A prolactin-dependent sexually dimorphic mechanism of migraine chronification

Abstract

Objective

Determination of possible sex differences in mechanisms promoting migraine progression and the contribution of prolactin and the prolactin long (PRLR-L) and short (PRLR-S) receptor isoforms.

Background

The majority of patients with chronic migraine and medication overuse headache are female. Prolactin is present at higher levels in women and increases migraine. Prolactin signaling at the PRLR-S selectively sensitizes nociceptors in female rodents, while expression of the PRLR-L is protective.

Methods

Medication overuse headache was modeled by repeated sumatriptan administration in male and female mice. Periorbital and hindpaw cutaneous allodynia served as a surrogate of migraine-like pain. PRLR-L and PRLR-S isoforms were measured in the trigeminal ganglion with western blotting. Possible co-localization of PRLR with serotonin 5HT1B and 5HT1D receptors was determined with RNAscope. Cabergoline, a dopamine receptor agonist that inhibits circulating prolactin, was co-administered with sumatriptan. Nasal administration of CRISPR/Cas9 plasmid was used to edit expression of both PRLR isoforms.

Results

PRLR was co-localized with 5HT1B or 5HT1D receptors in the ophthalmic region of female trigeminal ganglion. A single injection of sumatriptan increased serum PRL levels in female mice. Repeated sumatriptan promoted cutaneous allodynia in both sexes but down-regulated trigeminal ganglion PRLR-L, without altering PRLR-S, only in females. Co-administration of sumatriptan with cabergoline prevented allodynia and down-regulation of PRLR-L only in females. CRISPR/Cas9 editing of both PRLR isoforms in the trigeminal ganglion prevented sumatriptan-induced periorbital allodynia in females.

Interpretation

We identified a sexually dimorphic mechanism of migraine chronification that involves down-regulation of PRLR-L and increased signaling of circulating prolactin at PRLR-S. These studies reveal a previously unrecognized neuroendocrine mechanism linking the hypothalamus to nociceptor sensitization that increases the risk of migraine pain in females and suggest opportunities for novel sex-specific therapies including gene editing through nasal delivery of CRISPR/Cas9 constructs.

Introduction

Sex differences in neurological disorders have long been recognized (1) but the underlying mechanisms that may be involved are often unclear. Migraine is a common neurological disorder (2) and cause of disability (3). For unknown reasons, migraine has an approximately 3:1 female prevalence (2). Patients with migraine are classified on the basis of attack frequency; episodic and chronic migraine are defined as less than, or equal/more than, 15 headache days per month, respectively (2). Many risk factors have been identified that can promote the progression from episodic migraine to chronic migraine, but chief among them are female sex, headache frequency, and the overuse of acute medications (4,5). The development of chronic migraine in those overusing acute medications and the presence of medication overuse headache (MOH) is also more common in women (3.5:1) (6,7).

Migraine has been conceptualized as a sensory threshold disorder where normally innocuous stimuli can promote pain (8). Prolactin (PRL) (9) is a neuroendocrine hormone that is found in both sexes but circulates at higher levels in women. PRL has been associated with migraine progression and chronic migraine (10,11). PRL is secreted from pituitary lactotrophs and is under tonic inhibitory control by tuberoinfundibular dopamine (TIDA) cells in the arcuate nucleus. PRL signals through homodimers of prolactin receptor isoforms called long (PRLR-L) and short (PRLR-S). Signaling at PRLR-L has been shown to promote gene transcription, while PRLR-S activation on nociceptors elicits pronociceptive actions through phosphorylation of pro-nociceptive ion channels; heterodimers of these isoforms inhibit cellular signaling (12–15). PRL sensitizes nociceptors selectively, but not exclusively, in females promoting increased evoked release of calcitonin gene-related peptide (CGRP) that may be integral in eliciting migraine in some people (16).

Preclinical assessment of migraine pain has relied on surrogate outcome measures including cutaneous allodynia (CA). CA is readily observed in rodents following activation of nociceptive afferents in the cranial meninges (17–19). CA is also observed in approximately 70% of patients during a migraine attack (20) and is an independent risk factor for migraine chronification (21). Migraine chronification can be modeled preclinically with frequent administration of acute medications that promote MOH in humans (18,22). In these rodent models, repeated administration of triptans, ditans, opioids, and cannabinoids elicit transient CA that can be measured in the periorbital region as well as in the hindpaw, likely reflecting the development of both peripheral and central sensitization (18,23–25). Following termination of drug administration and resolution of CA, a state of vulnerability termed latent sensitization (22) is established so that CA is elicited by normally subthreshold stimuli including those that may promote migraine in humans (18,23–25).

Most episodic and chronic migraine sufferers are women and therefore acute medications for migraine are mainly used in women. Whether there are sex differences in the mechanism(s) promoting migraine chronification is unknown. We have previously demonstrated sumatriptan-induced MOH in both male and female rodents (18,26), but whether the mechanisms promoting this effect differ between the sexes is not known. Here, we determined whether there is a sexually dimorphic mechanism underlying sumatriptan-induced MOH and if this is linked to the PRL-PRLR pathway.

Materials and methods

Animals

All procedures were performed in accordance with the policies and recommendations of the International Association for the Study of Pain, the National Institutes of Health guidelines for the handling and use of laboratory animals, and by the Institutional Animal Care and Use Committee (IACUC) of the University of Arizona. Female and male 8–10-week-old mice were used in this study. Mice were housed two to three per cage and were maintained in a climate-controlled room on a 12-h light/dark cycle with access to food and water ad libitum. C57BL/6J mice were purchased from Jackson Laboratories. PRLR-L^Cre mice were generated by Dr. Ulrich Boehm (University of Saarland School of Medicine, Homberg, Germany) and kindly provided by Dr. Armen Akopian (University of Texas Health Science Center at San Antonio, San Antonio, Texas). In these mice, an internal ribosome entry site (IRES) followed by Cre recombinase cDNA is inserted immediately after exon 10 in the prlr gene (27). Ai6 reporter mice (B6.Cg-Gt(ROSA) 26Sortm6 (CAG-ZsGreen1) Hze/J; Stock No: 007906), which produce ZsGreen, a green fluorescent protein (GFP), in Cre-recombinase expressing cells, were purchased from Jackson Laboratories and were crossed with PRLR-L^Cre/− mice to generate PRLR-L^Cre/−;Ai6^+/− mice. In all experiments, mice were randomly allocated to treatment groups. Investigators were blinded to treatments for all behavioral experiments. No animals in behavioral experiments were excluded from analysis.

Drugs

Sumatriptan (10 mg/kg i.p. in saline) was obtained from Abmole Bioscience (Houston, TX, USA) and administered as a single dose or once daily for 9 days. Cabergoline (1.2 mg/kg i.p. in 10% DMSO, 10% Tween 80 and 80% saline) was from Tocris Bioscience (Batch# 2A/206222). Recombinant mouse prolactin (0.5 µg/5 µl in PBS) was obtained from Sigma Aldrich (SRP4688-50UG) and administered onto the dura mater of anesthetized mice (2% isoflurane) at the intersection of the lambdoidal and sagittal sutures using a modified internal cannula (part # 8IC313ISPCXC, Internal Cannula, standard, 28 gauge, fit to 0.7 mm) according to our previous study (14).

Periorbital and hindpaw tactile allodynia assessment

As described previously (14), mice were acclimated for 2 h individually in clear Plexiglas chambers (7.5 × 7.5 × 18 cm) wrapped with black poster board on elevated wire-mesh platforms to allow access to the forehead and ventral surface of hindpaws. For response frequency measurements, calibrated von Frey filaments (Stoelting Co) were applied 10 times to the forehead (force 0.4 g) or to the hindpaws (force 1.0 g) with a space of ∼30 sec between each application with just enough pressure to cause the filament to arch. Shaking of head, face washing, and scratching with front paws were considered as positive responses to facial stimulation. For hind paw measurements, sharp withdrawal of the paw, shaking and/or licking the paw were considered a positive response. Frequency response was calculated as [(number of positive responses/10) * 100%].

Mouse prolactin ELISA

Mice were anesthetized with isoflurane (2%) and whole blood was collected by inferior vena cava puncture and coagulated at room temperature for 2 h. Serum was isolated by centrifugation at 1000 g for 15 min at 4°C. Serum samples were collected and stored at −80°C until use. Serum prolactin was quantified by a mouse prolactin ELISA kit (Abcam, ab100736) according to the manufacturer instructions.

Immunoblot preparation and analysis

Bilateral trigeminal ganglia were dissected after the last behavioral testing. V3 and V2 branches were removed under a dissecting microscope and remaining V1 was immediately frozen in liquid nitrogen and stored at −80°C. Tissue lysates were prepared and analyzed with western blot as described before (14) with the liquid transfer step done using Tris-Glycine SDS Running Buffer (TGS; Cat# LC26754, Novex) 20% (vol/vol) methanol as transfer buffer. The following primary antibodies were used (anti-PRLR: Cat# ab2772, Abcam for western blots of PRLR-S, Cat# ab170935, Abcam for western blots of PRLR-L, and anti-βIII-tubulin: Cat# G7121, Promega). Horseradish peroxidase-conjugated light chain and Fc-specific secondary antibodies were from Jackson Immunoresearch. Samples from parallel groups within the experiment were always run on the same gels using the same antibodies, allowing relative comparisons and reliable results. For all experiments, protein expression was normalized to βIII-tubulin in the same sample.

Visualization of PRLR-L^Cre cells using fluorescence microscopy

Naïve PRLR-L^Cre/−;Ai6^+/− male and female mice were anesthetized and transcardially perfused with 4% paraformaldehyde. The TGs were collected and, following 2 h post-fixation, 10-µm thick sections were cut on a Microm HM 525 cryostat and mounted on Surgipath X-tra microscope slides (Leica Biosystems). The sections were stained with DAPI and examined under an Olympus BX51 microscope equipped with a Hamamatsu C8484 digital camera using HC Image Live Imaging Software (Hamamatsu Corporation, Version 4.1.6.0).

In situ hybridization with RNAscope

Co-localization of mRNA for prolactin receptor (prlr) and serotonin receptor 1b (htr1b) or 1d (htr1d) in the TG of male and female C57Bl/6J mice was assessed using RNAscope® Multiplex Fluorescent assay (Advanced Cell Diagnostics, ACDbio). Briefly, transcardial perfusion/fixation was performed in mice with 4% paraformaldehyde. The dissected TGs were postfixed for an additional 24 h according to the manufacturer’s protocol. Frozen sections 10 µm thick were cut and transferred to PBS for rehydration. This was followed by multiple steps of tissue pretreatment according to the ACDbio protocol. Hybridization was performed for 2 h at 40°C with a mixture of RNAscope® Probe- Mm-Prlr-C2 (Cat# 430791-C2) and either RNAscope® Probe- Mm Htr1b (Cat# 315861) or Mm-Htr1d (Cat# 315871). The hybridized RNA probes were then recognized and amplified three times using the amplifiers provided in the RNAscope® Multiplex Fluorescent Reagent Kit v2 (Cat# 323100). Finally, the amplified RNA probe signals were labeled by fluorescent dyes Opal 520 and Opal 570 (Akoya Biosciences, FP1487001KT and FP1488001KT) for microscopic viewing of prlr and htr1b or htr1d, respectively. Revolve microscope (Echo Laboratories Inc.) was used for fluorescence observation; images were taken with a 20× objective.

CRISPR/Cas9 gene editing of prlr

Deletion of the total PRLR (PRLR-S and PRLR-L) was achieved by targeting the exon 1 of the prlr gene (ENSMUST00000128921.7) with a previously validated guide RNA (gRNA total PRLR: GTGTCAGGGGAACGACATTTG, quality score 97) inserted between the Esp3I restriction sites (Cat# ER0451, ThermoFisher Scientific, Waltham, MA) of the pL-CRISPR.EFS.GFP plasmid (Cat# 57818, Addgene) (28). For in vivo transfection, the CRISPR plasmids were diluted to 0.4 µg/µl in 5% sterile glucose solution. Then, Turbofect in vivo transfection reagent (Cat# R0541, ThermoFisher) was added following manufacturer’s instructions. Finally, 15 µl of the plasmid complexes were delivered in each nostril. The nasal plasmid administration was repeated a second time after 5 days. Sumatriptan administration began 3 days after the second nasal plasmid administration.

Statistical analysis

Numbers of mice per group to obtain significance at the α = 0.05 at statistical power 0.9 for PRL ELISA and allodynia measurements were estimated to be n = 6 using power analysis (G-power). Data are expressed as means ± SEM. Each data point represents an individual score. Statistical analyses were performed using GraphPad Prism 8 (GraphPad, La Jolla, CA). Differences between mean values of two groups were evaluated by two-tailed Mann–Whitney test and two-tailed unpaired t test. Mean differences of more than two groups were analyzed using one-way analysis of variance (ANOVA) with Tukey’s multiple comparisons post hoc test. Differences in time-course experiments were assessed by two-way ANOVA with Bonferroni’s multiple comparison post hoc test. Differences were considered to be statistically significant when p < 0.05. p-values and sample sizes are indicated in the respective figure legends.

Results

Repeated administration of sumatriptan to both male and female mice produced periorbital and hindpaw cutaneous allodynia

We have previously characterized transient periorbital and hindpaw cutaneous allodynia (CA) resulting from prolonged or repeated administration of sumatriptan as a model of MOH in male rats (26). This model has also been replicated in our laboratory, and by others using male and female mice. However, whether there may be sex differences in the underlying mechanisms promoting MOH has not been explored. Therefore, we repeatedly administered sumatriptan to parallel groups of female and male mice and evaluated periorbital and hindpaw CA. Sumatriptan (10 mg/kg, i.p.) was administered once daily for 9 days at 0900 and allodynia was evaluated 24 h later before the next sumatriptan dose. Compared to the vehicle-treated control groups, time-dependent CA was observed after repeated administration of sumatriptan at periorbital (Figure 1(a) and (b)) and hindpaw (Figure 1(c) and (d)) regions in both female and male mice (Figure 1(a)–(d); *p < 0.05, **p < 0.01, ***p < 0.001; two-way repeated measures ANOVA with Bonferroni’s post hoc test for multiple comparisons; n = 6 animals per group). Under our experimental conditions, there was no difference in the time course and the degree of CA between male and female mice.

Figure 1.

Repeated systemic administration of sumatriptan elicits cutaneous allodynia in both female and male mice and promotes downregulation of PRLR-L in the V1 region of the trigeminal ganglion only in females. Sumatriptan (10 mg/kg, i.p.) or saline were given once a day to female and male mice. Mechanical periorbital ((a),(b)) and hindpaw ((c),(d)) allodynia were measured at baseline (BL) and over a time course of 9 days in female ((a),(c)) and male ((b),(d)) mice. Western blot quantification of PRLR isoform expression in TGV1 from female (e) and male (f) mice receiving repeated sumatriptan or saline. Insets show representative examples of PRLR-L, PRLR-S and βIII-tubulin immunoblots. Serum levels of PRL 30 min after a single dose of sumatriptan in female (g) and male (h) mice (note the different scale of the Y axes). Group data are expressed as means ± SEM; n = 6 mice per group per sex.

Repeated administration of sumatriptan down-regulated PRLR-L in female, but not male, mice

After the last behavioral testing following repeated sumatriptan treatment, we collected TGV1 and investigated whether protein expression of PRLR-L and PRLR-S isoforms in the TGV1 was altered by sustained sumatriptan exposure. In female mice, repeated administration of sumatriptan produced a significant down-regulation of TGV1 PRLR-L (saline: 100 ± 11.7%; sumatriptan: 42.8 ± 6.4%; **p = 0.002; n = 6/group) without change in expression of PRLR-S (saline: 100 ± 13.3%; sumatriptan: 108.6 ± 11.1%; p = ns; n = 6/group) (Figure 1(e)). The same sumatriptan treatment had no effect on expression of either PRLR-L (saline: 100 ± 15.1%; sumatriptan: 82.6 ± 6.1%, p = ns; n = 6/group) or PRLR-S (saline: 100 ± 17.1%; sumatriptan: 93.9 ± 19.6%, p = ns; n = 6/group) in TGV1 of male mice (Figure 1(f)).

Sumatriptan administration increased serum PRL levels in female mice

To investigate if increased PRL levels may contribute to dysregulation of PRLR isoforms, we measured PRL levels in serum of female and male mice 30 min after a single dose of sumatriptan (10 mg/kg, i.p.). As demonstrated previously, basal PRL levels were much higher in females (11.3 ± 3.2 ng/ml) than males (0.73 ± 0.22 ng/ml). Sumatriptan significantly increased PRL levels in females (to 27.8 ± 2.5 ng/ml; **p = 0.0043; n = 6/group; Figure 1(g)) while a small increase in males was not statistically significant (1.12 ± 0.36 ng.ml; p = ns; n = 6/group; Figure 1(h)).

PRLR-L receptor is found in TG of female and male mice and prlr⁺ neurons co-express htr1b or htr1d

PRLR-L^Cre/−;Ai6^+/− female and male mice were used to visualize PRLR-L^Cre expressing neurons in trigeminal ganglia (Figure 2(a)). PRLR-L^Cre expressing neurons were observed in all TG regions in females, but particularly strong expression was observed in the V1 region. PRLR-L^Cre expressing neurons were also observed in male mice but at lower numbers than in female mice. We used in situ hybridization with RNAscope to investigate if PRLR⁺ neurons also express 5-hydroxytryptamine (5HT) 1B or 1D receptors. Labeling of prlr mRNA probe confirms the expression of prlr in TG neurons including in the V1 region of both female and male mice (Figure 2(b) and (c)). Double labeling with mRNA probes to prlr and htr1b or htr1d demonstrates that many prlr⁺ neurons in female TGV1 region also express mRNA encoding the 5HT1B (Figure 2(b)) or 5HT1D (Figure 2(c)) receptors. Few neurons in the male TGV1 co-expressed transcripts for both PRLR and 5HT1B (Figure 2(b)) or PRLR and 5HT1D (Figure 2(c)) receptors.

Figure 2.

PRLR expression in TGs and colocalization with 5HT1B and 5HT1D receptors. (a) Representative images of PRLR-L^Cre/−;Ai6^+/− in TGs from female (top) and male (bottom) mice. The arrowheads indicate GFP positive (PRLR-L^Cre positive) cells. DAPI, 4′,6-diamidino-2-phenylindole is shown in blue. (b) Representative images of RNAscope in situ hybridization transcripts for PRLR receptor (red) and 5HT1B receptor (green) in female (top) and male (bottom) mice. (c) Representative images of RNAscope in situ hybridization transcripts for PRLR receptor (red) and 5HT1B receptor (green) in female (top) and male (bottom) mice. Scale bar: 50 µm.

Inhibition of circulating PRL prevented cutaneous allodynia in female, but not male, mice

To interrogate the possible role of circulating PRL in female and male mice with repeated administration of sumatriptan, we assessed the efficacy of cabergoline, a dopamine D2 agonist that inhibits secretion of PRL from pituitary lactotrophs (14,29,30). Consistent with our previous observations (13), repeated administration of sumatriptan with vehicle produced both periorbital (Figure 3(a) and (b)) and hindpaw (Figure 3(c) and (d)) allodynia in female and male mice (Figure 3(a)–(d) *p < 0.05, **p < 0.01, ***P < 0.001; two-way ANOVA with Bonferroni’s post hoc test for multiple comparisons; n = 6/group). However, administration of sumatriptan with cabergoline significantly inhibited both periorbital and hindpaw CA in female (Figure 3(a),(c); #p < 0.05, ##p < 0.01) but not in male (Figure 3(b),(d)) mice. As expected, repeated administration of saline with vehicle or with cabergoline did not produce significant changes in response frequency in either female or male mice.

Figure 3.

Cabergoline prevented periorbital and hindpaw allodynia and down-regulation of TG PRLR-L induced by repeated administration of sumatriptan in female, but not male mice. Effect of pretreatment with cabergoline on periorbital ((a),(b)) and hindpaw ((c),(d)) allodynia elicited by repeated administration of sumatriptan in female ((a),(c)) and male ((b),(d)) mice. Western blot quantification of PRLR isoform expression in TGV1s from female (e) and male (f) mice that received saline or sumatriptan and vehicle or cabergoline. Representative western blot images are shown in the insets. Group data are expressed as means ± SEM; n = 6 per group per sex.

Cabergoline prevented down-regulation of PRLR-L in TGV1 region selectively in females

We also determined whether cabergoline could alter protein expression of PRLR-L and PRLR-S isoforms in the TGV1 of female or male mice treated with saline or sumatriptan. We found that in vehicle-pretreated female mice repeated injection of sumatriptan suppressed the expression of PRLR-L protein but pretreatment with cabergoline prevented the down-regulation of TGV1 PRLR-L (saline/vehicle: 100 ± 6.4%; sumatriptan/vehicle: 52.2 ± 6.4%; saline/cabergoline: 106 ± 3.7%; sumatriptan/cabergoline: 99.0 ± 18.6%; one way ANOVA; p = 0.0059; Tukey’s post hoc analysis: *p = 0.0206 for saline/vehicle and sumatriptan/vehicle groups; #p = 0.024 for sumatriptan/vehicle and sumatriptan/cabergoline groups; n = 6/group) (Figure 3(e)). No effect of sumatriptan or cabergoline on TGV1 PRLR-S expression was observed in these female groups (saline/vehicle: 100 ± 8.5%; sumatriptan/vehicle: 81.3 ± 3.3%; saline/cabergoline: 107.6 ± 6.5%; sumatriptan/cabergoline: 81.9 ± 6.8%; n = 6/group) (Figure 3(e)).

In contrast, in male mice, neither sumatriptan nor cabergoline had any effect on the expression of PRLR-L (saline/vehicle: 100 ± 17.5%; sumatriptan/vehicle: 125.4 ± 27.9%; saline/cabergoline: 103.6 ± 25.7%; sumatriptan/cabergoline: 97.6 ± 53.1%; n = 6/group) or PRLR-S (saline/vehicle: 100 ± 16.7%; sumatriptan/vehicle: 83.5 ± 5.7%; saline/cabergoline: 81.6 ± 10.3%; sumatriptan/cabergoline: 73.0 ± 5.3%; n = 6/group) (Figure 3(f)).

Editing of total PRLR in TGs of female mice prevented periorbital allodynia induced by dural application of prolactin or by repeated sumatriptan treatment

To determine if the effects of sumatriptan in producing CA in female mice were dependent upon PRLR in the TG, prlr gene was edited by two administrations of CRISPR/Cas9 plasmid for total PRLR (PRLR-T) spaced by 5 days and delivered by the nasal route. The deletion of both PRLR-L and PRLR-S isoforms in the TGV1 was confirmed using western blotting (Figure 4(a)). Compared to control plasmid, PRLR-T CRISPR/Cas9 plasmid resulted in statistically significant downregulation of both PRLR isoforms (Figure 4(c); PRLR-L expression was 100 ± 8.0% for control and 37.5 ± 10.3% for PRLR-T CRISPR plasmid, **p < 0.01; PRLR-S expression was 100 ± 6.4% for control and 65.0 ± 9.6% PRLR-T CRISPR plasmid, **p < 0.01).

Figure 4.

Editing of total PRLR in TGs in female mice reduced periorbital allodynia induced by dural application of prolactin or by repeated administration of sumatriptan. (a) Nasal administration of total PRLR CRISPR resulted in down-regulation of both PRLR isoforms in the female TGs. The inset shows representative western blot images of PRLR-L, PRLR-S and βIII-tubulin. (b) Effect of nasal pretreatment with PRLR-T CRISPR on periorbital allodynia elicited by application of prolactin on the dura mater of female mice. Effect of pretreatment with PRLR-T CRISPR on sumatriptan-induced periorbital (c) and hindpaw (d) allodynia in female mice.

Baseline tactile responses were measured 16 days after the second plasmid delivery; mice then received dural injections of either vehicle or prolactin (0.5 µg/5 µl in PBS) and CA was measured 1 h later (Figure 4(b)). Prolactin elicited significant periorbital allodynia in mice receiving control plasmid, but PRL-induced CA was significantly reduced in mice receiving PRLR-T CRISPR (control/vehicle vs. control/PRL: ***p < 0.001; control/PRL vs. PRLR-T CRISPR/PRL: ^##p < 0.01; PRLR-T CRISPR/vehicle vs. PRLR-T CRISPR/PRL: p = ns; two-way repeated measures ANOVA with Tukey’s post hoc test for multiple comparisons; n = 6 mice per group) (Figure 4(b)).

Three days following the second nasal plasmid delivery, female mice received daily systemic administration of sumatriptan, or saline as described above and periorbital and hindpaw CA was assessed (Figure 4(c) and 4(d)). Time-related periorbital CA peaking on day 9 after sumatriptan was observed in mice receiving control plasmid, but not in mice receiving PRLR-T CRISPR (control/saline vs. control/sumatriptan: *p < 0.05, ***p < 0.001; PRLR-T CRISPR/saline vs. PRLR-T CRISPR/sumatriptan: p = ns; two-way repeated measures ANOVA with Bonferroni’s post hoc test for multiple comparisons; n = 6 mice per group) (Figure 4(c)). The difference between the control/sumatriptan group and the PRLR-T CRISPR/sumatriptan group was statistically significant (^##p < 0.01; n = 6/group). In contrast, both sumatriptan-treated female groups receiving either control plasmid or PRLR-T CRISPR demonstrated significant hindpaw CA (control/saline vs. control/sumatriptan or PRLR-L CRISPR/saline vs. PRLR-L CRISPR/sumatriptan: *p < 0.05, **p < 0.01; ***p < 0.001; two-way repeated measures ANOVA with Bonferroni’s post hoc test for multiple comparisons; n = 6/group) (Figure 4(d)) and there was no difference between control/sumatriptan and PRLR-T CRISPR/sumatriptan groups.

Discussion

Both women and men with episodic migraine are at risk of developing chronic migraine, but whether the same underlying mechanisms are involved has not been previously investigated. We used a well-established model of MOH (18,26) and chronic migraine (CM) to explore the possibility that mechanisms promoting these conditions may be sexually dimorphic. We found that a) repeated sumatriptan promotes periorbital and hindpaw CA in both male and female mice; b) a single administration of sumatriptan increases levels of circulating prolactin in female mice; c) the PRLR-L is expressed in the TGV1 in both sexes; d) transcripts for the PRLR and 5HT1B or 5HT1D receptors are co-localized in female TG neurons; e) repeated sumatriptan down-regulates TGV1 PRLR-L isoform without changes in PRLR-S only in females, f) inhibition of circulating PRL with cabergoline selectively inhibits the dysregulation of PRLR isoforms and blocks both periorbital and hind paw CA only in females and g) editing of TG PRLR with nasal PRLR-T CRISPR plasmid prevents sumatriptan-induced periorbital, but not hind paw, CA in females. These findings suggest that female selective actions of PRL and PRLR signaling provide a novel hypothalamic neuroendocrine mechanism of nociceptor sensitization that may increase the likelihood of migraine attacks and enhance the risk of migraine chronification in women.

Migraine is a neurological disorder that predominately affects women of child-bearing age (22,31). Differences in sex prevalence begin at menarche and decrease around the time of menopause implying a role of sex hormones, including estrogen and testosterone (32). Emerging data from preclinical studies have raised the previously unsuspected possibility of sex differences in the nociceptor (14,16,19,33). Sex differences in TG afferents that innervate the dura and likely contribute to pain have been reported (34,35).

Migraine is a functional pain disorder in which pain can occur in the apparent absence of injury (18,36). We have previously developed and characterized an injury-free model of migraine-related pain based on clinical observation that medications which are used for acute migraine treatment can model the clinical syndrome of chronic migraine and MOH (22–26,37). Triptans are known to increase migraine frequency and produce migraine chronification in some patients. In rodents, repeated administration of triptans produce CA, a surrogate measure of migraine-like pain (10,22,38,39). The number of days with migraine headache and acute medication consumption are major risk factors for migraine chronification (40–42), suggesting that each pain attack that is treated with acute medication can promote a priming effect to increase the likelihood of a subsequent attack (43,44). Importantly, the translational relevance of this model is enhanced by the lack of tissue injury consistent with nociplastic pain, and the assessment of CA, which is seen in the majority of migraine patients during an attack (20,45).

We have previously reported that circulating PRL elicits female-selective sensitization of nociceptors in the dorsal root ganglia by down-regulating PRLR-L, allowing increased signaling at PRLR-S homodimers (14). Lowered nociceptor thresholds can increase pain from normally innocuous stimuli. For this reason, we determined if sumatriptan could dysregulate PRLR isoforms establishing a sexually dimorphic mechanism relevant to migraine chronification. MOH occurs in both men and women. Consistent with this clinical observation, repeated administration of sumatriptan produced time-dependent periorbital and hindpaw CA in both male and female mice. We previously reported that repeated sumatriptan produces CA in male and female rats (18,26) and increases in the number of identified dural afferents that express CGRP (18). PRLR is mainly expressed on small diameter primary afferents that also express the TRPV1 channel, supporting their classification as nociceptors (16). Importantly, these cells also show high expression of CGRP (16). We found that the PRLR-L is expressed in the trigeminal V1 region of both female and male mice but the number of PRLR-L^Cre cells was greater in female PRLR-L^Cre/-;Ai6^+/- mice. RNAscope studies showed that prlr was co-localized in the same cells with htr1b or htr1d to a greater extent in the TG of female, than male, mice. Prlr, htr1b and htr1d transcripts were also found in human trigeminal ganglia from both men and women (46,47); however, possible sex-dependent expression and colocalization of these transcripts were not determined. Collectively, these data provide an anatomical basis for a functional role of PRLR and 5HT1B/1D activation in the modulation of meningeal nociceptors in females.

Serotonin is known to tonically stimulate PRL release and serotonergic agonists acting at 5HT1 receptors consistently increase plasma PRL in humans (48). Despite some conflicting reports, most studies, however, show that sumatriptan decreases circulating PRL in healthy human volunteers (49,50). In contrast, we observed that a single dose of sumatriptan produced a significant increase in plasma PRL in female mice. In preclinical studies, repeated administration of sumatriptan was shown to decrease expression of 5HT1B/D/F receptors in the trigeminal ganglion and in other migraine-related tissues (51) suggesting that a possible disruption of serotonergic autoreceptor inhibition could affect serotonergic PRL modulation. Additionally, while sumatriptan has highest affinity for 5HT1B/1D receptors, studies in vitro have demonstrated significant binding affinity and full agonist actions for other serotonergic receptors including 5HT1A (52,53). Differential regulation of serotonergic receptors and PRL modulation with repeated agonist exposure is possible. Neuroendocrine disturbances have been reported in patients with MOH (54) but there is a lack of information on PRL levels in these patients. Notably, however, patients with chronic migraine showed higher PRL levels than those with episodic migraine (55) and opioids, another class of drugs that elicit MOH, are also strongly associated with hyperprolactinemia (14,56). Further studies are required to determine the consequences of triptan overuse in MOH patients on circulating PRL.

We administered cabergoline, a dopamine D2 agonist used clinically for migraine related to prolactinomas, to determine a possible role of circulating PRL in sumatriptan-induced nociceptor sensitization (57). Cabergoline mimics the actions of endogenous dopamine to decrease circulating PRL in humans (9) and in mice of both sexes (14). Co-administration of cabergoline with sumatriptan prevented PRLR-L downregulation selectively in female mice. This suggests that circulating PRL may lower the threshold for nociceptor activation through PRLR-S signaling in females. PRL does not directly activate nociceptors in TG cultures but sensitizes them so that subthreshold stimuli can produce enhanced responses including the release of CGRP, known to be causal in promoting migraine headache (12). Incubation of cultured TG cells from female mice with PRL resulted in increased capsaicin-evoked CGRP release (16).

Co-administration of cabergoline with sumatriptan also prevented the expression of both periorbital and hindpaw CA in female mice but had no effect in males. These results support the effect of circulating PRL in promoting CA in female mice. Application of PRL and CGRP to the dura mater produces pain behaviors selectively in female mice (19). Therefore, PRL-induced sensitization of dural afferents could promote CA in females through enhanced CGRP release (35). We have previously demonstrated that opioid-induced hyperalgesia in female mice requires increased circulating PRL as well as signaling at both PRLR and mu opioid receptors in dorsal root ganglia (DRGs). Whether a similar mechanism involving concurrent TG activation of prolactin and serotonin receptors by circulating PRL and by repeated sumatriptan administration, respectively, is responsible for MOH-related CA will require further investigations. Our findings that prlr and htr1b or htr1d transcripts are co-expressed in a large number of TGV1 neurons in females suggest a possibility that both PRLR and 5HT1B/1D receptors participate in lowering the sensory threshold for nociceptor activation resulting in CA.

The possible causal effects of PRL signaling in the TG on sumatriptan-induced CA were additionally addressed by editing expression of both PRLR isoforms with nasal administration of a PRLR-T targeting CRISPR/Cas9 plasmid in female mice. Pilot experiments optimized the dosing procedure to demonstrate that two plasmid administrations would significantly inhibit expression of TG PRLR-T. Following PRLR-T editing in the TG, periorbital CA induced by dural application of PRL or by repeated sumatriptan administration was significantly inhibited. Importantly, however, nasal CRISPR plasmid did not affect sumatriptan-induced hindpaw CA. Hindpaw CA following repeated exposure to drugs associated with MOH such as sumatriptan or opioids has been thought to reflect central sensitization that can amplify the activity of primary afferents. The present results, however, suggest that systemic administration of drugs may influence the excitability of afferent fibers that express serotonin or opioid receptors throughout the body, possibly as a consequence of co-activation of PRLR by circulating PRL. Interestingly, the 5HT1B/1D receptors are expressed in dorsal root ganglion cells (58). In the present study, nasal CRISPR effects on CA were localized to the periorbital region, reflecting modulation of the innervation territory of the trigeminal nerve. Sparing of the increased sensitivity of hindlimb afferents innervated by fibers from the lumbar dorsal root ganglia supports a local action of the CRISPR gene disruption and provides an important internal control for specificity. This conclusion is consistent with the inhibition of both periorbital and hindpaw CA by cabergoline which reduces circulating PRL that can act at PRLR on TG and DRG afferents throughout the body.

Migraine may be a cycling sensory threshold disorder that arises from brain circuits including the hypothalamus and the brainstem (8). Hypothalamic activity increases in the premonitory phase and precedes brain activity changes observed during the headache (59). The anterior hypothalamus is an important region in migraine attack generation and chronification of migraine and has been suggested to act as a “hub” between attack-precipitating factors and promotion of the headache phase through descending pain modulatory circuits. While descending neural circuits can undoubtedly amplify nociceptive inputs from the trigeminal afferents to promote pain, an additional important possibility supported by our findings is that this region of the hypothalamus promotes nociceptor sensitization through a neuroendocrine mechanism in females (57). Hypothalamic PRL-induced modulation of PRLR isoform expression in nociceptors can result in increased signaling at PRLR-S homodimers (15), providing a mechanism by which sensory thresholds are lowered so that sub-threshold stimuli may provoke increased frequency of migraine attack and increased risk of migraine chronification in women.

Our data reveal an unexpected sexually dimorphic mechanism of MOH/CM (Figure 5) suggesting that acute migraine therapies, which are mostly used by women, may promote transformation to chronic states. Uncovering the mechanisms that can promote MOH/CM in males will require future investigation. However, our studies provide a basis for the development of sex-dependent mechanism-based therapies, some of which can be immediately investigated for the treatment of migraine and prevention of transition to chronic migraine in women. Finally, our studies highlight a novel strategy for targeting gene therapy to the TG for prevention of chronic migraine in high-risk episodic migraine subgroups, as well as for treatment of chronic migraine, and other craniofacial pain disorders, by nasal administration of CRISPR plasmid.

Figure 5.

Proposed female selective mechanism of trigeminal nociceptor sensitization that may help to explain increased prevalence of migraine in women. Repeated treatment with sumatriptan promotes allodynia and MOH in both males and females. However, only in females, a single systemic administration of sumatriptan increases serum levels of PRL. Repeated sumatriptan promotes down-regulation of PRLR-L in trigeminal nociceptors, resulting in increased formation of PRLR-S homodimers selectively in females. PRL signaling at PRLR-S represents a female selective mechanism of allodynia and MOH that can be blocked by reducing serum PRL with cabergoline or by reducing expression of PRLR-S with nasal administration of CRISPR/Cas9 targeting PRLR-T.

Article highlights

A sexually dimorphic neuroendocrine mechanism promoting migraine that involves dysregulation of trigeminal prolactin receptor isoforms was revealed in a model of chronic migraine.

In female mice, chronic sumatriptan administration induced down-regulation of prolactin receptor long isoform, promoting tactile allodynia through signaling at the short isoform of the receptor by circulating prolactin.

Dysregulation of prolactin receptor isoforms may explain the increased risk of migraine attack from subthreshold stimuli and chronification in females.

Novel sex-specific therapies including gene editing through nasal delivery of CRISPR/Cas9 constructs may be feasible.

Data and materials availability

All raw data, methods and analyses will be available upon request to the corresponding author.

Supplemental Material

sj-pdf-1-cep-10.1177_03331024211039813 - Supplemental material for A prolactin-dependent sexually dimorphic mechanism of migraine chronification

Supplemental material, sj-pdf-1-cep-10.1177_03331024211039813 for A prolactin-dependent sexually dimorphic mechanism of migraine chronification by Daigo Ikegami, Edita Navratilova, Xu Yue, Aubin Moutal, Caroline M Kopruszinski, Rajesh Khanna, Amol Patwardhan, David W Dodick and Frank Porreca in Cephalalgia