Subjects
Sprague-Dawley rats identified as female and male by external genitalia were purchased from Envigo Labs (Livermore, CA), and upon arrival were housed in same-sex pairs under a 12:12 hr light:dark cycle for approximately one week prior to beginning the experiment. Thereafter, rats were housed singly, in one of two test rooms (12 cages/test room), with lights off at 1400 or 1430 depending on the test room. Rats were tested in seven cohorts of 12 rats/cohort.
Rats were 47–61 days old when each was singly housed in a home cage with a running wheel. Each rat remained in this cage with ad libitum access to food and water except during weighing, vaginal lavage, capsule implantation, saline/TNBS instillation, and animal husbandry. All manipulations except surgery were conducted during the hour before the dark phase began; surgery was conducted during a 3-h period before the dark phase began. All procedures were conducted in accordance with the NIH Guide for Care and Use of Laboratory Animals[30], and were approved by the Washington State University Institutional Animal Care and Use Committee.
Apparatus
Each rat had continuous home cage access to a running wheel (Tecniplast, Starr Life Sciences Corp, PA). The wheel had a circumference of 1.04 m and was mounted to hang from the top of a standard rat cage. The number of wheel revolutions was recorded continuously in 5-min bins except for the last hour of the light phase, when no wheel running data were collected so that animal husbandry and experimental manipulations could be conducted. The number of wheel revolutions was recorded using VitalView software (Starr Life Sciences Corp, Oakmont, PA) on a computer located in a room adjacent to the testing room.
Hormones and drugs
Crystalline testosterone (T) was purchased from Steraloids (Newport, RI); T-filled and blank Silastic capsules (0.062 in. i.d./0.125 in. o.d.) were constructed in-house, in 5-mm and 10-mm lengths as previously described[31]. 2, 4, 6-Trinitrobenzene sulphonic acid (TNBS, Millipore Sigma, St. Louis, MO, USA) 30 mg/ml was prepared in a 1:4:5 ratio of TNBS:saline:ethanol. Sterile physiological saline served as the vehicle control.
Surgeries
To model T exposure as used in gender-affirming T therapy – in which the goal is to achieve and maintain cisgender male levels of T[32], typically in people who retain their ovaries – constant-release T-filled or blank capsules were implanted in gonadally intact female rats at doses that maintain reproductive behaviors and organ weights in orchidectomized male rats similarly to those in gonadally intact male rats[31, 33]. Rats (57–71 days old) were anesthetized with an i.p. injection of ketamine (90 mg/kg) and xylazine (10 mg/kg) (MWI Animal Health, Visalia, CA). Meloxicam (1 mg/kg; Patterson Veterinary Supply) was administered s.c. as a pre-operative analgesic. Constant-release, Silastic capsules were implanted s.c. between the shoulder blades based on body weight (approximately one 10-mm capsule/100 g body weight): Rats weighing 160–210 g were implanted with two 10-mm capsules; rats weighing 211–235 g were implanted with two 10-mm and one 5-mm capsules. Half of the female rats were implanted with capsules containing T and the other half received blank capsules (female control group). Each male was implanted with two blank capsules.
Behavioral procedure
The experimental timeline is illustrated in Fig. 1A. Rats were acclimated to single-housing with a running wheel for 10 days. Wheel running gradually increases during this habituation period, with a majority of wheel running occurring during the dark phase[34]. Baseline running was defined as the average number of wheel revolutions during the dark phase over the last 3 days of the 10-day habituation period. A matching protocol was then used to assign female rats to each of the 4 female treatment groups (Blank + Saline, Blank + TNBS, T + Saline, T + TNBS), and to assign males to each of the 2 male groups (Blank + Saline, Blank + TNBS), so that mean baseline wheel running was comparable across treatment groups within each sex, at the start of the experiment. Rats were weighed daily in g throughout the experiment.
Near the end of the 3-day baseline period (within the last few hours of the light phase on the third baseline day), rats were anesthetized and blank or T-filled capsules were implanted s.c. (see Surgeries). Rats were returned to their home cages, and wheel-running was recorded 23 h/day for the next 3 weeks (“Capsule Implant Phase”).
After the last day of the Capsule Implant Phase, rats were anesthetized with isoflurane during the last hour of the light phase for instillation of 0.6 ml sterile saline or TNBS, administered via PE60 tubing inserted 7 cm into the distal colon[27]. Rats were returned to their home cages, and wheel-running was recorded 23 h/day for the next 10 days (“Visceral Pain Phase”). Any rat that lost more than 20% body weight during this phase was euthanized.
Following the Visceral Pain Phase, within the first two hours of the dark phase, rats were euthanized with isoflurane and trunk blood was collected and centrifuged for 20 min at 3200 rpm. Serum was collected from all rats that completed the 41-day study, and stored at -80˚C for later determination of hormone levels. Capsules were removed to confirm number and type (blank or T), and the clitoral glands (females) or preputial glands (males) were removed and stored in 10% formalin for a minimum of 2 weeks before trimming and measuring length and width of the right and left glands by an experimenter blind to treatment group assignment.
Estrous cycle monitoring
For 8–9 female rats in each of the four treatment groups, vaginal lavage was conducted daily for up to 21 days, starting 1.5-2 weeks after capsule implantation and continuing 4–10 days into the Visceral Pain Phase. Slides were air-dried and later stained with Giemsa (Sigma Aldrich). Estrous stage was determined via microscope by a reader who was blind to treatment group assignment: proestrus was defined as approximately 75% or more of cells in the sample being nucleated epithelial cells; estrus was defined as approximately 75% or more of cells in the sample being cornified epithelial cells; diestrus was defined as an approximately equal distribution of nucleated and cornified epithelial cells plus leukocytes (diestrus day 1, also known as metestrus), or primarily leukocytes (diestrus day 2)[35].
Hormone analysis
Serum T and estradiol levels were determined in duplicate, using ELISA kits (IB79106: Immuno-Biological Laboratories, Inc., Minneapolis, MN, and 11-ESTHU-E01: American Laboratory Products Company, Salem, NH) according to the manufacturer’s protocol by a technician blind to treatment group assignment.
Data analysis
To quantify estrous cycling, percent time in proestrus and estrus was estimated as: # proestrus and estrus samples / number of sample days x 100. Percent time in proestrus and estrus was then compared among the female groups (blank vs. T) by ANOVA, with TNBS (saline vs. TNBS) entered as a co-variate, because TNBS might be expected to disrupt cycling but was not a primary variable of interest for this analysis.
Clitoral and preputial gland sizes were estimated by calculating the mean length and width of the right and left glands for each rat, and then calculating the area (mean length x mean width, in mm2). Because glands were expected to scale by size of the rat to some degree, gland area was also adjusted by final body weight, for each rat. Unadjusted and body weight-adjusted gland area were each compared among treatment groups (female, blank; female, T; male blank) by ANOVA, with TNBS entered as a co-variate. Glands were harvested from all rats except one female in the T + TNBS group.
Most wheel running occurs during the dark phase (Kandasamy et al., 2016), so only the number of wheel revolutions during daily 12-h dark phase periods was analyzed. The initial baseline was the mean number of dark phase wheel revolutions over the last 3 days of the 10-day habituation period, for each rat. One to four days before the end of the study, several rats were euthanized due to excessive weight loss after TNBS instillation (3 Females, Blank; 2 Females, T; 1 Male); in these cases, missing body weight and wheel running values were replaced with those obtained on the last day each rat was alive. Body weight and wheel running data were analyzed in two phases. Data during the 3-week Capsule Implant Phase were analyzed by ANOVA, with factors of Treatment Group (Females, Blank; Females, T; Males, Blank) and Day (repeated measure; baseline + 21 days post-capsule implant). Significant treatment group differences were followed by planned comparisons to test for sex differences (females implanted with blank capsules vs. males implanted with blank capsules), and to test for a T effect between the two female groups. Because there were group differences in body weight and wheel running at the end of the Capsule Implant Phase, body weight and wheel running data during the subsequent 10-day Visceral Pain Phase were converted to percent of baseline, for each rat; the second baseline was the mean body weight of the last 2 days, or mean number of dark phase wheel revolutions of the last 3 days of the Capsule Implant Phase. Percent baseline body weight and percent baseline wheel running were each analyzed by ANOVA with factors of Treatment Group, TNBS, and Day (repeated measure).
The mean of the two duplicates was calculated for each serum hormone sample. Hormone levels then were compared among the three treatment groups via 2-way ANOVA, with factors of Treatment Group and TNBS. To determine if hormone levels at the end of the study were associated with TNBS-suppressed body weight and wheel running during the 10-day visceral pain phase, mean body weight and mean dark phase wheel running during the 10 days after TNBS instillation were calculated for each rat (using percent of baseline values), and the associations between these values and T and estradiol levels were tested using Pearson correlation analyses.
It should be noted that in the case of serum estradiol, at least 1 of the 2 duplicates from 27 (of 73 total) samples yielded estradiol levels below the minimum level of detection (10 pg/ml). The proportion of samples with at least one duplicate below the minimum level of detection did not differ among treatment groups, but data were also analyzed after imputing mean values that were below the level of detection, using the minimum detectable concentration of 10 pg/ml (see 3.3.).
To determine whether the frequency of TNBS-related morbidity, and the frequency of estradiol values below the level of detection differed among the three treatment groups, non-parametric Pearson Chi Square tests were used, since these variables were nominal and neither continuous nor normally distributed.
SPSS version 29 was used for analyses. Post-hoc comparisons were conducted using Tukey’s test, or a Bonferroni-corrected t-test to compare groups on 22 days. For all repeated measures ANOVAs, Mauchley’s test of sphericity was used to test for homogeneity of variance; if this assumption was violated and Greenhouse-Geisser-adjusted p values were > 0.05 (i.e., if the unadjusted p value was p ≤ 0.05 but the adjusted p value was > 0.05), then adjusted df, F, and p values are reported. Partial eta squared (ηp2) values are provided as effect size estimates; ηp2 values of 0.01, 0.06, and 0.14 are considered small, medium, and large, respectively[36].