A Mouse Model to Investigate the Impact of Gender Affirming Hormone Therapy with Estradiol on Reproduction

Abstract

Study question:

Can the diverse reproductive outcomes of patients receiving estradiol (E) for gender-affirming hormone therapy (GAHT) be replicated in a mouse model?

Summary answer:

Implanting adult male mice with capsules containing 1.25, 2.5 or 5 mg of E results in overall and dose-dependent effects on reproductive hormones and anatomy and thus permits examination of E-GAHTs diverse impact on reproductive outcomes.

What is known already:

Prior examinations of E’s effects in male rodents utilized acute treatments at specific developmental timepoints or utilized gonadectomized animals; however, E-GAHT treatment is often given long term to adults before gonadectomy. Available E-GAHT rat models do not replicate the hormonal profiles of E-GAHT patients.

Study design, size, duration:

Adult male B6129S mice were implanted subcutaneously with capsules containing 1.25, 2.5 or 5 mg E powder and a control group received blank implants (n=5 mice/group). Treatment lasted six weeks involving both longitudinal and cross-sectional assessments.

Participants/materials, setting, methods:

Blood samples and genital images were collected from 6–8 week-old male mice prior to implanting capsules containing 1.25, 2.5 or 5 mg E powder or left blank. Weekly blood collections began after implantation while genital images were taken again at three and six weeks post-implantation. Implants were left for six weeks then animals sacrificed. Terminal measures included steroid hormone and gonadotropin levels, body/organ weights, seminiferous/epididymis tubule histology and gonad cellular morphology, including spermatogenesis.

Main results and the role of chance:

Compared with controls, E treated mice had suppressed testosterone (T) levels two weeks after implantation and genital changes by the third week, and these differences remained throughout treatment. The gonads of mice given 5 mg E implants weighed less than controls while E decreased seminal vesicle weight in all E-GAHT groups. E treatment elevated terminal E and decreased FSH in all E-GAHT groups but only the 5 mg E capsules significantly decreased LH from control levels. E-GAHT treatment did not alter the presence of mature spermatocytes in the cauda epididymis/vas deferens, though E-exposed sperm displayed altered motility. Negative impacts of E treatment on epididymis tubule size were similar across all E-GAHT groups, while dose-dependent effects appeared when examining seminiferous tubule morphology and bladder weight.

Limitations, reason for caution:

Mouse models can be powerful tools for informing future explorations but may not entirely match human physiology.

Wider implications of the findings:

Transgender and gender-diverse individuals identify E-GAHT’s effects on reproductive health as a research priority. This study represents the first mouse model that facilitates examinations of E-GAHT’s reproductive impact. Reproductive phenotype and effect on reproductive hormones varied by E dose in many instances, but reassuringly mature spermatocytes were seen at all doses. The dose-dependent effects of this model permit examinations of diverse patient outcomes (reproductive and otherwise) while future work may manipulate the duration of treatment or determine the reversibility of outcomes.

Graphical Abstract

Mouse models of gender-affirming hormone therapy (GAHT) are currently limited to testosterone GAHT while available estradiol (E)-GAHT rat models don’t replicate hormonal profiles of E-GAHT patients. Here, intact male mice were subcutaneously implanted with capsules containing one of three E doses, or left blank, for six weeks. Reproductive and endocrine outcomes were comparable to clinical data from E-GAHT patients.

Introduction:

Gender-affirming hormone therapies (GAHT) are potentially life-saving medical interventions known to improve mood and mental health outcomes (Coleman et al., 2012, Fishman et al., 2019) for many of the 1.2 million transgender and/or gender diverse (TGD) people in the United States (Flores, 2016). This consistent and growing population (Meerwijk and Sevelius, 2017) faces unique medical challenges to their reproductive systems that are currently informed by case reports alone (Matoso et al., 2018). Currently, TGD patients on GAHT must make reproductive health decisions based on limited data (ASRM, 2015). Such potential to change individuals’ lives warrants research that is responsive and reflective of their lived experiences and, through community-based studies, the TGD community identifies GAHTs influence on reproduction as a top research priority (Ross, 2021). While randomized controlled trials on reproductive effects of GAHT are not possible in a clinical setting, translational research offers a powerful alternative by allowing for controlled experiments in non-human models. Indeed, translational research on GAHT patient needs is an exciting and growing field (Bartels et al., 2021, Kinnear et al., 2021) using reverse translational methods to recapitulate clinical findings for GAHT. To date, researchers have published the reproductive impact of GAHT with testosterone (T) in mice (Goetz et al., 2018, Kinnear et al., 2019) and rats (Tassinari et al., 2023) while the impact of GAHT with estradiol (E) on these same systems is limited to rat studies (Alexander et al., 2022).

E-GAHT patients take E orally, intramuscularly, or transdermally, leading to increased circulating levels of E and suppression of the natal hypothalamic-pituitary-gonadal axis (HPG) by providing negative feedback to the system (Fishman, et al., 2019, Herndon et al., 2023). E may act on both the hypothalamus (Chakraborty et al., 2005, Smith et al., 2005) and pituitary (Bagatell et al., 1994), leading to lower gonadotropins and therefore lower T production. Both T (O’Donnell et al., 1994, O’Donnell et al., 1996) and E (Kula et al., 2001, Schulster et al., 2016) influence spermatogenesis so it is unsurprising that E-GAHT significantly impacts reproduction (Andrews et al., 2022, Matoso, et al., 2018); however, these changes are poorly understood and highly variable. For example, researchers examining the testicles of E-GAHT patients after orchiectomy found complete spermatogenesis in some (Venizelos and Paradinas, 1988) while another study found arrested spermatogenesis in all patients (Lu and Steinberger, 1978). Morphological outcomes with variable incidence also exist: decreased seminiferous tubule diameter (Rodriguez-Rigau et al., 1977) and tubule hyalinization (Venizelos and Paradinas, 1988), thicker and smaller epididymal tubules (Sapino et al., 1987) and increased interstitial medium (Barreno et al., 2020, Kisman et al., 1990). Such thickening of the reproductive tract is known to disrupt reproductive capacity (Enders et al., 1995) and altered fertility is likely compounded by the maturational arrest of spermatocytes seen in some E-GAHT patients (Lübbert et al., 1992). One study found E-GAHT patients had Leydig cell histology comparable to cisgender men (Payer et al., 1979), but cells consistent with Leydig cell morphology were absent in another study population (Schulze, 1988). E-GAHT’s impact on HPG function is also highly variable as the negative feedback from E alone may be insufficient to completely suppress natal HPG function (Gava et al., 2016, Valenta et al., 1992). These studies varied in the dosage and type of E used, duration of treatment, and pre-surgical procedures, which likely contribute to the lack of uniform results (Schneider et al., 2017, Schneider et al., 2015, Thiagaraj et al., 1987). Methodological differences aside, the degree of E-GAHTs impact on the gonads appears negatively correlated with increasing levels of circulating gonadotropins i.e., an unsuppressed HPG (Sapino, et al., 1987), implicating variable sensitivity to E feedback as another potential source of diverse outcomes. Controlling for and studying how these variables impact reproductive measures in rodents may help us to understand where the heterogeneity in clinical outcomes originates and how to address the health needs of individual E-GAHT patients (Schwartz and Moravek, 2021). Our current lack of knowledge restricts access to these life-saving interventions by limiting clinicians’ capacity to offer data-driven care for E-GAHT patients (Miyagi et al., 2021). Understanding the reproductive health outcomes of E-GAHT patients represents an essential goal to pursue.

Previous rodent studies documenting suppressive effects of E on circulating gonadotropin levels in male rats (Swerdloff and Walsh, 1973) have used treatment paradigms inconsistent with E-GAHT or failed to capitalize on gonad-intact animals, thus precluding translation to the integrated HPG changes that occur in E-GAHT patients. Studies with castrate male rats may help with understanding the biological significance of E in male rodents but not the clinical significance of E for E-GAHT patients. The objective of the current study was to create a mouse model to more closely mimic the E-GAHT paradigm in humans by studying the effects of varying doses of E treatment in gonad-intact male mice.

Materials and methods:

Ethical approval:

All animal procedures were approved by the Institutional Animal Care & Use Committee (IACUC) at the University of Michigan (PRO00009635).

Experimental Animals:

Twenty 6-week-old male C57BL/6NHsd mice (Envigo, Indianapolis, IN, USA) were co-housed five to a cage (L:D, 12:12) and provided food and water ad libitum. Implants were generated by cutting 16 mm of silastic tubing then sealing one end with 4 mm of adhesive (Factor II, Lakeside, AZ). Empty implants were weighed before E powder (Sigma Aldrich, St. Louis, MO) was dissolved in ethanol then pipetted into the open end of the tube; control tubes were filled with ethanol alone. After the ethanol evaporated, implants were weighed to ensure correct dose of E and additional E in ethanol was added if necessary, or the capsule remade if heavier than expected. After a final ethanol evaporation, the open end was sealed with 4 mm adhesive. At 8 weeks of age, mice (n=5/group) were implanted (supplemental figure 1) with a capsule filled with either 1.25 mg, 2.5 mg or 5 mg of E powder or a control capsule left blank. Silastic capsules were left implanted in the mice for 6 weeks, then the mice were euthanized and tissue collected for hormone and histological analysis. Sample size and study duration were determined by previous GAHT dose studies in mice (Kinnear, et al., 2019).

Blood Collection and Hormone Analysis:

One blood sample was collected from the lateral tail vein one week prior to implants and weekly draws began the week after (volume ≤ 0.5% body weight), alternating group order. Terminal cardiac blood was drawn after six weeks of treatment. Blood was kept at 4ᵒC for up to 24 hours before collecting serum and storing at −20ᵒC. Hormone analysis was performed by the Ligand Assay and Analysis Core Facility, University of Virginia Center for Research in Reproduction. Sensitivities for each assay were as follows: Testosterone 10 ng/dL (IBL ELISA, Minneapolis, MN), Estradiol 5 pg/mL (ALPCO ELISA, Salem, NH), Luteinizing Hormone 0.04 ng/mL (In house radioimmunoassay (Haavisto et al., 1993)) and Follicle Stimulating Hormone 3 ng/mL (In house radioimmunoassay (Gay et al., 1970)).

Body Weights and Measurements:

One week prior to implant placement, and every other week thereafter, mice were anesthetized and genital examinations performed. Given the need to palpate and hold structures in place, variation in pressure and position precluded precise quantitative analysis so several qualitative characteristics were established. Briefly, mice were weighed, their testes palpated from the abdominal cavity into the scrotum and the presence of a recognizable epididymis noted (Boersma et al., 2015). The external prepuce was retracted to examine the penis and whether the sheath recovered automatically or the penis required external pressure to return it to the sheath was noted.

Sperm Morphology and Reproductive Histology:

After 6 weeks of treatment, mice were euthanized and the epididymis was separated from one testicle, placed in saline and the testis weighed alone. The caudal portion of the epididymis was scored with a 31-gauge needle to release maturing spermatozoa. The morphology and movement of the spermatozoa were observed using a Makler Device (Sefi-Medical Devices, Israel). The seminal vesicles, penis, bladder (urine expelled) and right kidney were blotted and weighed. The remaining testicle was removed with the epididymis intact and placed into Bouin’s solution (Ricca Chemical, Arlington, TX) overnight, washed with ethanol dilutions, then stored in 70% ethanol before blocking in paraffin. The testis and epididymis were sectioned serially at 5 um and the number of seminiferous tubule cross-sections containing any vacuolitic Sertoli cells counted and epididymis tube cross-sectional areas analyzed (Weissbach and Ibach, 1976). For the testis, 5 random sections at least 120 um apart were stained with hematoxylin (Epredia, Kalamazoo, MI) and eosin (Ricca Chemical, Arlington, TX; H&E) then the number of cross-sections with vacuolitic Sertoli cells were counted in each section. Due to the limited number of seminiferous tubules within the testis, it is likely the same tubule was counted multiple times. However, the minimum distance between sections ensures that a new pool of Sertoli cells was analyzed in each section. For the cauda and caput epididymis, ten random sections were stained with H&E, imaged at 5X and the cross-sectional area of the afferent and efferent regions of the tube and its lumen were measured. A grid of four quadrants was placed on each image, one cross-section was measured in each and a fifth measured at the center. All cross-sectional areas were measured in sections containing four or less. These methods provided data from between 20 and 25 cross-sections of the epididymis tubule in its cauda or caput regions. While each cross-section is from the same tubule, a minimum distance of 70 um between sections ensured different epithelial cells were measured each time. Histological measures were taken by observer blind to treatment.

Statistical Analysis:

Data were analyzed with PRISM (v9.1.0) software (α=0.05). Differences between the four experimental groups showing normal distributions were analyzed with ANOVA followed by posthoc pairwise comparisons adjusted for multiple comparisons using Tukey’s method, which provided adjusted p-values. Hormone concentration values for T, E and LH were log transformed prior to statistical analysis based on significant Bartlett’s and Welch’s tests. No animals were excluded from analyses.

Results:

E-GAHT Treatment Suppressed Testosterone:

Prior to implants, no differences in serum T levels were seen between the four groups (Means and SD for all values found in supplemental Table I; Fig 1A, F_{(3, 16)}=0.019; p>0.99). E-GAHT treatment resulted in suppression of T levels from 192.8 ± 229.6 ng/dL before implantation to 47.2 ± 19.8 ng/dL one week and 40.9 ± 14 ng/dL two weeks after (averaged across all doses). While a significant main effect was evident one week post-implant (Fig 1A; F_{(3, 16)}=3.7; p=0.001), posthoc analysis indicated only a trend towards lower T in the 5 mg group (p=0.055) compared with controls. At week two, the significant effect of treatment (F_{(3, 16)}=8.3; p=0.0015) was apparent across all E-GAHT groups (p≤0.01). This pattern persisted across week three (p≤0.008), week four (p≤0.03) and week five (p≤0.03). Terminal levels confirmed E-GAHT suppression of T (Fig 1B; F_{(3, 16)}=40.8; p<0.0001) until the end of the experiment at week six (p<0.0001). No significant differences in T levels between E doses were evident at weekly timepoints or terminal levels.

Figure 1:

Hormone Measurements. A) Weekly and B) terminal testosterone levels. C) Terminal estradiol levels. D) Terminal follicle stimulating hormone levels. E) Terminal luteinizing hormone levels. Points represent individual animal values (n=5/group). Means ± SD, *p<0.05, **p<0.01, ****p<0.0001.

E-GAHT Increased Estradiol and Suppressed Gonadotropin Levels:

E-GAHT treatment increased terminal E (Fig 1C; Means and SD in supplemental Table I; F_{(3, 16)}=5.8; p=0.007) at all doses (p≤0.04). E-GAHT also significantly impacted the gonadotropins FSH (Fig 1D; F_{(3, 16)}=52.8; p<0.0001) and LH (Fig 1E; F_{(3, 16)}=5.49; p=0.009) with suppression evident for FSH in all dosage groups (p<0.0001) but only with the highest dose for LH (Fig 1D; p=0.006). The average T level for three adult female mouse samples sent for analysis as standards was 37.9 ng/dL and their average E level was 25.9 pg/mL.

E-GAHT Induced Anatomical Changes:

Genital anatomy examined prior to E implantation had no identifiable differences between groups. Overall, E-GAHT treatment diminished external anatomical evidence of the cauda epididymis and altered prepuce and/or penile morphological characteristics (Fig 2). One week following implants, the cauda epididymis (Fig 2A,D) was no longer visible in one mouse given high E-GAHT and three mice given the mid dose. From three weeks on, all mice in the high and mid dose groups and four mice in the low dose group no longer had visible cauda epididymis. After six weeks, one mouse given the low dose had at least one visible cauda epididymis (Fig 2B,E) while they were absent in all other E-GAHT mice (Fig 2C,F). The cauda epididymis of each control animal was apparent at all timepoints. Prior to implantation, the external prepuce automatically covered the penis after it was released following retraction (Fig 2G). At three weeks, the penis of all mice that were given mid doses and four mice given the high doses required external pressure to return to the sheath. By week five, all E-GAHT treated mice had a sheath that would not recover after retraction (Fig 2H,Ia), but retracted after applying pressure (Fig. 2lb). E-GAHT significantly impacted the combined weight of the testis and epididymis (Fig 3A,B; Means and SD for all values found in supplemental Table II; F_{(3, 16)}=4.58; p=0.02). Compared with controls, testis plus epididymis from mice implanted with high E-GAHT weighed less (p=0.012) while weights in the mid and low dose groups were not significantly altered. When weighed alone, testis showed a significant main effect of treatment (Fig 3C, F_{(3, 16)}=7.6; p=0.002) and posthoc analysis indicated testis weight was reduced in the high E-GAHT group compared with controls (p=0.002). Treatment also decreased seminal vesicle weight (Fig 3D,E; F_{(3, 16)}=546.4; p<0.001), all mice given E-GAHT had lighter seminal vesicles than controls (Fig 3E, p<0.001). No significant effect of treatment was seen on penis weight (F_{(3, 16)}=0.31; p=0.82), kidney weight (F_{(3, 16)}=1.73; p=0.2), nor terminal body weights (Supplemental Table II; F_{(3, 16)}=0.08; p=0.97). However, E-GAHT increased bladder weight (Fig 3F,G; F_{(3, 16)}=52.8; p<0.001) and all E-GAHT groups had heavier bladders than controls (p<0.001). Finally, E-GAHTs effect on bladder and testis weight were dose-dependent. Mice given high E-GAHT had a greater increase in bladder weight than those given mid (p=0.028) and low E-GAHT (p<0.001) while high E-GAHT mice had lower testis weights compared with the mid dose (p<0.01).

Figure 2:

Genital Anatomy. A) The cleft formed at the distal scrotum (yellow arrowheads) by the internal cauda epididymis in controls is B) reduced or C) absent in E-GAHT mice. D) The protrusions formed by the cauda epididymis (yellow arrows) are also E) reduced or F) absent in E-GAHT mice. (G) The penis of control mice is automatically covered by the external prepuce after retraction (as seen in D-F) and release. H,Ia) The external prepuce of E-GAHT mice did not recover after retraction and release. Ib) The penis of E-GAHT mice can be returned to the external prepuce manually.

Figure 3:

Terminal Organ Appearance and Weights. A) Testicles and epididymis from control and E-GAHT mice. B) Terminal testicle and epididymis weights. C) Terminal testicle weights. D) Seminal vesicles from control and E-GAHT treated mice. E) Terminal seminal vesicle weights. F) Bladders from control and E-GAHT treated mice. G) Terminal bladder weights. Means ± SD, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. Diagrams for all figures created using BioRender.com.

E-GAHT altered testicular and epididymis histology:

E-GAHT increased the number of seminiferous tubule cross-sections containing Sertoli cells with a vacuolitic morphology (Fig 4A–J; Means and SD for all values found in supplemental Table III; F_{(3, 16)}=13.14; p=0.0001). The high (p=0.0002) and mid E-GAHT (p=0.001) groups but not the low E-GAHT group (p=0.27) had significantly higher number of cross-sections containing vacuolitic Sertoli cells than controls. No significant morphological changes in Leydig cells were observed in E-GAHT groups (Fig 4K,L), however, shed germ cells were sometimes visible, adjacent to vacuolitic Sertoli (Fig 4M). All mice had spermatocytes with both immature and mature morphology (Fig 4N); however, no spermatocytes from E-GAHT treated animals had the mature locomotion seen in controls, instead their spermatocytes displayed uncoordinated shaking. The cauda (Fig 5A–J) and caput regions (Fig 5K–T) of the epididymal tubes were also significantly modified by E-GAHT but the effect was region-specific. E-GAHT decreased tube cross-sectional area in caput regions of the tube (Fig 5E; F_{(3, 16)}=20.96; p<0.0001) but not the cross-sectional area of the lumen (Fig 5J; F_{(3, 16)}=0.83; p=0.5). On the contrary, the cross-sectional areas measured from caudal portions of the tube were independent of treatment (Fig 5O; F_{(3, 16)}=1.01; p=0.42) while the cauda lumen is increased (Fig 5T; F_{(3, 16)}=13.7; p=0.0001). Posthoc analysis indicated low, mid and high E-GAHT groups differed from controls for the caput tube area (p≤0.002) and cauda lumen area (p≤0.003). One dose-dependent histological measure was evident as mice given high E-GAHT had significantly more cross-sections with vacuolitic Sertoli cells than the low E-GAHT group (p=0.026).

Figure 4:

Testicle and Sperm Histology. A-D) Hematoxylin and eosin-stained control and E-GAHT mouse testis. E) Sertoli cells with vacuoles (V) from E-GAHT mouse. F-I) Control and E-GAHT mouse seminiferous tubules and G-I) tubules with vacuolitic Sertoli cells from E-GAHT mice (black arrowheads). J) Number of tubules with vacuolitic Sertoli cells. K,L) Leydig cells (black arrows) from control and E-GAHT mice. M) Shed germ cells (GC) in the seminiferous tubule lumen adjacent to vacuolitic Sertoli cells in E-GAHT mouse. N) Mature spermatozoa, and immature spermatozoa with excess cytoplasm (red arrowheads), from the cauda epididymis and vas deferens of controls and E-GAHT mice. Black scale bars; 5X=100 um, 20X=100 um, 40X=50 um, 63X=50 um. Means ± SD, *p<0.05, **p<0.01, ***p<0.001.

Figure 5:

Epididymis Histology. A-D) Hematoxylin and eosin-stained caput epididymis from control and E-GAHT mice. E-J) Caput epididymal tubule cross-sectional images and caput tube and lumen area K-N) Hematoxylin and eosin-stained images of cauda epididymis from control and E-GAHT mice. O-T) Cauda epididymal tubule cross-sectional images and cauda tube and lumen area. Black sale bars; 5X=100 um, 40X=50 um. Means ± SD, **p<0.01, ***p<0.001.

Discussion:

Mimicking the treatment paradigm for E-GAHT in a mouse model led to outcomes similar to those of TGD patients on E-GAHT (Table II). At all doses, E-GAHT treatment suppressed T levels significantly for four weeks to a typical female range. Terminal E was high, though still at physiological levels typical of female mice, and terminal FSH low in E-GAHT mice compared with controls. This persistent change in hormonal milieu likely contributed to the anatomical and morphological changes to the reproductive tract, along with altered spermatogenesis. E-GAHT treated mice showed expansion of the interstitial medium along the reproductive tract and tubule changes in the testicles and epididymis. Conversely, Leydig cells in E-GAHT mice appeared similar to controls. While many Sertoli cells from E-GAHT treated animals appeared normal, others took on a vacuolitic appearance. Within seminiferous tubule cross-sections containing vacuolitic Sertoli cells, germ cells were seen being shed into the lumen. Notably, hormonal and reproductive outcomes are remarkably heterogeneous in E-GAHT patient populations (Schneider, et al., 2017) and the use of multiple doses allowed us to uncover similar variation. The low E-GAHT dose had fewer or minor effects on both circulating hormones and reproductive anatomy compared with the highest dose. Only mice given high E-GAHT had significantly lower levels of LH compared with controls. Conversely, the lowest E-GAHT dose inhibited T and FSH production without significantly decreasing LH and gonad size or increasing the number of tubules with vacuolitic Sertoli cells. Clinical data indicate both the former and latter (Schneider, et al., 2017) conditions occur among E-GAHT patients. These findings are consistent with differences in E levels achieved or susceptibility to E increase as potential contributors to the variable responses seen in E-GAHT patients.

Table II:

Comparison of experimental results to clinical outcomes,

Anatomy	Previous Clinical Findings, vs. cisgender men*	Our Mouse Model Findings, vs. control males
Testis	Lower Weight Spermatogenesis: Complete to absent (High variability between patients) More Seminiferous Tubules with Vacuolitic Sertoli Cells Leydig cells: Present to absent (High variability between patients)	Lower Weight† Spermatogenesis: Complete More Seminiferous Tubules with Vacuolitic Sertoli Cells† Leydig cells: Present
Epididymis	Smaller tubules	Smaller tubule cross-sectional area
Spermatocytes	Fully mature to absent (High variability between patients)	Mature Morphology Altered Movement†
Overall Histological Findings	Expanded Interstitium	Interstitium Expanded in Epididymis
Hormones	Lower T Lower FSH Lower LH Higher E2	Lower Weekly† and Terminal T Lower FSH Lower LH† Higher E2

^†dose effect (Clinical data citations: Matoso et al., 2018; Schneider et al., 2017; Sapino et al., 1987).

Studies treating male rodents with E concur with our current findings but are not always translatable to the reproductive care of E-GAHT patients. One group found implanting adult transgenic male mice with silastic capsules containing 5 mg of E was sufficient to suppress T and increase E (Van Steenbrugge et al., 1984); however, the transgene the mice carry disrupts HPG development (Rebar et al., 1982) and the treatment led to supraphysiological levels of E (250–400 pg/mL). Still, like our E-GAHT treated mice, E-treated males had lower seminal vesicle weights and, like our high E-GAHT group, lower testicle weights (Rebar, et al., 1982). One study paradigm does follow methods similar to E-GAHT by treating adult gonad-intact male rats with E for 16 days but this did not significantly impact their FSH (Swerdloff and Walsh, 1973), which is low in E-GAHT patients (Schneider, et al., 2017) and in mice in our study. Still, rats receiving higher E doses achieved suppression of T faster than rats at lower doses (Swerdloff and Walsh, 1973), an effect we replicate with a trend the week after E implant placement. Our model offers tunable conditions to rigorously examine the reproductive and hormonal care needs of E-GAHT patients. The variability between our highest and lowest E-GAHT doses mimics heterogeneous clinical outcomes (Table II), allowing us to ask a diversity of questions related to reproductive function.

Insufficient suppression of the endogenous HPG is a common concern for E-GAHT patients (Cunha et al., 2018). Gonadotropins released by an active HPG may allow the testis to continue producing T (Maheshwari et al., 2021), preventing the desirable effects of E-GAHT. This is often addressed through the addition of antiandrogens (Angus et al., 2021), various classes of drugs that further suppress the HPG, disrupt the ability of T to interact with its receptor or suppress conversion of T to dihydrotestosterone. The addition of antiandrogens in E-GAHT paradigms also contributes to heterogeneity between E-GAHT patient outcomes (Barreno, et al., 2020, Schneider, et al., 2017). Two research groups recently asked how E and different antiandrogens impact steroid hormone-sensitive tissues and metabolic parameters in adult male rats. The first group gave E with dihydroxyprogesterone acetophenide, an antiandrogen commonly used in Brazil (Krüger et al., 2019), and included gonadectomized controls (Gusmão-Silva et al., 2022). The other study utilized male rats given both E and cyproterone acetate and found this suppressed the HPG and significantly altered reproductive tissues (Tassinari, et al., 2023). Unfortunately, both paradigms preclude examination of E’s unique impact on the integrated HPG by including antiandrogens and/or gonadectomized animals. One group treating male rats with E alone found female-typical E levels but no other hormonal measures were taken (Alexander, et al., 2022). While rats offer an important translational model to understand metabolic changes on E-GAHT, it is unknown whether antiandrogens are serving the same purpose in these models as in clinical populations. By first identifying the unique effects of E on the male mouse HPG, our model will permit future analyses of the hormonal and reproductive influence of antiandrogens at doses and timepoints relevant to patient needs. For example, one may investigate whether adding antiandrogens to our lowest adult E-GAHT dose will lead to gonad changes comparable to mice given our highest dose.

Understanding how different modes of delivery influence HPG outcomes in mice is another important addition to this line of inquiry. While using subcutaneous (SQ) silastic implants simplifies E-GAHT delivery, patients often take E-GAHT orally or through IM injections (Fishman, et al., 2019). In one study population most patients, 56%, took E-GAHT orally while IM injections were utilized by 37%, the remainder used transdermal E (Chantrapanichkul et al., 2021). However, SQ injections are considered a more desirable delivery route for GAHT patients compared with IM (Spratt et al., 2017). Though SQ E-GAHT use is rare in clinical practice, recent evidence indicates SQ E injections are an effective route (Herndon, et al., 2023). Our outcomes support this assertion, and these methods offer an appropriate pre-clinical model to investigate SQ E-GAHT. The effects of SQ E on bladder weights is of particular interest given urinary concerns are not documented for other routes. Further, our prior work with T-GAHT mice indicates T levels drop quickly after SQ implant removal but remain elevated long after stopping IM injections (Hashim et al., 2022). Some E-GAHT users pause treatment for reproductive purposes or when seeking assistive reproductive technologies, however, stopping E-GAHT may be emotionally harmful to patients and potentially unnecessary. Indeed, we found spermatocytes in the cauda epididymis and vas deferens of E-GAHT mice at stages of morphological maturation capable of fertilizing an egg (Cooper, 2011). The impact of altered sperm motility on fertility can be analyzed with natural breeding and in-vitro fertilization studies (Schwartz and Moravek, 2021). Ensuring reproductive viability, especially in relation to expensive assisted reproductive technologies, with minimal disruption to GAHT treatments would be ideal. As GAHT is a life-saving intervention for some, controlled clinical studies that include withholding treatment would be unethical. Therefore, the effect of E-GAHT cessation on mouse E levels and fertility are critical next targets for this model. Further, the TGD community has identified priorities for GAHT research that include its impact on reproductive and cardiovascular health, along with topics like efficient delivery routes and long-term effects (Ross, 2021). It is essential that GAHT patient needs remain the center of questions (Aghi et al., 2022, Goetz et al., 2023) developed for use with this and future GAHT models.

There are some limitations to our model which future research and/or additions to the model may address. Once suppressed, weekly T, terminal T and terminal E measurements from E-GAHT mice matched physiological levels typical of female mice. However, the latency and duration of E elevation is unknown as E levels were not analyzed through-out the experiment, same for our evidence of suppressed gonadotropin levels. Pharmacokinetics data from rodents implanted with powdered E-filled silastic capsules suggest E is highest one to two days following implantation then slowly declines (Ingberg et al., 2012, Mannino et al., 2005). Therefore, circulating E likely elevated soon after implant surgeries in our study despite the two-week latency to a significant suppression in T. While previous data suggest E level is dependent on the amount of powdered E within the capsule (Mannino, et al., 2005), we found no significant effect of E dose on terminal E levels. However, we identified dose-dependent anatomical and histological measures. The lack of significant dose effects on circulating E levels may be due to the sensitivity of the radioimmunoassay used. Future studies should capitalize on more precise hormone testing to uncover any significant effects on E level. This is critical because not all E-GAHT patients have the same serum hormone level and embodiment goals. As such, additional strategies (Cocchetti et al., 2020) and endpoints (Kennis et al., 2022) must be investigated with particular attention to varying E levels. Finally, a common thread among mice but not humans appears to be consistent T suppression by E alone (Van Steenbrugge, et al., 1984). Human-specific factors, such as the sex-hormone binding globulin (SHBG) (Jänne, et al., 1998), may produce species-specific differences in HPG function that limit translation to those E-GAHT patients whose T remains high without antiandrogens.

In sum, we have created a mouse model of E-GAHT with suppressed circulating T levels and altered reproductive phenotypes at all three doses, which can be utilized as a translational model towards understanding the reproductive and hormonal impact of E-GAHT in TGD patients. The dose-dependent effects on LH and gonad histology suggest diverse clinical outcomes should be further examined, like variable evidence of disrupted spermatogenesis (Schneider, et al., 2017). Further, integrating transgenic mouse lines, such as those expressing the human SHBG (Jänne, et al., 1998),into GAHT research offers opportunities to control mediators not available in other rodent models. Though the use of gonad intact mice increases the complexity of endocrine outcomes, they offer more meaningful translation as E-GAHT treatments are generally initiated prior to orchiectomy in most United States clinical settings. Furthermore, some E-GAHT users previously received gender-affirming hormone care during adolescence. Limited clinical data regarding the reproductive health of E-GAHT patients who previously received peripubertal GnRH agonists are available (de Nie et al., 2022) and could be compared with a combined mouse model of adolescent-GAHT (Dela Cruz et al., 2022) and E-GAHT. Ultimately, translational GAHT research will allow clinicians to provide evidence-based counseling and thus empower individual GAHT patients with more personalized choices and information regarding their reproductive health.

Table I:

Significant fold changes (means±SD; p<0.05; n.s. not significant) in hormonal and histological measures following E-GAHT treatment relative to Blank (control).

Measure	Blank	Low E (1.25 mg)	Mid E (2.5 mg)	High E (5 mg)
Weekly Testosterone (ng/dL)	Mean ±SD	FC	FC	FC
Week -1 (from implant)	281.02 ±451.02	n.s.	n.s.	n.s.
Week 1	185.31 ±118.7	n.s.	n.s.	n.s.
Week 2	317.88 ±417.45	−7	−7	−9
Week 3	145.12 ±81.55	−2.5	−2.5	−3
Week 4	319.36 ±395.57	−6	−6.5	−7
Week 5	401.74 ±436.27	−9	−9	−10
Terminal Hormones
Testosterone (ng/dL)	634.15 ±205.81	−10	−20	−16.5
Estradiol (pg/mL)	>5 n.a.	7	5	6
Luteinizing Hormone (ng/mL)	0.614 ±0.57	n.s.	n.s.	−6
Follicle Stimulating Hormone (ng/mL)	28.71 ±4.75	−2	−2	−2.5
Histological Measures
Vacuolitic Seminiferous Tubule (avg #)	13.8 ±4.266	1.5	2.5	3
Caput Epididymis Tube (μm2)	21454 ±1531	−1.3	−1.4	−1.4
Caput Epididymis Lumen (μm2)	5266 ±554.2	n.s.	n.s.	n.s.
Cauda Epididymis Tube (μm2)	15089 ±2638	n.s.	n.s.	n.s.
Cauda Epididymis Lumen (μm2)	53579 ±4529	−1.5	−1.7	−2.5