The analgesic paracetamol metabolite AM404 acts peripherally to directly inhibit sodium channels

Abstract

Paracetamol has been used for decades to relieve mild-to-moderate pain. Its analgesic effect is mainly attributed to its metabolite, AM404, acting on cannabinoid receptors or TRPV1 channels in central nervous system (CNS) neurons. Here, we show that AM404 is produced by primary sensory neurons. It inhibits sodium current in nociceptor neurons, blocking action potential (AP) generation and reducing nocifensive behavior in naïve and inflamed rats. We demonstrated that this analgesic effect of AM404 is mediated by its direct inhibition of nociceptive voltage-gated sodium channels (Na_V) 1.8 and 1.7 via the local anesthetic binding site. The Na_V1.8 and 1.7 inhibition was specific for AM404 and not observed with other metabolites of paracetamol. Our findings suggest that the analgesic effect of paracetamol is mediated mainly by direct AM404-induced inhibition of nociceptive sodium channels at the peripheral nociceptor neurons. Our findings lay a foundation for the potential development of AM404 as a selective local analgesic.

Keywords: analgesics; local anesthetic; nociception; pain; sodium channels.

Fig. 1.

AM404 is produced in trigeminal ganglion neurons, but despite inducing a prominent inward current, it does not evoke action potential (AP) firing. (A) In situ formation of AM404 in rat trigeminal ganglion (TG) from its precursors 4-aminophenol. The Brown–Forsythe and Welch ANOVA test with the Dunnett T3 post hoc test was performed. Each data point represents AM404 concentration in a single measurement from a homogenized batch of 3 and 4 TGs. (B) Left: Representative whole-cell gap-free voltage-clamp recordings from HEK293T cells transiently expressing rat TRPV1 at a holding potential of −60 mV before and after application of AM404 (1 µM; pale red) followed by applications of capsaicin (1 µM; pale blue). Right: The peak of capsaicin-evoked current normalized to the peak of AM404-evoked current. (C) Same as B, but recorded from TG neurons. Note that the ratio is similar in capsaicin-sensitive TG neurons and the rat TRPV1 expression system. (D) Left: Representative whole-cell gap-free current-clamp recordings from TG neurons demonstrating the voltage changes following AM404 (1 µM; red trace) and capsaicin (0.2 µM; black trace). Note that the application of AM404 led to substantially lower membrane depolarization and no AP firing. Note different time scales for capsaicin and AM404 trace. Right: The graph depicting the individual values of the number of APs evoked by the application of capsaicin (gray) and AM404 (red). Note that the number of APs evoked by the application of capsaicin is significantly larger than that by AM404. One-way ANOVA, followed by Bonferroni’s post hoc test. In all experiments presented in B–D, the number of dots represents the number of cells; recordings were taken from one cell on each coverslip.

Fig. 2.

AM404 inhibits nociceptive firing and pain-related behavior. (A) Representative whole-cell current-clamp recording from acutely dissociated rat nociceptive TG neurons in response to a current ramp (300 pA in 0.5 s; Inset) before (Left), during exposure to 1 µM AM404 (Middle), and after washout of AM404 (Right). Upper panel, the effect of treatment of DRG neurons with 0.01 µM AM404; Lower panel, the effect of treatment of DRG neurons with 1 µM AM404. (B) Concentration–response relationship for inhibition of the AP firing by AM404 in nociceptor TG neurons presented as a fractional block (fractional block of “1” depicts a complete inhibition). Box plots and individual values demonstrate changes in the number of APs following 10 min of exposure to AM404 at the indicated concentrations. The number of APs was normalized to the number of evoked APs before applying AM404. The IC₅₀ was determined using Hill’s equation fit (black curve). One-way ANOVA, followed by Bonferroni’s post hoc test. Statistical comparisons are between the “Before” group and the AM404 concentrations (black) and between different AM404 concentrations (gray). (C) Box plot and individual values of changes in the number of APs with time during exposure to AM404 (1 µM). The number of APs was normalized to the number of APs before applying AM404. One-way ANOVA, followed by Bonferroni’s post hoc test. Statistical comparisons are between the “Before” group and the AM404 concentrations (black), between different AM404 concentrations (gray), and between the “2 min after AM404” group with other time points (pink). (D) Mean ± SEM of the hind limb withdrawal latencies to noxious mechanical (von Frey; Left) and noxious thermal stimuli (Hargreaves; Right) after intraplantar application of AM404 (10 mM, red) or vehicle (white) in female (Left, N = 6 rats in each group) and male (Right, N = 6 rats in each group) rats. Two-way ANOVA followed by Bonferroni’s post hoc test. Note that local application of AM404 is sufficient to inhibit the responses of the rats to noxious stimuli. Please refer to SI Appendix, Fig. S5 for non-normalized data from ipsilateral and contralateral limbs.

Fig. 3.

AM404 inhibits nociceptive sodium currents. (A) Representative whole-cell voltage-clamp recording of total sodium currents from acutely dissociated rat nociceptive TG neurons before (Left) and after exposure to AM404 (1 µM; Right). Currents were elicited by depolarizing steps from a holding potential of −80 mV to 10 mV in 10 mV increments (Inset). (B) Box plot and individual values of the changes in the peak amplitude of total sodium current (normalized to the current before the application of AM404) with time during exposure to AM404 (1 µM). One-way ANOVA, followed by Bonferroni’s post hoc test. (C) Box plot and individual values of changes in the peak total sodium current amplitude (normalized to the current before the application of AM404) following 10 min of exposure to AM404 at the indicated concentrations. One-way ANOVA, followed by Bonferroni’s post hoc test. (D) Same as A but depicting the effect of AM404 on TTX-r sodium current, elicited by depolarizing steps from a holding potential of −80 mV to 10 mV in 10 mV increments (Inset). (E) Same as B, but examining the time-dependent inhibition of AM404 on TTX-r current. One-way ANOVA, followed by Bonferroni’s post hoc test. (F) Same as C but examining the effect of indicated concentrations of AM404 on TTX-r current. One-way ANOVA, followed by Bonferroni’s post hoc test. In all experiments, the number of dots represents the number of cells; one cell from the coverslip was recorded. The box plots depict the mean and 25 to 75 percentiles, and the whiskers depict 1.5 SD. (G) Same as A but depicting the effect of AM404 on sodium current recorded from acutely dissociated rat non-nociceptive, large (≥35 µm) TG neurons before (Left) and during exposure to AM404 (1 µM; Right). Currents were elicited by depolarizing steps from a holding potential of −80 mV to 10 mV in 10 mV increments (Inset). (H) Box plot and individual values of the changes in the peak sodium current amplitude in non-nociceptor neurons with time (normalized to the current before the application of AM404) during exposure to AM404 (1 µM; blue) and compared to the AM404 effect on total sodium current in nociceptor neurons (shown in B, pale pink). One-way ANOVA, followed by Bonferroni’s post hoc test, was used to determine the effect of AM404 exposure time on the peak sodium current in non-nociceptor neurons (black asterisk and numbers). Two-way ANOVA was used to compare the effect of AM404 between nociceptive and non-nociceptor neurons (gray asterisks and numbers; ****P ≤ 0.0001). (I) Box plot and individual values of the changes in the peak sodium current amplitude in non-nociceptor neurons (normalized to the current before the application of AM404) following 10 min of exposure to AM404 at the indicated concentrations (blue) and compared to the AM404 effect on total sodium current in nociceptor neurons (shown in C, pale pink). One-way ANOVA, followed by Bonferroni’s post hoc test, was used to define the effect of the peak sodium current in non-nociceptor neurons (black asterisk and numbers). Two-way ANOVA was used to compare the effect of AM404 between nociceptive and non-nociceptor neurons (gray asterisks and numbers; ****P ≤ 0.0001). In all experiments, the number of dots represents the number of cells; one cell from the coverslip was recorded. The box plots depict the mean and 25 to 75 percentiles, and the whiskers depict 1.5 SD.

Fig. 4.

AM404 inhibits the nociceptive human isoforms of Na_V1.8 and Na_V1.7 channels. (A) Representative whole-cell current-clamp recording from ND7/23 cells overexpressing human Na_V1.8 (hNa_V1.8) channels before (Left) and after exposure to AM404 (1 µM; Right). Currents were elicited by depolarizing steps from a holding potential of −80 mV to 30 mV in 10 mV increments (Inset). (B) Same as A but demonstrating the effect of AM404 on hNa_V1.7 expressed in HEK293T cells and elicited by depolarizing steps from a holding potential of −80 mV to 10 mV in 10 mV increments (Inset). (C) Concentration–response relationship for inhibition of hNaV1.8 (red) and hNaV1.7 (orange) channels by AM404 presented as a fractional block (fractional block of “1” depicts a complete inhibition). Each dot represents the mean and SD of at least n = 6 cells. IC₅₀ were determined by fitting Hill’s function to the data (shown as red and orange curves for Na_V1.8 and Na_V1.7, respectively).

Fig. 5.

AM404 inhibits hNa_V1.8 channels via the local anesthetics binding site. (A) Left: Representative whole-cell voltage-clamp recording from ND7/23 cells expressing hNa_V1.8 comparing the response to depolarizing steps (pulse 1 vs. 60) from −80 mV to +10 mV applied at 20 Hz, before and after application of 1 µM AM404. Because the application of AM404 inhibits sodium currents, the currents obtained after treatment with AM404 were scaled to the currents before the treatment for better comparison between the differences in the responses to pulses. Note that the difference between pulse No. 1 and 60 is more prominent in the presence of AM404, suggesting use-dependent blockade. Right, same as Left, but shows the effect of bupivacaine (30 µM; blue). (B) The peak hNa_V1.8 current, normalized to the peak current of the first pulse and plotted vs. the activation pulse number, shows a use-dependent block of AM404 (1 μM; red) or bupivacaine (30 μM; blue) on hNa_V1.8 and compared to the control condition in which no drug was applied. Repetitive 10 ms −80 mV to 10 mV pulses at 20 Hz were used to evoke sodium currents (Inset). Each dot represents the mean ± SD of the normalized peak currents from at least n = 6 cells. The averaged data were fitted with a single exponential decay function, and the decay time of the function (τ) was determined. Please refer to SI Appendix, Fig. S9A for the use-dependent block assessed with the 5 Hz stimulation protocol. (C) Box plot and individual values of the peak current evoked by the 60th pulse in control conditions and after treatment with AM404 and bupivacaine. One-way ANOVA, followed by Bonferroni’s post hoc test. Note the prominent use-dependent block by AM404, which is similar to bupivacaine. Please refer to SI Appendix, Fig. S9B for the 5 Hz stimulation protocol. (D) Left: Representative whole-cell voltage-clamp recording from ND7/23 cells overexpressing WT and mutated hNa_V1.8 (F1759A) before (black) and after application of AM404 (1 µM; red). Right, same as Left, but shows the effect of bupivacaine (30 µM; blue). Currents were elicited by a depolarizing step from a holding potential of −80 mV to 10 mV. (E) Concentration–response curves of the inhibitory effect of AM404 (Left) and bupivacaine (Right) in WT and mutated hNa_V1.8 (F1759A) channels presented as a fractional block. Each dot represents the mean ± SD of the normalized peak currents recorded from at least n = 6 cells. Note the lack of the inhibitory effect of AM404 on the mutated hNa_V1.8 (F1759A) channel. (F) Same as A but recorded from mutated hNa_V1.8 (F1759A) channels. Because in these conditions, neither AN404 nor bupivacaine affects the amplitude of hNa_V1.8 currents, the responses were not scaled. Note that in these conditions, the application of AM404 does not change the difference between pulse No. 1 and 60. Right, same as Left, but shows the effect of bupivacaine (30 µM; blue). (G) Same as B but recorded from mutated hNa_V1.8 (F1759A) channels. Each dot represents the mean ± SD of the normalized peak currents recorded from at least n = 6 cells. The averaged data were fitted with a single exponential decay function, and the decay time of the function (τ) was determined. Please refer to SI Appendix, Fig. S9E for the use-dependent block assessed with the 5 Hz stimulation protocol. (H) Same as C but for the mutated hNa_V1.8(F1759A) channel. One-way ANOVA, followed by Bonferroni’s post hoc test. Note that F1759A point mutation annuls AM404-induced use-dependent block. The box plots depict the mean and 25 to 75 percentiles, and the whiskers depict 1.5 SD. Please refer to SI Appendix, Fig. S10A for the 5 Hz stimulation protocol.

Fig. 6.

Paracetamol and its derivatives NAPQI and 4-aminophenol do not inhibit hNa_V1.7 current. (A) Left: Representative whole-cell voltage-clamp recording from HEK293T cells overexpressing hNa_V1.7 channels before and after exposure to paracetamol. Currents were elicited by depolarizing steps from a holding potential of −80 to −10 mV in 10 mV increments. Right: Concentration–response curves of the effect of paracetamol on hNa_V1.7 current presented as a fractional block. Each dot represents the mean ± SD recorded from at least n = 6 cells. Note the lack of the effect of paracetamol on hNa_V1.7 current. Please refer to SI Appendix, Fig. S12 for the hNa_V1.8 response to paracetamol. (B) Same as A but for NAPQI (Upper panel) and 4-aminophenol (Lower panel). Each dot represents the mean ± SD recorded from at least n = 6 cells. Note that NAPQI does not inhibit hNa_V1.7 current, and 4-aminophenol inhibits hNa_V1.7 current but in the mM concentration range. Please refer to SI Appendix, Fig. S7 B and C for the hNa_V1.8 response to NAPQI and 4-aminophenol.

Fig. 7.

AM404 reduces formalin-induced inflammatory hyperalgesia. (A) Box plot and individual values of hind limb withdrawal latencies to noxious mechanical stimuli assessed before and 60 min after injection of formalin in animals injected intraplantary with vehicle (Left) or AM404 (10 mM; Right) in female (Left, N = 9 rats in each group) and male (Right, N = 9 rats in each group) rats. Note that treatment with AM404 diminished the formalin-induced decrease in the mechanical threshold. One-way ANOVA followed by Bonferroni’s post hoc test. Please refer to SI Appendix, Fig. S13A for non-normalized data. (B) Treatment with AM404 attenuates the CFA-induced decrease in mechanical threshold. Two days after the intraplantar injections of CFA, animals were injected with AM404 or vehicle. Note a significant decrease in CFA-induced mechanical hypersensitivity following AM404 injection that started 15 min after injection and lasted for 1 h. Black letters and asterisk, comparison of AM404 and vehicle group, two-way ANOVA followed by Bonferroni’s post hoc test (N = 6, 3 male and 3 female rats in each group). Red letters and asterisk, comparison of the effect of AM404 treatment on the sensitivity to the mechanical stimuli before the injection of AM404, one-way ANOVA followed by Bonferroni’s post hoc test (N = 6, 3 male and 3 female rats in each group). Please refer to SI Appendix, Fig. S13B for non-normalized data.