A Novel Tissue-Specific Insight into Sex Steroid Fluctuations Throughout the Murine Estrous Cycle

Abstract

Serum sex steroid levels fluctuate throughout the reproductive cycle. However, the degree to which sex steroid tissue content mimics circulating content is unknown. Understanding the flux and physiological quantity of tissue steroid content is imperative for targeted hormonal therapy development. Utilizing a gold-standard ultrasensitive liquid chromatography–mass spectrometry (LC/MS) method we determined sex steroid (17β-estradiol [E2], testosterone, androstenedione, and progesterone) fluctuations in serum and in 15 tissues throughout the murine estrous cycle (proestrus, estrus, and diestrus I) and in ovariectomized (OVX) mice. We observed dynamic fluctuations in serum and tissue steroid content throughout the estrous cycle with proestrus generally presenting the highest content of E2, testosterone, and androstenedione, and lowest content of progesterone. In general, the trend in circulating steroid content between the stages of the estrous cycle was mimicked in tissue. However, the absolute amounts of steroid levels when normalized to tissue weight were found to be significantly different between the tissues with the serum steroid quantity often being significantly lower than the tissue quantity. Additionally, we found that OVX mice generally displayed a depletion of all steroids in the various tissues assessed, except in the adrenal glands which were determined to be the main site of peripheral E2 production after ovary removal. This investigation provides a comprehensive analysis of steroid content throughout the estrous cycle in a multitude of tissues and serum. We believe this information will help serve as the basis for the development of physiologically relevant, tissue-specific hormonal therapies.

It is well-established that sex steroids play a fundamental role in a number of physiological processes, including metabolism, sexual differentiation, reproduction, as well as skeletal and immune function, to name a few (1, 2). Although both males and females possess estrogens, androgens, and progesterone, the absolute content of these steroids differs between the sexes with males presenting higher circulating androgens and lower estrogen and progesterone levels relative to females (3). With respect to females, the absolute levels of these hormones are known to fluctuate in serum throughout the human menstrual and the rodent estrous cycles (4‐7). However, due to limitations in previous and current methodologies, including instrument sensitivity to detect low (pg) levels of sex steroids, the degree to which tissue content of sex steroids mimics circulating levels during the reproductive cycle in females remains elusive.

An understanding of natural tissue fluctuation of sex steroids is of significant importance so that physiologically relevant hormone treatments may be used to elicit beneficial outcomes with the ultimate goal of creating tissue-specific hormonal therapies to minimize any potential off-target side effects. Utilizing an ultrasensitive liquid chromatography–mass spectrometry (LC/MS) method and transgenic mice, we have previously shown that different concentrations of circulating and tissue steroids can elicit behavioral and physiological changes such as food intake, metabolic outcomes, bone mineral density (BMD), and body composition (1, 8). This suggests that a certain tissue steroid content must be reached in order to elicit a given biological effect. However, before scientists move into the realm of tissue-specific hormonal therapies, it is imperative that we gain an understanding of how sex steroid tissue content changes throughout the reproductive cycle.

Therefore, the goal of this current study was to provide basic science researchers with a comprehensive assessment of tissue sex steroid levels throughout the murine estrous cycle and during estrogen deficiency utilizing an ovariectomized (OVX) mouse model. We believe this information will help serve as a foundation for the development of targeted tissue-specific hormone therapies reaching physiological levels.

Materials and Methods

Animals, Ovariectomy, and Diet

Female mice on a C57BL/6 background were bred at the University of South Carolina. Mice were housed 3 to 5/cage at 22 °C on 12:12 hours light (07:00-19:00 hours)–dark cycle (19:00-07:00 hours). All mice were 12-16 weeks of age at the start of the experiment. Bilateral ovariectomy was performed on mice at approximately 12 weeks of age (n = 15). OVX mice were sacrificed for tissue collection 3 weeks after ovariectomy. All mice were administered the purified, open-source AIN-76A diet over the course of the experiment (Bio-Serv, Frenchtown, NJ). The experimental design is presented in Fig. 1. All experimental protocols were approved by the University of South Carolina Institutional Animal Care and Use Committee.

Figure 1.

Description of experimental outline.

Vaginal Lavage and Cytology

To monitor estrous cycles, vaginal lavage was performed on mice between 07:30 and 9:30 hours daily for at least 2 consecutive cycles. Vaginal lavage was performed as described by Goldman et al with minor modifications (9). We focused our attention on proestrus, estrus, and diestrus I stages at the respective times in the day in order to capture steroid hormone changes across the estrous cycle (9). Metestrus was not consistently detected in mice in morning vaginal lavages, thus we chose to omit this stage. To obtain vaginal cells, 10 µL of sterile-filtered phosphate-buffered saline was applied to the vaginal orifice using a 200-µL pipet tip and pipettor. The liquid was allowed to sit for a few seconds then collected at the vaginal opening and placed on a glass microscope slide and examined immediately with a microscope using a 10× objective. Representative samples were dried for later staining and imaging. Proestrus was defined as the first day following diestrus and which had predominantly nucleated epithelial cells with smooth edges, which often appeared in clumps. In some cases, proestrus females had a combination of smooth nucleated epithelial cells and cornified epithelial cells, but the smooth nucleated cells were the major cell type. Estrus was defined as the stage following proestrus with the presence of mostly cornified epithelial cells that were either large anucleate cells with jagged edges or needle-like appearance. Metestrus was the period following estrus that was characterized by smooth nucleated and cornified irregular epithelial cells and small leukocytes; metestrus was not observed in some mice. Diestrus I was the first day following estrus or metestrus that exhibited predominantly leukocytes alone or in combination with a few smooth nucleated cells. Mice were typically 5-day cyclers with either 2 consecutive days of estrus or 3 consecutive days of diestrus.

Dried samples were stained using hematoxylin and 1% eosin and coverslipped with mounting media. Images were captured using a Nikon Eclipse E600 microscope and a 50× objective, SPOT camera, and software. Representative images of the proestrus, estrus, and diestrus I stages are presented in Fig. 2A.

Figure 2.

Impact of estrous cycle on uterine weight. (A) Representative images of vaginal lavage cytology pertaining to each stage of the estrous cycle examined, and (B) uterus weight of the experimental mice (n = 11-18). Data are presented as mean ± SE. Bar graphs not sharing a common letter are significantly different from one another (P < .05).

Tissue Collection

All mice were sacrificed between 09:00 and 10:30 hours after vaginal lavage and cytology had been completed in order to confirm the stage of the estrous cycle. After mice were euthanized via isoflurane inhalation, blood was immediately collected via the inferior vena cava and was subsequently processed for serum isolation. After blood removal, the animals were perfused with approximately 50 mL of mass-spectrometry-grade water (Fisher Scientific, Waltham, MA) through the inferior vena cava using a peristaltic pump. After perfusion, organs including the ovaries (except for the OVX mice), adrenal glands, uterus, adipose tissue depots (visceral [gonadal], subcutaneous [inguinal], and brown), pancreas, skeletal muscle (gastrocnemius), bone (tibia), liver, kidneys, spleen, colon, lungs, and heart were dissected and immediately flash frozen in liquid nitrogen.

Extraction of Steroids

Extraction of steroids was performed as previously described (8, 10). The average volume of serum and tissue weight used for each tissue analyzed are presented in Table 1. Internal standards (Sigma Aldrich, St. Louis, MO) for testosterone (-2,3,4-¹³C₃), E2 (D₅), progesterone (D₉), and androstenedione (-2,3,4-¹³C₃) were added to each tissue and serum sample prior to tissue processing. Calibration curves utilizing certified reference material (testosterone, Sigma Aldrich, Catalog# T037; E2, Sigma Aldrich, Catalog# E-060; progesterone, Sigma Aldrich, Catalog# P-069; and androstenedione, Sigma Aldrich, Catalog# A-075) were used to determine the quantity of each steroid. Steroids were extracted from serum via vortex using 500 mL of mass spectrometry grade methyl tert-butyl ether (MTBE) (×2) (Fisher Scientific, Waltham, MA). The resulting upper phase was removed and placed into a glass vial where it was dried down using N₂. The dried-down steroids were derivatized as previously described utilizing 1-methylimidazole-2-sulfonyl chloride (11). For tissue samples, tissues were weighed and homogenized in 1 mL of mass spectrometry grade acetonitrile (Fisher Scientific, Waltham, MA) for 3 to 5 minutes using a bead beater (Biospec Products, Bartlesville, OK) and 3.5-mm stainless steel UFO beads (Next Advance, Raymertown, NY). The samples were centrifuged 5 minutes × 12 000g and the resulting supernatant was removed and placed in glass vials. Subsequently, the tissue samples were resuspended in 1 mL of acetonitrile, homogenized, centrifuged, and the supernatant collected for a second time. The supernatants were dried down using N₂ and were resuspended in 200 mM sodium acetate. MTBE was added to the sample for liquid–liquid extraction of the steroids (×2). The MTBE layer was removed and dried under N₂ prior to derivatization as previously described (11). After derivatization, the samples were placed in 0.2 PVDF micro spin filters (Fisher Scientific, Waltham, MA) and centrifuged to remove any insoluble material prior to mass spectrometry analysis.

Table 1.

Average tissue weight and serum volume used for steroid determination

Tissue	Average weight (mg)/volume (µL) ± SE
Ovaries	12.9 ± 0.9
Adrenal glands	7.3 ± 0.3
Serum	167 ± 8.0
Uterus	84.0 ± 14.7
Visceral adipose tissue	274.5 ± 14.7
Subcutaneous adipose tissue	225.6 ± 12
Brown adipose tissue	66.2 ± 3.4
Pancreas	117.1 ± 3.6
Skeletal muscle	123.4 ± 7.4
Bone	37.4 ± 1.1
Liver	257.4 ± 15.1
Kidney	183.9 ± 4.4
Spleen	99.9 ± 7.2
Colon	149.1 ± 4.8
Lungs	156.1 ± 10.7
Heart	119.9 ± 3.1

Data are presented as mean ± SE.

Chromatography and Mass Spectrometry

Analyses were carried out on a Q Exactive HF-X hybrid quadrupole-orbitrap mass spectrometer with a Vanquish HPLC on the front end (Thermo Electron, Waltham, MA). Ten microliters of the derivatized sample was injected onto a Waters Xbridge C18 column (2.1 mm × 150 mm, 3.5-um particles; Waters Corp. Milford, MA) using a mobile phase consisting of 0.1% formic acid in water (A) and 0.1% formic acid in methanol (B). The sex steroids were separated using the following linear solvent gradient at a flow rate of 200 μL/minute: initial conditions 30% B ramping to 65% B over 6 minutes followed by a shallow ramp to 73% B ending at 20 minutes. At 20.1 minutes a quick ramp to 95% B holding until 23 minutes. Finally, a return to starting conditions of 30% B at 23.1 minute ending at 26 minutes.

Mass analysis was carried out using electrospray ionization in the positive ion mode. The H electrospray ionization source settings were as follows: sheath gas flow, 45 units; Aux gas flow, 10 units; Sweep gas flow, 2 units; electrospray voltage, 4 kV; capillary temperature, 320 °C; Funnel RF, 40; Aux gas heater, 400 °C. E2 was analyzed using targeted selected ion monitoring using the M + H ion for the derivative (M + H = 417.1843); simultaneously, the heavy isotopically labeled internal standard was similarly monitored (for D₅ E2 M + H = 422.2156). The targeted selected ion monitoring settings were microscans, 1; resolution, 240 K; AGC target, 1E6; maximum IT, 150 ms; isolation window, 1 m/z; data type, profile. Testosterone, androstenedione, and progesterone and their stable isotope labeled internal standard were analyzed using parallel reaction monitoring mode fragmenting the M + H ion in the HCD collision cell and measuring the fragments in the orbitrap analyzer. For testosterone, M + H precursor ion =289.2162 with quantifying product ion 97 m/z; for ¹³C₃ testosterone, M + H precursor ion = 292.2263 with quantifying product ion 100 m/z, for androstenedione, M + H precursor ion = 287.2006 with quantifying product ion 97 m/z; for ¹³C₃ androstenedione, M + H precursor ion = 290.2106 with quantifying product ion 100 m/z; for progesterone, M + H precursor ion = 315.2319 with quantifying production ion 97 m/z; for D₉ progesterone, M + H precursor ion = 324.2884 with quantifying product ion 100 m/z. The parallel reaction monitoring settings were microscans, 1; resolution, 30 K; AGC target, 1E6; maximum IT, 250 ms; isolation window, 2.0 m/z; stepped NCE, 35,45,54; data type, profile. The mass spectrometer was mass calibrated before each batch of runs and the data was collected and processed using the vendor-provided Xcalibur software with mass error filter set at 3 ppm. Peak area ratios of analyte and internal standard were used to determine steroid concentrations. In order to determine the limit of quantification for each analyte, the following amount of each steroid (6 replicates each) was spiked into 200 μL of charcoal-stripped (2×) fetal bovine serum: E2, 0.01, 0.025, 0.05, 0.1, 0.25, and 0.5 pg; testosterone and androstenedione, 0.01, 0.025, 0.05, 0.1, 0.25, 0.5, and 1 pg; and progesterone, 0.25, 0.5, 1, 2, 3 pg and was compared with nonspiked samples. It was determined that the limit of quantification for each of the 4 analytes was 0.1, 0.5, 0.5, and 1 pg for E2, testosterone, androstenedione, and progesterone, respectively, in 200 μL of charcoal-stripped fetal bovine serum with a coefficient of variation <15%. The E2, (0.1, 0.25, 0.5, 0.75, 1, 3, 5, and 10 pg), testosterone (0.5, 1, 2, 3, 4, 5, 6, 8, 10, 15, 20, and 185 pg), androstenedione (0.5, 0.75, 1, 3, 5, 10, 20, 30, 50, 100 pg), and progesterone, (0.005, 0.010, 0.05, 0.250, 0.500, 1 ng) calibration curves all yielded r² values > 0.98. For the detection of progesterone in the ovaries and adrenal glands we used a more robust calibration curve (1, 2.5, 7, 10, 100, and 250 ng) which yielded a r² values > 0.98. If a steroid was not detected in a respective tissue sample, it was given a value of “0.” If no steroid was detected in a respective group, the steroid was labeled as “not determined.” All data were normalized to pg or ng/g tissue or mL of serum.

Statistical Analysis

Data were analyzed using commercially available statistical software: Prism 9 (GraphPad Software, La Jolla, CA). A 1-way analysis of variance followed by a Newman–Keuls post hoc test was used to assess differences between groups. Any statistical test that did not pass the equal-variance test (Bartlett's test for equal variances) was log or square-root transformed and then reanalyzed. Data are presented as means ± SE, and the level of significance was set at P < .05.

Results

Impact of Estrous Cycle and Ovariectomy on Uterine Weights

Uterus weight was found to be greatest during the proestrus stage of the estrous cycle followed by the estrus and diestrus I phases (Fig. 2B) (P < .05). The uterus of OVX mice presented the lowest uterus weight (P < .05).

E2 Content in Circulation and Tissue is Greatest During Proestrus Followed by Diestrus I

The concentration of E2 (Fig. 3) was found to be greatest during proestrus in 14/16 tissues (ovaries [Fig. 3A], serum [Fig. 3C], uterus [Fig. 3D], visceral [Fig. 3E], subcutaneous [Fig. 3F], and brown adipose tissues [Fig. 3G], pancreas [Fig. 3H], skeletal muscle [Fig. 3I], bone [Fig. 3J], kidney [Fig. 3L], spleen [Fig. 3M], colon [Fig. 3N], lungs [Fig. 3O], and heart [Fig. 3P]) (P < .05). In the adrenal glands, no statistically significant difference was found in the E2 content between mice in proestrus and diestrus I (Fig. 3B) (P < .05). Additionally, no difference was found in hepatic E2 content across any of the estrous stages assessed, including OVX mice (Fig. 3K). Mice in diestrus I presented higher E2 levels relative to mice in estrus for 6/16 tissues (ovaries [Fig. 3A], subcutaneous [Fig. 3F] and brown adipose tissue [Fig. 3G], bone [Fig. 3J], kidney [Fig. 3L], and lungs [Fig. 3O]) (P < .05) and similar E2 content in 9/16 tissues (adrenal glands [Fig. 3B], serum [Fig. 3C], uterus [Fig. 3D], visceral adipose tissue [Fig. 3E], pancreas [Fig. 3H], skeletal muscle [Fig. 3I], spleen [Fig. 3M], colon [Fig. 3N], and heart [Fig. 3P]). The OVX mice had the lowest E2 content in 10/16 tissues (serum [Fig. 3C], uterus [Fig. 3D], visceral [Fig. 3E], subcutaneous (not detected) [Fig. 3F], and brown (not detected) adipose tissue [Fig. 3G], as well as the skeletal muscle [Fig. 3I], kidney [Fig. 3L], spleen (Fig. 3M), and heart [Fig. 3P]) (P < .05). However, the OVX mice, as well as mice in estrus and diestrus I, had similar levels of adrenal gland (Fig. 3B), pancreas (Fig. 3H), and colon E2 (Fig. 3N). Additionally, the OVX mice shared similar E2 content as mice in estrus for bone (Fig. 3J) and lung E2 (Fig. 3O).

Figure 3.

17β-Estradiol tissue content. (A) Ovaries, (B) adrenal glands, (C) serum, (D) uterus, (E), visceral adipose tissue, (F) subcutaneous adipose tissue, (G) brown adipose tissue, (H) pancreas, (I) skeletal muscle, (J) bone, (K) liver, (L) kidney, (M) spleen, (N) colon, (O) lungs, and (P) heart (n = 8-18). Data are presented as mean ± SE. Bar graphs not sharing a common letter are significantly different from one another (P < .05). NA, not applicable; ND, not detected.

Absolute E2 Content Is Variable Across Tissues

The ovaries followed by the uterus and adrenal glands displayed the greatest amounts of tissue E2 content for each of the respective estrous cycle stages evaluated when normalized to tissue weight (Fig. 4A-4C). For OVX mice, the adrenal glands presented the greatest E2 content (Fig. 4D). An interesting observation was that the pancreas displayed relatively higher levels of E2 content than other tissues examined only during proestrus. Additionally, the colon and liver displayed some of the lowest E2 content across the stages. Furthermore, the general trend was for the visceral and subcutaneous adipose tissue to have higher E2 content than the brown adipose tissue.

Figure 4.

Descending order of 17β-estradiol tissue content across estrous cycle stage and ovariectomy. (A) Proestrus, (B) estrus, (C), and diestrus stage of the estrous cycle, and (D) ovariectomy. NA, not applicable.

Testosterone Content Is Impacted by Stage of the Estrous Cycle and Ovariectomy

In 6/16 tissues assessed (ovaries [Fig. 5A], serum [Fig. 5C], subcutaneous adipose tissue [Fig. 5F], kidney [Fig. 5L], lungs [Fig. 5O], and heart [Fig. 5P]), testosterone was found to be significantly higher during proestrus than in all other stages (Fig. 5) (P < .05). Of these tissues, no differences were found when comparing estrus, diestrus, and OVX groups for serum (Fig. 5C), subcutaneous fat (Fig. 5F), and heart (Fig. 5P). The diestrus group had a significantly higher testosterone content in the ovaries (Fig. 5A), visceral fat (Fig. 5E), kidney (Fig. 5L), and lungs (Fig. 5O) than mice in the estrus stage and OVX mice (P < .05). In brown adipose tissue (Fig. 5G) and pancreas (Fig. 5H), the OVX mice presented the lowest level of testosterone relative to all other groups (P < .05). No differences were found with respect to testosterone content among any of the stages assessed, including OVX mice, for 3/16 tissues (bone [Fig. 5J], spleen [Fig. 5M], and colon [Fig. 5N]). In the uterus (Fig. 5D), the estrus mice had the lowest testosterone content relative to all groups (P < .05). For skeletal muscle, only the mice in estrus displayed lower testosterone content relative to the proestrus mice (Fig. 5I) (P < .05). Interestingly, the OVX mice had higher adrenal gland testosterone content than all other groups (Fig. 5B) (P < .05).

Figure 5.

Testosterone tissue content. (A) Ovaries, (B) adrenal glands, (C) serum, (D) uterus, (E) visceral adipose tissue, (F) subcutaneous adipose tissue, (G) brown adipose tissue, (H) pancreas, (I) skeletal muscle, (J) bone, (K) liver, (L) kidney, (M) spleen, (N) colon, (O) lungs, and (P) heart (n = 8-18). Data are presented as mean ± SE. Bar graphs not sharing a common letter are significantly different from one another (P < .05). NA, not applicable.

The Ovaries and Adrenal Glands Have the Highest Absolute Testosterone Content Across the Estrous Cycle

In terms of absolute values normalized to tissue weight, the ovaries and adrenal glands generally had the greatest amount of testosterone (Fig. 6A-6C). Of interest was the finding that the OVX mice presented 10× the amount of adrenal gland testosterone compared with all other groups (Fig. 6D). The general trend across the different estrous cycle stages and in OVX mice was for the adipose tissue depots (visceral, subcutaneous, and brown) to have a higher content of testosterone compared to other tissues analyzed. Serum, bone, colon, liver, and skeletal muscle were tissues with a lower content of testosterone across the different stages of the estrous cycle and in the OVX mice.

Figure 6.

Descending order of testosterone tissue content across estrous cycle stage and ovariectomy. (A) Proestrus, (B) estrus, (C) and diestrus stage of the estrous cycle, and (D) ovariectomy. NA, not applicable.

Androstenedione Tissue Content Fluctuates Throughout the Estrous Cycle and Is Impacted by Ovariectomy

Androstenedione variability throughout the estrous cycle mimicked fluctuations seen with testosterone (Fig. 7). The proestrus stage elicited the highest content of androstenedione in 9/16 tissues (ovaries [Fig. 7A], serum [Fig. 7C], visceral [Fig. 7E], subcutaneous [Fig. 7F], and brown adipose tissue [Fig. 7G], pancreas [Fig. 7H], skeletal muscle [Fig. 7I], bone [Fig. 7J], and lungs [Fig. 7O]) (P < .05). In the uterus (Fig. 7D), kidney (Fig. 7L), spleen (Fig. 7M), colon (Fig. 7N), and heart (Fig. 7P), mice in the proestrus and diestrus stages presented a similar content of androstenedione. In 5/16 tissues (ovaries [Fig. 7A], serum [Fig. 7C], uterus [Fig. 7D], lungs [Fig. 7O], and heart [Fig. 7P]), the diestrus mice had higher androstenedione content than mice in estrus and OVX mice (P < .05). OVX mice presented the lowest content of androstenedione in serum (Fig. 7C), uterus (Fig. 7D), visceral and brown adipose tissue (Fig. 7E and 7F), pancreas (Fig. 7H), and skeletal muscle (Fig. 7I) (P < .05). In the brown adipose tissue (Fig. 7G), lungs (Fig. 7O), and heart (Fig. 7P), the mice in estrus and OVX mice had a lower tissue content of androstenedione than mice in proestrus and diestrus (P < .05). No differences in hepatic androstenedione was found across any of the groups (Fig. 7K). Similar to what was found with testosterone, the OVX mice had the highest content of adrenal gland androstenedione (Fig. 7B).

Figure 7.

Androstenedione. (A) Ovaries, (B) adrenal glands, (C) serum, (D) uterus, (E), visceral adipose tissue, (F) subcutaneous adipose tissue, (G) brown adipose tissue, (H) pancreas, (I) skeletal muscle, (J) bone, (K) liver, (L) kidney, (M) spleen, (N) colon, (O) lungs, and (P) heart (n = 8-18). Data are presented as mean ± SE. Bar graphs not sharing a common letter are significantly different from one another (P < .05). NA, not applicable; ND, not detected.

Besides the Ovaries and Adrenal Glands, Adipose Tissue Depots Are the Largest Site of Androstenedione Content

The ovaries and adrenal glands possessed the greatest content of androstenedione normalized to tissue weight regardless of the stage of the estrous cycle (Fig. 8A-8C). Similarly, in the OVX mice, androstenedione content was greatest in the adrenal glands (Fig. 8D). Of interest was the finding that regardless of estrous stage or ovariectomy, androstenedione content was found to be highest in visceral, subcutaneous, and brown adipose tissues than in other analyzed tissues. Conversely, serum, liver, uterus, bone, and lungs tended to have the lowest androstenedione content across the groups.

Figure 8.

Descending order of androstenedione tissue content across estrous cycle stage and ovariectomy. (A) Proestrus, (B) estrus, (C), and diestrus stage of the estrous cycle, and (D) ovariectomy. NA, not applicable.

Ovariectomy and Proestrus Generally Present Similar Levels of Tissue Progesterone Content Relative to Estrus and Diestrus Stages

Mice in proestrus and OVX mice exhibited similarly low progesterone content (Fig. 9) in 12/16 tissues (serum [Fig. 9C], uterus [Fig. 9D], visceral [Fig. 9E], subcutaneous [Fig. 9F], and brown adipose tissues [Fig. 9G], pancreas [Fig. 9H], skeletal muscle [Fig. 9I], bone [Fig. 9J], kidney [Fig. 9L], spleen [Fig. 9M], colon [Fig. 9N], lungs [Fig. 9O], and heart [Fig. 9P]) relative to mice in the diestrus and estrus stages (P < .05). No differences across groups were found in ovary (Fig. 9A) and adrenal gland (Fig. 9B) progesterone content. In the OVX mice, liver (Fig. 9K) and lung (Fig. 9O) progesterone content was found to be lower than mice in any 3 of the estrous cycle stages (P < .05).

Figure 9.

Progesterone. (A) Ovaries, (B) adrenal glands, (C) serum, (D) uterus, (E), visceral adipose tissue, (F) subcutaneous adipose tissue, (G) brown adipose tissue, (H) pancreas, (I) skeletal muscle, (J) bone, (K) liver, (L) kidney, (M) spleen, (N) colon, (O) lungs, and (P) heart (n = 8-18). Data are presented as mean ± SE. Bar graphs not sharing a common letter are significantly different from one another (P < .05). NA, not applicable.

Progesterone Tissue Content Is Highest in Ovaries and Adrenal Glands Followed by Adipose Tissue Sites

With respect to tissue comparisons of normalized progesterone content, the ovaries and adrenal glands exhibited the greatest amount of progesterone regardless of the stage of the estrous cycle or ovariectomy (adrenal glands only) (Fig. 10A-10D). Similar to the finding with androstenedione, the visceral, subcutaneous, and brown adipose tissues presented high progesterone content relative to other tissues analyzed across the different groups. The spleen, skeletal muscle, bone, serum, and liver were the tissues that tended to exhibit the lowest progesterone content across the estrous cycle stages and in OVX mice.

Figure 10.

Descending order of progesterone tissue content across estrous cycle stage and ovariectomy. (A) Proestrus, (B) estrus, (C), and diestrus stage of the estrous cycle, and (D) ovariectomy. NA, not applicable.

Discussion

We performed an extensive study examining how the stage of the murine estrous cycle and the removal of the main site of sex steroid production in females—the ovaries—impacts circulating and tissue steroid content. To do this we used a gold-standard ultrasensitive LC-MS technique that we have previously established to measure sex steroids in murine serum and tissue (8, 10). The sensitivity of our method allowed us to detect E2, progesterone, androstenedione, and testosterone in circulation and tissue even at very low levels. This has never been accomplished before in female mice due primarily to limitations with respect to methodological limits of quantification (12).

The general consensus is that female mice present the highest concentration of circulating E2 during proestrus, which we confirm based upon our study design (7, 9). However, a recent publication has challenged this as it was found that E2 peaks during the morning hours of diestrus (12). The discrepancy between our findings and that of Wall et al may be explained by the fact that we only assessed steroid levels during diestrus I whereas it is unclear what stage (I or II) of diestrus E2 was examined in the contrary publication (12). Additionally, we only assessed steroid levels at a single time-point (morning), whereas Wall et al examined steroid levels at multiple time points throughout the day. Lastly, the light cycle of our mice commenced at 07:00 hours whereas that of the mice from Wall et al started at 05:30 hours. These differences between study designs may explain discrepant results.

For the majority of tissues analyzed, we show that tissue steroid content mimics circulating levels when comparing trends across the groups assessed. This is evident in the fact that mice in proestrus presented higher circulating E2, testosterone, androstenedione, and lower progesterone relative to mice in estrus and diestrus, which was generally true for a multitude of the tissues analyzed. However, similar to the findings of others who did a comprehensive sex steroid study in male mice, the absolute amount of each steroid varied significantly between tissues (13). Besides the primary sites of steroid synthesis (the ovaries and adrenal glands), the adipose tissue, particularly visceral and subcutaneous sites, were found to be a significant depot for all steroids examined. This is consistent with the findings in male mice (13). Of interest, in most of the tissues examined was the dynamic changes in the absolute amounts of the steroids when progressing from proestrus to estrus indicating their rapid intratissue metabolism. These fluctuations seemed to be tissue specific and dependent upon the steroid examined. For instance, visceral adipose tissue and pancreas E2 dropped approximately 90% between proestrus and estrus whereas in skeletal muscle and serum the drop was only approximately 66%. Alternatively, visceral adipose tissue and pancreas testosterone content dropped approximately 58% between proestrus and estrus. These contrasts suggest a difference in steroid flux with respect to steroid production and metabolism in a tissue-specific manner across the estrous cycle. It should be noted that just because tissue steroid content is changed does not mean that steroid receptor activation is consequently modified. Future studies would need to be performed to determine the degree to which tissue steroid content impacts steroid receptor activation.

We also observed that OVX mice presented dramatically higher contents of adrenal testosterone and androstenedione compared to gonadally intact mice. Older non-specific and less sensitive steroid analyses perpetrated the belief that the adrenal glands in rodents do not synthesize androgens like humans do (12, 14). However, a 2018 publication clearly showed that male murine adrenals do indeed synthesize significant amounts of androgens (15). Our data in females support the finding that murine adrenals are a major source of androgen production, especially when the ovaries are removed.

OVX mice presented vastly lower serum and tissue content of E2, but testosterone and androstenedione levels generally on par with estrus and diestrus mice. A limitation of using mouse models to translate to the human condition of menopause (estrogen deficiency) is the difference in the expression of aromatase (CYP19A1) between humans and mice. Aromatase is an important enzyme in the steroid pathway that is responsible for converting androstenedione and testosterone to estrogens (E1 and E2, respectively). In human premenopausal females, the primary site of E2 production is the gonads where aromatase activity is high (16). During postmenopause, however, the primary site of E2 production is in peripheral tissues, such as adipose tissue, via the conversion of testosterone to E2 (16). It was recently shown that adipose tissue E2 is approximately 3 times higher than circulating E2 in postmenopausal females (17). We have found that mice do not exactly mimic humans with respect to these characteristics during ovariectomy-induced E2 deficiency. We find that in mice E2 is still the predominant estrogen in circulation and in tissue during E2 deficiency—not E1, as evidenced in our previous publication (8). Others have also shown that serum E1 is not significantly higher than E2 in the circulation of OVX mice (5). Additionally, we show that the circulating E2 concentration is similar to the adipose tissue E2 content on a milliliter to gram comparison (<5 pg/mL or pg/g) in OVX mice, which suggests that peripheral aromatization of testosterone to estrogens in female mice in adipose tissue is minimal or nonexistent. This finding is supported by a previous investigation where it was found that aromatase is expressed in male but not female gonadal fat (18). The fact that we found the highest content of E2 in the adrenal glands of OVX mice paired with the elevated adrenal testosterone and androstenedione content suggests that the main site of E2 production in OVX mice is from the adrenal glands via aromatization of testosterone. One caveat of our experiment, is that we waited only 3 weeks after ovariectomy to assesses serum and tissue steroid content. It may be that further adaptations to steroid production/metabolism may occur with a longer period of estrogen deficiency induced by ovariectomy.

It was of interest to see that ovary and adrenal gland progesterone content was similar between the stages of the estrous cycle and OVX mice (adrenal gland only), but differed significantly between the groups with respect to serum and tissue content. Luteinizing hormone is well-known for eliciting ovulation, corpus luteum formation, and progesterone secretion from the ovaries (19). During the estrous cycle, in the absence of mating (or cervical stimulation), tissue progesterone produced by new corpora lutea during estrus and diestrus is rapidly metabolized to its inactive metabolite 20α-hydroxyprogesterone within the ovary by the increase in 20α-hydroxysteroid dehydrogenase (AKR1C1) expression (20). Thus the increase in progesterone in the serum reflects the transient release of progesterone from the tissue (9).

Another interesting observation was that the liver, when normalized by tissue weight, was one of the tissues to consistently present a low steroid content for each of the steroids assessed. This was also found to generally true for the sex steroid content in liver relative to other tissues in a comprehensive study performed in male mice (13). It is well established that the liver is a primary site of steroid metabolism as it initiates the process of steroid deactivation for ultimate excretion (21). This may explain why hepatic steroid content is relatively low.

The tissue-specific effect that a steroid has and the content needed to elicit such effect is an unresolved area of research. This is largely due to limitations in the ability to detect steroids at extremely low picogram levels, which we have achieved. We believe this knowledge is a necessity for the development of safe and effective hormonal treatments. For instance, we have previously shown that a level of circulating E2 reaching approximately 26 pg/mL was associated with improvements in BMD in OVX female mice, but approximately 12 pg/mL was insufficient (8). Similarly, serum E2 at a concentration of approximately 26 pg/mL was not associated with changes in food intake or glucose metabolism, but approximately 64 pg/mL E2 impacted both of these outcomes as evidenced by decreased food intake and improved glucose metabolism. Hormone replacement therapy can be provided in many different forms (eg, ester-conjugated) and modes of delivery (eg, oral tablets, subcutaneous injections, implants, skin patches, intranasal, sublingual, gels, and intervaginal) which have been shown to impact bioavailability as assessed by circulating steroid concentrations (22‐24). These changes in bioavailability have been shown to differentially impact health outcomes, such as BMD, hypertension, and c-reactive protein levels in humans (25‐27). From a basic science perspective, a better understanding of how the form and mode of steroid delivery impacts not only circulating but also tissue bioavailability is needed. Furthermore, we need an improved comprehension of the tissue-specific physiological processes impacted by varying levels of steroids to improve therapies for clinical application.

In summary, we have established a gold-standard ultrasensitive LC/MS methodology to comprehensively examine steroid content in a multitude of tissues through the murine estrous cycle. We hope this information provides a foundation for researchers to develop tissue-specific hormonal therapies that reach physiological levels. Future studies are necessary to determine the tissue-specific steroid quantity required to impact physiological outcomes to enhance human health by minimizing any potential unwanted off-target effects.

Abbreviations

BMD

bone mineral density

E2

17β-estradiol

LC/MS

liquid chromatography-mass spectrometry

MTBE

methyl tert-butyl ether

OVX

ovariectomized

Disclosures

None declared.

Data Availability

Original data generated and analyzed during this study are included in this published article.