Testosterone protects against the development of widespread muscle pain in mice

Abstract

Chronic widespread pain conditions are more prevalent in women than men suggesting a role for gonadal hormones in the observed differences. Previously, we showed female mice, compared to male, develop widespread, more severe, and longer duration hyperalgesia in a model of activity-induced muscle pain. We hypothesized testosterone protects males from developing the female pain phenotype. We tested if orchiectomy of males prior to induction of an activity-induced pain model produced a female phenotype and if testosterone administration produced a male phenotype in females. Orchiectomy produced longer lasting, more widespread hyperalgesia, similar to females. Administration of testosterone to females or orchiectomized males produced unilateral, shorter lasting hyperalgesia. Prior studies show that the serotonin transporter (SERT) is increased in the nucleus raphe magnus (NRM) in models of chronic pain, and that blockade of SERT in the NRM reduces hyperalgesia. We examined potential sex differences in the distribution of SERT across brain sites involved in nociceptive processing using immunohistochemistry. A sex difference in SERT was found in the NRM in the activity-induced pain model; females had greater SERT-immunoreactivity than males. This suggests testosterone protects against development of widespread, long-lasting muscle pain and that alterations in SERT may underlie the sex differences.

Introduction

Sex differences are present in both the symptomology and prevalence of chronic pain conditions. Females show a greater incidence of chronic pain conditions including headache, osteoarthritis, rheumatoid arthritis, irritable bowel syndrome, and fibromyalgia [22;45;63]. In animals, numerous studies show sexually dimorphic pain behavior with females typically demonstrating more severe and longer-lasting pain compared to males within models of arthritic, neuropathic, cancer-induced, stress-induced, and muscle pain [15;20;24;32;46;60;65]. Also, after induction of chronic pain, females require higher doses of mu-opioid agonists to see similar pain reduction levels as males [3;29;40;67]. These sexual differences found for severity, duration, and alleviation of pain suggest the presence of a sexually dimorphic mechanism underlying chronic pain.

Varying levels of sex hormones could mediate the observed sexual dimorphism in pain. Sex differences in the prevalence of pain conditions are not present in children until puberty suggesting a role for sex hormones [6;30]. Numerous studies show that the female sex hormone estrogen/estradiol mediates the increased severity of hyperalgesia since removal of the ovaries and blockade of estrogen receptors reduces pain in models of inflammation and visceral pain [27;36]. However, more recent studies show that testosterone also plays a role in the sexual dimorphism of pain. Specifically, testosterone modulates the spinal toll like receptor 4 (TLR4) neuroimmune response following induction of inflammatory and neuropathic pain [48;57;58]. Activation of androgen receptors transcriptionally increases expression of mu-opioid and cannabinoid type 1 receptors on peripheral nociceptors both of which have anti-nociceptive effects [3;34;35;47;70]. Lastly in humans, testosterone levels alter sites of brain activation following administration of noxious stimuli, with lower levels of testosterone resulting in less activation of regions involved in descending pain inhibition and higher activation of brain regions responsible for negative pain affect [13;66]. This research has demonstrated the anti-nociceptive effects of testosterone are multi-factorial. It is unclear, however if the effects of testosterone play a role in muscle pain.

We recently demonstrated sexual dimorphism in a model of activity-induced pain [24]. Specifically, female mice present with bilateral, long-lasting hyperalgesia that is easier to induce when compared to males who developed unilateral, short-lasting hyperalgesia. Interestingly, ovariectomy had no effect on the hyperalgesia in female mice suggesting estradiol did not influence development of chronic widespread pain. We therefore hypothesized that testosterone protects males against the development of activity-induced muscle pain. Also, since widespread hyperalgesia is thought to reflect changes in the central nervous system, we explored a possible mechanism responsible for the sex-dependent pain phenotype. In neuropathic and widespread muscle pain models we previously show increases in the serotonin transporter (SERT) in the nucleus raphe magnus [5;8], with blockade of SERT reducing hyperalgesia. Sex hormones modulate mRNA and protein expression of SERT in several brain regions with mixed results in uninjured animals [25;43;44]. Thus, we tested if alterations in SERT were sex-dependent in supraspinal sites involved in nociceptive transmission in response to the activity-induced pain model. We hypothesized that females would show a greater increase in SERT in the nucleus raphe magnus in the activity-induced pain model when compared to males.

Materials and Methods

Mice

All experiments were approved by the University of Iowa Animal Care and Use Committee and were conducted in accordance with National Institutes of Health guidelines. A total of 221 C57BL/6J mice (139 male, 82 female) (19–22 g)(Jackson Laboratories, Bar Harbor, ME) aged 6–8 weeks were used in this study. Mice were housed in transparent plastic cages on a 12-hour light–dark cycle with access to food and water ad libitum.

Orchiectomy

To test the role of testosterone on sex-dependent effects observed in previous studies, male mice were orchiectomized (Orx)(n=40) or received sham orchiectomy surgery (sham)(n=37). Animals were deeply anesthetized with 2–4% isoflurane and a single scrotal incision was made. The testes were exposed, and the vas deferens and testicular blood vessels were ligated with a synthetic non-absorbable monofilament suture. The testes and epididymis were then removed, and the incision closed with silk sutures. Sham orchiectomies were performed in a similar way except the testicles were not ligated or removed. Following either orchiectomy or sham surgery, all animals were moved to single cage housing to protect the surgical sites. Induction of activity induced pain model and subsequent behavior testing was performed two weeks after surgeries.

Hormone Administration

To further test the role of testosterone on the observed sex-dependent effects, testosterone pellets (n=24) or placebo pellets (n=21) were administered to female and orchiectomized male mice. Time release testosterone pellets (7.5 mg testosterone/pellet, 60-day release; Innovative Research of America, Sarasota, FL) or placebo pellets were inserted subcutaneously. Animals were deeply anesthetized with 2–4% isoflurane and a small incision was made at the base of the neck and the pellet was inserted through the incision. The incision was closed with a synthetic non-absorbable monofilament suture. In orchiectomized male mice, the pellets were inserted at the same time as the orchiectomy surgery. Following either testosterone or placebo pellet implantation, all animals were moved to single cage housing to protect the surgical sites. Induction of the activity-induced pain model and behavior testing was performed, or serum blood samples were collected two weeks following pellet implantation.

Blockade of Androgen Receptors

To determine if the sex differences observed in male mice required continued activation of the androgen receptor, we gave flutamide (Sigma Aldrich, St. Louis, Missouri) systemically 24hr after induction of the activity-induced pain model. The androgen receptor antagonist flutamide was delivered via subcutaneous injection (50mg/kg) at the nape of the neck. This dose was chosen based on prior research which showed strongest effects at this dose [37;71]. Flutamide was made fresh the day of injection by dissolving flutamide in ethanol/saline (2:1, vol/vol) and then diluting by a factor of 10 in a saline solution. Control animals received injections of the vehicle. Behavior testing was re-administered 30 minutes and 2 hours post injection.

Behavioral Assessments

Muscle withdrawal thresholds (MWT) were measured by applying custom-built force sensitive tweezers to the belly of the gastrocnemius muscle as previously described, where lower force thresholds indicate higher sensitivity [55]. Mice were acclimated in two, five-minute sessions over a two-day period by placing them in a gardener’s glove with their hindlimbs held into extension outside of glove. On the testing day, the gastrocnemius muscle was squeezed with force sensitive tweezers until the animal withdrew its hindlimb, MWT were recorded as the amount of force required to elicit this withdrawal. In order to prevent behavioral sensitization to testing, five minutes elapsed between each MWT assessment. Each animal was tested bilaterally and the average of three trials was used to determine withdrawal thresholds for each limb. A decrease in withdrawal thresholds was interpreted as muscle hyperalgesia. All testers were blinded to treatment group during pain assessments.

Fatigue activity paradigm

Fatigue was induced in the gastrocnemius muscle using electrically simulated muscle contractions as previously described [24]. Briefly, mice were deeply anesthetized with 2–4% isoflurane. Needle electrodes coupled to a Grass S88 solid-state square wave form generator (Grass Technologies, West Warwrick, RI) were inserted into the belly of the gastrocnemius muscle in an orientation to run parallel with the muscle fibers. Baseline maximum force production was quantified through three 100Hz trains at 7V. To induce muscle fatigue, six minutes of submaximal contractions were administered using 7V stimulations at 40 Hz for 3.75 s with 4.25 s of rest between stimulations. Following the six minutes of fatiguing contractions, three additional maximum force contractions were administered to measure post-fatigue maximum force production. Force was measured by adhering a force plate connected to a iWorx FT-302 force transducer (iWorx, Dover, NH) to the planter surface of the foot. Data was collected on LabView software and analyzed with FreeMat and Python scripts. Force transducer data was converted to mN using a standard curve of 1-g weights applied to the apparatus. Fatigue was defined as the decline in maximum force production from baseline following the six minutes of fatiguing stimulation. This protocol results in approximately 60% reduction in muscle force and a significant decrease in muscle pH [24].

Activity-induced muscle hyperalgesia

Muscle pain was induced through combination of low-intensity muscle insults with six minutes of fatiguing contractions. This model was administered using three different protocols. In all three protocols, on day one, mice were anesthetized with 2–4% isoflurane and given intramuscular (i.m.) injection into the left gastrocnemius of 20-μl normal saline adjusted to pH 5.0 ± 0.1 with HCl. The three protocols then differed in their combination of muscle fatigue and second muscle insult. In our original paper which characterized this model, the temporal and spatial separation experiments were used to demonstrate that the hyperalgesia was easier to induce in females [24]. We included these experiments here for completeness in testing the role of testosterone in prevention of widespread pain.

Protocol 1: Standard Protocol

On day five, mice were anesthetized with 2–4% isoflurane and underwent six minutes of fatiguing contractions in the gastrocnemius. Immediately following the fatiguing contractions, a second i.m injection of pH 5.0 was delivered into the same gastrocnemius. Muscle withdrawal thresholds were measured bilaterally 24hr after the second pH 5.0 saline injection. This protocol produces sex-dependent effects in which males have unilateral, shorter lasting decrease in muscle withdrawal threshold compared to females [24].

Protocol 2: Spatial Separation

On day five, mice received the same protocol for six minutes of fatiguing contractions into the gastrocnemius. Immediately following the fatiguing contractions, the second i.m. injection of pH 5.0 was delivered into the contralateral gastrocnemius muscle. Muscle withdrawal thresholds were measured bilaterally 24hr after the second pH 5.0 saline injection. This protocol produces sex-dependent effects in which males do not develop decreases in muscle withdrawal thresholds while females have significant decreases in muscle withdrawal thresholds bilaterally [24].

Protocol 3: Temporal Separation

On day four, mice received the same protocol for six minutes of fatiguing contractions in the gastrocnemius. On day five, mice received the second pH 5.0 i.m. injection into the same gastrocnemius muscle 24hr after the fatiguing contractions. Muscle withdrawal thresholds were measured bilaterally 24hr after the second pH 5.0 i.m. saline injection. This protocol was found to produce sex-dependent effects in which males do not develop decreases in muscle withdrawal thresholds while females have significant decreases in muscle withdrawal thresholds bilaterally [24].

Measurement of Testosterone

Mouse blood was collected into 3.0 mL serum blood collection tubes (BD vacutainer, Franklin Lakes, NJ) by cardiac stick of mice anesthetized with pentobarbital (50mg/mL). The blood was allowed to clot for 30 minutes and the serum was separated from blood by centrifugation for 10 minutes at room temperature. Total testosterone concentrations from serum samples were tested in duplicate with a testosterone rat/mouse enzyme-linked immunosorbent assay (ELISA) kit (MyBioSource, Inc. San Diego, CA).

Immunohistochemistry

Immunohistochemistry was done in animals that received two injections of pH 5.0 with muscle stimulation and compared to a control group that received two injections of pH 7.2 with muscle stimulation. Two injections of pH 7.2 with muscle stimulation does not produce hyperalgesia in either male or female mice [24].

Immunohistochemistry was performed to analyze serotonin transporter (SERT) in the spinal cord and various brain regions. We have previously shown increases in SERT in the rostral ventral medial medulla (RVM) following induction of neuropathic pain [5]. Animals were deeply anesthetized with pentobarbital (50mg/mL) and transcardially perfused with heparinized saline followed by freshly prepared 4% paraformaldehyde (PFA) in 0.1 M phosphate buffered saline (PBS). The brains and lumbar spinal cords were removed and stored in 30% sucrose in PBS overnight (4°C), tissues were then frozen in cryomolds (OCT, Tissue Tek, Fisher Scientific, Waltham, MA) at −20°C. Sections from the raphe magnus (NRM), raphe obscurus (NRO) and raphe pallidus (NRP), dorsal raphe, periaqueductal grey (PAG), cingulate cortex, insular cortex, and amygdala were cut at 20μm with a cryostat and placed onto slides.

Sections from brain and spinal cord locations for all animals were processed simultaneously and stained for SERT using previously described protocols [5]. Briefly, during the first day, sections were blocked with 5% normal goat serum followed by an Avidin/Biotin blocking kit (Vector Labs, Burlingame, Ca) and then incubated overnight at room temperature with the primary antibody against SERT (Rabbit anti-5-HT transporter; ImmunoStar, Hudson, WI, 1:500). During the second day, sections were rinsed, blocked with 5% normal goat serum, and incubated for one hour at room temperature in the secondary antibody (Biotinylated Goat anti-Rabbit immunoglobulin G (IgG), Invitrogen, Carlsbad, CA 1:1000). The sections were then reacted with streptavidin conjugated to Alexa Fluor 568 (Life Technologies, Carlsbad, CA, 1:500) for one hour at room temperature. Slides were then cover slipped with Vectashield (Vector Labs, Burlingame, CA).

Images of sections were taken on an Olympus BX-51 light microscope (Diagnostic Instruments, Sterling Heights, MI) in the Central Microscopy Facility at the University of Iowa. Specific brain locations were determined from stereotaxic coordinates from Paxinos and Watson Mouse Atlas [49] and were the following: NRM: bregma −6.48 to −5.52, NRO: bregma −6.96 to −6.24, NRP: bregma −6.48 to −5.52, PAG: bregma −3.28 to −2.54, dorsal raphe: bregma −4.84 to −4.36, insular cortex: bregma 0.98 to 1.54, cingulate cortex: bregma 0.26 to 0.98, and amygdala: bregma −1.46 to −1.22. All images were taken under the same conditions and stored for analysis. Optical density of SERT staining was analyzed using Image J software (NIH, Bethesda, MD) as previously described [8]. For each location three to five sections were quantified and averaged for each animal. The investigator was blinded to sex and treatment group during staining, image acquisition, and optical density quantification.

Experiments

Experiment 1: Effect of orchiectomy on male pain behavior

This experiment tested if orchiectomy could reverse the sex-dependent effects observed in pain behavior between males and females. Animals were grouped randomly as orchiectomized males (Orx) (n=8), sham orchiectomized males (sham) (n=7) or females (n=7). Two weeks following surgeries, the fatigue induced pain model was administered and MWT measurements were assessed at 24hr 1wk, 2wks, 3wks, 4wks, 5wks, and 6wks after induction of the model. To test if orchiectomy surgery had any effect on circulating levels of testosterone, serum was collected from the mice and analyzed with a testosterone ELISA kit. Animals were grouped as Orx (n=4) or sham (n=4). Three weeks following surgeries, serum samples were collected and analyzed.

Experiment 2: Effect of testosterone on pain behavior in orchiectomized male and female mice

This experiment tested if testosterone administration could reverse the sex-dependent effects observed in pain behavior between intact males and orchiectomized males and females. Animals were grouped randomly as orchiectomized males with testosterone pellets (Orx+T) (n=8), orchiectomized males with placebo pellets (Orx+P) (n=8), females with testosterone pellets (female+T) (n=5), or females with placebo pellets (female+P) (n=8). Two weeks following surgeries and pellet implantations, the fatigue induced pain model was administered and MWT measurements were assessed at 24hr 1wk, 2wks, 3wks, and 4wks after induction of the model. Three animals in female+P group had to be euthanized during experimental procedures due to surgical complications. Two during the initial pellet implantation surgery and the third after two weeks of testing. These animals were excluded from data analysis. To test if pellet implantation had any effect on circulating levels of testosterone, serum was again collected analyzed with a testosterone ELISA kit. A separate group of animals were grouped as Orx+T (n=4), Orx+P (n=4), female+T (n=3), or female+P (n=4). One female mouse with testosterone was lost during collection of the blood sample. Three weeks following surgeries, serum samples were collected and analyzed. Four animals were used in the female+T group, however a clot formed in vial during analysis thus data was discarded leaving three samples for final analysis.

Experiment 3: Effect of androgen receptor antagonist flutamide after induction of activity induced pain model

This experiment tested the effects of a subcutaneous injection of flutamide to reverse the male pain phenotype following induction of activity induced pain. Intact male mice received the standard activity induced pain model protocol and were assessed for hyperalgesia 24 hours induction of model. Animals then were randomly divided into two groups which received either subcutaneous flutamide injections (50mg/kg) (n=5) or vehicle (n=5) and MWT were reassessed at 30 minutes and 2 hours following the injection.

Experiment 4: Effect of orchiectomy on pain behavior when the muscle fatigue and final muscle insult were given in opposite muscles

This experiment tested if orchiectomy surgery could reverse the sex-dependent effects observed in pain behavior between males and females when the muscle fatigue and final muscle insult are delivered in opposite muscles. Animals were randomly grouped into receiving the final muscle insult into either the ipsilateral (n=24) or contralateral muscle (n=24). The ipsilateral group was further divided into Orx (n=8), sham (n=8) or females (n=8). The contralateral group was also divided into Orx (n=8), sham (n=8) or females (n=8). Two weeks following surgeries, the fatigue induced pain model was administered and MWT measurements were assessed at 24hr after the induction of the model.

Experiment 5: Effect of orchiectomy on pain behavior when the muscle fatigue was delivered 24 hr before the final muscle insult

This experiment tested if orchiectomy surgery could reverse the sex-dependent effects observed in pain behavior between males and females when the muscle fatigue is administered 24 hours before the final muscle insult. Animals were randomly grouped as Orx (n=8), sham (n=8), or females (n=8). Two weeks following surgeries, the fatigue induced pain model was administered and MWT measurements were assessed at 24 hours after the induction of the model.

Experiment 6: Sex differences in SERT immunohistochemistry following induction of activity-induced pain model

This experiment tested if there were sex differences in supraspinal sites in SERT expression in response to our activity-induced pain model. Animals were randomly grouped into receiving either the pH 7.2 model (n=4 M, n=4 F) or pH 5.0 (n=7 M, n=7 F). Brains were harvested 24hr after induction of model. Expression of SERT was quantified in the raphe magnus (NRM), raphe obscurus (NRO) and raphe pallidus (NRP), dorsal raphe, periaqueductal gray (PAG), cingulate cortex, insular cortex, and amygdala. Animals in which three or more sections for a certain brain site were not obtained were removed from analysis (see Supplemental Table 1 for sample sizes for each brain region).

Statistical analysis

All data are presented as mean ± S.E.M, with range included for testosterone values. For MWT data values are presented as millinewtons (mN) of force and testosterone values are presented as nanogram per milliliter (ng/mL). A Student’s t-test revealed baseline MWT differences between orchiectomized males and females on the ipsilateral (males 1934.5±46.4 mN vs females 1650.5±74.3 mN; p=0.005) and contralateral side (males 2102.1±70.9 mN vs females 1741.7±64.7 mN; p=0.01) To account for differences at baseline a repeated measures analysis of variance (ANOVA) with baseline values as a covariate was performed. A repeated measures ANOVA with a Tukey’s post hoc test on MWT change scores from baseline were used to assess differences between groups. In order to maintain consistency, a change score from baseline was utilized for all MWT values. Duration of hyperalgesia was calculated by examining the MWT force at each time period. If an animal returned to within 1 standard deviation from its group’s baseline score it was considered no longer hyperalgesic. A one-way ANOVA with a Tukey’s post hoc test was used to compare duration of hyperalgesia between groups. Student’s t-test was used to compare total testosterone concentrations between relevant groups (sham vs Orx; Orx+T vs Orx+P; Female+T vs Female+P). In the animals who received subcutaneous flutamide injections or vehicle a one-way ANOVA with Tukey’s post hoc test was used to test for group differences at the 30 minute and 2-hour time points. For immunohistochemistry data, a two-way ANOVA tested differences by the pain model (pH 5.0 vs. pH 7.2), sex (male vs. female) and an interaction between the pain model and sex. If an interaction between the pain model and sex was found, a post-hoc Tukey’s test examined for individual group differences. For all experiments, p<0.05 was considered statistically significant. Statistical analyses were performed on SPSS Version 25.0 (SPPS Inc. Chicago, IL).

Results

Experiment 1 – Orchiectomy of male mice resulted in increased severity of hyperalgesia that was similar to female mice

To determine if circulating testosterone levels were responsible for sex differences observed in behavior experiments, we compared orchiectomized males to intact males and females in the activity-induced pain model. There was a significant group difference for the decrease in MWT following induction of the activity-induced pain model when MWT was measured on the ipsilateral side (F_2,18 = 26.3, p<0.001). Analysis of MWT change scores from baseline on the ipsilateral side revealed similar group differences (F_2,19 = 5.4, p=0.013) with post hoc testing demonstrating significant differences between sham and Orx groups (p=0.02) and sham and female groups (p=0.03) but not between Orx and female groups (p=0.99) (Fig. 1A). Similarly, significant group differences for decrease in MWT was found on the contralateral side following induction of the model (F_2,18 = 28.4, p<0.001). Again, analysis of change scores revealed similar group differences (F_2,19 = 8.7, p=0.002) with post hoc testing showing significant differences between sham and Orx groups (p=0.017), and sham and female groups (p=0.002) but not between Orx and female groups (p=0.53) (Fig. 1B).

Figure 1.

Group differences in the changes in muscle withdrawal thresholds (MWT) following induction of activity induced pain model. (A&B) Differences in withdrawal thresholds from baseline values over a 6-week time period for the ipsilateral (A) and the contralateral sides (B). Orchiectomy (Orx) of males 2 weeks prior to induction of pain model resulted in MWT changes similar to females. *p<0.05, compared to sham orchiectomy group (sham). (C&D) Duration of muscle hyperalgesia in days for the ipsilateral (C) and the contralateral sides (D). Orchiectomy of males resulted in pain duration that was similar to females. *p<0.05, compared to sham group

Significant differences between groups for duration of hyperalgesia were demonstrated on both the ipsilateral (F_2,21 = 10.9, p = 0.001) and the contralateral side (F_2,21 = 20.38, p<0.001). Post hoc analysis of duration demonstrated group differences between sham and Orx groups (ipsilateral p=0.001, contralateral p <0.001) and between sham and female groups (ipsilateral p=0.003, contralateral p<0.001) but not between Orx and female groups (ipsilateral p=0.95, contralateral p=0.59) (Fig. 1C,D). Testosterone concentrations in animals who underwent orchiectomy surgery (0.161±0.061 ng/mL; range 0.13–0.25 ng/mL; n=4) were significantly lower than those who underwent sham surgery (0.929±0.37 ng/mL; range 0.41–1.28 ng/mL; n=4) (p<0.01) confirming that orchiectomy surgery decreases testosterone levels. Taken together this data demonstrates orchiectomized males develop a pain phenotype similar to that of females.

Experiment 2 – Testosterone administration in orchiectomized males and females decreases severity of hyperalgesia

To determine if administration of testosterone modulated the pain phenotype, testosterone pellets were implanted subcutaneously two weeks prior to induction of activity-induced pain model in female and orchiectomized male mice. Following induction of the model, significant group differences were found in the changes in withdrawal thresholds on the ipsilateral side (F_3,25 = 33.2, p<0.001). Analysis of MWT change scores from baseline on the ipsilateral side revealed similar group differences (F_3,25 = 13.85, p<0.001) with post hoc testing demonstrating significant differences between female+T and female+P groups (p=0.003) and between Orx+T and Orx+P groups (p=0.015) (Fig. 2A). Similar results were demonstrated on the contralateral side with a significant group effect for changes in MWT (F_3,25 = 42.9, p<0.001). Again, MWT change scores from the contralateral side revealed similar group differences (F_3,25 = 20.9, p<0.001) with post hoc testing demonstrating significant differences between female+T and female+P groups (p<0.001) but not between Orx+T and Orx+P (p=0.41) (Fig. 2B). In females, implantation of testosterone pellets successfully increased testosterone concentrations (25.0±0.00 ng/mL; range 25.0 ng/mL; n=3) to levels which were significantly higher than those receiving implantation of placebo pellets (0.151±0.052 ng/mL; range 0.10–0.22 ng/mL; n=4) (p<0.001). Of note, all three females with implanted testosterone pellets reached ceiling of limit of detection for ELISA kit used. Similar to females, orchiectomized males with testosterone pellets demonstrated significantly higher concentrations of testosterone (23.474±1.942 ng/mL; range 20.63–25.0 ng/mL; n=4) compared to those receiving the placebo pellets (0.168±0.123 ng/mL; range 0.03–0.32 ng/mL; n=4) (p<0.001). Taken together this data demonstrates testosterone’s role in protecting against the development of widespread hyperalgesia which is independent of sex.

Figure 2.

Group differences in MWT following 2-week administration of testosterone prior to induction of activity induced pain model. (A-D) Changes in muscle withdrawal threshold from baseline values over a 4-week time period for the ipsilateral and contralateral muscle in females (A&B) or orchiectomized males (C&D) receiving either testosterone or placebo administration. Testosterone administration in females (F+T) significantly reduced the changes in muscle withdrawal thresholds that was different from females receiving placebo (F+P) treatment on both the ipsilateral and contralateral side. Testosterone administration in orchiectomized males (Orx+T) significantly reduced the changes in muscle withdrawal thresholds that was different from males receiving placebo (Orx+P) on the ipsilateral side. #p<0.05, compared to females with placebo treatment; *p<0.05, compared to orchiectomy with placebo treatment

Experiment 3 – Flutamide temporarily reverses the male pain phenotype

To determine if the blockade of androgen receptors altered the sex-dependent pain phenotype in males, we administered flutamide systemically to intact males 24hr after induction of the activity-induced pain model. Prior to flutamide, withdrawal thresholds of the muscle decreased ipsilaterally but not contralaterally after induction of the model, consistent with the male pain phenotype [24]. A single subcutaneous flutamide injection (50mg/kg) had no effect on the decreased withdrawal threshold ipsilaterally when compared to animals receiving a vehicle 30 minutes after injection (F_1,8 = 2.4, p=0.15) (Fig. 3A). However, a significant decrease in MWT occurred on the contralateral side 30 minutes following flutamide injection when compared to those receiving vehicle (F_1,8 = 12.1, p=0.008). The effect of flutamide was reversed after two hours (Fig. 3B).

Figure 3.

Group differences in MWT at 30 minutes and 2 hours post subcutaneous flutamide (50mg/kg) or vehicle injection in male mice. (A) Changes in muscle withdrawal threshold following the activity-induced pain model resulted in hyperalgesia ipsilaterally. Flutamide had no effect on the muscle withdrawal thresholds. (B) 24h after induction of the activity-induced pain model had no effect on muscle withdrawal thresholds contralaterally. 30 minutes after systemic administration of flutamide there was a significant reduction in muscle withdrawal thresholds observed as a greater difference in threshold on the contralateral side which reversed after 2hr. *p<0.05, compared to vehicle at 30 minutes after injection

Experiment 4 - Orchiectomized male mice develop hyperalgesia similar to females when the muscle fatigue was applied contralateral to the muscle insult

To determine if testosterone mediates the sex-dependent effect seen when the muscle insult and fatigue are administrated in opposite muscles, we compared orchiectomized males to intact males and females in the activity-induced pain model when the insult is given in the right gastrocnemius and muscle fatigue induced in the left gastrocnemius muscle. Following induction of the model, significant group differences were found on the left (injected) side for changes in withdrawal threshold (F_2,21 = 24.4, p<0.001). Post hoc testing demonstrated significant differences between sham and Orx groups (p<0.001) and sham and female groups (p<0.001) but not between Orx and female groups (p=0.075) (Fig. 4). On the right (fatigued) side, similar significant group differences were found for changes in withdrawal threshold (F_2,21 = 12.5, p<0.001) with post hoc testing revealing significant differences between sham and Orx groups (p=0.04) and sham and female groups (p<0.001) but not between Orx and female groups (p=0.06) (Fig. 4). Taken together this data demonstrate hyperalgesia is more easily induced in orchiectomized mice similar to females.

Figure 4.

Group differences in MWT 24hr following induction of spatial or temporal separation paradigms of activity induced pain model. In both paradigms, orchiectomy of males (Orx) 2 weeks prior to induction of pain model resulted in decreases in withdrawal thresholds compared to baseline bilaterally similar to females. *p<0.05, compared to sham orchiectomy group (sham)

Experiment 5 - Orchiectomized male mice develop hyperalgesia similar to females when the muscle fatigue is applied 24hr prior to the muscle insult

To determine if testosterone mediates the sex-dependent effect seen when muscle fatigue is induced 24hr before the second muscle insult, we compared orchiectomized males to intact males and females in the activity induced pain model where the muscle fatigue was induced 24hr prior to the second intramuscular acidic saline injection. Following induction of the model, significant group differences found on the ipsilateral side for changes in withdrawal threshold (F_2,21 = 20.2, p<0.001) with post hoc testing demonstrating significant differences between sham and Orx groups (p<0.001) and sham and female groups (p<0.001) but not between Orx and female groups (p=0.36) (Fig. 4). On the contralateral side, similar significant group differences were found for changes in withdrawal threshold (F_2,21 = 14.5, p<0.001) with post hoc testing revealing significant differences between sham and Orx groups (p=0.006) and sham and female groups (p<0.001) but not between Orx and female groups (p=0.17) (Fig. 4). This data again demonstrate it is easier to induce hyperalgesia in orchiectomized males similar to that of females.

Experiment 6 – Activity-induced pain alters SERT expression in NRM and dorsal raphe

To determine if there were sex differences in SERT staining following induction of the activity-induced pain model, we compared SERT immunoreactivity in various brain regions. There was a significant difference in SERT density in the NRM for sex (F_3,16 = 9.93, p=0.006), but not for the pain model (F_3,16 = 0.42, p=0.52). However, there was a significant interaction between the pain model and sex (F_3,16 = 5.9, p= 0.027). Post hoc testing revealed females in the pH 5.0 group had significantly higher levels of SERT (0.051±0.01) when compared with males in the pH 5.0 (0.022±0.01)(p=0.005) (Figure 5A).

Figure 5.

Group differences in SERT staining density in the raphe magnus (NRM) and dorsal raphe following induction of activity-induced pain model between males and females. Females who underwent the pH 5.0 model had significantly higher SERT levels than males who underwent the pH 5.0 model in the NRM (A&B). Males and females who underwent the pH 5.0 model had significantly reduced SERT when compared with males and females who underwent the pH 7.2 model in the dorsal raphe (C&D). *p<0.05, compared with pH 5.0 males; ^#p<0.05, pain model differences

In the dorsal raphe, there was a significant effect for the pain model for SERT density (F_3,14 = 11.8, p=0.004), but not for sex (F_3,14 = 1.37, p=0.26). There was no interaction between the pain model and sex (F_3,14 = 1.5, p=0.23). Animals in the pH 5.0 group (M:0.021±0.02; F:0.021±0.01) showed a lower density of SERT immunoreactivity when compared to those in the pH 7.2 group (M:0.044±0.03; F:0.069±0.01)(Figure 5B). No significant differences were found at any other brain site (Supplemental Table 1).

Discussion

The current study shows testosterone mediates the sex-dependent phenotype observed in an activity-induced pain model. Removal of testosterone by orchiectomy in males produced bilateral, longer-lasting hyperalgesia similar to females. Administration of testosterone to orchiectomized males or females produced unilateral, shorter-lasting hyperalgesia similar to intact males. In low testosterone groups, bilateral hyperalgesia developed when the muscle insult and fatigue were administered in opposite muscles or when the muscle insult and fatigue were 24h apart; testosterone prevented development of bilateral hyperalgesia. The current data, together with prior data showing ovariectomy had no effect on hyperalgesia in this model, suggests testosterone mediates the pain phenotype in the activity-induced pain model [24]. Lastly, the current study showed females have significantly higher levels of SERT in the RVM when compared to males following induction of the activity-induced pain model. The current study is the first to show testosterone impacts development of widespread muscle pain suggesting that testosterone could protect against development of widespread pain and potentially modulate SERT expression in the NRM. The fact that hyperalgesia is easier to induce, bilateral, and longer-lasting in animals with lower testosterone could explain why a higher percentage of women develop widespread pain when compared to males [22;63].

The current study is consistent with prior literature showing that testosterone is protective in animals models of pain including inflammatory, formalin-induced, and stress-induced pain [21;27]. For example, hyperalgesia in orchiectomized male rats, after formalin injection into the temporomandibular joint, is enhanced and similar to females; testosterone protected females and rescued the phenotype in orchiectomized male rats [21]. Similarly, the longer-lasting and greater hyperalgesia observed in female rats to a stress-induced pain model rats is altered by modulation of testosterone with orchiectomy of males and administration of testosterone to females [27]. Contrary to our prior work, ovariectomy of female mice produced a pain response which was similar to intact males in a stress-induced pain model [27]. Thus, prior studies consistently show testosterone is protective against development of hyperalgesia.

While some studies show a sex-dependent phenotype, others show no difference in the phenotype in a variety of animal models of pain including neuropathic and inflammatory [48;57]. The reason for these differences is unclear and could be related to the animal model, the behavioral test used, or different underlying mechanisms producing the same phenotype. Several more recent studies show different underlying mechanisms for the observed hyperalgesia. Classically sex-dependent effects are thought to be a result of either organizational or activational effects. Organizational effects are permanent changes produced by gonadal hormones during development. Higher amounts of mu-opioid receptors in the PAG and RVM account for greater morphine-induced analgesia in males compared to females [39;41;42]. Orchiectomy of neonatal male mice or testosterone administration to neonatal female mice produces the organization effects of the opposite sex, while orchiectomy and testosterone administration in adult mice has no effect on of mu-opioid receptor expression [7;14]. Thus, testosterone during development alters the central nervous system to protect male mice from development of hyperalgesia.

Activational effects are transient, can acutely modify a variety of systems, and can result in a phenotypic difference or alter underlying mechanisms explaining the same phenotype [1]. The current study modulated testosterone in adult mice and therefore examined if activational effects were responsible for induction of a phenotypic sex difference. Several studies show the immune system is differentially activated in male and female mice. In males, testosterone mediates involvement of microglia in the spinal cord after inflammatory and neuropathic insults leading to hyperalgesia through a pathway that is dependent on TLR4 activation, purinergic receptors or p38 mitogen-activated protein kinase signaling [48;57;58]. While female hyperalgesia is produced by an alternative pathway activating the adaptive immune system [58]. Further, administration of testosterone in females activates the spinal TLR4 pathway similar to males while orchiectomy of males results in a pain mechanism that is independent of spinal TLR4 activation [57;58]. Synovial plasma extravasation in response to bradykinin injection is greater in males, and ovariectomy of males or administration of testosterone to females reverses this sex-dependent difference [23]. Lastly, sex steroids have immunomodulatory properties with testosterone increasing release of anti-inflammatory cytokines (interleukin-10) and decreasing release of pro-inflammatory cytokines (interleukin-1β, interleukin-6) from macrophages and lymphocytes [2;38;69]. In the peripheral nervous system, activational effects of testosterone mediate transcriptional upregulation of the receptors involved in inhibition of nociception, mu-opioid and cannabinoid type-1, on peripheral nociceptors [34;35;47;70]. Thus, testosterone can modulate protective mechanisms in the immune system and the peripheral and central nervous system.

Testosterone is also protective in experimental and clinical pain studies in humans. In experimental studies, lower circulating levels of testosterone are associated with higher rated pain, unpleasantness, and anxiety in response to noxious stimuli [4;11–13]. Pain-free individuals with lower testosterone levels have less activation of the RVM in response to noxious heat applied to forearm [66], and higher activation of the pregenual anterior cingulate and orbitofrontal cortex, areas associated with negative affect component of pain, in response noxious cold water finger submersion [13]. Also, lower testosterone levels are associated with greater symptom severity in fibromyalgia and rheumatoid arthritis [18;50;51;53;61]. Few studies have examined the impact of testosterone administration in clinical pain populations. Of note, four-weeks of transdermal testosterone administration to women with fibromyalgia improved muscle pain, stiffness, and fatigue [68]. Thus, testosterone modulates activity in areas involved in pain inhibition, and reduces pain responses in experimental and clinical conditions.

It is difficult to compare circulating testosterone values to prior literature because variables such as animal strain, age, and experimental techniques can impact values [9;28;33]. However, prior literature reports similar testosterone values for C57BL/6 females, gonadally intact males, and orchiectomized males [9;19]. The current study measured testosterone values 19 days after implanting subcutaneous slow-release pellets or orchiectomy since it is the same time we induce the activity-induced pain model. It is unknown if these values were consistent throughout our testing paradigm; it should be noted that slow-release pellets were maintained through the duration of the experiment. After orchiectomy, testosterone values show significant and similar reductions for six weeks following surgery [56]. Interestingly, in the current study testosterone pellet implantation produced supraphysiological levels of testosterone in males and females but had no effect on the ipsilateral hyperalgesia. We speculate that supraphysiological levels are not needed to see the positive effects of testosterone since intact males do not develop secondary hyperalgesia and blockade of androgen receptors with flutamide in intact males produces a transient contralateral hyperalgesia; however, this is a limitation of the current study. It is possible that blockade of androgen receptors could lead to increased aromatization of testosterone into estradiol to cause this effect; however, a prior report shows that flutamide does not change estradiol levels [59]. Together this suggests testosterone protects against development of secondary hyperalgesia. The lack effect on the ipsilateral side could suggest that the ipsilateral hyperalgesia uses a non-testosterone mechanism or could reflect a ceiling effect on the hyperalgesic response.

The current study showed increases in SERT in the NRM following induction of the activity-induced pain model only in females. The RVM is a key location in brainstem that inhibits and facilitates nociception [52;62;64]. We previously show increases in SERT in the RVM, and blockade of SERT in the RVM reduces hyperalgesia, in models of neuropathic and widespread muscle pain [5;8]. We also showed voluntary wheel running increases SERT expression in the dorsal horn of the lumbar spinal cord; however, there was no increase following induction of a non-inflammatory muscle pain model [8]. Testosterone and estradiol modulate SERT levels in brain locations such as the dorsal raphe, amygdala, and hypothalamus; however, it is unclear if sex hormones have similar effects in the RVM, or if pain modulates these effects [25;43;44]. Thus, increases in SERT seen in females but not males could be responsible for the development of the sex-dependent widespread hyperalgesia. Further work will confirm that testosterone modulates SERT levels in the NRM in models of pain.

We also show a decrease in SERT expression in the dorsal raphe. Serotonergic responses to noxious stimuli are heterogeneous, with decreases and increases observed in different regions [31]. These differences might be related to the ascending and descending components of the serotonergic system [10;17;54]. The dorsal raphe is part of the ascending nociceptive pathway involved in affective processing [17] with increased serotonin expression associated with nociception. Local application of 5-HT into the dorsal raphe of rats produces a nociceptive response [16] and neuropathic pain increases serotonin in dorsal raphe neurons of mice [26]. Our data are consistent with prior data since a decrease in SERT would be reflected as an increase in 5-HT. Future work will need to elucidate the sex differences in serotonergic modulation in the ascending and descending pathways.

In summary, the current study demonstrated the role of testosterone in mediating the sex-dependent pain phenotype seen in an animal model of activity-induced widespread muscle pain. Administration of testosterone produced a pain phenotype which was unilateral and short-lasting regardless of sex. This demonstrates the role of testosterone in activational changes to protect against development of chronic pain.