Introduction

Insulin plays a crucial role in regulating energy metabolism in nonneuronal cells by binding insulin receptors. Insulin may also modulate distinct neural functions through insulin receptors, which are widely expressed in the central nervous system1,2. For example, insulin can facilitate critical brain functions including learning, cognition and motivated behaviour3,4,5. Several lines of emerging evidence indicate that insulin may function as an important neuromodulator by regulating ion channels. For instance, insulin modulates GABAA receptor-mediated currents in neurons from the hippocampus, amygdala, cerebral cortex and prefrontal cortex6,7,8,9. Insulin receptors are predominantly located in the spinal cord and primary sensory neurons, which suggests their involvement in sensory pathways10,11,12. In spinal neurons, insulin modulates glycine receptor-mediated currents and AMPA excitatory postsynaptic currents through insulin receptors13,14. In primary sensory neurons including those in the dorsal root ganglion (DRG) and trigeminal ganglion, the activation of insulin receptors by insulin sensitizes TRPV1 receptors15,16. In addition, insulin potentiates the response to mechanical stimuli in small DRG neurons17. In addition to the effects on TRPV1 receptors, little is known about the effects of insulin on other ion channels, although they are also expressed in primary sensory neurons.

Acid-sensing ion channels (ASICs) are ion channel members that detect changes in extracellular pH18. Four genes encode at least six ASIC subunits (ASIC1a, -1b, -2a, -2b, -3, and -4), which form homotrimeric or heterotrimeric channel complexes19. Among all ASIC subunit configurations, those containing ASIC3 are the most sensitive to acid20. ASICs are widely expressed in peripheral sensory neurons including DRG cell bodies and sensory terminals21,22. In particular, ASIC3 is the most abundant in sensory neurons and has emerged as a critical pH sensor23. Tissue acidosis is a hallmark of several painful diseases, such as inflammation, and tissue injury24. Direct infusion of pH 6.0 acidic solution into human skin caused localized pain, which is significantly alleviated by the nonselective ASIC inhibitor amiloride, suggesting that pain sensation is mainly mediated by ASICs25. ASIC3 expression is increased in DRG neurons after plantar incision and blockade of ASIC3 ameliorates the incision-induced pain26. Multiple pain model mice lacking ASIC3 exhibit reduced nociceptive behaviors27. However, other studies have shown that mice lacking ASIC3 exhibit increased mechanosensitivity28. These conflicting findings may be due to differential dysregulation of alternative ASIC subunits in different mouse genotypes29. The latter may be attributed to changes in the activity of ASIC1a and 2a, which also play a role in pain28,30. Tissue acidification occurs in a variety of pain conditions. ASICs, especially ASIC3, are major players in pain associated with tissue acidification23,31,32,33.

Since both insulin receptors and ASICs have been shown to be present in DRG neurons and participate in sensory processes, the purpose of this investigation was to determine the effects of insulin on the functional activity of ASICs. Herein, we report that insulin potentiated the acid-evoked ASIC currents and action potentials in rat DRG neurons. This potentiation was dependent on insulin receptors and intracellular signaling pathways involving tyrosine kinase and phosphatidylinositol-3 kinase (PI3 kinase). Furthermore, we investigated whether findings from ex vivo experiments could be translated to in vivo conditions. Local administration of insulin exacerbated the spontaneous nociceptive behaviors induced by intraplantar acid injection and the mechanical hyperalgesia induced by intramuscular acid injections through peripheral insulin receptors.

Materials and methods

Preparation of DRG neurons

We purchased female Sprague-Dawley rats (5–6 weeks old) from Hubei Biont Biological Technology Co. Ltd. (Wuhan, China). All experiments were approved by the Animal Research Ethics Committee of Hubei University of Science and Technology (2016-03-005). All methods were carried out in accordance with the American Veterinary Medical Association (AVMA) Guidelines for the Euthanasia of Animals (2020) and the National Institute of Health’s Guide for the Care and Use of Laboratory Animals. All animal experiments were designed in accordance with the ARRIVE guidelines (https://arriveguidelines.org). One female rat was anesthetized with a mixture of ketamine (90 mg/kg) and xylazine (10 mg/kg) and sacrificed by decapitation. The DRGs of lumbar segments 4–6 were removed and chopped. The minced ganglia were transferred to a test tube containing Dulbecco’s modified Eagle’s medium (DMEM) and incubated with shaking for 25–30 min at 35 °C. The incubation solution contained DNase (type IV, 0.1 mg/mL), trypsin (type II-S, 0.5 mg/mL), and collagenase (type I-A, 1.0 mg/mL). The digestion was terminated by the addition of soybean trypsin inhibitor 1.25 mg/ml. The cells were cultured in a 35 mm culture dish for 12–24 h at 37 °C in DMEM containing nerve growth factor (100 ng/ml) and fetal bovine serum (10%).

Electrophysiological recordings

A MultiClamp-700B amplifier and Digidata-1440A A/D converter (Axon Instruments, CA, USA) were used for whole-cell patch clamp and voltage-clamp recordings. Before performing the electrophysiological experiments, we replaced the DMEM with an external solution in the DRG cell culture dish, and then kept the cultured DRG cells in the external solution for at least 60 min. The external solution contained the following (in mM): 150 NaCl, 5 KCl, 2 MgCl2, 2.5 CaCl2, 10 HEPES, and 10 d-glucose. The pH and osmolarity were adjusted to 7.4 with NaOH and 330 mOsm/L with sucrose, respectively. The recording pipettes were pulled using a Sutter P-97 puller (Sutter Instruments, CA, USA), whose resistance was in the range of 3–6 MΩ. The micropipette solution contained the following (in mM): 140 KCl, 2 MgCl2, 11 EGTA, 10 HEPES, 4 ATP, and 0.3 Na2GTP. The pH and osmolarity were adjusted to 7.2 with KOH and 310 mOsm/L with sucrose, respectively. After the whole-cell configuration was established, 70–80% of the series resistance and membrane capacitance current were compensated. The recorded currents were sampled at 10 kHz and filtered at 2 kHz. Electrophysiological measurements were only performed on small- and medium-sized nociceptive DRG cells (with a diameter of 15–40 μm), which are thought to be nociceptive neurons34. The membrane potential of the neurons was clamped at − 60 mV. Only DRG neurons with a resting membrane potential more negative than − 50 mV were used for current-clamp recordings.

Drug application

For the experiments, drugs were purchased from Sigma Chemical Co. (St. Louis, MO, USA). Acidic solutions were prepared daily with HCl and an external solution that contained 10 mM HEPES (pH ≥ 6.0) or 10 mM 4-morpholineethanesulfonic acid (MES, pH ≤ 5.5). All working drugs were adjusted to a pH of 7.4 with NaOH and an external solution containing 10 mM HEPES. Each working drug was stored in a series of independent reservoirs and applied by gravity. The distance between the drug outlet and the recording neuron was ∼ 30 μm. To ensure that the recorded current was ASIC, AMG9810 (5 μM) was added to the extracellular solution to block acid-induced TRPV1 activation. To block PI3 kinase, the PI3 kinase inhibitor wortmannin was dissolved in the internal solution and applied for intracellular dialysis through recording patch pipettes.

Nociceptive behavior

Considering the gender differences in pain caused by acidosis in our previous reports35, only female rats were used for the present study. Rats were first habituated for 30 min in a Plexiglas chamber during the nociceptive behavioral experiment. The rats received two intraplantar injections, each with a volume of 50 μl. Rats in five different groups (n = 10/group) were injected with 50 μl of 10 μM AMG 9810+ vehicle, 50 μl of cocktail containing10 μM AMG 9810 + different doses of (5, 50 and 500 ng) insulin, or 50 μl of cocktail containing AMG 9810 + 100 ng of GSK1838705 + 500 ng of insulin. After 10 min, another experimenter injected 50 μl of acetic acid solution (1% v/v, the pH was adjusted to 6.0 with NaOH and external solution) into the ipsilateral hind paws and tested nociceptive behavior. The assessor of the behavioral measures was blinded to the prior treatment conditions. Nociceptive behavior (i.e., flinching, shaking and licking) was expressed as the number of flinches within 5 min after injection23,36. In the acid-induced muscle pain model, rats were injected twice with 100 ul of pH 4.0 sterile saline (0.059 M citric acid, 0.04 M trisodium citrate, pH value was adjusted to 4.0 with NaOH and HCl) into one gastrocnemius muscle 5 days apart at the same site33. The mechanical sensitivity of the ipsilateral hind plantar was determined by applying a series of calibrated von Frey filaments to the plantar aspect of the hind paw using the ‘up–down’ method to determine the paw withdrawal threshold (PWT)37.

Data analysis

All the data are presented as the mean ± S.E.M. The normality of the data distribution and homogeneity of variance were initially investigated, and the result was that all distributions were normal. One-way and repeated-measures analysis of variance (ANOVA) followed by Tukey's post hoc test for multiple comparisons were used to determine significant differences when comparing multiple groups. Two-way repeated-measures ANOVA followed by Bonferroni post hoc correction was used to analyze bivariate parameters. Comparisons between two groups were carried out using unpaired t-test. P values < 0.05 were considered to indicate statistical significance. The nonlinear curve-fitting program ALLFIT was used for statistical analysis of the concentration–response data.

Ethical approval

The experimental protocol was approved by the Animal Research Ethics Committee of Hubei University of Science and Technology (2016-03-005).

Results

Insulin enhances ASIC currents in rat DRG neurons

In the presence of AMG9810 (5 μM), capsaicin (100 nM) failed to evoke any membrane currents in any of the recorded DRG neurons. In contrast, a 5 s exposure of pH 6.0 acidic solution produced a rapid inward current (IpH6.0) in the majority of DRG neurons (84.2%, 16/19; Fig. 1A). The IpH6.0 was blocked by the broad-spectrum ASIC channel blocker amiloride (100 μM), and by the ASIC3 blocker APETx2 (2 μM), indicating that the acid-evoked current was an ASIC current or an ASIC3-like current. We observed that insulin enhanced the peak amplitude of the IpH6.0 to 150.7 ± 6.1% of control (p < 0.01, one-way ANOVA followed by Tukey’s post hoc test, n = 10; Fig. 1B,C) when the DRG cells were pretreated with insulin (300 nM) for 5 min prior to the next current recording. A cutoff value was considered indicative of a neuron responsive to insulin only when increase in the IpH6.0 exceeded 7%. The present results showed that 62.5% (10/16) of neurons sensitive to pH 6.0 acidic stimuli also responded to insulin. After a 6 min washout of insulin, the IpH6.0 amplitude was similar to that of the control (104.6 ± 2.3% of the control, n = 10; Fig. 1B,C), suggesting rapid and reversible enhancement. In contrast, heat-inactivated insulin (300 nM) failed to enhance IpH6.0 (98.1 ± 1.9% of the control, n = 10; Fig. 1B,C).

Figure 1
figure 1

Enhancement of ASIC currents by insulin in rat DRG neurons. (A) In the presence of 5 μM AMG9810, capsaicin (Cap, 100 nM) induced no membrane currents. However, a 5 s exposure of pH 6.0 acidic solution produced a rapid inward current (IpH6.0) in the same DRG neuron. The IpH6.0 was blocked by amiloride (Amil, 100 μM) and APETx2 (2 μM). (B) The IpH6.0 was enhanced by insulin pretreatment (300 nM for 5 min) and recovered after the washout of insulin. In contrast, the IpH6.0 was not affected by heat-inactivated insulin (300 nM). Insulin (300 nM) or heat-inactivated insulin (300 nM) was pre-applied to a recorded DRG cell for 5 min. (C) The bar graph shows that the amplitude of the IpH6.0 under different conditions. **p < 0.01, n.s., not significant, compared with the control. The data were analyzed by one-way ANOVA followed by Tukey’s post hoc test. n = 10 cells from 6 rats. (D) The enhancing effect of insulin (300 nM) on the IpH6.0 increased with increasing duration of insulin preapplication from 0 to 5 min. Each point represents the mean ± S.E.M. of 7–10 cells from 5–6 rats. (E) The graph shows the concentration-effect curve of insulin on IpH6.0 with an EC50 value of 142.6 ± 11.7 nM. Each point represents the mean ± S.E.M. of 8–10 cells from 5–6 rats.

Full size image

To explore the relationship between the effect of insulin on IpH6.0 and the duration of the pre-application of insulin, different durations of insulin preapplication were tested. As shown in Fig. 1D, the amplitude of IpH6.0 increased as the duration of insulin (300 nM) pre-application increased from 0 to 5 min. Although the preapplication of insulin for 1 min had no significant effect on the IpH6.0, a 2 min or longer duration of insulin administration increased the IpH6.0. The enhancing effect of insulin approached its maximum at 4 min of pretreatment, with a mild further effect for longer durations.

We first investigated whether the enhancement of IpH6.0 by insulin was dependent on its concentration. The concentration-effect curve of insulin on the IpH6.0 in Fig. 1E indicates that the threshold concentration of insulin that increased the IpH6.0 was 30 nM, the EC50 (half-maximal effective concentration) value was 142.6 ± 11.7 nM, and the Hill coefficient was 1.1 ± 0.1. The results indicated that insulin produced a concentration-dependent increase in the amplitude of the IpH6.0.

We then investigated whether the enhancement of ASIC currents by insulin was dependent on acidic pH. The current traces in Fig. 2A show that insulin pretreatment (300 nM for 5 min) had an enhancing effect on all three ASIC currents, which were evoked by different acidic solutions (pH 6.5, pH 5.5, and pH 4.5). Figure 2B shows that the concentration–response curves for acid were constructed for various acidic solutions in the absence and presence of 300 nM insulin. Insulin shifted the concentration–response plot for ASIC currents an increase of 46.2 ± 7.6% in the normalized maximal current response (IpH4.5) of the curve, which was induced by an acidic solution at pH 4.5 (p < 0.01, two-way repeated measures ANOVA followed by Bonferroni post hoc correction, n = 9). However, insulin had no effect on the pH0.5 (pH for half-maximal activation) values, and there was no significant difference in the pH0.5 values between the two curves (pH: pH0.5 = 5.7 ± 0.2; insulin + pH: pH0.5 = 5.9 ± 0.3). The slope or Hill coefficient was also unchanged (pH: n = 1.4 ± 0.1; insulin + pH: n = 1.5 ± 0.2). These results indicated that insulin increased the maximum response of ASICs to acid but did not change the sensitivity to acid.

Figure 2
figure 2

Effects of insulin on the concentration–response curve for acidic stimuli. (A) Representative current traces showing that insulin pretreatment (300 nM for 5 min) enhanced the three ASICs currents stimulated by acidic solutions at pH 6.5, pH 5.5 and pH 4.5. (B) The graph shows that the concentration‒response curve for acid was shifted upward in the presence of 300 nM insulin, showing a significant increase in the normalized maximum response to acid. The curves were drawn according to the logistic equation I = Imax/[1 + (10^pH0.5/10^pH)n], where I is the normalized current response value, pH0.5 is the pH for half-maximal activation, and n is the Hill coefficient. All current values from the same cell were normalized to the current response, which was induced by pH 4.5 applied alone (marked with asterisk). Each point represents the mean ± S.E.M. of 8–10 cells from 5–6 rats.

Full size image

Insulin-mediated enhancement of ASIC currents involves insulin receptors, intracellular tyrosine kinases and PI3 kinases

We further investigated whether activation of the insulin receptors was required for the insulin-mediated enhancement of ASIC currents. The insulin receptor antagonist GSK1838705 was used to treat DRG neurons with insulin. The neurons were pretreated for 1 min with GSK1838705, and then GSK1838705 and insulin were coadministered for 5 min. Figure 3A presents a sample recording showing that pretreatment with 20 nM GSK1838705 prevented the insulin-induced potentiation of ASIC currents. The group mean data demonstrated that application of insulin (300 nM for 5 min) alone increased the IpH6.0 amplitude (150.7 ± 6.1% of the control, n = 10), whereas the enhancement of IpH6.0 by insulin did not occur in the GSK1838705 treated cells (106.6 ± 2.3% of the control, n = 10; Fig. 3B). In contrast, Fig. 3C,D show that the enhancement of IpH6.0 by insulin still occurred in cells treated with vehicle (0.1%DMSO, v/v) instead of GSK1838705 (152.8 ± 4.9% of the control, n = 7). In addition, GSK1838705 alone did not alter the IpH6.0 amplitude in separate experiments (98.6 ± 1.1% of the control, n = 5). These results showed that the enhancement of ASIC currents required the activation of insulin receptors.

Figure 3
figure 3

Participation of insulin receptors in the insulin-mediated enhancement of ASIC currents. Representative current traces in (A and C) showing the recorded IpH6.0 in a DRG neuron under control, treatment with insulin (300 nM for 5 min) alone, co-treatment with insulin (300 nM for 5 min) and the insulin receptor antagonist GSK1838705 (20 nM for 6 min, A) or its vehicle (C) conditions. GSK1838705 or its vehicle was pre-applied to the recorded neuron for 1 min, and followed by coapplication with insulin for 5 min. The bar graph in (B and D) shows that the amplitude of the peak IpH6.0 was enhanced by the application of insulin (300 nM for 5 min) alone, and this effect was blocked by co-treatment with GSK1838705 (20 nM for 6 min), but not by co-treatment with vehicle (0.1%DMSO, v/v). ***p < 0.001, paired t-test, n = 10 cells from 6 rats; n.s., not significant, paired t-test, n = 7 cells from 4 rats.

Full size image

We next sought to identify the intracellular signaling downstream of the insulin receptor that might be involved in the insulin-ASIC interaction. The insulin receptor is a member of the superfamily of receptor tyrosine kinases. Activation of the insulin receptor by binding insulin results in intracellular signaling cascades that involve tyrosine kinase and PI3 kinase. First, we pre-incubated DRG cells with the tyrosine kinase inhibitor lavendustin A or its inactive analog lavendustin B to determine whether tyrosine kinases contribute to the enhancement of ASIC currents by insulin. In the presence of either lavendustin A (5 μM) or lavendustin B (5 μM), insulin (300 nM for 5 min) was subsequently applied, and the change in the IpH6.0 amplitude was measured. Compared with insulin, which increased the IpH6.0 amplitude to 150.7 ± 6.1% of the control, lavendustin A completely abolished the insulin-induced increase in the IpH6.0 amplitude (104.5 ± 1.8% of the control, p < 0.001, one-way ANOVA followed by Tukey's post hoc test, n = 10; Fig. 4A,B). In contrast, the inactive analog lavendustin B had no effect on insulin enhancement of ASIC currents as the IpH6.0 amplitude increased to 149.7 ± 5.9% of the control (p > 0.1, one-way ANOVA followed by Tukey's post hoc test, n = 10; Fig. 4A,B). These results indicate that tyrosine kinases were involved in the intracellular signaling pathways that contributed to the insulin-induced enhancement of ASIC currents in DRG neurons. We then added the PI3 kinase inhibitor wortmannin (50 nM) to the pipette solution and applied it internally to DRG cells to determine whether PI3 kinase was involved in the insulin-induced potentiation of ASIC currents. In these wortmannin treated DRG cells, pre-application of insulin (300 nM for 5 min) failed to increase the IpH6.0 amplitude (p > 0.1, unpaired t-test, compared with the normal internal solution, n = 10 cells; Fig. 4C,D). Thus, a PI3 kinase pathway was likely involved in the insulin-induced enhancement of ASIC currents.

Figure 4
figure 4

Involvement of intracellular tyrosine kinase and PI3 kinase in the insulin-mediated enhancement of ASIC currents. The current traces in (A) and the bar graph in (B) show that the insulin (300 nM) mediated increase in ASIC currents was prevented by the tyrosine kinase inhibitor lavendustin A (5 μM), but not by its inactive analog lavendustin B (5 μM). ***p < 0.001, n.s., not significant, compared with the normal column. One-way ANOVA followed by Tukey's post hoc test. n = 10 cells from 6 rats in each column. The current traces in (C) and the bar graph in (D) show that IpH6.0 was enhanced by insulin (300 nM for 5 min) under conditions of normal internal solution, but not under conditions in which recording pipettes were filled with wortmannin (50 nM). ***p < 0.001, unpaired t-test, n = 10 cells from 6 rats in each column.

Full size image

Insulin potentiates acid-triggered action potentials in rat DRG neurons

We also probed whether insulin affects acid-triggered action potentials (APs) in rat DRG neurons under current-clamp conditions. In the presence of 5 μM AMG9810, an acidic stimulus at pH 6.0 not only induced a significant inward current under voltage-clamp conditions but also triggered a burst of APs under current-clamp conditions in the same DRG cell (Fig. 5A). Consistent with its enhancing effect on ASIC currents under voltage-clamp conditions, insulin also enhanced the bursts of acid-evoked APs. The number of APs evoked by acidic stimuli at pH 6.0 increased from 3.0 ± 0.7 in the control group to 6.3 ± 1.1 in the insulin pretreated group (300 nM for 5 min) in seven examined DRG neurons (p < 0.01, one-way ANOVA followed by Tukey's post hoc test, n = 7; Fig. 5B). Conversely, following the co-application of GSK1838705 (20 nM for 6 min) and insulin (300 nM for 5 min) to DRG neurons, the number of APs evoked by acidic stimuli at pH 6.0 was 3.4 ± 0.8, which was not significantly different from that of the control group (p > 0. 1, one-way ANOVA followed by Tukey's post hoc test, n = 7; Fig. 5B). These results indicated that insulin also potentiated acid-evoked APs via insulin receptors.

Figure 5
figure 5

Potentiation of acid-triggered action potentials by insulin in rat DRG neurons. (A) In the presence of 5 μM AMG9810, an acid stimulus at pH 6.0 caused an inward current and action potentials (APs) in the same DRG cell under voltage-clamp and current-clamp conditions, respectively. The original traces in (A) and the graphs in (B) illustrate that the number of acid-triggered APs was increased by the application of insulin (300 nM for 5 min) alone. The enhancing effect of insulin was blocked by co-treatment with GSK1838705 (20 nM for 6 min). **p < 0.01, one-way ANOVA followed by Tukey's post hoc test, n = 7 cells from 5 rats.

Full size image

Insulin exacerbates acid-induced nociceptive behaviors in rats

Finally, we investigated whether the insulin-induced enhancement of ASIC currents in vitro led to acid-induced nociceptive behaviors in vivo. Rats displayed spontaneous flinching/shaking responses when acetic acid (1% v/v, 50 μl) was injected into the hind paws, even with blockade of TRPV1 by AMG 9810 (10 μM). We observed exacerbated acid-elicited nociceptive behaviors in rats pretreated with insulin in the ipsilateral hind paws. Insulin administration resulted in a dose-dependent (5, 50 and 500 ng in 50 μl, with corresponding concentrations of 8.65 nM,86.5 nM and 865 nM) increase in the number of acid-elicited flinching episodes, as per the quantitative analysis in Fig. 6A (p < 0.05 and 0.01, one-way ANOVA followed by Tukey's post hoc test, n = 9). The exacerbating effect of insulin (500 ng) on acid-triggered nociceptive behaviors was significantly attenuated in rats cotreated with the insulin receptor antagonist GSK1838705 (100 ng) in the ipsilateral hind paws (p < 0.01, compared with 500 ng insulin, one-way ANOVA followed by Tukey's post hoc test, n = 9), and the mean number of flinches in these rats was not different from that in the control rats treated with acetic acid alone. These results suggested that insulin can dose-dependently increase acid-induced spontaneous nociceptive behaviors through peripheral insulin receptors.

Figure 6
figure 6

Exacerbation of acid-induced nociceptive behaviors by insulin in rats. (A) Intraplantar injection of acetic acid (1%, 50 μl) resulted in significant flinching behaviors even in the presence of the TRPV1 inhibitor AMG9810 (10 μM). Intraplantar pretreatment with insulin (5, 50 or 500 ng) dose-dependently increased the number of acid-induced flinching. The effect of insulin (500 ng in 50 μl) on flinching behaviors was prevented by co-treatment with the insulin receptor antagonist GSK1838705 (100 ng). *p < 0.05, **p < 0.01, compared with the control; ##p < 0.01, compared with 500 ng insulin; one-way ANOVA followed by Tukey's post hoc test. Each column represents the mean ± S.E.M. of 9 rats. (B) Effects of intramuscular administration of insulin on mechanical hypersensitivity in the acid-induced muscle pain model. Acidic saline (pH 4.0) was injected into the gastrocnemius muscle of each rat twice, 5 days apart. Insulin (1 μg in 100 μl) was administered prior to the second acid injection. Mechanical hypersensitivity was assessed by measuring the paw withdrawal threshold (PWT, in g) to von Frey filament stimulation of the ipsilateral hind paws. **p < 0.01, compared with the vehicle group; ##p < 0.01, compared with the insulin group; two-way repeated measures ANOVA followed by Bonferroni's post hoc correction, n = 9 rats per group.

Full size image

Intramuscular acid injections result in mechanical nociceptive responses. The effects of insulin were further investigated in an acid-induced muscle pain model. In this model, two acid injections (pH 4.0) into the left gastrocnemius muscle of rats 5 days apart produced long-lasting mechanical hypersensitivity, as indicated by a decrease in the paw withdrawal threshold (PWT). Figure 6B shows that the PWT of the rats decreased at 4 h after the second acid injection on the fifth day and on the sixth and ninth days. Insulin (1 μg in 100 μl) was administered by intramuscular injection prior to the second acid injection. We observed that insulin treatment significantly decreased the PWT after the second acid injection in the ipsilateral hind paws (treatment: F1,133 = 33.4, p < 0.05 and 0.01, two- way ANOVA with Bonferroni's post hoc correction, compared with the vehicle group, n = 9; Fig. 6B), suggesting a significant exacerbation of acid-induced persistent mechanical hypersensitivity. However, co-application of GSK1838705 (200 ng) and insulin (1 μg) to the ipsilateral gastrocnemius muscle had no effect on acid-induced persistent mechanical hypersensitivity, and the PWTs of these rats were similar to those produced by vehicle injection (treatment: F1,333 = 0.86, p > 0.1, two-way ANOVA with Bonferroni's post hoc correction, compared with vehicle, n = 9; Fig. 6B). These data suggested that insulin exacerbated the persistent mechanical hyperalgesia induced by repeated acid injection through peripheral insulin receptors.

Discussion

Electrophysiological recordings and behaviors were used to investigate the effects of insulin on ASICs. The results showed that insulin enhanced the functional activity of ASICs. Exposure to insulin for 5 min increased the acid-evoked whole-cell current response and APs in dissociated rat DRG neurons. This potentiation was dependent on insulin receptors and intracellular signaling pathways involving tyrosine kinase and PI3 kinase. Furthermore, insulin exacerbated acid-induced spontaneous nociceptive behaviors and mechanical hyperalgesia induced by intramuscular acid injections through peripheral insulin receptors.

In the presence of AMG9810, TRPV1 was not activated, just as capsaicin failed to induce any membrane currents. In the present study, ASIC3 homomeric or ASIC3-containing heteromeric channels were found to be responsible for mediating acid-evoked currents, because these currents were blocked not only by the ASIC blocker amiloride but also by the ASIC3 blocker APETx2. This finding was consistent with the morphological evidence that ASICs are expressed in DRG neurons, with ASIC3 being the most prevalent ASIC subunit21,22,23. We therefore considered that these acid-evoked currents may be ASIC or ASIC3-like currents, although precise ASIC subunits need to be identified. We observed that ASIC currents were rapidly enhanced by the preapplication of insulin to rat DRG neurons. Insulin-induced potentiation has the following characteristics. The enhancement was rapid and reversible. Insulin enhanced ASIC currents in a concentration-dependent manner. Insulin increased the maximum response of ASICs to acid without affecting their sensitivity to acidic stimuli. This characteristic contrasts with the insulin modulation of glycine receptors because the agonist potency rather than the number of functional glycine receptors is increased13. The effects of insulin on ASICs may be similar to those that modify NMDA and GABAA receptors. The enhancement of NMDA and GABAA receptor expression by insulin results from the delivery of new channels to the cell membrane38,39. Insulin also facilitates the capsaicin response of TRPV1 and the translocation of TRPV1 and TRPV2 to the cell surface in DRG neurons and pancreatic β cells40,41. However, it remains to be determined whether the effect of insulin on ASICs is mediated by a trafficking mechanism or other mechanisms, as the effect of insulin occurs rapidly and is reversible within a few minutes. Consistent with the results of the voltage-clamp experiments, insulin also rapidly increased the number of APs evoked by acidic pH solution under current-clamp conditions, suggesting enhanced membrane excitability of rat DRG neurons.

Were the effects of insulin on ASICs direct or indirect? The present study indicated that the insulin-mediated increase in the functional activity of ASICs required the activation of insulin receptors, since the insulin-induced increase in acid-evoked currents and APs in nociceptive DRG neurons and acid-induced nociceptive behaviors in rats were completely blocked by the insulin receptor antagonist GSK1838705. In contrast, heat-inactivated insulin failed to change ASIC currents. In addition, only some, but of all, ASIC-like responses were enhanced by insulin, which may be related to the degree of the co-expression of insulin receptors and ASICs in rat DRG neurons. Therefore, we believe that insulin indirectly acts on ASICs, which occurs only in DRG neurons coexpressing insulin receptors and ASICs. Early studies demonstrated the presence of insulin receptor immunoreactivity in a population of small DRG neurons and peripheral terminals10,12. Studies on rat cultured primary sensory neurons have shown that approximately 50% of neurons express insulin receptors42. ASIC3 is expressed in more than 65% of DRG neurons23. It is possible that the insulin receptor and ASICs/ASIC3 are co-expressed in some DRG neurons, although morphological evidence is needed.

The insulin receptor is a member of the receptor tyrosine kinase family43. The tyrosine kinase inhibitor lavendustin A suppresses the autophosphorylation of tyrosine residues and the subsequent phosphorylation of insulin receptor substrate proteins. The present results indicated the involvement of tyrosine kinase phosphorylation in the insulin-induced enhancement of ASIC currents in DRG neurons, since this enhancement was blocked by the tyrosine kinase inhibitor lavendustin A but not by its inactive analog lavendustin B. Insulin binding to its receptor activates the intracellular transduction pathway of PI3 kinase, which is a Src homology 2 ___domain-containing lipid kinase, by phosphorylating intracellular insulin receptor substrate proteins44,45. The present data showed that PI3 kinase may underlie the insulin-induced increase in the activity of ASICs since the administration of the PI3 kinase inhibitor wortmannin prevented the insulin-induced increase in ASIC currents. It is known that insulin can modulate the trafficking of proteins through PI3 kinase signaling, thereby inserting some ion channels into the plasma membrane, such as GABAA receptors46. In addition, activation of PI3 kinase induces membrane insertion of transient receptor potential channels and voltage-dependent calcium channels47,48. In both rat spinal dorsal horn neurons and heterologous cell cultures, brain-derived neurotrophic factor (BDNF) increases the ASIC currents mediated by homomeric ASIC1a, ASIC2a or heteromeric ASIC1a + ASIC2a channels49. BDNF can enhance forward trafficking and increase the surface expression of these ASIC channels via PI3 kinase signaling, resulting in enhanced ASIC currents49. However, BDNF does not influence the ASIC currents mediated by homomeric ASIC1b or ASIC3 channels49. The present data showed that the insulin-mediated enhancement of ASIC3-like currents is dependent on PI3 kinase signalingt. One explanation may be that the present ASIC3-like currents were due to activation of ASIC3-containing heteromeric channels rather than ASIC3 homotrimers, such as heteromeric ASIC1a + ASIC3 and heteromeric ASIC2a + ASIC3. In addition, we observed that the Hill coefficient for ASIC3-like current activation in DRG neurons was lower than the slope or Hill coefficient of 2.36-4 in homomeric ASIC3 expressing CHO cells50,51, suggesting that the present ASIC3-like currents may be mediated by ASIC3-containing heteromeric channels.

Tissue acidosis commonly develops in a variety of painful conditions. ASICs, especially ASIC3, localized on peripheral nociceptors can detect pH changes and respond to acidosis52. Exposure of ASICs to an acidic pH causes cation influx, neuronal depolarization and action potential firing in peripheral nociceptors, resulting in acid-induced nociceptive behaviors53. Behavioral studies have shown that rats display an intense flinching/shaking response after acetic acid is injected into their hind paws23,36. We observed that ipsilateral intraplantar injection of insulin dose-dependently augmented acid-induced spontaneous nociceptive responses in rats through peripheral insulin receptors. The enhancement of ASIC-mediated electrophysiological activity by insulin/insulin receptor signaling in DRG neurons could underlie the exacerbation of acid-induced spontaneous nociceptive behaviors. Repeated intramuscular acid injections produce persistent mechanical hyperalgesia. ASICs are necessary for the development of muscle-induced hyperalgesia33,54. Local intramuscular injection of insulin was also found to exacerbate the persistent mechanical hyperalgesia in a model of muscle pain through peripheral insulin receptors. Hotta et al. reported that intramuscular injection of acid or insulin sensitizes thin-fiber muscle afferents to mechanical stimuli17,55. Primary sensory neurons innervating the gastrocnemius muscle express insulin receptors56. The present results indicated that the interactions between ASICs and insulin signaling may contribute to the exacerbating effect of insulin on the mechanical hyperalgesia produced by repeated intramuscular acid injections. Our results did not exclude the possibility that peripheral insulin exacerbates acid-induced nociceptive behaviors through other mechanisms, such as regulation of voltage-gated channels. In diabetic patients, the intensive administration of insulin is reported to induce severe and acute neuropathic pain, which is referred to as insulin neuritis or treatment-induced diabetic neuropathy57,58. However, the mechanisms of insulin neuritis remain to be clarified59. The present findings may provide insight into the mechanisms underlying the development of insulin neuritis.

We also acknowledge the limitations of the present study. The insulin concentrations used in our experiments were higher than the plasma insulin concentrations measured in rats or humans60,61. Even insulin concentrations in hyperinsuline patients are in the nanomolar range62. The concentration of insulin in the central nervous system is unknown but may be as high as 200 nM63. The lowest concentration of insulin that increased ASIC currents was 30 nM, suggesting that a pharmacological rather than a physiological concentration of insulin enhances ASIC currents. We do not know the exact tissue concentration of insulin that reaches primary sensory afferents under physiological and pathophysiological conditions. Therefore, it remains to be determined whether the effect of insulin on ASICs contributes to acid-induced pain under pathophysiological conditions.

Conclusions

In summary, our results indicated that ASICs, particularly ASIC3, are downstream modulation targets of insulin/insulin receptor signaling. Insulin enhanced ASIC-mediated electrophysiological activity, spontaneous nociceptive behaviors and muscle pain via the intracellular tyrosine kinase and PI3 kinase signaling pathways.