Intra-channel bi-epitopic crosslinking unleashes ultrapotent antibodies targeting Na_V1.7 for pain alleviation

Summary

Crucial for cell activities, ion channels are key drug discovery targets. Although small-molecule and peptide modulators dominate ion channel drug discovery, antibodies are emerging as an alternative modality. However, challenges persist in generating potent antibodies, especially for channels with limited extracellular epitopes. We herein present a bi-epitopic crosslinking strategy to overcome these challenges, focusing on Na_V1.7, a potential analgesic target. Aiming to crosslink two non-overlapping epitopes on voltage-sensing domains II and IV, we construct bispecific antibodies and ligand-antibody conjugates. Enhanced affinity and potency are observed in comparison to the monospecific controls. Among them, a ligand-antibody conjugate (1080-PEG₇-ACDTB) displays a two-orders-of-magnitude improvement in potency (IC₅₀ of 0.06 ± 0.01 nM) and over 1,000-fold selectivity for Na_V1.7. Additionally, this conjugate demonstrates robust analgesic effects in mouse pain models. Our study introduces an approach to developing effective antibodies against Na_V1.7, thereby initiating a promising direction for the advancement of pain therapeutics.

Graphical abstract

Highlights

• •Bi-epitopic crosslinking enhances the potency of antibodies targeting ion channels
• •A ligand-antibody conjugate shows ultrapotent and highly selective Na_V1.7 inhibition
• •The ligand-antibody conjugate exhibits potent analgesic effects in mouse pain models

Zhang et al. introduce a bi-epitopic crosslinking strategy to enhance antibody potency against ion channels, particularly for Na_V1.7, a pain relief target. This approach, using bispecific antibodies and ligand-antibody conjugates, significantly improves affinity and potency, with one conjugate showing a 1,000-fold selectivity for Na_V1.7 and strong analgesic effects in mice, offering a promising direction for pain therapeutics.

Introduction

Ion channels play an essential role in maintaining membrane potential and participating in a myriad of cellular functions, including signal transduction, secretion, and volume regulation. They have emerged as the second most frequently pursued target for drug development, with nearly 18% of globally approved drugs deriving their therapeutic efficacy through ion channels.¹ Despite the widespread acknowledgment of ion channels’ critical involvement in numerous diseases, the current therapeutic arsenal addresses only 8% of all known ion channels, underscoring the substantial untapped potential for therapeutic innovation.²

Although small-molecule and peptide modulators have been the primary focus of ion channel drug development, monoclonal antibodies (mAbs) present a promising alternative modality.^3,4 Unlike small-molecule modulators, which often display unsatisfactory subtype selectivity and off-target binding,^5,6,7 antibodies generally demonstrate exceptional selectivity. As a result, antibodies have been gaining attention in the search for novel ion channel drugs.^3,4

Despite the advantages of mAbs as ion channel modulators, no mAbs targeting ion channels have been approved to date. The lack of success in generating such mAbs is attributable to several challenges. A significant challenge facing many ion channels is the limited availability of potential epitopes in their short extracellular loops, which are most likely targeted by mAbs. For example, voltage-gated ion channels and transient receptor potential channels have short extracellular loops.^8,9,10,11,12 Despite their potential as drug targets, the antibody discovery for these channels is challenging, often yielding antibodies with low affinities. Moreover, antibodies that bind to the ion channels do not necessarily modulate the channel activity.¹³ Developing a robust strategy to generate potent functional antibodies against such ion channels remains challenging.

The intrinsic cross-arm avidity of mAbs may improve the apparent binding to their targets.^14,15 Conventional mAbs achieve avidity through bivalent binding to two adjacent cell surface receptors. However, such intermolecular avidity decreases with decreasing receptor densities due to a shortage of receptors in the vicinity.¹⁶ Ion channels generally have a low cellular density,^17,18 which partially explains the low apparent affinity of the ion-channel-targeting antibodies. To overcome this limitation, we rationalized that bi-epitopic crosslinking of an ion channel can result in antibodies with high binding affinity to these low-density membrane proteins. Such intramolecular avidity has been demonstrated in several bispecific antibodies against viruses, G protein-coupled receptors (GPCRs), and intracellular proteins.^{19,20,21,22,23} Here, we present that targeting two non-overlapping epitopes is a viable strategy to obtain antibodies with high affinity and selectivity against ion channels.

In this study, we selected voltage-gated sodium channel Na_V1.7 as a model target. Na_V1.7 is a potential target for a novel class of analgesic therapeutics.^24,25 However, current anti-Na_V1.7 antibodies have low affinity and potency.^26,27,28 Here, we leveraged the bi-epitopic crosslinking strategy to construct potent and selective anti-Na_V1.7 antibodies. First, we constructed bispecific antibodies targeting the voltage-sensing domains II and IV (VSD_II and VSD_IV), which displayed increased inhibitory activity against Na_V1.7 compared to the parental monospecific antibodies. To provide another example, ligand-antibody conjugates were created to specifically target the voltage-sensing domains VSD_II and VSD_IV. In both formats, we observed that the linker length is a key determinant of effective bi-epitopic crosslinking. Furthermore, one ligand-antibody conjugate (1080-PEG₇-ACDTB) showed the highest potency, with IC₅₀ of 0.06 ± 0.01 nM at the half-inactivated state. Its affinity to Na_V1.7 was significantly increased, by two orders of magnitude, compared to the monovalent controls due to its interaction with both VSD_II and VSD_IV. Moreover, it showed more than 1,000-fold selectivity over other Na_V channels. The ligand-antibody conjugates showed efficient analgesic effects on mice in multiple different pain models. Overall, our research introduces a method for developing effective and specific antibodies targeting Na_V1.7 and other difficult targets with low expression levels and limited extracellular epitopes.

Results

A bispecific antibody with proper linker length exhibited increased Na_V1.7-inhibiting activity over the monospecific antibodies

To test the bi-epitopic crosslinking strategy, we first constructed biparatopic anti-Na_V1.7 antibodies in a heterodimeric single-chain variable fragment Fc fusion (scFv-Fc) format (Figure 1A). In particular, Knobs-into-Holes technology was used to promote the Fc heterodimerization,²⁹ leading to a “1 + 1” bispecific antibody format. Prior to constructing the bispecific antibodies, we first expressed a series of anti-Na_V1.7 mAbs generated in-house or by previous researchers (Figure S1A).³⁰ Whole-cell voltage-clamp recording on Na_V1.7-overexpressing HEK293 cells (Na_V1.7-HEK293) revealed that two candidates with the highest inhibitory activities were Ab-1B53 and Ab-1080 (Figure 1B), with IC₅₀s of 75.23 ± 8.15 and 137.8 ± 8.67 nM. Ab-1B53 binds to the S3-S4 loop of VSD_IV, and Ab-1080 binds to the S1-S2 loop of VSD_II. It is noteworthy that ligands binding to VSD_II and VSD_IV reportedly stabilize the activated/inactivated state of Na_V1.7.^31,32,33 Therefore, we chose Ab-1B53 and Ab-1080 to construct the two arms of the heterodimeric scFv-Fc.

Figure 1.

Design and optimization of heterodimeric scFv-Fc for Na_V1.7

(A) Schematic diagram of the interaction between the biparatopic antibody and Na_V1.7. The binding mode of conventional monoclonal antibodies is depicted as a comparison.

(B) Concentration-response curves and the table depicting the Na_V1.7 inhibitory activity of the monoclonal antibodies, determined with whole-cell patch-clamp recording on Na_V1.7-HEK293 cells. Data are shown as mean ± SEM (n = 3).

(C) Schematic diagram of biparatopic antibodies with hinge regions with different lengths. The amino acid sequences of the hinge regions are shown on the right.

(D) Concentration-response curves of Na_V1.7-inhibitory activity of 1080×1B53-21, 1080×1B53-35, 1080×1B53-64, and the monospecific controls Ab-1080 and Ab-1B53. The IC₅₀ of 1080×1B53-21 was 84.32 ± 23.46 nM. The IC₅₀ of 1080×1B53-35 was 209.53 ± 42.73 nM. The IC₅₀ of 1080×1B53-64 was 6.07 ± 0.97 nM. Data are shown as mean ± SEM (n = 3).

(E) Side view of the simulated structure of 1080 × 1B53-64/Na_V1.7 complex.

To achieve intramolecular bi-epitopic crosslinking, linker length is a critical design element. In this study, we designed three biparatopic antibodies with different hinge regions (21-aa human IgA hinge region, 35-aa camel IgG2 hinge region, and 64-aa human IgD hinge region) (Figures 1C and S1B). The three designed bispecific antibodies (1080×1B53-21, 1080×1B53-35, and 1080×1B53-64) were recombinantly expressed in HEK293F cells. SDS-PAGE was used to confirm the correct construction of the protein (Figure S1C). Next, whole-cell voltage-clamp recording was performed on Na_V1.7-HEK293 cells treated with the three bispecific antibodies. 1080×1B53-21 and 1080×1B53-35 exhibited similar inhibitory activity against human Na_V1.7 as the parental antibodies Ab-1080 and Ab-1B53 (Figure 1D). In contrast, 1080×1B53-64 showed significantly stronger inhibitory activity than the parental antibodies, with an IC₅₀ of 6.07 ± 0.97 nM (Figure 1D). The strong apparent binding affinity of 1080×1B53-64 suggested increased cross-arm avidity . The superior performance of 1080×1B53-64 over 1080×1B53-21 and 1080×1B53-35 indicates that sufficient linker length is critical for achieving high cross-arm avidity.

We next performed a molecular simulation to model the possible binding mode of 1080×1B53-21, 1080×1B53-35, and 1080×1B53-64 to Na_V1.7. The protein-protein docking revealed that 1080×1B53-21 and 1080×1B53-35 did not allow simultaneous binding of the two arms on a single Na_V1.7 protein (Figures S1D and S1E), likely due to the inadequate hinge length. In contrast, the hinge in 1080×1B53-64 exhibits a lengthy and flexible structure, allowing the simultaneous binding to the two epitopes on Na_V1.7 (Figure 1E). The simulation results corroborate with the Na_V1.7-inhibiting assay, validating that adequate linker length is required to bridge the two separating epitopes on Na_V1.7.

Bridging two distinct epitopes with a ligand-antibody conjugate resulted in ultrapotent Na_V1.7 inhibition

After demonstrating the bi-epitopic crosslinking strategy with bispecific antibodies, we explored whether we could employ ligand-antibody conjugates to bridge two epitopes. Several synthetic ligands have been discovered that are highly effective in inhibiting Na_V1.7, some with IC₅₀ in the sub-nanomolar range.³⁴ We chose 4-(2-(3-aminobenzo[d]isoxazole-5-yl)-4-chlorophenoxy)-2,5-difluoro-N-(1,2,4-thiadiazol-5-yl)benzenesulfonamide (ACDTB) as the synthetic ligand.³⁵ ACDTB is a potent state-dependent Na_V1.7 inhibitor with an IC₅₀ of 0.4 nM against the half-inactivated state. According to the structure of Na_V1.7 bound to an ACDTB analog,³⁶ we speculated that ACDTB binds to a pocket near the S3-S4 loop in VSD_IV. Molecular docking confirmed that ACDTB could bind in the extracellular aqueous cleft of the VSD_IV (Figure S2). As ACDTB and Ab-1080 bind to different extracellular regions, we reasoned that conjugating ACDTB to Ab-1080 potentially enables intramolecular crosslinking of VSD_II and VSD_IV of Na_V1.7 (Figure 2A).

Figure 2.

Construction and optimization of biparatopic ligand-antibody conjugates

(A) Schematic illustration of the interaction between the biparatopic antibody-ligand conjugate and Na_V1.7.

(B) The chemical structure of ACDTB and ACDTB-PEG_(n+3)-Mal. n = 0, 4, 12.

(C) Concentration-response curves of ACDTB-PEG_(n+3)-Mal (n = 0, 4, 12) on HEK293 cells. Data are shown as mean ± SEM (n = 2).

(D) Schematic illustration of the structure of 1080-PEG_(n+3)-ACDTB. The zoomed-in partial view on the left displays the Fab region of Ab-1080 (PDB: 8YHZ). The blue mark indicates the paired disulfide bond in the LC, and the green indicates the unpaired cysteine located in Cys80.

(E) The LC-TOF-MS of the deglycosylated and reduced conjugates was used to identify the conjugation efficiency of ACDTB-PEG_(n+3)-Mal (n = 0, 4, 12).

(F) Concentration-response curves of Na_V1.7 inhibitory effect of 1080-PEG_(n+3)-ACDTB on Na_V1.7-HEK293 cells. Data are shown as mean ± SEM (n = 3).

Ab-1080 is a rabbit-human chimeric antibody, possessing a free cysteine at position 80 within the light chain (LC) framework, allowing site-specific conjugation. Therefore, we designed three thiol-reactive linkers to conjugate ACDTB to Ab-1080. Molecular docking indicates ACDTB binds in the extracellular aqueous cleft of the VSD_IV with its anionic sulfonamide head group sitting deep in the cleft (Figure S2), similar to the other aryl sulfonamides. The anionic head group interacts with the gating charges on the S4 through electrostatic as well as van der Waals interactions. The aminobenzo[d]isoxazol group of ACDTB points out toward the opening of the pocket, presenting a potential site for linker extension. Moreover, structure-activity relationship studies revealed that modification of the 3-amino group on the benzo[d]isoxazole ring did not reduce the Na_V1.7-inhibiting activity.³⁵ Based on these results, we designed to extend polyethylene glycol (PEG) linkers with a maleimide (Mal) head group for conjugation. As shown in the studies on biparatopic 1080×1B53, a suitable linker length between the two binding paratopes is crucial for achieving intramolecular crosslinking. Therefore, we designed three linkers with 3, 7, and 15 PEG repeats to screen for an optimal length (Figure 2B). The three ligand-linker conjugates (ACDTB-PEG_(n+3)-Mal, n = 0, 4, 12, compounds 17A–C) were synthesized using a convergent approach (Scheme S1). ¹H-NMR and ¹³C-NMR spectrometry confirmed the successful synthesis of the compounds (Figures S3–S13). Whole-cell voltage-clamp analyses revealed that these ligand-linker conjugates (compounds 17A–C) exhibited similar Na_V1.7-inhibiting activity as ACDTB (Figure 2C).

Next, we conjugated ACDTB-PEG_(n+3)-Mal (n = 0, 4, 12) to Ab-1080 by Michael addition, yielding 1080-PEG₃-ACDTB, 1080-PEG₇-ACDTB, and 1080-PEG₁₅-ACDTB (Figure 2D). Mass spectrometry showed that ACDTB-PEG_(n+3)-Mal was conjugated to Ab-1080 successfully (Figures 2E and S14A–S14C). The ligand-to-antibody ratios were 1.81, 1.98, and 1.82, respectively. In addition, we performed proteomic analysis on 1080-PEG₇-ACDTB (Figure S14D), confirming that ACDTB was conjugated to Cys80 within the light chain of Ab-1080.

Once we constructed the ligand-antibody conjugates, we tested and compared their inhibitory activities against Na_V1.7 in the half-inactivated state. All conjugates showed an efficient Na_V1.7-inhibiting effect. Among them, 1080-PEG₇-ACDTB showed the strongest inhibitory activity, with an IC₅₀ of 0.06 ± 0.01 nM, almost 19-fold lower than 1080-PEG₁₅-ACDTB and 211-fold lower than 1080-PEG₃-ACDTB (Figure 2F). These results showed that PEG₇ is optimal among the three linkers tested. PEG₃ linker may not be long enough to bridge the two epitopes, incapable of achieving efficient intramolecular crosslinking. On the other hand, excessive linker length resulted in an unfavorable entropy change upon binding, possibly causing the decreased inhibitory activity of 1080-PEG₁₅-ACDTB in comparison to 1080-PEG₇-ACDTB. Since 1080-PEG₇-ACDTB has the best inhibitory effect, we chose it for further in-depth investigations.

1080-PEG₇-ACDTB effectively inhibits Na_V1.7 by simultaneously engaging two epitopes

To demonstrate the contribution of Ab-1080 and ACDTB to Na_V1.7 inhibitory activity, we chose two monospecific controls, each of which only binds to one epitope on Na_V1.7. Ab-1080 serves as a good control to bind to VSD_II. To construct a VSD_IV-binding control, we conjugated ACDTB to an inert antibody without Na_V1.7 binding affinity. The ideal inert antibody should have negligible Na_V1.7 binding affinity and a similar structure to Ab-1080. From the co-crystal structure of the Fab fragment of Ab-1080 with Na_V1.7 VSD_II peptide (EHHPMTEEFKN) (Figure 3A), we identified R52, S53, and S55 in Ab-1080 complementarity-determining region H2 (CDR-H2) and S91, and Y92 in CDR-L3 are in direct contact with Na_V1.7 (Figure S15A). Therefore, we constructed a series of control antibodies by mutating these key amino acids to alanine (Table S1). The flow cytometry results indicate that an Ab-1080 mutant, H2L3, has the weakest affinity (Figure S15B). Its mutation sites are R52A, S53A, and S55A in CDR-H2 and S91A, and Y92A in CDR-L3 (Figure 3A). Therefore, we chose H2L3 to construct a control ligand-antibody conjugate, namely H2L3-PEG₇-ACDTB.

Figure 3.

1080-PEG₇-ACDTB has improved affinity and excellent subtype selectivity

(A) The co-crystal structure of Ab-1080 Fab and VSD_II S1-S2 extracellular epitope (PDB: 8YHZ). The green box highlights the key amino acids in the CDR-H2 region of the 1080 antibody that interacts with the VSD_II peptide. Similarly, the key amino acids in the CDR-H3 region are displayed in the orange box. The S1-S2 extracellular epitope is shown in blue.

(B) Dose-response curves of Na_V1.7-inhibiting activity of 1080-PEG₇-ACDTB, Ab-1080, ACDTB, and H2L3-PEG₇-ACDTB. Data are shown as mean ± SEM (n = 2–3)

(C) Binding traces for Alexa Fluor 488-labeled antibodies to Na_V1.7-HEK293 cells (gray) and the result from globally fitting curves to a “one-to-one two-state” model (black).

(D) The sodium current pulses of Na_V1.1–1.6 and Na_V1.8 with and without the treatment of 1080-PEG₇-ACDTB (1 μM).

Next, we compared the inhibitory activity of 1080-PEG₇-ACDTB with the two controls (Ab-1080 and H2L3-PEG₇-ACDTB) using whole-cell voltage clamp. Ab-1080 has an IC₅₀ of 137.8 ± 8.67 nM, substantially higher than 1080-PEG₇-ACDTB (Figure 3B). Moreover, at a concentration of 1 μM, H2L3-PEG₇-ACDTB only inhibited approximately 20% of the sodium current (Figures 3B and S15C). We note that the inhibition activity of H2L3-PEG₇-ACDTB is drastically weaker than ACDTB and ACDTB-PEG₇, possibly due to the steric hindrance caused by the bulky antibody conjugated to ACDTB. These observations suggest that the simultaneous binding of both Ab-1080 and ACDTB is essential for the potent Na_V1.7 inhibition activity.

Next, we evaluated the affinity of 1080-PEG₇-ACDTB and the two controls on Na_V1.7-HEK293 with a LigandTracer assay. We first conjugated Alexa Fluor 488 to the antibodies. Electrophysiology studies revealed that the activity of the labeled 1080-PEG₇-ACDTB was comparable to 1080-PEG₇-ACDTB (Figure S16A), indicating that fluorophore labeling did not affect the target engagement. Subsequently, the labeled antibodies were co-incubated with Na_V1.7-HEK293 cells, and the kinetics of the association and dissociation processes were monitored with LigandTracer Green, a device for real-time interaction measurement. In comparison to Ab-1080, 1080-PEG₇-ACDTB showed faster association and slower dissociation (Figure 3C). Moreover, H2L3-PEG₇-ACDTB did not show observable binding on Na_V1.7-HEK293 cells at a concentration of 30 nM. Very weak interaction can be detected at 500 nM and 1 μM, which may be attributed to non-specific binding (Figure S16B). The association and dissociation curve can be fitted into a two-state model (Table S2).³⁷ The apparent dissociation constant (K_D) of 1080-PEG₇-ACDTB is 2.78 × 10⁻⁹ M, which is two orders of magnitude lower than that of Ab-1080. Therefore, the aforementioned results indicate that simultaneous binding of both Ab-1080 and ACDTB resulted in improved apparent affinity over monospecific controls. Note that the IC₅₀s of the Na_V1.7 inhibition activity positively correlated with the K_D values (Table S2), suggesting that the improved apparent affinity (avidity) is an important determinant of improved Na_V1.7 inhibition.

Once we demonstrated that dual-epitope engagement is critical for the potency of 1080-PEG₇-ACDTB, we sought to understand whether the avidity observed in 1080-PEG₇-ACDTB is due to intermolecular crosslinking or intramolecular crosslinking (Figure S16C). As demonstrated in Figure 2F, the Na_V1.7-inhibiting activities of the 1080-PEG_(n+3)-ACDTB conjugates largely depend on the linker length, indicating that the avidity might be caused by intramolecular bi-epitopic crosslinking. However, the PEG₇ linker we used to connect ACDTB and Ab-1080 (approximately 40 Å) is too short to bridge the epitopes in VSD_II and VSD_IV, which are approximately 65 Å apart. The only possible way to achieve intramolecular epitope bridging is via cross-arm avidity, i.e., one arm of Ab-1080 and the ACDTB on the other arm simultaneously engage the VSD_II and VSD_IV on a single Na_V1.7 protein. To demonstrate whether the cross-arm avidity is essential, we constructed two 1080×H2L3 antibodies (Figure S16D), one with a C80A mutation on the Ab-1080 arm and the other with a C80A mutation on the Ab-H2L3 arm. Conjugating these two antibodies with ACDTB-PEG₇-Mal resulted in two bivalent ligand-antibody conjugates, namely 1080-PEG₇-ACDTB(SA) and 1080-PEG₇-ACDTB(CA). 1080-PEG₇-ACDTB(SA) conjugated to the same arm of 1080, while 1080-PEG₇-ACDTB(CA) has ACDTB conjugated to the H2L3 arm. After recombinant expression and conjugation, SDS-PAGE confirmed the purity of the antibodies and conjugates (Figure S16E). Whole-cell patch-clamp recordings revealed that while 1080-PEG₇-ACDTB(CA) effectively inhibited Na_V1.7, 1080-PEG₇-ACDTB(SA) exhibited poor Na_V1.7 inhibition activity (Figures S16F and S16G). These observations demonstrated that cross-arm intramolecular avidity is essential for effective Na_V1.7 inhibition of the ligand-antibody conjugates.

1080-PEG₇-ACDTB is a subtype-selective Na_V1.7 inhibitor

Previous studies have shown that ACDTB has off-target inhibitory effects on two other voltage-gated sodium channels, Na_V1.2 and Na_V1.6.³⁵ We also verified the inhibitory activity of ACDTB-PEG₇-Mal for Na_V1.2 and Na_V1.6. The results indicate that adding PEG linkers did not attenuate the Na_V1.2 or Na_V1.6 inhibitory effect (Figure S16H). To study whether 1080-PEG₇-ACDTB has improved selectivity over ACDTB, we tested the inhibitory effect of 1080-PEG₇-ACDTB on other subtypes through electrophysiology. The results showed that there was no observable inhibitory activity at 1 μM on Na_V1.1-Na_V1.6 and Na_V1.8 (Figure 3D). Comparing the inhibitory activity of 1080-PEG₇-ACDTB with ACDTB/Ab-1080 mixture on Na_V1.2, Na_V1.6, and Na_V1.7 (Figure S16I), we found that 1080-PEG₇-ACDTB had superior inhibitory activity on Na_V1.7 compared to the ACDTB/Ab-1080 mixture, indicating the synergy caused by the intramolecular avidity. In contrast, 1080-PEG₇-ACDTB showed much weaker inhibition than the ACDTB/Ab-1080 mixture on Na_V1.2 and Na_V1.6, signifying its improved selectivity. As we observed in H2L3-PEG₇-ACDTB, once conjugated to an inert antibody, the activity of the ligand is dramatically decreased, possibly due to the steric hindrance. The same effect also hindered 1080-PEG₇-ACDTB from interacting with the off-targets of ACDTB. Therefore, although the parent synthetic ligand ACDTB has multiple off-target effects, the ligand-antibody conjugate can achieve excellent specificity by simultaneously targeting two epitopes.

1080-PEG₇-ACDTB preferred to bind the inactivation state of Na_V1.7

Since the aryl sulfonamide analogs inhibit Na_V1.7 in a state-dependent manner,^35,36,38 we next characterized whether 1080-PEG₇-ACDTB has a similar state preference. Toward this aim, we examined the dose-inhibition curves of ACDTB, Ab-1080, and 1080-PEG₇-ACDTB on Na_V1.7 in resting and half-inactivated states (Figure 4A). IC₅₀ of ACDTB-PEG₇-Mal at the resting state is approximately 82-fold higher than that in the half-inactivated state, similar to the parental ACDTB.³⁵ Ab-1080 had no apparent preference for either state. The IC₅₀ of 1080-PEG₇-ACDTB in the resting state was approximately 19-fold higher than that in the half-inactivated state. The preferential binding of 1080-PEG₇-ACDTB to the inactivated state of Na_V1.7 aligns with ACDTB and other aryl sulfonamide analogs.

Figure 4.

1080-PEG₇-ACDTB preferentially binds the inactive state of Na_V1.7 and affects the slow inactivation process

(A) The Na_V1.7 inhibition activity of ACDTB-PEG₇-Mal, Ab-1080, and 1080-PEG₇-ACDTB in resting and half-inactive states on stable Na_V1.7-HEK293 cells (n = 2–3).

(B) Schematic diagram of Na_V1.7 gating kinetics.

(C) Normalized conductance versus voltage of Na_V1.7 channels with or without treatment of ACDTB, Ab-1080, and 1080-PEG₇-ACDTB. Curves were fitted by the Boltzmann function (n = 4–6).

(D) Top: the voltage protocol of steady-state fast-inactivation curves of Na_V1.7 channels. Bottom: steady-state fast-inactivation curves of Na_V1.7 channels with or without treatment of ACDTB, Ab-1080, and 1080-PEG₇-ACDTB (n = 5–7).

(E and F) Top: the protocols for entry into and recovery from the fast inactivation state. Bottom: the entry into and recovery from the fast inactivation state of Na_V1.7 channels with or without treatment of ACDTB, Ab-1080, and 1080-PEG₇-ACDTB (n = 5–6).

(G) Top: protocol of voltage-dependent steady-state slow inactivation of Na_V1.7 currents. Bottom: steady-state slow-inactivation curves of Na_V1.7 channels with or without treatment of ACDTB, Ab-1080, and 1080-PEG₇-ACDTB (n = 5–6).

(H and I) Top: the protocols for entry into and recovery from the slow inactivation state. Bottom: the entry into and recovery from the slow inactivation state of Na_V1.7 channels with or without treatment of ACDTB, Ab-1080, and 1080-PEG₇-ACDTB. The curves were fitted with a one-phase exponential or two-phase exponential function. All data were expressed as the mean ± SEM, n = 5–8, unless otherwise specified.

To further understand the mechanism of 1080-PEG₇-ACDTB, we examined the effect of 1080-PEG₇-ACDTB on Na_V1.7 gating kinetics (Figure 4B). The effects of 1080-PEG₇-ACDTB, Ab-1080, and ACDTB on steady-state activation were first investigated. As shown in Figure 4C, G-V curves revealed that 1080-PEG₇-ACDTB at 1 nM had no obvious effect on the activation gating of Na_V1.7 with half-activation voltage of −30.88 ± 0.88 mV, similar to −28.76 ± 0.77 mV in the absence of 1080-PEG₇-ACDTB. Ab-1080 at 100 nM and ACDTB at 1 nM also showed similar results (Figure 4C). This indicates that 1080-PEG₇-ACDTB has no obvious effect on the activation of Na_V1.7.

Once the Na_V channel is opened, it enters fast and slow inactivation processes (Figure 4B). Fast inactivation is an important process for Na_V channels to allow repolarization and return to the resting state in preparation for firing the next action potential within a few milliseconds (ms).^39,40,41 In contrast, slow inactivation develops in a time range of 100 ms to several minutes and reduces the excitability of neurons. We then determined the influence of 1080-PEG₇-ACDTB on the steady-state inactivation. The treatment of 1080-PEG₇-ACDTB caused a rightward shift of the steady-state fast inactivation curve, with V_1/2 shifted from -74.08±0.63 mV to -68.91±0.76 mV (Figure 4D). The treatment of ACDTB, Ab-1080, and 1080-PEG₇-ACDTB resulted in slight acceleration of the development of fast inactivation, with no observable impact on the recovery from fast inactivation (Figures 4E and 4F). In addition, 1080-PEG₇-ACDTB at 1 nM caused a leftward shift of the slow inactivation curve, with V_1/2 shifted from −40.27 ± 0.8 mV to −54.03 ± 3.29 mV (Figure 4G). A further kinetics study revealed that while 1080-PEG₇-ACDTB treatment did not alter the kinetics of the entry into slow inactivation (Figure 4H), it delayed the recovery from slow inactivation (Figure 4I). Interestingly, while ACDTB significantly prolonged the second phase of the recovery, 1080-PEG₇-ACDTB primarily delayed the first phase. These data indicate that 1080-PEG₇-ACDTB inhibits Na_V1.7 mainly by affecting the channel’s fast and slow inactivation.

1080-PEG₇-ACDTB has good serum stability and long circulation time in mice

To investigate the potential of 1080-PEG₇-ACDTB for therapeutic applications, we first studied the serum stability of 1080-PEG₇-ACDTB. ACDTB is coupled to the light chain (LC) of Ab-1080 via a non-cleavable linker. As the thiol-maleimide adduct was reported to undergo reverse-Michael addition, we tested the serum stability of the ligand-antibody conjugates (Figure 5A). Mass spectrometry detected three variations of LC ACDTB conjugate (LC-PEG₇-ACDTB). The first one is the intact LC-PEG₇-ACDTB, the percentage of which gradually decreased from 83% to 27%. We noted that a species with a molecular weight of LC-PEG₇-ACDTB+18 was detected in the mass spectrometry. The species corresponds to the product resulting from the ring-opening hydrolysis of the succinimide group. The hydrolysis reaction reportedly stabilizes the linker.⁴² These two species constitute over 80% of the total LC species, indicating the linker has a high stability in serum. The third species corresponded to Ab-1080 LC, which may be the product of reverse Michael addition. The percentage of Ab-1080 LC increased slightly over time, yet remained below 20% throughout the 7-day incubation. Collectively, 1080-PEG₇-ACDTB retained more than 80% conjugated form after incubation in mice serum for 7 days.

Figure 5.

Characterization of the pharmaceutical properties of 1080-PEG₇-ACDTB

(A) In vitro serum stability of 1080-PEG₇-ACDTB. The cartoon on the left shows ring-opening hydrolysis of 1080-PEG₇-ACDTB. The plot is the percentage change of the light chain and the ring-opening product of 1080-PEG₇-ACDTB in serum over time.

(B) Pharmacokinetic study in C57BL/J6 mice. 1080-PEG₇-ACDTB was administered at 30 mg/kg (mpk) by intravenous injection, and 1080-PEG₇-ACDTB concentrations in serum samples were determined by LC-MS/MS. Data are presented as mean ± SEM (n = 3).

(C) Immunofluorescence of mouse plantar frozen sections 2 days after subcutaneous injection of 1080-PEG₇-ACDTB or H2L3-PEG₇-ACDTB. The tissue sections were stained with anti-PGP9.5 antibodies and DAPI. The scale bars are 50 μm.

(D) Inhibitory effect of 1080-PEG₇-ACDTB on A-803467-insensitive currents in mouse DRG neurons. Data are presented as mean ± SD. A two-tailed unpaired Student's t test was performed for statistical analysis, ∗∗∗p < 0.001.

Pharmacokinetic studies were conducted to discern the circulating half-life of 1080-PEG₇-ACDTB in vivo. 1080-PEG₇-ACDTB concentration was determined by liquid chromatography-tandem mass spectrometry (LC-MS/MS) with a detection limit of 20 μg/mL (Figure S17A). LC-PEG₇-ACDTB was detectable in 4 days post dosing, with a distribution half-life (t_½α) of 0.98 h and an elimination half-life (t_½β) of 74.93 h (Figure 5B; Table S3). Although the half-life of 1080-PEG₇-ACDTB is shorter than typical mAbs,⁴³ it exhibited a similar half-life to antibody-drug conjugates.⁴⁴ Overall, the pharmacokinetic performance of 1080-PEG₇-ACDTB is comparable to those of antibody-drug conjugates.

Live imaging experiments were conducted to evaluate the local retention behavior of the 1080-PEG₇-ACDTB in C57BL/6J mice upon subcutaneous injection. Before injection, a fluorescent dye (AF647-C2-Mal) was used to label 1080-PEG₇-ACDTB, yielding 1080-PEG₇-ACDTB-AF647. Fluorescent images of hind paws from C57BL/6J mice were taken at specified time intervals following the intraplantar administration of 1080-PEG₇-ACDTB-AF647 in the left hind paws (Figures S17B and S17C). The fluorescence intensity reached its peak on day 1 post injection, likely attributed to the protein diffusing toward skin-proximal area. After day 1, the fluorescence intensity started to decay, with a half-life around 2 days (for the 10 mpk group). This observation indicated that 1080-PEG₇-ACDTB had a long retention time in tissue, indicating its potential utility for local injection.

We examined the colocalization of 1080-PEG₇-ACDTB with PGP9.5 marker in mouse tissue after subcutaneous injection of 1080-PEG₇-ACDTB in the hind paws (Figure 5C). Clear colocalization of 1080-PEG₇-ACDTB and PGP9.5 indicated that 1080-PEG₇-ACDTB engaged peripheral nerve upon injection. To ascertain whether 1080-PEG₇-ACDTB exerts a potent effect on mouse neurons, we extracted mouse dorsal root ganglion (DRG) cells and performed voltage-clamp experiments. Before the test, 100 nM A-803467, a Na_V1.8 inhibitor, was applied to selectively block Na_V1.8 currents. We found that the A-803467-insensitive current was effectively inhibited by 1 nM 1080-PEG₇-ACDTB (Figure 5D). These findings suggest that 1080-PEG₇-ACDTB can potently inhibit Na_V1.7 expressed in mouse DRG neurons.

1080-PEG₇-ACDTB has analgesic activity in mice

Following the assessment of the in vitro activity, stability, and pharmacokinetic performance of 1080-PEG₇-ACDTB, we proceeded to examine its analgesic effect across different administration routes and animal models. First, we employed a thermal pain model, in which latency to respond to a hot plate is measured by the amount of time it takes for the mouse to lick one of its paws. After detection of pre-dose latency, 1080-PEG₇-ACDTB, Ab-1080, or H2L3-PEG₇-ACDTB were administered by intrathecal injection at a dosage of 4 mpk. 2 h, 24 h, and 48 h after the injection, the latency of first response at was recorded and plotted in Figure 6A. The maximum possible effect (%MPE) of 1080-PEG₇-ACDTB for thermal hyperalgesia reached 63.29% at 2 h and 74.81% at 24 h. However, this effect only lasted for 24 h, and after 48 h there was no significant difference between the groups (Table S4). This may be due to the drug clearance from the cerebrospinal fluid (CSF).⁴⁵ Surprisingly, we observed comparable analgesic effects of 1080-PEG₇-ACDTB and Ab-1080. Given the short half-life of mAbs and derivatives in CSF,^46,47 we speculated that both 1080-PEG₇-ACDTB and Ab-1080 reached maximum efficacy at 2 and 24 h. Overall, intrathecal 1080-PEG₇-ACDTB treatment attenuated thermal hyperalgesia in mice.

Figure 6.

1080-PEG₇-ACDTB reduces pain in mice

(A) The first response latency of mice on the hot plate at 53°C after intrathecal administration of 1080-PEG₇-ACDTB, H2L3-PEG₇-ACDTB, Ab-1080, and PBS (n = 6).

(B) Mechanical pain thresholds in the hind paws of mice after subcutaneous administration of 1080-PEG₇-ACDTB, H2L3-PEG₇-ACDTB, Ab-1080, and PBS (n = 5–6).

(C) The numbers of writhings on different days after intravenous administration of 1080-PEG₇-ACDTB, H2L3-PEG₇-ACDTB, Ab-1080, and PBS. Acetic acid (0.7%) is intraperitoneally injected to induce abdominal inflammatory pain (n = 6).

(D) In vivo dose-dependent analgesic effects of 1080-PEG₇-ACDTB in acetic acid-induced writhing test with PF05089771 as a positive control (n = 6). IV, intravenous; IG, intragastric. All data are presented as mean ± SEM. Two-tailed Student’s t tests with Welch’s correction were performed for statistical analysis, with n.s., not significant, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

After demonstrating the efficacy of 1080-PEG₇-ACDTB via intrathecal injection, we then investigated the effect of 1080-PEG₇-ACDTB on pain thresholds via local injection. We tested whether 1080-PEG₇-ACDTB has an effect on the mechanical pain threshold of hind paws in mice after local 1080-PEG₇-ACDTB administration (Figure 6B). After detecting the baseline pain threshold of the mice, we subcutaneously injected PBS, Ab-1080, H2L3-PEG₇-ACDTB, and 1080-PEG₇-ACDTB (10 mpk) into the left hind paws of mice and measured the pain threshold from day 0 (1 h after the injection) to day 7 (Figures 6B and S18; Table S5). 1080-PEG₇-ACDTB showed a significant increase in the threshold for mechanical pain from 1 h to 3 days after administration. Results indicated that Ab-1080 had an increase in mechanical pain thresholds, though the effect was not as pronounced as in the 1080-PEG₇-ACDTB group. On day 2, the difference between the 1080-PEG₇-ACDTB group and other control groups reached a peak. The analgesic effect of 1080-PEG₇-ACDTB on the withdrawal threshold was maintained until day 3, while it was restored to the same level as the PBS group on day 4. These observations indicate that local subcutaneous administration of 1080-PEG₇-ACDTB can increase the mechanical pain threshold in mice.

We next tested the analgesic effect of 1080-PEG₇-ACDTB via intravenous injection in the acetic acid-induced abdominal writhing model (Figure 6C; Table S6). First, PBS, Ab-1080, H2L3-PEG₇-ACDTB, and 1080-PEG₇-ACDTB of 50 mpk were injected via tail vein injection. 24 h later, 0.7% acetic acid was given to these mice by intraperitoneal injection. We then counted the number of writhing times within 20 min after acetic acid injection. Writhing is defined as a stretch, extension of hind limbs, and contraction of the abdomen (Video S1). 1080-PEG₇-ACDTB treatment led to a 66.7% decrease in writhing times, indicating 1080-PEG₇-ACDTB can suppress acetic acid-induced abdominal pain via systemic administration. Moreover, the analgesic activity was dose dependent (Figure 6D), demonstrating that analgesic effects are caused by 1080-PEG₇-ACDTB. PF-05089771, a selective aryl sulfonamide Na_V1.7 inhibitor previously tested in phase 2 clinical trials, was selected as a positive control. Following oral administration, the analgesic efficacy of 25 mpk PF-05089771 at 24 h was modest, in line with previous observations on aryl sulfonamides (Figure 6D). The results from the three pain models collectively suggest that the 1080-PEG₇-ACDTB has analgesic activity in different pain models.

Discussion

Many ion channels have limited extracellular epitopes and low cell surface density, posing challenges for antibody discovery for these ion channels. The lack of accessible epitopes impedes the generation of mAbs with strong intrinsic affinity. The low cell surface density diminishes the inherent cross-arm avidity of conventional mAbs.^48,49,50 The biparatopic antibody approach circumvents these issues by introducing intramolecular bivalent binding. In our research, we obtained highly potent and selective Na_V1.7 inhibitors by utilizing intramolecular bi-epitopic crosslinking. Moreover, we highlighted that the linker length is a key determinant to allow the bi-epitopic binding to the two epitopes. Such intramolecular crosslinking has been documented in antibody discovery for viruses and GPCRs.^{19,20,21,22,23} Our work first demonstrates its potential in antibody discovery for ion channels.

Na_V1.7 is a highly sought-after target in the pursuit of non-opioid analgesics. Human genetic data have provided strong support for the development of selective inhibitors of Na_V1.7 as potential analgesic drugs. However, recent clinical setbacks and unsatisfactory performance of preclinical compounds in animal pain models have raised concerns about the potential of Na_V1.7 inhibitors as effective human therapeutics. There is an ongoing debate regarding the reasons for the failure of Na_V1.7 inhibitors in clinical trials. Hypotheses include high plasma protein binding, insufficient target engagement, the requirement for suitable subtype cross-reactivity, inappropriate state dependency, the involvement of opioid receptors, etc.^51,52 The early clinical candidates, however, were hindered by a lack of selectivity toward other Na_V isoforms, high plasma protein binding, or high unbound clearance, thus rendering the level of target engagement inadequate to generate substantial analgesia. In comparison, our 1080-PEG₇-ACDTB, being a distinct modality, has potent Na_V1.7 inhibition, high serum stability, and a long plasma half-life. Consequently, 1080-PEG₇-ACDTB presents an alternative strategy for effective Na_V1.7 inhibition, thus aiding in the elucidation of the clinical failure of Na_V1.7 inhibitors.

In summary, this study introduces an intramolecular bi-epitopic crosslinking strategy to selectively inhibit the Na_V1.7 ion channel, a key target for non-opioid analgesics. Utilizing bispecific scFv-Fc and ligand-antibody conjugate formats, we achieved potent and selective inhibition, with 1080-PEG₇-ACDTB exhibiting an IC₅₀ of 0.06 ± 0.01 nM. Further, our research underscores the significance of linker length in intramolecular crosslinking. Addressing challenges in Na_V1.7 inhibitors, our modality showed reliable pain relief in murine models due to its potent inhibition, high serum stability, and extended plasma half-life. Beyond Na_V1.7, the study suggests the bi-epitopic crosslinking approach as a solution for antibody discovery challenges in ion channels with limited epitopes and low cellular density.

Limitations of the study

Study limitations include the lack of suitable antibodies as positive controls in the pain relief models. As a proof-of-concept study, the analgesic efficiency of the designed antibody-ligand conjugate was tested only in three animal models of acute pain. Further neuropathic pain models will provide a more comprehensive assessment of the potential of 1080-PEG₇-ACDTB in treating neuropathic pain. While PF-05089771 was included for comparison in our study, further extensive studies are required to make conclusions on whether 1080-PEG₇-ACDTB exhibits superior efficacy as a pain reliever. Despite outperforming PF-05089771, 1080-PEG₇-ACDTB fails to demonstrate in vivo efficacy commensurate with its potent in vitro activity, necessitating further investigation into the underlying reasons.

Due to the flexibility of the linkers in the biparatopic antibody and antibody-ligand conjugate format, this study lacks direct structural evidence for intra-channel bi-epitopic crosslinking. In addition, channel activation and inactivation are dynamic processes involving multiple conformations. Unfortunately, structure biology studies have not provided sufficient evidence to depict the key conformations in these processes, making it difficult for us to understand the interaction between 1080-PEG₇-ACDTB and Na_V1.7.

Resource availability

Lead contact

Further correspondence and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Juanjuan Du (dusps@tsinghua.edu.cn).

Materials availability

Plasmids and cell lines used in this study are available under an appropriate material transfer agreement.

Data and code availability

• •The structure factors and atomic coordinates for the peptide-antibody complex have been deposited in the Protein DataBank under accession number PDB: 8YHZ. Other data generated in this study are available from the lead contact upon request.
• •Data generated in this study are available from the lead contact upon request.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

The authors acknowledge the financial support from the National Science Foundation of China (82373774), China; Tsinghua University Initiative Scientific Research Program; Beijing Advanced Innovation Center for Structural Biology, China; and the Tsinghua-Peking Joint Center for Life Sciences (CLS) Program. We thank the staff at the X-ray Crystallography Facility of Tsinghua University, and the staff at Shanghai Synchrotron Radiation Facility for providing technical support and assistance in X-ray crystallography data collection and analysis. We thank the Center of Pharmaceutical Technology, Tsinghua University, for data collection in mass spectrometry and serum stability testing and the Protein Chemistry and Histology Platform for proteomics analysis.We thank Prof. Nieng Yan and Prof. Xiaojing Pan for providing the plasmids of human Na_V channels and Prof. Zhuo Huang at Peking University and Prof. Bailong Xiao for the helpful discussion on electrophysiology.

Author contributions

J.D., Y.Z., and Y.D. conceived the study, designed the experiments, analyzed the data, and wrote the manuscript. Y.D. and R.Z. conducted the ligand syntheses and antibody conjugation. P.Z. performed computational simulations. S.F. assisted with the structural studies and data analysis. Z.Z. performed the mouse DRG electrophysiology. Y.Z. performed the rest experiments with assistance from Y.D., P.Z., Q.C., H.C., W.R., M.W., and L.W. All authors contributed to manuscript editing.

Declaration of interests

J.D., Y.Z., and R.Z. have a patent pending to Tsinghua University.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-NeuN Antibody, clone A60, Alexa Fluor® 555 Conjugate	Merck	Cat# MAB377A5; RRID:AB_2814948
Donkey anti-Rabbit IgG (H + L) Highly Cross-Adsorbed Secondary Antibody	Thermo Scientific	Cat# A-21207; RRID:AB_141637
Anti-PGP9.5 antibody	Abcam	Cat# ab108986; RRID:AB_10891773
Goat Anti-Human IgG-Fc Secondary Antibody	Sino biological	Cat# SSA015;RRID:AB_3662834
Goat anti-Human IgG (H + L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 488	Thermo Scientific	Cat# A-11013; RRID:AB_2534080
Ab-1A64	This paper	RRID: AB_3662835
Ab-1B53	This paper	RRID: AB_3662836
Ab-1D38	This paper	RRID: AB_3662837
Ab-1C78	This paper	RRID: AB_3662838
Ab-1D54	This paper	RRID: AB_3662839
Ab-1080	patent US 9,266,953 B2 (variable region)	RRID: AB_3662840
Ab-1066	patent US 9,266,953 B2 (variable region)	RRID: AB_3662841
Ab-932	patent US 9,266,953 B2 (variable region)	RRID: AB_3662842
Chemicals, peptides, and recombinant proteins
(L)-dehydroascorbic acid	SigmaAldrich	Cat# 261556
NaV1.7 VSDII domain S1-S2 loop peptide - EHHPMTEEFKN	Genewiz Corporation	N/A
PNGase F	NEB	Cat# P0704S
Papain from papaya latex	SigmaAldrich	Cat# P3125
Alexa Fluor™ 488 C5 maleimide	SigmaAldrich	Cat# A10254
EZ-Link™ Sulfo-NHS-LC-Biotin	Thermo Scientific	Cat# 21335
Streptavidin Magnetic Beads	NEB	Cat# S1420S
HBS-EP+ Buffer 10×	Cytiva	Cat# BR100826
ACDTB	Focken, T. et al.34	N/A
Chymotrypsin, Sequencing Grade	Promega Corporation	Cat# V1061
PEI Max	Polysciences	Cat# 24765
A-803467	MedChemExpress	HY-11079
PF-05089771	Bide Pharmatech.	BD292811
Critical commercial assays
Pierce™ BCA Protein Assay Kit	Thermo Scientific	Cat# 23227
Q5 High-Fidelity 2X Master Mix	NEB	Cat# M0492L
Gibson Assembly® Master Mix	NEB	Cat# E2611L
Deposited data
Co-crystal structure of Ab-1080 Fab and NaV1.7 VSDII S1-S2 extracellular epitope	This paper	PDB: 8YHZ
Experimental models: Cell lines
HEK293 FreeStyle (Invitrogen) cells	Gibco™	Cat# R79007
NaV1.7-HEK293 cells	Donated by Peter G. Schultz lab	N/A
NaV1.2-HEK293 cells	This paper	N/A
NaV1.6-HEK293 cells	This paper	N/A
Experimental models: Organisms/strains
C57/BL-6J mice	THU-LARC	N/A
Software and algorithms
GraphPad Prism 8.0	GraphPad software	https://www.graphpad.com
ZEISS ZEN3.3 (blue edition)	ZEISS	https://www.zeiss.com
AlphaFold Multimer v2.0	AlphaFold Multimer v2.0	https://alphafold.ebi.ac.uk/
PyMol	Schrödinger	https://pymol.org/2/
pClamp 10 software	Axon Instruments	https://support.moleculardevices.com/
TraceDrawer Software Version 1.9.2	Ridgeview Instruments	https://tracedrawer.com/release-of-tracedrawer-1-9-2/
Schrödinger 2022/2023	Schrödinger	https://newsite.schrodinger.com/life-science/download/release-notes/release-2022-1/
Proteome Discovery searching algorithm version 1.4	Thermo Scientific	https://www.thermofisher.cn
Octet analysis software version 9.0	FortéBio, Inc.	https://www.sartorius.com

Experimental model and study participant details

Cell lines and culture methods

HEK293 FreeStyle (Invitrogen) cells were cultured in SMM 293-TII (Sino Biological) and grown in humidified incubators at 37°C and 5% CO₂.

Na_V1.2, Na_V1.6, and Na_V1.7 HEK293 cell lines were cultured in DMEM (Gibco) and grown in humidified incubators at 37°C and 5% CO₂. Culture media was supplemented with 10% Fetal Bovine Serum and 1% Penicillin-Streptomycin (Thermo Scientific).

Animals

Adult C57/BL-6J mice (female) were used for all the behavioral and PK studies. All the animal procedures were approved by the Tsinghua University Institutional Animal Care and Use Committee (THU-LARC) with a protocol number of 17-DJJ1. For behavioral tests, animals were habituated to the environment for at least a week before the testing. All the behaviors were tested blindly by participants Y. Z., Y. D., R. Z., and H. C.

Method details

N-(2,4-dimethoxybenzyl)-1,2,4-thiadiazol-5-amine (rzhu001, 3)

Amine 1 (1 g, 9.89 mmol), and aldehyde 2 (1.81 g, 1.1 eq.), were mixed in 100 mL toluene. The reaction mixture was sealed and heated for 16 h under reflux and azeotropic removal of water using a Dean-Stark apparatus. Then the solution was cooled to ambient temperature and evaporated in high vacuo yielding a colorless oil which was dissolved in 50 mL of MeOH. Sodium borohydride (NaBH₄) (1.5 eq.) was added at 0°C. The reaction mixture was allowed to warm to ambient temperature and stirred overnight. After concentration under vacuum, water (50 mL) was added to the residue. The resulting aqueous solution was extracted with EtOAc (3 × 50 mL). The combined organic phase was washed with water (50 mL), brine (2 × 50 mL), and dried over anhydrous Na₂SO₄. Concentrated under vacuum, the residue was purified by silica gel chromatography (PE/EtOAc, 7:1) provided 3 as a light-yellow solid in 49.2% yield (1.22 g). ¹H NMR (400 MHz, DMSO-d₆) δ 8.69 (s, 1H), 7.90 (s, 1H), 7.17 (d, J = 8.3 Hz, 1H), 6.58 (s, 1H), 6.49 (d, J = 8.3 Hz, 1H), 4.38 (d, J = 3.9 Hz, 2H), 3.80 (s, 3H), 3.74 (s, 2H); ¹³C NMR (101 MHz, DMSO-d₆) δ 183.47, 160.60, 158.56, 130.09, 117.99, 104.81, 98.85, 55.92, 55.68, 44.02; HRMS-ESI⁺ [M + H]⁺: 252.0807, found: 252.0815.

N-(2,4-dimethoxybenzyl)-2,4,5-trifluoro-N-(1,2,4-thiadiazol-5-yl)benzenesulfonamide (DYCC14, 5)

To a mixture of 3 (0.7 g, 2.79 mmol) in THF (20 mL) was added LHMDS (2.36 mL of a 1.3 M solution in THF, 1.1 eq.) at −60°C. The reaction mixture was stirred at −60°C for 30 min, warmed to ambient temperature and stirred for additional 15 min. The mixture was cooled to −60°C, and Sulfonyl chloride (0.43 mL, 1.1eq.) was slowly added to it. The reaction mixture was allowed to warm to ambient temperature, stirred overnight, and quenched by addition of saturated NH₄Cl solution (50 mL). The mixture was extracted with EtOAc (3 × 50 mL). The combined organic phase was washed with brine (3 × 50 mL), dried over anhydrous Na₂SO₄, and concentrated in vacuo. Purification using silica gel chromatography (PE/EtOAc, 5:1) provided the title compound 5 as a light yellow oil in 84.6% yield (1.05 g).¹H NMR (400 MHz, DMSO-d₆) δ 8.47 (s, 1H), 7.88–7.79 (m, 2H), 7.10 (d, J = 8.4 Hz, 1H), 6.43 (dd, J = 8.4, 2.4 Hz, 1H), 6.35 (d, J = 2.3 Hz, 1H), 5.25 (s, 2H), 3.71 (s, 3H), 3.66 (s, 3H); ¹³C NMR (101 MHz, DMSO-d₆) δ 178.92, 161.21, 159.37, 158.34, 130.13, 119.96, 119.73, 114.30, 108.83, 105.13, 98.19, 55.68, 49.09. HRMS-ESI⁺ [M + H]⁺: 446.0456, found: 446.0489.

tert-butyl (2-(2-(2-hydroxyethoxy)ethoxy)ethyl)carbamate (DYCB55, 7)

Di-tert-butyl dicarbonate (2.93 g, 2 eq.) was added to a solution of amino alcohol (1 g, 6.7 mmol) in a 9:1 (v/v) mixture of methanol/triethylamine (20 mL). The reaction was left stirring under reflux and upon completion, the solvent was removed under reduced pressure and the residue extracted with DCM/water. The combined organic layer was dried over anhydrous Na₂SO₄ and concentrated under reduced pressure to yield title product 7 as a colorless oil and for further use (Crude form).

2,2-Dimethyl-4-oxo-3,8,11-trioxa-5-azatridecan-13-yl 4-methylbenzenesulfonate (DYCB56, 8)

A mixture of p-toluenesulfonyl chloride (10.05 mmol, 2eq.) in DCM (20.0 mL) was added to a DCM solution [10 mL, 7 (6.7 mmol, crude form), Et₃N (13.4 mmol, 2eq.) and 4-DMAP (0.67 mmol, 0.1eq.)] in a dropwise fashion over a period of 1 h. The resulting mixture was stirred overnight at room temperature, then poured into water. The organic layer was washed with brine (3 × 50 mL), and dried over anhydrous Na₂SO₄. Filtration, evaporation of the solvent in vacuo and purification of the residue by silica gel chromatography provided the title compound 8 as a light brown oil in 66% yield (1.78 g). ¹H NMR (400 MHz, DMSO-d₆) δ 7.78 (d, J = 6.8 Hz, 2H), 7.48 (d, J = 5.7 Hz, 2H), 6.75 (s, 1H), 4.10 (d, J = 2.7 Hz, 3H), 3.60–3.54 (m, 2H), 3.43 (s, 4H), 3.17 (dd, J = 5.6, 2.5 Hz, 2H), 3.04 (d, J = 6.6 Hz, 2H), 2.42 (s, 3H), 1.37 (s, 9H); ¹³C NMR (101 MHz, DMSO-d₆) δ 156.04, 145.36, 132.88, 130.59, 128.09, 78.04, 70.44, 70.10, 69.88, 69.75, 68.36, 49.06, 28.68, 21.55. HRMS-ESI⁺ [M + H]⁺: 404.1743, found: 404.1738.

tert-butyl (2-(2-(2-(4-formylphenoxy)ethoxy)ethoxy)ethyl)carbamate (DYCB58, 10)

To an anhydrous DMF (25 mL) solution of 8 (1.75g, 4.3372 mmol) were successively added K₂CO₃ (1.2 g, 2 eq.) and aldehyde 9 (583 mg, 1.1 eq.). After being stirred for overnight at 75°C, the reaction mixture was poured into water (50 mL), extracted with EtOAc (3 × 50 mL), the combined organic layer was washed with brine (3 × 50 mL). Then, the organic extract was dried over Na₂SO₄ and filtered off from an insoluble fraction. The filtrate was evaporated to dryness, and the residue was purified by silica gel chromatography (PE/EtOAc, 2:1) provided the title compound 10 as a colorless oil in 70.13% yield (1.075 g). ¹H NMR (400 MHz, DMSO-d₆) δ 9.86 (s, 1H), 7.86 (d, J = 7.4 Hz, 2H), 7.14 (d, J = 7.3 Hz, 2H), 6.76 (s, 1H), 4.25–4.16 (m, 2H), 3.81–3.74 (m, 2H), 3.62–3.55 (m, 2H), 3.55–3.49 (m, 2H), 3.41–3.34 (m, 2H), 3.06 (dd, J = 12.5, 6.3 Hz, 2H), 1.36 (s, 9H); ¹³C NMR (101 MHz, DMSO-d₆) δ 191.73, 163.93, 156.03, 132.24, 130.12, 115.41, 78.03, 70.36, 69.96, 69.66, 69.16, 68.14, 49.06, 28.68. HRMS-ESI⁺ [M+Na]⁺: 376.1731, found: 376.1724.

tert-butyl (2-(2-(2-(4-(((5-bromobenzo[d]isoxazol-3-yl)amino)methyl)phenoxy)ethoxy) ethoxy)ethyl)carbamate (DYCC10, 12)

4 Å Molecular Sieve (560 mg) was added in two-necked flask, heated for 2 min by hot air gun, allowed to cool to ambient temperature and degassed for 10 min. To a solution of 10 (0.33 g, 0.934 mmol), 11 (0.2 g, 1 eq.) in DCM (10 mL) were added Et₃SiH (300 μL, 2eq.) and TFA (390 μL, 3 eq.). The reaction mixture was reflux for overnight, allowed to cool to ambient temperature. The mixture was filtrated with DCM (3 × 15 mL) to remove 4 Å Molecular Sieve and the filtrate was concentrated under vacuum. The residue was purified by silica gel chromatography (PE/EtOAc, 3:1) provided the title compound 12 as a colorless oil in 78.79% yield (405 mg). ¹H NMR (400 MHz, DMSO-d₆) δ 8.16 (d, J = 2.4 Hz, 1H), 7.67 (dd, J = 8.8, 2.2 Hz, 1H), 7.47 (d, J = 9.3 Hz, 2H), 7.33 (d, J = 8.6 Hz, 2H), 6.93 (d, J = 8.7 Hz, 2H), 6.75 (s, 1H), 4.40–4.32 (m, 2H), 4.10–4.04 (m, 2H), 3.73 (dd, J = 5.8, 3.3 Hz, 2H), 3.61–3.55 (m, 2H), 3.55–3.48 (m, 2H), 3.38 (t, J = 6.3 Hz, 2H), 3.06 (d, J = 5.6 Hz, 2H), 1.36 (s, 9H).¹³C NMR (101 MHz, DMSO-d₆) δ 161.45, 158.03, 156.06, 133.04, 131.33, 129.54, 124.68, 119.01, 114.77, 114.11, 112.04, 78.04, 70.34, 70.21–70.03, 69.82, 69.41 (s), 67.61, 49.07, 46.34, 28.69. HRMS-ESI⁺ [M + H]⁺: 550.1553, found: 550.1550.

tert-butyl (2-(2-(2-(4-(((5-(5-chloro-2-hydroxyphenyl)benzo[d]isoxazol-3-yl)amino) methyl)phenoxy)ethoxy)ethoxy)ethyl)carbamate (DYCC12, 14)

To a solution of 12 (0.4 g, 0.73 mmol), 13 (150 mg, 1.2 eq.) and Na₂CO₃ (231 mg, 3 eq.) in mixed solution (DME/H₂O, v/v, 15:1, 19.2 mL) was added Pd(PPh₃)₄ (84 mg, 0.1 eq.). The reaction mixture was heated to 90°C for overnight, allowed to cool to ambient temperature and poured into water (50 mL), diluted with EtOAc (3 × 30 mL), the combined organic layer was washed with brine (3 × 50 mL), dried over Na₂SO₄ and then the organic layer was concentrated under vacuum. The residue was purified by silica gel chromatography (PE/EtOAc, 2:1) provided the title compound 14 as a light-yellow oil in 48.78% yield (212 mg). ¹H NMR (400 MHz, DMSO-d₆) δ 9.89 (s, 1H), 8.07 (s, 1H), 7.75 (dd, J = 8.7, 1.7 Hz, 1H), 7.47 (d, J = 8.7 Hz, 1H), 7.42 (t, J = 5.8 Hz, 1H), 7.35 (d, J = 8.3 Hz, 2H), 7.30 (d, J = 2.7 Hz, 1H), 7.22 (dd, J = 8.6, 2.7 Hz, 1H), 6.98 (d, J = 8.6 Hz, 1H), 6.93 (d, J = 8.5 Hz, 2H), 6.73 (t, J = 6.0 Hz, 1H), 4.39 (d, J = 5.7 Hz, 2H), 4.07 (dd, J = 6.1, 3.1 Hz, 2H), 3.78–3.68 (m, 2H), 3.58 (dd, J = 6.2, 3.7 Hz, 2H), 3.52 (dd, J = 6.0, 3.6 Hz, 2H), 3.39 (t, J = 6.1 Hz, 2H), 3.07 (q, J = 6.0 Hz, 2H), 1.36 (s, 9H); ¹³C NMR (101 MHz, DMSO-d₆) δ 161.53, 158.57, 157.98, 156.22–155.76, 153.49, 131.79, 131.74, 131.65–131.64, 130.02, 129.51, 128.34, 123.19, 122.10, 118.11, 117.01, 114.76, 109.21, 77.93, 70.35, 69.99, 69.67, 69.42, 67.63, 60.21, 28.69. HRMS-ESI⁺ [M + H]⁺: 598.2320, found: 598.2336.

tert-butyl (2-(2-(2-(4-(((5-(5-chloro-2-(4-(N-(2,4-dimethoxybenzyl)-N-(1,2,4-thiadiazol-5-yl) sulfamoyl)-2,5-difluorophenoxy)phenyl)benzo[d]isoxazol-3-yl)amino)methyl)phenoxy) ethoxy)ethoxy)ethyl)carbamate (DYCC15, 15)

To a mixture of 14 (0.77 g, 1.29 mmol) and K₂CO₃ (0.53 g, 3eq.) in DMSO (40 mL) was added 5 (0.63 g, 1.1 eq.) and the mixture was stirred overnight at ambient temperature. After dilution with water (50 mL), the resulting mixture was extracted with EtOAc (3 × 50 mL), the combined organic layer was washed with brine (3 × 100 mL) and dried over Na₂SO₄. Concentration in vacuo provided a residue which was purified by silica gel chromatography (PE/EtOAc, 4:1) provided 15 as an off-white solid in 84% yield (1.1 g). ¹H NMR (400 MHz, DMSO-d₆) δ 8.43 (s, 1H), 8.04 (d, J = 1.8 Hz, 1H), 7.63 (ddd, J = 7.0, 6.3, 3.9 Hz, 3H), 7.55 (dd, J = 8.7, 2.6 Hz, 1H), 7.48 (d, J = 8.7 Hz, 1H), 7.43 (t, J = 5.8 Hz, 1H), 7.33 (d, J = 8.4 Hz, 2H), 7.27 (d, J = 8.7 Hz, 1H), 7.11–7.03 (m, 2H), 6.92 (d, J = 8.5 Hz, 2H), 6.73 (t, J = 5.8 Hz, 1H), 6.42 (dd, J = 8.6, 2.3 Hz, 1H), 6.29 (d, J = 2.3 Hz, 1H), 5.17 (s, 2H), 4.38 (d, J = 5.7 Hz, 2H), 4.09–4.05 (m, 2H), 3.76–3.71 (m, 2H), 3.68 (s, 3H), 3.62–3.55 (m, 5H), 3.52 (dd, J = 6.0, 3.7 Hz, 2H), 3.39 (t, J = 6.1 Hz, 2H), 3.07 (q, J = 6.0 Hz, 2H), 1.36 (s, 9H); ¹³C NMR (101 MHz, DMSO-d₆) δ 178.86, 162.08, 161.05, 159.36, 158.67, 158.25, 158.10, 156.04, 150.17, 134.88, 131.60, 131.52, 131.25, 130.69, 129.86, 129.82, 129.49, 122.79, 122.49, 119.23, 119.01, 118.78, 117.37, 114.75, 114.52, 109.83, 108.56, 108.29, 104.96, 98.28, 78.03, 70.35, 69.99, 69.67, 69.42, 67.63, 55.66, 49.07, 48.80, 46.36, 28.68. HRMS-ESI⁺ [M + H]⁺: 1023.2636, found: 1023.2628.

4-(2-(3-((4-(2-(2-(2-aminoethoxy)ethoxy)ethoxy)benzyl)amino)benzo[d]isoxazol-5-yl)-4-chlorophenoxy)-N-(2,4-dimethoxybenzyl)-2,5-difluoro-N-(1,2,4-thiadiazol-5-yl) benzenesulfonamide (DYCC16, 16)

To a solution of 15 (296 mg, 0.29 mmol) in DCM (10 mL) was added TFA (2 mL). The pink reaction mixture was stirred overnight at ambient temperature and concentrated in vacuo. Purification of the residue by silica gel chromatography (MeOH/DCM, 1:20) provided the title product as an off-white solid in 31.75% yield (71 mg). ¹H NMR (400 MHz, DMSO-d₆) δ 8.07 (s, 1H), 7.90 (s, 1H), 7.68 (dd, J = 8.7, 1.8 Hz, 1H), 7.57 (dd, J = 9.2, 4.7 Hz, 3H), 7.51–7.43 (m, 3H), 7.33 (d, J = 8.2 Hz, 2H), 7.18 (d, J = 8.8 Hz, 1H), 7.03 (dd, J = 9.9, 6.4 Hz, 1H), 6.92 (d, J = 8.3 Hz, 2H), 4.39 (d, J = 5.8 Hz, 2H), 4.13–4.05 (m, 2H), 3.75 (dd, J = 5.6, 3.6 Hz, 2H), 3.60 (dt, J = 10.6, 3.7 Hz, 6H), 2.96 (t, J = 5.2 Hz, 2H), 2.70 (q, J = 7.2 Hz, 2H), 1.03 (t, J = 7.2 Hz, 2H); ¹³C NMR (101 MHz, DMSO-d₆) δ 161.99, 158.65, 158.04, 157.98, 151.26, 151.22, 146.69, 146.67, 134.11, 131.66, 131.50, 131.48, 131.31, 129.82, 129.63, 129.60, 129.49, 129.42, 122.67, 121.43, 117.27, 117.16, 116.94, 114.71, 109.86, 109.48, 109.21, 70.23, 70.18, 69.41, 67.58, 67.17, 46.34, 46.23. HRMS-ESI⁺ [M + H]⁺: 773.1430, found: 773.1415.

ACDTB-PEG₃-Mal (DYCB04, 17A)

A mixture of 16 (Hydrochloride form), Mal-PEG₃-NHS ester (1.5 eq.), Et₃N (3.0 eq.) and 1 mL of super-dry DMF was stirred overnight at 40°C. After completion of the reaction, the resulting mixture was diluted with ultrapure water and acetonitrile (in total 5 mL), injected into sampler and then purified by HPLC. Finally, the appropriate fractions were pooled and lyophilized to obtain the final product as a white solid (8 mg, 27% yield). ¹H NMR (400 MHz, DMSO-d₆) δ 8.36 (s, 1H), 8.06 (s, 1H), 7.97 (t, J = 5.5 Hz, 1H), 7.70–7.64 (m, 2H), 7.59 (d, J = 2.3 Hz, 1H), 7.50–7.45 (m, 2H), 7.42 (t, J = 5.5 Hz, 1H), 7.32 (d, J = 8.3 Hz, 2H), 7.25 (d, J = 8.7 Hz, 1H), 7.08 (dd, J = 10.4, 3.9 Hz, 1H), 6.97 (s, 1H), 6.91 (d, J = 8.4 Hz, 2H), 4.37 (d, J = 5.4 Hz, 2H), 4.10–4.05 (m, 2H), 3.75–3.70 (m, 2H), 3.58 (t, J = 7.1 Hz, 4H), 3.53–3.49 (m, 2H), 3.37 (s, 2H), 3.15 (dd, J = 11.5, 5.7 Hz, 2H), 2.32 (t, J = 7.3 Hz, 2H); ¹³C NMR (101 MHz, DMSO-d₆) δ 171.18, 169.94, 162.03, 158.70, 158.08, 150.89, 134.99, 134.38, 134.40, 130.03, 129.73, 129.67, 129.48, 122.74, 117.33, 114.76, 109.81, 70.32, 70.03, 69.48, 69.43, 67.63, 46.35, 38.96, 34.53, 34.41. HRMS-ESI^- [M-H]^-: 922.1543, found: 922.1617.

ACDTB-PEG₇-Mal (DYCB06, 17B)

The compound was synthesized using Mal-PEG₄-NHS ester and was isolated as a white solid (9 mg, 39.6% yield). ¹H NMR (400 MHz, MeOD) δ 7.96 (s, 1H), 7.93 (d, J = 1.1 Hz, 1H), 7.65 (dd, J = 8.7, 1.7 Hz, 1H), 7.60 (dd, J = 10.1, 6.3 Hz, 1H), 7.54 (d, J = 2.6 Hz, 1H), 7.42 (dd, J = 8.7, 2.6 Hz, 1H), 7.35 (dd, J = 12.3, 8.8 Hz, 3H), 7.14 (d, J = 8.7 Hz, 1H), 6.93 (d, J = 8.6 Hz, 2H), 6.78 (s, 2H), 6.67 (dd, J = 10.1, 6.4 Hz, 1H), 4.43 (s, 2H), 4.18–4.09 (m, 2H), 3.88–3.80 (m, 2H), 3.74 (t, J = 6.9 Hz, 2H), 3.69 (dt, J = 13.0, 5.1 Hz, 4H), 3.63 (dd, J = 5.7, 2.9 Hz, 3H), 3.59–3.49 (m, 13H), 3.46 (t, J = 5.4 Hz, 2H), 3.35 (t, J = 5.4 Hz, 2H), 3.30–3.26 (m, 2H), 2.43 (dt, J = 8.1, 6.5 Hz, 4H). ¹³C NMR (101 MHz, MeOD) δ 172.62, 171.67, 170.80, 170.75, 162.06, 158.51, 158.15, 157.13, 157.03, 150.62, 134.53, 134.07, 131.18, 131.07, 131.05, 130.59, 130.02, 128.92, 128.78, 121.58, 121.47, 121.44, 121.40, 116.87, 116.84, 116.65, 114.29, 109.15, 107.48, 107.21, 70.36, 70.22–69.68 (m), 69.50, 69.16, 69.05, 67.24, 66.82, 46.18, 39.03, 38.96, 36.14, 34.37, 34.06. HRMS-ESI⁺ [M + H]⁺: 1171.5937, found: 1171.6028.

ACDTB-PEG₁₅-Mal (DYCB05, 17C)

The compound was synthesized using Mal-PEG₁₂-NHS ester and was isolated as a white solid (10.7 mg, 36.2% yield). ¹H NMR (400 MHz, DMSO-d₆) δ 8.07 (s, 1H), 8.01 (s, 1H), 7.89 (s, 2H), 7.67 (dd, J = 8.7, 1.7 Hz, 1H), 7.60–7.53 (m, 2H), 7.52–7.41 (m, 3H), 7.32 (d, J = 8.6 Hz, 2H), 7.15 (d, J = 8.7 Hz, 1H), 7.07 (dd, J = 9.9, 6.5 Hz, 1H), 7.00 (s, 1H), 6.91 (d, J = 8.6 Hz, 2H), 4.37 (d, J = 5.6 Hz, 2H), 4.09–4.04 (m, 2H), 3.75–3.70 (m, 2H), 3.57 (dd, J = 8.1, 4.7 Hz, 6H), 3.49 (d, J = 5.3 Hz, 54H), 3.17 (dt, J = 18.0, 5.7 Hz, 4H), 2.31 (dd, J = 15.0, 7.2 Hz, 4H).; ¹³C NMR (101 MHz, DMSO-d₆) δ 185.84, 171.20, 170.57, 169.92, 162.00, 159.42, 158.69, 158.05, 151.40, 135.01, 133.98, 131.52, 131.44, 131.27, 129.86, 129.54, 129.49, 129.45, 122.72, 121.12, 117.30, 114.71, 109.80, 109.71, 70.31, 70.22, 70.16, 70.12, 70.05, 70.01, 69.96, 69.60, 69.45, 69.42, 67.59, 67.25, 46.33, 40.62, 40.41, 40.20, 39.99, 39.79, 39.58, 39.37, 38.98, 36.48, 34.52, 34.39, 25.69. HRMS-ESI⁺ [M + H]⁺: 1523.5217, found:1523.5123.

The cloning, production, and purification of monoclonal antibodies

The variable region sequences of the Ab-1080, Ab-1066, and Ab-932 were obtained from patent US 9,266,953 B2. The genes encoding the variable regions were synthesized by GENEWIZ, Inc. and cloned into the pFUSE vector (Invivogen) to construct the plasmids for the chimeric antibody. The light chain variable region of the rabbit IgG (PDB: 4HBC) contains a Cys80 that can form a disulfide bond with Cys172 in the CH1 region. Once the constant regions were replaced with human counterparts, Cys80 became a free unpaired cysteine that can be coupled with maleimide.

Ab-1A64, Ab-1B53, Ab-1D38, Ab-1C78, and Ab-1D54 were derived from mouse hybridoma immunized with the peptide DLIETYFVSP (S3-S4 loop in VSD_IV human Na_V1.7). Peptide immunization in mice and hybridoma production were conducted by Hangzhou HuaAn Biotechnology Co., Ltd. (Hangzhou, China). The DNA fragments encoding the variable regions were amplificated by multiplexed PCR with pooled primers for murine antibody frameworks. Subsequently, the DNAs encoding the variable regions were inserted into the pFUSE vector to construct the plasmids for chimeric antibody expression.

The plasmids for Ab-1080 mutants (Table S1) were generated by site-directed mutagenesis (Q5 High-Fidelity 2X Master Mix and Gibson Assembly Master Mix, NEB).

To express the monoclonal antibody, the plasmids of heavy and light chains were mixed in a 1:1 M ratio and transfected with PEI Max (Polysciences) in HEK293 FreeStyle (Invitrogen) cells according to the manufacturer’s protocol. The supernatant was collected and subject to Protein G (GE Healthcare) affinity chromatography. After the antibodies were bound to the Protein G resin, PBS was used to wash the column. Antibodies were then eluted with 0.1 M Glycine-HCl, pH 2.5, and immediately neutralized with 1 M Tris-HCl, pH 8.0. Antibodies quality was assessed by SDS-PAGE.

The cloning, production, and purification of bispecific antibodies

The DNA fragments encoding scFvs of the 1080 × 1B53 Series antibodies were generated by overlap PCR to include a linker SSGGGGSGGGGSGGGGS. The Fc fragments were mutated to introduce T350V, T366L, K392L, T394W (1080) and T350V, L351Y, F405A, Y407V (1B53) to promote Fc heterodimerization. The hinge region fragments were synthesized by Beijing RuiBiotech Company (Beijing, China) and assembled with the scFv fragments by overlap PCR. The scFv-hinge and Fc-vector were assembled with Gibson Assembly Master Mix (NEB).

1080×H2L3 antibody is constructed with a Knobs-into-Holes and DuetMab format.^29,53 In addition to the mutations in Fc, we introduced S127C (LC), C220V (LC), F121C (HC), and C215V(HC) mutations in the Ab-1080 arm. The light chain of the H2L3 arm of 1080-PEG₇-ACDTB (SA) was introduced a C80A mutation to prevent conjugation. Similarly, 1080-PEG₇-ACDTB (CA) has a C80A mutation in the light chain of the Ab-1080 arm. The mutations were generated by site-directed mutagenesis Gibson Assembly Master Mix (NEB).

To express the monoclonal antibody, the plasmids of heavy and light chains were mixed in a 1:1 M ratio and transfected with PEI Max (Polysciences) in HEK293 FreeStyle (Invitrogen) cells according to the manufacturer’s protocol. The supernatant was collected and subject to Protein G (GE Healthcare) affinity chromatography. After the antibodies were bound to the Protein G resin, PBS was used to wash the column. Antibodies were then eluted with 0.1 M Glycine-HCl, pH 2.5, and immediately neutralized with 1 M Tris-HCl, pH 8.0. Antibody quality was assessed by SDS-PAGE. Antibodies were purified using Superdex200 Column (GE Healthcare). After PNGase F (NEB) treatment and DTT reduction, Xevo G2-XS Q Tof (Waters Corporation) was used to analyze the molecular weights of heavy and light chains.

Antibody conjugation

Antibodies were adjusted to 5 mg/mL with PBS containing 1mM EDTA and reacted with a 50-fold molar excess of TCEP to reduce the capped and paired cysteines at room temperature overnight. The excess TCEP was removed and buffer-exchanged with PBS containing 1 mM EDTA (pH 7.4) using Zeba Spin Desalting Columns (Thermo Scientific, 89882) and Amicon Ultra-0.5 Centrifugal Filter (Millipore, UF₅₀1096). To re-oxidize the intramolecular disulfide bonds, the reduced antibodies were adjusted to 1.5 mg/mL and reacted with a 20-fold molar excess of (L)-dehydroascorbic acid (DHAA, SigmaAldrich) at 28°C for 2–3 h. The excess DHAA was removed by buffer-exchanging with PBS (pH 7.4) using Amicon Ultra-0.5 Centrifugal Filter (Millipore, UF₅₀1096). Re-oxidized antibodies were conjugated with a 3-fold molar excess of ACDTB-PEG_(n+3)-Mal in the presence of a 10% v/v DMSO, and the reaction was shaken at 28°C for 2–3 h. The excess ACDTB-PEG_(n+3)-Mal was removed by ultrafiltration. The formation of conjugates was confirmed by Xevo G2-XS Q Tof (Waters Corporation) analysis after deglycolysation with PNGase F (NEB, P0704S) and reduction with DTT (1mM). The concentrations of conjugates were detected by a bicinchoninic acid (BCA) assay (Thermo Fisher Scientific, Pierce BCA Protein Assay Kit) and stored in a −80°C freezer for further characterization.

Whole-cell patch-clamp recordings in Na_V1.7-HEK293 cell line

HEK 293 cells stably expressing human Na_V1.7 were isolated using 0.05% trypsin-EDTA treatment (Gibco) from the culture dish. Treated cells were cultured on poly-L-lysine-coated glass slides before recordings. Whole-cell recordings were performed at room temperature using a Multiclamp 700B amplifier with a Digidata 1550B (Axon Instruments). Pipette resistance was 3–4 MΩ. The extracellular solution consisted of (mM) 1 CaCl₂, 1 MgCl₂, 5 KCl, 10 HEPES, 10 Glucose, and 138 NaCl, adjusted to pH 7.4 with NaOH. The internal solution consisted of 10 EGTA, 10 HEPES, 5 NaCl, 2 MgCl₂, and 135 CsF, adjusted to pH 7.2 with CsOH. For tests at resting state, cells were voltage-clamped at a holding potential of −120 mV with a test pulse to 0 mV for 20 ms. For ACDTB, ACDTB-PEG_3,7,15, Ab-PEG_3,7,15-ACDTB, 1080-PEG₇-ACDTB (CA/SA), and ACDTB/1080 mixture, the Na_V1.7 current in the half-inactivated state is detected because of the state-dependency of ACDTB. Cells were clamped at a holding voltage of −120 mV, then a voltage of −60 mV for 8 s long was applied, and then the current was detected with a 0 mV pulse for 20 ms after giving a voltage of −120 mV for 20 ms. pClamp 10 software (Axon Instruments) and Graphpad Prism 8 were used for data analysis and fitting the IC₅₀ curve with a Hill (four-parameter logistic) fit. Each curve contained 5 concentrations, with two or three data points per concentration. ICE Bioscience Corporation completed the specificity tests on Na_V1.1 to Na_V1.6, and Na_V1.8. For channel kinetics, the internal solution consisted of 1 EGTA, 10 HEPES, 10 NaCl, and 140 CsF, adjusted to pH 7.2 with CsOH. The pipettes were pulled with a PC-10 puller with a resistance of 3–4 MΩ. Whole-cell membrane clamp recordings were performed using a Multiclamp 700B amplifier and an Axon Digidata 1550B (Axon Instruments). Sampling was performed using pClamp 10 software. The sampling frequency was 50 kHz with a low-pass filter of 10 kHz. All recordings were made at room temperature, and series resistance was compensated by 60%–80%. Data processing was performed using Clampfit 10. The voltage step stimulation protocols are shown in Figures 4C–4I.

Steady-state activation curves were generated using a Boltzmann equation.

$$\frac{G}{G_{\mathit{max}}}=\frac{1}{1+\exp \left[\left(V-V_{0.5}\right)/k\right]}$$ (Equation 1)

where G is the conductance, G_max is the maximal conductance of Na_V1.7 during the protocol, V is the test potential, V_0.5 is the half-maximal activation potential, and k is the slope.

Fast inactivation curves were generated using a Boltzmann equation.

$$\frac{I}{I_{\mathit{max}}}=\frac{1}{1+\exp \left[\left(V-V_{0.5}\right)/k\right]}$$ (Equation 2)

where I is the current at indicated test pulse, I_max is the maximal current of Na_V1.7 activation during the test-pulse, V is the test potential, V_0.5 is the half-maximal inactivation potential and k is the slope factor.

Recovery curves from fast and slow inactivation were fit using a two-phase exponential of the following equation.

$$\frac{I}{I_{\mathit{max}}}={Span}_{Fast}\ast [1-\exp \left(-\frac{t}{{\tau }_{Fast}}\right)]+{Span}_{slow}\ast [1-\exp \left(-\frac{t}{{\tau }_{Slow}}\right)]+y_0$$ (Equation 3)

Entry curves into fast and slow inactivation were fit using a single exponential of the following equation.

$$\frac{I}{I_{\mathit{max}}}=\left(y_0-Plateau\right)\ast \exp \left(-\frac{t}{\tau }\right)+Plateau$$ (Equation 4)

where I_max is the current at prepulse, I is the current at test pulse, y₀ is the non-inactivated current at the first pulse, t is the delay time between prepulse and test pulse, and τ is the time constant.

Preparation of Ab-1080 Fab and the complex for crystallization

The purified Ab-1080 antibody was digested using papain at a molar ratio of 132:1 (Sigma, P3125) and incubated with 40 mM EDTA (pH 8.0) and 40 mM L-cysteine at 4°C for 4 h. The digested Fab was purified using protein A column (GE Healthcare) and collected in the flow through. The Fab fragment was further purified using Superdex 75 column (GE Healthcare), and the buffer was replaced with 20 mM HEPES buffer (pH 7.40). The Na_V1.7 VSD_II domain S1-S2 loop peptide (EHHPMTEEFKN) was synthesized by Genewiz Corporation with over 98% purity. The Fab fragment and Na_V1.7 peptide were mixed at a molar ratio of 1:1.5 and incubated on ice for 30 min. The 1080 Fab-peptide complex was concentrated to 12 mg/mL in HEPES buffer for crystallization. Crystals were grown using the hanging drop method in hanging drops at 16°C. Drops with successful crystals contained equal volumes of protein mix and reservoir solution, including 0.2 M potassium thiocyanate and 25% W/V Polyethylene glycol 3,350. Crystals were protected by soaking in a reservoir solution with 20% V/V glycerol before being frozen in liquid nitrogen and data collection. Diffraction data were collected on the BL17U beamline at the Shanghai Synchrotron Research Facility (SSRF). Structural analysis was assisted by the Innovation Center for Structural Biology of Tsinghua University.

Docking of biparatopic antibodies and Na_V1.7

In order to explore the possible binding modes of Na_V1.7, a series of computational experiments were set up enrolling a protein-protein docking between Na_V1.7 and biparatopic antibodies. For Na_V1.7, the Cryo-EM structure of human Na_V1.7 at 2.2 Å resolution was set as the receptor (PDB ID:7W9K), while the biparatopic antibody structures predicted by AlphaFold-Multimer were selected as the ligands. Before docking, the model of biparatopic antibodies was relaxed by Desmond with a 200 ns Molecular dynamics simulation.

The Na_V1.7 and biparatopic antibodies structures were loaded as PDB files in Schrödinger 2022 and prepared with the embedded Protein Preparation Wizard application using default settings. The optimization of hydrogen bonds was performed to resolve structural ambiguities, and a final restrained minimization of the system was carried out under the OPLS4 force field.

The protein structures were loaded in PDB format to the Schrodinger software, and the docking process was submitted using the protein-protein docking protocol. The docking results and the PDB structures of the complexes for the 30 top-ranked docking poses were then retrieved and the best complex was selected for visualization with PyMol.

Docking of Na_V1.7 and ACDTB

To elucidate the structure of the Na_V1.7/ACDTB complex, a series of computational experiments involving Induced-fit docking (IFD) studies was conducted. Cryo-EM structure of human Na_V1.7, with a resolution of 2.2 Å (PDB ID: 7W9K), was employed as the receptor. The chosen binding site was the VSD_IV pocket. The structures of the Na_V1.7 protein and ligand were loaded into Schrödinger 2023 software as PDB files and prepared using the embedded Protein Preparation Wizard with default settings. Hydrogen bonds were optimized to resolve structural ambiguities, followed by a final restrained minimization of the system using the OPLS4 force field.

For the docking process, the induced-fit docking protocol within the Schrödinger software was employed. The Na_V1.7 and ACDTB ligands were subjected to IFD and Prime MM-GBSA studies to observe the docking score. After the docking process, the top 30 ranked docking poses and the PDB structures of the corresponding complexes were retrieved. The optimal complex, determined by the docking score, was selected for visualization using PyMOL.

Synthesis of ACDTB-PEG_(n+3)-Mal

Unless otherwise noted, chemicals, building blocks, and solvents were purchased from J&K Scientific, Confluore, Accela ChemBio, Sigma Aldrich, Bidepharm, and others and were used without further purification. Flash column chromatography was performed using silica gel (200−300 mesh, Weina) columns. ¹H NMR and ¹³C NMR spectra were recorded on a Bruker Avance 400 (¹H: 400 MHz, ¹³C: 101 MHz) NMR spectrometer. Chemical shifts were calculated as δ (ppm) relative to tetramethylsilane (TMS) in dimethyl sulfoxide (DMSO)-d6. Coupling constants (J) are reported in Hertz. High-resolution mass spectra (HRMS) were measured with Waters Synapt G2-Si Qtof MS using ESI as the ionization source. Preparative HPLC was performed on an SHIMADZU LC20AT system using a GL Sciences ODS-3 14 × 250 mm, 5 μm column. The gradient was 5% acetonitrile in 5 min and 5–95% acetonitrile at 55 min, hold on 95% acetonitrile for 5min, and then 95-5% acetonitrile in 5 min, the flow rate was 10 mL/min. The other mobile phase was ultrapure water with 0.5% formic acid or 50 mM NH₄HCO₃. All of the final compounds reported were >95% pure based on ¹H NMR. The detailed synthetic methods are listed in the supplementary information.

Proteomic study

SDS-PAGE was performed before the proteomic study. Gel bands corresponding to the light chain were cut out and subjected to in-gel digestion. Firstly, the proteins were reduced with 25 mM DTT and then alkylated with 55 mM iodoacetamide (SigmaAldrich). Next, sequencing-grade chymotrypsin (Promega Corporation) was used for in-gel digestion carried out at 37°C overnight in 50 mM ammonium bicarbonate. The peptides were extracted twice using a 1% trifluoroacetic acid in a 50% acetonitrile aqueous solution for 30 min and then concentrated using SpeedVac centrifugation.

To conduct LC-MS/MS analysis, peptides were separated through a 120-min gradient elution at a flow rate of 0.300 μL/min using a Thermo-Dionex Ultimate 3000 HPLC system. A homemade fused silica capillary column (75 μm ID, 150 mm length; Upchurch, Oak Harbor, WA) packed with C-18 resin (300 A, 5 μm; Varian, Lexington, MA) was used as the analytical column. The mobile phase A consisted of 0.1% formic acid, while mobile phase B consisted of 100% acetonitrile and 0.1% formic acid. The HPLC was connected directly to the Thermo Orbitrap Fusion mass spectrometer. The Orbitrap Fusion mass spectrometer was operated through Xcalibur3.0 software in the data-dependent acquisition mode. The analysis began with a single full-scan mass spectrum in the Orbitrap (350–1550 m/z, 120,000 resolution) followed by 3 s of data-dependent MS/MS scans in an Ion Routing Multipole at 30% normalized collision energy (HCD). Every LC-MS/MS run’s MS/MS spectra were searched against the chosen database using the Proteome Discovery searching algorithm (version 1.4).

Affinity measurement with bio-layer interferometry

We evaluated antibody affinity by measuring K_D of Na_V1.7 VSD_II domain S1-S2 loop peptide (EHHPMTEEFKN) to Ab-1080 antibody and 1080 Fab. K_D was measured using bio-layer interferometry method by Octet RED 96 (FortéBio, Inc.). The peptide was incubated with EZ-Link Sulfo-NHS-LC-Biotin (Thermo Scientific) at a molar ratio of 1:0.8 at room temperature for 30 min and diluted to 25 μg/mL in PBS. After recording the baseline with PBS, the peptide was loaded onto SA sensor chips until the capture levels reached 0.5 nm. After the signal stabilized for at least 100 s, the association and dissociation kinetics of Ab-1080 and Ab-1080 Fab with different concentrations were measured for 300 s. Due to the presence of non-specific binding in the Ab-1080 Fab sample, the chip was blocked by adding 1% BSA to the buffer, while the buffer for Ab-1080 was PBS. All experiments were performed at room temperature. Data processing and analysis used Octet analysis software version 9.0 (FortéBio, Inc.). Data is shown in Table S2.

Affinity determination with LigandTracer green

EDTA was added to 1080-PEG₇-ACDTB, H2L3-PEG₇-ACDTB, and Ab-1080 at a final concentration of 1 mM. 50×TCEP was added and the reaction was carried out at room temperature for 2 h. 10×Alexa Fluor 488 C5 maleimide was added for conjugation at 4°C overnight. The fluorescently labeled antibodies were then purified by ultrafiltration. The binding kinetics of Ab-1080-488, 1080-PEG₇-ACDTB-488 and H2L3-PEG₇-ACDTB-488 to Na_V1.7-HEK293 cells were measured using LigandTracer Green (Ridgeview Instruments, Uppsala, Sweden). After trypsin digestion of Na_V1.7-HEK293, 10⁶ cells were diluted into 2 mL of medium. The Petri dishes were padded 1 cm high and the cells were placed on the tilted bottom of the Petri dishes. After the cells were adhered to the dish bottom for 6 h, the medium was replenished to 10 mL and the dishes were placed horizontally for overnight incubation. Cells were plated in a dedicated area of a 10 cm Petri dish 24 h before the experiment. Cells were fixed with 95% ethanol, washed with PBS, and then blocked with PBS containing 5% BSA before the test. Labeled antibodies were added stepwise to the dish, starting from the lowest concentration. The next concentration was added when the signal had reached equilibrium. The solution was replaced by fresh PBS (5% BSA) to record the dissociation from the cells. Data were analyzed and ka, kd, and K_D were determined by TraceDrawer Software using the One-to-One Two-State model (Version 1.9.2, Ridgeview Instruments, Uppsala, Sweden).

Immunofluorescence

Mice were injected with 10 mpk of 1080-PEG₇-ACDTB or H2L3-PEG₇-ACDTB in hind paws. After 48 h, hind paw tissue was taken and covered with cryo-embedding media OCT in a tissue box. Then the box was frozen in liquid nitrogen for approximately 20 s. Next, we transferred the tissue to a cryotome cryostat (Leica CM1950) and sectioned it into 8 μm slices. After fixing with 95% ethyl alcohol, tissue sections were blocked using 5% BSA in PBS at room temperature for 1 h.

Mouse DRG sections with a thickness of 6 μm had been pre-fixed using 4% PFA for 2 h and stored at −80°C. After restoring the frozen sections to room temperature, the sections were fixed again with 4% PFA for 10 min. The sections were incubated for 1 h at room temperature with PBS solution containing 5% BSA.

The sections were incubated with primary antibodies at 4°C overnight. The primary antibodies include anti-PGP9.5 (1/200 dilution, Abcam), 1080-PEG₇-ACDTB, and H2L3-PEG₇-ACDTB with a final concentration at 5 μg/mL. Then the sections were washed three times with PBST (0.2% Tween 20) for 5 min each. The sections were incubated with secondary antibodies at room temperature for 1 h. The secondary antibodies include donkey anti-rabbit IgG 594 (Thermo Scientific) with a final concentration of 5 μg/mL and goat anti-human IgG 488 (Thermo Scientific) with a final concentration of 2 μg/mL washing steps were repeated. After anti-fluorescence quenching treatment, the images were acquired using Zeiss LSM880 with Airyscan.

Serum stability assay

10 μg of 1080-PEG₇-ACDTB were incubated at 37°C in 50 μL mouse serum, resulting in a final concentration of 200 ng/μL. During the incubation, aliquots were collected at different time points and frozen at −80°C. Goat anti-human IgG Fc secondary antibody (Sino biological, SSA015) was biotinylated by 5×sulfo-NHS-biotin at room temperature for 30 min to generate the capture antibody. The capture antibody was added into the aliquots to allow a 2-h incubation at room temperature. Then the mixture was incubated with streptavidin magnetic beads for 1 h at room temperature under gentle shaking. The magnetic beads were washed with PBS containing 0.1% Tween 20 two times and washed with HBS-EP buffer at 37°C overnight to remove non-specific binding. 1080-PEG₇-ACDTB was eluted with 100 mM Glycine (pH 2.5), and neutralized with Tris-HCl (pH 8.0). The samples were then reduced with 1 mM DTT and subjected to LC-MS analysis. The standard sample of 1080-PEG₇-ACDTB was diluted to 10 μL with PBS and mixed with 40 μL of mouse serum to give final concentrations of 20 μg/mL, 100 μg/mL, 200 μg/mL, 300 μg/mL, and 500 μg/mL. After the separation of the antibody by the magnetic beads and LC-MS analysis, the Area Under Curve (AUC) of the characteristic light chain peak of 1080-PEG₇-ACDTB was calculated. The peak areas were then plotted against the concentrations in a standard curve. The Area Under the Curve (AUC) of light chain peaks was used to quantify the conjugate.

Pharmacokinetic (PK) study

A pharmacokinetic (PK) study was conducted with 8 week-old female C57/BL-6J mice from THU-LARC. 1080-PEG₇-ACDTB in PBS (10 mM, pH 7.4) was dosed to mice at 30 mpk by intravenous injection. Blood samples were taken by retro-orbital sampling at various time points using a composite sampling scheme with no more than 2 samples taken from an individual mouse. For each sample, approximately 200 μL of blood was collected into a serum separator tube via retro-orbital blood collection using a glass pipette. Samples were allowed to clot at room temperature for 1h and then centrifuged under refrigerated conditions (4°C) for 10 min at approximately 3000 rpm. Serum was transferred to a 1.5 mL tube and frozen at −80°C until analysis. Goat anti-human IgG1 Fc secondary antibody (Sino Biological Inc, SSA015) was biotinylated by a 5-fold molar excess of biotin reagent (Thermo Fisher, 21335) at room temperature for 30 min. The excess biotin is removed by the desalting column (Thermo Fisher, 89882) and ultrafiltered with a 10K ultrafiltration tube (Millipore, UF₅₀10BK). 125 μL (500 μg) pre-washed streptavidin magnetic beads (NEB, S1420S) were added into a 1.5 mL microcentrifuge tube. The serum containing 1080-PEG₇-ACDTB and 20 μg of biotinylated antibody were added to the beads to allow an incubation for 1 h at room temperature. Subsequently, the beads were washed with wash/binding buffer 3 times. 20 μL Elution buffer (0.1 M glycine, pH 2.0) was applied to elute the 1080-PEG₇-ACDTB. After neutralization with Tris-HCl (1M, pH 8.0) and reduction with 1mM DTT, the samples were analyzed with LC-MS/MS. The Area Under the Curve (AUC) of light chain peaks was used to quantify the conjugate.

Preparation of fluorescence-labeled 1080-PEG₇-ACDTB

1080-PEG₇-ACDTB (1 mg, 6.7 nmol, 500 μL) in PBS (0.01 M, pH7.4, 1 mM EDTA) was mixed with TCEP (26.8 nmol in PBS, pH 7.0) and then shaken at room temperature (RT) for 1.5 h. To the reduced antibody was added AF647-C2-Mal (26.74 nmol in DMSO, ThermoFisher). The resulting mixture was shaken at RT for 1 h. After incubation, Cysteine (5 equivalents) was added to block the remaining reactive maleimide group. The reaction mixture was rotated at room temperature for another 1 h. Unbound dye molecules were removed by repetitive centrifugation (12,000 rpm, 9 min at a time, 8 times) with an ultrafiltration centrifuge tube (MWCO 30 kD, 0.5 mL, Millipore) until no fluorescence was detected in the supernatant. The concentrations of conjugates were detected by a BCA assay (Thermo Fisher Scientific, Pierce BCA Protein Assay Kit) and stored in a −80°C freezer for further usage.

Local retention behavior of 1080-PEG₇-ACDTB

To determine the local retention behavior of 1080-PEG₇-ACDTB, C57/BL-6J mice were randomly divided into three group (3 mice in each group), in which the C57/BL-6J mice were given PBS (20 μL), 1080-PEG₇-ACDTB-AF647 (20 μL, 2.5 and 10 mpk) intraplantarly into the left footpad. The plantar imaging was visualized by IVIS Lumina III in vivo imaging system (PerkinElmer) at predetermined time points after injection, and the relative fluorescent intensity was calculated in the same system.

Mouse DRG preparation

The dorsal root ganglia (DRGs) of mice were extracted from the thoracic and lumbar regions of C57/BL-6J mice (9–12 weeks old, female) using a previously described method.⁵⁴ Following isolation, the DRGs were enzymatically digested with collagenase (Gibco) and papain (Worthington). The harvested DRG cells were then seeded onto glass coverslips pre-coated with poly-L-lysine and laminin. Subsequently, the cells were cultured in a DMEM/F12 (1:1) medium supplemented with Pen Strep (1%, Gibco), NGF (Nerve Growth Factor, 100 ng/mL, Sino Biological), GDNF (Glial cell line-derived Neurotrophic Factor, 50 ng/mL, Sino Biological), and Ara-C (cytosine arabinoside, 10 μM, Sigma-Aldrich). The DRG cells can be maintained for one week under this condition.

Whole-Cell voltage-clamp recording on mouse DRG

Voltage-clamp recordings were conducted at room temperature using a Multiclamp 700B amplifier coupled with a Digidata 1550B acquisition system (Molecular Devices). The extracellular solution contained (in mM) 1 CaCl₂, 1 MgCl₂, 5 KCl, 10 HEPES, 10 glucose, and 138 NaCl, with the pH adjusted to 7.4 using NaOH. The intracellular solution was composed of (in mM) 2 MgCl₂, 10 EGTA, 10 HEPES, 5 NaCl, and 135 CsF, with the pH adjusted to 7.2 using CsOH. Pipette resistances ranged from 3 to 4 MΩ, and series resistances were compensated by 80%. The mouse DRG cells were cultured for at least one day prior to the experiment. As Na_V1.7 is more prominently expressed in small-diameter mouse DRG neurons, we recorded neurons with a whole-cell capacitance ranging from 10 to 25 pF. DRG neurons were held at a potential of −60 mV, and the half-inactivated state current was measured using a 0 mV pulse for 20 ms following a −120 mV pulse for 20 ms. Before the test, A-803467 (MedChemExpress) was applied at a concentration of 100 nM to selectively block Na_V1.8 currents, and BSA was added at a concentration of 5 μg/mL to prevent adsorption.

Von frey mechanical sensitivity test

Before the baseline test, mice were placed into the testing arenas and habituated for at least a week, 20 min per day. After habituation, the filament was positioned below the mid-planter surface of the foot when the mouse had all four feet on the platform. The response was marked as “O” if the mouse didn’t react, or “X” if it moved its foot. The next filament was tested using the up-down pattern until 6–9 trials had been completed. 24 h after the baseline measurement, mice were anesthetized with isoflurane. We injected H2L3-PEG₇-ACDTB, Ab-1080, and 1080-PEG₇-ACDTB with a dose of 10 mpk into the right hind paw and re-tested mice with filaments using the same procedure as described above at different time points after the injection. 50% paw withdrawal threshold was calculated and analyzed.

Acetic acid-induced writhing test

H2L3-PEG₇-ACDTB, Ab-1080, and 1080-PEG₇-ACDTB (50 mpk, 200 μL in volume) were injected into mice by i.v. injection 24 h before the acetic acid writhing experiment. And 0.7% glacial acetic acid was injected intraperitoneally. The videos were recorded within 60 min after injection and counted the number of writhing times within the first 20 min. Writhing is defined as a stretch, extension of hind limbs, and contraction of the abdomen so that the percentage inhibition of writhing was calculated. The inhibition% = (mean number of writhing for control - mean number of writhing for test)/(mean number of writhing for control) × 100%.

Dose-dependent analgesic effect in acetic acid-induced writhing test

Based on the previously established method, groups of C57/BL-6J mice (n = 6) were administrated with vehicle (PBS, 2.5 mL/kg, IV), 1080-PEG₇-ACDTB (10, 25, or 50 mpk, IV), and positive control PF-05089771 (Bide Pharmatech, 25 mpk, IG). After 24 h, each mouse was intraperitoneally injected with 0.7% acetic acid and the number of writhing was recorded for 20 min after acetic acid injection. The dose-dependent analgesic effect was expressed as number of writhing compared to the control.

Hot plate test

The mice were habituated for one week in advance, and each time the mice were placed on hot plate at room temperature for 10–20 min. The hot plate was set to a temperature of 53°C. Mice were placed on the surface of a hot plate (BIO-CHP, Bioseb) and covered by a transparent box. The time between placement and hind paws licking or shaking or jumping was recorded as a base latency of response. Subsequently, the mice received intrathecal injections (10 μL) of H2L3-PEG₇-ACDTB, Ab-1080, and 1080-PEG₇-ACDTB with a dosage of 4 mpk using 29 G needle. The injection location was the L5/L6 intervertebral space of the mouse spine. Mouse tail flopping was used as a criterion for successful injection. Response latency was tested again 24 h and 48 h after the injection. The %MPE for thermal pain = (postdrug latency - baseline latency)/(baseline latency) × 100%.

Quantification and statistical analysis

All graphs with error bars or statistical significance in this study were generated by Graphpad Prism. Two-tailed Student’s t test with or without Welch’s correction was performed in this study to compare values between different groups. The statistical methods were also indicated in the figure captions. The p values, t values, and degree of freedom (df) are listed in Figures S4–S6. Results are presented as mean ± SEM of at least three replicates unless otherwise stated. Biological replicates (n) are indicated in the figure captions. Statistical significances are as follows: n.s., not significant; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Published: October 25, 2024