Abstract
Accurate measurement of very low circulating estradiol (E2) (<5 pg/ml) in postmenopausal women and in mice is essential to investigating sex steroid action in target tissues. However, direct immunoassays are too inaccurate and conventional mass spectrometry-based measurement too insensitive at these serum E2 levels. We report application of an ultrasensitive method using a novel estrogen-selective derivatization in liquid chromatography-mass spectrometry to measure serum E2, with a detection limit of 0.25 pg/ml in small (0.2 ml) serum volumes that can quantify serum E2 in 98% and serum E1 in 100% of healthy postmenopausal women. Aromatase inhibitor (AI) treatment of postmenopausal women with breast cancer further reduces serum E2 by 85% and serum estrone (E1) by 80%. The wide scatter of circulating E2 in AI-treated women suggests that the degree of sustained E2 depletion, now quantifiable, may be an efficacy or safety biomarker of adjuvant AI treatment. This ultrasensitive method can also measure serum E2 in most (65%) female but not in any male mice. Further studies are warranted using this and comparable ultrasensitive liquid chromatography-mass spectrometry estrogen measurements to investigate the relationship of circulating E2 (and E1) in male, postmenopausal female, and childhood health where accurate quantification of serum estrogens was not previously feasible. This will focus on the direct impact of estrogens as well as the indirect effects of androgen aromatization on reproductive, bone, and brain tissues and, notably, the efficacy and safety of AIs in adjuvant breast cancer treatment.
Keywords: estradiol, testosterone, aromatase inhibitor, steroid mass spectrometry, post-menopause, mice
The measurement of circulating estradiol in conditions in which its concentrations are very low, such as postmenopausal women receiving aromatase inhibitors (AIs), and in small volume samples such as those from rodents, has, to date, been problematic due to limited sensitivity of available methods.
Recognizing the estrogen dependence of breast cancer led to the development of adjuvant antiestrogen treatments to prevent disease recurrence, thereby prolonging disease-free survival and a potential long-term cure of breast cancer, especially in postmenopausal women with estrogen receptor (ER)-positive disease, the most frequent category of lethal breast cancers [1, 2]. There are 2 classes of antiestrogens, ER blockers (also known as selective ER modulators) and AIs. Estrogen receptor blockers compete with estradiol (E2), the major potent bioactive estrogen, for binding to ERs and block estradiol’s growth-promoting effects on breast cancer cells [3]. Aromatase inhibitor drugs are a newer class of antiestrogens that inhibit aromatase, the unique enzyme which converts circulating androgen precursors, testosterone, and androstenedione into corresponding estrogens, estradiol (E2), and estrone (E1). Despite the success of AI therapy in lowering circulating E2 and E1 concentrations [4], some women experience disease relapse [5] or toxicity (including symptomatic estrogen deficiency) [6]. The mechanisms of AI resistance or adverse effects remain little understood. Important pathophysiological reasons include differences in efficacy of AI treatment in suppressing E2 synthesis due to medication noncompliance and/or pharmacogenetic differences in metabolism or efficacy of AI drugs or aromatase enzyme activity [7]. Accurate measurement of circulating E2 is required for elucidating whether the degree of E2 depletion determines efficacy or safety of AI treatment. However, circulating E2 levels are very low in postmenopausal women, making it technically challenging to measure [8, 9]. Consequently, for women taking an AI, undetectable serum E2 concentrations are considered an index of effective AI treatment, but this categorization may be indiscriminate with regard to the extent of E2 depletion [10]. Measurement of serum E2 by direct (nonextraction) E2 immunoassays is unreliable due to their nonspecificity, leading to overestimating concentrations, most prominent at low E2 levels [8, 11]. Conventional liquid chromatography-mass spectrometry (LC-MS) methods have characteristic detection limits for serum E2 of 3–5 pg/ml that are not sufficiently sensitive to quantify serum E2 in all postmenopausal women [8, 9, 12] making them unable to estimate the further depression of serum E2 concentration induced by AI drugs. Similarly, circulating E2 concentrations in rodents are also very low so that measurement is unreliable with direct (nonextraction) estradiol immunoassays [8, 13] and too low (< 3–5 pg/ml) for conventional LC-MS methods [14, 15]. One study of gas chromatography-mass spectrometry using relatively large (250 µl) mouse serum samples reported mean serum E2 of 2.7 pg/ml in female but undetectable levels (< 0.3 pg/ml) in male mice [16].
We reported an ultrasensitive method to measure serum E2 using a novel estrogen-selective derivatization featuring the theoretical sensitivity to achieve a serum E2 measurement in postmenopausal women and mice [17]. The present study demonstrates that this ultrasensitive method has sufficient sensitivity so that, using small serum sample volume (0.2 ml in humans, 0.1 ml in mice), it can measure serum E2 in virtually all postmenopausal women as well as quantifying further lowering of serum E2 due to AI treatment. Furthermore, we have applied this ultrasensitive method to measure of serum E2, E1, and testosterone (T) in female and male mouse serum samples.
Materials and Methods
Patient samples
Patient samples were provided with ethical approval from 2 studies of postmenopausal women treated with AIs for breast cancer.
Study 1 was a randomized placebo-controled clinical trial (principal investigator: Professor SR Davis; Australian New Zealand Clinical Trials registry # ACTRN12615000083594) reported previously [18] of postmenopausal women with breast cancer taking an AI treated with vaginal cream (containing testosterone or placebo) for genitourinary symptoms and taking no systemic hormone therapy for at least 6 months. Participating women (aged 57 ± 9 years, mean ± SD) had breast cancer diagnosis at mean age 53 ± 9 years and whose treatment for breast cancer (median 2.7 years) included AI drugs (letrozole 34%, anastrozole 57%, exemestane 9%) for a median duration of 2.0 years. As a group, 32% had prior oophorectomy, 38% had prior tamoxifen treatment, and their weight (68.4 ± 11.8 kg), height (164 ± 5.9 cm), body mass index (26.6 ± 4.2 kg/m2), and systolic (125 ± 14 mm Hg) and diastolic (75 ± 12 mm Hg) blood pressures were within normal limits.
Study 2 was an observational study investigating the occurrence of dry eye disease during AI treatment (principal investigator: Associate Professor Blanka Golebiowski) [19]. Serum samples were provided by women (n = 45) undergoing breast cancer treatment with AI drugs at an ambulatory hospital clinic and by age-matched control women (n = 42) not taking AI drugs recruited from a university optometry clinic. Recruited women with breast cancer had no hormone treatment in the past 12 months. The AI-treated and control groups were well matched for age (66 ± 8.5 years vs 64 ± 7.5 years, respectively), weight (73.1 ± 14.4 kg vs 70.8 ± 16.4 kg), body mass index (26.8 ± 6.2 kg/m2 vs 26.4 ± 5.3 kg/m2), and ethnicity (Caucasian: Asian: Other 35 [78%]: 1 [2%]: 9 [20%] vs 38 [91%]:0 [0%]:4 [9%]), respectively. Women taking AI drugs (letrozole [32, 71%], anastrozole [7, 16%], exemestane [6, 13%]) had been treated for a median of 1.3 years (IQR 2.85 years, range 0.2–25 years).
Mouse samples
Mice (C57BL6/J) were maintained under standard housing conditions (ad libitum access to food and water in a temperature- and humidity-controled, 12-hour light/dark environment) with new incoming mice acclimatized for 1 week prior to experiments at the approved animal facilities of the ANZAC Research Institute or the Biological Resources Centre, UNSW. All procedures were approved by the Sydney Local Health District Animal Welfare Committee or the UNSW Animal Ethics Committee within NHMRC guidelines for animal experimentation. Serum samples were obtained by terminal cardiac puncture under anaesthesia from sexually mature female and male mice. One experiment involved terminal blood sampling without regard to estrus stage in females (n = 114, 87 female). A second experiment involved 40 mature female mice (~3 months of age) that underwent vaginal smears for a minimum of 2 consecutive days prior to terminal blood sampling to ensure they were actively progressing through the estrus cycle. Estrus cycle stage was determined from vaginal epithelial cell smears collected using 15 µL of 0.9% sterile saline and transferred to glass slides to air dry [20]. Dry smears were stained with 0.5% toluidine blue before being examined under light microscope. Estrus cycle stage was determined based on the presence or absence of leukocytes, cornified epithelial cells, and nucleated epithelial cells. Proestrus was characterized by the presence of mostly nucleated and some cornified epithelial cells; estrus was identified by the presence of primarily cornified epithelial cells; at metestrus, both cornified epithelial cells and leukocytes were present; and at diestrus, predominantly leukocytes were present.
Steroid analysis
All steroid analyses were performed in a single batch for each study. Each sample was measured with the standard nonderivatized as well as ultrasensitive estrogen-specific derivatization LC-MS method.
The standard nonderivatization LC-MS method and validation are described elsewhere originally [21], with modification in switching from high pressure to ultrapressure liquid chromatography [22]. For Study 1, serum testosterone (T), dihydrotestosterone (DHT), estradiol (E2), estrone (E1), and dehydroepiandrosterone (DHEA) were measured. Subsequently, for Study 2 samples, this profile was extended to also quantify 11 keto testosterone (11KT), 11 keto dihydrotestosterone (11KDHT), 11 keto androstenedione (11KA), and androstenedione (A4) [22]. All steroids were measured in extracts from a single serum sample stored at -80°C until analysis. All steroid standards and internal standards (IS) were obtained from recognized suppliers of high-quality steroids (National Measurement Institute, Sydney, Australia; Steraloids, Newport, RI; Cerilliant, Round Rock, TX; Cambridge Isotopes, Tewksbury, MA; RTI International, Research Triangle Park, NC; Sapphire Biosciences, Redfern, Sydney, Australia). All IS were deuterated (d2-4) isotopes and certified reference materials were used as assay standards where available.
Aliquots of serum (200 µL human, 100 µl mice), standards and quality control samples fortified with 50µl of steroid IS, were extracted by adding 1 ml of methyl tert-butyl ether with vigorous mixing (1 minute) to extract steroids into the organic layer. After phase separation and freezing the lower aqueous layer, the upper organic layer was decanted into clean glass tubes and solvent evaporated at 37°C overnight. Dried extracts were resuspended in 75 µl of 20% methanol:water, with 50 µL transferred into a 96-well microtitre plate for injection into the LC-MS equipment. Ultrapressure liquid chromatography conditions comprised a methanol/water gradient on a Kinetex Phenyl Hexyl column (100 mm × 2.1 mm × 1.7 µm) with a Phenomenex guard cartridge at a column temperature of 45°C and flow rate of 0.35 mL/minute. The LC running conditions were split into 3 periods to match the ionization requirements of the steroids and provided the baseline separation for each steroid with a total run time of 13 minutes. The run times were, in the first period for the keto-androgens (positive ionization) 11KT (4.49 minutes), 11KA (4.68 minutes), and 11KDHT (5.14 minutes), in the second period for estrogens (negative ionization) with E2 (5.74 minutes) and E1 (6.56 minutes), and in the third period (positive ionization) testosterone (7.07 minutes), A4 (7.25 minutes), DHEA (7.39 minutes), and DHT (8.73 minutes). The eluant was introduced into the mass spectrometer without splitting.
Mass spectrometry was performed on an API-5000 triple-quadrupole mass spectrometer (Applied Biosystems/MDS SCIEX, Foster City, CA [and Concord, Ontario, Canada]) equipped with an atmospheric pressure photoionization source capable of fast switching between positive and negative ion modes. Multiple reaction monitoring with qualifier and quantifier transitions was used to quantify the steroids. Unit mass resolution was set in both mass-resolving quadrupole Q1 and Q3. The quantifier transitions and collision energies (CE, eV) were testosterone (289->109, 35), DHT (273->123, 31), E2 (271->145, -57), E1 (269->145, -57), DHEA (253->197, 28), 11KT (306->123, 40), 11KDHT (308->190, 26), 11KA (301.5->121, 30), and A4 (287.1->97.1, 34). For all analytes, the external recovery was 88% to 96% and the matrix effect was 85% to 93%. The limits of detection (LOD) and limits of quantitation (LOQ), between-run (15 replicates) and within-run (10 replicates) reproducibility (range of the percentage of coefficient of variation [CV] over 3 QC levels) for each analyte were T (10 and 25 pg/ml, 2–9%, 4–8%), DHT (50 and 100 pg/ml, 4–12%, 4–9%), E2 (2.5 and 5 pg/ml, 4–8%, 5–8%), E1 (2.5 and 5 pg/ml, 3–6%, 5–9%), DHEA (20 and 50 pg/L, 3–6%, 8–12%), 11KT (20 and 50 pg/mL, 3–9%, 4–8%), 11KDHT (50 and 100 pg/ml, 5–7%, 5–7%), 11KA (20 and 50 pg/mL, 3–8%, 5–7%), and A4 (30 and 50 pg/ml, 5–9%, 4–12%).
The ultrasensitive LC-MS method uses a novel, estrogen-selective derivatizing reagent 1,2-dimethylimidazole-5-sulfonyl chloride (DMIS), as described and validated elsewhere [17] using the same regents, extraction, and ultrapressure mass spectrometry as the method described above. This method has an LOD, LOQ, and CV of E2 (0.25 and 0.50 pg/ml), E1 (0.25 and 0.50 pg/ml), and T (10 and 25 pg/ml).
Validation of sex steroid measurement in mouse serum samples was performed by spike-recovery experiments measuring endogenous steroids (E2, E1, T) before and then after fortification with 4 levels of spike (very low, low, mid-, and high range) in quadruplicate (n = 16 replicates). Initial validation, using pooled male and female sera, displayed high accuracy and reproducibility, respectively, for E2 (107%, 1.2–4.7%), for E1 (106%, 0.8–3.9%), and for T (111%, 3.4–5.4%). Additional validation for male sera (n = 16) confirmed similar high accuracy and reproducibility for E2 (103%, 1.2–6.9%), for E1 (99%, 1.6–5.7%), and for T (99%, 2.5–8.0%).
Data analysis
Data were described as mean and standard error of the mean and compared between 2 groups by an unpaired t-test (comparing 2 groups of women), analysis of variance (comparing 4 estrus cycle stages in mice), or analysis of covariance (to evaluate confounder or modifier variables in an analysis of 2 groups of women), as required, using NCSS 2019 software (NCSS, Kaysville, UT). Imputation for undetectable concentrations of serum T, E2, and E1 was performed using a validated substitution method for circulating sex steroids [23].
Results
In Study 1, the ultrasensitive method was able to quantify serum E2 in 102/107 (95%) of samples from AI-treated post-menopausal women (Table 1). By contrast, the conventional nonderivatization method was able to quantify serum E2 in only 46/107 (43%) and serum E1 in 5/107 (5%) of the same serum samples. Neither serum E2 or T were significantly correlated with weight, height, body mass index (BMI), pulse rate, orsystolic or diastolic blood pressure, nor was prior tamoxifen, oophorectomy, or antidepressant use associated with any differences in serum E2 or T (data not shown). There were positive correlations between weight or BMI, respectively, with systolic (r = 0.51, P < 0.001; r = 0.57, P = 0.001) and diastolic (r = 0.32, P < 0.001; r = 0.38, P < 0.001) blood pressure. Serum E1 was not measured in Study 1.
Table 1.
Ultrasensitive and conventional LC-MS measurements in Studies 1 and 2
| Study 1 | Study 2 | |||||||
|---|---|---|---|---|---|---|---|---|
| Treatment | Detectable N (%) | Mean ± SEM | Median (IQR) | Treatment | Detectable N (%) | Mean ± SEM | Median (IQR) | |
| Standard E2 (pg/ml) | AI + Placebo | 19/46 (41) | 4.1 ± 0.4 | 3.6 (3.0, 5.0) | Control | 4/42 (10) | 15 ± 4 | 15 (8.2, 22) |
| AI + Testosterone | 27/61 (44) | 4.3 ± 0.3 | 4.1 (3.3, 4.7) | AI | 1/46 (2) | 19 | 19 | |
| AI Total | 46/107 (43) | 4.3 ± 0.2 | 3.9 (3.3, 4.8) | Total | 5/88 (6) | 16 ± 3 | 15 (10, 21) | |
| Ultrasensitive E2 (pg/ml) | AI + Placebo | 42/46 (91) | 3.5 ± 0.4 | 3.1 (1.1, 6.2) | Control | 41/42 (98) | 3.8 ± 1 | 2.0 (1.3, 4.7) |
| AI + Testosterone | 60/61 (98) | 4.8 ± 0.3 | 4.6 (3.0, 6.9) | AI | 28/46 (61) | 1.5 ± 1 | 0.4 (0.3, 0.6) | |
| AI Total | 102/107 (95) | 4.2 ± 0.3 | 4.0 (1.9. 6.4) | Total | 69/88 (78) | 2.9 ± 1 | 1.4 (0.5, 3.7) | |
| Standard E1(pg/ml) | AI + Placebo | 2/46(4) | 23 ± 18 | 23 (4.7, 41) | Control | 39/42 (93) | 23 ± 3 | 20 (12, 29) |
| AI + Testosterone | 3/61 (5) | 25 ± 22 | 2.5 (1.7, 70) | AI | 12/46 (26) | 8.6 ± 0.9 | 8.5 (5.8, 11) | |
| AI Total | 5/107 (5) | 24 ± 14 | 4.7 (2.1, 55) | Total | 51/88 (58) | 20 ± 2 | 16 (9, 26) | |
| Ultrasensitive E1 (pg/ml)a | AI + Placebo | -- | -- | -- | Control | 42/42 (100) | 23 ± 2 | 21 (12, 29) |
| AI + Testosterone | -- | -- | -- | AI | 43/46 (93) | 5.0 ± 0.5 | 4.8 (2.7, 7.0) | |
| AI Total | -- | -- | -- | Total | 85/88 (97) | 13.8 ± 1.5 | 8.0 (4.1, 21) | |
| Standard T (ng/ml) | AI + Placebo | 46/46 (100) | 0.12 ± 0.01 | 0.12 (0.91, 0.16) | Control | 42/42 (100) | 0.06 ± 0.01 | 0.05 (0.04, 0.08) |
| AI + Testosterone | 61/61 (100) | 0.14 ± 0.01 | 0.13 (0.11, 0.15) | AI | 46/46 (100) | 0.05 ± 0.0 | 0.05 (0.04, 0.07) | |
| AI Total | 107/107 (100) | 0.14 ± 0.01 | 0.12 (0.11, 0.15) | Total | 88/88 (100) | 0.06 ± 0.0 | 0.05 (0.04, 0.07) | |
| Ultrasensitive T (ng/ml) | AI + Placebo | 46/46 (100) | 0.17 ± 0.01 | 0.18 (0.12, 0.22) | Control | 42/42 (100) | 0.07 ± 0.01 | 0.06 (0.05, 0.08) |
| AI + Testosterone | 61/61 (100) | 0.19 ± 0.01 | 0.18 (0.15, 0.20) | AI | 46/46 (100) | 0.06 ± 0.0 | 0.05 (0.04, 0.07) | |
| AI Total | 107/107 (100) | 0.18 ± 0.01 | 0.18 (0.14, 0.21) | Total | 88/88 (100) | 0.06 ± 0.0 | 0.060.04, 0.08) | |
Standard refers to conventional nonderivatized LC-MS measurement. Ultrasensitive refers to estrogen-selective derivatization LC-MS measurement. “--” indicates ND, not done.
Abbreviations: AI, aromatase inhibitor; IQR, interquartile range (Q1, Q3); ND, not done; SEM, standard error of mean.
a Ultrasensitive E1 was not measured in Study 1.
In study 2, the ultrasensitive method was able to quantify serum E2 in 41/42 (98%) and serum E1 in 42/42 (100%) of samples from healthy postmenopausal women compared with 28/46 (61%) for serum E2 and 43/46 (93%) for serum E1 in serum samples from AI-treated postmenopausal women (Table 1). By contrast, the conventional nonderivatization method could quantify serum E2 in only 4/42 (10%) of healthy postmenopausal women and 2/46 (4%) of serum samples from AI-treated postmenopausal women.
Aromatase inhibitor treatment was associated with an 85% lower median serum E2 (0.3 pg/ml vs 2.0 pg/ml; Fig. 1) and 80% for serum E1 (4.2 pg/ml vs 21 pg/ml, Fig. 2). There was no significant difference in serum E2, E1, or T between women taking letrozole (n = 32), anastrozole (n = 7), or exemestane (n = 6) nor—among women taking letrozole—any difference according to prior cytotoxic drug treatment or not (data not shown).
Figure 1.
Combination box and dot plot of serum E2 measured by ultrasensitive LC-MS method (using DMIS derivatization) in healthy postmenopausal women (left plot) and in postmenopausal women with breast cancer undergoing adjuvant treatment including aromatase inhibitor drugs. Note the logarithmic y-axis scale. The box covers the interquartile range at its extremities, the median at its notched narrowest width and whiskers length at 1.5 times the interquartile range. Each dot represents a single serum sample. To convert E2 concentrations (pg/ml) to SI units (pmol/L), multiply E2 by 3.67.
Figure 2.
Combination box and dot plot of serum E1 measured by ultrasensitive LC-MS method (using DMIS derivatization) in healthy postmenopausal women (left plot) and in postmenopausal women with breast cancer undergoing adjuvant treatment including aromatase inhibitor drugs. Note the logarithmic y-axis scale. The box covers the interquartile range at its extremities, the median at its notched narrowest width and whiskers length at 1.5 times the interquartile range. Each dot represents a single serum sample. To convert E1 concentrations (pg/ml) to SI units (pmol/L), multiply E1 by 3.7.
In the ultrasensitive method and pooling samples from both studies (Table 1), serum E2 was quantifiable in 171/195 (88%) samples with the median serum E2 of 2.7 pg/ml (IQR 0.6, 5.3 pg/ml).
Using the conventional nonderivatized LC-MS steroid profile and pooling samples from both studies (Table 2), AI treatment was associated with significant reduction in serum 11KDHT (0.14 ± 0.07 ng/ml vs 0.26 ± 0.04 ng/ml, P = 0.009), significant increase in serum 11KA4 (2.38 ± 1.25 ng/ml vs 1.53 ± 0.68 ng/ml, P < 0.001) but no significant changes in serum 11KT (0.45 ± 0.04 ng/ml vs 0.43 ± 0.03 ng/ml, P = 0.58), DHT (P = 0.16), DHEA (0.39), or A4 (P = 0.36).
Table 2.
Serum steroids data pooled from Study 1 and Study 2 measured with the nonderivatized LC-MS method
| Serum | Treatment | Detectable (%) | Mean ± SEM | Median (IQR) |
|---|---|---|---|---|
| 11KT (ng/ml) | Control | 42/42 (100) | 0.43 ± 0.03 | 0.39 (0.25,0.56) |
| Aromatase inhibitor | 153/153(100) | 0.45 ± 0.04 | 0.39 (0.29, 0.53) | |
| 11KDHT (ng/ml) | Control | 42/42 (100) | 0.26 ± 0.04 | 0.16 (0.11,0.22) |
| Aromatase inhibitor | 61/153 (40) | 0.14 ± 0.01 | 0.13 (0.09,0.17) | |
| 11KA4 (ng/ml) | Control | 42/42 (100) | 1.53 ± 0.10 | 1.35 (1.06,1.92) |
| Aromatase inhibitor | 153/153 (100) | 2.38 ± 0.10 | 2.11 (1.40, 3.11) | |
| DHT (ng/ml) | Control | 35/42 (83) | 0.56 ± 0.26 | 0.26 (0.12,0.43) |
| Aromatase inhibitor | 118/153 (77) | 0.35 ± 0.02 | 0.29 (0.15, 0.50) | |
| DHEA (ng/ml) | Control | 42/42 (100) | 0.79 ± 0.07 | 0.71 (0.48, 1.06) |
| Aromatase inhibitor | 153/153 (100) | 0.86 ± 0.04 | 0.77 (0.52, 1.18) | |
| A4 (ng/ml) | Control | 42/42 (100) | 0.22 ± 0.02 | 0.22 (0.15, 0.28) |
| Aromatase inhibitor | 153/153 (100) | 0.24 ± 0.01 | 0.22 (0.17, 0.30) |
In the first mouse study (Table 3), in female samples, serum E2 was detectable in 65%, estrone in 76% ,and T in 97% of samples, regardless of estrus cycle stage. In male serum samples, T was detectable in 100% and E1 in 33% of samples, but E2 was undetectable (0.25 pg/ml) in all 27 samples.
Table 3.
Sex steroids in mice measured with ultrasensitive estrogen-selective derivatization LC-MS
| Mouse Study 1 | Testosterone (T) | Estradiol (E2) | Estrone (E1) |
|---|---|---|---|
| Female (n = 87) | |||
| n (% detectable) | 84 (97%) | 55 (65%) | 64 (76%) |
| Mean ± SD | 0.28 ± 0.25 | 6.56 ± 9.46 | 3.00 ± 3.44 |
| Median (IQR) | 0.19 (0.11, 0.39) | 2.08 (0.12, 10.7) | 2.04 (0.12, 4.36) |
| Male (n = 27) | |||
| n (% detectable) | 27 (100%) | 0 (0%) | 9 (33%) |
| Mean ± SD | 7.88 ± 8.73 | < 0.25 | 2.93 ± 1.99 |
| Median (IQR) | 4.21 (1.81, 13.0) | < 0.25 | 3.74 (0.77, 4.75) |
| Mouse Study 2 | |||
| Diestrus | |||
| n (% detectable) | 10 (90%) | 10 (20%) | 10 (0%) |
| Mean ± SD | 0.12 ± 0.12 | 0.79 ± 1.59 | < 0.25 |
| Median (IQR) | 0.09 (0.02, 0.19) | 0.18 (0.18, 0.46) | < 0.25 |
| Proestrus | |||
| n (% detectable) | 10 (70%) | 10 (100%) | 10 (80%) |
| Mean ± SD | 0.05 ± 0.08 | 19.4 ± 13.5 | 5.2 ± 5.9 |
| Median (IQR) | 0.01 (0.01, 0.06) | 13.8 (9.7, 30) | 2.9 (0.7, 9.8) |
| Estrus | |||
| n (% detectable) | 11 (91%) | 11 (9%) | 11 (9%) |
| Mean ± SD | 0.03 ± 0.02 | <0.25 | <0.25 |
| Median (IQR) | 0.03 (0.02, 0.04) | <0.25 | <0.25 |
| Metestrus | |||
| n (% detectable) | 9 (100%) | 9 (89%) | 9 (44%) |
| Mean ± SD | 0.06 ± 0.06 | 9.2 ± 11.2 | 2.2 ± 3.6 |
| Median (IQR) | 0.03 (0.02, 0.07) | 5.0 (3.3, 12.5) | 0.18 (0.18, 3.1) |
Undetectable values were imputed as described in [1] to estimate mean, standard deviation (SD) ,and interquartile range (IQR, 1st and 3rd quartiles)
In a second mouse study evaluating estrus cycle stage in mature female mice (Table 3), serum E2 and E1 were both significantly higher in proestrus than in estrus or diestrus but not different from metestrus (Fig. 3). Serum T did not differ significantly between estrus cycle stages.
Figure 3.
Combination box and dot plots of serum E2 (left panel), serum E1 (middle panel), and T (right panel) measured by ultrasensitive LC-MS method (using DMIS derivatization) in sexually mature female mice across the 4 stages of the estrus cycle. Note the logarithmic y-axis scale. The box covers the interquartile range at its extremities, the median at its notched narrowest width and whiskers length at 1.5 times the interquartile range. Each dot represents a single serum sample. To convert concentrations to SI units multiply E2 (pg/ml) by 3.67 (pmol/L), E1 (pg/ml) by 3.7 (pmol/L), and T (ng/ml) by 3.47 (nmol/L).
Discussion
The present study demonstrates that the estrogen-selective derivatization using the novel derivatizing regent DMIS allows ultrasensitive measurement by LC-MS of serum E2 and E1 in 98% and 100%, respectively, of healthy postmenopausal women using only a small (0.2 ml) serum sample. This method with an LOD of 0.25 pg/ml (detecting 50 fg E2 on the column with 0.2 ml samples) improves on our nonderivatized LC-MS method featuring LOD limits (3–5 pg/ml), but which can quantify serum E2 in less than half of the same serum samples [12]. Further studies comparing the sensitivity and practical application of this ultrasensitive method with others reporting similar sensitivity [9, 24] to postmenopausal women, notably those on AI treatment, as well as to mice are desirable.
Using this ultrasensitive method, the present study also shows that AI treatment reduces median serum E2 by 85% and serum E1 by 80% in postmenopausal women undergoing adjuvant AI treatment for breast cancer. Although a similar extent of serum E2 suppression was reported for letrozole [25] and anastrozole [26] in short-term clinical studies using an estradiol immunoassay with preassay solvent extraction but no chromatography [27], the present studies were not powered to test noninferiority. Yet, serum E2 concentrations varied widely between women so that some had little or no apparent suppression of serum E2. This raises the question of whether these represent differences in efficacy of the prescribed AI drugs. It may be speculated that the variations in clinical efficacy, resistance, and toxicity of adjuvant AI drug treatment in women may reflect the degree of sustained suppression of serum E2 achieved by the AI treatment. This may be influenced by adherence to a prescribed adjuvant regimen while being asymptomatic, thereby lacking symptom reinforcement. Alternatively, pharmacogenetic differences in AI drug absorption, metabolism, or enzyme inhibition may be contributing factors that warrant further analysis. Finally, duration of drug-induced enzyme inhibition may also vary, as blood sampling was not timed to the taking of the AI drugs. In any case, accurate quantification of serum E2 is an essential prerequisite to identify and/or differentiate these possibilities.
The present study demonstrates that this ultrasensitive method can quantify serum E2 in most female rodents with widely varying concentrations between mice due to variations in the estrus cycle stage [16], as rising circulating E2 stimulates LH surge, the trigger to ovulation. The present findings confirm those of a previous report based on gas chromatography-mass spectrometry [16], with peaks of serum E2 at proesterus and similar high levels at metestrus but much lower levels at diestrus and estrus. This application used only 0.1 ml serum samples consistent with the requirement for most mouse studies to measure multiple analytes in the limited volume of terminal blood obtainable. Yet, using the present ultrasensitive method serum E2 remains undetectable (< 0.25 pg/ml) in all 27 male serum samples consistent with a previous report of an ultrasensitive method [16]. Our demonstration of high recovery and accuracy of E2 spiked into male mouse sera indicates that the undetectable measurements reflect endogenous circulating E2 in male mice being lower than the detection limit of this assay (1 pmol/l) and not an artefact in male mouse serum. These serum E2 concentrations are much lower than the concentrations required to activate ER (1–10 nmol/l) [28, 29]. This implies that circulating E2 is unlikely to exert direct endocrine effects on bone or other estrogen-sensitive male mouse organs or tissues. However, it is well established that aromatization of T to E2 is critical to testosterone action on bone and brain [30]. Hence, the present findings imply that the site of aromatization for testosterone-derived estrogen action on male bone must be located within bone to create high enough E2 concentrations to act locally on ERs, but with minimal spillover into the general circulation, as previously inferred [31]. Further, it is also possible that regional aromatase expression in bone may serve as a local control, paracrine diversification mechanism for androgen action as it does in the brain [32, 33]. Although circulating E2 (usually by direct [nonextraction] immunoassay) in men is correlated with various age-related bone changes in men [34], the circulating levels are much lower than in premenopausal women and may also reflect unregulated spillover rather than a direct endocrine mechanism of androgen action on male bone.
The present study also demonstrates that all three 11keto androgens (11 keto testosterone, 11 keto dihydrotestosterone, 11 keto androstenedione), which are measurable in premenopausal women [22, 35], remain detectable in postmenopausal women. Furthermore, they are not lower in women taking AI drugs consistent with their adrenal origins [36].
Although AI drugs were introduced primarily for adjuvant treatment of breast cancer, there are other off-label and mostly unproven indications for these drugs including female and male infertility, induction of ovulation and male hypogonadism. Furthermore, AI drugs may have important side-effects such as bone loss [31] and male sexual dysfunction [37] when used for off-label applications. Appraisal of whether the degree of suppression of circulating E2 is a determinant of the efficacy or side-effects of AI drug treatment in these settings is now feasible with methods such as the ones used presently and other [9] ultrasensitive MS-based estrogen assays.
Acknowledgments
S.R.D. is an NHMRC Senior Principal Research Fellow (Grant no 1135843). E.G. was supported by a University International Postgraduate Award and Science Writing scholarships from Faculty of Science, University of NSW. The authors thank Professor Fiona Stapleton (UNSW Sydney) and Associate Professor Rachael Dear (Kinghorn Cancer Centre) for input for Study 2. The authors are grateful to Melissa Edwards and Valentina Rodriguez Paris for excellent technical assistance.
Additional Information
Disclaimer Summary: There was no external funding for this study. Study 1 was supported by an Accelerator Grant of the National Breast Cancer Foundation (Australia) Grant #: NT-16-006. Study 2 was supported by a University International Postgraduate Award, UNSW Sydney. D.J.H. has received institutional grants without personal income for investigator-initiated androgen pharmacology studies and has provided expert testimony to antidoping and professional standards tribunals and testosterone litigation. No other authors have any relevant declarations.
Data Availability: Restrictions apply to the availability of data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will, on request, detail the restrictions and any conditions under which access to some data may be provided.
References
- 1. Anderson WF, Chatterjee N, Ershler WB, Brawley OW. Estrogen receptor breast cancer phenotypes in the surveillance, epidemiology, and end results database. Breast Cancer Res Treat. 2002;76(1):27–36. [DOI] [PubMed] [Google Scholar]
- 2. Kaklamani VG, Gradishar WJ. Endocrine therapy in the current management of postmenopausal estrogen receptor-positive metastatic breast Cancer. Oncologist. 2017;22(5):507–517. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Jordan VC. Tamoxifen (ICI46,474) as a targeted therapy to treat and prevent breast cancer. Br J Pharmacol. 2006;147(Suppl 1):S269–S276. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. van Hellemond IEG, Geurts SME, Tjan-Heijnen VCG. Current status of extended adjuvant endocrine therapy in early stage breast cancer. Curr Treat Options Oncol. 2018;19(5):26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Rani A, Stebbing J, Giamas G, Murphy J. Endocrine resistance in hormone receptor positive breast cancer-from mechanism to therapy. Front Endocrinol (Lausanne). 2019;10:1–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Yang G, Nowsheen S, Aziz K, Georgakilas AG. Toxicity and adverse effects of Tamoxifen and other anti-estrogen drugs. Pharmacol Ther. 2013;139(3):392–404. [DOI] [PubMed] [Google Scholar]
- 7. Hertz DL, Henry NL, Rae JM. Germline genetic predictors of aromatase inhibitor concentrations, estrogen suppression and drug efficacy and toxicity in breast cancer patients. Pharmacogenomics. 2017;18(5):481–499. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Rosner W, Hankinson SE, Sluss PM, Vesper HW, Wierman ME. Challenges to the measurement of estradiol: an endocrine society position statement. J Clin Endocrinol Metab. 2013;98(4):1376–1387. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Wang Q, Mesaros C, Blair IA. Ultra-high sensitivity analysis of estrogens for special populations in serum and plasma by liquid chromatography-mass spectrometry: Assay considerations and suggested practices. J Steroid Biochem Mol Biol. 2016;162:70–79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Hertz DL, Speth KA, Kidwell KM, et al. Variable aromatase inhibitor plasma concentrations do not correlate with circulating estrogen concentrations in post-menopausal breast cancer patients. Breast Cancer Res Treat. 2017;165(3):659–668. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Handelsman DJ, Newman JD, Jimenez M, McLachlan R, Sartorius G, Jones GR. Performance of direct estradiol immunoassays with human male serum samples. Clin Chem. 2014;60(3):510–517. [DOI] [PubMed] [Google Scholar]
- 12. Davis SR, Bell RJ, Robinson PJ, et al. ; ASPREE Investigator Group Testosterone and estrone increase from the age of 70 years: findings from the sex hormones in older women study. J Clin Endocrinol Metab. 2019;104(12):6291–6300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Haisenleder DJ, Schoenfelder AH, Marcinko ES, Geddis LM, Marshall JC. Estimation of estradiol in mouse serum samples: evaluation of commercial estradiol immunoassays. Endocrinology. 2011;152(11):4443–4447. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. McNamara KM, Harwood DT, Simanainen U, Walters KA, Jimenez M, Handelsman DJ. Measurement of sex steroids in murine blood and reproductive tissues by liquid chromatography-tandem mass spectrometry. J Steroid Biochem Mol Biol. 2010;121(3–5):611–618. [DOI] [PubMed] [Google Scholar]
- 15. Toran-Allerand CD, Tinnikov AA, Singh RJ, Nethrapalli IS. 17alpha-estradiol: a brain-active estrogen? Endocrinology. 2005;146(9):3843–3850. [DOI] [PubMed] [Google Scholar]
- 16. Nilsson ME, Vandenput L, Tivesten Å, et al. Measurement of a comprehensive sex steroid profile in rodent serum by high-sensitive gas chromatography-tandem mass spectrometry. Endocrinology. 2015;156(7):2492–2502. [DOI] [PubMed] [Google Scholar]
- 17. Keski-Rahkonen P, Desai R, Jimenez M, Harwood DT, Handelsman DJ. Measurement of estradiol in human serum by LC-MS/MS using a novel estrogen-specific derivatization reagent. Anal Chem. 2015;87(14):7180–7186. [DOI] [PubMed] [Google Scholar]
- 18. Davis SR, Robinson PJ, Jane F, White S, White M, Bell RJ. Intravaginal testosterone improves sexual satisfaction and vaginal symptoms associated with aromatase inhibitors. J Clin Endocrinol Metab. 2018;103(11):4146–4154. [DOI] [PubMed] [Google Scholar]
- 19. Gibson E, Stapleton F, Dear R, Wolffsohn JS, Golebiowski B. Dry eye signs and symptoms in aromatase inhibitor treatment and the relationship with pain. Ocul Surf. 2019;18:108–113. [DOI] [PubMed] [Google Scholar]
- 20. Bertoldo MJ, Caldwell ASL, Riepsamen AH, et al. A hyperandrogenic environment causes intrinsic defects that are detrimental to follicular dynamics in a PCOS mouse model. Endocrinology. 2019;160(3):699–715. [DOI] [PubMed] [Google Scholar]
- 21. Harwood DT, Handelsman DJ. Development and validation of a sensitive liquid chromatography-tandem mass spectrometry assay to simultaneously measure androgens and estrogens in serum without derivatization. Clin Chim Acta. 2009;409(1–2):78–84. [DOI] [PubMed] [Google Scholar]
- 22. Skiba MA, Bell RJ, Islam RM, Handelsman DJ, Desai R, Davis SR. Androgens during the reproductive years: what is normal for women? J Clin Endocrinol Metab. 2019;104(11):5382–5392. [DOI] [PubMed] [Google Scholar]
- 23. Handelsman DJ, Ly LP. An accurate substitution method to minimize left censoring bias in serum steroid measurements. Endocrinology. 2019;160(10):2395–2400. [DOI] [PubMed] [Google Scholar]
- 24. Bertelsen BE, Kellmann R, Viste K, et al. An ultrasensitive routine LC-MS/MS method for estradiol and estrone in the clinically relevant sub-picomolar range. J Endocr Soc. 2020;4(6):1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Iveson TJ, Smith IE, Ahern J, Smithers DA, Trunet PF, Dowsett M. Phase I study of the oral nonsteroidal aromatase inhibitor CGS 20267 in healthy postmenopausal women. J Clin Endocrinol Metab. 1993;77(2):324–331. [DOI] [PubMed] [Google Scholar]
- 26. Banerjee S, Pancholi S, A’hern R, et al. The effects of neoadjuvant anastrozole and tamoxifen on circulating vascular endothelial growth factor and soluble vascular endothelial growth factor receptor 1 in breast cancer. Clin Cancer Res. 2008;14(9):2656–2663. [DOI] [PubMed] [Google Scholar]
- 27. Dowsett M, Goss PE, Powles TJ, et al. Use of the aromatase inhibitor 4-hydroxyandrostenedione in postmenopausal breast cancer: optimization of therapeutic dose and route. Cancer Res. 1987;47(7):1957–1961. [PubMed] [Google Scholar]
- 28. Blair RM, Fang H, Branham WS, et al. The estrogen receptor relative binding affinities of 188 natural and xenochemicals: structural diversity of ligands. Toxicol Sci. 2000;54(1):138–153. [DOI] [PubMed] [Google Scholar]
- 29. Zhu BT, Han GZ, Shim JY, Wen Y, Jiang XR. Quantitative structure-activity relationship of various endogenous estrogen metabolites for human estrogen receptor alpha and beta subtypes: Insights into the structural determinants favoring a differential subtype binding. Endocrinology. 2006;147(9):4132–4150. [DOI] [PubMed] [Google Scholar]
- 30. Russell N, Grossmann M. Mechanisms in endocrinology: estradiol as a male hormone. Eur J Endocrinol. 2019;181(1):R23–R43. [DOI] [PubMed] [Google Scholar]
- 31. Vanderschueren D, Laurent MR, Claessens F, et al. Sex steroid actions in male bone. Endocr Rev. 2014;35(6):906–960. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Naftolin F. Brain aromatization of androgens. J Reprod Med. 1994;39(4):257–261. [PubMed] [Google Scholar]
- 33. Biegon A. In vivo visualization of aromatase in animals and humans. Front Neuroendocrinol. 2016;40:42–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Khosla S, Melton LJ 3rd, Riggs BL. Clinical review 144: estrogen and the male skeleton. J Clin Endocrinol Metab. 2002;87(4):1443–1450. [DOI] [PubMed] [Google Scholar]
- 35. O’Reilly MW, Kempegowda P, Jenkinson C, et al. 11-oxygenated C19 steroids are the predominant androgens in polycystic ovary syndrome. J Clin Endocrinol Metab. 2017;102(3):840–848. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Turcu A, Smith JM, Auchus R, Rainey WE. Adrenal androgens and androgen precursors-definition, synthesis, regulation and physiologic actions. Compr Physiol. 2014;4(4):1369–1381. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Finkelstein JS, Lee H, Burnett-Bowie SA, et al. Gonadal steroids and body composition, strength, and sexual function in men. N Engl J Med. 2013;369(11):1011–1022. [DOI] [PMC free article] [PubMed] [Google Scholar]