Report

The Molecular Basis of Acid Insensitivity in the African Naked Mole-Rat

See allHide authors and affiliations

Science  16 Dec 2011:
Vol. 334, Issue 6062, pp. 1557-1560
DOI: 10.1126/science.1213760

Abstract

Acid evokes pain by exciting nociceptors; the acid sensors are proton-gated ion channels that depolarize neurons. The naked mole-rat (Heterocephalus glaber) is exceptional in its acid insensitivity, but acid sensors (acid-sensing ion channels and the transient receptor potential vanilloid-1 ion channel) in naked mole-rat nociceptors are similar to those in other vertebrates. Acid inhibition of voltage-gated sodium currents is more profound in naked mole-rat nociceptors than in mouse nociceptors, however, which effectively prevents acid-induced action potential initiation. We describe a species-specific variant of the nociceptor sodium channel NaV1.7, which is potently blocked by protons and can account for acid insensitivity in this species. Thus, evolutionary pressure has selected for an NaV1.7 gene variant that tips the balance from proton-induced excitation to inhibition of action potential initiation to abolish acid nociception.

Acid is a noxious and painful stimulus in all vertebrates examined, with one known exception among mammals: the African naked mole-rat (1, 2). The absence of acid nociception can be explained by the acid-insensitivity of C-fiber nociceptors and may have resulted from exposure to high-CO2 conditions that accompanies the communal living of many naked mole-rats underground (2, 3). We sought to identify factors that render naked mole-rat nociceptors insensitive to acid. Two main acid sensors are proposed to participate in acid-mediated nociception: the acid-sensing ion channel (ASIC) family and the transient receptor potential vanilloid-1 ion channel (TRPV1) (1, 4). We first asked whether naked mole-rat nociceptors simply lack proton-gated ion channels by using whole-cell patch-clamp to compare proton-gated currents in nociceptors isolated from naked mole-rat and mouse dorsal root ganglia (DRG) [supporting online material (SOM) methods]. Surprisingly, both ASIC-like currents, which are benzamil-sensitive and transient, as well as TRPV1-like currents that are sustained-only and blocked by N-(4-tertiarybutylphenyl)-4-(3-chloropyridin-2-yl)tetrahy-dropyrazine-1(2H)-carboxamide (BCTC) (5) were present both in naked mole-rat and mouse nociceptors (Fig. 1, A and B, and fig. S1). Only minor differences were noted in the incidence and kinetic properties of proton-gated currents between mouse and naked mole-rat (fig. S1 and table S1). Real-time polymerase chain reaction (PCR) analysis revealed that mRNA encoding both TRPV1 and ASIC subunits were similarly abundant in mouse and naked mole-rat DRGs (fig. S2). Therefore, the acid insensitivity of naked mole-rat nociceptors cannot be explained by minor differences observed in proton-gated ion currents in this species as compared with mice (6, 7).

Fig. 1

Cloned and native naked mole-rat acid sensors are similar to those of mice. (A) Acid evokes transient (T), benzamil-sensitive, inward currents followed by a small sustained phase (S) and sustained-only currents, (B) 52% of which are capsaicin-sensitive and inhibited by the TRPV1-selective antagonist BCTC. (C) nmrASIC1a and mASIC1a are acid-sensitive, pH50 6.03 (n = 12 to 18 cells) and 6.28 (n = 12 to 13 cells), respectively, and undergo SSI. nmrASIC1a, IC50 = 7.38 (n = 3 to 22 cells); mASIC1a, IC50 = 7.46 (n = 10 cells). (D) nmrTRPV1 and mTRPV1 are acid-sensitive, pH50 5.31 (n = 11 to 13 cells) and 5.22 (n = 10 cells). **P < 0.01, ***P < 0.001.

We next cloned and expressed recombinant ASIC1a, ASIC1b, and TRPV1 channels to determine whether naked mole-rat channels exhibit unusual pH sensitivity. The proton sensitivity and pharmacology of naked mole-rat ASIC1a (nmrASIC1a), ASIC1b (nmrASIC1b), and TRPV1 (nmrTRPV1) were essentially the same as their mouse counterparts (Fig. 1, C and D, fig. S3, and table S2). Both mouse and naked mole-rat ASICs displayed steady-state inactivation (SSI) (Fig. 1C, fig. S3, and table S2), a phenomenon in which mild acidification (pH 7.0) inactivates the channels, impairing their ability to open after challenge with much lower pH (8). The rate at which recombinant nmrASIC1a- and nmrASIC1b-mediated currents inactivated was similar to that found for native naked mole-rat ASIC-like currents (fig. S1) and was significantly faster than their mouse orthologs (table S2). nmrTRPV1 had properties characteristic of other mammalian TRPV1s; it was capsaicin-sensitive, outwardly rectifying, and heat-gated (threshold 42.6° ± 0.5°C, n = 15 cells) (Fig. 1D and fig. S3). Based on BCTC-sensitivity, ~52% of sustained-only currents in naked mole-rat DRG neurons were mediated via nmrTRPV1 channels (Fig. 1B).

Consistent with these results, increasingly acidic stimuli evoked graded increases in inward current and increasing depolarization in both mouse and naked mole-rat DRG neurons (fig. S4). However, despite pronounced membrane depolarization few neurons in the mouse (4.3%) or naked mole-rat (13.3%) fire an action potential (AP) after acid stimulation (Fig. 2, A and B), although the depolarization was sufficient to activate voltage-gated sodium channels (NaVs) to trigger APs. Approximately 20% of C-fibers are excited by acid in skin-nerve studies (9, 10), and so we made additional recordings from identified skin DRG neurons, retrogradely labeled from mouse skin. Stimulation with pH 5.0 evoked APs more commonly in identified skin neurons (18.5%), and these neurons exhibited larger peak current densities and larger, transient, ASIC-like currents at pH 5.0 than did nonidentified mouse neurons (fig. S5 and table S1). Thus, larger current/bigger depolarization increases AP probability, but there does not appear to be a voltage threshold above which neurons always spike (fig. S5). AP spiking DRG neurons were, however, on average more depolarized by acid than were non-spikers (–10.4 ± 1.3 mV versus –26.0 ± 1.3 mV, P < 0.001, unpaired t test) (fig. S6). In both mouse and naked mole-rat DRG neurons, we observed that acid-evoked transient currents were significantly more likely to generate an AP than were sustained-only currents (P < 0.05 and 0.01 respectively, Fisher’s exact test). Why can acid evoke APs in naked mole-rat DRG neurons but still not excite C-fiber nociceptors? (2). This discrepancy may arise because in our in vitro experiments, pH stimuli are applied rapidly (<0.5 ms), but tissue acidification is probably buffered so that pH falls only gradually. Gradual acidification will favor SSI of ASIC-mediated currents (Fig. 1C), and indeed, transient currents in naked mole-rats DRG neurons inactivated with a half-maximal inhibitory concentration (IC50) of pH 7.31 (Fig. 2C). In current-clamp mode, gradual acidification to pH 5.0 produced a similar absolute depolarization as step acidification from pH 7.4 to pH 5.0 (Fig. 2D), but gradual acidification never evoked APs (n = 0 of 30 cells), whereas step acidification did (n = 12 of 69 cells, P < 0.05 Fisher’s exact test). This suggests that ASIC-mediated APs only occur when there are rapid shifts in pH. Moreover, in the same mouse DRG neurons large acid-induced depolarization produced no AP, but 50 μM adenosine 5′-triphosphate (ATP) triggered a small depolarization and many APs (Fig. 2E). Thus, during acid-evoked depolarization a concomitant inhibitory process prevents AP firing.

Fig. 2

Proton-gated currents poorly drive AP spiking in DRG neurons. In (A) mouse and (B) naked mole-rat, larger proton-gated currents produce larger depolarizations, but no threshold exists above which all cells spike. (C) Transient currents undergo SSI in naked mole-rat DRG neurons (IC50 = 7.31, n = 17 cells). (D) Gradual acidification does not produce spiking, although the same depolarization is reached with a step from pH 7.4 to 5.0 (compare black curve, n = 30 cells, with red point, n = 69 cells). (E) 50 μM ATP evokes APs in cells, whereas pH 5.0 does not.

Protons can inhibit NaV channels (11, 12), and so we next examined acid inhibition of macroscopic, voltage-gated inward currents in mouse DRG neurons. At pH 6.0, there was a decrease in inward current amplitude and a depolarizing shift in the half-maximal voltage for activation (V1/2); peak inward conductance was inhibited by 42% (Fig. 3, A and B, and fig. S7). The IC50 for this inhibition was pH 5.95, which was accompanied by a pH-dependent depolarizing shift in the V1/2 for activation (fig. S7 and table S3). Unlike voltage-gated inward currents, macroscopic outward currents only showed a large inhibition at pH 5.0 (fig. S7). Pharmacologically isolated NaV currents exhibited a 46% reduction in conductance at pH 6.0 (fig. S7)—a pH sensitivity similar to the macroscopic current. These data predict that acid would raise AP firing thresholds, and AP threshold was increased significantly in neurons exposed to pH ≤6.0, and AP amplitudes decreased (fig. S8). The characteristic hump on the falling phase of the AP in nociceptors (1, 13, 14) was often absent at low pH, so that 41% lacked a hump at pH 6.0, and all APs at pH 5.0 were humpless (fig. S8).

Fig. 3

Increased acid block of voltage-gated inward currents and AP initiation in naked mole-rat neurons. (A) pH 6.0 inhibits voltage-gated inward current amplitude and produces a depolarizing shift in V1/2 for activation [outward currents digitally removed (fig. S7)]. (B) pH 6.0 inhibits macroscopic voltage-gated inward conductance by 42% in mouse and (C) 63% in naked mole-rat, which is (D) a significant difference. (E) Mechanically evoked APs in C-fibers are inhibited by acid (red traces) in (left) mouse and (right) naked mole-rat, but (F) significantly more in naked mole-rat. **P < 0.01.

We next tested whether voltage-gated inward currents in naked mole-rat are more susceptible to acid inhibition than are those in mice. Macroscopic voltage-gated inward currents in naked mole-rat neurons were inhibited by pH 6.0, coupled with a depolarizing shift in V1/2 for activation, resulting in a 63% decrease in conductance (Fig. 3C). The 63% inhibition of conductance at pH 6.0 was significantly greater than the 42% inhibition observed in mouse neurons (P < 0.01, unpaired t test) (Fig. 3D). If acid is more potent in inhibiting voltage-gated inward currents in naked mole-rat neurons, then low pH should also inhibit nociceptor firing to natural stimuli more profoundly than in the mouse. Indeed, pH 4.0 and pH 6.0 solutions inhibited mechanically evoked C-fiber nociceptor firing more potently in naked mole-rat than in mouse [P < 0.05, 2-way analysis of variance (ANOVA)] (Fig. 3, E and F). Furthermore, the overall effect of exposure to pH 6.0 in mouse C-fibers was one of sensitization, as has been observed in the rat (10), whereas most naked mole-rat C-fibers were inhibited at pH 6.0 (fig. S9). Electrically evoked APs were also blocked by acid (fig. S9). Thus, proton inhibition of NaV currents in nociceptors directly inhibits natural activation of nociceptors in vivo.

The threshold for AP generation in nociceptors is largely determined by NaV1.7 (15); nonsense mutations in the human SCN9A gene encoding NaV1.7 produce complete insensitivity to pain (16, 17); and extreme pain disorders are associated with activating NaV1.7 mutations (1820). The outer carboxylate ring in the pore loop of NaV channels is responsible for determining acid sensitivity (11). Therefore, we sequenced the naked mole-rat NaV1.7 transcript (nmrNaV1.7, 94.7% sequenced based on human NaV1.7) and looked for variants in regions that might alter the pH sensitivity of NaV1.7. In nmrNaV1.7, domain IV—a highly conserved, positively charged amino acid motif (KKV, amino acids 1718 to 1720 in human NaV1.7)—was negatively charged (EKE) (Fig. 4A). We introduced the EKE motif into hNaV1.7 and compared acid sensitivity of hNaV1.7 and hNaV1.7EKE. In the presence of acid (pH 6.0), hNaV1.7EKE was significantly more inhibited, so that the conductance was inhibited by 73% compared with 58% for hNaV1.7 (P < 0.001) (Fig. 4, B to D). Thus, naked mole-rats harbor a distinct amino acid motif within the NaV1.7 gene, and the exchange of two amino acids renders the channel more sensitive to proton block than other vertebrate NaV channels. The degree of proton inhibition of naked mole-rat–like hNaV1.7EKE channels was similar to that of macroscopic voltage-gated inward currents in naked mole-rats. Sequencing of the nmrNaV1.8 transcript did not identify any charge-changing motifs around the outer carboxylate ring.

Fig. 4

A sequence variation in naked mole-rat NaV1.7 enhances acid block. (A) Alignment of NaV1.7 domain IV. nmrNaV1.7 harbors a negatively charged EKE motif, in contrast to more positively charged residues in other mammalian NaV1.7s (green boxes, charges next to alignment; orange square, outer carboxylate ring; and yellow star, EKE motif). (B) Wild-type hNaV1.7 and (C) hNaV1.7EKE currents are inhibited by pH 6.0; (D) hNaV1.7EKE is significantly more inhibited. (E) Proton-mediated C-fiber nociceptor activation is a balance between (red arrow) depolarizing input and (gray bar) proton-mediated inhibition of NaV1.7. In mouse, acid can evoke nociceptor AP firing; however, the increased nmrNaV1.7 proton sensitivity in naked mole-rats largely prevents acid-driven AP initiation and behavior. ***P < 0.001.

Lack of proton sensors in naked mole-rat nociceptors cannot explain acid insensitivity in this species, but rather, protons act to put a brake on AP initiation by potently blocking nmrNaV1.7. Our data supports a new model for understanding how acid excites mammalian nociceptors: Depolarizing input (from TRPV1 and ASICs) must overcome the concurrent inhibition of NaV channels (Fig. 4E). This balance is shifted in naked mole-rat neurons because nmrNaV1.7 is more potently blocked by acid, resulting in no acid-evoked APs. In their subterranean niche, naked mole-rats are exposed to an unusually high-CO2/low-O2 environment (3), and high CO2 can drive tissue acidosis (21). Another vertebrate species exposed to a high-CO2 environment is the cave-roosting microbat Myotis lucifugus, and it carries a motif charged similarly to the naked mole-rat EKE motif in its NaV1.7 gene. In contrast, the tree-roosting non–CO2-challenged megabat Pteropus vampyrus carries a more typical, less negatively charged motif NKV in its NaV1.7 gene (Fig. 4A). Thus, convergent evolution on species with similar high-CO2 living conditions may have selected for NaV1.7 gene variants that reduce acid nociception. Our study illustrates that the extreme physiology of the naked mole-rat offers particular insights into normal physiology relevant to inflammatory pain processes in which tissue acidosis plays a role (22).

Supporting Online Material

www.sciencemag.org/cgi/content/full/334/6062/1557/DC1

Materials and Methods

Figs. S1 to S9

Tables S1 to S3

References (23, 24)

References and Notes

  1. Acknowledgments: E.St.J.S. carried out DRG electrophysiology. E.St.J.S. and D.O. carried out CHO cell electrophysiology. S.L. carried out skin-nerve electrophysiology. D.O., L.L., G.A., and E.St.J.S. carried out molecular biology. E.St.J.S. and G.R.L. designed experimental studies and wrote the paper. E.St.J.S. was supported by the Alexander von Humboldt foundation, and D.O. was supprted by the MolNeuro Helmholtz Research School. Additional funding was obtained from Deutsche Forschungsgemeinschaft collaborative research center 665 and Exc 257 (NeuroCure). We thank H. Thränhardt, A. Wegner, and K. Barda for technical support. Sequences have been submitted to GenBank (SOM).
View Abstract

Navigate This Article