key: cord-0276173-pwtfmqbl
authors: Hall, Steven R.; Rasmussen, Sean A.; Crittenden, Edouard; Dawson, Charlotte A.; Bartlett, Keirah E.; Westhorpe, Adam P.; Albulescu, Laura-Oana; Kool, Jeroen; María Gutiérrez, José; Casewell, Nicholas R.
title: Repurposed drugs and their combinations prevent morbidity-inducing dermonecrosis caused by diverse cytotoxic snake venoms
date: 2022-05-20
journal: bioRxiv
DOI: 10.1101/2022.05.20.492855
sha: 8041ae1b386b9ff1a44ceeda9da603813de00127
doc_id: 276173
cord_uid: pwtfmqbl

Morbidity from snakebite envenoming affects approximately 400,000 people annually. Tissue damage at the bite-site often leaves victims with catastrophic life-long injuries and is largely untreatable by currently available antivenoms. Repurposing small molecule drugs that inhibit specific snake venom toxins offers a potential new treatment strategy for tackling this neglected tropical disease. Using human skin cell assays as an initial model for snakebite-induced dermonecrosis, we show that the repurposed drugs 2,3-dimercapto-1-propanesulfonic acid (DMPS), marimastat, and varespladib, alone or in combination, reduce the cytotoxic potency of a broad range of medically important snake venoms up to 5.7-fold. Thereafter, using a preclinical murine model of dermonecrosis, we demonstrate that the dual therapeutic combinations of DMPS or marimastat with varespladib potently inhibit the dermonecrotic activity of three geographically distinct and medically important snake venoms. These findings strongly support the future translation of repurposed drug combinations as broad-spectrum therapeutics for preventing morbidity caused by snakebite.

Current estimates suggest that 1.8-2.7 million people are envenomed due to snakebite every year, resulting in 81,000-138,000 deaths and 400,000 cases of morbidity annually, predominantly affecting those in the tropics and sub-tropics 1-3 . Snakebite has been labelled 'the most neglected of neglected tropical diseases (NTDs)' 4 , with the late UN Secretary General Kofi Annan calling it 'the biggest public health crisis you've never heard of' 5 . In 2017, snakebite envenoming was added to the World Health Organization (WHO)'s formal list of NTDs; the WHO has since elevated snakebite to a 'priority category A NTD' and has created a roadmap with the goal of reducing the global burden of snakebite by one-half by the year 2030 6 . One of the proposed methods to accomplish this is to develop novel treatments for snakebite; an ambitious task considering the myriad issues associated with developing snakebite therapies, including the variability and complexity of toxins that make up different snake venoms 7, 8 .

Snake venoms are comprised of dozens of different toxins at varying concentrations, which differ both inter-and intra-specifically and induce a range of pathological and pathophysiological effects 7 . However, there are four primary toxin families that are dominant across many different venoms and thus represent attractive targets for toxin-inhibiting therapeutics: phospholipases A 2 (PLA 2 s), snake venom metalloproteinases (SVMPs), snake venom serine proteases (SVSPs), and three-finger toxins (3FTxs) 9 . The main syndromes of snakebite envenoming are generally categorised as haemotoxic (e.g. haemorrhage and coagulopathy), neurotoxic (e.g. muscle paralysis), and/or cytotoxic (e.g. local tissue necrosis) 10, 11 . Haemotoxicity is a particularly common sign of envenoming, especially following bites from viperid (family Viperidae) snakes, and is largely caused by SVMPs, SVSPs, and PLA 2 s [10] [11] [12] . Neurotoxic envenoming is more commonly caused by elapid (family Elapidae) snakes and is primarily associated with neurotoxic 3FTxs and PLA 2 s 11, 13 . While 4 haemotoxicity and neurotoxicity are primarily responsible for snakebite-induced mortality, life-altering morbidity in survivors is most frequently caused by severe local tissue damage in and around the bite-site. This pathology is caused by cytotoxic 3FTxs, SVMPs, and PLA 2 s and frequently leads to permanent disability, often requiring surgical debridement or even amputations of the affected limb or digit 14, 15 .

The only treatments currently available for snakebite envenoming are animal-derived polyclonal antibody therapies called antivenoms. These therapies have conceptually remained unchanged for over a century and are associated with a multitude of issues including high cost, requirement for a consistent cold-chain, limited cross-snake species efficacy due to venom variation, and high frequency of adverse events post-administration 1, 7, 8, [16] [17] [18] [19] . In addition, due to the large size of antivenom antibodies or their fragments (i.e. typically ~110 kDa, F(ab') 2 ; or ~150 kDa, IgG) these treatments are unable to efficiently penetrate into peripheral tissue surrounding a bite-site thus reducing their efficacy against local tissue cytotoxicity, and they need to be administered intravenously (IV) in a clinical environment by a medical professional which severely restricts their utility in rural communities where snakebite victims are often hours or even days away from appropriate facilities 1, 8, 20 . To address some of these challenges, next-generation snakebite therapies such as toxin-specific monoclonal antibodies 21, 22 and toxin-inhibiting small molecule drugs [23] [24] [25] [26] [27] [28] [29] , have received considerable attention in recent years.

Small molecule drugs (hereafter simply called drugs) offer many desirable characteristics in comparison to existing antivenoms, such as increased cross-species efficacy, tolerability, stability, and affordability 8, 26, 27, 29 . Additionally, small molecule drugs are more able to penetrate peripheral tissue than large IgG-derived antibodies, meaning they should exhibit enhanced tissue distribution dynamics to better inhibit cytotoxins at the bite site. Further, they can be formulated as oral or topical therapies which could be administered in the field much 5 more quickly after a snakebite victim is envenomed compared to an IV-administered antivenom 8, 25, 26, [29] [30] [31] [32] .

Three repurposed drugs initially developed for other conditions 26, 33, 34 have shown particular promise as potential new drug therapies for snakebite envenoming based on in vitro and rodent in vivo data: the SVMP-inhibiting metal chelator, DMPS (Unithiol) 26 , the hydroxamic acid, marimastat 27, 28, [35] [36] [37] , and the secretory PLA 2 -inhibiting drug, varespladib 23, 28, [38] [39] [40] [41] .

Additionally, it has been shown that combining marimastat with varespladib improves their pan-geographic utility, resulting in superior prevention of venom-induced lethality in mice compared with either drug alone against diverse snake venoms 27 . While these studies have demonstrated such drugs can effectively protect against snake venom-induced lethality in animal models, there is limited published evidence of their efficacy or potential utility against the pathology most likely to cause the permanent, life-changing injuries often seen in snakebite survivors: tissue necrosis caused by snake venom cytotoxins. Herein we explore the therapeutic potential of small molecules drugs against the local tissue damage stimulated by cytotoxic snakebite envenoming. Using a variety of geographically diverse snake venoms, we demonstrate that DMPS, marimastat, and varespladib individually provide protection against snake venom cytotoxins to different extents, but that drug combinations are highly effective at preventing local tissue damage in vivo, and thus represent promising leads for combatting the local dermonecrotic effects caused by snakebite envenoming.

Prior to exploring the inhibitory capability of drugs against the cytotoxic effects of snake venoms, we first defined the effect of 11 venoms sourced from distinct snake species and geographic regions on the viability of adherent human skin cells. Using 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assays 42, 43 and immortalised human epidermal keratinocytes (HaCaT 44,45 ), we generated venom dose-response curves ( Fig. 1a- (Fig. 1m) , suggesting likely 'positive cooperativity' between venoms toxins 47, 48 . Naja haje, (i) East African Naja nigricollis, (j) West African Naja nigricollis, and (k) Naja pallida. (l) IC 50 and (m) Hill slope values were calculated for each independent trial. Red-coloured data denotes viperid snakes, while blue-coloured data denotes elapid snakes. * Signifies that the value is significantly higher than all other tested venoms, and † signifies that the value is significantly higher than B. asper, C. atrox, C. rhodostoma, E. carinatus, and E. ocellatus, as determined by a one-way ANOVA comparing all values to each other followed by a Tukey's multiple comparisons test (P < 0.05, n=4). ANOVA statistics for individual statistically analysed 

Prior to investigating the inhibitory potency of toxin-inhibiting drugs in the MTT assay, we first determined the cellular 'maximum tolerated concentration (MTC)' of the repurposed drugs DMPS, marimastat, and varespladib. Thus, HaCaT cells were treated with two-fold serial dilutions of each drug until a significant reduction in cell viability was observed after 24 hours of exposure. The highest concentration of each drug that did not significantly reduce cell viability when compared to the vehicle control (labelled '0') was determined to be the MTC. Then, to ensure that cells would be treated with a sub-toxic amount of drug in the venom-inhibition experiments, one half of this dose (MTC ½ ) was selected for the venom-drug co-treatment experiments 49, 50 . The MTC ½ for DMPS, marimastat, and varespladib used in the following experiments were 625, 2.56, and 128 µM, respectively (Supplementary Fig. 1) .

Next, using a drug pre-incubation model 26, 27 followed by MTT assays in the HaCaT cells, we tested the inhibitory effect of the three toxin-inhibiting drugs (using their MTC ½ values)

against six of our previously tested cytotoxic snake venoms. Our results demonstrated that the SVMP inhibitors DMPS and marimastat 26, 27, 35 significantly (P < 0.05) reduced the cell-9 damaging potency of venom from C. atrox, E. carinatus, and E. ocellatus ( Fig. 2b-d, respectively), as demonstrated by the increased IC 50 values. Additionally, DMPS slightly, albeit significantly, increased the IC 50 of East African N. nigricollis venom (P = 0.0053) (Fig.   2e) , though its effect was not significant against West African N. nigricollis venom (P = 0.0501) (Fig. 2f) . In contrast, the PLA 2 -inhibitor varespladib 23 did not display an inhibitory effect on any of the six tested venoms. The cell viability-inhibitory effects of B. arietans and West-African N. nigricollis (Fig. 2a,f) venom were not significantly inhibited by any of the tested drugs. Fig. 2 , we decided to repeat these experiments using venoms from D. russelii and B. asper, which have higher PLA 2 toxin abundances proportionally than the other six tested venoms 9 , and to increase the concentration of varespladib from its MTC ½ (128 µM) to its MTC (256 µM). In addition, propidium iodide (PI) cell death assays 51, 52 were multiplexed with the MTT assays as secondary measures of the cytotoxic potencies of the venoms, in case varespladib was incompatible with the MTT assays. Despite the potential for more abundant PLA 2 toxins to contribute to cell cytotoxic effects, varespladib again showed no inhibition as measured by either MTT or PI assays against either of these viper venoms ( Fig. 3) . None of the drugs significantly inhibited D. russelii venom potency as measured with MTTs, though DMPS reduced its potency as measured with PI ( Fig. 3a-b) . Both DMPS and marimastat inhibited B. asper venom potency as measured by MTT, while only marimastat did so as measured by PI ( Fig. 3c-d) . toxins. Thus, we repeated the MTT and PI assays using B. asper venom and compared the 13 protective effects of combination treatments with those conferred by individual drug therapies. While no drug-potentiation effect was observed when varespladib was combined with DMPS ( Fig. 4a-b) , when combined with marimastat such potentiation was apparent as the potency of B. asper venom was significantly reduced compared to the marimastat-alone treatment, as measured with both MTT and PI assays ( Fig. 4c-d) . Fig. 5b) . B. asper venom caused a mean lesion area of 41.9 mm 2 which, in contrast to the cell data, was not significantly reduced by marimastat (55.1 mm 2 ) but was by varespladib (12.2 mm 2 ). Although DMPS (21.1 mm 2 ) visually appeared to reduce the mean lesion area caused by B. asper venom, this was not statistically significant (P=0.1535) (Fig. 5c) . C. atrox venom caused a mean lesion area of 19.1 mm 2 , which was significantly reduced in size by all 15 three drug treatments: DMPS (3.1 mm 2 ), marimastat (4.4 mm 2 ) and, again in contrast to the cell data, varespladib (5.8 mm 2 ) (Fig. 5d) . E. ocellatus venom caused a mean lesion area of 5.0 mm 2 . In contrast with the other two venoms, varespladib was ineffective at reducing the lesion size (7.0 mm 2 ). Both SVMP inhibitors appeared to substantially reduce E. ocellatus venom-induced lesions, with all five marimastat-treated and four of the five DMPS-treated mice displaying no lesions; however, only marimastat's effects were significant (0 mm 2 ) while those of DMPS were not, due to the single outlier value in this treatment group (1.0 mm 2 , P=0.0856) (Fig. 5e ).

Using the same in vivo methods, we then tested combination therapies consisting of the PLA 2 -inhibiting varespladib with the SVMP-inhibiting DMPS or marimastat against these same three venoms. In contrast to the single drug therapies, which displayed variable efficacies depending on the snake species and rarely completely inhibited lesion formation in individual mice, both combination therapies significantly inhibited lesion formation caused by all three venoms tested, with many individual mice displaying no lesion development at all (Fig. 5, Supplementary Fig. 3) . Thus, mean B. asper venom-induced lesions (41.9 mm 2 ) were decreased to 2.7 and 6.7 mm 2 (Fig. 5c) , C. atrox lesions (19.1 mm 2 ) to 0.3 and 0.3 mm 2 ( Fig. 5d) , and E. ocellatus lesions (5.0 mm 2 ) to 0.1 and 0.4 mm 2 (Fig. 5e ) by the DMPS-plusvarespladib (DV) and marimastat-plus-varespladib (MV) combination therapies, respectively. representing between 25-50%, 3 representing between 50-75%, and 4 being the most severe and representing >75% of the skin layer (Supplementary Fig. 4 ). An overall dermonecrosis score was then calculated from the mean of the resulting scores obtained for the various layers (Fig. 6) . Representative photomicrographs of no, partial, and heavy dermonecrosis are shown in Fig. 6a-c. The drug treatments plus venom vehicle control induced no dermonecrosis ( Fig. 6d and   Supplementary Fig. 4a) . The varespladib, DV, and MV treatments decreased B. asper venom-induced dermonecrosis in the epidermis, dermis, hypodermis, and panniculus carnosus layers, though not in the adventitia, while neither DMPS nor marimastat alone inhibited the effects of B. asper venom in any of the skin layers (Supplementary Fig. 4b) .

This collectively resulted in the varespladib, DV, and MV treatments decreasing the overall mean dermonecrosis score induced by B. asper venom from 2.57 to 0.72, 0.06, and 0.32, respectively, while DMPS and marimastat were ineffective (Fig. 6e) . All treatments decreased C. atrox venom-induced dermonecrosis in the epidermis and dermis, and all but varespladib did so in the hypodermis, though no treatment had a significant effect in the panniculus carnosus or adventitia (Supplementary Fig. 4c) . This resulted in the DMPS, marimastat, varespladib, DV, and MV treatments decreasing the overall mean dermonecrosis score induced by C. atrox venom from 2.86 to 0.69, 0.38, 1.32, 0.10, and 0.04, respectively ( Fig. 6f) . Lastly, the marimastat, DV, and MV treatments significantly decreased E. ocellatus venom-induced dermonecrosis only in the dermis while DMPS and varespladib did not; no significant results were calculated from any treatment in any other skin layer (Supplementary Fig. 4d) . While the mean overall dermonecrosis score induced by E.

ocellatus venom was not significantly decreased by any treatment, there was a trend towards inhibition with DMPS, marimastat, DV, and MV resulting in mean overall dermonecrosis scores of 0.04, 0.00, 0.02, and 0.12, respectively, versus 1.04 for the drug-vehicle control and 0.74 for the varespladib treatment (Fig. 6g) . Note that minimal necrosis was observed in the adventitia even in the absence of drug treatment, suggesting that histological scoring of necrosis in this layer is likely less informative than in other skin layers. (Fig. 1) . These findings were unexpected given the extensive variation in toxin composition among these snake species 9,56 .

While reductions in cytotoxic potency observed with D. russelii were minor, N. haje venom was significantly less cytotoxic than all other venoms, including those from the congeneric spitting cobra species N. nigricollis and N. pallida; however, this observation is in line with the distinct, predominantly neurotoxic, composition and functional activity of this nonspitting cobra venom 56, 57 . As an additional pharmacological measure, the Hill slopes of all venoms were calculated and compared (Fig. 1m) , and while no slope was significantly different from another, the magnitudes of all 11 were greater than 1.5 and thus considered 'steep' 58 , meaning a small change in venom concentration can lead to a large change in overall pathological effect. In classical pharmacology, steep Hill slope values can be explained by the activity of one bioactive factor (e.g. a toxin) agonising the activity of another, thus allowing them to reach maximum pathological efficacy over a shorter concentration range. This indicates a phenomenon known as 'positive cooperativity' 47, 48 which suggests probable pathological synergy between certain snake venom toxins, something that has been previously evidenced [59] [60] [61] .

Our skin cell assays demonstrated that the SVMP-inhibitors DMPS and marimastat may be effective anti-cytotoxic drugs as individual therapies, although their inhibitory effects were not universal across all cytotoxic snake venoms (Fig. 2) . Unexpectedly, the PLA 2 inhibitor varespladib was ineffective against any of the venoms tested, despite it displaying impressive results against systemic venom-induced toxicity previously 23, 28, [38] [39] [40] [41] . To explore whether MTT assays are simply a poor assay choice for testing PLA 2 -inhibitors against cytotoxic venoms, we multiplexed them with a secondary cytotoxicity assay using PI to measure cell membrane disruption 51, 52 . Nevertheless, varespladib remained ineffective in these assays, suggesting that much of the cytotoxicity observed in these studies is mediated by SVMP toxins rather than PLA 2 s ( Fig. 2 and 3) ; however, when we treated the cells with varespladib in combination with marimastat we observed significant reductions in the potency of B. asper venom versus the marimastat-alone treatment (Fig. 4) . These findings suggest that PLA 2 toxins may indeed, at least to some extent, contribute to cytotoxic venom effects, and that combining an SVMP-inhibitor with a PLA 2 -inhibitor may improve overall treatment efficacy.

Interestingly, this anti-cytotoxic potentiation of marimastat by varespladib was not observed with DMPS despite this drug also being a SVMP-inhibitor. This dichotomy is likely due to the fact that these drugs' mechanisms of action are different, as marimastat directly inhibits metalloproteinases by acting as a peptidomimetic and binding covalently to the Zn 2+ ion present in the active site 25, 37, [62] [63] [64] , while the inhibitory mechanism of action of DMPS is solely the result of Zn 2+ chelation 26, 62 . These mechanistic variations likely underpin the previously described differences in SVMP-inhibiting potencies of these drugs in vitro 27, 65, 66 .

For subsequent in vivo neutralisation experiments, we tested three venoms whose cytotoxic potencies were reduced by both DMPS and marimastat in the cell assays, that were sourced from three different genera of snakes displaying considerable inter-species toxin variability 9 , and that inhabit distinct localities: B. asper (Latin America), C. atrox (North America), and E. ocellatus (West Africa). Using a drug pre-incubation 26,27 model of venom dermonecrosis in mice 53 , we showed that DMPS, marimastat, and varespladib are each individually capable of inhibiting venom-induced skin lesion formation in vivo, though the efficacy of each drug was restricted to only certain venoms (Fig. 5) . Thus, in line with the cell cytotoxicity findings, DMPS significantly reduced dermonecrotic lesions induced by C. atrox venom, while marimastat was effective against both C. atrox and E. ocellatus. However, contrasting with our cell data, at the therapeutic doses tested marimastat did not significantly reduce lesions caused by B. asper venom, nor did DMPS against B. asper or E. ocellatus venom.

Further, varespladib was effective at significantly inhibiting B. asper and C. atrox venom in vivo, though not E. ocellatus venom, likely due to this venom being dominated by SVMPs 9 .

In summary these findings clearly evidence that, while they are undoubtedly informative and could be valuable screening tests, cell-based cytotoxicity assays do not fully recapitulate findings obtained through in vivo dermonecrosis experiments, and thus additional models are required downstream to robustly assess drug efficacy against this aspect of local snakebite envenoming.

While the results of the single drug trials against select cytotoxic snake venoms in vivo are certainly promising, they also evidence how single drugs are not paraspecifically effective. In contrast, the two combination therapies tested (MV and DV) both effectively and substantially reduced the development of macroscopic dermal lesions caused by all three tested venoms despite major differences in their toxin compositions 9 , with many of the drug combination-treated mice displaying no dermal lesion development at all (Fig. 5,  Supplementary Fig. 3) . This could be explained by the variable roles that SVMPs and PLA 2 s play in the pathogenesis of skin damage induced by different snake venoms.

Histopathological analysis of the resulting lesions confirmed the efficacy of the drug combinations with both DV and MV significantly reducing the severity of overall dermonecrosis observed in the skin layers of mice envenomed with B. asper and C. atrox venoms, albeit the results in mice injected with E. ocellatus venom were not significant due to the dose being too low to consistently induce sufficient dermonecrosis in the drug-vehicle control group (Fig. 6) . These data strongly evidence the utility of small molecule drug combination therapies targeting SVMP and PLA 2 toxins as largely preclinically efficacious and pan-species effective therapies for the treatment of the local skin-damaging effects of viperid snakebite envenoming.

When combined with the results of Albulescu, et al. 27 Such an approach must, however, ensure that a robust access plan for LMIC communities is developed in parallel to avoid potential future distribution pitfalls, like those recently reported around the inequitable distribution of COVID vaccines 69, 70 .

There remains much work to be done to translate these drugs and their combinations into approved snakebite therapies. This includes additional preclinical research, for example against the venoms of additional snake species (e.g. other viperids and cytotoxic Naja spitting cobras 56, 71, 72 ) , assessment in 'challenge-then-treat' models of envenoming 26 , trials testing different routes of therapeutic administration 26 , and experiments to better understand their pharmacokinetics and pharmacodynamics to elucidate informed dosing regimens and potential drug-drug interactions. Since a major anticipated benefit of drug therapies for snakebite is their potential to be orally or topically formulated 8 (i.e. in contrast with intravenously-injected antivenom), considerable research effort should focus around this space to pursue the translation of safe, affordable, community-level interventions to reduce existing treatment delays in rural tropical communities, thus improving patient outcomes. To that end it is worth noting that DMPS is already undergoing Phase I clinical trials to determine both its safety and a PK-informed oral dosing regimen for snakebite indication 73 , while methyl varespladib has recently entered Phase II trials to assess its safety, tolerability, and efficacy in snakebite victims (https://clinicaltrials.gov/ct2/show/NCT04996264). These studies emphasise the growing confidence the research community has in specific small molecule drugs as novel oral treatments for snakebite envenoming, though the data presented here highlight that additional research to develop these (among other) drugs into combination therapies is likely to yield treatments with superior pan-snake species effectiveness than any single drug alone.

In conclusion, our data provide strong evidence that the small molecule drugs DMPS, marimastat, and varespladib can significantly protect against dermonecrosis associated with local snakebite envenoming, though their efficacy is limited to certain snake species. This limitation is largely overcome when the SVMP-inhibiting drugs DMPS or marimastat are used in combination with the PLA 2 -inhibiting drug varespladib, most likely due to the dual 26 role of SVMPs and PLA 2 s in the pathogenesis of tissue damage across snake species. Our data illustrate that toxin-inhibiting small molecule drugs show considerable potential as novel broad-spectrum treatments against the local skin-damaging effects of cytotoxic snake venoms, an important outcome when considering current antivenom therapies have limited efficacy against severe local envenoming. Our findings therefore advocate for further research to translate these drugs and their combinations into community-deliverable snakebite treatments with the goal of significantly reducing the morbidity associated with one of the world's most neglected tropical diseases. East-African Naja nigricollis (Tanzania), West-African Naja nigricollis (Nigeria), and Naja pallida (Tanzania).

Note that the Indian E. carinatus venom was collected from a specimen that was inadvertently imported to the UK via a boat shipment of stone, and then rehoused at LSTM on the request of the UK Royal Society for the Prevention of Cruelty to Animals (RSPCA). Crude venoms were lyophilized and stored at 4 °C to ensure longterm stability. Prior to use, venoms were resuspended to 10 mg/ml in PBS and then kept at -80 °C until used in the described experiments, with freeze-thaw cycles kept to a minimum to prevent degradation. MTT assays alone. HaCaT cells were seeded (5,000 cells/well, clear-sided 96-well plates) in standard medium, then left to adhere overnight at standard conditions. The next day, serial dilutions were prepared in standard medium of (a) venom treatments (1-1,024 µg/mL; i.e. Fig. 1) 

The concentration that resulted in a 50% reduction in adherent cell viability (IC 50 ) was calculated from the log 10 concentration versus normalised response curves using the 'log(inhibitor) vs. normalized response -Variable slope' in GraphPad Prism, which uses the following equation:

where y is the normalised %-cell viability values and x is the log 10 of the venom concentrations.

MTT assays multiplexed with PI assays. HaCaT cells were seeded (20, 

where y is the normalised PI (RFU treatment minus RFU blanks ) values and x is the log 10 of the venom concentrations.

After the PI assays were completed, the PI-containing treatment solutions were aspirated from each well and replaced with 100 µL/well of minimally fluorescent medium containing 0.833 mg/mL of filtered MTT solution, and MTT assays completed and analysed as described above. asper venom and 100 µg of C. atrox venom (i.e. Supplementary Fig. 2 necrosis, disarray with complete loss of architecture and hyalinization. In the epidermis, ulceration with superficial debris was interpreted as evidence of necrosis. In the dermis, loss of skin adnexal structures (e.g. hair follicles and sebaceous glands) and extracellular matrix disarray were also interpreted as evidence of necrosis.

Expanding upon methods originally published by Ho, et al. 55 , the %-necrosis of each skin layer (epidermis, dermis, hypodermis, panniculus carnosus, and adventitia) within each image was assessed by two independent and blinded pathologists and scored between a 0 and 4, with a 0 meaning no observable necrosis in the layer within that image, a 1 meaning up to 25% of the layer in that image exhibiting signs of necrosis, a 2 meaning 25-50% necrosis, a 3 meaning 50-75%, and a 4 meaning more than 75% exhibiting indicators of necrosis. The mean scores of the pathologists for each layer from each image were determined, and the highest scores-permouse used for our data analysis as these represented the maximum necrotic severity within each lesion (i.e. Supplementary Fig. 4) . The 'mean overall dermonecrosis severity' was determined for each lesion by taking the mean of the individual layer scores (i.e. Fig. 6d-g) .

Statistical Analysis. All data are presented as mean ± standard deviation 75 of at least three independent experimental replicates. For cell experiments, 'n' is defined as an independent experiment completed at a separate time from other 'n's within that group of experiments; all drug and/or venom treatments within an 'n'

were completed in triplicate wells and the mean taken as the final value for that one trial. For in vivo experiments, 'n' is defined as the number of mice in that specific treatment group 76 . Two-tailed t-tests were performed for dual comparisons, one-way analysis of variances (ANOVAs) performed for multiple comparisons with one independent variable followed by Dunnett's or Tukey's multiple comparisons tests when the trial data were compared to a single control group or to all other groups, respectively, as recommended by GraphPad Prism, and two-way ANOVAs performed for multiple comparisons with two independent variables followed by Dunnett's multiple comparisons tests. A difference was considered significant if P ≤ 0.05.

Funding was provided by a (i) Newton International Fellowship (NIF\R1\192161) from the Royal Society to SRH, (ii) a Sir Henry Dale Fellowship

iv) a UK Medical Research Council research grant (MR/S00016X/1) to NRC. This research was funded in whole, or in part, by the Wellcome Trust. For the purpose of open access, the authors have applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission

The global burden of snakebite: a literature analysis and modelling based on regional estimates of envenoming and deaths

Vulnerability to snakebite envenoming: a global mapping of hotspots

The London declaration on neglected tropical diseases: 5 years on

Snakebite: the biggest public health crisis you've never heard of

Strategy for a globally coordinated response to a priority neglected tropical disease: snakebite envenoming

Causes and consequences of snake venom variation

Small molecule drug discovery for neglected tropical snakebite

A review and database of snake venom proteomes

Haemotoxic snake venoms: their functional activity, impact on snakebite victims and pharmaceutical promise

Multifunctional toxins in snake venoms and therapeutic implications: from pain to hemorrhage and necrosis

Hemorrhage caused by snake venom metalloproteinases: a journey of discovery and understanding

Neurotoxicity in snakebite -the limits of our knowledge

Skin necrosis due to snakebites

Cytotoxic effects of snake venoms

Antivenom: the most cost-effective treatment in the world?

Adverse reactions to snake antivenom, and their prevention and treatment

Snake antivenoms

Preclinical antivenom-efficacy testing reveals potentially disturbing deficiencies of snakebite treatment capability in East Africa

Biosynthetic oligoclonal antivenom (BOA) for snakebite and next-generation treatments for snakebite victims

Recombinant snakebite antivenoms: a cost-competitive solution to a neglected tropical disease?

How can monoclonal antibodies be harnessed against neglected tropical diseases and other infectious diseases?

Varespladib (LY315920) appears to be a potent, broadspectrum, inhibitor of snake venom phospholipase A2 and a possible pre-referral treatment for envenomation

Exploration of the inhibitory potential of varespladib for snakebite envenomation

Repurposing cancer drugs batimastat and marimastat to inhibit the activity of a group I metalloprotease from the venom of the western diamondback rattlesnake, Crotalus atrox

Preclinical validation of a repurposed metal chelator as a community-based therapeutic for hemotoxic snakebite

A therapeutic combination of two small molecule toxin inhibitors provides broad preclinical efficacy against viper snakebite

Neutralizing effects of small molecule inhibitors and metal chelators on coagulopathic viperinae snake venom toxins

Developing small molecule therapeutics for the initial and adjunctive treatment of snakebite

The global snake bite initiative: an antidote for snake bite

The urgent need to develop novel strategies for the diagnosis and treatment of snakebites

Topical and transdermal drug delivery: from simple potions to smart technologies

Varespladib and cardiovascular events in patients with an acute coronary syndrome: the VISTA-16 randomized clinical trial

Matrix metalloproteinase inhibitors in cancer therapy: turning past failures into future successes

Peptidomimetic hydroxamate metalloproteinase inhibitors abrogate local and systemic toxicity induced by Echis ocellatus (saw-scaled) snake venom

Inhibition of local hemorrhage and dermonecrosis induced by Bothrops asper snake venom: effectiveness of early in situ administration of the peptidomimetic metalloproteinase inhibitor batimastat and the chelating agent CaNa2EDTA

Neutralization of the haemorrhagic activities of viperine snake venoms and venom metalloproteinases using synthetic peptide inhibitors and chelators

Neutralizing properties of LY315920 toward snake venom group I and II myotoxic phospholipases A2

Delayed Oral LY333013 Rescues Mice from Highly Neurotoxic, Lethal Doses of Papuan Taipan (Oxyuranus scutellatus) Venom. Toxins (Basel)

Delayed LY333013 (oral) and LY315920 (intravenous) reverse severe neurotoxicity and rescue juvenile pigs from lethal doses of micrurus fulvius (Eastern coral snake) venom

Varespladib (LY315920) and methyl varespladib (LY333013) abrogate or delay lethality induced by presynaptically acting neurotoxic snake venoms

Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays

Cell viability assays

Eli Lilly & Company and the National Center for Advancing Translational Sciences

HaCaT cells as a reliable in vitro differentiation model to dissect the inflammatory/repair response of human keratinocytes

Growth and differentiation of HaCaT keratinocytes

In vitro activity of BaH1, the main hemorrhagic toxin of Bothrops asper snake venom on bovine endothelial cells

Metrics other than potency reveal systematic variation in responses to cancer drugs

A personal view of pharmacology

A lipid-based oral supplement protects skin cells in culture from ultraviolet light and activates antioxidant and anti-inflammatory mechanisms

Leukaemia xenotransplantation in zebrafishchemotherapy response assay in vivo

Evaluation of impermeant, DNA-binding dye fluorescence as a real-time readout of eukaryotic cell toxicity in a high throughput screening format

High-throughput method for dynamic measurements of cellular viability using a FLUOstar OPTIMA

Development of simple standard assay procedures for the characterization of snake venom

Preclinical assessment of the efficacy of a new antivenom (EchiTAb-Plus-ICP®) for the treatment of viper envenoming in sub-Saharan Africa

Analysis of the necrosis-inducing components of the venom of Naja atra and assessment of the neutralization ability of freeze-dried antivenom

Convergent evolution of defensive venom components in spitting cobras

Neurotoxic effects of bites by the Egyptian cobra (Naja haje) in Nigeria

Interpreting steep dose-response curves in early inhibitor discovery

Synergism between basic Asp49 and Lys49 phospholipase A2 myotoxins of viperid snake venom in vitro and in vivo

Unity makes strength: exploring intraspecies and interspecies toxin synergism between phospholipases A2 and cytotoxins

Phospholipase A2 enhances the endothelial cell detachment effect of a snake venom metalloproteinase in the absence of catalysis

Snake venom metalloproteinases: structure, function and relevance to the mammalian ADAM/ADAMTS family proteins

Recent advances in matrix metalloproteinase inhibitor research

A phase II trial of marimastat in advanced pancreatic cancer

In vitro and in vivo preclinical venom inhibition assays identify metalloproteinase inhibiting drugs as potential future treatments for snakebite envenoming by Dispholidus typus

Venom-induced blood disturbances by palearctic viperid snakes, and their relative neutralization by antivenoms and enzyme-inhibitors

The importance of bites by the saw scaled or carpet viper (Echis carinatus): epidemiological studies in Nigeria and a review of the world

New insights into snakebite epidemiology in Costa Rica: a retrospective evaluation of medical records

The fight to manufacture COVID vaccines in lower-income countries

Global COVID-19 vaccine inequity: failures in the first year of distribution and potential solutions for the future

Observations on the bite of the Mozambique spitting cobra (Naja mossambica mossambica)

Pathogenesis of dermonecrosis induced by venom of the spitting cobra, Naja nigricollis: an experimental study in mice

TRUE-1: trial of repurposed Unithiol for snakebite envenoming phase 1 (safety, tolerability, pharmacokinetics and pharmacodynamics in healthy Kenyan adults)

Jadomycins are cytotoxic to ABCB1-, ABCC1-, and ABCG2-overexpressing MCF7 breast cancer cells

Mean SEM' or 'mean (SD)'

What exactly is 'N' in cell culture and animal experiments?

We would like to give our thanks to Paul Rowley for maintaining the snakes at the LSTM's herpetarium and for routine venom extractions, Dr. Cassandra Modahl and Dr. Amy Marriott for their help with animal welfare observations during the in vivo experimentation, Valerie Tilston and her team at the University of Liverpool for preparing the histopathology slides,

The authors declare no competing interests.Correspondence and requests for materials should be addressed to NRC.