Humanized Na_V1.8 rats overcome cross-species potency shifts in developing novel Na_V1.8 inhibitors

Highlights

• •MSD199 is a potent and selective Nav1.8 inhibitor.
• •Humanized rats can overcome species potency shifts in developing NaV1.8 inhibitors.
• •MSD199 demonstrates preclinical efficacy in SNL and CFA.

Abstract

Voltage-gated sodium channel isoform 1.8 (Na_V1.8) has emerged as a promising pharmaceutical target for the treatment of acute and chronic pain. However, highly selective and potent inhibitors for this channel have been difficult to develop and only recently have advanced to clinical testing. Our efforts to develop Na_V1.8 small molecule inhibitors yielded a series of molecules with favorable in vitro potency and selectivity against the human Na_V1.8 channel but exhibited dramatic rightward potency shifts against the rodent channel, severely limiting in vivo screening and candidate selection. In anticipation of supporting drug discovery efforts, a transgenic rat line expressing the human Na_V1.8 channel in lieu of the rodent channel was developed. Utilizing these humanized animals, the in vitro potency of our chemical matter in freshly isolated humanized rat DRG neurons was consistent with in vitro human potency values, enabling in vivo work to progress. We demonstrate capsaicin-induced nocifensive behaviors (CNB) as a moderate throughput in vivo screening assay, from which we demonstrate pharmacokinetic-pharmacodynamic (PK-PD) and in vitro-in vivo correlation (IVIVC) relationships. We identified MSD199 as a potent Na_V1.8 inhibitor with acute pain efficacy and assessed it in traditional inflammatory (Complete Freund’s Adjuvant) and neuropathic (spinal nerve ligation) behavioral chronic pain assays where it was shown to significantly reduce pain-related behaviors. Overall, we demonstrate the utility of humanized transgenic animals when cross-species potency shifts are observed within an otherwise promising chemical series.

Introduction

Voltage-gated sodium (Na_V) channel inhibitors hold significant therapeutic potential in the non-opioid treatment of acute and chronic pain. Among the Na_V channel family, Na_V isoform 1.8 (Na_V1.8) has emerged as a particularly attractive pharmaceutical target due to predominant expression in thinly myelinated Aδ and unmyelinated small-diameter sensory neurons (C-fibers) with limited expression elsewhere (Akopian et al., 1996, Pristera et al., 2012). These fibers relay nociceptive information to the central nervous system, and their hyperactivity is believed to play a significant role in the development and maintenance of persistent and chronic pain (Finnerup et al., 2021, Walters et al., 2023). Within these fibers, the Na_V1.8 channel is a key contributor to action potential generation, particularly the upstroke of the action potential in which it exhibits fast activation and slow inactivation kinetics (Blair and Bean, 2002, Renganathan et al., 2001, Sangameswaran et al., 1996). Unlike other sodium channels, Na_V1.8 exhibits recovery from inactivation quickly, facilitating repetitive or burst firing in response to stimuli (Akopian et al., 1996, Dib-Hajj et al., 1997). Importantly, studies linking Na_V1.8 to human pain conditions have demonstrated gain-of-function mutations within the Na_V1.8 gene that are associated with painful peripheral neuropathies, highlighting its critical role in pain modulation. The individuals harboring these mutations tend to display hypersensitivity to pain in a manner resembling small fiber neuropathy such as a burning or stabbing pain in extremities and sensitive skin (Eijkenboom et al., 2019, Faber et al., 2012). Such mutations were also shown to result in hyperexcitability of dorsal root ganglion neurons (DRG) (Faber et al., 2012). Additionally, increased Na_V1.8 expression has been observed in certain painful conditions, such as human lingual neuromas, where its expression correlates with the severity of pain symptoms (Bird et al., 2013).

Consequently, for these reasons it has long been hypothesized that Na_V1.8 inhibition may provide analgesia for the treatment of acute and chronic pain conditions. However, successful Phase II and III clinical trials involving selective Na_V1.8 inhibition in pain have only recently been reported (NCT05558410, NCT05553366) (Jones et al., 2023), culminating with the approval of the first selective Na_V1.8 inhibitor, Suzetrigine, in January 2025. One challenge in developing potent and selective small molecule Na_V1.8 inhibitors is pronounced species differences in Na_V1.8 channel homology, which can lead to cross-species shifts in potency which slows preclinical development and lead optimization efforts (Gilchrist et al., 2024, Jarvis et al., 2007, Krafte et al., 2007, McGaraughty et al., 2008, Payne et al., 2015).

Our internal efforts to develop Na_V1.8 small molecule inhibitors yielded a series of candidates with promising in vitro potency and selectivity against the human Na_V1.8 channel but exhibited dramatic rightward-shifted potency against the rodent Na_V1.8 channel, presumably due to differences in protein homology between rodent and human Na_V1.8 channels. This cross-species potency shift hindered lead selection and optimization efforts as achieving sufficient in vivo exposure levels to meaningfully engage the rodent Na_V1.8 channel was not feasible. To overcome this obstacle, we developed a custom-designed transgenic rat model expressing the human Na_V1.8 channel. Utilizing this model, we demonstrate Na_V1.8 inhibitor potency in these animals in line with human potency, enabling the establishment of pharmacokinetic-pharmacodynamic (PK-PD) relationships. This facilitated the identification of MSD199, which was subsequently advanced into assays assessing inflammatory and neuropathic pain behaviors.

Methods

Animals

Male wild-type Sprague-Dawley rats and humanized Na_V1.8 rats (homozygous Nav1.8 KO/ homozygous Na_V1.8 BAC Tg, see below) maintained in a Sprague-Dawley background were obtained from Charles River Laboratories (CRL, Wilmington, MA). Rats were maintained on a standard 12 h light/dark cycle and given access to food and water ad libitum. All aspects of the work including housing, experimentation, and animal disposal was performed in general accordance with the “Guide for the Care and Use of Laboratory Animals: Eighth Edition” (The National Academies Press, Washington, DC, 2011) in AAALAC-accredited laboratory animal facilities. All animal protocols were approved by the Merck & Co., Inc., Rahway, NJ, USA Institutional Animal Care and Use Committee.

Generation of a humanized Na_V1.8 rat

Transgenic (Tg) Na_V1.8 KO rats were generated by SAGE/Envigo/Inotiv (Indianapolis, IN) on behalf of Merck & Co., Inc., Rahway, NJ, USA (official gene symbol for Na_V1.8 is Scn10a). Exon 6, which encodes for the S4 voltage sensor domain I (IS4) of the Na_V1.8 rat gene, was chosen as the target, as an equivalent exon was disrupted in a previously published Na_V1.8 KO mouse model resulting in a functional null mutation (Akopian et al., 1999). To generate a Na_V1.8 knockout rat, synthesized sgRNA (GGTGCAGCGATAGACCTCCGAGG) were ordered from Integrated DNA Technologies (Coralville, Iowa) and co-injected with Cas9 protein into one cell stage embryos harvested from female Sprague Dawley rats. Injected embryos were implanted into pseudo-pregnant females for pregnancy. One female founder was identified having a 7 bp frame-shift deletion resulting in disruption of the IS4 domain and introduction of a downstream stop codon in exon 6. The female founder was bred to germ line transmission and heterozygous offspring used to establish homozygous Na_V1.8 KO and wild-type controls. Next, rats containing a bacterial artificial chromosome (BAC) encoding the human Na_v1.8 channel (Na_V1.8 BAC Tg rats) were generated at SAGE/Envigo/Inotiv on behalf of Merck & Co., Inc., Rahway, NJ, USA. The BAC clone RP11-1114A3 contained approximately 50 kb upstream, 100 kb of human Na_V1.8 and 10 kb downstream containing 3′UTR and regulatory regions. An additional 30 kb DNA sequences from BAC clone RP11-831 J6 were added to the 3′ region by DNA recombineering comprising of 3′UTR and regulatory regions. The fused RP11-1114A3 construct (BAC1) was then introduced into fertilized Sprague Dawley embryos by microinjection. Four independent BAC Tg founder lines were generated and evaluated for expression of the human Na_V1.8 transgene. Only one line showed high expression levels of the human BAC transgene and was further bred to homozygosity for the human Na_V1.8 transgene. Finally, humanized Na_V1.8 rats were generated by breeding Na_V1.8 CRISPR KO rats with the Na_V1.8 BAC Tg mice over several generations to establish double homozygous rats lacking the rodent Na_V1.8 channel and expressing the human channel.

Wild-type and humanized DRG isolation

Lumbar DRG (L4-L6) from wild type or humanized Na_V1.8 rats were collected in ice cold Hibernate A media without calcium (BrainBits, Springfield, Il). A total of 4 ganglion were carefully trimmed under a dissecting microscope, in ice cold Hibernate A media without calcium, to remove the capsules and roots. Following trimming, DRG were digested in Hibernate A media with calcium (BrainBits, Springfield, Il) containing collagenase 2 (1 mg/ml, Worthington Biochemical Corp.) and neutral protease (1 mg/ml, Worthington Biochemical Corp.) for 40 min at 37 °C. The ganglia were then triturated with fire-polished Pasteur pipets and pelleted at 1000 rpm for 10 min. The supernatant was aspirated, and the cell pellet was resuspended in Neurobasal A media (Gibco, Thermo Fisher Scientific) containing 30 ng/ ml each of mouse NGF (Alamone Labs) + rat GDNF (Peprotech). Dissociated neurons were seeded on glass coverslips coated with poly-D-lysine and laminin (Corning, Thermo Fisher Scientific), and maintained in a 37 °C, 5 % CO2 incubator prior to electrophysiology experiments.

Manual whole-cell patch clamp electrophysiology

In vitro activity of MSD199 was assessed by measurement of TTX-resistant currents (TTX-r) on acutely dissected wild-type or humanized Na_V1.8 rat L4-L6 DRG using manual voltage clamp electrophysiology. The internal solution consisted of (in mM): 30 CsCl, 5 HEPES, 10 EGTA, 120 CsF, 5 NaF, 2 MgCl₂, pH = 7.3 with CsOH. The external solution for TTX-r recordings consisted of (in mM): 40 NaCl, 97.5 choline-Cl, 5 KCl, 2 MgCl₂, 2 CaCl₂, 10 glucose, 30 TEA Cl, 0.1 CdCl₂, and 10 HEPES, pH to 7.4 with NaOH. Whole-cell voltage-clamp recordings were performed at room temperature, and cell capacitance and series resistance were compensated at levels approximately 80 % of the maximum resistance before recording. Cells were held at −80 mV and depolarized to + 10 mV for 50 ms to measure peak amplitudes of TTX-r sodium inward current at a frequency of 0.1 Hz. Once a stable current was reached, MSD199 was applied in ascending concentrations to DRGs, with 3 concentrations tested. Peak amplitudes of TTX-r sodium inward currents were measured in 300 nM TTX (control) (Sigma-Aldrich Co., St Louis, MO, USA) and in the presence of ascending concentrations of MSD199 together with 300 nM TTX. The percent inhibition of TTX-r Na_V channels by MSD199 was determined as follows: 100 % x [1-(I/I_o)], whereas I equals the peak current in the presence of MSD 199 and 300 nM TTX, and I_o is the peak current in 300 nM TTX control. The last three steady state sweeps for each addition were averaged prior to percent inhibition calculations. A 4-parameter logistic curve was used to determine the IC₅₀ value for each experiment using Graphpad Prism 8. Minimum was set to 0 and maximum to 100. The hill slope was not constrained.

Automated whole-cell patch-clamp recordings

Qube voltage-clamp experiments were performed using HEK293 cell lines or CHO cell lines stably expressing human Na_V1.1, Na_V1.2, Na_V1.3, Na_V1.4, Na_V1.5, Na_V1.6, Na_V1.7, Na_V1.8 or Na_V1.9 on the Qube or Qpatch automated patch-clam platforms (Sophion Bioscience A/S, Ballerup, Denmark) as previously described (Kraus et al., 2021, Vardigan et al., 2024).

MSD199 binding location assessment using chimeric Na_V channels

We systematically exchanged the domains of Na_V1.8 and Na_V1.5 channels, one at a time, followed by the exchange of specific segments and their associated loops, to accurately identify the binding site of MSD 199. Chimera 1 was engineered by modifying Na_V1.5 sodium channel where the entire second domain (DII) – was substituted with the corresponding domain from the Na_V1.8 channel (Chimera 1 structure: Na_V1.5(1–711) – Na_V1.8(660–890) – Na_V1.5(940–2015)). Conversely, Chimera 2 was constructed on the backbone of the Na_V1.8 sodium channel, with DII replaced with that from Na_V1.5 channel (Chimera 2 structure: Na_V1.8(1–399) – Na_V1.5(416–1199) – Na_V1.8(1148–1956)). Chimera A was engineered by modifying Na_V1.5 sodium channel where the second domain (DII) – segment 3 (S3) along with the associated intracellular and extracellular loop regions was substituted with the corresponding domains from the Na_V1.8 channel (Chimera A structure: Na_V1.5(1–770) – Na_V1.8(719–751) – Na_V1.5(804–2015)). Conversely, Chimera B was constructed on the backbone of the Na_V1.8 sodium channel. In this variant, DII segment 3 and its connecting intracellular and extracellular loops was replaced with that from Na_V1.5 channel (Chimera B structure: Na_V1.8(1–718) – Na_V1.5(771–803) – Na_V1.8(752–1956)). Following the electroporation of mammalian over-expression plasmids carrying these constructs into FS-293 HEK cells, chimeric sodium currents were recorded utilizing the Sophion QPatch II system. Internal Solution was made using 30 CsCl, 5 HEPES, 10 EGTA, 120 CsF, 5 NaF, 2 MgCl₂ (in mM), pH = 7.3 with CsOH and the External Solution comprised of 150 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, 12 Dextrose (in mM), pH 7.3 with NaOH (330 Osm). Sodium currents were measured using a voltage protocol stepping from −90 mV to +10 mV to assess inhibition by MSD199.The resultant data was analyzed and fitted using non-linear regression models to determine the IC₅₀. Upper constraint was set to 1, lower constraint was set to 0, hill slope was not constrained. Subsequent rat and human Na_V1.8 sequence alignment was conducted utilizing the UniProt alignment feature. Feature boundaries were adopted based on the UniProt identifiers for Na_V1.8 human (Q9Y5Y9) and rat (Q62968) entries.

Tissue processing

Wild type and humanized rat DRG consisting of L4, L5, and L6 were collected and transferred to tubes containing lysis buffer (Cell Signaling 9803) with protease inhibitor (Pierce Catalog 78442) and PMSF (Sigma 78330). To homogenize tissue, 3 mm beads (Qiagen, 69997) were added to the DRG tubes and processed for 3x1 min at 30 Hz via the Tissuelyser II (Qiagen). The tissue homogenate was centrifuged for 15 min at 12,000×g) at 4 °C. The resulting supernatant was pipetted into 96-well polypropylene storage plates for analysis.

qRT-PCR

Total RNA was extracted from freshly-isolated wild type or humanized Na_V1.8 rat DRGs using TRIzol (TRI reagent, Ambion)−chloroform method (chloroform, Sigma-Aldrich) and purified with Arcturus PicoPure RNA isolation kit (Applied Biosystems) according to the manufacturer's protocol. Next, RNA was reverse transcribed using high-capacity cDNA reverse transcription kit (Applied Biosystems). Specific cDNAs were amplified with TagMan probe and primer sets designed against rat (Rn00568393_m1 Scn10a; lot number 1422945) or human (Hs01045137_m1 SCN10A; lot number 1339855) Na_V1.8 or housekeeping GAPDH (4352338E-1211016) utilizing TaqMan FAST advanced master mix (Applied Biosystems) in a Quant Studio 7 Flex Real Time PCR System (Applied Biosystems). The relative levels of message were calculated using the ΔΔct method.

ELISA

A custom ELISA was developed using previously described methods (Price et al., 2017). Maxisorp plates (Nunc, 446471) were coated with anti-pan Na_V (Sigma, S8809) capture antibody at 3 ug/ml concentration in carbonate-bicarbonate buffer overnight at 4C. Plates were washed once with TBS (Tris-buffer saline 20 mM Tris, 150 mM NaCl, pH 7.4) and blocked for one hour with 3 % BSA/TBS and stored at 4C until use. 100 ul of tissue homogenate was added per well and plates were incubated overnight at 4C. Following, plates were washed 3X with TBST (TBS/ 0.05 % tween-20) and detection antibodies diluted 1:1000 in 1 % BSA/TBST targeted to rat (Alomone Labs, ASC-016) or human (LS Bio, LS-C154767) Na_V1.8 were incubated for 2 h at room temperature. Plates were subsequently washed 3X with TBST and Goat anti-rabbit alkaline phosphatase secondary antibody (Jackson Immuno, 111–055-144) diluted 1:20,000 in 1 % BSA/TBST was added for 1 h at room temperature. Plates were finally washed 5X in TBST, then 100 ul of luminescent substrate (CDP-star substrate, ABI, T2214) was added per well. Plates were incubated protected from light for 30 min, then read on a Perkin Elmer Envision Multilabel reader using an enhanced luminescent protocol with a read time of 0.5 s.

Blinding

For all behavior studies, the experimenter was blinded at all times, such that each animal’s group assignment and treatment was unknown until data collection was complete. In addition, the testing of animals was randomized in regard to test order so that groups were not tested all at once.

Capsaicin nocifensive behavior (CNB)

Studies were conducted either internally or at CRL under the same protocol. Animals were acclimated to the observation plexiglass chamber twice prior to the study day and for 15 min prior to capsaicin injection on study day. Test compounds were dosed PO one-hour prior to capsaicin injection. The dosing timepoint of one hour prior to behavioral testing throughout was chosen as it is the T_max for the given dose and dose route. Immediately before assessment, an intra-dermal injection of 25 μL 0.01 % capsaicin (M2028, Sigma-Aldrich) dissolved in ethanol/ Tween80/ Saline (20:7:73) was delivered into the left hind paw. Animals were then immediately placed back in the observation chamber. Capsaicin-evoked spontaneous nociceptive behaviors in the rats were scored over 5 min either in real time if done at Merck & Co., Inc., Rahway, NJ, USA or recorded using a commercial camcorder for later scoring if at CRL. The total number of flinches (animal raising its injected hind paw and putting it back to ground is counted as 1 flinch) of the capsaicin-injected paw was scored over 5 min. Blood samples for pharmacokinetic analysis were collected following testing via tail vein.

CFA-induced thermal hyperalgesia testing

Studies were conducted externally at CRL. Rats were acclimated to the test chambers for at least 10 min prior to thermal testing. Baseline thermal latencies were obtained via the Hargreaves’ method (Hargreaves et al., 1988). The average of two trials were taken, and the cutoff point was set at 20 s. Following baseline measurements, animals received an intra-plantar subcutaneous injection of Complete Freund’s Adjuvant (CFA; 50 μL, 1 mg/mL) into the mid-plantar region of the hind paw and were subsequently returned to their cages. 24 h following CFA injection, thermal latencies were reassessed. Afterwards, animals were randomized into treatment groups and dosed PO with vehicle or MSD199 (1 or 10 mg/kg), or intraperitoneal with diclofenac (80 mg/kg) as a positive control. Diclofenac dosage was selected based on internal studies and previous published studies in the CFA model (Nagakura et al., 2003). Thermal latencies were reassessed at one-hour post-dosing which corresponds to the time of the maximum blood concentration of the compounds (T_max). Blood samples for PK analysis were collected at four hours post dosing via tail vein.

Spinal nerve ligation (SNL) mechanical allodynia testing

Studies were conducted internally. Rats were habituated to the test environment for 2 days (∼60 min/day) and approximately 30 min prior to testing. Rats were placed on an elevated mesh floor in Perspex boxes and calibrated von Frey filaments of increasing forces were applied to the plantar surface of the paw using – the Up-Down method to determine baseline mechanical paw withdrawal thresholds (PWT) (Chaplan et al., 1994, Dixon, 1980). Rats with baseline PWTs < 9 g were not included for surgery. Surgery: Rats were anesthetized with isoflurane. Depth of anesthesia was verified via lack of reflex to strong toe pinch prior to surgery. Under sterile conditions a small incision was made and the L6 transverse process was removed to expose the L5 spinal nerve. The L5 spinal nerve was tightly ligated and cut between ligatures. Muscles were sutured using silk thread and the skin was closed with metal clips. Rats were allowed to recover for one week prior to mechanical testing. Rats with post-surgery PWT measurements of >6 g were excluded from further testing. On test day 1, vehicle (10 % Tween 80) or 10 mg/kg MSD199 was dosed one hour prior to mechanical testing. After a washout period of four days, rats were reassessed for mechanical allodynia. Animals that met the inclusion criteria were dosed PO one day later, with vehicle or 1 mg/kg MSD199 and subjected to mechanical testing one-hour post-dosing which corresponds to the time of the maximum blood concentration of the compounds (T_max). On both test days, blood samples for PK analysis were collected immediately following mechanical testing via tail vein draws.

Results

Potency of MSD199 in wildtype and humanized rats

We first characterized the novel Na_V1.8 inhibitor 2-(4,4-difluoroazepan-1-yl)-6-methyl-N-(2-sulfamoylpyridin-4-yl)-nicotinamide (MSD199) using manual patch clamp electrophysiology in freshly-isolated wildtype rat DRG neurons. TTX-R currents were isolated via bath application of 300 nM tetrodotoxin and evoked via a voltage step to + 10 mV from a holding voltage of −80 mV. MSD199 was subsequently bath applied to examine the degree of TTX-R inhibition. The median inhibitory concentration (IC₅₀) of MSD199 on the rodent Na_V1.8 channel in DRG was determined to be 4826 nM (95 % CI: 4064 – 5732 nM; n = 4). This was significantly rightward-shifted when compared to the IC₅₀ of MSD199 on the human Na_V1.8 channel, which was previously found to be 4.70 nM in HEK293 cells overexpressing recombinant human Na_V1.8 (Vardigan et al., 2024). MSD199 is additionally ∼2000-fold selective over recombinant human Na_V1.4 channel (8370 ± 1429 nM; n = 3) and showed no inhibition at the highest concentrations tested (30–34 uM) on all other recombinant human Na_V channel isoforms (Vardigan et al., 2024), summarized here in Table 1.

Table 1.

In vitro potencies and rat plasma protein binding values for novel Na_V1.8 inhibitors. Na_V1.1–––1.9 in vitro potencies were determined via automated patch clamp electrophysiology systems. N.d. = not determined.

Molecule	Structure	NaV1.8 IC50 (nM)	Plasma Protein Binding (%)	NaV1.1 IC50 (nM)	NaV1.2 IC50 (nM)	NaV1.3 IC50 (nM)	NaV1.4 IC50 (nM)	NaV1.5 IC50 (nM)	NaV1.6 IC50 (nM)	NaV1.7 IC50 (nM)	NaV1.9 IC50 (nM)
MSD199		4.20	85.0	> 33,570	> 33,570	> 33,570	8370	> 33,570	> 33,570	> 33,570	>33000
MSD869		1.10	96.5	32,860	> 33,570	> 33,570	20,060	14,820	> 33,570	> 33,570	>33000
MSD537		32.1	97.2	32,930	n.d.	> 33,570	n.d.	> 33,570	> 33,570	> 33,570	>33000

The marked selectivity of MSD199 to Na_V1.8 over other Na_V isoforms allowed for a chimeric-based approach to identify its approximate binding domain. Specific structural domains between Na_V1.8 and Na_V1.5 were successively exchanged to create chimeric constructs to study its effects on the potency of MSD199. This resulted in the disruption of sodium current inhibition by MSD199 following voltage-sensing domain 2 (VSDII) swap of Na_V1.5 into the Na_V1.8 channel (Chimera 2), as well as robust current inhibition following the VSDII swap of Na_V1.8 into the Na_V1.5 channel (Chimera 1 – IC₅₀ of 1.1 nM; 95 % CI 0.6–1.9 nM). These findings (Fig. 1C) identified VSDII as the primary locus for MSD199 binding. Further elucidations swapping specifically VSDII segment 3 (VSDIIS3) demonstrated Chimera B, containing VSDIIS3 of Na_V1.5, abolished the potency of MSD 199 while Chimera A, containing VSDIIS3 of Na_V1.8, restores it to nanomolar levels (IC₅₀ of 8.6 nM; 95 % CI 6.2–11.84 nM), suggesting VSDIIS3 plays a key role in MSD199 binding.

Fig. 1.

MSD199 potency in wildtype and humanized Na_V1.8 rats. A. Structure of MSD199, a novel Na_V1.8 small molecule inhibitor. B. Left, representative manual patch clamp traces of TTX-resistant evoked currents from WT rat DRG in the presence of vehicle or increasing concentrations of MSD199. Right, MSD199 exhibits poor potency against TTX-R currents in freshly-isolated wild type rat DRG neurons (IC₅₀ = 4826 nM, 95 % CI: 4064–5732 nM). C. Left, cartoon illustration of chimeric channels created to determine MSD199 binding location. Structure of chimeric channels are as follows: Chimera 1: Na_V1.5(1–711) – Na_V1.8(660–890) – Na_V1.5(940–2015); Chimera 2: Na_V1.8(1–399) – Na_V1.5(416–1199) – Na_V1.8(1148–1956); Chimera A: Na_V1.5(1–770) – Na_V1.8(719–751) – Na_V1.5(804–2015); Chimera B: Na_V1.8(1–718) – Na_V1.5(771–803) – Na_V1.8(752–1956). Right, patch clamp electrophysiology demonstrating MSD199 inhibits currents arising from Chimeras 1 and A but not 2 and B, suggesting a binding site of MSD199 to be within voltage-sensing domain 2, within segment 3 (VSDIIS3) or its associated loops (chimera 1 IC₅₀ = 1.1 nM; 95 % CI: 0.6–1.9 nM; chimera A IC₅₀ = 8.6 nM; 95 % CI: 6.2–––11.84 nM). D. Species alignment of VSDIIS3 transmembrane domain and associated cytoplasmic and extracellular linkers of rat and human Na_V1.8. Alignment yielded a sequence of three amino acids towards the extracellular aspect of the transmembrane domain of VSDIIS3 that were non-homologous between rat and human. E. Genomic localization of the final Na_V1.8 bacterial artificial chromosome clone (BAC-1) used to express the human Na_V1.8 channel in the humanized rat. F. Quantitative RT-PCR against either rat or human Na_V1.8 mRNA showing expression of human Na_V1.8 channel transcript in the humanized Na_V1.8 rat (1.12 ± 0.08; n = 4) with a significant decrease of rat Na_V1.8 channel transcript in humanized Na_V1.8 rat (0.17 ± 0.02; n = 4) and KO (0.14 ± 0.04; n = 4). Data expressed as fold housekeeper (GAPDH) transcript, presented as mean ± SD. G. Custom ELISA targeted against either rat or human Na_V1.8 channel showing expression of the human Na_V1.8 channel in the humanized Na_V1.8 rat (99.40 ± 18.91; n = 4) with a concurrent significant decrease of the rat Na_V1.8 channel in humanized Na_V1.8 rat (23.21 ± 4.35; n = 4) and KO (29.34 ± 3.70; n = 4). RLUs = relative light units. H. Left, representative manual patch clamp traces of TTX-resistant evoked currents from humanized Na_V1.8 rat DRG in the presence of vehicle or increasing concentrations of MSD199. Right, MSD199 is highly potent against TTX-R currents in freshly-isolated humanized Na_V1.8 rat DRG neurons expressing human Na_V1.8 channel (IC₅₀ = 5.60 nM, 95 % CI: 5.00–6.28 nM). Data represent mean ± SD. Cartoon illustrations in panels C and D created via BioRender.

Alignment of rat and human Na_V1.8 VSDIIS3 sequences yielded a non-homologous three amino acid sequence towards the extracellular aspect of the transmembrane domain of VSDIIS3 (745Gly, 746Val, 747Ala in human, 744Ser, 745Ala, 746Ser in rat; Fig. 1D) that may explain the loss of potency observed against the rat Na_V1.8 channel. Interestingly, this three amino acid sequence is directly neighboring a KKGS (Lys748, Lys749, Gly750, Ser751 in human) sequence of the extracellular S3–S4 linker unique to Na_V1.8 which was shown to be important for the binding of two other small molecule Na_V1.8 inhibitors, VX-548 and LTGO-33. Both of which also demonstrate reduced potency in rats, further suggesting this region as a potential driver for the of loss of potency at rat Na_V1.8 channels (Gilchrist et al., 2024, Osteen et al., 2025).

The observed loss of potency against the rodent Na_V1.8 channel constrained further efforts at evaluating MSD199 due to challenges in reaching sufficient exposures to engage the channel in vivo. To overcome this, a rat expressing the human Na_v1.8 channel in lieu of the rodent channel was engineered (see methods section for details). Utilizing quantitative RT-PCR against either rat or human Na_V1.8 mRNA demonstrated human Na_V1.8 channel transcript in the humanized Na_V1.8 rat (1.12 ± 0.08; n = 4) with a concurrent significant decrease of rat Na_V1.8 channel transcript in the humanized Na_V1.8 rat (0.17 ± 0.02; n = 4) and KO (0.14 ± 0.04; n = 4; data expressed as fold housekeeper (GAPDH) transcript). Next, using a custom ELISA targeted against either rat or human Na_V1.8 channel we found expression of the human Na_V1.8 channel in the humanized Na_V1.8 rat (99.40 ± 18.91; n = 4) with a concurrent significant decrease of the rat Na_V1.8 channel in humanized Na_V1.8 rat (23.21 ± 4.35; n = 4) and KO (29.34 ± 3.70; n = 4; relative light units x 1000). The small signal detected in the KO and humanized animals in both qRT-PCR and ELISA is likely to be residual, non-functional truncated protein that arises due to the KO strategy employed. Such a finding was likewise reported by Akopian et al, who used the same KO strategy in successfully generating a functional Na_V1.8 KO mouse line (Akopian et al., 1999). Together, these data suggest successful generation of the humanized Na_V1.8 rat. To next determine if the potency of MSD199 was similar to the human channel potency in these newly developed humanized Na_V1.8 rat, we repeated the above manual patch clamp electrophysiology experiment with primary DRG neurons isolated from the humanized animals. We determined the TTX-resistant IC₅₀ of MSD199 within the humanized Na_V1.8 rat-derived DRG neurons as 5.60 nM (95 % CI: 5.00 – 6.28 nM; n = 5), closely resembling the human Na_V1.8 IC₅₀ of 4.70 nM.

CNB as a moderate throughput in vivo screening assay

Next, with potency of the chemical matter established in humanized Na_V1.8 rat ex-vivo DRG neurons, we aimed to establish an acute pain in vivo PD assay that could be used to determine IVIVC relationships and enable lead optimization and candidate selection. In rodents, expression of the Transient Receptor Potential Vanilloid 1 (TrpV1) ion channel is largely constrained to a subpopulation of DRGs that are nociceptive and co-express the Na_V1.8 channel (Jung et al., 2023, Usoskin et al., 2015) and injection of the TrpV1 agonist capsaicin into the paw of rats results in robust and easily quantifiable nocifensive behaviors (Carey et al., 2017, Carey et al., 2016). Therefore, capsaicin-induced nocifensive behavior (CNB) was chosen to test as an in vivo readout with moderate throughput. Capsaicin injection into the hind paw resulted in robust nocifensive behaviors (Fig. 2). In line with our electrophysiology data, MSD199 at 10 mg/kg had no effect on CNB measured in wildtype rats (unpaired t test, p = 0.7250) but significantly reduced nocifensive behaviors in humanized Na_V1.8 rats (Mann Whitney test, p < 0.0001). Furthermore, in humanized Na_V1.8 rats, we found MSD199 dose-dependently blocked nocifensive behaviors such that responses were significantly lower than vehicle at all tested doses (one-way ANOVA, Bonferroni multiple comparisons test: vehicle vs 1 mg/kg MSD199, p = 0.0066; vehicle vs 3 mg/kg MSD199, p < 0.0001; vehicle vs 10 mg/kg MSD199, p < 0.0001). To determine a PK-PD relationship, responses of individual animal percent inhibition of nocifensive behaviors relative to vehicle control were plotted against their respective MSD199 unbound plasma exposure levels. Following this, the calculated in vivo IC₅₀ of MSD199 in the CNB assay was 6.78 nM (Fig. 2D). Additionally, the effects of two similar yet distinct molecules (MSD869, human Na_V1.8 IC₅₀ = 1.10 nM; MSD537, human Na_V1.8 IC₅₀ = 32.10 nM) within our chemical series was also examined in the CNB assay. Following normalization for each molecule’s plasma protein binding and potency (see Table 1), we found strong overlap within each molecule’s PK-PD curve, demonstrating good IVIVC for this chemical series in CNB assay (Fig. 2E). Due to favorable potency and plasma protein binding, MSD199 was then selected for further testing in two classical rodent models of chronic pain, complete Freund’s adjuvant (CFA)-induced thermal hyperalgesia and spinal nerve ligation (SNL)-induced mechanical allodynia, representing inflammatory and neuropathic pain, respectively.

Fig. 2.

MSD199 blocks capsaicin-induced nocifensive (CNB) behavior. A. In wild type rats, 10 mg/kg MSD199 has no observable impact on CNB (p = 0.725, Student’s T-test, n = 10/group). B. In humanized Na_V1.8 rat (n = 10/group), MSD199 (10 mg/kg) significantly blocked CNB (p < 0.0001, Student’s T-test). C. MSD199 dose-dependently reduced CNB in humanized Na_V1.8 rats, 8–9 rats /treatment group. D. MSD199 unbound plasma concentrations from CNB dose–response (panel C) vs pharmacodynamic response, represented as percent inhibition of the vehicle mean flinch response, with an IC₅₀ = 6.783 nM. E. Mean flinch response percent inhibition vs fold exposure (unbound concentration / in vitro IC₅₀) of MSD199 compared to two other small molecule Na_V1.8 inhibitors (MSD869, MSD537) belonging to the same chemical series, showing strong in vitro – in vivo correlation between Na_V1.8 inhibition and inhibition of CNB. Data represent mean ± SD. Asterisk (**) denotes P < 0.01, (****) denotes P < 0.0001.

MSD199 is efficacious in the CFA model of inflammatory pain

The analgesic efficacy of MSD199 in the CFA model was assessed via the Hargreaves’s method (Hargreaves et al., 1988). Following CFA injection, withdrawal latencies were significantly decreased vs. baseline responses (Fig. 3). At one hour post dosing, 10 mg/kg MSD199 but not 1 mg/kg significantly increased the hind paw withdrawal latency as compared to vehicle controls (10 mg/kg MSD199 vs vehicle, two-way ANOVA, Bonferroni multiple comparisons test, p = 0.0393). Total plasma exposures for 1 and 10 mg/kg treatment groups were 20.10 ± 5.10 nM and 198.74 ± 27.26 nM, respectively, at time of blood collection four-hours post-dosing. MSD199 thermal hyperalgesia reversal at 10 mg/kg was similar in magnitude to the nonsteroidal anti-inflammatory positive control Diclofenac at 80 mg/kg (Diclofenac vs vehicle, two-way ANOVA, Bonferroni multiple comparisons test, p < 0.0001) such that there was no significant difference between 10 mg/kg MSD199 and Diclofenac (Diclofenac vs MSD199, two-way ANOVA, Bonferroni multiple comparisons test, p = 0.6823).

Fig. 3.

Reversal of inflammatory and neuropathic pain by MSD199. A. MSD199 reduces thermal hyperalgesia following Complete Freund’s Adjuvant (CFA) injection. Following CFA injection, humanized Na_V1.8 rats developed thermal hyperalgesia. MSD199 dosed at 10 mg/kg, but not at 1 mg/kg, significantly increased the withdrawal latency to radiant heat stimulus as compared to vehicle controls one-hour post-dosing; 10 mg/kg MSD199 (n = 20) vs vehicle (n = 20), two-way ANOVA, Bonferroni multiple comparisons test, p = 0.0393. In addition, the nonsteroidal anti-inflammatory positive control Diclofenac reversed thermal hyperalgesia at 80 mg/kg; Diclofenac (n = 20) vs vehicle (n = 20), two-way ANOVA, Bonferroni multiple comparisons test, p < 0.0001. There was no significant difference between 10 mg/kg MSD199 and Diclofenac; Diclofenac (n = 20) vs 10 mg/kg MSD199 (n = 20), two-way ANOVA, Bonferroni multiple comparisons test, p = 0.6823). Data represent mean ± SD. Asterisk (*) denotes P < 0.05, (****) denotes P < 0.0001. B. MSD199 reverses mechanical allodynia following spinal nerve ligation (SNL) surgery. Humanized Na_V1.8 rat developed mechanical allodynia one week following SNL surgery. MSD199 dosed at 1 mg/kg showed no significant antiallodynic effects when compared to vehicle controls. 1 mg/kg MSD199 (n = 10) vs vehicle (n = 9), mixed-effects model, Bonferroni multiple comparisons test, p > 0.9999. MSD199 dosed at 10 mg/kg demonstrated significant antiallodynic effects when compared to vehicle controls. 10 mg/kg MSD199 (n = 9) vs vehicle (n = 9), mixed-effects model, Bonferroni multiple comparisons test, p = 0.0301. Data represent mean ± SD. Asterisk (*) denotes P < 0.05.

MSD199 is efficacious in the SNL model of neuropathic pain

It has been previously demonstrated that Na_V1.8 expression increases following spinal nerve ligation in uninjured unmyelinated axons, and knockdown of the Na_V1.8 ion channel reverses mechanical allodynia (Gold et al., 2003, Lai et al., 2002). Therefore, we sought to assess whether MSD199 is analgesic within this model of neuropathic pain by utilizing the Dixon Up-Down method of assessing mechanical allodynia. One week following surgery, paw withdrawal thresholds (PWTs) were significantly reduced from 15 g (cut off value) to approximately 5 g (vehicle group: 4.20 ± 0.44 g; n = 9; MSD199 group: 4.29 ± 0.72 g, n = 10). MSD199 or vehicle was PO dosed and PWTs were measured one-hour post-dosing. MSD199 dosed at 1 mg/kg did not significantly alter paw withdrawal thresholds as compared to vehicle (MSD199 vs vehicle, mixed-effects model, Bonferroni multiple comparisons test, p > 0.9999). However, MSD199 dosed at 10 mg/kg resulted in a significant reversal of mechanical allodynia as compared to vehicle (MSD199 vs vehicle, mixed-effects model, Bonferroni multiple comparisons test, p = 0.0301) demonstrating analgesic efficacy. Total plasma exposures for 1 and 10 mg/kg treatment groups were measured to be 73.06 ± 4.71 nM and 1117 ± 84.19 nM at the time of the blood collection approximately 1-hour post-dosing.

Discussion

There have been many efforts by pharmaceutical companies to develop selective Na_V1.8 channel inhibitors, with only a few compounds progressing into clinical development (Hijma et al., 2021, Hijma et al., 2022, Jarvis et al., 2007, Jones et al., 2023, Payne et al., 2015). One issue that has slowed development is the significant species differences in channel homology observed within the Na_V1.8 ion channel, which often leads to species-specific shifts in molecule potency and presents challenges in conducting in vivo experiments (Gilchrist et al., 2024, Jarvis et al., 2007, Krafte et al., 2007, McGaraughty et al., 2008, Payne et al., 2015). We present here pharmacological activity of novel, potent and selective small molecule Na_V1.8 inhibitors in a humanized Na_V1.8 rat in multiple traditional assays of rodent pain behavior, including clear PK-PD relationships in CNB, an assay with throughput sufficient for discovery compound screening. Importantly, the in vivo experiments conducted within this manuscript would have proved impossible in wild type animals, as reaching sufficient plasma exposure levels to effectively engage the Na_V1.8 ion channel was limited by solubility of the chemical matter. Therefore, utilizing genetic engineering, we developed a transgenic rat in which the rodent Na_V1.8 channel is knocked out and replaced by the human channel. We validated successful transgenic rat generation via qRT-PCR and a custom ELISA assay and confirmed potency in freshly isolated humanized Na_V1.8 rat DRG neurons. We believe this serves as a strong example of when generating humanized animals to pharmacologically engage the human target can overcome species-shifts in potency.

In the CNB model, significant PD impact was observed following administration of 1 mg/kg MSD199. This contrasts with the chronic pain models tested, where significant pharmacodynamic impact was only observed at the higher 10 mg/kg dose despite matching dosing regiments (PO, one-hour post dosing test time). The reasons behind this apparent rightward shift in PD impact relative to PK exposure are unclear, although it may underlie differing fiber types contributing to the pain phenotype. In CNB, capsaicin selectively activates TrpV1-expressing fibers, which in rodents are largely constrained to nociceptors having significant overlap with Na_V1.8 expression (Jung et al., 2023). Additionally, capsaicin-induced activation of c-fiber nociceptors via intraplantar bolus injection may result in high-frequency firing within c-fiber nociceptors (Wooten et al., 2014), of which Na_V1.8 is thought to play a critical role in enabling. Therefore, it is not surprising that this assay would be particularly sensitive to Na_V1.8 inhibition. In contrast, the chronic inflammatory and neuropathic pain models tested here may have ectopic, unsynchronized firing of a lower frequency (Han et al., 2000, Serra et al., 2012, Sun et al., 2005), and thus may require a higher degree of Na_V1.8 inhibition to show efficacy due to other Na_V channels being able to drive action potential firing at lower frequencies. In line with this, we have recently demonstrated that at lower firing frequencies (0.25 – 2 Hz), selective Na_V1.8 inhibition via MSD199 has only modest impact on c-fiber nociceptor firing in the rhesus monkey (Vardigan et al., 2024). Finally, it has been demonstrated that central sensitization of dorsal horn circuitry following SNL contributes to the mechanical allodynia phenotype (Liu et al., 2019). This central sensitization may necessitate a greater degree of Na_V1.8 inhibition when compared to CNB, in which TrpV1-driven nociceptor excitability is relayed via “naïve” second order neurons within the dorsal horn. In support of this, utilizing a selective Na_V1.8 inhibitor, analgesic efficacy was demonstrated at a lower dose in neuropathy and inflammatory pain models when compared to formalin paw phase II, which is believed to be driven in part by central sensitization within the dorsal horn of the spinal cord (Coderre et al., 1990, Payne et al., 2015, Tjolsen et al., 1992).

Utilizing MSD199 here, we demonstrate selective Na_V1.8 inhibition significantly reverses thermal hyperalgesia following CFA injection. This finding is in good agreement with previous literature demonstrating that knockdown of Na_V1.8 via antisense RNA or antisense oligodeoxynucleotide reduced CFA-induced thermal hyperalgesia (Porreca et al., 1999, Yu et al., 2011), and selectively silencing Na_V1.8 positive fibers reduced thermal hyperalgesia following CFA injections in mice (Daou et al., 2016). In contrast, Na_V1.8′s role in preclinical neuropathic pain models has remained less certain, in part because of a lack of available Na_V1.8 inhibitors with good potency at the rodent Na_V1.8 channel and various caveats that are associated with other experimental approaches. It has been shown that Na_V1.8 knockout mice develop thermal and mechanical hypersensitivity comparable to wildtype animals following the partial ligation of the sciatic nerve model of neuropathic pain (Dib-Hajj et al., 1997, Kerr et al., 2001), and furthermore, neuropathic pain develops normally in mice lacking both Na_V1.7 and Na_V1.8 from birth (Nassar et al., 2005). Overall, these findings would suggest a Na_V1.8-independent mechanism of neuropathic pain development. However, these results are confounded by the apparent increase in Na_V1.7 expression observed within Na_V1.8 knockout animals, suggesting compensatory mechanisms within the pain neurocircuitry when genetic manipulation is used (Akopian et al., 1999). And while co-knockout of Na_V1.7 would presumably alleviate this confounding variable, other changes in gene expression within the pain neurocircuitry cannot be ruled out. Opposing these findings, selective knockdown of Na_V1.8 via small interfering or antisense RNA reverses mechanical allodynia following CCI and SNL surgery in rats, respectively (Dong et al., 2007, Porreca et al., 1999). However, these approaches may suffer from off-target effects and variable silencing efficacy. In a different neuropathic model, spared nerve injury (SNI), selectively silencing Na_V1.8 positive fibers via optogenetics reverses SNI-induced mechanical allodynia (Daou et al., 2016). However, the neuronal inhibition produced by an optogenetic approach is likely greater than pharmacological block of Na_V1.8 alone. Ideally, selective pharmacological blockage of Na_V1.8, as demonstrated here, provides direct evidence of Na_V1.8′s role in mediating neuropathic pain.

Our findings are not the first to show pharmacological inhibition of Na_V1.8 reverses pain states in rodents. A previously published selective Na_V1.8 small molecule inhibitor, A-803467, demonstrated reversal of SNL-induced mechanical allodynia among other pain assays in rodents. However, like MSD199, this molecule exhibited reduced potency in wildtype rodents and may have been less selective at the efficacious exposures reached in vivo (Jarvis et al., 2007). More recently, another small molecule Na_V1.8 inhibitor, PF-01247324, reversed neuropathic pain in the rat SNL model as well. In contrast to our findings, reversal of pain-related behaviors of this molecule occurred at unbound plasma exposure levels significantly below the rat IC₅₀ at TTX-R channels, whereas for MSD199 we found greater exposure levels necessary for behavioral reversal (Payne et al., 2015). The reasons behind this are unclear, however, MSD199 does not display frequency dependency, with the calculated IC₅₀ following a train of 10 pulses at 1 Hz (4.3 ± 0.8 nM, n = 3) closely mirroring the IC₅₀ of the first pulse. This contrasts with PF-01247324, which demonstrates frequency dependency, blocking more current as the stimulation rate increases (Payne et al., 2015). As Na_V1.8 is thought to play a role in repetitive firing during pain states (Hameed, 2019, Han et al., 2016, Renganathan et al., 2001), it is possible that frequency dependency would increase a molecule’s effective potency during repetitive firing and thus provide greater efficacy at lower plasma exposures, however more experiments are needed to explore this concept.

A limitation of our current study is the inclusion of only male rats for all experiments. It is understood that biological sex can influence pain sensitivity and perception, analgesic efficacy, and development and prevalence of painful conditions across species (DeLeo and Rutkowski, 2000, Mogil, 2012, Mogil et al., 2024, Niesters et al., 2010, Presto and Neugebauer, 2022, Vierck et al., 2008). It is possible the effects observed here may differ by sex, as previous studies have demonstrated sex differences in responses to neuropathic and inflammatory pain in rats (Boullon et al., 2021, DeLeo and Rutkowski, 2000, Ghazisaeidi et al., 2023, Jones et al., 2025). Therefore, future research with the inclusion of both male and female sex is thus warranted to provide a more comprehensive understanding of the findings and to evaluate whether these results are applicable across sexes.

In total, utilizing potent and selective Na_V1.8 small molecule inhibitors and humanized Na_V1.8 rats, we demonstrate significant efficacy of Na_V1.8 inhibition in CNB, CFA, and SNL preclinical pain models. The preclinical data we present here is in line with proof-of-mechanism clinical data in which selective Na_V1.8 inhibition demonstrates an analgesic profile in small fiber neuropathy patients (NCT03304522) and, in a more recent trial, efficacy in diabetic neuropathic pain (NCT05660538). Overall, these results further establish Na_V1.8 as a key player in the transmission of acute and chronic pain. These results highlight the utility of humanized transgenic animals when cross-species potency shifts are observed within an otherwise promising chemical series. The broader use of humanized animals may improve the success of translational drug discovery efforts toward better and safer pain medications.

CRediT authorship contribution statement

Dillon S. McDevitt: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Joshua D. Vardigan: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Conceptualization. Xiaoping Zhou: Writing – review & editing, Investigation, Formal analysis, Conceptualization. Thomas W. Rosahl: Writing – review & editing, Investigation, Formal analysis, Conceptualization. Heather Zhou: Writing – review & editing, Investigation, Formal analysis. Eric A. Price: Writing – review & editing, Investigation, Formal analysis, Conceptualization. Michelle K. Clements: Writing – review & editing, Investigation, Formal analysis, Conceptualization. Yuxing Li: Investigation. Nissi Varghese: Investigation. Alicja Krasowska-Zoladek: Investigation. Shawn J. Stachel: Writing – review & editing, Resources, Conceptualization. Michael J. Breslin: Writing – review & editing, Resources, Conceptualization. Christopher S. Burgey: Writing – review & editing, Resources, Conceptualization. Richard L. Kraus: Investigation, Methodology, Supervision. Parul S. Pall: Writing – review & editing, Writing – original draft, Investigation, Formal analysis, Conceptualization. Darrell A. Henze: Writing – review & editing, Writing – original draft, Formal analysis, Conceptualization. Vincent P. Santarelli: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Conceptualization.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: All authors reports financial support was provided by Merck & Co Inc. All authors reports a relationship with Merck & Co Inc that includes: employment and equity or stocks. Michael Breslin, Chris Burgey, Shawn Stachel has patent #11377438B2 issued to Merck Sharp & Dohme LLC. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.