Progesterone Receptor Activation Regulates Sensory Sensitivity and Migraine Susceptibility

Abstract

Women develop chronic pain during their reproductive years more often than men, and estrogen and progesterone regulate this susceptibility. We tested whether brain progesterone receptor (PR) signaling regulates pain susceptibility. During the estrous cycle, animals were more sensitive to mechanical stimulus during the estrus stage than in the diestrus stage, suggesting a role for reproductive hormones, estrogen, and progesterone. Progesterone treatment of ovariectomized and estrogen-primed mice caused a delayed reduction in the mechanical threshold. Segesterone, a specific agonist of PRs replicated this effect, whereas, the segesterone-induced reduction in mechanical threshold was blocked in the mice lacking PRs in the nervous system. Segesterone treatment also did not alter mechanical threshold in adult male and juvenile female mice. PR activation increased the cold sensitivity but did not affect the heat and light sensitivity. We evaluated whether PR activation altered experimental migraine. Segesterone and nitroglycerin when administered sequentially, reduced the pain threshold but not when given separately. PRs were expressed in several components of the migraine ascending pain pathway, and their deletion blocked the painful effects of nitroglycerin. PR activation also increased the number of active neurons in the components of the migraine ascending pain pathway. These studies have uncovered a pain-regulating function of PRs. Targeting PRs may provide a novel therapeutic avenue to treat chronic pain and migraine in women.

Female reproductive hormones regulate pain susceptibility and render women more susceptible to chronic pain conditions, including migraine, tension-type headaches, neuropathic pain, and temporomandibular pain.^1–7 Women also have a lower threshold for experimental pain than men.^1,8 Increased pain susceptibility in women often emerges following puberty, which marks cyclic hormonal fluctuations. Furthermore, the pain symptoms frequently worsen during the perimenstrual phase when hormone levels are low^9–12 and improve during pregnancy when hormone levels are high.^13,14

Progesterone exerts acute analgesic actions. Its administration a few hours before testing increased the pain threshold in animals with neuropathic pain or inflammatory allodynia.^15–17 Progesterone metabolite allopregnanolone likely exerts these analgesic actions by potentiating GABAergic inhibition,¹⁸ since agents that potentiate GABAergic inhibition are also potent analgesics (reviewed in ref.¹⁹). However, the pain-suppressing actions of progesterone are not uniform. It did not alleviate pain induced by sciatic nerve cuffing²⁰ and was also ineffective in suppressing perimenstrual migraine headaches in a small number of patients.^21,22 Progesterone also has other neuronal targets, including progesterone receptors (PRs).^23,24 PR activation increases neuronal excitability, whereas allopregnanolone lowers it.²³ Also, unlike the rapid actions of allopregnanolone, PR activation induces gene expression with effects emerging in several hours to days.²³ This complexity could explain the discrepant results of progesterone treatment in treating chronic pain conditions. However, unlike allopregnanolone, the pain regulation by PRs remains poorly understood.

PRs are ligand-activated transcription factors, and their regulated effects emerge slowly but last longer.^24,25 We have described PR activation’s slow, delayed excitatory effect.^26,27 Since the pain paroxysms are thought to occur due to increased firing of the neurons in the pain pathway or because of a lowering of their firing threshold, the excitatory action of PR activation may be critical for the pain pathophysiology.

We performed this study to evaluate the role of PRs in regulating sensory sensitivity. We found that PR activation increased sensitivity to mechanical and thermal stimuli, and primed animals to the pain-promoting effects of nitroglycerin (NTG).

Methods

Materials

Segesterone, a progesterone receptor agonist sold under the brand name Nestorone, progesterone, and allopregnanolone were purchased from Sigma-Aldrich. Nitroglycerin was obtained from American Reagent. All the other common chemicals were also obtained from Sigma-Aldrich.

Animals

Animals were handled according to a protocol approved by the University of Virginia Animal Care and Use Committee, compliant with the ARRIVE guidelines. All the animals had ad libitum access to food and water and were maintained on a 12-hour light and 12-hour dark cycle (lights on at 6 AM, lights off at 6 PM). The experiments were performed with adult female (50–70 day-old) C57Bl6 mice or female mice lacking progesterone receptor expression. Mice with a floxed first exon of Pgr (PR^fl/fl) were a kind gift from Dr. M. Luisa Iruela-Arispe (University of California, Los Angeles, CA).^26,28 We and others have characterized these mice before.^26,29,30 These mice were crossed with nestin-Cre mice (B6. Cg-Tg(Nes-Cre)1Kln/J, The Jackson Laboratory #003771) to generate animals lacking PR expression in the neurons and glia in the central and peripheral nervous system (progesterone receptor knockout, PRKO). The colony was maintained by breeding the PR^fl/fl-cre^+ve male mice with PR^fl/fl-Cre^−ve females. Throughout the manuscript, we will refer to PR^fl/fl-cre^+ve mice as PRKO and PR^fl/fl-Cre^−ve as wild-type (WT) mice.

Double transgenic TRAP2 mice,³¹ which express an estrogen receptor (ER)-tagged Cre recombinase under the control of immediate early gene cFos promoter and lox-P flanked stop codon that prevents CAG-regulated tdTomato expression from the Rosa26 locus in the absence of Cre recombinase were used to evaluate whether PR activation altered the activity of migraine pain pathway.

Mice generated by crossing the mice expressing a cre recombinase under the control of Pgr promoter (Pgr-cre, B6.129S(Cg)-Pgrtm1.1(cre)Shah/AndJ)^32,33 with the Ai9 mice (B6. Cg-Gt(ROSA)26Sortm9(CAG-tdTomato) Hze/J) were also used.

Supplementary Fig 1 summarizes all the sensory sensitivity testing experiments.

Estrous Cycle Monitoring

Vaginal smears of adult female C57Bl6 mice were obtained daily between 8 AM and 10 AM. The cytology of the wet smears was checked under a light microscope to determine the percentage of nucleated cells, non-nucleated cornified cells, or small round leukocytes. The animals in estrus phase (smear contained primarily cornified cells) and diestrus phase (presence of mostly leukocytes in the smear) were used for behavioral testing and blood collection for hormone measurements (Fig 1A).

Figure 1.

Mechanical sensitivity during estrous cycle and following progesterone treatment. (A) Representative images showing vaginal cytology of estrus and diestrus stages, the estrus stage was characterized by the abundance of cornified cells whereas the leukocytes were plenty during diestrus stage. (B) Responses to the presentation of von Frey filaments in ascending order in a representative animal tested in estrus and diestrus phases; + represents a positive response and – represents no response. The red arrow marks the mechanical threshold. (C and D) Mean and SD of the circulating progesterone and estradiol levels respectively in animals in estrus and diestrus stages (n = 4), The p-values show results of the student’s t-test. (E) The mechanical threshold in animals in estrus and diestrus stages. The values represent mean ± SD of the Von Frey filament weight that produced a response on both the hind paws at least 50% of the times, n = 8, P-value represents the results of Wilcoxin matched-pairs signed-rank test. The same set of animals was tested during the 2 phases. (F) A schematic showing treatment and testing schedule. Adult C57Bl6 female mice were ovariectomized and primed with estrogen (17β-estradiol 10 μg, sc). Two days later the animals received progesterone (10 mg/kg, sc) or vehicle (20% β-hydroxycyclodextrin). The same group of animals was tested daily. (G) Mean and SD of the mechanical threshold in progesterone- or vehicle-treated mice, n = 6 each, the P-values of Šídák’s multiple comparisons test are shown on the graph. (H) Effect of allopregnanolone (THP) on mechanical threshold, n = 9 each.

The serum progesterone levels (n = 4) were measured using an ELISA assay (# K025-H1/H5 Arbor Assays, Ann Arbor, MI; detection range 50–3,200 pg/mL). The intra-assay variation was 7% in these studies. The estradiol levels (n = 4) were also measured using an ELISA assay (#ES180S-100 Calbiotech, detection range 3–300 pg/mL). The intra-assay variation was 5.9% in these assays.

Ovariectomy and Hormone Treatment

Bilateral ovaries were removed under isofluorane anesthesia.²⁷ The experiments were performed after 10 to 12 days of recovery.

The animals were primed with 17β-estradiol (10 μg/animal, subcutaneous), and 2 days later with progesterone (10 mg/kg, subcutaneous, sc). The control animals received vehicle, 20% β-hydroxycyclodextrin. This progesterone treatment caused an acute increase in serum progesterone levels (40–80 ng/mL, n = 4), but the levels were back to baseline (3–5 ng/mL, n = 4) 16 hours after the injection. Another cohort of C57Bl6 females were treated with PR agonist segesterone (10 mg/kg, sc) or vehicle (20% β-hydroxycyclodextrin). A third cohort of C57Bl6 mice were treated with allopregnanolone (THP, 10 mg/kg, sc) or vehicle (dimethyl sufoxide, DMSO). PRKO females and littermate wild-type animals were also treated with vehicle or segesterone or progesterone.

Evaluation of Mechanical Threshold

Mechanical sensitivity was tested using manual Von Frey monofilaments.³⁴ Briefly, the animals were individually placed in Plexiglas boxes (5″ × 5″ × 7″) with a wire mesh floor. The testing was performed after 30 minutes of acclimatization. The midplantar surface of each of the hind paws was tested thrice with von Frey filaments presented in an ascending order (.4, .6, 1.0, 2.0, 4.0, 6.0, 8.0, 10.0 g). The filament was applied perpendicularly to the paw until the filament buckled slightly and held for 3 seconds. Each stimulation is separated from the previous by 10 seconds to prevent sensitization. A quick withdrawal of the paw and/or its licking marked a positive response. The filament that produced a positive response for 3 out of the six presentations was considered as the threshold (Fig 1B).

We also evaluated the periorbital mechanical threshold using the up-down method.³⁵ The animals were acclimatized to 4 Oz paper cups, and the von Frey filaments were presented to the periorbital area. Wiping of the face or moving away from the filament were considered positive responses.³⁶ The first filament tested was .4 g. In the absence of a response, the next heavier filament (up) in the series was presented, whereas the next lighter filament (down) in the series was presented in the presence of a response. This pattern was repeated 4 times after the first positive response.

Evaluation of Thermosensitivity

Cold sensitivity was measured by the latency to escape from a cold floor (15 °C). The animals are kept in a custom-built Plexiglas chamber (5 × 5 × 7 in.) with a peltier-controlled cooling plate forming half the floor of the chamber. The remaining half of the floor was elevated (2.5 cm) to form an escape platform. The animals were acclimatized to the chamber for 5 minutes on the day before the testing. On the day of the experiment, the animals were introduced to the chamber on the cooling plate side and the peltier turned on to cool the plate from room temperature (23–25 °C) to 15 °C over a period of 3 seconds. The time to jump away from the cold floor to the elevated platform was measured. To prevent injury to the animal, any animal that did not jump away in 30 seconds was manually removed and the latency marked as 30 seconds. Each animal was tested twice with 5 minutes interval during which the animal was returned to the home cage. Mean latency to climb onto the elevated platform was determined for each animal on each testing day.

Warm sensitivity was determined using the tail-flick test.³⁷ The animals were restrained in a cylindrical tube and 2 minutes later, 3 cm of the end of the tail was dipped in hot water maintained at 48 ± 1 °C. The latency to flick the tail was recorded, and as done for the cold sensitivity assay, the assay ended at 30 seconds. Each animal was tested twice with a 5 minutes interval between tests, and average latency was determined (n = 7).

Evaluation of Light Sensitivity

The light sensitivity was evaluated using a light-dark box, with a light (15 × 11 × 20 cm) compartment with transparent walls and floor, connected to a dark chamber of the same dimensions but with black walls through a 4 × 4 cm opening in the wall. The animals were introduced to the light compartment and allowed to explore both the compartments for 7 minutes. The movement of animals was video recorded and analyzed post hoc using AnyMaze software. The time in light and dark compartments and the latency to the first entry to the dark compartment were evaluated.

Nitroglycerin Treatment

Migraine-like central sensitization was evoked using NTG. A 5 mg/mL stock of NTG (in 30% alcohol, 30% propylene glycol, and water, American Reagent) was diluted 1:1 with saline and then injected into animals. NTG 5, 10, or 15 mg/kg or vehicle were intraperitoneally injected in experiments that measured paw withdrawal. Microliter volumes corresponding to 2× the body weight for 5 mg/kg dose and 6× the body weight for 15 mg/kg dose were injected. The control animals received injections of the vehicle diluted 1:1 with saline.

NTG doses of 10, 1, and 0.1 mg/kg were tested in the experiment evaluating the periorbital withdrawal threshold. For 1 mg/kg dose, a 1 mg/mL solution was prepared by diluting the 5 mg/mL stock solution with saline and microliters corresponding to the body weight in g were injected. For 0.1 mg/kg dose, the 1 mg/mL stock solution was further diluted 1:10, and microliters corresponding to the body weight in g were injected.

Evaluation of Active Neurons

TRAP2 female mice were ovariectomized and treated with 17β-estradiol (10 μg, subcutaneous) 10 to 12 days post ovariectomy. The animals were treated with vehicle or segesterone 2 days after the estrogen priming. To TRAP the neurons activated by segesterone, 4 hydroxy-tamoxifen (4OHT, 50 mg/kg, subcutaneous) in sesame seed oil (100 μl) was injected in the home cages 2 days after the treatment. Seven days later, the animals were transcardially perfused with 4% PFA.³³ The brain and trigeminal ganglia (TG) were isolated and post-fixed in the same solution overnight, and then kept in 30% sucrose for cryoprotection. Forty micron-thick coronal brain sections and horizontal TG sections were collected in 4 groups. Sections in 1 group were counterstained with 4’,6-diamidino-2-phenylindole (DAPI) and imaged using a Nikon Eclipse Ti-U microscope equipped with a Nikon confocal C2 scanner, DU3 High Sensitivity Detector System, and Nikon LUN4 4 Line Solid State Laser System under 10 × .45 numerical aperture lens with 512 × 512 frame size. Ten μm optical sections were acquired, and images were tiled and stitched using NIS-Elements software (Nikon). The images were processed and analyzed using Imaris 8.3.1 software (Bit-plane Scientific, Zurich, Switzerland). The regions of interest were marked using Paxinos Atlas (Paxinos G, 1986).

Open Field Activity

The mice were tested in an open field for their locomotory activity. The animals were handled for 5 minutes on 3 days prior to the exposure to the arena (45 × 45 × 40 cm, Harvard Apparatus). The animal’s activity in the arena was tracked for 5 minutes using a video recorder, and the total distance traveled during 5 minutes and the speed was determined using AnyMaze software (Stoelting).

Real-Time PCR

PR mRNA expression was quantified from the TG, microdissected trigeminal nucleus caudalis (TNC), hypothalamus, and combined primary and secondary somatosensory, anterior cingulate, and insular cortices using a qRT-PCR assay as described in our prior study.³³ The GAPDH, β-actin, and HPRT expression was used as a control. The primers used in the real-time PCR assay are listed in Supplementary Table 1.

Statistical Analysis

Graphpad Prism 9 was used to perform statistical comparisons. The data are plotted as mean ± standard deviation. Wilcoxon matched-paired sign rank test was used to evaluate differences in mechanical threshold during the estrous cycle. Student’s t-test and repeated measures two-way ANOVA or ordinary one-way ANOVA with post hoc Šídák’s multiple comparisons test or Dunnet’s multiple comparison test were used to evaluate whether the treated and untreated animals differed from each other. The differences were considered significant when the P-value was less than .05.

Results

Pain Susceptibility Fluctuates During the Estrus Cycle

Women experience elevated pain sensitivity during the perimenstrual period. We assessed mechanical sensitivity using manual Von Frey monofilaments (Fig 1B) during 2 estrous cycle phases determined by vaginal cytology (Fig 1A). The progesterone and estrogen levels were higher in the diestrus than in the estrus phase (Fig 1C, 1D, n = 4, t(6) = 3.079, P = .02) for progesterone and t(6) = 3.081, P = .022 for estrogen, Student’s t-test). We tested 5 animals in the estrus stage first and subsequently in the diestrus stage, and 3 animals in reverse order. Animals in the estrus phase responded to a smaller force than those in the diestrus phase (Fig 1B, 1E, n = 8, P = .016, Wilcoxin matched-pairs signed-rank test). Thus, similar to the heightened perimenstrual pain susceptibility, the mechanical threshold was lower during the low hormone level estrus phase. The lower threshold during the estrus stage could be due to the slow progesterone effect observed on its withdrawal or estrogen withdrawal. We focused on the PR-mediated slow impact of progesterone.

Lasting Effects of Progesterone Lowered Mechanical Threshold

The animals were ovariectomized and primed with 17β-stradiol to avoid the confounding effects of endogenous hormonal fluctuations (Fig 1F). The animals received a single injection of progesterone or vehicle 2 days after estradiol treatment and subsequently tested on 4 days (Fig 1F, n = 6 each). The thresholds seemed comparable a day after progesterone or vehicle treatment; however, on subsequent days, the threshold in the progesterone treatment group appeared lower than that in the vehicle treatment group (Fig 1G). Repeated measures of two-way ANOVA uncovered a main effect of duration and treatment (F (3, 30) = 3.115, P = .041). Post hoc Šídák’s multiple comparisons test revealed differences between the vehicle and progesterone treatment groups on days 2, 3, and 4 (P = .03, .03, and .02, respectively).

Thus, progesterone exerted a slow-emerging reduction in the mechanical threshold, possibly due to allopregnanolone withdrawal or PR activation. Allopregnanolone is a progesterone metabolite with analgesic actions,^38,39 and because of rapid metabolism, its levels closely follow those of progesterone.^40,41 Thus allopregnanolone withdrawal could lower the mechanical threshold on days 2 to 4 following progesterone treatment. Alternately progesterone could activate PR-regulated cellular signaling with slow-onset but lasting effects. We evaluated whether allopregnanolone replicated the effects of progesterone on the mechanical threshold.

Allopregnanolone Withdrawal and Pain Susceptibility

Ovariectomized and estrogen-primed animals were treated with allopregnanolone (10 mg/kg) or vehicle; the acute effects were evaluated an hour after its administration and the lasting effects were evaluated for 4 subsequent days (Fig 1H, n = 9 each). A main effect of duration and treatment was observed (F (4, 64) = 3.866, P = .0071 repeated measures two-way ANOVA); however, post hoc comparisons did not uncover significant differences between the 2 groups at any time point. Thus, allopregnanolone withdrawal did not explain the delayed pain-promoting effect of progesterone.

PRs and Regulation of Sensory Sensitivity

We found that PR-mediated effects emerge 24 to 48 hours after progesterone administration, consistent with the genomic action of these receptors.^26,27 If PRs regulated the reduction in mechanical threshold in progesterone-treated animals, their deletion should block its effects, and their activation should reduce the mechanical threshold. We first evaluated the effect of progesterone treatment in the mice lacking PR expression in the nervous system (PRKO, PR^fl/fl-cre+ve) and used littermate wild-type (WT, PR^fl/fl-Cre-ve) mice as controls. The pre-progesterone mechanical threshold in the WT and PRKO mice was similar (n = 7 each, t(12) = .000, P > .99, student’s t-test), indicating that when the endogenous hormonal fluctuations were eliminated, the 2 groups were comparable.

Comparison of the mechanical threshold following progesterone treatment with that before uncovered an effect of the treatment duration (n = 7 each, F (3.128, 37.54) = 4.608, P = .0070, repeated measures two-way ANOVA). Post hoc comparisons revealed significant differences in the WT mice on days 3 and 4 (P = .0098 and P = .0097, respectively Bonferroni’s multiple comparisons test) (Fig 2A). A similar difference was not observed in the PRKO mice (P = .8803 and P > .9999 respectively). Post hoc comparisons also revealed significant differences in the mechanical threshold of WT and PRKO mice on day 4 of the testing (P = .0315).

Figure 2.

PR activation and mechanical threshold in female mice. (A) Mechanical threshold in the PRKO and littermate WT mice treated with progesterone, n = 7 each, the P-value of uncorrected Fisher’s LSD showing significant difference in vehicle and progesterone-treated mice on day 4 is shown. (B) A schematic showing segesterone or vehicle treatment of the animals. (C) The mechanical threshold in the ovariectomized and estrogen-primed C57 females treated with vehicle (n = 9) or segesterone (10 mg/kg, n = 10). The P values represent post hoc Šídák’s multiple comparisons test between vehicle and segesterone-treated animals. (D) The mechanical threshold in cycling animals treated with vehicle (n = 7) or segesterone (n = 7). The P value shows a mechanical threshold comparison of the vehicle and segesterone-treated mice (Šídák’s multiple comparisons test). (E) The mechanical threshold in the ovariectomized and estrogen-primed WT and PRKO mice treated with segesterone (n = 7 WT and 8 PRKO). The P-value shows the comparison of the WT and PRKO mice on day 1 of the testing (Šídák’s multiple comparisons test).

We used segesterone, a specific PR agonist, to confirm further the role of PRs in regulating pain sensitivity. We first tested the effect of segesterone in the ovariectomized and estrogen-primed C57Bl6 mice (Fig 2B, n = 9 vehicle and 10 segesterone-treated), and the mechanical threshold was evaluated one and 2 days later. The main effect of treatment was observed (F (1, 17) = 23.52, P = .0002, repeated measures two-way ANOVA) (Fig 2C), and post hoc comparisons revealed significant differences in the mechanical threshold between the vehicle and segesterone-treated mice on both the days of testing (P = .0007 and P = .0034 respectively).

We also treated intact cycling C57Bl6 mice with segesterone because ovariectomy has short and long-term effects, which could confound the PR activation effects described above.⁴² Animals in the estrus stage were injected and tested for 2 days (n = 7 each, Fig 2D). An effect of the treatment (F (1, 12) = 14.36, P = 0.0026, repeated measures two-way ANOVA) was observed and post hoc comparisons uncovered a significant difference in the mechanical threshold between the vehicle and segesterone-treated animals on day 1 of the testing (P = .0006).

Then we tested the effect of segesterone in the WT and PRKO mice (Fig 2E). Since the PRKO mice have irregular estrous cycles (data not shown), the animals were ovariectomized and estrogen-primed (n = 7 WT and 8 PRKO). The main effect of genotypes was observed (F (1, 13) = 4.700, P = .0493, repeated measures two-way ANOVA) and post hoc comparisons showed a significant difference in the mechanical threshold between the WT and PRKO mice on day 1 of the testing (P = .0293). These studies together revealed that PRs regulated sensory sensitivity.

PR Activation and Thermal Sensitivity

PR activation lowered heat but not cold sensitivity in female mice. We compared the sensitivity to warmer (48 °C) and colder (15 °C) temperatures in the segesterone and vehicle-treated mice to evaluate whether PR activation also regulated thermal nociception. We measured the latency to escape the cold floor. The vehicle and segesterone-treated mice were distinct in their sensitivity to the cold floor (Fig 3A, n = 7 each). The main effect of treatment was observed (F (1, 12) = 11.45, P = .0054, repeated measures two-way ANOVA). Post hoc comparisons revealed a significant difference in the escape latency between the vehicle- and segesterone-treated mice on the 2nd day of testing (P = .0064).

Figure 3.

The effect of PR activation on thermal and light sensitivity in female mice. (A) Latency to escape (mean ± SD) the cold floor in mice treated with vehicle or segesterone, n = 7 each, P-value of the post hoc Šídák’s multiple comparisons test is shown. (B) The tail-flick latency in the vehicle- and segesterone -treated mice, n = 7 each. (C) The time spent in the light (L) and dark (D) chambers respectively after treatment with segesterone (10 mg/kg) or vehicle, n = 6 each. (D) The number of transitions between the light and dark compartments. (E) Latency to the first entry into the dark chamber.

The animals had to climb on the elevated platform to escape the cold floor requiring motor function, which may have been altered by segesterone treatment. To obtain insights into the potential effects of segesterone treatment on motor performance, we evaluated animals’ behavior in the open field assay (n = 6 each, Supplementary Fig 2A, 2B). The vehicle and progesterone-treated mice walked comparable distances in the arena, and the speed was also similar (distance covered: t(10) = .7537, P = .4684 and speed: t(10) = .6788, P = .5126, student’s t-test).

We used a tail-flick assay to measure warmer temperature sensitivity following segesterone treatment (Fig 3B, n = 7 each). No treatment effect on the tail-flick latency was observed in this assay (F (1, 12) = 3.750, P = .0767, repeated measures two-way ANOVA). Thus, PR activation increased sensitivity to cold but did not alter the sensitivity to warmth.

PR Activation and Light Sensitivity

Migraines and tension-type headaches are often associated with photophobia.⁴³ Therefore, to extend the findings that uncovered the role of PRs in regulating tactile and thermal sensitivity, we also evaluated if PR agonist segesterone altered light sensitivity. The animals’ behavior in the light-dark box was tested after the administration of segesterone or vehicle (n = 6 each). Increased light sensitivity following segesterone treatment could lead to increased time in the dark compartment, fewer transitions between the 2 compartments, and/or a shorter latency to enter the dark compartment. Thus, we compared these parameters in the vehicle- and segesterone-treated animals. The time spent in the dark compartment did not differ between the segesterone or vehicle-treated mice (Fig 3C, F (1, 10) = 1.114, P = .3161, repeated measures two-way ANOVA). Furthermore, neither the number of transitions between the 2 compartments (Fig 3D; F (1, 10) = 2.568, P = .1401, repeated measures two-way ANOVA) nor the latency to enter the dark compartment changed (Fig 3E, F (1, 10) = .001729, P = .9677 repeated measures two-way ANOVA). Thus, PR activation did not affect light sensitivity. Since the PR agonist segesterone exerted only a modest effect on thermo and light sensitivity, we used only the mechanical threshold measurements in the subsequent studies.

Sex- and Age-Specific Effects of PR Activation on Mechanical Threshold

Gender differences that disproportionately affect females exist for various pain conditions. To evaluate whether PR activation regulated mechanical threshold in males, we treated adult male mice with segesterone and evaluated the paw withdrawal threshold. The threshold seemed comparable between segesterone- and vehicle-treated males (Fig 4A, n = 7). An overall effect of the treatment was observed (F (1, 12) = 6.353, P = .0269, repeated measures two-way ANOVA), but post hoc comparisons did not reveal differences between the 2 treatment cohorts on either of the days (at 1 day P = .1106 and at 2 days P = .1106 Šídák’s multiple comparisons test). Thus, PR activation did not lower the mechanical threshold in adult males.

Figure 4.

The effect of PR activation on mechanical sensitivity in males and juvenile females. (A) The mechanical threshold in adult male mice treated with vehicle or segesterone, n = 7 each. (B) The mechanical threshold in juvenile, pre-pubertal female mice (23 day-old) treated with vehicle (n = 8) or segesterone (n = 7).

Puberty marks the onset of cyclic hormonal changes in females, accompanied by the emergence of sex differences in the prevalence of clinical pain. We evaluated whether the onset of puberty increased the sensitivity to the effects of PR activation. The vaginal opening occurred in 30 to 32-day-old C57Bl6 females. Hence, we evaluated the effect of segesterone in 23 to 26-day-old females, before the onset of puberty (n = 8 vehicle and 7 segesterone). Overall, the mechanical threshold was lower in the juvenile animals compared to that in the gonadally-intact adult females (t(15) = 7.095, P < .0001, student’s t-test). However, the threshold did not differ between the vehicle- and segesterone-treated mice, and it also remained stable over the 2 days of testing (Fig 4B, F (1, 13) = .1454, P = .7091, repeated measures two-way ANOVA). Together these studies showed that PR signaling was not associated with sensory sensitivity regulation in adult males and juvenile females.

Nitroglycerin (NTG) Reduced the Mechanical Threshold

The regulation of sensory sensitivity by PR activation raised the possibility that PR signaling may also regulate pathological pain. Therefore, in the subsequent studies, we evaluated whether PRs regulated migraine-accompanying sensory hypersensitivity. NTG has been extensively used to induce migraine-like alterations in experimental animals and it reduces mechanical threshold.^44–48 Ovariectomized and estrogen-primed mice were treated with 3 doses of NTG (5, 10, 15 mg/kg, n = 7 each) and the mechanical threshold was evaluated one and 4 hours later. An overall effect of the treatments was observed (Fig 5A, F (3, 24) = 7.839, P = .0008, repeated measures two-way ANOVA). Post hoc comparisons revealed that 10 mg/kg and 15 mg/kg doses of NTG reduced the mechanical threshold at both the time points compared to the threshold in vehicle-treated mice at the respective time points (at 1 hour: P = .0017 vehicle vs 10 mg/kg NTG, P = .0023 vehicle vs 15 mg/kg NTG; at 4 hours: P = .0036 vehicle vs 10 mg/kg NTG and P = .018 vehicle vs 15 mg/kg NTG, Šídák’s multiple comparisons test).

Figure 5.

PR activation primed female mice for the nitroglycerin-induced reduction in mechanical threshold. (A) The mechanical threshold in mice treated with NTG at 5 mg/kg, 10 mg/kg, or 15 mg/kg or with vehicle. The testing was done at 1 hour and 4 hours after the injections, n = 7 each, the P-values of the post hoc Sidak’s multiple comparisons are shown. (B) The mechanical threshold in animals that received a low dose of segesterone (3 mg/kg) followed by a low dose of NTG (5 mg/kg). The threshold was determined before, and 1 and 4 hours after NTG administration, n = 7, and the P-values of the post hoc Šídák’s multiple comparisons test are shown.

Subsequently, we evaluated the effect of segesterone pretreatment on NTG-induced pain. The animals were treated with segesterone (10 mg/kg) and a day later received NTG. However, the mechanical threshold in the segesterone+NTG-treated mice (n = 6) was similar to that in segesterone+vehicle-treated mice (n = 5) (Supplementary Fig 2C, F (3, 19) = 9.142, P = .006 repeated measures two-way ANOVA).

Synergistic Action of Segesterone and NTG on Paw Withdrawal Threshold

Since segesterone treatment itself reduced the mechanical threshold, it may not have caused a further reduction in the mechanical threshold following NTG administration. However, if segesterone and NTG exert a synergistic action, lower doses of these agents, at concentrations that individually don’t alter the mechanical threshold, could lower the mechanical threshold when administered together. To test, this we first evaluated the effect of 3 mg/kg segesterone on the mechanical threshold. This dose did not impact the threshold (4.29 ± 0.76, n = 7), which was similar to that in animals treated with vehicle (3.83 ± 1.9, n = 7, t(12) = 0.5956, P = .5625, student’s t-test).

We then evaluated the effect of low dose NTG (5 mg/kg) administration on the animals treated with low-dose (3 mg/kg) segesterone. NTG was administered one day after segesterone. NTG reduced the mechanical threshold in these animals (Fig 5B, n = 7, F(2, 18) = 56.82, P < .0001, repeated measures one-way ANOVA). The posthoc comparison revealed that the mechanical threshold 1 and 4 hours after NTG administration was lower than that measured before its injection (P < .0001 at both the time points, Šídák’s multiple comparisons test). Thus, a combination of low-dose segesterone and NTG exerted a synergistic action to reduce the mechanical threshold.

Synergistic Action of Segesterone and NTG on Periorbital Mechanical Threshold

We also confirmed the synergistic action of segesterone on NTG-induced migraine-like pain by evaluating the periorbital pain. We first evaluated the periorbital withdrawal threshold in ovariectomized and estrogen-primed animals treated with different doses of segesterone (10, 3, or 1 mg/kg, n = 9 each for 10 and 3 mg/kg doses and n = 7 for 1 mg/kg dose) or vehicle (n = 9). An overall effect of the treatment and time was observed (Fig 6A, F (7, 60) = 7.175, P < .0001, repeated measures ANOVA). Post hoc Šídák’s multiple comparisons test revealed a significantly lower withdrawal threshold on both the testing days in animals that received 10 (day 1 P = .045, day 2 P = .0002) and 3 mg/kg (day 1 P = .003, day 2 P = .0001) segesterone compared to vehicle-treated animals.

Figure 6.

PRs regulated periorbital pain following NTG administration to female mice. (A) The effect of different doses of segesterone (10, 3, 1 mg/kg, sc) on the periorbital mechanical threshold, n = 9 each for vehicle and 10 and 3 mg/kg segesterone, and n = 7 for 1 mg/kg segesterone. The P values of post hoc Šídák’s multiple comparisons test are displayed on the graph. (B) The effect of different doses of NTG (10, 1, and 0.1 mg/kg) on periorbital mechanical threshold measured 2 hours later, n = 8 for 10 mg/kg and 7 each for 1 and 0.1 mg/kg, and n = 5 for vehicle. The P values of post hoc Dunnet’s multiple comparison test are shown. (C) The effect of the vehicle and NTG (0.1 mg/kg) on the periorbital mechanical threshold in the mice primed with segesterone (1 mg/kg). The P value of student’s t-test comparison between vehicle and NTG-treated mice is shown on the graph. (D) The effect of NTG (10 mg/kg) on periorbital pain in the WT and PRKO mice. The periorbital mechanical threshold was evaluated in gonadally-intact female WT (n = 710) and PRKO (n = 9) mice before and 2 hours after administration of NTG. The P values of post hoc Šídák’s multiple comparisons test are shown on the graph.

Then we treated ovariectomized and estrogen-primed animals with varying doses of NTG (10, 1, or 0.1 mg/kg, n = 8, 7, and 7 respectively) to determine an ineffective dose in lowering the withdrawal threshold. Vehicle-treated animals (n = 5) were used as controls, and the thresholds were assessed 2 hours after the injections (Fig 6B). An overall effect of the treatment was observed (F (3, 23) = 8.442, P = .0006, ordinary one-way ANOVA). Post hoc analyses uncovered significant differences in the withdrawal threshold between vehicle and NTG 10 mg/kg and NTG 1 mg/kg animal groups (P = .0002 and P = .019, Dunnett’s multiple comparisons test).

After identifying the doses of segesterone and NTG ineffective in altering the mechanical threshold on their own, we treated animals with segesterone (1 mg/kg) and 2 days later with NTG (0.1 mg/kg, n = 7) or vehicle for NTG (n = 6). The withdrawal threshold was evaluated before and 2 hours later (Fig 6C). The mechanical threshold in the segesterone+NTG-treated mice was substantially lower than that in segesterone+vehicle-treated animals (t(11) = 6.303, P < .0001, student’s t-test). Thus, PR activation also primed the animals to the NTG-induced reduction in periorbital pain. Together these studies showed a similar effect of PR activation on cephalic and peripheral nociception.

PRKO Mice Were Resistant to the Effects of NTG

The pain-promoting effect of PR agonist raised the possibility that PRKO mice may be resistant to migraine. This we evaluated by treating the PRKO and WT mice with NTG (10 mg/kg, n = 7 WT and 6 KO). As seen for the paw withdrawal threshold, the periorobital mechanical threshold was also comparable between the WT and PRKO mice before NTG administration (P = .9935). However, the mechanical threshold 2 hours after NTG administration was substantially lower in the WT mice than that in the PRKO mice (Fig 6D, row and column effect F (1, 17) = 16.11, P = .0009 repeated measures two-way ANOVA, post hoc Šídák’s multiple comparisons test P < .0001 WT pre vs post and P = .0012 WT-post vs PRKO-post-NTG). Thus, the PRKO mice were resistant to the effects of NTG.

PR Agonist Segesterone Increased the Number of Active Neurons in the Ascending Pain Pathway

Our prior studies have found that PR activation increased the number of active neurons in the hippocampus and entorhinal cortex and their down-stream targets.³³ A lowering of the mechanical threshold in segesterone-treated animals raised the possibility that the number of active neurons in the components of the pain matrix may also be altered in them. We evaluated whether this was the case using TRAP2 mice. Ovariectomized and estrogen-primed mice were treated with vehicle or segesterone. Four-hydroxytamoxifen (4OHT) was administered 2 days later, when a prominent behavioral effect of the treatment was observed. Since the neurons active during 60 to 90 minutes prior to 4OHT administration express tdTomato,^49,50 this protocol labeled the neurons active on day 2 of the treatment.

We focused on the 1st and 2nd order neurons of TG and TNC, and somatosensory, anterior cingulate, and insular cortices. The labeled neurons were rare in the TG and sparse in the TNC of vehicle-treated animals, whereas, several tdTomato-positive neurons were present in the TG and TNC of segesterone-treated mice (Fig 7A–D). Similarly, while the labeled neurons were present in the somatosensory, insular, and anterior cingulate cortices of vehicle-treated mice, they appeared denser in the segesterone-treated animals (Fig 7E–L). More labeled neurons were present in the deeper layers of the anterior cingulate and somatosensory cortices of segesterone-treated animals compared to those in the vehicle-treated mice. On the other hand, labeled neurons in the insular cortex were spread across all the layers in segesterone-treated animals. We quantified the number of labeled neurons in these regions (Fig 7M, n = 4 each for TG and n = 5 each for the other regions). It confirmed an increase in the number of active neurons in TNC (t(8) = 2.730, P = .026), insular (t(8) = 3.283, P = .011), anterior cingulate (n = 5, t(8) = 2.559, P = .034), and somatosensory cortices (n = 5, t(8) = 2.856, P = .021). On the other hand, the number of TRAPed neurons in the TG of segesterone-treated mice was similar to that in controls (t(6) = 1.284, P = .25).

Figure 7.

PR agonist segesterone increased the number of active neurons in the migraine ascending pain pathway of female mice. (A, B) Images from representative vehicle- and segesterone-treated TRAP2 mice respectively showing active neurons in the trigeminal ganglia. The red fluorescence of tdTomato corresponds to cFos-expressing neurons and DAPI is used as a counterstain. The scale bars in this and other images are in μm. (C, D) Images from the vehicle- and segesterone-treated TRAP2 mice show active neurons in the TNC. (E, F) Images of sections encompassing the somatosensory, anterior cingulate, and insular cortices from the vehicle- and segesterone-treated TRAP2 mice respectively. (G, H) The magnified boxed areas c1 and d1 in images C and D encompassing the anterior cingulate cortex. (I, J) The magnified boxed areas c2 and d2 from images C and D showing the active neurons in the somatosensory cortex. (K, L) The boxed regions c3 and d3 from images C and D show active neurons in the insular cortex. (M) The average number of tdTomato-positive active neurons per slice in the vehicle- and segesterone-treated mice, n = 4 each for TG and n = 5 each for TNC, insular, anterior cingulate, and somatosensory cortices. The P values of student’s t-test showing comparisons between the vehicle- and segesterone-treated animals are shown.

In contrast to the moderate to high presence of labeled neurons in the TNC and cortex, not many labeled neurons were present in the ventropostero medial and posterior thalamic nuclei which represent the 3rd order neurons in the migraine ascending pain pathway (Supplementary Fig 3).

Besides these regions, activated neurons were also present in the periaqueductal gray, amygdala, and prelimbic cortex of vehicle and segesterone-treated mice (Supplementary Fig 3) and their density tended to be more in the segesterone-treated animals. Thus, segesterone treatment led to increase in the number of active neurons in selected regions of the migraine pain network.

PR Expression in the Migraine Ascending Pain Pathway

We evaluated PR expression in the regions of the migraine ascending pain pathway using mice which expressed tdTomato reporter protein under the control of Pgr promoter (Fig 8A) (Pgr-Cre mice crossed to Ai9 mice, see methods for details). The PR-expressing neurons appeared concentrated in superficial layers of the anterior cingulate cortex (Fig 8B). On the other hand, the labeled neurons were present in the superficial and deeper layers of the somatosensory cortex (Fig 8C). The labeled neurons in the insular cortex were denser in the deeper layers (Fig 8D). Besides the cellular tdTomato expression, diffused tdTomato expression, which likely arose from tdTomato protein present in the dendrites and axons of PR-expressing, was also detected.

Figure 8.

PR expression in the migraine ascending pain pathway of female mice. (A) A section of the frontal cortex of a representative animal showing tdTomato expression in the PR-expressing neurons. DAPI was used for counterstaining. The scale bar shown is in μm. (B, C, D) magnified boxed areas from A encompassing anterior cingulate, somatosensory, and insular cortices. (E) A section of medulla showing PR-expressing neurons of the TNC. (F) magnified boxed region from E. (G) PR expression in the trigeminal ganglia. (H) The average number of tdTomato-labeled neurons in the regions of interest. The neurons were counted from 3 slices from each animal and presented as an average, n = 3 animals each. The scale for TNC, anterior cingulate, and insular cortices is shown on the left and that for the somatosensory cortex is shown on the right. (I) PR mRNA expression in the TG, TNC and combined somatosensory, cingular, and insular cortices evaluated using real-time PCR assay. The expression is shown relative to that in the hypothalamus (red line). the P values correspond to post hoc Dunnet’s multiple comparison test compared to the expression in the hypothalamus.

The neurons of the TNC also expressed tdTomato protein indicating the expression of PRs in them (Fig 8E, 8F). More labeled neurons were present in the superficial laminae of than those in the deeper laminae. On the other hand, tdTomato expression in the TG was sparse (Fig 8G). A similar expression pattern was seen in 3 animals and the average number of labeled neurons in the cortical regions and in TNC is presented in Fig 8H.

To evaluate quantitative differences in the PR expression in the cortex (combined somatosensory, insular, and anterior cingulate), TNC, and TG, we performed a qRT-PCR assay (Fig 8I). Hypothalamus, which is known to express PRs was used as a positive control. The PR mRNA expression in the cortices (n=4) was comparable to that in the hypothalamus. On the other hand, PR mRNA expression in the TNC was 73 ± 10% of that in the hypothalamus (n=5, P=.02 post-hoc Dunnett’s multiple comparison test), and that in the TG was less than half (Fig 8I, n=4, 40 ± 3%, P < .0001, one-way ANOVA with post-hoc Dunnett’s multiple comparison test with the expression in hypothalamus). Thus, PR expression was high in the cortices and moderate in the TNC.

Discussion

We have demonstrated a pain-promoting effect of PR activation in female mice. Progesterone and PR receptor agonist segesterone reduced mechanical threshold and this effect was blocked by the deletion of PRs in neurons and glia. Furthermore, PR activation also primed the animals to the painful effects of nitroglycerin.

Pain susceptibility or its perception varies across the menstrual cycle in women^9,51; it is higher during late luteal/early follicular phases when the hormone levels are low compared to that during the follicular or ovulatory phases when the hormone levels are high. We found a similar fluctuation in the mechanical stimulus sensitivity across the estrous cycle in mice; the response threshold was higher during diestrus phase compared to that during the estrus phase of the cycle. Other studies have found similar fluctuations in the susceptibility to thermal stimuli and visceromotor pain during the estrous cycle.^52–54

Several prior studies have evaluated the pain-modulating effects of progesterone, but no study had focused on the effects dependent on PR activation. Many of these studies reported a pain-suppressing action of progesterone^{15–17,55,56,57}; however, other studies found that progesterone treatment was ineffective in suppressing neuropathic pain.^20,58 The pain-promoting action of PR activation uncovered here may oppose the analgesic effects of progesterone metabolite allopregnanolone, which could make progesterone treatment ineffective in suppressing pain. Furthermore, Ungrad et al found that progesterone was robustly protective in males with neuropathic pain but had only a moderate effect in females. We found that PR agonist segesterone did not exert a pain-promoting effect in males and pre-pubertal females, raising the possibility that the expression of PRs and downstream signaling molecules could be distinct in adult females and adult males and juvenile females.

Although PR deletion prevented the lowering of the mechanical threshold following segesterone treatment and blocked the painful effects of nitroglycerin, PR deletion did not impact the basal mechanical threshold. Sensory sensitivity is regulated by complex signaling involving a balance between the inhibitory and excitatory neurotransmission, expression of pro- and anti-inflammatory molecules and peptides including calcitonin gene-related peptide, pituitary adenylate cyclase-activating peptide, and the function of Trp channels.^59–62 PR may regulate only some signaling molecules, so the PRKO and WT mice may not differ at baseline. Two prior studies have found that PR antagonists RU-486 and ICI 182,780 could protect from hypersensitivity associated with neuropathic pain. In 1 study, RU-486 treatment blocked mechanical allodynia and thermal hyperalgesia induced by spinal nerve injury.⁶³ Although this study correlated the improvement in pain to glucocorticoid receptors, RU-486 also blocks PR actions. In another study, ICI 182,780 an anti-PR and anti-estrogen receptor antagonist prevented neuropathic pain.⁶⁴ However, in both these studies the protective effects were noted soon after the antagonist administration. Thus, PR suppression also seems to have a pain-suppressing acute impact.

Besides regulating the mechanical threshold in healthy animals, we have also uncovered a potential role of PRs in regulating migraine-accompanying sensory hypersensitivity. The perimenstrual period is associated with a heightened susceptibility to migraine headaches, whereas some women experience headaches exclusively during menstruation.^65–67 The role of progesterone in regulating perimenstrual migraine is underexplored. In 1 clinical trial exogenous progesterone administration did not suppress migraine headaches in a limited number of patients.^21,22 Based on the findings presented here, we propose that the luteal rise in progesterone would activate PRs. However, the pain-promoting effect of this activation would remain suppressed as long as the strength of inhibitory neurotransmission is maintained by high levels of circulating progesterone. With the perimenstrual decline in progesterone levels, the inhibitory neurotransmission would weaken and the effects of PR activation would emerge. A single injection of progesterone caused an acute elevation in circulating progesterone levels which returned to the baseline levels after 16 hours. However, this single administration of progesterone caused a delayed lowering of the mechanical threshold. Thus, the PR-regulated effects could last even after the progesterone levels had declined.

Progesterone receptor gene polymorphism, which is predicted to affect the receptor function is seen in migraine patients. However, whether these mutations exacerbate or alleviate this condition is unclear.^68–73 The diverse study population and distinctions in the clinical characterization of patients could have influenced these results. The current study in experimental animals found that PR activation could exacerbate this disorder and blockade of PR signaling could alleviate it. Thus, PRs may provide a novel therapeutic target to treat migraine-accompanying sensory alterations.

Headaches are a common side effect of oral contraceptives in women, and the use of combined hormonal contraceptives may increase the risk of migraines.^74–77 On the other hand, retrospective studies have found reduction in the number of headache and migraine days in women taking progestin-only contraceptive pills.^78–80 Although this finding seems at odds with the role of PRs uncovered here, estrogens induce PR expression, whereas progesterone exerts a suppressive effect.^81,82 A potential reduction in the PR expression together with allopregnanolone-mediated analgesic effects in these women may improve headaches. In 1 study average pain tolerance was more in healthy women taking combined oral contraceptives compared to that in men or menstruating women.⁸³ This indicates that hormonal contraceptives may protect healthy women from pain. However, whether the contraceptives have a similar effect in conditions like migraine or chronic neuropathic pain in which the activity of the pain matrix is altered remains unexplored.

These studies were informed by the emerging studies elucidating the role of PRs in regulating neuronal excitability and glutamatergic neurotransmission. We have found that PR activation potentiates glutamatergic transmission through AMPA receptors and neuronal excitability.^26,27 The pain matrix is composed of sensory neurons of the dorsal horn, periaqueductal gray, parabrachial nucleus, thalamus, somatosensory, insular, cingulate, and prefrontal cortices, and amygdala.⁸⁴ Additionally, TG, spinal trigeminal caudal nucleus (TNC), and hypothalamus are also associated with migraine pain.⁶² A potential PR-regulated increase in excitability and/or strengthening of glutamatergic transmission of one or more components of the pain matrix could increase pain sensitivity. This regulation could happen directly or indirectly. PRs are expressed in the cortices, amygdala, and hypothalamus.^27,85–89 A prior study has also raised the possibility of PR expression in the TNC.⁹⁰ We also found moderate to high PR expression in these regions. Thus, the observed effects of segesterone could be a direct effect of the activation of PRs expressed in one or more components of the migraine pain network. Further in-depth characterization is necessary to identify critical nodes of the pain matrix that are under PR control. Additional mechanisms that may regulate pain downstream of the progesterone-PR signaling may involve anti- and pro-inflammatory molecules. Two prior studies have found that the anti-inflammatory actions of progesterone are blocked in the PRKO mice.^55,91 PRs also interact with synaptic proteins in the hypothalamus,⁹² and similar actions at synapses in the pain pathway may also be important.

We found an increased number of active neurons in the TNC, somatosensory, insular, and anterior cingulate cortices in the segesterone-treated animals using TRAP mice which provide a high signal-to-noise labeling of active neurons in a short time window (usually 60–90 minutes prior to the administration of 4-hydroxytamoxifen). In contrast, many cortical and TNC neurons seemed to express PRs marked by the tdTomato labeling. Thus, only a fraction of PR-expressing neurons seems to be activated by segesterone treatment. One limitation of Pgr-Cre X Ai9 mice is that the neurons are permanently marked once the stop codon preceding the tdTomato gene is removed. The real-time PCR assay showed the presence of PR mRNA expression, that matched the extent of neuronal tdTomato labeling, in the regions studied; however, additional studies are needed to evaluate PR expression in the principal neurons and interneurons and to determine what fraction of PR-expressing neurons are activated following segesterone treatment.

Conclusions

In conclusion, we demonstrate a novel pain-promoting effect of progesterone that is slow to emerge. This effect depends on PR expression. PRs also seem to regulate migraine-accompanying sensory hypersensitivity and could represent a potential target for the treatment of migraine and other chronic pain conditions that disproportionately affect women of reproductive age.

Perspective:

This article provides evidence for the role of progesterone receptors in regulating pain sensitivity and migraine susceptibility in females. Progesterone receptors may be a therapeutic target to treat chronic pain conditions more prevalent in women than men.