DRG neurons express a variety of VGSCs. Despite lacking a specific population of channels for new inhibitor studies, these cells provide an outstanding platform to study VGSCs natively expressed in their physiological environment. Additional benefits of studying VGSCs in DRG neurons are the practicality of cell preparation, the cell sizes ranging from 20 to 50 μm, and the absence of dendrites that may result in space clamp issues when these structures are present. We used voltage-gated K⁺ channel blockers, voltage gated-Ca²⁺ channel blockers, and online leak subtraction to isolate the voltage-activated I_Na for our experiments (see Methods).

We tested the fast inhibition by EUG on I_Na expressed in DRG neurons and compared the results with the inhibition produced by LID, a local anesthetic used in clinics that inhibits VGSCs to produce anesthesia. In a separate batch of experiments, we used 300 nM of tetrodotoxin (TTX) during the whole experiment to inhibit only the TTX-sensitive part of the total I_Na in DRG neurons (Roy and Narahashi, 1992; Elliott and Elliott, 1993; Ogata and Tatebayashi, 1993; Tan et al., 2014). This procedure enables the recording of only the TTX-R fraction of I_Na that is important in the peripheral nervous system as it relates to sensorial information such as the ones produced by noxious stimuli (Djouhri et al., 2003; Gudes et al., 2015; Bennett et al., 2019).

EUG and LID quickly inhibit the total I_Na and TTX-R I_Na in a concentration-dependent manner, and the inhibition is fully reversible after drug washout (Figure 1A, B). By using an in-house fast single-cell superfusion system, we show that EUG inhibits I_Na in less than 5 s upon its addition to the experiments. The inhibition may be even faster than 5 s, but we used a time series of 0.2 Hz to avoid inhibition accumulation during faster series. In addition, we used a holding potential of −110 mV in all our experiments in an effort to remove VGSCs from their inactivated states, to maximize the currents we recorded.

EUG was solubilized with ethanol and then diluted in experimental (external) solution to the desired concentration. The maximal final concentration of ethanol was 0.73% (vol/vol), and this amount was used to solubilize EUG for 10 and 30 mM working solutions only. For lower final concentrations of EUG, we used a molar solubilization ratio of 1:21 (EUG:ethanol). At 0.73% (vol/vol), ethanol reversibly blocks 10% of the total I_Na peak (Figure 2). For reference, the concentration of ethanol used to solubilize EUG at 2 mM was 0.2%.

EUG inhibits the total I_Na with an IC₅₀ value of 2.27 ± 0.07 mM (n = 30) and TTX-R I_Na with an IC₅₀ value of 2.21 ± 0.08 mM (n = 32). Using an identical approach, we found that LID inhibited the total I_Na with an IC₅₀ value of 1.42 ± 0.17 mM (n = 24) and TTX-R I_Na with an IC₅₀ value of 0.44 ± 0.09 mM (n = 26) (Figure 1C). All IC₅₀ values for the inhibition of I_Na and TTX-R I_Na by EUG and LID are shown in Table 1. It is noteworthy to mention that none of the EUG concentrations utilized in the present study affected the membrane resistance of the neurons. As an example, we show paired data with absolute membrane resistance values before and after 3 mM EUG is added to the experiment (Figure 1D). For these experiments, we used a hyperpolarizing pulse to −130 mV from the usual −110 mV holding potential. In addition, we tested EUG on action potential firing in a representative neuron. As predicted, 1 mM EUG prevents neurons from firing action potentials, a neuronal capability fully recovered a few seconds after EUG washout (Figure 1E).

Remarkably, the inhibition profile of the total I_Na by EUG differs from that of LID in the slope of the dose-dependent inhibition curves, the Hill coefficient (EUG slope = 1.95 ± 0.120 vs. LID slope = 1.02 ± 0.169; extra sum-of-squares F-test p < 0.0001), suggesting that at least part of the inhibiting mechanism is different between these two drugs.

From the holding potential of −110 mV, we applied depolarizing voltage steps from −90 mV to +50 mV in increments of +5 mV every 5 s (0.2 Hz). EUG at a concentration of 2 mM inhibits the total I_Na and TTX-R I_Na, as activated by all voltages without significantly affecting the reversal potential of the currents. Comparable results were found when LID at a concentration of 1 mM was tested as an inhibitor using identical procedures (Figure 3).

Interestingly enough, the persistent total I_Na after 50-ms depolarization was remarkably inhibited by EUG but not by LID (Figures 4A–D). Nevertheless, neither EUG nor LID changed the kinetics of the inactivation process of total I_Na or TTX-R I_Na during a test pulse (Figures 4E–H), as discussed later.

We transformed I_Na peak values from each separate cell into Na⁺ conductance by using Ohm’s law (Eq. 4; see Methods) before normalizing all values by their maxima. We then plotted these cell-specific data against the activating membrane potential to produce conductance–voltage (G-V) curves in the absence and presence of 2 mM EUG or 1 mM LID for comparison. We also averaged data to highlight voltage-dependence shifts when they existed (Figures 5A, D, G, J). To the individual cell normalized Na⁺ conductance curves, we fitted Boltzmann’s equation (inactivation kinetics; Eq. 5; see Methods) for the individual voltage dependence of the activation process (V _1/2-act ) and the voltage sensitivity of the process (max slope-act). On average, EUG significantly shifted V _1/2-act of the total I_Na but not of the TTX-R I_Na to more depolarized membrane potentials, and it decreased the max slope-act of both total I_Na and TTX-R I_Na (Figures 5A–F). Data from the inhibition of I_Na by LID show no changes in V _1/2-act of both total I_Na and TTX-R I_Na and a change in the max slope-act of TTX-R I_Na only (Figures 5G–L). Figures 5A, D, G, J show the average curves that hide the individual variations in V _1/2-act and max slope-act. Averaged fitting parameters and statistical analysis are shown in Table 2.

In addition to activation, one of the greatest properties of VGSCs is their fast inactivation process. This process limits the action of a VGSC to a few milliseconds, thus preventing long depolarizing periods that would impair the right functioning of cells in most cases. We tested EUG on the inactivation process of I_Na by studying the level of inactivation of the total I_Na and TTX-R I_Na after many depolarized (−140 mV–0 mV at every 5 mV) 100-ms conditioning pre-pulses. The inactivation level was tested by a pulse to +20 mV right after the conditioning pulse (Figure 6A). Graphically, we expressed the normalized and average levels of non-inactivated I_Na against the conditioning pre-pulse voltage to build steady-state inactivation curves (Figures 6B, E, H, K). Individual inactivation curves were fitted with Boltzmann’s equation (Eq. 6), similar to what was done with G-V curves for their voltage dependence (V _1/2-inact ) and voltage sensitivity (max slope-inact). EUG enhances the voltage-dependent inactivation of the total I_Na and TTX-R I_Na by shifting their voltage dependence to more negative voltages. The shift in the curves from the total I_Na was highly significant, nearly −15 mV on average. The shift in the curves from TTX-R I_Na was also significant and −5 mV on average. LID also shifted the inactivation curves of both the total I_Na and TTX-R I_Na to more negative potentials but not to the extent that EUG did, with −5-mV and −4-mV shifts on average, respectively. Neither EUG nor LID effectively changed the voltage sensitivity of the fast inactivation process of I_Na (Figure 6). Averaged fitting parameters and statistical analysis results for I_Na inactivation analysis are shown in Table 2.

We evaluated the speed of the recovery from the fast inactivation of the total I_Na and TTX-R I_Na in DRG neurons. To do so, we used the classic three-pulse voltage-clamp protocol, which consists of a pulse P₁ to +20 mV lasting 50 ms to fully fast inactivate I_Na, a duration-varying pulse P₂ at a holding potential of −110 mV to recover the currents from inactivation, and a pulse P₃ to +20 mV once again to activate the fraction of the I_Na recovered during P₂. Typical current recordings show small differences in the kinetics of recovery from inactivation before and after EUG is added to the experiment (Figure 7A). All I_Na peaks recorded during P₃ were normalized by the corresponding (same sweep) I_Na peak during P₁, averaged, and plotted against the duration of P₂ (Figures 7B, F, J, N). Individual plots from the same cell before and after EUG was added to the experiment were fitted with two exponentials for a fast and slow kinetic component and a proportion amplitude component that we chose to represent as a percent of the fast component (Eq. 7). The statistical analyses of these parameters show that in the experiments with the total I_Na, but not with TTX-R I_Na, the percentage of fast component of the recovery from inactivation was significantly decreased from 80% to 74% (Figure 7C). In this same experiment, the slow recovery component of the recovery from inactivation was significantly delayed by EUG (Figures 7B–E). Because most of the recovery of the inactivation process happens at a fast pace (fast component), the delayed slow component did not strongly affect the overall recovery from inactivation. EUG did not affect any parameter of recovery from the inactivation of the TTX-R I_Na (Figures 7F–I). For comparison, we ran the same experiments with LID, which is known to delay I_Na recovery from inactivation. LID substantially and significantly delays the recovery from the inactivation of both total I_Na and TTX-R I_Na, mainly by affecting the fast component of the process. In the total I_Na, LID significantly decreases the percent fast component, and it delays both the fast and slow components of the process (Figures 7J–Q). In TTX-R I_Na, LID acts by significantly delaying the fast component of the process only. Averaged fitting parameters and statistical analysis results for I_Na recovery from the inactivation analysis are shown in Table 3.

We tested the inhibition intensity of the total I_Na and TTX-R I_Na by EUG in a time series of depolarizations at 2 and 5 Hz (Figure 8). This test shows the inability of the VGSCs to recover from inactivation in a period between two depolarization events. The voltage clamp protocols used here consisted of 50-ms depolarizing pulses; therefore, they were characterized by a duty cycle of 10% when stimulation was at 2 Hz and 25% when stimulation was at 5 Hz. In other words, at 2 Hz, the I_Na represented the recovered currents during a 450-ms period, and at 5 Hz, the recovered currents during a 150-ms period. It is noteworthy that the first pulse in the series was taken after at least 1 min at a holding potential of −110 mV. Typically, 2 mM EUG does not have its fractional inhibition of total I_Na significantly increased at 2 Hz, from 0.53 ± 0.049 in the first pulse of the series to 0.55 ± 0.061 in the 20th pulse of the series. At a stimulation of 5 Hz, the fractional inhibition of the total I_Na was significantly increased from 0.49 ± 0.056 in the first pulse to 0.55 ± 0.062 in the 20th pulse (two-tailed paired t-tests, p = 0.20 for 2 Hz and p < 0.01 for 5 Hz). When EUG was used to inhibit TTX-R I_Na, the fractional inhibition was significantly increased from 0.37 ± 0.051 to 0.44 ± 0.064 at 2 Hz and from 0.37 ± 0.046 to 0.49 ± 0.061 at 5 Hz (two-tailed paired t-tests, p < 0.01 for 2 Hz and 5 Hz). For comparison, we also tested LID under the same conditions. At a stimulation of 2 Hz, LID performs similar to EUG both on total and TTX-R I_Na. With LID as an inhibitor, the total I_Na inhibition was significantly increased from 0.46 ± 0.034 to 0.54 ± 0.044 at a stimulation of 2 Hz and from 0.47 ± 0.021 to 0.81 ± 0.055 at 5 Hz (two-tailed paired t-tests, p < 0.01 for 2 Hz and 5 Hz). On TTX-R I_Na, inhibition by LID was significantly potentiated from 0.40 ± 0.025 to 0.55 ± 0.040 at 2 Hz and from 0.37 ± 0.027 to 0.64 ± 0.054 at 5 Hz (two-tailed paired t-tests, p < 0.01 for 2 Hz and p < 0.001 for 5 Hz). Fractional inhibition data and statistical analysis results for the I_Na inhibition series at 2 Hz and 5 Hz are shown in Table 4.

VGSCs continue to undergo conformational changes after fast inactivation is completed. After long depolarizations, the channels are in the commonly called slow inactivated states (Vilin and Ruben, 2001). Because of the fast inactivation, slow inactivation cannot be observed directly as current decay like the fast inactivation. We studied slow inactivation with a three-pulse protocol consisting of a duration-varying P₁ pulse to +20 mV intended to activate and inactivate VGSCs, followed by a 20-ms period at holding P₂ to recover I_Na from fast inactivation, and finally, a P₃ pulse to +20 mV once again that is intended to activate the available I_Na (Figure 9). We plotted averaged P₃/P₁ current ratios against the duration of P₁ to estimate the kinetics of inactivation entry. P₁ pulses up to 50 ms in duration mostly measure fast inactivation entry. Longer depolarizations drive the channels into slow inactivated states. We used two exponentials (Eq. 8) to fit the data to extract the level of inactivation when P₁ = 2 ms, as well as the percent of the fast component and the time constants of the P₃/P₁ current ratio decay that we use to infer about the slow-inactivation entry kinetics. EUG does not significantly affect any of these two parameters from the total I_Na or TTX-R I_Na. In turn, our data using LID as an inhibitor show it significantly affects both the level of inactivation when P₁ = 2 ms and the fast and the slow inactivating components of the slow inactivation process. All parameters from the analysis above are summarized in Table 5.

Our data showed that EUG significantly and consistently shifts the total I_Na activation curves to more depolarized potentials (Figure 5). In addition, we showed that EUG shifts the steady-state inactivation curve to more negative potentials (Figure 6), although without inducing remarkable changes in the recovery from fast inactivation (Figure 7) and on the use-dependent inhibition that is a hallmark for local anesthetics like LID (Figure 8). One of our hypotheses that would explain the abovementioned data is that EUG interacts with and stabilizes pre-open–closed states of VGSCs (Sunami and Hiraoka, 1996). These pre-open–closed states are short-lived non-conducting states of the channels that are, therefore, not trivial to study (Armstrong, 2006; Lenaeus et al., 2017; Xiao et al., 2021; Gamal El-Din and Lenaeus, 2022). We designed a protocol, which, to the best of our knowledge, has not been used before exactly as is to study the possible interaction between EUG and the pre-open states of VGSCs. The protocol consists of depolarizing the membrane of the neurons from a −110-mV holding potential to −70 mV for various different durations (0–420 m) in a time series. The −70 mV was chosen based on the steady-state inactivation curves as the membrane potential that does not lead to current activation (Figure 5) and that, however, produces little inactivation (Figure 6) in the absence of drugs. Upon depolarizing the membrane potential to −70 mV, from our standard holding potential of −110 mV, we quickly shift the equilibrium of VGSCs to pre-open–closed states (Armstrong, 2006). The equilibrium quickly favors pre-open–closed states over the resting states upon stepping the voltage to −70 mV since the voltage sensors of VGSCs move within a few milliseconds at −70 mV (Lacroix et al., 2013). In addition, a much slower process takes channels from those pre-open–closed states to an inactive state of the VGSCs (Figure 10) (Armstrong, 2006). Our rationale was that if EUG interacts with such pre-open–closed states, the available current to be activated by P₂ in our custom protocol would decrease in amplitude with the extended duration of P₁. We plotted the normalized average inhibition intensity of the total I_Na recorded with P₂, induced by EUG, against P₁ duration (Figure 10). For comparison, we repeated the same experiment and data analysis but using LID as the total I_Na inhibitor. Our data show that EUG, but not LID, inhibited I_Na after a conditioning pulse (P1) to −70 mV shorter than 100 ms in an exponential time course with a time constant of approximately 72 ms. For comparison, LID induces similar effects in decreasing the current at pulse P₂ of this protocol, with kinetics of 709 ms. The kinetics of EUG inhibition potentiation by P₁ was statistically different from that of LID (Student’s t-test, p = 0.04). In fact, at −70 mV, some level of inactivation exists (Figure 10A), but we cannot rule out the notion that EUG may bind to pre-open–closed states, and this effect would be reflected in the steady-state inactivation curves as well. This matter is currently being worked out in terms of a kinetic model to further strengthen this notion in a separate paper. There is currently limited information about drug binding to pre-open–closed states of mammalian VGSCs (Gilliam et al., 1989; Chernoff, 1990; Carmeliet and Mubagwa, 1998; Nesterenko et al., 2011), and this is one of our lines of investigation moving forward.

Our data suggest that EUG and LID may inhibit I_Na with a different mechanism. To further corroborate this notion, we built additional I_Na-inhibiting dose–response curves by EUG and LID but in the presence of LID and EUG, respectively. First, we established a pre-inhibition of 25% of total I_Na with LID at 0.5 mM. This was planned to leave the other 75% of the current amplitude for the building of an inhibiting dose–response curve using EUG. All currents inhibited in these experiments were fully recovered upon drug washout. A ∼25% pre-inhibition of the total I_Na with LID was associated with a statistically significant change in the IC₅₀ value for the inhibition by EUG. In the presence of LID at 0.5 mM, EUG inhibits the total I_Na with an IC₅₀ value of 1.84 ± 0.274 mM (vs. 2.27 ± 0.075 mM in the absence of LID; extra sum-of-squares F-test, p < 0.01). In addition, the Hill coefficient for the inhibition of I_Na by EUG was reduced to 0.98 ± 0.140 mM (n = 12) by pre-inhibiting the total I_Na with LID (vs. 1.95 ± 0.120 with EUG alone; extra sum-of-squares F-test, p < 0.0001; Table 1; Figures 11A, B).

Next, we carried out the opposite experiment with regard to inhibitors. First, we pre-inhibited around 25% of the total I_Na with 1.3 mM EUG and proceeded with a dose–response curve for the inhibition of the remaining 75% of the currents with LID. This time, the remaining 75% of I_Na was inhibited by LID with a significantly lower IC₅₀ value of 0.73 ± 0.126 mM (n = 10) vs. 1.42 ± 0.169 with LID alone (extra sum-of-squares F-test, p < 0.0001) and with no significant changes in the Hill coefficient compared with LID alone (Hill coefficient of 1) (Table 1; Figures 11C, D).

Previous studies have reported the inhibiting properties of EUG on VGSCs (Park et al., 2006; Cho et al., 2008; Park et al., 2009; Moreira-Lobo et al., 2010). In one of these studies, EUG was tested on the total I_Na and TTX-R I_Na, both activated from a holding membrane potential of −80 mV (Cho et al., 2008). Their recordings show a considerable rundown of currents during the experiments. In addition, the inhibition of I_Na induced by EUG shown in that study was not fully recovered to its initial values. Finally, that study exposed the neurons to EUG for 10 min, which, together with the drug application technique, could have caused I_Na rundown during the course of the experiments. We propose that these four drawbacks contributed to a possible overestimation of EUG affinity to block I_Na in the study conducted by Cho et al. (2008).

Our dose–response curves were built with the fast inhibition of I_Na that was activated from a holding potential of −110 mV chosen to maximize VGSC recovery from inactivation, therefore I_Na maximization (Figures 1B–C). Our data demonstrate an immediate inhibition of the total I_Na or TTX-R I_Na by EUG or LID upon inhibitor perfusion onset, with 100% recovery upon inhibitor washout. Note that the I_Na amplitude of the currents we studied is stable throughout the time series, and they are rundown-free (Figure 1B). Our data shown in Figure 3 also demonstrate a rundown-free recording both under control conditions and in the presence of EUG or LID as I_Na inhibitors: when at maximal Na⁺ conductance activation at test potentials of +20 mV or more, all I-V curves are linear. Our approach was intended to estimate the affinity of the fast inhibition of I_Na by EUG. Our experiments using LID as the positive control serve as a validation for our method.

Our data show that EUG inhibits the total I_Na in DRG neurons with an IC₅₀ value of 2.27 mM and a Hill coefficient of 1.95. EUG inhibits TTX-R I_Na with virtually the same properties, with an IC₅₀ value of 2.21 mM and a Hill coefficient of 1.84 (Figure 1C; Table 1). The affinity of the binding of EUG to the resting states of VGSCs (by using a holding potential of −110 mV and a test pulse to +20 mV) was statistically different from the inhibition induced by LID, with an IC₅₀ value of 1.42 mM, and TTX-R I_Na with an IC₅₀ value of 0.44 mM, which agrees with previous reports (Kistner et al., 2010). EUG affinity to inhibit VGSCs is far enough for local anesthesia. For comparison, a popular 2% LID formulation for human local anesthesia is a solution of LID at 85 mM. Interestingly, the Hill coefficient of the dose responses when EUG is used to inhibit I_Na, ∼2, is remarkably different from that when LID is used to inhibit the same currents (∼1). The Hill coefficient of dose–response curves is often related to binding sites, but the use of that knowledge is limited to orthosteric inhibitors such as pore blockers and not allosteric or state-dependent inhibitors (Bindslev, 2008; Prinz, 2010). Therefore, the only suggestion we can make, based on our data presented here and the Hill coefficients, is that EUG and LID may have different VGSC-inhibiting mechanisms. Additionally, in the case of a Hill coefficient of 2, for the inhibition of I_Na by EUG, we can rationalize that the coefficient does not arise from different EUG affinities to different VGSCs expressed in the neurons studied here. If that was the case, the slope of the dose–response curve would, instead, be shallower, thus showing an apparent Hill coefficient smaller than 1. Hence, we reinforce that since we deal with different VGSC isoforms here, with different kinetics and voltage dependencies, we simply speculate that the Hill coefficient we found for the I_Na inhibition by EUG, ∼2, indicates a different inhibiting mechanism compared to LID.

Our data shown in Figure 11 suggest that EUG and LID may interact when blocking I_Na. The 25% I_Na pre-inhibition with either drug induces a significant improvement on the inhibition by the other, which blocks the remaining 75% of the currents with a lower IC₅₀ value compared to the drug alone (Table 1). In addition, pre-blocking I_Na with LID reduces the Hill coefficient of the EUG inhibition curve from ∼2 with EUG alone to ∼1 in the presence of LID. We are extremely cautious when interpreting these data, especially because they come from experiments with a natively expressed variety of different VGSCs that could be inhibited by EUG or LID with different mechanisms. Therefore, we took a very conservative approach and limited the conclusion of these experiments to report that EUG and LID inhibit I_Na with likely different mechanisms. New experiments using a single VGSC subunit expressed heterologously are necessary to continue the investigation of this aspect of I_Na inhibition by EUG and LID simultaneously.

Our dose–response curves of inhibition of the total I_Na and TTX-R I_Na by EUG relate to the interaction of the inhibitor with the resting states of VGSCs. However, our detailed investigation revealed that EUG interacts with other states of VGSCs. Our data suggest that EUG may interact with VGSC pre-open–closed states possibly with a higher affinity than that of resting states (Figure 10). This notion is suggested by our experiments with a conditioning period at a membrane potential of −70 mV, which is known to populate pre-open–closed states of VGSCs (Armstrong, 2006), before I_Na is activated by a membrane potential step of +20 mV. Our data demonstrate that I_Na inhibition by EUG is intensified by such conditioning voltage steps, with a time constant of ∼70 ms, and we propose that this finding may be due to an interaction of EUG with the pre-open–closed states of VGSCs that are associated with a membrane potential of −70 mV. Specific experiments are still needed, but this piece of data alone suggests that such an interaction may take place with higher affinity than that with the resting states of the channels (when we activate the currents straight from a membrane potential of −110 mV). Our data using LID, an inhibitor known to bind to the inactivated states of VGSCs with higher affinity than to resting states, show the remarkable difference from that produced with EUG after time-varying −70-mV conditioning pulses. LID has its inhibiting effects on I_Na enhanced by the −70-mV conditioning pulse with a time constant of ∼700 ms. We propose that LID has its inhibiting potency increased in these experiments as more channels become inactivated, a slow process at −70-mV membrane potential (Armstrong, 2006). Differently, EUG would have its inhibiting potency increased faster (with a ∼70-ms time constant) as the pre-open–closed states, quickly achieved after stepping the membrane potential to −70 mV, are the states to which EUG would have an increased affinity. This notion will be a matter of further investigation in the near future using computation biology for kinetic models of this system.

We also propose that the right shift in the activation curves induced by EUG might be due to the interaction of EUG with pre-open–closed states of the VGSCs (Figure 5). Weak depolarizations would provide the channels with time to populate their pre-open–closed states, and that would be enough for EUG to further interact with the channels if the affinity to those states is higher than to resting states. LID does not show such a right shift in the voltage dependence of I_Na activation, in agreement with the notion that this drug would not interact with those pre-open–closed states. The remarkable effect of EUG on the steady-state inactivation curve of the total I_Na might also have been caused by the binding of EUG to pre-open–closed states. We propose the novel idea that a 100-ms conditioning pre-pulse, as shown in Figure 6A, is not a perfect measure of the binding of drugs to the fast inactivated states of VGSCs, even when we consider inactivation from pre-open–closed states. Other studies show that 1,000-ms or 10-s (10-fold or 100-fold longer than ours) conditioning pulses indeed show a remarkable shift in the V _1/2 -inact of the available currents, as induced by LID (Clarkson et al., 1988; Sheets et al., 2010). In particular, with a 10-s conditioning pulse, LID shifted the available curve by 10 mV only. For comparison, EUG shifted the inactivation curve by 16 mV with a 100-ms conditioning pulse. Our 100-ms conditioning pre-pulses may not last enough to inactivate the channels and for an interaction with LID after that. Once again, LID, known to interact with the inactivated states of VGSCs, does not shift the inactivation curves with the intensity or with the short conditioning pulse as EUG does (Figure 6). We suggest that this voltage clamp protocol with a 100-ms conditioning pre-pulse may also evidence the drug binding to the pre-open–closed states of the channels, and we think that EUG induces a consistent 16-mV shift in these curves because it might interact with the pre-open–closed states of the channels during the conditioning pre-pulses around −70 mV.

In our view, protocols to test the recovery from inactivation at a −110-mV membrane potential (Figure 7), high frequencies of depolarizing time series (Figure 8), and rate of slow inactivation entry (Figure 9) are better ways to study the interaction of inhibitors with the inactivated states of VGSCs (Clarkson et al., 1988). We sustain this notion because in all these protocols, the membrane potential is held at −110 mV (the holding potential) or at +20 mV when I_Na are activated and inactivated. Neither of these voltages is associated with the pre-open–closed states of VGSCs (Lenaeus et al., 2017). EUG underperformed in all three voltage clamp protocols compared with LID. These results show that EUG is not an inhibitor that binds to the inactivated states of VGSCs, fast and slow, with a higher affinity than binds to the resting states of VGSCs.

Our data show that EUG does not possess the ability to interact with pre-open–closed states of TTX-R VGSCs. Therefore, this EUG property must be related to the TTX-S VGSCs that are expressed in DRG neurons, i.e., Nav1.1, Nav1.6, and Nav1.7. Interestingly, since we observed the state-dependent effects of EUG on the total I_Na, i.e., on TTX-S plus TTX-R currents, we hypothesize that the observable effects on a TTX-S VGSC subunit would be even more clear.

EUG interacts with VGSCs, and this interaction modifies the voltage dependence of the channels. EUG may interact directly with the voltage sensors of the channels, as shown before for small molecules binding on extracellular residues of the homologous domain IV voltage sensor to cause a shift in the inactivation curve of the currents mediated by Nav1.3 (McCormack et al., 2013). EUD may also bind to the canonical local anesthetic binding site on S6 segments (the activation gate) that is allosterically connected to the voltage sensors of homologous domains III and IV (Sheets and Hanck, 2003; Sheets and Hanck, 2007). It is known that the activation of the voltage sensor of domain IV in VGSCs is necessary for the fast inactivation of these VGSCs (Kirsch et al., 1989; Campos et al., 2008). It seems that EUG does not interact with those regions since our data do not indicate stabilization of the inactivated states of VGSCs. Nevertheless, EUG may interact with pre-open–closed states of VGSCs as configured by the activation of voltage sensors of domains I and II or either one individually. Importantly, we cannot rule out that the voltage-dependent effect of EUG on the total I_Na reported here might be due to EUG interactions with an allosteric biding site elsewhere in the channel, including in the pore domain of the channels, a pathway we previously reported to be possible in Shaker K⁺ channels (Bassetto et al., 2021).

Another interesting and surprising piece of data in the present work is that EUG, but not LID, inhibits persistent total I_Na in DRG neurons (Figures 4A–D). It is important to mention that the inhibition of persistent I_Na could only be apparent if the inactivation kinetics of I_Na were accelerated. For example, we showed that neither EUG nor LID changed the kinetics of the inactivation process during a test pulse of the currents we studied (Figures 4E–H). Therefore, it is possible that EUG inhibits the specific channels that yield persistent currents more intensively than LID. This notion can lead to the inhibition of specific channels by EUG, and it will be investigated in our next studies. A previous study expressing the predominant cardiac VGSCs showed LID as an effective inhibitor of persistent voltage-activated Na⁺ currents (Dumaine and Kirsch, 1998). We propose that EUG might have a similar effect on the persistent I_Na observed in the present study.

All salts and drugs used in this work were purchased from Sigma (Sigma, St. Louis, MO). EUG was at least 98% pure, as stated by vendors. The pipette solution and the bath solutions were prepared directly from the salts and maintained at −20 °C until the day of the experiment, when they were thawed. A solution of eugenol (stock solution) was also prepared in advance to a concentration of 1 M in ethanol and stored at −20 °C. Just before the experiment, the stock solution was diluted in the bath solution to the desired final concentration of eugenol and sonicated for 5 min. The maximal final concentration of ethanol used in this study was 0.73% vol/vol, which, according to our own data (Figure 2) and data obtained by others (Wu and Kendig, 1998; Horishita and Harris, 2008), inhibits I_Na in approximately 10% and without any change in the voltage dependence of the activation and inactivation processes. This solution containing eugenol was applied over a single cell through a home-made pressurized perfusion system connected to a double-barreled capillary (Harvard Apparatus, Holliston, MA) to deliver the solutions with and without eugenol in the vicinity of the tested cell, with very fast interchangeable capability.

DRG neurons from 1–3-day-old rats were used as a model of natively expressed VGSCs. The cell preparation procedure was adapted from the study by Kostyuk et al. (1981). In brief, the rats were euthanized by decapitation, the vertebral canal was opened, and DRGs were quickly removed. All animals were handled in compliance with the Guide for the Care and Use of Laboratory Animals by the US National Institutes of Health (NIH Publication 85–23, revised 1996; http://www.nap.edu/.readingroom/books/labrats/index.html). In addition, all experimental protocols were approved by the Comissão de Ética no Uso de Animais (CEUA), a collegiate body linked to the Congregation of the Institute of Biomedical Sciences of the University of São Paulo (ICB-USP). Ganglia were digested with trypsin 0.25% in a Ca²⁺, Mg²⁺-free Earle´s balanced salt solution (EBSS) containing (mM) 132.8 NaCl, 5.3 KCl, 1 NaH₂PO₄, 5.5 glucose, and 10 HEPES, pH 7.4. After digestion, the ganglia were mechanically reduced using a fire-polished Pasteur pipette in a Ca²⁺, Mg²⁺-free EBSS containing 5 U/mL DNAse (type I; Sigma) and 0.15% trypsin inhibitor (type IS; Sigma), supplemented with 10% fetal calf serum. After pelleting, the cells were resuspended in Dulbecco's modified Eagle's medium (DMEM; Sigma) supplemented with 10% fetal calf serum, 100 UI/mL penicillin, and 100 mg/mL streptomycin and seeded on glass coverslips previously treated with poly-L-lysine. The cells were kept in a water-jacket incubator at 37 °C and a 5% CO₂ atmosphere until just before experiments, which were carried out during the first 7 days after cell isolation.

Voltage-activated Na⁺ currents passing through VGSCs were recorded using voltage clamping with the conventional whole-cell patch-clamp configuration (Hamill et al., 1981). Patch electrodes were fabricated from borosilicate glass capillaries using a model PB-7 micropipette puller (Narishige, Tokyo, Japan). Recording pipettes were pulled from borosilicate glass to achieve initial bath resistances averaging 2 MΩ and were filled with an intracellular solution containing (mM) 10 NaCl, 150 CsF, 10 TEA chloride, 1 ATP, 4.5 MgCl₂, 9 EGTA, and 10 HEPES, pH 7.3. CsCl and TEA chloride were used in order to eliminate K⁺ currents. Cells were bathed during the recordings in an extracellular solution containing (mM) 82 choline chloride, 50 NaCl, 1.2 MgCl₂, 1.8 CaCl₂, 1 CoCl₂, 4 KCl, 5 glucose, and 10 HEPES, pH 7.4. Na⁺ was partially replaced by choline (82 mM) to decrease the sodium driving force, avoid overload of the amplifier, and ensure a good voltage clamp. Co²⁺ in the extracellular solution was used to eliminate Ca²⁺ currents. After achieving a high-resistance seal, the whole-cell configuration was usually established by applying negative pressure to the pipette. Typical access resistance values were below 5 MΩ. Command voltage waveforms were generated in a computer using Clampex 10 software (Molecular Devices, Foster City, CA), and a DAC interface (model 1322; Molecular Devices, Foster City, CA) delivered the analog voltage signal to the Axopatch 200B patch-clamp amplifier (Molecular Devices, Sunnyvale, CA), which maintained the pipette voltage. The P/-4 protocol was used to remove voltage-independent leak currents and uncompensated capacitance (Bezanilla and Armstrong, 1977). The current recordings were low-pass filtered at 2 kHz by a built-in Bessel filter in the amplifier, sampled at 25 kHz in 16-bit digital levels by an ADC, and recorded on a hard disk for ulterior analysis. The cell membrane capacitance was canceled, and access resistance was routinely compensated (85% for both prediction and compensation; lag set to 10 ms). The holding potential in all experiments was set to −110 mV to avoid inactivation of the VGSC. All recordings were performed between 20°C and 23 °C. The recording chamber was continuously perfused with the bath solution to avoid unstirred layers. Recording cells exhibiting an I_Na peak of less than 500 pA throughout several depolarizing voltages were discarded. AP were recorded using the same cell preparation but in the current-clamp mode and using a pipette solution containing (mM) 10 NaCl, 150 KCl, 1 ATP, 4.5 MgCl₂, 9 EGTA, and 10 HEPES, pH 7.3, and a bath solution, including the solution where eugenol was dissolved, containing (mM) 132 NaCl, 1.2 MgCl₂, 1.8 CaCl₂, 4 KCl, 5 glucose, and 10 HEPES, pH 7.4.

Scientific data were processed, analyzed, and plotted using Clampfit (Molecular Devices, Foster City, CA), GraphPad Prism (GraphPad Software, LLC, La Jolla, CA), Origin (OriginLab, Northampton, MA), and Microsoft Excel (Microsoft, Redmond, WA). All the graphs represent mean values from at least six independent experiments, unless otherwise noted. The vertical bars in the graphs are the standard error of the mean (SEM).

The fitting functions mentioned in the Results section are as follows:

Dose–response curves were fitted with the Hill formalism: N o r m a l i z e d   u n i n h i b i t e d   c u r r e n t = E U G n H E U G n H + I C 50 n H , (1) where the Normalized (uninhibited) current is the remaining I_Na after inhibition by EUG or LID (the Fractional current in the graphs), IC ₅₀ is the concentration of eugenol that inhibits 50% of the I_Na, and nH is the Hill coefficient.

I_Na inactivation kinetics were fitted by the following exponentials: I N a = A 1 × 1 − e − t / τ 1 + A 2 × 1 − e − t / τ 2 + y 0 , (2) where I _Na is the Na⁺ current at a given moment t, A _1-2 are the weights of the respective exponentials, τ _1-2 are their time constants, and y ₀ is an adjusting factor for persistent currents. Inactivation time constants were calculated as follows: I n a c t i v a t i o n   t i m e   c o n s t a n t = A 1 × τ 1 + A 2 × τ 2 A 1 + A 2 . (3)

Current–voltage relationships were transformed into Na⁺ conductance activation by voltage (G-V) curves by using Ohm’s law: G m = I p e a k V m − V r , (4) where G _m is the equivalent conductance at Ipeak, which, in turn, is the current peak value at the given voltage, V _m is the membrane potential, and V _r is the reversal potential for Na⁺.

Na⁺ conductance activation by voltage (G-V) curves was fitted by the following equation: N o r m a l i z e d   N a + c o n d u c t a n c e = 1 1 + e V 1 / 2 − a c t − V m Max   s l o p e − a c t , (5) where Normalized Na ⁺ conductance is the fractional conductance activated at a given membrane potential V _m , V _1/2-act is the membrane potential for half-maximal Na⁺ conductance activation (the midpoint), and Max slope-act is the voltage sensitivity of the activation by voltage process.

I_Na inactivation by voltage curves (inactivation curves) was fitted by the following equation: A v a i l a b l e   N a +   c u r r e n t = 1 − 1 1 + e V 1 2 − i n a c t − V m p − p Max   s l o p e − i n a c t , (6) where Available Na ⁺ current is the Na⁺ current after the conditioning pre-pulse voltage period V _m p-p, V _1/2-inact is the V _m p-p that inactivates half of I_Na, and Max slope-inact is the voltage sensitivity of the inactivation by voltage process.

The kinetics of I_Na recovery from inactivated states were fitted with the following sum of exponentials: P 3 P 1 c u r r e n t   r a t i o = % f a s t 100 × 1 − e − P 2 d u r a t i o n τ f a s t + 1 − % f a s t 100 × 1 − e − P 2 d u r a t i o n τ s l o w , (7) where P ₃ /P ₁ current ratio is the fractional I_Na recovered from inactivation during P ₂ , %fast is the % fraction of the fast recovery from inactivation component amplitude, τ ₁ is the faster time constant of the I_Na recovery from inactivation, and τ ₂ is the slower time constant of the I_Na recovery from inactivation. P 3 P 1 c u r r e n t   r a t i o = P l a t e a u + Y P 1 = 2 m s − P l a t e a u × % f a s t 100 × e − P 1 d u r a t i o n / τ f a s t + Y P 1 = 2 m s − P l a t e a u × 1 − % f a s t 100 × e − P 1 d u r a t i o n / τ s l o w , (8)

where P ₃ /P ₁ current ratio is the fractional I_Na inactivated during P ₁ and not recovered from inactivated during P ₂ (see text for details), Plateau is the current level at P₁>>10000 ms, Y _P1=2ms is the P ₃ /P ₁ current ratio when P₁ = 2 ms, %fast is the % fraction of the fast inactivation component, τ ₁ is the faster time constant of the slow inactivation component, and τ ₂ is the slower time constant of the slow inactivation component.

Data from individual cells were treated individually, including for fitting purposes. Pooled fitting parameters from different groups, e.g., EUG vs. control (its absence), were compared using a paired t-test to detect consistent changes in the parameters that relate to the drugs. Levels of significance were *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.