in the RFU between SAIMR Polyvalent and the two test antivenoms, at equivalent volumes of antivenom, were identified using a one-way ANOVA with multiple comparisons using Prism 9.

2.10 Phospholipase A₂ (PLA₂) activity assay

PLA₂ activity was measured using an EnzCheck Phospholipase A2 assay kit (E10217, ThermoFisher Scientific), as previously described [21]. Bee venom standards were run alongside according to manufacturer’s instructions to determine specific PLA₂ activity. Prior to analyses of the inhibitory activity of antivenoms, dilution curves of venoms only and antivenoms only were performed to (i) identify the concentrations of venom for each snake species that fall within the linear range of the activity curve, and (ii) identify whether the addition of antivenom causes assay interference. For data analysis, buffer only well values were subtracted from all other values. From these results (shown in S2 Fig), the optimal venom amounts were determined as 2 μg for B. arietans, D. polylepis and N. annulifera, 0.1 μg for H. haemachatus and 0.05 μg for N. mossambica, and antivenom volumes of less than 1.25 μL per well did not interfere with the assay readout. Using these venom amounts, the specific PLA₂ activity was calculated from a standard curve using a bee venom positive control as per the manufacturers recommendations.

Two-fold dilutions of antivenom (diluted in PBS in a 12.5 μL volume) were incubated with the pre-defined amount of venom (in 2.5 μL volume) at 37°C for 30 minutes, cooled to room temperature and added to a clear, polystyrene 384-well plate (Greiner BioOne). 12.5 μL PLA₂ substrate (reconstituted as per manufacturers instructions) was added to each well, the plate was then incubated in the dark at room temperature for 10 minutes and then read at excitation 485 nm and emission 520 nm with gain set to 50% of the highest amount of bee venom control, using a FLUOstar plate reader (BMG Labtech). RFU measurements were converted to percentage of PLA₂ activity, where the venom only (no antivenom) control was 100% activity, and the data from venom plus test antivenom wells were expressed as a percentage of 100% activity. Neutralisation by varying doses of antivenom were compared to the ‘no antivenom’ control using a one-way ANOVA, and comparisons between SAIMR Polyvalent and the two test antivenoms, at equivalent volumes of antivenom, were identified using a one-way ANOVA with multiple comparisons.

3.1 In vivo neutralisation of venom bioactivity

3.1.1 Venom toxicity profiling.

The amount of venom that kills 50% of mice when injected intravenously was determined for each venom, as shown in Table 2. The five venoms tested demonstrated a wide range in lethal dose. Venoms from D. polylepis and N. mossambica demonstrated the greatest lethal toxicity whereas H. haemachatus, N. annulifera and B. arietans showed weaker toxicity in the intravenous murine model.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Lethal and necrotic toxicity from Eswatini snake venoms in a murine model.

The amount of venom that caused lethality in 50% of mice (LD₅₀) following intravenous injection was determined by Probit analyses, and is reported as μg per mouse, with 95% confidence intervals in parentheses. The minimum amount of venom that caused a lesion approximately 5x5 mm in diameter (MND) following intradermal injection is reported as μg per mouse, with mean lesion area in parentheses. Due to signs of systemic envenoming at higher doses of H. haemachatus venom injection, a smaller lesion area is reported. N.D. = not determined. All experiments used groups of 5 animals.

https://doi.org/10.1371/journal.pntd.0010496.t002

Envenoming by only three (B. arietans, H. haemachatus and N. mossambica) of the five Eswatini venoms causes tissue damage (painful, progressive swelling sometimes leading to necrosis) in Eswatini snakebite victims. We therefore only tested these three venoms in the MND assays and the amount of venom causing a 5x5 mm tissue-destructive lesion in the injected mice is illustrated in Table 2. Due to overwhelming neurotoxicity induced by H. haemachatus venom at higher doses, the maximum lesion size obtained was lower than a typical MND (5mm x 5 mm). Prior preclinical research reported that 60 μg of Eswatini H. haemachatus venom did not result in a dermonecrotic lesion in venom-injected mice—we are unable to account for this difference to our results [22].

3.1.2 Antivenom efficacy profiling—prevention of lethal venom effects.

To compare the efficacy of the antivenoms at neutralising the lethal effects of each venom, we used the WHO-recommended antivenom effective dose (ED₅₀) assays, in which the amount (volume) of antivenom required to prevent venom-induced lethality in 50% of animals is determined, typically using venom concentrations of 3–5 x LD₅₀.

As shown in Table 3 and Fig 1, B. arietans venom was most effectively neutralised by SAIMR Polyvalent (6.4 μL), whilst Panafrican (ICP) required a three-fold higher dose (19.4 μL), and a drastically larger volume (47.5 μL) of PANAF (PS&V) was required to prevent lethality in 50% of the group. Similarly, the venom of H. haemachatus was most effectively neutralised by SAIMR Polyvalent (33.5 μL) and Panafrican (ICP) required an almost four-fold higher dose (122.5 μL), whereas we had to use a lower challenge venom dose (3 x LD₅₀) to determine the ED₅₀ for PANAF (PS&V), due to reaching the maximum volume limit of antivenom in the assay which this maximal volume did not confer 50% survival. Even at a reduced amount of venom, PANAF (PS&V) antivenom was markedly less efficacious than SAIMR Polyvalent (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. ED₅₀ of the three antivenoms against each of the five Eswatini venoms, expressed as μL per mg of venom.

SAIMR Polyvalent (SAVP) shown in teal, Panafrican (ICP) shown in magenta, PANAF Premium (PS&V) shown in blue. ED₅₀ was determined using Probit analysis, data represents the calculated ED₅₀ and error bars represent 95% confidence intervals. The units used (μL per mg of venom) attempts to normalise reporting to the amount of venom used to compensate for differences in number of venom LD₅₀s used, however for clarity, a 3 x LD₅₀ venom dose was used for the following groups; H. haemachatus with PANAF (PS&V) antivenom, N. annulifera with Panafrican (ICP) antivenom, and N. annulifera with PANAF (PS&V) antivenom. All other experiments used 5 x LD₅₀ venom dose.

https://doi.org/10.1371/journal.pntd.0010496.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. The preclinical venom-neutralising efficacy (antivenom ED₅₀) of the PANAF (PS&V) and Panafrican (ICP) antivenoms compared to SAIMR Polyvalent in the pre-incubation model of envenoming.

ED₅₀ is defined as the volume of antivenom which protects 50% of mice from the lethal effects of venom. Each experiment used five mice per dose group. The assays utilised a venom challenge dose of 5 x venom LD₅₀s (Table 2), except when it was necessary to reduce the challenge dose 3 x venom LD₅₀s (indicated by red text). In these instances, the maximum volume limit of antivenom that can be injected intravenously was reached without reducing the venom lethality of the 5 x LD₅₀ venom dose, therefore it was necessary to reduce the venom dose to determine an ED₅₀ for the antivenom. Results are reported as ED₅₀ determined by Probit analyses and expressed in (i) volume of antivenom and, in italics, (ii) as μL of antivenom per mg of venom. 95% confidence intervals are reported in parentheses.

https://doi.org/10.1371/journal.pntd.0010496.t003

Both test antivenoms were able to neutralise the lethal effects of D. polylepis venom, although at higher volumes than SAIMR Polyvalent, whilst for N. annulifera we were required to lower the venom concentration to 3 x LD₅₀ as both test antivenoms failed to confer 50% survival when tested at the maximum volume of antivenom with 5 x LD₅₀. Even at these lower venom amounts, the volume of both test antivenoms to confer 50% survival was higher than the ED₅₀ volume of SAIMR Polyvalent against 5 x venom LD₅₀ doses. In contrast, both test antivenoms performed equally as well or slightly better than SAIMR Polyvalent when tested against N. mossambica venom, with PANAF (PS&V) requiring the least antivenom.

As summarised in Fig 1, SAIMR Polyvalent was overall the most potent antivenom against the venoms of Eswatini snakes, with the exception of N. mossambica venom, which was more effectively neutralised by PANAF (PS&V) and Panafrican (ICP) antivenom was equally as efficacious as SAIMR Polyvalent.

3.1.3 Antivenom efficacy profiling—prevention of tissue-destructive venom effects.

We next determined the neutralising efficacy the volume of the two test antivenoms one MND dose of B. arietans, H. haemachatus and N. mossambica venoms (Table 4). We first determined the volume of SAIMR Polyvalent to reduce the lesion size to less than 5 mm diameter. At the same dose of antivenom, against Panafrican (ICP) was more effective than SAIMR Polyvalent at reducing lesion size in B. arietans and N. mossambica venoms, whereas PANAF (PS&V) outperformed SAIMR Polyvalent against N. mossambica venom but was less effective against B. arietans venom. All three antivenoms completely prevented lesion development induced by H. haemachatus venom.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. The resulting in mean lesion size after intradermal injection with one MND dose of venom preincubated with each of the antivenoms.

The dose of SAIMR Polyvalent in μL was first determined, and the test antivenoms at this fixed volume were then tested to compare lesion size.

https://doi.org/10.1371/journal.pntd.0010496.t004

3.2 In vitro analyses of the interactions of antivenoms with venoms and specific venom proteins

We undertook a series of in vitro assays to better understand the reasons explaining the distinct efficacies of the different antivenoms to neutralise lethal and tissue destructive effects of the various venoms. The immunological analyses examined the immunoglobulin titre and venom protein-specificity of each antivenom to each venom.

3.2.1 Immunological analysis of antivenoms.

3.2.1.1 ELISA analysis to determine the immunoglobulin titre of each antivenom to each venom. Serial dilutions of each antivenom or normal horse IgG control were incubated with each of the five venoms in ELISA assays to compare the antivenom Ig titre to the Eswatini venoms, as shown in Fig 2. To represent the clinical efficacy of antivenoms, where neutralising ability is reported as volume of antivenom, we did not standardise antivenom concentrations and instead compared equivalent volumes of antivenom. The OD readings (venom-binding) of the antivenoms at the 1 in 62,500 dilution (indicated by vertical line in Fig 2 panels), were used to compare the titres of each antivenom.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The titre of three antivenoms against five Eswatini venoms determined by end-point titration ELISA.

SAIMR Polyvalent (SAVP) shown in teal, Panafrican (ICP) shown in magenta, PANAF (PS&V) shown in blue, normal horse IgG shown in purple. Panel A: B. arietans. Panel B: D. polylepis. Panel C: H. haemachatus. Panel D: N. annulifera. Panel E: N. mossambica. Data points represent the mean of two replicates and error bars show the standard deviation. The vertical line at the 1 in 62,500 dilution represents the point at which Ig titres of each antivenom were compared.

https://doi.org/10.1371/journal.pntd.0010496.g002

The highest overall antivenom Ig titres were against B. arietans venom, closely followed by H. haemachatus venom. At the 1 in 62,500 dilution ‘antivenom-comparison’ point, PANAF (PS&V) showed a notably lower titre for B. arietans venom (OD₄₀₅ = 1.33) than SAIMR Polyvalent (OD₄₀₅ = 2.32) and Panafrican (ICP) (OD₄₀₅ = 2.23) antivenoms. Narrower differences in Ig titre were observed for H. haemachatus venom, although again PANAF (PS&V) showed a lower titre (OD₄₀₅ = 1.16) than the SAIMR Polyvalent (OD₄₀₅ = 1.75) and Panafrican (ICP) (OD₄₀₅ = 1.55) antivenoms. Immunoglobulin titres were markedly lower for the D. polylepis, N. annulifera and N. mossambica venoms, with a highest OD₄₀₅ reading of 1.03 at the 1 in 62,500 dilution (Panafrican [ICP] with N. mossambica venom) (Fig 2). The three antivenoms exhibited near-equivalent Ig titres for the D. polylepis and N. annulifera venoms—differences in OD₄₀₅ <0.3 for all antivenoms at the 1 in 62,500 dilution. The Panafrican (ICP) antivenom exhibited the highest Ig titre (OD₄₀₅ = 1.03) for N. mossambica venom—almost double the titres of the PANAF (PS&V) (OD₄₀₅ = 0.57) and SAIMR Polyvalent (OD₄₀₅ = 0.53) antivenoms. It must be noted that despite PANAF (PS&V) displaying the lowest Ig titres against four out of five venoms, this antivenom product contains substantially lower protein content (51.9 mg/mL, Table 1), which must be taken into consideration. Similarly, the antivenom with the highest protein concentration (113.7 mg/mL) SAIMR Polyvalent showed the highest Ig titres against four of the five venoms.

3.2.1.2 Immunoblot analysis to determine the venom protein-specificity of each antivenom to each venom. We used immunoblot analysis to assess whether Ig titre differences between the three antivenoms were also detectable in their ability to bind distinct proteins within each venom (Fig 3). As in the ELISA analyses and for the same reason, we again used the same dilution of the antivenoms in each of the immunoblots (i.e. they were standardised by volume not antibody content).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Protein profiles of the venom protein components of five Eswatini snake venoms.

A: Venom proteins separated by SDS-PAGE under reducing conditions and stained with Coomassie blue to show all proteinaceous components. B: Immunoblot using normal horse IgG used as primary antibody negative control. C: Immunoblot using SAIMR Polyvalent as primary antibody. D: Immunoblot using Panafrican (ICP) antivenom as primary antibody. E: Immunoblot using PANAF (PS&V) antivenom as primary antibody.

https://doi.org/10.1371/journal.pntd.0010496.g003

All three antivenoms showed recognition of the full molecular weight range of venom proteins (depicted in Fig 3, panel A—Coomassie blue stained protein content of each venom). The PANAF (PS&V) antivenom showed an overall lower protein-binding intensity, however, and as per the ELISA results, this could be attributed to the lower protein concentration of this antivenom. The most notable differences in toxin recognition by the three antivenoms was observed for B. arietans venom, with SAIMR Polyvalent showing stronger recognition of the higher molecular weight (MW) proteins (>25 kDa) whereas Panafrican (ICP) and PANAF (PS&V) recognised fewer of the higher MW proteins. For the other four venoms, the toxin recognition profiles were similar across the three antivenoms, with no obvious differences in binding to specific toxins.

The immunoblot analyses, using a chromogenic assay failed to identify venom protein-binding distinctions that might account for the differential venom-neutralising efficacies of venom-binding Ig titres of the three antivenoms. We therefore repeated this assay using fluorescent immunoblots, which outputs quantifiable protein-binding intensities (S3 Fig). We were limited to performing these as semi-quantitative, rather than quantitative, measurements of protein binding, due to the varying formats of the antivenom Igs (F[ab’]₂ and intact IgG), which may affect recognition by the secondary antibody thus limiting the comparisons. Aside from a more-pronounced binding of the Panafrican (ICP) to the higher molecular weight components of H. haemachatus venom, the results using fluorescent immunoblotting did not identify any additional venom protein-binding specificities.

Altogether the serological analysis of antivenom binding to venom toxins showed similar protein binding characteristics and, with the exception of B. arietans venom, these data did not explain the distinct in vivo efficacies of the three antivenoms. Whilst the data obtained with these immunological assays are informative, they do not differentiate between recognition of toxin and non-toxin venom components and between binding and neutralising antibodies. We therefore sought to further characterise the neutralising ability of the antivenoms using in vitro phenotypic and toxin-specific enzymatic assays.

3.2.2 In vitro neutralisation of venom bioactivity

3.2.2.1 Plasma clotting activity. Severe anticoagulant disturbances to the normal clotting of bovine plasma were induced by H. haemachatus and N. mossambica venoms (Fig 4A), while venoms of B. arietans, D. polylepis and N. annulifera showed no significant alterations to plasma clotting (p > 0.05 compared to PBS control). SAIMR Polyvalent failed to neutralise the effects of N. mossambica venom on plasma clotting at the antivenom volumes tested, whilst Panafrican (ICP) and PANAF (PS&V) both restored plasma clotting to ‘normal’ levels (Fig 4C) (indicated by no significant differences in the AUC for the test wells compared to the normal clotting PBS control). The differences in restoration of plasma clotting between SAIMR Polyvalent and the test antivenoms were compared for the highest volume of antivenom (2.5 μL). The restoration of plasma clotting levels in the presence of Panafrican (ICP) and PANAF (PS&V) were both significantly different to SAIMR Polyvalent (p = 0.02 and 0.004, respectively). All three antivenoms effectively neutralised the anticoagulant effects of H. haemachatus venom to restore plasma clotting to ‘normal’ levels at the three doses tested (Fig 4B), with no significant differences between the test wells and PBS control, and no significant differences were measured between the two test antivenoms compared to SAIMR Polyvalent.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Plasma clotting activity of Eswatini venoms and their neutralisation by three antivenoms.

Data show the mean area under the curve of nine replicates and error bars represent standard deviation. **** indicates p < 0.0001, ** indicates p < 0.01 and * indicates p < 0.05 by one-way ANOVA. A: Plasma clotting activity of B. arietans, D. polylepis, H. haemachatus, N. annulifera and N. mossambica in comparison to PBS normal clotting control (white bar). B: Effects of antivenoms on neutralising the anticoagulant effects of H. haemachatus venom. C: Effects of antivenoms on neutralising the anticoagulant effects of N. mossambica venom. Normal clotting is shown by ‘PBS control’ bars in white (wells contain no venom), and venom-induced clotting deficiency shown by ‘no antivenom control’ bars in black.

https://doi.org/10.1371/journal.pntd.0010496.g004

Although the three antivenoms showed comparable efficacy at neutralising H. haemachatus venom-induced anticoagulant effects in vitro, SAIMR Polyvalent significantly outperformed the other two antivenoms in the in vivo ED₅₀ tests of venom lethality. The data presented here suggested that these in vivo differences may not be a direct consequence of clotting abnormalities.

3.2.2.2 Snake venom metalloproteinase (SVMP) activity. The SVMP toxins possess potently disruptive activities against diverse mammalian cardiovascular targets [23] and we therefore compared (i) the extent to which the Eswatini venoms expressed this bioactivity in vitro and (ii) the SVMP-neutralising ability of the antivenoms. Strong SVMP activity was detected for B. arietans venom as shown in Fig 5A, whilst N. annulifera venom demonstrated weak but detectable SVMP activity (statistically greater than PBS control, p = 0.0004) and the remaining three venoms exhibited no significant activity (p > 0.2 for all).

All three antivenoms significantly reduced the SVMP activity of B. arietans in a dose-dependent manner (p < 0.0001 for all conditions compared to no antivenom 100% activity control), with SAIMR Polyvalent demonstrating the greatest reduction in toxin activity at the highest dose (2.5 μL), as shown in Fig 5B. The neutralisation of SVMP activity by SAIMR Polyvalent at 2.5 μL was significantly greater than that of Panafrican (ICP) and PANAF (PS&V) (p = 0.0004 and p < 0.0001, respectively). The abundance of SVMP activity detected in B. arietans is at least partially responsible for the venom pathology observed in vivo, and as demonstrated here, the ranking of antivenom neutralising ability correlates well between the in vitro and in vivo results, with SAIMR Polyvalent performing best in vitro and in vivo, and PANAF (PS&V) performing the weakest.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. SVMP activity of Eswatini venoms and their neutralisation by three antivenoms.

*** indicates p < 0.001, **** indicates p < 0.0001 by one-way ANOVA. A: SVMP activity of B. arietans, D. polylepis, H. haemachatus, N. annulifera and N. mossambica in comparison to PBS control. Data show the mean of four replicates and error bars represent standard deviation. Neutralisation of (B) B. arietans and (C) N. annulifera SVMP activity by the three antivenoms SAIMR Polyvalent (SAVP), Panafrican (ICP) and PANAF (PS&V) at different doses, in comparison to a no antivenom control (venom only) showing 100% activity. Data show the mean of four replicates and error bars represent standard deviation.

https://doi.org/10.1371/journal.pntd.0010496.g005

All three antivenoms showed poor neutralising ability against N. annulifera SVMP activity (Fig 5C) and were unable to neutralise more than half of the activity. This was unexpected because SVMP activity of N. annulifera venom was considerably weaker than B. arietans venom, which was effectively neutralised by the antivenoms. As shown in Fig 5C, SAIMR Polyvalent outperformed the comparator antivenoms at all doses tested, yet only reached approximately 50% reduction in SVMP activity at the top dose tested. At the highest dose tested (2.5 μL) SAIMR Polyvalent showed significantly greater reduction in N. annulifera SVMP activity than Panafrican (ICP) and PANAF (PS&V) (p < 0.0001 against both).

3.2.2.3 Venom Phospholipase A2 activity. The PLA₂s are a diverse and complex group of venom proteins that target a variety of mammalian cardiovascular and neurological targets [24], and we therefore compared (i) the extent to which the Eswatini venoms expressed this bioactivity in vitro and (ii) the PLA₂-neutralising ability of the antivenoms. Strong enzymatic PLA₂ activity was detected in the venoms of H. haemachatus and N. mossambica as shown in Fig 6A, whilst B. arietans, D. polylepis and N. annulifera demonstrated negligible PLA₂ activity at the venom doses tested.

All three antivenoms showed potent inhibition of both H. haemachatus and N. mossambica PLA₂ activity, with significant neutralisation of PLA₂ at antivenom doses as low as 0.06 μL (p < 0.0001 for each antivenom, compared to no antivenom). SAIMR Polyvalent was overall the most effective antivenom at neutralising H. haemachatus PLA₂ activity and was able to fully neutralise activity at the 0.06 μL volume, whereas the two other antivenoms showed lower neutralising ability at 0.13 μL and 0.06 μL doses (Fig 6B). At 0.25 μL of antivenom, PANAF (PS&V) showed no significant difference in neutralising efficacy compared to the neutralisation observed for SAIMR Polyvalent (p > 0.05), whilst Panafrican (ICP) showed significantly less neutralising ability compared to SAIMR Polyvalent (p = 0.037). At the lowest volume of antivenom tested (0.06 μL) SAIMR Polyvalent was marginally less effective than the other antivenoms in neutralising N. mossambica PLA₂ activity (Fig 6C), however at the dose of 0.25 μL, all three antivenoms fully neutralised the toxin activity with no significant differences observed between the antivenoms.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. PLA₂ activity of five venoms from Eswatini and their neutralisation by different antivenoms.

** indicates p < 0.01 by one-way ANOVA. A: PLA₂ activity of B. arietans, D. polylepis, H. haemachatus, N. annulifera and N. mossambica. Data show the mean of three replicates and error bars represent standard deviation. Neutralisation of (B) H. haemachatus and (C) N. annulifera PLA₂ activity by the three antivenoms SAIMR Polyvalent (SAVP), Panafrican (ICP) and PANAF (PS&V) at different doses, expressed as a percentage of a no antivenom control showing 100% activity. Data show the mean of three replicates and error bars represent standard deviation.

https://doi.org/10.1371/journal.pntd.0010496.g006

The ability of the antivenoms to neutralise PLA₂ activity correlates well with the in vivo ED₅₀ data. As we previously demonstrated in Fig 4 that all the antivenoms performed equally well at restoring plasma clotting of H. haemachatus, the differences in the in vivo antivenom ED₅₀s may therefore be mediated by PLA₂ activity, and these toxins are known to potentiate the pathological effects of other toxins found in spitting elapid venoms [25].