Amiloride is a nonspecific blocker of acid-sensing ion channels (ASICs) that have been recently implicated in the mediation of mechanical and chemical/inflammatory nociception. Preliminary data using a transgenic model are suggestive of sex differences in the role of ASICs. We report here that systemic administration of amiloride (10–70 mg/kg ip) produces a robust, dose-dependent blockade of late/tonic phase nociceptive behavior on the mouse formalin test (5%; 20 μl) in female but not male mice, completely abolishing the known sex difference in formalin test response. Adult gonadectomy produced a “switching” of sex differences in amiloride efficacy, with castrated males displaying an amiloride blockade and ovariectomized females rendered less sensitive to amiloride. Gonadectomized mice could be switched back to their intact status using chronic estrogen benzoate or testosterone propionate replacement via osmotic minipump (6 μg/day or 250 μg/day, respectively). It is unclear whether this striking sex difference is due to sex-specific involvement of ASICs in pain processing, but the present data represent one of the first demonstrations of pain-related sex differences with no obvious opioid involvement.
- hormone replacement
women are greatly overrepresented among the sufferers of chronic pain syndromes (6, 46). This clinical reality may have a basis in perceptual biology, since women also exhibit greater sensitivity to experimental pain across a number of modalities (40). To identify underlying mechanisms, rodent models have been studied, and rat and mouse sex differences in nociception are now well appreciated, even though males continue to be overwhelmingly chosen as the sole subjects of basic science research in pain (34). Studying rodent sex differences in acute, thermal nociception (the most commonly used stimulus) may not be the best choice of model, both in terms of its questionable clinical relevance and because of lingering confusion in the literature as to which sex is, in fact, more sensitive (see Ref. 35). By contrast, in the formalin test of chemical/inflammatory pain (16), female rats and mice have consistently been found, by a number of different investigators, to display greater sensitivity than males (1, 21, 25, 39; see also Ref. 12) when differences were observed. Females also appear to be more sensitive to the hypersensitivity produced by inflammatory injury (3, 8, 14). A number of mechanisms have been proposed to explain rodent sex differences in nociception and its inhibition by opioids, ranging from body size and blood pressure to neurosteroids, receptors, and signal transduction molecules (see Ref. 32). Gonadal hormones obviously play a primary role in mediating these sex differences, although debate continues as to the relative involvement of estrogen, progesterone, and testosterone, as well as the relative importance of organizational vs. activational effects (see Ref. 32). Thus far, however, the existing evidence suggests that the sexes modulate pain differently, perhaps using distinct, sex-specific neural circuitry in the midbrain, brain stem, and spinal cord (28, 36, 45). In the present study, we find evidence suggesting that sex differences in pain processing may also exist at the very first stages of sensory transduction.
Acid-sensing ion channels (ASICs) are members of the ENaC/DEG superfamily of amiloride-sensitive epithelial sodium channels (see Ref. 26). ASICs are distinct among the ENaC/DEGs in that they are proton-gated sodium channels proposed as transducers of acid-evoked pain, mechanical pain, and innocuous touch. Furthermore, ASICs contribute to pain processing and central sensitization during pathological states, such as inflammation and ischemia, both of which are accompanied by moderate-to-severe tissue acidosis (29, 47). Previously, we examined the pain phenotype of mice bearing a dominant-negative mutation of the ASIC3 gene, which renders ASIC1, ASIC2, and ASIC3 channel subunits inactive through oligomerization, thereby abolishing all ASIC-related currents in sensory neurons (33). We found that mutant mice exhibited increased sensitivity to acute mechanical and chemical/inflammatory pain, as well as increased mechanical hypersensitivity following chronic inflammation and muscle acidification (33). However, on further analysis, we discovered that the ASIC3 dominant-negative mutation exerted a sex-specific impact on nociception: only male, but not female, mutants had heightened nociceptive sensitivity relative to their wild-type counterparts (9).
Thus, in an initial attempt to further investigate sex differences in the contribution of ASICs to nociception, we administered the nonspecific ASIC-blocker (and K+-sparing diuretic), amiloride, to outbred mice of both sexes. Administration of amiloride has been shown to inhibit pain in laboratory animals across a variety of nociceptive assays (15, 18, 41), as has a novel ASIC blocker, A-317567 (15). Although the mechanisms of its analgesic action are not fully understood, amiloride has been shown to effectively block acid-evoked ASIC-like currents in sensory neurons (15). In the present study, we characterized sex differences in amiloride inhibition of responding in the formalin test of chemical/inflammatory nociception and then investigated the hormonal mechanisms underlying this difference through gonadectomy and hormone replacement experiments.
MATERIALS AND METHODS
Naïve, adult male and female CD-1 (Crl:ICR) mice were obtained from Charles River Laboratories (Boucherville, QC, Canada). Upon our request, some 6-wk-old mice were gonadectomized or sham gonadectomized by personnel at Charles River at least 1 wk before shipment. Mice were housed in same-sex groups of four and acclimatized to our vivarium at McGill University for at least 1 wk before any experiments. The vivarium was temperature controlled and maintained on a 12:12-h light-dark cycle (lights on at 0700). Access to food (Harlan Teklad 8604) and water was ad libitum. All procedures were in accordance with the guidelines specified by the Canadian Council on Animal Care.
Amiloride hydrochloride hydrate, 2-hydroxypropyl-β-cyclodextrin (cyclodextrin), estrogen benzoate, polyethylene glycol (PEG), and testosterone propionate were all obtained from Sigma (St. Louis, MO). Amiloride or PEG vehicle were administered in every experiment via intraperitoneal injection. In hormone replacement experiments, estrogen benzoate, testosterone propionate, or cyclodextrin vehicle were administered via subcutaneously implanted osmotic minipumps (ALZET model 2002) at a rate of 0.5 μl/h over 7 (estrogen) or 14 (testosterone) days. Mice were anesthetized with isoflurane/oxygen, and osmotic minipumps were implanted via a small (∼1 cm) incision into the upper back. The incision was closed with sterile 9-mm stainless steel wound clips, and animals were placed in a recovery chamber for 30 min before being returned to their home cages.
The formalin test was employed using procedures described in considerable detail previously (37). Mice were habituated singly for 30 min to Plexiglas observation chambers (15 cm diameter; 22.5 cm high) atop a glass floor and were then treated with amiloride (0–70 mg/kg ip, dissolved in 30% PEG in physiological saline) or vehicle. The 30-mg/kg dose of amiloride administered was previously reported to induce half-maximal antinociception on the formalin test in mice (18). Thirty minutes after drug administration, all mice received 20 μl of 5% formalin injected into the plantar surface of the right hind paw. Behavioral responses to formalin were captured by video cameras located underneath the glass floor and scored later using Observer software (Noldus). Mice were sampled for 5 s at the start of every minute for 1 h for the occurrence of licking/biting of the affected paw (10). The percentage of total number of samples with licking/biting behavior in the acute, early phase (0–10 min) and the tonic, late phase (10–60 min) were recorded. Immediately after behavioral testing, mice were killed, and hind paws were severed and weighed to quantify inflammation. Edema was defined by the difference in hindpaw weights divided by the weight of the contralateral paw, expressed as a percentage. Data obtained from mice with <20% inflammation (n = 13) were discarded before analysis.
Four separate experiments were conducted. The first compared formalin test sensitivity of male and female mice given various doses of amiloride (n = 7–18/dose/drug/sex). The second experiment investigated the effects of gonadectomy (i.e., castration and ovariectomy) on amiloride’s actions in male and female mice, respectively (n = 20–29/sex/surgery/drug). The third experiment focused on the possible role of estrogen in mediating the sex differences in amiloride efficacy, and attempted to reverse the effect of ovariectomy via hormone replacement with estrogen benzoate (n = 16/hormone/drug). The final experiment focused on the possible role of testosterone in mediating the sex difference in amiloride efficacy and attempted to reverse the effect of castration via hormone replacement with testosterone propionate (n = 15–22/hormone/drug).
Ovariectomized female mice were treated with estrogen benzoate (6 μg/day, dissolved in 40% cyclodextrin in physiological saline) or vehicle via osmotic minipump. On day 7 following the beginning of this treatment, mice were administered either amiloride (30 mg/kg ip) or vehicle and tested on the formalin test as described above. Immediately following testing and death, the uteri of all mice in this experiment were removed and weighed to confirm the efficacy of the estrogen replacement.
Castrated male mice were treated with testosterone propionate (250 μg/day dissolved in 40% cyclodextrin in physiological saline) or vehicle via osmotic minipump. On day 14 following the beginning of this treatment, mice were administered either amiloride (30 mg/kg ip) or vehicle and tested on the formalin test as described above. Immediately following testing and death, the seminal vesicles of all mice in this experiment were removed and weighed to confirm the efficacy of the testosterone replacement.
Dose-response data were analyzed by the method of Tallarida and Murray (44), as implemented by the FlashCalc 21.5 program (M. Ossipov, University of Arizona). Data were analyzed by ANOVA, with sex, surgery, hormone, and/or drug as between-subjects factors as appropriate (see results). ANOVAs were followed by two-tailed Student’s t-tests where appropriate. A significance level of α = 0.05 was adopted.
In all experiments, the expected biphasic pattern of formalin recuperative behaviors (in mice; mainly licking behavior; see Ref. 43) was observed (see Figs. 1 –⇓⇓4), and most mice were displaying minimal levels of licking behavior by 60 min postinjection. Besides decreasing formalin licking in some conditions, amiloride did not appear to affect either the qualitative or timing aspects of formalin responding.
Amiloride dose-response experiment.
As can be seen in Fig. 1, amiloride affected formalin licking behavior in the late (10–60 min) but not early (0–10 min) phase. In female mice, the half-maximal antinociceptive dose (AD50) of amiloride against late-phase licking was calculated as 34 mg/kg (95% confidence interval: 24–49 mg/kg). By extrapolation, the AD50 in males was estimated as 136 mg/kg, but this estimate should be treated with great caution because only the highest dose (70 mg/kg ip) produced any evidence of an effect. We were unable to try higher doses because we observed a number of lethalities in a separate group of mice given 70 mg/kg and observed for several hours. Inspection of Fig. 1B suggested that 30 mg/kg amiloride was the most appropriate single dose for further experimentation, because it produced a half-maximal inhibition of responding in female mice, produced absolutely no effect whatsoever in male mice, and was fully able to abolish the preexisting sex difference in formalin behavior. In comparison of vehicle and 30 mg/kg amiloride, two-way ANOVAs (sex × dose) revealed no main effects or interactions in the early phase, a significant main effect of drug (F1,67 = 9.1, P < 0.01) and a significant sex × drug interaction (F1,67 = 4.7, P < 0.05) in the late phase. Figure 1, C and D illustrates the robust blockade of formalin-induced licking behavior throughout the late phase by 30 mg/kg amiloride in female, but not male, CD-1 mice. This was confirmed statistically by the fact that a two-between, one-within repeated-measures ANOVA revealed a significant repeated-measures effect (F1,759= 22.9, P < 0.001) but no interaction of repeated measures with either sex, drug, or sex × drug conditions.
In gonadally intact mice, there was no effect of sex or amiloride on the inflammation produced by formalin injection assessed immediately postmortem at 60 min postinjection (data not shown).
Effects of gonadectomy.
Figure 2 shows the results of the gonadectomy experiment. In the early phase, we observed a significant drug × surgery interaction effect on formalin responding (F1,173 = 6.5, P < 0.05). Inspection of Fig. 2, A–D reveals that this interaction was driven largely by a decrease in formalin licking produced by amiloride in castrated males. However, this effect was not reliable, as it was not replicated in a subsequent experiment (see also Fig. 4A).
In the late phase, we observed a significant main effect of drug (F1,178 = 20.2, P < 0.001) but also a significant three-way (sex × surgery × drug) interaction (F1,178 = 4.5, P < 0.05). Subsequent t-tests revealed a higher level of formalin responding in sham + vehicle females vs. sham + vehicle males (as was observed in the previous experiment in intact vehicle-treated females and males) that only just failed to reach significance (t42 = 2.1, P = 0.051). More importantly, t-tests revealed highly significant effects of amiloride in female (t40 = 3.3, P < 0.005), but not male (t44 = 1.0, P = 0.32), mice, precisely replicating the results of the first experiment. Neither castration nor ovariectomy significantly altered formalin responding per se, although there was a trend toward lower sensitivity in ovariectomized females relative to sham females (t39= 1.53, P = 0.13). However, in castrated males, amiloride reduced formalin responding compared with vehicle (t53 = 3.2, P < 0.001), producing an antinociceptive effect comparable to that of sham females (see Fig. 2E). Female ovariectomy also produced a “switch” of amiloride efficacy. That is, in ovariectomized females, amiloride no longer significantly reduced late-phase formalin responding (t41 = 1.5, P = 0.14) (Fig. 2E).
Effects of estrogen replacement in ovariectomized females.
A two-way ANOVA performed on early-phase data revealed a main effect of amiloride (F1,59 = 5.4, P < 0.05). The modest inhibition of early-phase responding by amiloride in this experiment (see Fig. 3, A and B) stands in contrast, however, to the results of the gonadectomy experiment, in which no such inhibition was observed in ovariectomized female mice (see Fig. 2C).
A two-way ANOVA performed on late-phase data revealed a significant effect of drug (F1,59 = 16.9, P < 0.001) and a trend toward a hormone × drug interaction (P = 0.20). As can be seen in Fig. 3B, estrogen was clearly successful in restoring the efficacy of amiloride, but in this experiment, amiloride also trended strongly toward producing a significant inhibition of formalin licking in ovariectomized females given vehicle replacement (P = 0.06).
Estrogen treatment to ovariectomized females significantly increased uterine weight equally across both drug conditions (main effect of hormone: F1,58 = 226.0, P < 0.001).
Effects of testosterone replacement in castrated males.
A two-way ANOVA performed on early-phase data revealed a main effect of hormone (F1,69 = 9.0, P < 0.005), with testosterone-replaced castrated males displaying greater early-phase formalin responding relative to their vehicle-treated counterparts (t71 = 3.0, P < 0.05) (see Fig. 4, A and B).
A two-way ANOVA performed on late-phase data revealed a significant effect of drug (F1,69 = 5.5, P < 0.05) but also a significant hormone × drug interaction (F1,69 = 4.1, P < 0.05). There was a trend toward lower formalin sensitivity in testosterone-replaced castrated mice receiving vehicle (t35 = 1.4, P = 0.17). As can be seen in Fig. 4, chronic testosterone treatment completely reversed the effect of gonadectomy on amiloride efficacy, such that amiloride reduced formalin responding during the late phase in vehicle-treated castrated males (precisely replicating the findings of the gonadectomy experiment), but not in testosterone-treated males.
Testosterone treatment to castrated males significantly increased seminal vesicle weight equally across both drug conditions (main effect of hormone: F1,69 = 344.9, P < 0.001).
The main finding of this study is a robust sex difference in the ability of amiloride to inhibit licking behavior in the late-phase formalin test. AD50s were estimated as 34 mg/kg and 136 mg/kg for the two sexes, respectively, a potency ratio of 4.0. However, we were unable to demonstrate amiloride antinociception in males at any dose < 70 mg/kg, a dose found to be toxic, suggesting that the sex difference observed may be qualitative as opposed to quantitative. We have some reason to believe as well that the sex difference is not specific to the formalin test or even to chemical/inflammatory nociception, since 30 mg/kg amiloride also increased von Frey fiber withdrawal thresholds in female but not male mice (data not shown).
We observed significantly greater formalin responses during the late phase in gonadally intact females relative to males, which is consistent with previous findings in the rodent literature from a number of different groups (1, 21, 25, 39). This quantitative sex difference has been attributed to the actions of gonadal steroid hormones, in which testosterone exerts a blunting effect on formalin pain, whereas estrogen heightens sensitivity through a reduction of pain inhibition mechanisms (2, 21, 22). The present study revealed a strong trend toward lower pain responding during the late phase in ovariectomized females relative to their sham-operated counterparts, consistent with the notion that estrogen is responsible for the sex difference in sensitivity to formalin. Castration has been shown to produce an increase in formalin responding in male rats (21, 22), but we did not observe any significant difference in sensitivity between castrated and sham-operated CD-1 males. Testosterone treatment to castrated males produced a significant increase in responding in the early phase, but a strong trend toward decreased responding in the late phase.
To our knowledge, only one study has ever tested amiloride against formalin test nociception (18). That study reported dose-dependent inhibition of formalin test licking behavior by amiloride injected systemically, supraspinally, or spinally in both the early and late phases with full efficacy against late-phase licking. Subjects were male Swiss mice of unreported origin. The only other studies to test amiloride in a behavioral nociceptive assay were performed by Sluka et al. (41), who reported dose-dependent amiloride inhibition of acid-induced mechanical hypersensitivity in male C57BL/6 mice, and by Dube et al. (15), who demonstrated dose-dependent amiloride reversal of both thermal and mechanical hypersensitivity produced by inflammation or skin incision, respectively, in male Sprague-Dawley rats.
Ours is, therefore, the first study to report a sex difference in the antinociceptive effects of amiloride in which drug treatment markedly and dose dependently attenuated late-phase responding in female, but not male, CD-1 mice, except at a toxic dose. Our data are obviously in contradiction with those of Ferreira and colleagues (18). Possible reasons for the discrepancy may include the more concentrated formalin solution used presently (5 vs. 2.5%), which may have rendered the amiloride dose less efficacious and/or genotypic differences between the two different outbred populations used (see Ref. 11). We have shown on a number of prior occasions that analgesic potency and efficacy are dependent on genetic factors (5, 38, 48, 49) and also that sex differences interact importantly with genotype, being observed in some strains but not others (24, 35; see also Ref. 31).
Gonadectomy reversed the sex difference in the present study, producing an unmasking of late-phase amiloride antinociception in males and attenuating amiloride’s efficacy in females. In interpreting these effects in Figs. 2–4, it is important to remember that the efficacy of amiloride to reduce formalin response is properly compared only to the analogous vehicle-treated group. This is because the baseline shifts from condition to condition, due to the often nonsignificant but nonetheless statistically influential effects of sex, gonadectomy, and hormone treatments.
Estrogen treatment reversed the effects of gonadectomy in ovariectomized females, such that the not-quite-significant (i.e., P = 0.14 and P = 0.06 by t-test in Figs. 2E and 3C) effect of 30 mg/kg amiloride after ovariectomy reverted to a strongly significant amiloride blockade (characteristic of intact females) after hormone replacement. These findings implicate estrogen in the modulation of amiloride’s antinociceptive effects. The possible involvement of progesterone, shown by us previously to play an important role in the “switching” of N-methyl-d-aspartate dependence of swim stress-induced and κ-opioid antinociception in male and female CD-1 mice (42), has yet to be evaluated.
Testosterone treatment reversed the effects of gonadectomy in castrated males such that hormone-replaced males reverted to the amiloride-insensitive status characteristic of intact males. These findings, therefore, implicate testosterone as well in the mediation of amiloride’s actions. A point that remains to be clarified is whether the robust effects of testosterone are due to conversion to dihydrotestosterone, aromatization to estrogen, or direct actions at androgen receptors.
In addition to the hormonal mediators involved, the site of the sex-specific effects of amiloride warrant further investigation. Both amiloride and the newly reported drug, A-317567, block ASIC-like currents in dorsal root ganglia. The common pharmacological actions of these two drugs reported by Dube et al. (15) implicate ASICs in the antinociceptive effects of amiloride. This hypothesis is perhaps contradicted by our findings using ASIC3 dominant-negative mutants, in which blockade of all ASIC-like transient currents produced increases in sensitivity on the formalin test (among other assays) in males but not females (9). The difference between the direction of the sex difference in these two cases does not appear to be because of genetic background issues, since amiloride given to FVB/J mice (the background strain of the ASIC mutants) also produces antinociception in females but not males (data not shown). It is possible that the transgenic data actually reflect compensatory mechanisms, and not the direct impact of the absence of ASIC-like currents in these animals.
Thus it is perfectly conceivable that the sex-specific effects of amiloride on nociception shown here are not related to the drug’s known effects on ASIC channel activity and that we have uncovered sex differences in another neurochemical system. Amiloride is known to act on other pharmacological target sites implicated in pain processing, such as biosynthetic pathways for proinflammatory cytokines (23), the Na+/H+ exchanger (NHE) (30), ATP-sensitive K+ channels (7), the adenosine A1 receptor (20), and the GABA-A receptor complex (19). None of these systems have ever been directly implicated in the mediation of sex differences in formalin nociception, but in many cases, there is evidence of their regulation by gonadal hormones (4, 13, 17, 27). We have begun to investigate these possibilities one by one. Preliminary experiments using ethylisopropyl-amiloride (3 and 10 mg/kg ip), a high-potency NHE-blocking compound (30), do not reveal efficacy against late-phase formalin nociception in either sex.
Whatever the mechanism of the sex differences in amiloride efficacy shown here turns out to be, this finding represents one of the first demonstrations of sex differences in pain processing not obviously related to opioid modulation. We believe that further study will eventually reveal sex differences in pain mechanisms to be pervasive and evident at all levels of the neuraxis. This finding is also representative of a class of recent pain-related discoveries that are only made possible by the simultaneous testing of both sexes, still a strategy only rarely used in the field (34).
This research was supported by a Louise Edwards Foundation Award in Pain Research (to J. S. Mogil).
We thank Dr. Jim Pfaus for his generous contributions.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society