Repurposed drugs and their combinations prevent morbidity-inducing dermonecrosis caused by diverse cytotoxic snake venoms

Snakebite envenoming causes 1.8–2.7 million envenomings and 81,000–138,000 deaths annually, with around 400,000 survivors experiencing morbidity, often severe local tissue necrosis requiring debridement or amputation. Venoms are complex mixtures dominated by four toxin families—phospholipases A2 (PLA2s), snake venom metalloproteinases (SVMPs), snake venom serine proteases (SVSPs), and three-finger toxins (3FTxs)—that drive haemotoxic, neurotoxic, and cytotoxic syndromes. Conventional antivenoms are life-saving for systemic effects but have limitations: cost, cold-chain dependence, IV administration in clinical settings, variable cross-species efficacy, adverse reactions, and poor penetration into peripheral tissues, rendering them largely ineffective at preventing local dermonecrosis. Small molecule toxin inhibitors offer advantages in stability, potential cross-species activity, and amenability to rapid field delivery (oral, topical, subcutaneous or intradermal). Prior work has identified three repurposed drugs—DMPS (SVMP-inhibiting metal chelator), marimastat (hydroxamate SVMP inhibitor), and varespladib (secretory PLA2 inhibitor)—as promising systemic antivenoming agents, with marimastat+varespladib improving survival in mice across diverse venoms. The present study investigates whether these drugs, alone or in combination, can prevent venom-induced local dermal cytotoxicity and dermonecrosis, including when treatment is delayed post-envenoming.

The paper reviews the dominance of PLA2s, SVMPs, SVSPs, and 3FTxs across venoms and their roles in haemotoxicity, neurotoxicity, and cytotoxic dermonecrosis. It summarizes limitations of current antivenoms, especially for local tissue damage, due to molecular size and pharmacological properties that hinder peripheral tissue penetration. Recent interest has grown in next-generation snakebite therapies, including monoclonal antibodies and small molecule toxin inhibitors. Prior studies showed DMPS and marimastat inhibit SVMPs and reduce haemotoxicity; varespladib inhibits sPLA2s and protects against systemic toxicity and neurotoxicity in animal models. Combination therapy (marimastat+varespladib) previously improved survival against diverse viperid venoms. However, evidence for efficacy against local dermonecrosis was limited, motivating the current work.

In vitro cell assays:

• Cell line: Immortalised human epidermal keratinocytes (HaCaT) cultured under standard conditions.
• Venoms: 11 geographically diverse species: Bitis arietans (Nigeria), Bothrops asper (Costa Rica, Caribbean), Crotalus atrox (USA lineage), Calloselasma rhodostoma (Malaysia), Daboia russelii (Sri Lanka), Echis carinatus (India), Echis ocellatus (Nigeria; likely E. romani), Naja haje (Uganda), Naja nigricollis East African (Tanzania), Naja nigricollis West African (Nigeria), Naja pallida (Tanzania). Venoms stored lyophilized, reconstituted to 10 mg/mL in PBS.
• Drugs: DMPS (metal chelator), marimastat (hydroxamate SVMP inhibitor), varespladib (sPLA2 inhibitor). Working stocks: DMPS 400 mM in PBS (fresh), marimastat 10 mM in water, varespladib 65.7 mM in DMSO.
• MTT viability assays: HaCaT seeded (5,000 cells/well). Venom dose-response (1–1024 µg/mL) for 24 h to determine IC50 and Hill slope. Determined maximum tolerated concentration (MTC) for drugs and used half MTC (MTC50) for inhibition tests: DMPS 625 µM, marimastat 2.56 µM, varespladib 128 µM. Venoms (2.5–200 µg/mL) pre-incubated with drug or vehicle for 30 min before 24 h exposure; IC50 calculated.
• PI cell death assays multiplexed with MTT: HaCaT seeded (20,000 cells/well, black-sided plates). Serial dilutions of D. russelii or B. asper venom (2.2–127 µg/mL) pre-incubated 30 min with drug(s): DMPS 625 µM, marimastat 2.56 µM, varespladib 256 µM, DMPS+varespladib (625/256 µM), marimastat+varespladib (2.56/256 µM) or vehicle, in minimally fluorescent medium with PI (50 µg/mL). After 24 h, PI fluorescence recorded to compute EC50; then MTT performed.

In vivo dermonecrosis models:

• Ethics: UK (pre-incubation model) under Home Office PPL #P58464F90; Costa Rica (rescue model) under CICUA 82-08.
• Animals: CD-1 mice, 18–20 g, grouped in fives, standard housing.
• Pre-incubation model (minimum necrotic dose principles): Determined intradermal (ID) venom doses yielding consistent local lesions without systemic effects: B. asper 150 µg, C. atrox 100 µg; E. ocellatus 39 µg (previously published). Venom or PBS pre-incubated 30 min at 37 °C with: vehicle (98.48% PBS, 1.52% DMSO), DMPS 110 µg, marimastat 60 µg, varespladib 19 µg, DMPS+varespladib (110/19 µg), or marimastat+varespladib (60/19 µg). Fifty µL injected ID into shaved flank skin (n=5/group). Mice euthanised at 72 h (some earlier at 24 h per humane endpoint). Lesions excised, photographed, and area (mm²) quantified.
• Histopathology: Cross sections of ID sites fixed, paraffin-embedded, H&E stained. Two blinded scorers graded necrosis in epidermis, dermis, hypodermis, panniculus carnosus, and adventitia on a 0–4 scale (0–100% involvement categories). Overall dermonecrosis score calculated as mean of layer maxima.
• Rescue model (post-venom treatment): Mice ID-injected with B. asper (150 µg) or E. ocellatus (39 µg). Marimastat+varespladib (60/19 µg) administered ID at the same site at 0, 5, 15, or 60 minutes post-venom; vehicle given immediately post-venom to separate control animals. Lesions quantified at 72 h.

Statistics: One-way ANOVA with appropriate multiple comparisons (Dunnett’s or Tukey’s) reported for IC50/Hill slopes, lesion areas, and histopathology scores; data presented as mean ± SD with individual points.

• Venom cytotoxicity in HaCaT cells: 11 venoms showed dose-dependent inhibition of adherent cell viability with broadly similar potencies for 9/11 venoms. IC50 ranges: viperids B. arietans, B. asper, C. atrox, C. rhodostoma, E. carinatus, E. ocellatus at 7.5–19.6 µg/mL; spitting cobras N. nigricollis (East and West) and N. pallida at 23.1–27.2 µg/mL; D. russelii less potent (IC50 45.1 µg/mL); N. haje least potent (IC50 86.8 µg/mL). Hill slopes for all venoms were >1 in magnitude and not significantly different, indicating steep, cooperative responses.
• Drug inhibition in cell assays: At MTC50 concentrations (DMPS 625 µM; marimastat 2.56 µM; varespladib 128 µM), SVMP inhibitors (DMPS, marimastat) significantly increased IC50s (reduced potency) for C. atrox, E. carinatus, and E. ocellatus; DMPS modestly increased IC50 for East African N. nigricollis, but not West African N. nigricollis. Varespladib showed no significant protection in these six-venom screens. For PLA2-rich D. russelii and B. asper, DMPS and marimastat reduced loss of viability and PI-measured cell death, whereas varespladib alone remained ineffective. Combining varespladib with marimastat, but not with DMPS, significantly potentiated protection against B. asper in both MTT and PI assays.
• In vivo dermal lesion pre-incubation model: • Bothrops asper (150 µg): Mean lesion area 41.9 mm². Marimastat did not reduce lesions (55.1 mm², ns); varespladib significantly reduced to 12.2 mm²; DMPS reduced to 21.1 mm² (ns). Both combinations significantly reduced lesions: DMPS+varespladib to 2.7 mm²; marimastat+varespladib to 6.7 mm². • Crotalus atrox (100 µg): Mean lesion 19.1 mm². All monotherapies significantly reduced lesions: DMPS 3.1 mm², marimastat 4.4 mm², varespladib 5.8 mm². Combinations reduced to 0.3 mm² (both DMPS+V and M+V). • Echis ocellatus (39 µg): Mean lesion 5.0 mm². Varespladib ineffective (7.0 mm²). Marimastat significantly reduced to 0 mm²; DMPS largely protective (four of five with no lesion) but not statistically significant due to one outlier (1.0 mm², P=0.0856). Combinations significantly reduced to 0.1 mm² (DMPS+V) and 0.4 mm² (M+V).
• Histopathology: For B. asper, overall dermonecrosis score decreased from 2.57 to 0.72 with varespladib, 0.06 with DMPS+V, and 0.32 with M+V; DMPS and marimastat alone were ineffective across layers. For C. atrox, all treatments improved scores in superficial layers, with overall scores reduced from 2.86 to a 0.04–1.32 range across treatments. For E. ocellatus, trends favored SVMP inhibition and combinations, though overall ANOVA for scores was not significant in the presented dataset.
• Rescue model (post-venom treatment): Marimastat+varespladib administered ID at 0, 5, 15, or 60 minutes post-envenoming significantly reduced dermal lesion sizes induced by B. asper and E. ocellatus at all timepoints tested, demonstrating a clinically relevant treatment window up to 1 hour. Overall, combinations targeting both SVMPs and PLA2s provided broad, pan-species protection, outperforming any monotherapy and often preventing lesions entirely in individual animals.

The study addresses the key challenge that current antivenoms poorly prevent local tissue necrosis following snakebite. While SVMP inhibitors (DMPS, marimastat) mitigated cytotoxic effects in vitro and, in some cases, in vivo, their efficacy varied by species and did not universally prevent dermonecrosis as monotherapies. Varespladib, despite limited impact in cell assays, significantly reduced dermal injury in vivo for certain venoms, underscoring that in vitro cytotoxicity assays do not fully predict in vivo dermal outcomes. Crucially, dual inhibition of SVMPs and PLA2s via combinations (DMPS+varespladib or marimastat+varespladib) provided robust, cross-species protection, frequently abolishing lesion formation and reducing histopathological severity. The rescue experiments further demonstrated efficacy when treatment was delayed up to 60 minutes post-envenoming, supporting real-world applicability where delays to care are common. These findings validate a therapeutic strategy that pairs complementary toxin inhibitors to cover the dominant drivers of dermonecrosis across diverse viperid venoms and suggest locally deliverable formulations (e.g., intradermal, transdermal) could meaningfully reduce morbidity in resource-limited settings.

DMPS, marimastat, and varespladib can each protect against venom-induced dermonecrosis, but their monotherapy efficacy is species-dependent and incomplete. Combining an SVMP inhibitor (DMPS or marimastat) with a PLA2 inhibitor (varespladib) yields broad-spectrum efficacy, markedly reducing lesion size and histopathological severity across geographically and taxonomically diverse viperid venoms, and remains effective even when administered up to one hour after envenoming. These results support advancing dual small-molecule toxin inhibitor combinations toward community-deliverable therapies to prevent snakebite-associated morbidity. Future work should expand testing to additional species (including cytotoxic elapids), optimize routes of administration (oral, topical, transdermal, local injection), and define pharmacokinetics/pharmacodynamics and dosing, including potential drug–drug interactions. Ongoing clinical evaluation of DMPS (Phase I) and methyl varespladib (Phase II) underscores the translational potential of such approaches.

• In vitro assays (MTT/PI) did not predict all in vivo outcomes; varespladib was inactive in cell assays yet effective in vivo for some venoms, highlighting model limitations.
• A limited set of venoms was tested in vivo (B. asper, C. atrox, E. ocellatus); broader taxonomic coverage, including additional viperids and cytotoxic Naja species, is needed.
• The principal in vivo efficacy model involved venom–drug pre-incubation, which may not fully mirror clinical envenoming; although a rescue model was used, it evaluated only the marimastat+varespladib combination.
• Drug delivery was via intradermal injection at the bite site; other clinically feasible routes (oral, topical, transdermal) require evaluation, along with PK/PD, dosing regimens, and potential interactions.
• Some statistical outcomes were influenced by small group sizes and outliers (e.g., DMPS with E. ocellatus histology), warranting larger studies.