Pharmacological Inhibition of the Voltage-Gated Sodium Channel Na_V1.7 Alleviates Chronic Visceral Pain in a Rodent Model of Irritable Bowel Syndrome

Abstract

The human nociceptor-specific voltage-gated sodium channel 1.7 (hNa_V1.7) is critical for sensing various types of somatic pain, but it appears not to play a primary role in acute visceral pain. However, its role in chronic visceral pain remains to be determined. We used assay-guided fractionation to isolate a novel hNa_V1.7 inhibitor, Tsp1a, from tarantula venom. Tsp1a is 28-residue peptide that potently inhibits hNa_V1.7 (IC₅₀ = 10 nM), with greater than 100-fold selectivity over hNa_V1.3–hNa_V1.6, 45-fold selectivity over hNa_V1.1, and 24-fold selectivity over hNa_V1.2. Tsp1a is a gating modifier that inhibits Na_V1.7 by inducing a hyperpolarizing shift in the voltage-dependence of channel inactivation and slowing recovery from fast inactivation. NMR studies revealed that Tsp1a adopts a classical knottin fold, and like many knottin peptides, it is exceptionally stable in human serum. Remarkably, intracolonic administration of Tsp1a completely reversed chronic visceral hypersensitivity in a mouse model of irritable bowel syndrome. The ability of Tsp1a to reduce visceral hypersensitivity in a model of irritable bowel syndrome suggests that pharmacological inhibition of hNa_V1.7 at peripheral sensory nerve endings might be a viable approach for eliciting analgesia in patients suffering from chronic visceral pain.

Irritable bowel syndrome (IBS) is a chronic gastrointestinal disorder characterized by abdominal pain, discomfort, bloating, and altered bowel habits but without abnormalities in gut morphology.^1,2 IBS affects 11–12% of the global population and it has an annual economic cost of $30 billion in the USA alone.^3,4 Visceral pain is the cardinal symptom experienced by IBS patients, with 75% reporting continuous or frequent abdominal pain.³ About 50% of IBS patients suffer from chronic gut hypersensitivity, and the chronic visceral hypersensitivity (CVH) of colonic afferents is implicated in development and maintenance of visceral pain in these patients.⁵ The lack of efficacious treatments for chronic visceral pain is a major contributor to the debilitating nature of IBS and its devastating socioeconomic impact.^4,6

Humans have nine subtypes of voltage-gated sodium (Na_V) channel, denoted Na_V1.1–Na_V1.9, and several are implicated in nociception.^7,8 Many frontline analgesics are nonselective Na_V channel inhibitors, but they typically have a narrow therapeutic window and are unsuited to long-term treatment of chronic pain.⁹ Hence, much effort has been directed toward development of subtype-selective Na_V channel inhibitors, with a particular focus on Na_V1.7 because of genetic studies linking this channel to pain.^8,10 Loss-of-function mutations in SCN9A, the gene encoding Na_V1.7, lead to a congenital indifference to pain, whereas gain-of-function mutations underlie paroxysmal extreme pain disorder (PEPD) and erythromelalgia, which are characterized by episodic bouts of extreme burning pain.¹¹ Consistent with these human genetic studies, gene ablation and pharmacological investigations have confirmed a key role for Na_V1.7 in mediating somatic pain.^12,13 However, it was recently shown that although Na_V1.7 is highly expressed in the majority of colonic sensory afferents, it is not an important contributor to acute visceral pain.¹²

Here we aimed to explore whether Na_V1.7 plays an important role in the chronic abdominal pain associated with IBS.⁵ The complement of Na_V channels in gut nociceptors is known to be altered under these conditions; for example, we recently showed functional upregulation of Na_V1.1 in colonic afferents in a mouse model of IBS.¹⁴ Thus, the relative contribution of Na_V channel subtypes to nociception may differ in chronic and acute visceral pain.¹⁵ To address this issue, we first screened a panel of arachnid venoms, which are the richest known source of Na_V channel modulators,^16,17 to isolate a potent and selective peptide inhibitor of Na_V1.7. Intracolonic administration of this peptide reversed hypersensitivity to noxious colorectal distension in a mouse model of IBS. Our data suggest that Na_V1.7 plays a key role in chronic mechanical hypersensitivity of the gut and that Na_V1.7 inhibitors might be a viable therapeutic option for treating chronic visceral pain.

Results

Discovery and Isolation of μ-Theraphotoxin-Tsp1a

In a search for venom compounds that inhibit hNa_V1.1, we noted that venom from an uncharacterized Peruvian tarantula strongly inhibited the channel, and this venom was therefore chosen for isolation of the active venom component. Morphological characterization of the spider revealed it to be an undescribed species in the genus Thrixopelma (see Supporting Information); we therefore refer to this spider as Thrixopelma spec. (Peru). Crude venom was fractionated using analytical reversed-phase (RP) HPLC, and electrophysiological screening revealed five fractions that inhibited hNa_V1.1 (Figure S2). Fraction F22 (Figure 1A) exhibited selectivity for hNa_V1.1 over hNa_V1.6 (Figure S2); therefore, it was further deconvoluted using hydrophilic interaction chromatography to obtain a pure active peptide (Figure 1B) with monoisotopic mass of 3388.589 Da (Figure 1C). N-Terminal sequencing revealed a 28-residue peptide with sequence YCQKFLWTCDSERPCCEGLVCRLWCKIN and monoisotopic mass of 3388.526 Da, assuming that the six cysteine residues form three disulfide bonds and the C-terminus is amidated (Figure 1D). The predicted mass based on this sequence is an excellent match (0.063 Da difference) with the experimentally determined mass of the native peptide.

Figure 1.

Isolation of μ-TXTR-Tsp1a from Thrixopelma spec. (Peru) spider venom. (A) Chromatogram resulting from fractionation of crude venom using C18 RP-HPLC. Fraction 22 (F22) containing Tsp1a is colored red and highlighted by an arrow. (B) Chromatogram resulting from HILIC fractionation of F22. The peak containing pure Tsp1a is colored red and highlighted by an arrow. The inset is a MALDI-TOF MS spectrum of purified native Tsp1a showing the [M + H]⁺ ion. (C) Amino acid sequence of Tsp1a obtained via Edman degradation. NH₂ indicates C-terminal amidation. (D) Alignment of the sequence of Tsp1a with spider venom peptides that exhibit at least 70% sequence identity. Identical noncysteine residues are indicated in green bold type; conserved cysteines are highlighted in orange. Asterisks indicate C-terminal amidation. The percent identity relative to Tsp1a, UniProt accession codes, and reported molecular target of each peptide are indicated. MSC = mechanosensitive channels.

The peptide was named μ-theraphotoxin-Tsp1a (hereafter Tsp1a) based on the rational nomenclature for spider-venom peptides.¹⁸ The μ prefix denotes Na_V channel inhibition, “theraphotoxin” indicates the toxin is from a spider in the Theraphosidae family, and the initials “Tsp” denote the genus/species. A BLAST search of the Tsp1a sequence against the ArachnoServer¹⁹ and UniProt databases indicated that Tsp1a is novel but has similarity to spider venom peptides in the NaSpTx family 3 (NaSpTx3) (Figure 1E). Many NaSpTx3 peptides inhibit Na_V channels, including β/ω-TRTX-Tp2a (ProTx-II) and β-TRTX-Gr1b (GsAFI) that potently inhibit hNa_V1.7 with IC₅₀ values of 0.3 and 40 nM, respectively.²⁰⁻²² Tsp1a is 84% identical to these peptides, suggesting that it might also inhibit hNa_V1.7.

Production of Synthetic and Recombinant Tsp1a

Chemical synthesis was employed to produce Tsp1a with C-terminal amidation (sTsp1a). A single peak was obtained following analytical RP-HPLC purification, with the area under the peak indicating purity of 96% (Figure 2A). Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) yielded a monoisotopic mass of 3388.589 Da, which corresponds to the mass of C-terminally amidated Tsp1a with three disulfides (Figure 2A, inset).

Figure 2.

Production of synthetic and recombinant Tsp1a. (A) Chromatogram showing final analytical RP-HPLC purification of synthetic Tsp1a (sTsp1a). The arrow indicates the peak containing sTsp1a. Inset: MALDI-TOF MS spectrum showing the [M + H]⁺ ion for sTsp1a (monoisotopic mass 3388.589 Da). (B) RP-HPLC chromatogram showing the final analytical RP-HPLC purification of recombinant Tsp1a (rTsp1a). Arrow indicates peak containing rTsp1a. Inset is a MALDI-TOF MS spectrum showing the [M + H]⁺ ion for rTsp1a (monoisotopic mass 3476.614 Da). (C) Chromatograms showing elution of native Tsp1a (black), sTsp1a (red), and rTsp1a (green) under identical conditions from an analytical PEPTIDE XB-C18 column. (D) Primary structure of native Tsp1a, sTsp1a, and rTsp1a. The disulfide connectivities, as determined using NMR spectroscopy (Figure 5), are shown above the sequence alignment.

Recombinant Tsp1a (rTsp1a) was produced by expression as a His₆-maltose binding protein (MBP) fusion protein in the periplasm of Escherichia coli.²³ A single peak was obtained after analytical RP-HPLC purification, indicating that rTsp1a forms a single isomer following liberation from the His₆-MBP tag with tobacco etch virus (TEV) protease (Figure 2B). The yield of rTsp1a was ∼50 μg·L^–1 cell culture. An additional non-native serine residue was introduced at the N-terminus of rTsp1a to assist with TEV protease cleavage of the fusion protein (Figure 2D), as Ser is a preferred residue in the P1′ position of the TEV protease recognition site, unlike the native Tyr.²⁴ The monoisotopic mass of rTsp1a measured using MALDI-TOF MS (3476.614 Da) is consistent with an additional Ser at the N-terminus and a nonamidated C-terminus (Figure 2B, inset).

To determine whether sTsp1a and rTsp1a adopt the native tertiary fold (and thus, by inference, the native disulfide connectivity), they were eluted from an analytical RP-HPLC column with the same solvent gradient as native Tsp1a (Figure 2C). sTsp1a and native Tsp1a eluted with identical retention time, indicating that they adopt the same fold. Despite the additional N-terminal Ser residue and the lack of C-terminal amidation, rTsp1a eluted with the same retention time as sTps1a and native Tsp1a. We therefore conclude that rTsp1a very likely adopts the native three-dimensional (3D) fold.

sTsp1a Inhibits hNa_V1.7 More Potently than Other Na_V Subtypes

Because sTsp1a is identical to native Tsp1a, it was used to determine activity on various hNa_V subtypes using patch-clamp electrophysiology. Analysis of concentration–response data revealed that sTsp1a most potently inhibits hNa_V1.7 (pIC₅₀ 7.99 M; IC₅₀ 10.3 nM) with 45-fold selectivity over hNa_V1.1 (pIC₅₀ 6.34 M; IC₅₀ 452 nM), 24-fold selectivity over hNa_V1.2 (maximum current inhibition of 60% at 2 μM; pIC₅₀ 6.61 M; IC₅₀ 245 nM), and >100-fold selectivity over hNa_V1.3–hNa_V1.6 and Na_V1.8 (maximum current inhibition < 50% at 2 μM) (Figure 3A). Saturating concentrations of sTsp1a did not yield complete block of hNa_V1.7; at 200 nM, sTsp1a inhibited 70% of the hNa_V1.7 current and 30% of hNa_V1.1 and hNa_V1.2 currents. It did not significantly inhibit hNa_V1.3 to hNa_V1.6 (Figure 3C).

Figure 3.

Inhibitory activity of sTsp1a on hNa_V1.1–hNa_V1.8. (A) Concentration–response curves for inhibition of hNa_V1.1–hNa_V1.8 by sTsp1a. Data points are mean ± SEM from n = 5 for Na_V1.1–Na_V1.6, n = 8 for Na_V1.7, and n = 3 for Na_V1.8. (B) pIC₅₀ values for inhibition of hNa_V1.1, hNa_V1.2, and hNa_V1.7 by sTsp1a. The peptide inhibits hNa_V1.7 with 45-fold selectivity over hNa_V1.1 and 24-fold selectivity over hNa_V1.2. *, p < 0.05 as determined using an unpaired t-test with Welch’s correction. (C) Representative traces for hNa_V currents in the absence (black) and presence of 200 nM sTsp1a (red).

sTsp1a had activity very similar to that of native Ts1pa, with no significant difference in pIC₅₀, whereas rTsp1a, which lacks C-terminal amidation and has an additional non-native N-terminal Ser residue, was 20-fold less potent than sTsp1a and native Tsp1a (Figure S3). Given the functional importance of C-terminal amidation for other spider venom peptides that inhibit hNa_V1.7,²⁵⁻²⁸ we conclude that it is most likely the lack of C-terminal amidation rather than the additional N-terminal Ser residue that leads to the lower potency of rTsp1a.

sTsp1a Is a Gating Modifier That Stabilizes the Inactivated State of hNa_V1.7

The mechanism by which sTsp1a inhibits hNa_V1.7 was determined using patch-clamp electrophysiology. We used a nonsaturating concentration of sTsp1a (60 nM) that reduced hNa_V1.7 currents by >50%, but it did not completely inhibit the channel. Conductance–voltage relationships revealed that sTsp1a causes a statistically insignificant depolarizing shift in the V_0.5 of activation (Figure 4A). In contrast, the current–voltage relationship for steady-state inactivation revealed that sTsp1a induces a hyperpolarizing shift in the V_0.5 of inactivation (ΔV_0.5 = −11.13 mV; Figure 4A). sTsp1a also decreased the rate of recovery from fast inactivation (Δτ = 1 ms; Figure 4B). The inhibitory activity of sTsp1a on hNa_V1.7 is only slowly reversible as less than 50% of the current was recovered after ∼15 min washout in a peptide-free solution (Figure 4C). We conclude that sTsp1a is a gating modifier that most likely targets one or more of the voltage-sensor domains (VSDs) to allosterically modulate hNa_V1.7 activity.

Figure 4.

Mechanism of action of sTsp1a on hNa_V1.7. (A) Left panel: Conductance–voltage (G–V) relationship before (black) and after (red) addition of 60 nM sTsp1a. Currents were evoked using the voltage protocol shown in the inset: From a holding potential of −120 mV, the voltage was stepped from −80 to + 60 mV in 5 mV increments. The V_0.5 for activation was −16.0 ± 0.2 mV in the absence and −12.0 ± 2.3 mV in the presence of sTsp1a (n = 5; mean ± SEM). Middle panel: Steady-state inactivation before (black) and after (red) addition of 60 nM sTsp1a. Currents were evoked using the voltage protocol shown in the inset: A depolarization to 0 mV was followed by step potentials from −120 mV to 20 mV in 5 mV increments. The V_0.5 inactivation was −67.7 ± 1.0 mV in the absence and −79.2 ± 1.1 mV in the presence of 60 nM sTsp1a (n = 5; mean ± SEM). Right panel: Comparison of V_0.5 of activation and inactivation in the absence (gray) and presence (red) of 60 nM sTsp1a (n = 5; mean ± SEM). Statistical significance was determined using a paired t-test: n.s., not significant; *, p < 0.05. (B) Recovery from fast inactivation before (black) and after (red) addition of 60 nM sTsp1a. Currents were evoked from a holding potential of −120 mV by applying a 0 mV pulse followed by repolarization to −120 mV. A second pulse to 0 mV was applied after a period ranging from 0 to 50 ms, increasing in 1 ms increments. Data are mean ± SEM; n = 5. *, p < 0.05 based on paired t-test. (C) hNa_V1.7 currents (I/I_max) following application of 200 nM sTsp1a for 4 min followed by washout for 12 min. Depolarizing pulses were applied from −120 to 0 mV every 6 s. The holding potential was −80 mV. (D) hNa_V1.7 currents during repeated pulsing at either 0.1 Hz (left panel) or 40 Hz (right panel) using the illustrated voltage protocol in the absence (black) and presence (red) of 30 nM sTsp1a. (E) Pulsing at 40 Hz led to the accumulation of inactivated channels but did not enhance sTsp1a inhibition of hNa_V1.7 currents relative to control cells (mean ± SEM, n = 7 at 0.1 Hz and n = 4 at 40 Hz). Statistical significance was determined using ordinary one-way ANOVA with multiple comparisons. ****, p < 0.0001. (F) Inhibitory potency of sTsp1a evaluated in the presence of partially inactivated channels induced by pulsing at the V₅₀ of steady-state inactivation. Fitting a Hill equation to the data yielded IC₅₀ = 64.0 ± 3.6 mV (n = 3). (G) At a concentration of 1 μM, sTsp1a potency was slightly enhanced at partially inactivated channels compared to resting/closed state channels, but the difference was not statistically significant (n = 3; t-test with Welch’s correction). Representative peak current amplitudes for these experiments are shown in panel (H). Application of 1 μM tetrodotoxin (TTX) at the end of the experiment yielded full current block. The activity on partially inactivated hNa_V1.7 channels was tested during the periods indicated by black dashed lines, while resting state activity was tested during the periods indicated by gray dotted lines. Data are mean ± SEM from 3 independent experiments.

To test whether sTsp1a inhibition of hNav1.7 is use-dependent, we examined the impact of high-frequency depolarizations (i.e., pulses shorter than the rate of recovery from inactivation) that lead to accumulation of partially and fully inactivated channels (Figure 4D). At a pulse frequency of 40 Hz, sTsp1a superficially appears to be a more potent inhibitor of hNa_V1.7 (Figure 4E), but the level of current inhibition, relative to the control, was not enhanced compared to that of 0.1 Hz stimulation (Figure 4E). High-frequency pulsing did not lead to a cumulative increase in channel inhibition relative to the control, a feature that is characteristic of use-dependent small-molecule Na_V channel inhibitors such as lidocaine,²⁹ chlorpromazine,³⁰ and vixotrigine.³¹ When applied to partially inactivated channels, sTsp1a was in fact a less potent inhibitor (IC₅₀ = 64.0 ± 3.6 nM; n = 3; Figure 4F) than when applied to closed/resting state channels (IC₅₀ = 10.3 nM; Figure 3). Thus, sTsp1a potency is not enhanced in a use-dependent manner.

NMR Structure of sTsp1a

The 3D structure of sTsp1a was elucidated using NMR spectroscopy (Figure 5A; Table 1). The three disulfide bonds form a classical inhibitor cystine knot (ICK), or knottin, motif^32,33 in which the Cys2–Cys16 and Cys9–Cys21 disulfides along with the intervening sections of the peptide backbone form a loop that is pierced by the Cys15–Cys25 disulfide in a similar fashion to ProTx-II³⁴ (compare Figure 5A,B). There are no clearly defined β sheets or α helices in the 3D structure of sTsp1a.

Figure 5.

3D structure of sTsp1a and comparison with ProTx-II. (A) Ensemble of 20 structures chosen to represent the solution structure of Tsp1a (PDB 7A64). Disulfide bonds (C2–C16, C9–C21, and C15–C25) are shown as orange tubes, and the backbone is shown in gray. (B) Comparison of the structures of sTsp1a (gray backbone with orange disulfides) and ProTx-II (cyan backbone with yellow disulfides; PDB 2N9T). The core ICK disulfides of these two peptides overlay closely. The poorly aligned C16–C21 and C2–C9 loops are indicated by arrows in the left and right panels, respectively. (C) Comparison of the molecular surfaces of sTsp1a (top) and ProTx-II (bottom). The pharmacophore of ProTx-II is highlighted in green (hydrophobic residues) and blue (positively charged residues). Corresponding residues in sTsp1a are shown in the same colors, while ProTx-II pharmacophore residues that are different in Tsp1a (F5, L6, and I27) are highlighted in dark gray.

Table 1. Structural Statistics for the Final Ensemble of 20 sTsp1a Structures.

Energies (kcal/mol)a
overall	–825.6 ± 21.0
bonds	14.3 ± 1.6
angles	42.3 ± 3.4
improper	19.8 ± 2.1
dihedral	142.6 ± 2.1
van der Waals	–121.7 ± 6.7
electrostatic	–927.6 ± 23.7
NOE	0.12 ± 0.04
constrained dihedral (cDih)	4.55 ± 1.0

MolProbity Statisticsa
clash score (>0.4 Å/1000 atoms)	16.7 ± 6.6
poor rotamers (%)	0.4 ± 1.2
Ramachandran outliers (%)	0 ± 0
Ramachandran favored (%)	97.7 ± 4.1
MolProbity score	1.88 ± 0.25
MolProbity percentile	80.6 ± 12.4b

Atomic RMSD (Å)a
mean global backbone (2–25)d	0.57 ± 0.17
mean global heavy (2–25)d	1.30 ± 0.19
mean global backbone (1–28)	0.68 ± 0.16
mean global heavy (1–28)	1.71 ± 0.27

Distance Restraints
intraresidue (i – j = 0)	124
sequential (|i – j| = 1)	135
medium range (|i – j| < 5)	76
long range (|i – j| > 5)	77
hydrogen bondsc	2
total	414
Dihedral Angle Restraints
ϕ	20
ψ	18
χ1	2
total	40
Violations of Experimental Restraints
total NOE violations exceeding 0.3 Å	0
total dihedral violations exceeding 3°	2

^aMean ± SD

^bThe 100th percentile is the best among structures of comparable resolution.⁷⁴

^cTwo restraints were used per hydrogen bond.

^dThis alignment excludes the disordered N- and C-terminal regions.

It is informative to compare the structures of sTsp1a and ProTx-II, since in addition to being the most potent known inhibitor of hNa_V1.7, ProTx-II is the Na_V1.7 inhibitor with highest similarity to Tsp1a (84% identity) and the only peptide similar to Tsp1a for which a 3D structure of the complex with Na_V1.7 is available.³⁵ Overall, the structures of sTsp1a and ProTx-II are very similar, but the Cys16–Cys21 and Cys2–Cys9 loops (arrowed in Figure 5B) do not align well despite the high level of sequence homology. ProTx-II also contains a longer and more dynamic C-terminal region than sTsp1a (Figure 5B). The crystal structure of the Na_V1.7/Na_VAb-ProTx-II complex revealed that Trp5, Met6, Trp7, Trp24, Lys26–28, and Trp30 are key residues for anchoring ProTx-II to the DII VSD of hNa_V1.7, enabling ProTx-II to insert residues Arg22 and Lys26 into the DII VSD and act as an electrostatic gating modifier.³⁵ Tsp1a also contains Trp7, Arg22, Trp24, and Lys26, while the Trp5 and Met6 in ProTx-II are replaced by similarly hydrophobic Phe5 and Leu6 residues in Tsp1a. However, Tsp1a lacks the key C-terminal residues in ProTx-II (Lys27–28 and Trp30), which might explain why its potency on hNa_V1.7 (IC₅₀ 10 nM) is less than that of ProTx-II (IC₅₀ 0.3 nM)^20,21 (Figure 5C). Previous mutagenesis studies showed that residues Met6, Trp7, Arg13, Val20, Arg22, Trp24, Ly27, Leu29, and Trp30 form the ProTx-II pharmacophore for inhibition of hNa_V1.5.³⁶ As mentioned above, Tsp1a lacks two of the key C-terminal pharmacophore residues (Leu29 and Trp30), and residues Met6 and Lys27 are replaced with Leu6 and Ile27 in Tsp1a (Figure 5C). These differences likely explain why hNa_V1.5 is insensitive to Tsp1a.

sTsp1a is Highly Stable in Human Serum and Simulated Gastric Fluid

Venom peptides that adopt an ICK motif are often highly resistant to thermal, chemical, and proteolytic degradation.³⁷⁻³⁹ Given the promising in vitro activity of sTsp1a against hNa_V1.7, it was important to understand its stability in biological fluids prior to in vivo testing. Thus, we first examined the stability of sTsp1a in human serum in comparison to those of the analgesic drug ziconotide, an ICK peptide that is very stable in serum and cerebrospinal fluid,⁴⁰ and human atrial natriuretic peptide (hANP), which has a short half-life in human serum.⁴¹ Remarkably, there was no observable degradation of sTsp1a after 24 h in human serum at 37 °C, whereas only 50% of ziconotide remained intact after 24 h. hANP was fully degraded within the first 1–2 h (Figure 6A,C). Similarly, sTsp1a was remarkably stable at 37 °C in simulated gastric fluid (pepsin, pH 1.86), with a half-life of more than 24 h (Figure 6B,D).

Figure 6.

Biological stability of sTsp1a. (A) Stability of sTsp1a, ziconotide and hANP at 37 °C in human serum. The right panel shows data for the full 24 h of incubation while the left panel is an expansion of data for the first hour. Data are mean ± SEM, n = 3. (B) Stability of sTsp1a at 37 °C in simulated gastric fluid (pepsin, pH 1.86). Data are mean ± SEM, n = 5. (C, D) RP-HPLC chromatograms showing elution of sTsp1a samples following incubation at 37 °C in (C) human serum and (D) simulated gastric fluid for 0, 4, 8, and 24 h.

Effect of sTsp1a on Chronic Visceral Hypersensitivity

Noxious distension of the colorectum triggers the visceromotor response (VMR), a nociceptive brainstem reflex that leads to contraction of the abdominal muscles. Thus, we examined the analgesic effect of sTsp1a in a mouse model of CVH by using abdominal electromyography (EMG) to monitor the VMR, which allows assessment of visceral sensitivity in vivo in fully awake animals.⁴² On the day of VMR assessment, mice were briefly anaesthetized using isoflurane before receiving a 100 μL enema of vehicle (sterile saline) or sTsp1a (200 nM). As shown in Figure 7A, CVH animals treated with vehicle (CVH + Veh) exhibited hyperalgesia, characterized by increased sensitivity to noxious distensions (≥50 mmHg) compared to the effect in healthy control mice (HC + veh). In this model of chronic visceral pain, a single intracolonic treatment with sTsp1a significantly reduced the VMR to colorectal distension in CVH mice, normalizing the responses to those of healthy control mice (Figure 7B,C). The colonic compliance of CVH mice was not altered by sTsp1a treatment compared to that of CVH mice receiving vehicle (Figure 7D), suggesting that the reduced VMR to colorectal distension induced by sTps1a in CVH mice is not due to changes in smooth muscle function. In addition, intracolonic treatment with sTsp1a did not reduce the VMR to colorectal distension in healthy mice (Figure S4), suggesting differences in the role of Na_V1.7 in signaling visceral pain between healthy and CVH states.

Figure 7.

In vivo intracolonic administration of sTsp1a reverses CVH in a mouse model of IBS. (A) Raw EMG traces showing the VMR to colorectal distension in healthy mice treated with vehicle (HC + Veh), CVH mice treated with vehicle (CVH + Veh), and CVH mice treated with 200 nM sTsp1a (CVH + sTsp1a). (B) VMR response was determined by measuring the area under the curve (AUC) of EMG recordings at each distension pressure (20–80 mmHg). The VMR to CRD was significantly increased in CVH mice treated with vehicle (red curve; n = 10, mean ± SEM) compared to that in healthy control mice treated with vehicle (black curve; n = 10, mean ± SEM), particularly at distensions of ≥50 mmHg (*, p < 0.05). Intracolonic administration of sTsp1a (200 nM) significantly reduced the VMR in CVH mice (blue curve; n = 8, mean ± SEM), normalizing the responses to those seen in healthy mice. (C) Group data expressed as the total area under the curve (AUC; sum of VMR at all distensions) reveals that CVH mice treated with vehicle have a significantly elevated VMR compared to that in healthy mice (red bar vs black bar; *, p < 0.05). Intracolonic treatment with sTsp1a (200 nM) was able to reverse the VMR in CVH mice to healthy control levels (blue vs red bars; *, p < 0.05). (D) In CVH mice, colonic compliance was not altered by intracolonic administration of sTsp1a relative to intracolonic administration of vehicle. A small but significant difference in compliance was observed between vehicle-treated healthy and CVH mice at a distension pressure of 80 mmHg (*p < 0.05).

Systemic Tsp1a Is Not Toxic

To examine whether systemic Tsp1a affects the respiratory musculature we monitored oxygen saturation in the peripheral blood of mice before and after injection of either phosphate-buffered saline (PBS) or sTsp1a at a dose of 7 nmol/mouse (∼0.95 mg/kg) (Figure 8A). The percentage oxygen saturation preinjection was 89.7 ± 7.5, and postinjection (p.i.) it was 90.3 ± 6.0 (15 s p.i.), 89.2 ± 7.4 (5 min p.i.), and 90.3 ± 5.6 (15 min p.i.). There was no difference between pre- and postinjection data points (p values were 0.62, >0.99, and 0.81, respectively). Moreover, no difference was observed between mice injected with sTsp1a and control mice (p = 0.84 for preinjection, and p > 0.99, 0.57, and 0.31 at 15 s, 5 min, and 15 min p.i., respectively).

Figure 8.

Pharmacodynamic effects of sTsp1a. (A) Continuous monitoring of oxygen saturation (%), heart rate (beats min^–1, bpm), and core temperature of mice before, and 15 s, 5 min, and 15 min after injection of sTsp1a (orange) or PBS (blue). Mice were anesthetized with isoflurane and placed on a heated pad before sTsp1a was injected through a catheter placed into the tail vein (7 nmoles Tsp1a in 100 μL of PBS, which equates to a dose of ∼0.95 mg/kg). No statistically significant changes were observed in these parameters compared to preinjection values, suggesting that sTsp1a does not cause acute side effects at pharmacologically relevant concentrations. Error bars are mean ± SD, n = 5. For oxygen saturation and core temperature, the red dotted lines represent literature baseline values. For heart rate, the dotted lines represent physiologic minima and maxima. (B) Representative EKG recording (lead III) of a mouse before (blue) and 15 s, 5 min, and 15 min after injection of sTsp1a (orange). No significant changes were observed in the EKG pattern suggesting that the electrical activity of the heart was not perturbed by injection of sTsp1a (n.s., p > 0.05).

No changes were observed in the electrocardiogram (ECG) following Tsp1a injection (Figure 8B), suggesting that it does not perturb electrical activity of the heart. We also monitored heart rate pre- and postinjection. The average heart rate before injection was 388 ± 83 bpm, and it was unchanged following injection of sTsp1a (395 ± 82, p = 0.88 at 15 s; 387 ± 89, p = 0.63 at 5 min; and 409 ± 91, p = 0.99 at 15 min p.i.). Although the mean heart rate trended slightly upward in the treated animal cohort (409 ± 91 bpm and 429 ± 80 bpm at 15 min p.i. for injected and control animals, respectively), there was no significant difference in heart rate between sTsp1a and control groups (p = 0.56 in the preinjection group, and p = 0.84, 0.39, and 0.73 in the 15 s, 5 min, and 15 min p.i. groups, respectively).

In parallel, we continuously monitored body temperature, keeping the heated platform constant at 39 °C. The average core temperature of mice before sTsp1a injection was 34.3 ± 0.9 °C. There was no difference in core temperature after injection (34.5 ± 0.63 °C, p = 0.63 at 15 s p.i.; 34.6 ± 0.66 °C, p > 0.99 at 5 min p.i.; 34.6 ± 0.64 °C, p > 0.99 at 15 min p.i.) and no difference between sTsp1a-injected mice and controls (p > 0.99 preinjection; p = 0.42, 0.57, 0.55 at 15 s, 5 min, 15 min p.i., respectively).

Discussion

Tsp1a Is a Potent and Selective Inhibitor of Na_V1.7

The primary structure of Tsp1a is similar to other members of the NaSpTx3 family of spider-venom peptides that target Na_V channels.¹⁶ In particular, the sequence of Tsp1a is 84% identical to that of ProTx-II, the most potent hNa_V1.7 inhibitor reported to date.²¹ Thus, it is not surprising that Tsp1a potently inhibits hNa_V1.7 even though it was discovered in a screen against hNa_V1.1. Although some gating modifier peptides such as ProTx-II and ProTx-III are reported to be selective for hNa_V1.7, most of these peptides also potently inhibit other Na_V channel subtypes.^21,27,43 These off-target effects are a therapeutic liability since it is critical for a Na_V1.7-directed analgesic to avoid effects on hNa_V1.4 in skeletal muscle, hNa_V1.5 in the heart, and hNa_V1.6 in both the peripheral and central nervous systems. As a result, ProTx-II is lethal to rats when administered intravenously at ≥1 mg/kg or intrathecally at ≥0.1 mg/kg.²¹ In contrast to other venom peptides that inhibit hNa_V1.7, with the notable exception of Pn3a,⁴³ Tsp1a is highly selective for Na_V1.7, with a selectivity of >100-fold over hNa_V1.3–hNa_V1.6 and hNa_V1.8, 45-fold over hNa_V1.1, and 24-fold over hNa_V1.2.

It is interesting to compare the potency and selectivity of Tsp1a with venom-derived peptides that have been engineered to have improved potency and selectivity for hNa_V1.7. Tsp1a is 66% identical to JNJ63955918, an analogue of ProTx-II designed to have improved selectivity for Na_V1.7.⁴⁴ The two peptides have almost identical potency on hNa_V1.7 (IC₅₀ ∼ 10 nM) and have similarly low activity at the off-target subtypes Na_V1.4 and Na_V1.5. However, Tsp1a has a potency considerably lower than that of JNJ63955918 at Na_V1.6, which we consider to be the key off-target subtype for an intracolonically or orally delivered gut-acting analgesic due to the ubiquitous expression of Na_V1.6 in motor neurons. Tsp1a has higher potency at Na_V1.1 (IC₅₀ ∼ 450 nM) than does JNJ63955918 (IC₅₀ ∼ 2.5 μM⁴⁴), but we do not consider this to be a therapeutic liability as Na_V1.1 is highly expressed in colonic sensory neurons, and we showed previously that it plays a role in chronic visceral mechanical hypersensitivity and may itself be an analgesic target for visceral chronic pain.^14,45JNJ63955918 does have a lower potency than that of Tsp1a at Na_V1.2, but this is not a key off-target subtype as Na_V1.2 is found primarily in the CNS where it is the most abundant Na_V isoform.⁴⁶

GpTx-1 is a spider venom peptide that falls within the NaSpTx1 family,¹⁶ and it has only 35% sequence identify with Tsp1a. It inhibits hNa_V1.7 with a potency similar to those of Tsp1a and JNJ63955918 (IC₅₀ = 10 nM), but its activity on the muscle subtype Na_V1.4 (IC₅₀ 200 nM) is a serious therapeutic liability.⁴⁷ A positional alanine scan of GpTx-1 followed by more focused analoguing led to the identification of an analogue with very high potency at Na_V1.7 (IC₅₀ 1.6 nM) and much lower potency at Na_V1.4 (IC₅₀ = 1.9 μM).⁴⁷ However, the effects of this analogue on the key off-target subtype Na_V1.6 were not reported. It is also unclear whether this analogue retains the high activity of the parent compound on Na_V1.3 (IC₅₀ ∼ 70 nM⁴⁷).

Tsp1a Has a Novel Mechanism of Action

In addition to its high subtype selectivity, Tsp1a inhibits hNa_V1.7 via a different mechanism than that of most other Na_V channel inhibitors.⁸ Venom-derived Na_V channel inhibitors fall into two broad classes: pore blockers, such as the μ-conotoxins from fish-hunting cone snails, physically occlude the channel pore, while gating modifiers allosterically modulate channel activity by binding to one or more of the VSDs in order to trap them in a particular conformational state.⁴⁸ Gating modifiers that inhibit Na_V1.7 typically target the VSD in channel domain II (VSDII) and induce a depolarizing shift in the voltage dependence of channel activation, whereas those that activate Na_V1.7 either bind to VSDIV and inhibit channel inactivation or bind to VSDII and trap the channel in an activated state.^16,48,49 In contrast, Tsp1a inhibits hNa_V1.7 by stabilizing the channel in its inactivated state, leading to a hyperpolarizing shift in the voltage dependence of steady-state inactivation and slower recovery from fast inactivation, whereas the voltage dependence of activation is not significantly altered (Figure 4). This mechanism of action is quite different from that of the ProTx-II analogue JNJ63955918, which is 66% identical to Tsp1a. JNJ63955918 is a closed-state inhibitor of Na_V1.7 that causes a significant depolarizing shift in the voltage dependence of channel activation without significantly impacting inactivation.⁴⁴ However, unlike small molecules that preferentially bind to the inactivated state of Na_V channels,²⁹⁻³¹ Tsp1a inhibition of Na_V1.7 is not use dependent.

Functional⁵⁰ and structural⁵¹ studies indicate that ProTx-II interacts with both VSDII and VSDIV of hNa_V1.7. Given that Tsp1a has 84% sequence identity with ProTx-II, it is possible that Tsp1a also targets both of these VSDs. However, the fact that Tsp1a affects channel inactivation but not activation suggests that it might primarily target VSDIV. Interestingly, several small molecules interact with VSDIV of Na_V channels and induce a hyperpolarizing shift in the voltage dependence of slow inactivation.⁵²

Although sTsp1a and ProTx-II have 84% sequence identity and both potently inhibit hNa_V1.7, their selectivity profiles and mechanisms of action are quite different. Mutation of the positively charged K27 to glutamine and the two hydrophobic residues L29 and W30 to alanine at the C-terminus of ProTx-II cause a decrease in the toxin’s affinity for hNa_V1.5 of ∼20- to 100-fold.³⁶ While the lack of these residues in Tsp1a possibly explains its inactivity on hNa_V1.5, the C-terminus of Tsp1a is also somewhat truncated compared with other NaSpTx3 peptides (Figure 1D). We therefore propose that residues (or lack thereof) at the C-terminus of Tsp1a are critical for its remarkable selectivity, consistent with C-terminal amidation increasing Tsp1a potency against Na_V1.7 by >20-fold (Figure S3).

Tsp1a Implicates Na_V1.7 in Maintenance of Chronic Visceral Hypersensitivity

Although it has proven difficult to pharmacologically recapitulate the pain-free phenotype experienced by individuals with congenital indifference to pain, gene ablation and some pharmacological interventions have unequivocally demonstrated a role for Na_V1.7 in somatic pain.^12,53,54 In contrast, the role of Na_V1.7 in visceral pain is less clear. However, in a recent seminal investigation, Na_V1.7-knockout mice and pharmacological blockade of Na_V1.7 were used to show that Na_V1.7 does not play a pivotal role in the visceral pain evoked by acute noxious stimuli, even though it is expressed in most colonic afferents.¹² This is perhaps not surprising: colonic nociceptors differ from their somatic counterparts in morphology, extent of myelination, varicosity, and their complement of channels and receptors,^55,56 and many stimuli that are noxious to the skin (e.g., mild heat, cold, and shear stress) do not evoke colorectal visceral pain.² In contrast, it is hard to reconcile this observation with the rectal pain associated with the gain-of-function mutations in Na_V1.7 underlying PEPD. However, PEPD is a chronic pain disorder, and it is entirely possible that Na_V1.7 plays an important role in chronic visceral hypersensitivity while not being pivotal for the nociceptive response to acute noxious stimuli. We decided to test this hypothesis using a mouse model of IBS.

Tsp1a did not affect the responses evoked by noxious colorectal distension in healthy control mice, consistent with previous work demonstrating that Na_V1.7 does not contribute to acute visceral pain.¹² In our model of CVH, mice develop chronic visceral mechanical hypersensitivity but display normal gut morphology and have increased sensitivity to pain evoked by colorectal distension, similar to humans with IBS.^42,57 Remarkably, intracolonic administration of a single bolus dose of Tsp1a reverted the visceromotor response in CVH mice to that of healthy controls, without affecting colonic compliance. Importantly, at the concentration of Tsp1a employed in these experiments, we would not expect significant inhibition of any Na_V channel subtype aside from Na_V1.7 as only small amounts of peptide would be expected to breach the intestinal epithelium and target sensory neurons. Thus, our data strongly implicate Na_V1.7 in the chronic visceral pain associated with IBS. However, this does not exclude the possibility that other Na_V channels might be involved in the generation of chronic visceral pain. Indeed, we recently demonstrated that inhibition of Na_V1.1 alleviates CVH in the same rodent model of IBS.⁴⁵ Thus, the simultaneous blockade of multiple Na_V channel isoforms might prove to be a better therapeutic strategy for alleviating chronic visceral pain than inhibition of a single Na_V channel subtype.

Is It Feasible to Develop an Orally Active Analgesic Peptide for Gut Pain?

Surprisingly, Tsp1a was active via intracolonic administration. Although Tsp1a is resistant to proteolytic degradation and thus may survive intact in the intestinal lumen for an extended period of time, a peptide of this size would not be expected to breach the intestinal epithelium in order to reach the peripheral endings of colonic sensory neurons. However, intestinal barrier dysfunction plays a pathogenic role in IBS, and the intercellular spaces between epithelial cells are enlarged in diarrhea-predominant IBS, providing a morphological basis for the observed increase in intestinal permeability.^4,58 As in human IBS, CVH mice also have persistently enhanced colonic epithelial barrier permeability.⁵⁹ Thus, we hypothesize that Tsp1a might use a paracellular route to access the peripheral endings of colonic nociceptors, which can be superficially localized below the intestinal epithelium and in the submucosa.⁴

This raises the intriguing possibility that a peptidic inhibitor of Na_V1.7 might serve as an orally active analgesic for treating chronic gut pain. The relatively superficial localization of colonic nociceptors means that such a peptide would not need to be orally bioavailable in the conventional sense of reaching the systemic circulation; indeed, from the viewpoint of minimizing potential side effects, it would be desirable for such a peptide drug to breach the intestinal epithelium sufficiently to target colonic nociceptors but not robustly enough to reach the systemic circulation. Notably, the peptide drug linaclotide, which is used to provide relief from constipation and abdominal pain in IBS patients with constipation,⁶⁰ is orally active but undetectable in the systemic circulation at therapeutic doses.^61,62 Linaclotide targets guanylate cyclase C on the luminal surface of intestinal epithelial cells,⁶⁰ and like Tsp1a, it contains three disulfide bonds and is resistant to proteases and acidic pH.⁶¹

Future Directions

The current study suggests that pharmacological inhibition of Na_V1.7 might be a viable therapeutic strategy for alleviating chronic gut pain. Tsp1a appears to be good lead for development of an orally active Na_V1.7-directed analgesic for treatment of chronic visceral pain, but it will be critical to explore its pharmacokinetics when delivered orally and intracolonically as well as its potential immunogenicity. Notably, injection of Tsp1a into the bloodstream at a dose of ∼1 mg/kg, at which ProTx-II is lethal,²¹ did not affect heart rate or respiratory patterns, indicating that acute toxicity would be limited even in the unlikely event that intestinal absorption was sufficient to yield circulating levels of peptide.

Materials and Methods

Isolation of Tsp1a from Spider Venom

Lyophilized crude venom from the spider Thrixopelma sp. (2 mg in 200 μL H₂O) was loaded onto an analytical RP-HPLC column (Jupiter C18, particle size 5 μm, pore size 300 Å; 250 × 4.6 mm²; Phenomenex) attached to a Prominence HPLC system (Shimadzu, Rydalmere, New South Wales, Australia). Venom components were eluted at a flow rate of 1 mL·min^–1 using a gradient of solvent B [90% acetonitrile (ACN) and 0.05% trifluoroacetic acid (TFA) in H₂O] in solvent A (0.05% TFA in H₂O). Elution began with an isocratic elution using 5% solvent B for 5 min, followed by a gradient of 5–20% solvent B over 5 min, then 20–40% solvent B over 40 min, and finally a gradient of 40–80% solvent B over 5 min. RP-HPLC fractions that inhibited hNa_V1.1 were further fractionated using a VisionHT HILIC column (Grace, 150 × 4.6 mm²; Fisher Scientific). Fractions were eluted at a flow rate of 1 mL·min^–1 using isocratic elution at 95% solvent B for 3 min, followed by a gradient of 95–75% solvent B over 20 min. For all HPLC experiments, absorbance was measured at 214 and 280 nm using a Shimadzu Prominence SPD-20A detector.

Determination of the Amino Acid Sequence of Tsp1a

Peptide mass was determined via MALDI-TOF MS using a Model 4700 Proteomics Analyzer (Applied Biosystems, Foster City, CA). The HPLC fraction containing Tsp1a was spotted with α-cyano-4-hydroxycinnamic acid (7.5 mg·mL^–1 in 50% ACN). Tsp1a was reduced and alkylated before being sequenced via Edman degradation at the Australian Proteome Analysis Facility (Macquarie University, Sydney, Australia).

Chemical Synthesis of Tsp1a

Synthetic Tsp1a (sTsp1a) was assembled using standard Fmoc chemistry on a Symphony peptide synthesizer (Gyros Protein Technologies, Tuscon, AZ) at 0.125 mmol scale. C-Terminal amidation was achieved using a Rink-amide resin. Simultaneous release of peptide from resin and removal of side chain protecting groups was achieved using a solution of TFA/triisopropylsilane/water (48:1:1 v/v/v) for 2.5 h. Crude sTsp1a was triturated in chilled diethyl ether, then precipitated peptide was dissolved in solvent (45% ACN and 0.05% TFA), lyophilized, and purified via RP-HPLC using a linear gradient of 10–60% solvent B over 50 min at 50 mL·min^–1. Fractions were collected and analyzed using electrospray ionization (ESI) MS, then fractions of interest were pooled, lyophilized, and stored at −20 °C. sTsp1a (0.1 mg·mL^–1) was oxidatively folded for 16 h at room temperature in a buffer containing 2 M urea, 0.1 M Tris pH 8, 0.15 mM reduced glutathione, and 0.3 mM oxidized glutathione. Oxidation was quenched by acidification to pH 3 using neat TFA, then the peptide was filtered and purified using preparative and semipreparative RP-HPLC. The mass of sTsp1a was verified via MALDI-TOF MS.

Production of Recombinant Tsp1a

Recombinant Tsp1a (rTsp1a) was produced using a method that we previously optimized for production of disulfide-rich venom peptides.²³ A DNA fragment encoding Tsp1a, with codons optimized for high-level expression in Escherichia coli, was synthesized by GeneArt (ThermoFisher) and cloned into the pLIC-C vector that encodes a His₆-MBP tag with a TEV protease recognition site preceding the peptide-encoding region. Overexpression of the His₆-MBP-Tsp1a fusion protein was induced at 25 °C with 0.5 mM isopropyl-β-d-thiogalactoside (IPTG) when the cell density (OD_600nm) reached 0.6–0.8. Following Ni-NTA purification of the fusion protein, rTsp1a was cleaved from the His₆-MBP fusion tag with His₆-tagged TEV protease produced in-house using a published protocol.⁶³ The liberated rTsp1a was then purified via RP-HPLC using a semipreparative Ascentis C4 column. The column was pre-equilibrated with 10% solvent B, and the peptide was then eluted using a gradient of 10–60% solvent B over 30 min at 3 mL·min^–1. rTsp1a was further purified using an analytical Ascentis C18 RP-HPLC column pre-equilibrated with 15% solvent B. Peptide was eluted with a gradient of 15–45% solvent B over 40 min at 0.8 mL·min^–1. The mass of rTsp1a was verified via MALDI-TOF MS.

RP-HPLC Elution of sTsp1a, rTsp1a, and Native Tsp1a

Native Tsp1a, sTsp1a, and rTsp1a were eluted from an AerisPEPTIDE XB-C18 column (3.6 μm particle size, 100 Å pore size, 50 × 2.1 mm²; Phenomenex) using a gradient of 10–50% solvent B over 30 min.

Automated Patch-Clamp Electrophysiology

HEK293 cells stably expressing the α subunit of human Na_V1.8 or the α subunit of subtypes ranging from Na_V1.1 to Na_V1.7 in combination with the human β1 subunit (SB Drug Discovery, Glasgow, UK) were used to examine the effect of Tsp1a on hNa_V channels. HEK293-hNa_V cells were seeded into a 175 cm² cell culture flask 2 days prior to patching and were detached at 60% confluency using 2 mL of Detachin (Genlantin, San Diego, CA). After pelleting cells at 400g for 8 min, the supernatant was discarded, and the cells resuspended in 5 mL of QPatch media containing 96.5% CD293 medium, 25 mM HEPES (Gibco), and 1× glutamine (Gibco).

Whole-cell patch-clamp experiments were performed at room temperature on a QPatch 16X automated electrophysiology platform (Sophion Bioscience, Denmark) using 16-channel planar patch-chip plates (QPlates) with a patch-hole diameter of 1 μm and resistance of 2 MΩ. Whole-cell currents were filtered at 5 kHz (8-pole Bessel) and digitized at 25 kHz. A P4 online leak-subtraction protocol was used with non-leak-subtracted currents acquired in parallel. The extracellular solution was 2 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, 4 mM KCl, 145 mM NaCl, and 10 mM sucrose at pH 7.4, and the intracellular solutions was 140 mM CsF, 1 mM/5 mM EGTA/CsOH, 10 mM HEPES, and 10 mM NaCl at pH 7.3. All peptides were dissolved in extracellular solution with 0.1% bovine serum albumin (BSA). Concentration–response data were obtained using seven concentrations of peptide (2, 6, 20, 60, 200, 600, and 2000 nM). HEK293-hNa_V cells were clamped at a holding potential of −80 mV then 10 μL of peptide was added for 6 s before applying the following voltage protocol: −80 mV for 10 ms, −120 mV for 200 ms, 0 mV for 20 ms (or + 10 mV for 50 ms in the case of Na_V1.8), then return to the holding potential of −80 mV. The total incubation time for each peptide concentration was 5 min. The voltage protocol was repeated 10 times, and between each application of the voltage protocol, the cells were clamped at a holding potential of −80 mV for 30 s.

Experiments designed to determine whether Tsp1a is a use-dependent inhibitor of hNa_V1.7 were performed using HEK293 cells stably expressing hNa_V1.7 and the β1 subunit. Cells were clamped at a holding potential of −120 mV in the presence or absence of 30 nM Tsp1a and submitted to 0 mV depolarization pulses of 20 ms at 0.1 Hz followed by 0 mV depolarization pulses of 20 ms at 40 Hz. We investigated whether Tsp1a has higher affinity for partially inactivated channels using a pulse frequency of 0.1 Hz. Cells were clamped at −120 mV in the presence or absence of sTsp1a and subjected to a prolonged pulse of 10 s at −60 mV (equivalent to the V₅₀ of hNa_V1.7 inactivation; see Figure S5), followed by 0.1 Hz depolarization pulses at 0 mV for 20 ms.

Manual Patch-Clamp Electrophysiology

Coverslips containing HEK293 cells stably expressing hNa_V1.7α and hNa_Vβ1 were placed in a recording chamber on the stage of an inverted microscope with the extracellular solution containing 140 mM NaCl, 2 mM CaCl₂, 4 mM KCl, 1 mM MgCl₂, and 10 mM HEPES (pH 7.4 with NaOH). Recording patch pipettes were filled with an intracellular solution containing 120 mM CsCl, 30 mM NaCl, 1 mM/5 mM EGTA/CsOH, and 10 mM HEPES (pH 7.2 with CsOH) and had a resistance of 1–3 MΩ. All recordings were made at room temperature (22–24 °C) using a MultiClamp 700B amplifier, Axo Digidata 1550B, and pCLAMP software (Molecular Devices). Na_V currents were measured using the whole-cell patch-clamp technique. All peptides were dissolved in extracellular solution with 0.1% BSA. Conductance-voltage relationships were obtained using a protocol in which cells were depolarized from a holding potential of −120 mV, stepping from −80 to +60 mV in 5 mV increments. Steady-state inactivation was obtained by applying a depolarization to 0 mV following a series of step potentials from −120 to 20 mV in 5 mV increments. Recovery from fast inactivation was obtained by first applying a 0 mV pulse followed by repolarization to −120 mV and a second pulse to 0 mV after a period ranging from 0 to 50 ms, increasing in 1 ms increments.

NMR Spectroscopy

NMR spectra of sTsp1a (1 mg·mL^–1 in 10% D₂O, 90% H₂O, pH ∼4) were acquired at 25 °C on a cryoprobe-equipped Bruker Avance III 600 MHz spectrometer (Bruker Biospin, Billerica, MA). 1D ¹H and 2D ¹H–¹H TOCSY (80 ms mixing time) and NOESY (200 ms mixing time) spectra were acquired and processed using TopSpin version 3.5 (Bruker Biospin) and used to make sequence-specific resonance assignments using CCPNMR Analysis, v.2.4.1.⁶⁴ 1D ¹H NMR spectra were also collected at a range of temperatures (10–35 °C in 5 °C increments) for measurement of amide-proton temperature coefficients. 1D ¹H, 2D ¹H–¹³C HSQC, and 2D ¹H–¹H ECOSY spectra were also acquired using a sample of sTsp1a dissolved in 100% D₂O.

CYANA 3.97⁶⁵ was used for automated NOESY assignment and calculation of structures. TALOS-N⁶⁶ was used to estimate ϕ and ψ dihedral angles using Hα, Cα, Cβ, and HN chemical shifts, and these estimates were used as restraints in structure calculations with the error range set to twice the standard deviation estimated by TALOS-N. In addition to distance and dihedral angle restraints, structure calculations included hydrogen-bond restraints obtained from 1D ¹H D₂O exchange experiments, temperature coefficient data, and direct measurements from preliminary structures. χ₁ angles were derived from the E.COSY spectrum in combination with NOE intensities, and these χ₁ angles (±30°) were also used as restraints in the structure calculations. The structure was further refined in a water shell using protocols in the RECOORD database,⁶⁷ and the final ensemble of 20 structures was selected based on the lowest energy, best MolProbity scores, and fewest distance and dihedral angle violations. Atomic coordinates and NMR chemical shifts for sTsp1a have been deposited in the Protein Data Bank and BioMagResBank with accession numbers 7A64 and 34554, respectively.

Stability of sTsp1a in Human Plasma

sTsp1a (10 μM) was added to human serum (Sigma Aldrich) and incubated at 37 °C for up to 24 h. The reaction mixture was precipitated at the desired time by addition of 5% TFA (5 μL) and 5% formic acid (FA; 10 μL). A 5 μL sample from each time point was analyzed via LC-MS using a Phenomenex C18 column (150 mm × 2.1 mm, particle size 5 μm, 100 Å pore size) with a gradient of 1–50% solvent D (90% ACN, 0.1% FA) in solvent C (0.1% FA) over 14 min at 0.25 mL·min^–1, coupled with an AB SCIEX 5600 Triple TOF mass spectrometer (cycle time 0.2751 s). The area of peaks corresponding to the quadruple-charge state of sTsp1a were measured using PeakView and MultiQuant (Applied Biosystems). hANP (GenScript, Piscataway, NJ) and ziconotide (Alomone laboratories, Jerusalem, Israel) were used for stability comparisons. Three replications were performed for each peptide.

Stability of sTsp1a in Simulated Gastric Fluid

The stability of sTsp1a in simulated gastric fluid (SGF) was examined using a modified version of a published protocol.⁶⁸ Tsp1a was dissolved in SGF (0.5 mg/mL pepsin, 30 mM NaCl, pH 1.86) to give a final peptide concentration of 5 μM, and the mixture was then incubated at 37 °C for up to 24 h. Samples (30 μL) were collected at various time points, and the reaction quenched by addition of 200 mM NH₄HCO₃. Samples (5 μL) were then analyzed via LC-MS using an Agilent ZORBAX SB-C₁₈ column (100 mm × 2.1 mm, particle size 1.8 μm) coupled to a TripleTOF 5600 mass spectrometer (AB SCIEX) using a linear gradient of 1–60% solvent B (0.1% FA in 90% ACN) in solvent A (0.1% FA) over 7 min with a cycle time of 0.5 s (flow rate 0.2 mL/min). The area of peaks corresponding to the triple-charge state of sTsp1a were quantified using Analyst software and MultiQuant.

Animal Studies

Acute in vivo toxicity experiments were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee of Memorial Sloan Kettering Cancer Center and followed the U.S. National Institutes of Health guidelines for animal welfare.

All other experiments were performed in accordance with the guidelines of the Animal Ethics Committees of the South Australian Health and Medical Research Institute (SAHMRI) and Flinders University. Male C57BL/6 mice were used in all experiments. These experiments conformed to the relevant regulatory standards and the ARRIVE guidelines. Male C57BL/6 mice at 10–13 weeks of age were used in all experiments. Mice were acquired from an in-house C57BL/6J breeding program (JAX strain no. 000664; originally purchased from The Jackson Laboratory (breeding barn MP14; Bar Harbor, ME)) within SAHMRI’s specific and opportunistic-pathogen-free animal care facility. Mice were group-housed (maximum five mice per cage) in individual ventilated cages (IVC), which were filled with coarse chip dust-free aspen bedding (PURA; Cat no. ASPJMAEB-CA, Niederglatt, Switzerland). These cages were stored on IVC racks in specific housing rooms within a temperature-controlled environment of 22 °C and a 12 h light/12 h dark cycle. Mice had free access to LabDiet JL Rat and Mouse/Auto6F chow (Cat# 5K52, St Louis, MO) and autoclaved reverse-osmosis-purified water.

Chronic Visceral Hypersensitivity Model of IBS

Colitis was induced by administration of dinitrobenzene sulfonic acid (DNBS) as described previously.^14,42 Briefly, 13 week old C57BL/6J mice were fasted overnight with access to 5% glucose solution. After the fasting period, isofluorane-anaesthetized mice were administered an intracolonic enema of 0.1 mL of DNBS (6.5 mg in 30% ethanol) via a polyethylene catheter inserted 3 cm past the anus. Mice were then individually housed with unlimited access to soaked food and 5% glucose solution and observed daily for changes in body weight, physical appearance, and behavior. Histological examination of mucosal architecture, cellular infiltrate, crypt abscesses, and goblet cell depletion confirmed that DNBS induced significant damage by 3 days post-treatment, with mucosal architecture largely recovered by day 7 and fully recovered by 28 days post-treatment. At 28 days post-DNBS-treatment, these mice display sprouting of afferent central terminals in the dorsal horn,⁶⁹ they have prominent hyperalgesia and allodynia to colorectal distension,^42,45,70,71 and high-threshold nociceptors from these mice display significant mechanical hypersensitivity^14,45,60,72 and lower mechanical activation thresholds.⁵⁷ Thus, these mice are termed “CVH mice”.^14,57,60,72

In Vivo Pain Assessment: Visceromotor Response to Noxious Colorectal Distension

We used abdominal EMG to monitor the visceromotor response to noxious colorectal distension in fully awake animals.⁴² Under isoflurane anesthesia, the bare endings of two Teflon-coated stainless-steel wires (Advent Research Materials Ltd., Oxford, UK) were sutured into the right abdominal muscle and tunnelled subcutaneously to be exteriorized at the base of the neck for future access. At the end of the surgery, mice received a prophylactic antibiotic (Baytril; 5 mg/kg s.c.) and analgesic (buprenorphine; 0.4 mg/10 kg s.c.) and then were housed individually and allowed to recover for at least 3 days before assessment of VMR.⁴² On the day of VMR assessment, mice were briefly anaesthetized using isoflurane before receiving a 100 μL enema of vehicle (sterile saline) or sTsp1a (200 nM). A lubricated balloon (2 cm length) was gently introduced through the anus and inserted into the colorectum up to 0.25 cm past the anal verge. The balloon catheter was secured to the base of the tail and connected to a barostat (Isobar 3, G&J Electronics, Willowdale, Ontario, Canada) for graded and pressure-controlled balloon distension. Mice were allowed to recover from anesthesia in a restrainer with dorsal access for 15 min prior to initiation of the distension sequence. Distensions were applied at 20–40–50–60–70–80 mmHg (20 s duration each) at 2 min intervals so that the last distension was performed 30 min after intracolonic treatment. Following the final distension, colonic compliance was assessed as described below. The EMG electrodes were relayed to a data acquisition system and the signal was recorded (NL100AK headstage), amplified (NL104), filtered (NL 125/126, Neurolog, Digitimer Ltd., bandpass 50–5000 Hz), and digitized (CED 1401, Cambridge Electronic Design, Cambridge, UK) for off-line analysis using Spike2 (Cambridge Electronic Design). The analog EMG signal was rectified and integrated. To quantify the magnitude of the VMR at each distension pressure, the area under the curve (AUC) during the distension (20 s) was corrected for baseline activity (AUC predistension, 20 s).⁴² We also calculated the total AUC, which is the summation of data points across all distension pressures for each individual animal.

Colonic Compliance

Immediately after VMR assessment, colonic compliance was assessed by applying graded volumes (40–200 μL, 20 s duration each) to the balloon in the colorectum of fully awake mice, while recording the corresponding colorectal pressure as described previously.⁴² Following completion of the colonic compliance protocol mice were humanely sacrificed by cervical dislocation.

Acute In Vivo Toxicity

Six athymic nude mice (6–8 weeks old) were purchased from Envigo RMS (Indianapolis, IN). An intravenous catheter was placed in the tail vein of each mouse and they were anesthetized using isoflurane (Novaplus, Telangana, India). Anesthesia was maintained using 1.0–1.5 L·min^–1 of isoflurane and 2 L·min^–1 of oxygen. Three mice were injected with sTsp1a (7 nmoles in 100 μL of PBS), and three with 100 μL of PBS. Animals were monitored using a rodent surgical monitor (Indus instruments, Houston, TX) before, during, and up to 16 min after injection. Numeric data points represent averaged values of 5 ± 1 s duration. Data were collected before injection, and at 15 s, 5 min, and 15 min postinjection. Mouse body core temperature was measured via a rectal probe and electronically regulated via a surgical platform. High-resolution ECGs were obtained by placing noninvasive electrodes on the four paws, and electrical contact was assured using a conducting gel (Electrode cream, Indus instruments, TX). Heart rate was automatically calculated from R–R peaks in the ECGs. Peripheral capillary oxygen saturation (SpO₂) was noninvasively measured by placing a clip sensor on the animal’s left thigh.

Statistical Analyses

Automated Patch-Clamp Electrophysiology

Concentration–response data were fitted with a Hill equation using Prism v8 (GraphPad, San Diego, CA) to obtain IC₅₀ values. Data are mean ± standard error of the mean (SEM), where n represents the number of patched HEK293 cells. A p value < 0.05 was considered statistically significant, and it was calculated in Prism v8 using an unpaired t-test with Welch’s correction or one-way ANOVA with multiple comparisons.

Manual Patch-Clamp Electrophysiology

The voltage for half-maximal activation and inactivation (V_0.5) was calculated by fitting data with a Boltzmann equation in Prism v8. The time constant of recovery from fast inactivation was calculated by fitting data using a one-phase decay in Prism v8. Data are presented as mean ± SEM, where n represents the number of patched HEK293 cells. A p value < 0.05 was considered statistically significant, and it was calculated by paired t-test in Prism v8.

VMR Data

Data are presented as mean ± SEM, where n represents the number of animals. AUC data were statistically analyzed by generalized estimating equations followed by a Least Significant Difference (LSD) post hoc test when appropriate using SPSS 25.0. Total AUC data were analyzed by unpaired two-tailed t-tests (for differences between two groups) or mixed-effects model (REML) with post hoc analysis conducted with Bonferroni tests (for differences between more than two groups). A p value of <0.05 was considered statistically significant. Statistical analyses were performed using Prism v7.

Toxicity Studies

Data points representing loss of signal, as defined by a brief contact loss between the monitoring equipment and the mice, were excluded from the analysis. Loss of signal was corrected by placing the electrode in close contact to the mice. Data points represent mean ± standard deviation. Mann–Whitney tests were used for analysis of unpaired samples (e.g., vital signs in mice injected with sTsp1a and mice from the control group), and the Wilcoxon test was used for analysis of paired samples (e.g., vital signs from same mouse before and after injection). A p value of <0.05 was considered statistically significant. Prism v7 was used for data analysis.

Taxonomy and Ion Channel Nomenclature

Details of the taxonomic identification of Thrixopelma spec. (Peru) are provided in the Supporting Information and Figure S1. All ion channel nomenclature conforms to that outlined in the 2019/2020 Concise Guide to Pharmacology of Ion Channels.⁷³

Glossary

Abbreviations

ACN

acetonitrile

AuC

area under the curve

BSA

bovine serum albumin

CNS

central nervous system

CRD

colorectal distension

CVH

chronic visceral hypersensitivity

DNBS

dinitrobenzene sulfonic acid

EMG

electromyography

ESI

electrospray ionization

FA

formic acid

hANP

human atrial natriuretic peptide

HEK293

human embryonic kidney 293

hNa_V

human voltage-gated sodium channel

IBS

irritable bowel syndrome

ICK

inhibitor cystine knot

MALDI-TOF MS

matrix-assisted laser desorption/ionization time-of-flight mass spectrometry

MBP

maltose binding protein

MVIIA

ω-conotoxin MVIIA

PEPD

paroxysmal extreme pain disorder

p.i.

postinjection

PNS

peripheral nervous system

rTsp1a

recombinant Tsp1a

SGF

simulated gastric fluid

sTsp1a

synthetic Tsp1a

TEV

tobacco etch virus

TFA

trifluoroacetic acid

TNBS

trinitrobenzene sulfonic acid

VMR

visceromotor response

VSD

voltage sensor domain

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.1c00072.

• Taxonomic identification and anatomical details of Thrixopelma spec. (Peru); electrophysiological screening of fractions from crude venom against hNa_V1.1 and hNa_V1.6 S-5; activities of native Tsp1a, synthetic Tsp1a (sTsp1a), and recombinant Tsp1a (rTsp1a) on hNa_V1.7 S-6; in vivo intracolonic administration of sTsp1a does not alter sensitivity to noxious colorectal distension in healthy mice S-7; V₅₀ of steady-state inactivation for hNa_V1.7 channels stably co-expressed with the auxiliary β1 subunit in HEK293 cells (PDF)

Author Present Address

^¶ Laboratory of Membrane Proteins and Structural Biology, Biochemistry and Biophysics Center, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD 20892, USA

Author Present Address

^■ National Cancer Institute, National Institutes of Health, Frederick, MD 21702, USA

Author Present Address

^◇ School of Science & Engineering, University of the Sunshine Coast, Sippy Downs, Queensland 4556, Australia

Author Contributions

Conceptualization, funding, and project management: G.F.K. and S.M.B.; supervision: L.V.B, G.F.K., and S.M.B.; design, conduct, and analysis of experiments: Y.J., J.C., A.A., L.B., J.M., G.S., C.Y.C., F.C.C., J.G., C.I.S., T.R., and S.M.B.; provision of venoms: V.H.; taxonomy of Thrixopelma spec. (Peru): S.E.; drafting of manuscript: Y.J., L.V.B., V.H., and G.F.K.. All authors contributed to review and editing of the manuscript.

The authors declare the following competing financial interest(s): J.G., P.D.S.F., G.F.K., and T.R. are co-inventors on U.S. patent application 62,794,520 that covers the sequence, derivatives, and methods of use for Tsp1a. All other authors have no conflicts to declare. This arrangement has been reviewed and approved by Memorial Sloan Kettering Cancer Center in accordance with its conflict of interest policies.