Dynamic change of estrogen and progesterone metabolites in human urine during pregnancy

Abstract

Endogenous estrogen and progesterone and their metabolites play a key role in regulating and maintaining the normal pregnancy process. However, the dynamic change of these estrogen and progesterone metabolites’ level across the entire gestational period is not fully revealed. This study systematically measures the temporal changes of estrogen, progesterone, and their metabolites in human urine during normal pregnancy using ultrahigh performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). We find that the levels of estrone, estradiol, estriol, 16-epiestriol, 17-epiestriol, 2-methoxyestradiol, and 4-hydroxyestrone gradually increase during pregnancy. The levels of 2-hydroxyestrone, 2-hydroxyestradiol, and 4-hydroxyestradiol rapidly decrease in the early pregnancy and then maintain at lower levels. The levels of 4-methoxyestradiol, 4-methoxyestrone and 2-methoxyestrone peak in the mid-pregnancy and then gradually decrease. The levels of pregnenolone, 17α-hydroxy pregnenolone, 17α-hydroxy progesterone, pregnanolone, epipregnanolone gradually increase during pregnancy. The levels of progesterone, 20α-hydroxy progesterone, 5α-dihydroprogesterone and 5β-dihydroprogesterone first increase in the mid pregnancy and then decrease in the late pregnancy. In sum, this study comprehensively depicts the dynamic change of estrogen, progesterone, and their metabolites in human urine during pregnancy, which lays the foundation for hormone monitoring and early diagnosis of pregnancy complications, and further mechanistic study of the roles of these metabolites in pregnancy.

Introduction

Estrogen, a class of important sex hormones present in vertebrates, plays vital roles in the growth, development, and function of multiple organ systems, especially in the reproductive system^1,2, cardiovascular system³, endocrine system^4–6, and nervous system^7–9. Estrogen has diverse metabolism pathways and different estrogen metabolites are produced in these pathways¹⁰ (illustrated in Fig. 1). Progesterone is the precursor for the synthesis of androgen and estrogen in the body^11,12, which also plays important roles in the regulation of diverse organ systems^13–15. The metabolism of progesterone is rapid and extensive¹⁶ (illustrated in Fig. 2). The main metabolic pathways of progesterone involve reduction by 5α-reductase¹⁷ and 5β-reductase into 5α-dihydroprogesterone and 5β-dihydroprogesterone, respectively¹⁸. It is of note that a number of studies have found that some estrogen and progesterone metabolites have unique biological functions which are different from estradiol and progesterone, the two well-studied estrogen and progestin^19–22.

Fig. 1.

Estrogen metabolism pathway in humans. The red colored 14 estrogen metabolites were measured by UPLC-MS/MS in this study. 17β-HSD, 17β-hydroxysteroid dehydrogenase; CYP450, Cytochromes P450; COMT, Catechol-O-methyltransferase.

Fig. 2.

Progesterone metabolism in humans. The red colored 9 progesterone metabolites were measured by UPLC-MS/MS in this study. 3β-HSD, 3β-hydroxysteroid dehydrogenase; 3α-HSD, 3α-hydroxysteroid dehydrogenase; 20β-HSD, 20β-hydroxysteroid dehydrogenase.

Endogenous estrogen and progesterone and some of their metabolites play a key role in regulating and maintaining the normal pregnancy process. Estradiol (E₂) plays key roles in placental development and function^23–25, and fetal growth^26,27. They also significantly affect the maternal circulatory system²⁸, metabolic regulation²⁹, and immune function³⁰. Progesterone is crucial for maintaining a stable uterine environment, preventing early labor, and supporting the normal function of the placenta³¹.

As a result, the levels of estrogen, progesterone and their metabolite can serve as an important indicator for assessing normal pregnancy³². It is widely recognized that the concentrations of estrogen and progesterone in maternal serum and urine increase significantly throughout pregnancy³³. During the first nine weeks of pregnancy, the corpus luteum in maternal ovaries and adrenal cortex jointly maintain circulating concentrations of maternal estrogen and progesterone^32,34. Subsequently, the placenta gradually becomes the main source of maternal steroids³⁵.

Changes in the levels of estrogen and its metabolites such as estrone (E₁) and 2-methoxyestradiol (2-MeO-E₂) may indicate the risk of pregnancy complications, such as gestational hypertension³⁶, pre-eclampsia³⁷, and gestational diabetes³⁸. In addition, abnormal changes in progesterone levels are closely related to various pregnancy-related complications, such as miscarriage^39,40, preterm birth^41,42, pre-eclampsia⁴³, and fetal growth restriction⁴⁴. Understanding the changes in these hormones during pregnancy can better elucidate the normal physiological processes of pregnancy and mechanisms of potential abnormalities^45,46. Therefore, early identification of these changes can help predict the occurrence of pregnancy complications, thereby enabling preventative or interventional measures to improve pregnancy outcomes.

It is revealed that the change of steroid hormone metabolite levels in urine has a parallel pattern with that in the bloodstream⁴⁷. This suggests that urinary hormone measurements could serve as an alternative, noninvasive and convenient method to track hormonal changes in serum. Previous studies have detected changes in hormone concentrations in serum or urine during pregnancy^48–51. These studies analyzed 2 to 14 different types of metabolites. In addition, their focus is mainly on different stages of pregnancy, and a systematic overview of dynamic changes of hormone level across the entire gestational period is lacking. To address these gaps, this study aims to investigate the dynamics of hormone changes in urine during pregnancy, which include the levels of 14 types of estrogen and its metabolites and 9 types of progesterone metabolites in pregnant women at different stages of pregnancy.

Materials and methods

Chemicals and reagents

Fourteen estrogen and its metabolites were detected in this study, and the standard chemicals were obtained from different commercial suppliers, which include estrone (E₁) (Macklin, Shanghai, China), estradiol (E₂) (Macklin, Shanghai, China), estriol (E₃) (Macklin, Shanghai, China), 16-epiestriol (16-epiE₃) (ChemeGen, Shanghai, China), 17-epiestriol (17-epiE₃) (ChemeGen, Shanghai, China), 2-methoxyestrone (2-MeO-E₁) (Macklin, Shanghai, China), 4-methoxyestrone (4-MeO-E₁) (Cato, Guangdong, Guangzhou province, China), 2-hydroxyestrone-3-methyl ether (2-OH-3-MeO-E₁) (TRC, Toronto, ON, Canada), 2-methoxyestradiol (2-MeO-E₂) (Macklin, Shanghai, China), 4-methoxyestradiol (4-MeO-E₂) (Macklin, Shanghai, China), 2-hydroxyestrone (2-OH-E₁) (MCE, Monmouth Junction, NJ, USA), 4-hydroxyestrone (4-OH-E₁) (ChemeGen, Shanghai, China), 2-hydroxyestradiol (2-OH-E₂) (MCE, Monmouth Junction, NJ, US), 4-hydroxyestradiol (4-OH-E₂) (MCE, Monmouth Junction, NJ, US).

Standard chemicals for nine progesterone and its metabolites used in this study include progesterone (MCE, Monmouth Junction, NJ, US), 17α-hydroxy progesterone (MCE, Monmouth Junction, NJ, US), 17α-hydroxy pregnenolone (MCE, Monmouth Junction, NJ, US), pregnenolone (MCE, Monmouth Junction, NJ, US), 5α-dihydroprogesterone (MCE, Monmouth Junction, NJ, US), 5β-dihydroprogesterone (MCE, Monmouth Junction, NJ, US), epipregnanolone (GLPBIO, Montclair, CA, US), pregnanolone (Macklin, Shanghai, China), 20α-hydroxy progesterone (Cato, Guangdong, Guangzhou province, China).

Internal standard chemicals tanshinone IIA (MCE, Monmouth Junction, NJ, US), E₂-d₃ (GLPBIO, Montclair, CA, US) and progesterone-d₉ (Macklin, Shanghai, China) were used for the detection of estrogen and its metabolites and progesterone and its metabolites, respectively. The purity of all chemicals are higher than 97%. Methanol, acetic acid, sodium bicarbonate, l-ascorbic acid, sodium acetate, dansyl chloride, acetone and formic acid were obtained from Macklin (Shanghai, China). β-Glucuronidase/sulfatase from Helix pomatia (Type H-2) was obtained from Sigma-Aldrich (St Louis, MO, US).

Study population and sample collection

The study followed the principles in the Declaration of Helsinki for medical research involving human subjects. This study is approved by the Research Ethics Committee of Longgang District Maternity & Child Healthcare Hospital of Shenzhen City (Approval No. KYXMLL-01-CZGC-14-1-0). All procedures were carried out in accordance with the approved guidelines. Written informed consent was obtained from all participants.

Urine samples were obtained from pregnant women receiving perinatal care at the Longgang District Maternity & Child Healthcare Hospital of Shenzhen City in China. A total of 68 pregnant women were enrolled in this study between April 2023 and May 2024. Participants were excluded from the study if they had heart disease or nephropathy, received estrogen-related or progesterone-related endocrine therapy, had multiple pregnancies, were diagnosed with pregnancy complications including preeclampsia, gestational diabetes, gestational hypertension, etc., experienced spontaneous or induced abortion, or delivered stillborn infants. Detailed characteristics of all participants are presented in Table 1. During the study period, fifty participants were followed through to delivery and included in this study. The remaining participants were excluded due to missing clinical data, or other reasons.

Table 1.

Demographic characteristics of pregnant women at five different gestational weeks.

	8-12 weeks N = 10	16-20 weeks N = 10	30-32 weeks N = 10	35-37 weeks N = 10	38-40 weeks N = 10
Maternal age at birth, years	33.3±2.7	33.8±2.5	32.8±3.9	31.7±2.6	31.2±2.2
Previous births, No. (%)
0	7 (70)	9 (90)	9 (90)	10 (100)	8 (80)
1	2 (20)	1 (10)	1 (10)	0 (0)	2 (20)
≥ 2	1 (10)	0 (0)	0 (0)	0 (0)	0 (0)
Pre-pregnancy BMI, kg/m	22.2±1.7	22.0±3.2	22.7±3.4	19.8±1.8	21.3±1.5
Smoking during pregnancy, No. (%)
Yes	0 (0)	1 (10)	0 (0)	1 (10)	0 (0)
No	10 (100)	9 (90)	10 (100)	9 (90)	10 (100)
Alcohol during pregnancy, No. (%)
Yes	0 (0)	1 (10)	0 (0)	1 (10)	0 (0)
No	10 (100)	9 (90)	10 (100)	9 (90)	10 (100)
Gestational age, days	272.5±7.5	270.6±7.4	275.5±5.3	275.5±7.9	271.6±9.7
Mode of delivery, No. (%)
Spontaneous vaginal birth	7 (70)	7 (70)	8 (80)	7 (70)	7 (70)
Induced vaginal birth	0 (0)	0 (0)	1 (10)	1 (10)	1 (10)
C-section during labor	3 (30)	3 (30)	1 (10)	2 (20)	2 (20)

Data are presented as means ± SD.

Urine samples from these 50 pregnant women were collected at five different stages of pregnancy with 10 pregnant women in each stage: 8 + 0 to 12 + 6 weeks, 16 + 0 to 20 + 6 weeks, 30 + 0 to 32 + 6 weeks, 35 + 0 to 37 + 6 weeks, and 38 + 0 to 40 + 6 weeks. All samples were immediately frozen at -80 °C upon collection until analysis. Pregnancy is defined as at term from 37 + 0 to 42 + 0 weeks of gestation. Gestational age was determined by calculating the time interval from the first day of the last menstrual period to the day of delivery.

Urine sample processing

All urine samples were first thawed at 4 °C and centrifuged at 6,000 × g at 4 °C for 5 min. Then, 1 mL supernatant was collected, and 1 mL freshly prepared enzymatic hydrolysis buffer was added, which contains 10 µL β-glucuronidase/sulfatase (85,000 units/mL), 2 mg L-ascorbic acid, and 0.15 M sodium acetate buffer (pH 4.6) and 10 µL tanshinone IIA, E₂-d₃ and progesterone-d₉ in methanol (100 ng/mL). These hydrolysis reactions were then incubated for 20 h at 37 °C.

Tanshinone IIA is a natural compound that cannot be produced endogenously⁵². As a compound that has been extensively studied, the stability of tanshinone IIA has been well-documented in the literature⁵³. Tanshinone IIA, although not a deuterated compound, was used as an internal standard for detecting estrogen metabolites due to its structural similarity to estrogens, lack of interference under the analytical conditions, and excellent stability during sample preparation and analysis in liquid chromatography-tandem mass spectrometry^54,55.

Then, solid-phase extraction (SPE) with a SPE cartridge (Agilent, Bond Elut Plexa) was used to enrich and purify estrogen, progesterone and their metabolites from the hydrolyzed urine samples. SPE cartridges were activated with 3 mL methanol containing 0.5% formic acid. Then, cartridges were equilibrated with 3 mL water. After the samples were loaded, the cartridges were washed with 1 mL 30% methanol, then dried under a minimum pressure of 10 inches of mercury for 5 min. Finally, elution was performed four times with 0.25 mL methanol each time. Total 1 mL eluate was divided into two 0.5 mL aliquots: one aliquote was dried under nitrogen at 37^oC and derivatized and then used to analyze estrogen and its metabolites in one injection of UPLC/MS/MS; and the other aliquote was directly used to analyze progesterone and its metabolites in the second injection for UPLC/MS/MS. In total, two injections were carried out for each urine sample. For derivatization of estrogen and its metabolites, the dried residues were re-dissolved in 150 µL sodium bicarbonate buffer (0.1 M, pH 9) and 150 µL of dansyl chloride in acetone (1 mg/mL) for derivatization⁵⁶. After vortexing, the samples were heated at 60 °C for 6 min for the formation of dansyl derivatives. Before mass spectrometric analysis, samples were filtered through a 0.22 μm membrane.

UPLC–MS/MS analysis

The samples were analyzed by UPLC/MS/MS (AB SCIEX QTRAP Enabled Triple Quad 5500+, Framingham, MA, US). The chromatographic column was obtained from Agilent (Agilent ZORBAX RRHD Eclipse Plus 95Å C18, CA, US). The MS conditions were as follows: source type, turbo spray; source temperature (at setpoint), 550.0 ℃; collision-activated dissociation, 8.000 psi; curtain gas, 35.000 psi; nebulizer gas, 55.000 psi; heating gas, 55.000 psi; ion spray voltage, 4500.000 V; source temperature, 550.000 ℃; cell exit potential, 22.000 psi; entrance potential, 10.000 psi. The Multiple reaction monitoring (MRM) conditions for the protonated molecules (MH+) of estrogen-dansyl are shown in Table 2. The MRM conditions for the protonated molecules (MH+) of progesterones are shown in Table 3.

Table 2.

MRM condition for estrogen and its metabolites.

Analyte	Precursor Ion (m/z)	Product Ion (m/z)	Collision Energy (eV)	Declustering Potential (V)	Retention time (min)
E1	504.2	171.1	45	70	10.40
E2	506.2	171.1	48	70	10.70
E3	522.2	171.1	45	75	7.82
16-epiE3	522.2	171.1	45	75	8.78
17-epiE3	522.2	171.1	45	75	9.01
2-OH-E2	755.3	170.1	50	85	11.78
4-OH-E2	755.3	170.1	50	85	11.93
2-MeO-E1	534.2	171.1	45	75	10.59
2-OH-3-MeO-E1	534.2	171.1	45	75	9.98
4-MeO-E1	534.2	171.1	45	75	10.99
2-MeO-E2	536.2	171.1	45	75	9.89
4-MeO-E2	536.2	171.1	45	75	10.08
2-OH-E1	753.3	170.1	50	85	12.98
4-OH-E1	753.3	170.1	50	85	13.40

Table 3.

MRM condition for progesterone and its metabolites.

Analyte	Precursor Ion (m/z)	Product Ion (m/z)	Collision Energy (eV)	Declustering Potential (V)	Retention time (min)
17α-hydroxy progesterone	331.4	97.2	29	124	5.11
17α-hydroxy pregnenolone	333.4	239.1	38	106	5.07
Progesterone	315.4	81.2	60	80	5.44
Pregnenolone	317.4	102.0	28	101	4.86
5α-dihydroprogesterone	317.4	102.0	29	100	4.46
5β-dihydroprogesterone	317.4	102.0	26	110	4.56
Epipregnanolone	319.5	257.3	32	146	4.76
Pregnanolone	319.5	257.4	33	149	4.59
20α-hydroxy progesterone	317.4	109.0	37	134	5.38

The UPLC mobile phase is made up of the linear gradient mixture of Mobile phase A: Water with 0.1% formic acid and 0.2 M ammonium acetate; and Mobile phase B: Methanol. The linear gradient of mobile phases was listed in Table 4 for estrogen and its metabolites and Table 5 for progesterone and its metabolites. The injection volume was 5 µL. The flow rate was 0.2 mL/min.

Table 4.

Time-dependent gradient of the mobile phase for estrogen and its metabolites.

Time	A (v%)	B (v%)
0	70	30
1	70	30
4	20	80
13	10	90
17	10	90
18	70	30
19	70	30

Table 5.

Time-dependent gradient of the mobile phase for progesterone and its metabolites.

Time	A (v%)	B (v%)
0	80	20
1	80	20
4	10	90
7	10	90
7.5	80	20
8.5	80	20

Accuracy and precision

To validate the above methods for detecting estrogen, progesterone and their metabolites in urine, we evaluated the recovery and accuracy of the method using known amounts of estrogen, progesterone and their metabolites in charcoal stripped human urine. In addition, we evaluated the precision of our method using six replicated charcoal stripped human urine samples supplemented with known amounts of estrogen, progesterone and their metabolites across six different batches. Known amounts of steroid hormone standards, including estrogens, progesterone, and their metabolites, were spiked into these charcoal treated urine samples at the following concentrations 0.1, 2, and 20 ng/mL for estrogen and its metabolites; and 0.05, 1, and 10 ng/mL for progesterone and its metabolites. The range of the testing concentrations was chosen to cover the concentration range of clinical samples to ensure that our method is robust when used in clinical scenarios. Accuracy was measured as the percentage match between the calculated amount and the expected amount in the charcoal treated urine samples. Intra- and inter-batch precision was determined by calculating the relative standard deviation. The calibration curves and sample quantification were performed using SCIEX OS software, version 2.2.0 (Framingham, MA, USA).

Statistical analysis

The estrogen, progesterone and their metabolites levels in urine was normalized to the creatinine level in urine. Data are presented as means ± standard deviation (SD). Data were analyzed by one-way ANOVA followed by Dunns post-test for multiple comparisons. GraphPad-Prism 9.0 (San Diego, CA, USA) was used to analyze data. Values of P < 0.05 were considered statistically significant.

Results

Method validation

The lower limit of quantification (LLOQ) for estrogen and its metabolites was 12 pg/mL (signal-to-noise ratio greater than 10), and for progesterone and its metabolites, the LLOQ was 20 pg/mL (signal-to-noise ratio greater than 10). LLOQ values and calibration curves were determined using a pooled approach, and the concentrations of the calibrators were as follows: 0.02, 0.1, 0.5, 1, 5, 50, 200, and 500 ng/mL for estrogen and its metabolites; and 0.06, 0.12, 0.3, 1, 3, 12, 60, and 120 ng/mL for progesterone and its metabolites. Urine with endogenous steroid hormones removed by charcoal stripping was used as the matrix to prepare calibration curves. A weighted linear regression was used to construct the calibration curves. The linear range for estrogen and its metabolites was 0.02 ng/mL to 500 ng/mL (r > 0.99) and for progesterone and its metabolites was 0.06 ng/mL to 120 ng/mL (r > 0.99). As shown in Table 6, the recovery (accuracy) of added estrogen and its metabolites at 0.1, 2, and 20 ng/mL was 93 − 103%, 98 − 104%, and 97 − 101%, respectively (Table 6). The recovery (accuracy) of a known added amount of progesterone and its metabolites at 0.05, 1, and 10 ng/mL was 97 − 107%, 96 − 106%, and 94 − 104%, respectively (Table 7). For the measurement of trace steroid hormones, accuracy within 70–125% and precision less than 15% are generally considered acceptable based on the U. S. Food and Drug Administration guidelines. These results indicate that the urine sample processing method has good recovery rate.

Table 6.

Recovery (accuracy), intra- and inter-batch precision of estrogen and its metabolites.

	0.1 ng/mL			2 ng/mL			20 ng/mL
	accuracy (%)	intra-batch (%)	inter-batch (%)	accuracy (%)	intra-batch (%)	inter-batch (%)	accuracy (%)	intra-batch (%)	inter-batch (%)
E1	93.2	7.3	7.5	104.5	9.5	11.8	100.5	5.2	2.8
E2	102.7	5.4	1.4	102.0	2.5	2.0	100.9	1.8	0.9
E3	95.3	1.8	2.2	101.1	1.3	1.3	98.8	2.1	1.0
16-epiE3	100.4	6.2	9.7	103.2	2.2	2.0	101.7	4.2	5.6
17-epiE3	102.4	1.4	2.5	101.0	1.6	2.1	98.3	2.5	3.1
2-OH-E2	94.0	3.6	4.1	99.6	1.8	0.6	97.6	2.7	0.6
4-OH-E2	96.5	3.6	6.2	101.3	1.0	1.4	100.5	1.7	2.0
2-MeO-E1	101.4	5.7	6.7	100.0	3.1	2.9	100.6	2.5	3.8
2-OH-3-MeO-E1	99.8	5.0	3.1	99.5	3.0	1.5	101.9	2.5	3.9
4-MeO-E1	99.3	5.2	6.6	101.6	3.7	1.3	99.8	2.4	3.1
2-MeO-E2	100.2	3.6	7.0	101.8	1.2	1.3	101.4	1.1	0.9
4-MeO-E2	99.9	3.7	7.4	101.4	1.0	1.8	98.4	1.2	1.8
2-OH-E1	95.7	2.7	3.5	98.9	1.5	0.6	97.5	1.1	2.0
4-OH-E1	103.2	1.4	1.2	101.5	0.8	1.6	98.4	1.4	3.0

Table 7.

Recovery (accuracy), intra- and inter-batch precision of progesterone and its metabolites.

	0.05 ng/mL			1 ng/mL			10 ng/mL
	accuracy (%)	intra-batch (%)	inter-batch (%)	accuracy (%)	intra-batch (%)	inter-batch (%)	accuracy (%)	intra-batch (%)	inter-batch (%)
17α-hydroxy progesterone	97.1	9.3	9.1	105.1	4.4	4.1	96.5	6.6	11.3
17α-hydroxy pregnenolone	107.9	5.9	3.2	105.6	3.1	3.4	98.5	4.7	7.9
Progesterone	102.2	2.4	4.1	100.0	6.5	8.7	104.6	8.0	12.1
Pregnenolone	101.0	7.2	5.1	95.9	5.4	10.2	98.8	3.7	5.5
5α-dihydroprogesterone	108.7	7.4	3.9	99.1	2.9	2.4	99.3	3.2	6.3
5β-dihydroprogesterone	100.6	5.1	7.7	96.0	3.2	3.9	98.9	4.0	4.1
Epipregnanolone	101.7	3.0	5.5	101.2	2.6	0.9	99.3	5.5	5.8
Pregnanolone	100.2	2.9	1.9	98.4	3.8	7.0	97.7	5.1	5.7
20α-hydroxy progesterone	102.6	9.3	7.1	106.6	5.0	10.6	94.9	2.1	3.3

The intra-batch precision, determined by the relative standard deviation (RSD) of six replicate analyses of added estrogen and its metabolites at 0.1, 2, and 20 ng/mL, was 1 − 7%, 1 − 9%, and 1 − 5%, respectively (Table 6). The intra-batch precision determined by the RSD of six replicate analyses of added progesterone and its metabolites at 0.05, 1, and 10 ng/mL was 2 − 9%, 2 − 6%, and 2 − 8%, respectively (Table 7). The inter-batch precision determined by the RSD of six independent batch analyses of added estrogen and its metabolites at 0.1, 2, and 20 ng/mL, was 1 − 9%, 1 − 11%, and 1 − 5%, respectively (Table 6). The inter-batch precision determined by the RSD of six independent batch analyses of added progesterone and its metabolites at 0.1, 2, and 20 ng/mL, was 3 − 9%, 1 − 10%, and 3 − 12%, respectively (Table 7). In sum, these results indicate that the sample processing and mass spectrometry detection methods are suitable for the quantification of estrogen, progesterone and their metabolites.

Changes of estrogen and its metabolites during pregnancy

Using the optimized UPLC/MS/MS methods, we found that 14 types of derivatized estrogen and its metabolites were well separated within 17 min (Fig. 3), and 9 types of progesterone and its metabolites were well separated within 8 min (Fig. 4). For metabolites with the same m/z values, their peaks were well separated with different retention times (Figs. 3 and 4).

Fig. 3.

MRM profiles of derivatized estrogens and their metabolites by UPLC-MS/MS in the estrogen-free urine spiked with 50 ng/mL estrogen metabolites standards. The numbers labeled in the right side are mass-to-charge ratios of the precursor and product ions.

Fig. 4.

MRM profiles of progesterone and its metabolites by UPLC-MS/MS in the progesterone-free urine spiked with 12 ng/mL progesterone metabolites standards. The numbers labeled in the right side are mass-to-charge ratios of the precursor and product ions.

We found that the concentration of estrogen and its metabolites in urine underwent significant changes during pregnancy (Fig. 5). Among the 14 analyzed estrogen and its metabolites, 7 had a time-dependent increase during pregnancy, which include E₁, E₂, E₃, 16-epiE₃, 17-epiE₃, 2-MeO-E₂ and 4-OH-E₁. In the late pregnancy (38–40 weeks), the concentrations of E₁, E₂, E₃, 16-epiE₃, 17-epiE₃ increased by more than tenfold, 2-MeO-E₂ increased by eight fold, and 4-OH-E₁ increased by two fold, compared to those in early pregnancy (8–12 weeks) (Figs. 5 and 6). E₁, E₂, E₃, 16-epiE₃, 17-epiE₃ levels in pregnancy were much higher than other estrogen metabolites, and the levels of 4-MeO-E₁ and 4-OH-E₁ were the lowest (Fig. 6).

Fig. 5.

The levels of 14 estrogen metabolites during pregnancy. Data are shown in mean ± SD with n = 10. *p < 0.05, **p < 0.01, compared to the 8-12 group; ^##p <0.01, compared to the 30-32 weeks group.

Fig. 6.

Urinary estrogen metabolites levels in 8-12 weeks and 38-40 weeks pregnancy. Data are shown in mean ± SD with n = 10. *p < 0.05, **p < 0.01, compared to the 8-12 group; ^##p <0.01, compared to the 30-32 weeks group.

The concentrations of 2-OH-E₁, 2-OH-E₂, 4-OH-E₂ were relatively higher in early pregnancy (8–12 weeks) and decreased continuously by more than 75% in the late pregnancy (Fig. 5). The concentration of 4-MeO-E₂ and 2-MeO-E₁ peaked in mid-pregnancy (30–32 weeks) and then gradually decreased. The concentration of 4-MeO-E₁ and 2-OH-3-MeO-E₁ increased at 16–20 weeks and then remained almost constant (Fig. 5). Throughout pregnancy, the ratio of 2-MeO-E₁/2-OH-E₁ and 2-MeO-E₂/2-OH-E₂ showed a continuous increase (Fig. 7). In contrast, the ratio of 4-MeO-E₁/4-OH-E₁ and 4-MeO-E₂/4-OH-E₂ initially increased but later declined during pregnancy (Fig. 7).

Fig. 7.

Ratios of methoxyestrogens and corresponding hydroxyestrogens during pregnancy. Data are shown in mean ± SD with n = 10. **p < 0.01, compared to the 8-12 weeks group; ^##p <0.01, compared to the 30-32 weeks group; ⁺⁺p <0.01, compared to the 16-20 weeks group.

Changes of progesterone metabolites during pregnancy

We found that the levels of 17α-hydroxy progesterone, 17α-hydroxy pregnenolone, pregnanolone, pregnenolone and epipregnanolone increased time-dependently throughout pregnancy (Fig. 8). Intriguingly, progesterone sharply increases in the mid pregnancy (30–32 weeks) and gradually decreases in the late pregnancy (Fig. 8). 20α-hydroxy progesterone, 5α-dihydroprogesterone, and 5β-dihydroprogesterone increase until 35–37 weeks, then decrease (Fig. 8). In the late pregnancy, the levels of 17α-hydroxy pregnenolone and 17α-hydroxy progesterone are the highest and lowest among all progesterone metabolites, respectively (Fig. 9).

Fig. 8.

The levels of 9 progesterone metabolites during pregnancy. Data are shown in mean ± SD with n = 10. *p < 0.05, **p < 0.01, compared to the 8-12 weeks group.

Fig. 9.

Urinary progesterone metabolites levels in 8-12 weeks and 38-40 weeks pregnancy. Data are shown in mean ± SD with n = 10. *p < 0.05, **p < 0.01, compared to the 8-12 weeks group.

Discussion

In this study, we revealed profound changes in maternal estrogen, progesterone and their metabolites pregnancy by using UPLC/MS/MS analysis of urine samples. We observed that the ratios of 2-MeO-E₁/2-OH-E₁ and 2-MeO-E₂/2-OH-E₂ consistently increased throughout pregnancy (Fig. 7), which indicates the activity of catechol-O-methyltransferase (COMT) increases during pregnancy. COMT is responsible for the enzymatic O-methylation of catechol estrogens (2-OH-E₁, 2-OH-E₂, 4-OH-E₁, 4-OH-E₂), leading to the formation of corresponding methoxyestrogens (2-MeO-E₁, 2-MeO-E₂, 4-MeO-E₁, 4-MeO-E₂)⁵⁷ (Fig. 1).

A study has shown that COMT enzyme activity in the liver of rats undergoes significant changes during pregnancy⁵⁸. However, research on the expression and activity of COMT during pregnancy in humans remains limited. Human COMT exhibits significantly higher catalytic activity toward 2-OH-E₂ than 4-OH-E₂⁵⁹. Indeed, we found that the 2-MeO-E₁ and 2-MeO-E₂ levels were over 10 times higher than those of 4-MeO-E₁ and 4-MeO-E₂, respectively (Fig. 5). Preeclampsia is associated with low COMT activity in the human placenta⁶⁰. A study shows that plasma levels of 2-MeO-E₁ and 2-MeO-E₂ are significantly reduced in preeclampsia, while levels of 2-OH-3-MeO-E₁ and 4-MeO-E₁ remain unchanged⁶¹. Consistently, we found that 2-MeO-E₁ and 2-MeO-E₂ levels increased in the mid and late pregnancy (Figs. 5 and 6), indicating that they play important roles in the pregnancy maintenance and placenta development.

The study on the role of 4-MeO-E₁ and 4-MeO-E₂ in pregnancy is limited. One study finds that 4-MeO-E₂ stimulates the proliferation of pregnant, but not non-pregnant, ovine uterine artery endothelial cells⁶². However, we found that there was a drastic decrease of 4-MeO-E₂ and 4-MeO-E₁ from the mid pregnancy to late pregnancy, which indicates that these two hormones may have unique effects in the late pregnancy, differently from 2-MeO-E₂ and 2-MeO-E₁. The causes of this decrease may be related to other metabolism enzymes besides COMT, which merits further investigation.

It is well known that E₁, E₂, E₃ are important hormones in pregnancy⁶³. Consistently, we find that E₃ becomes the dominant estrogen later in pregnancy with its level five times higher than E₁ and 50 times higher than E₂ (Figs. 5 and 6). Although it has been suggested that E₃ serves to protect the fetus during pregnancy, its exact function remains to be revealed². 16-epiE₃ and 17-epiE₃ are stereoisomers of E_3, whose levels increased dramatically in the late pregnancy (Figs. 5 and 6). Previous studies have shown that 16-epiE₃ has anti-inflammatory activity⁶⁴, and 17-epiE₃ inhibits TNFα-induced vascular cell adhesion molecule 1 (VCAM-1) expression⁶⁵. However, their roles during pregnancy remain unclear. A clinical study shows that the serum levels of the cell adhesion molecule VCAM-1 are significantly higher in preeclampsia patients⁶⁶. Given the drastic increase of 17-epiE₃ during pregnancy, we hypothesize that 17-epiE₃ may play an important role in pregnancy by inhibiting VCAM-1, which merits further investigation.

The hydroxyestrogens 2-OH-E₂ and 2-OH-E₁ are biologically active estrogens with lower estrogenic activity than E₂ and E₁, and are hypothesized to act as anti-estrogens through competitive inhibition⁶⁷. 4-OH-E₁ is the only one of the four hydroxyestrogens whose levels continually increase throughout pregnancy (Fig. 5), possibly due to reduced conversion to 4-MeO-E₁. The decrease of 2-OH-E₁ and 2-OH-E₂ may be due to the increased conversion to 2-MeO-E₁ and 2-MeO-E₂ by COMT. Research has shown that 4-OH-E₁ is an endogenous neuroestrogen that can strongly protect neurons against oxidative damage by inhibiting oxidative stress-induced ferroptosis through non-estrogen receptor pathways^68,69. Therefore, we hypothesize that 4-OH-E₁ may contribute to maintaining maternal neurological health or promoting fetal neurodevelopment during pregnancy. It has been reported that during the redox cycling of 4-OH-E₂, excessive reactive oxygen species are generated⁷⁰. Since oxidative stress is closely linked to pregnancy complications, such as gestational diabetes and preeclampsia^71–73, this may explain the very low level of 4-OH-E₂ in the late pregnancy (Fig. 6).

In most animal models, progesterone concentrations decline before the onset of labor, which is consistent with our results^74,75. 17α-hydroxy pregnenolone is produced from pregnenolone through 17-hydroxylase (Fig. 2). Our finding that 17α-hydroxy pregnenolone has the highest level among all progesterone metabolites indicates that 17-hydroxylase may play important roles in pregnancy. In humans, adrenal 17-hydroxylase is essential for the biosynthesis of cortisol, a key glucocorticoid necessary for regulating glucose metabolism, immune responses, and stress adaptation⁷⁶. In addition, a deficiency in 17-hydroxylase can lead to serious clinical conditions, including hypertension, hypokalemia, hypogonadism⁷⁷, and infertility⁷⁸, underscoring its importance in maintaining both endocrine and reproductive health. However, its role in pregnancy maintenance is not clear at present. It has been shown that levels of 17α-hydroxy pregnenolone in the umbilical artery of preterm infants are significantly lower than those in full-term infants⁷⁹. Thus, the very high level of 17α-hydroxy pregnenolone in the late pregnancy indicates that it may play important roles in the maintenance of pregnancy, which merits further investigation.

A study found that the levels of progesterone metabolites are often higher in patients with depression⁸⁰, suggesting changes in progesterone metabolites may be related to mood changes during pregnancy and the postpartum period. Pregnenolone is the main steroid synthesized from cholesterol in mammals and acts as an endogenous modulator of type-1 cannabinoid receptor⁸¹. The evidences from both animal studies and human clinical research support the neuroprotective role of pregnenolone⁸². Pregnanolone is an endogenous inhibitory neurosteroid produced from progesterone⁸³. Pregnanolone has potent sedative/hypnotic and anesthetic effects^77,84. The continuous increases of pregnenolone and pregnanolone during pregnancy (Fig. 8) may indicate its important role in promoting mental health of the mother.

5α-dihydroprogesterone and 5β-dihydroprogesterone are agonists of the progesterone receptor and positive allosteric modulators of the GABA_A receptor⁸⁵. Additionally, 5β-dihydroprogesterone has been found to act as a negative allosteric modulator of the GABA_A receptor⁸⁶ and has also been shown to regulate uterine contractility by activating the pregnane X receptor⁸⁷. Epipregnanolone is a negative allosteric modulator of the GABA_A receptor⁸⁸. We found that the negative regulators of GABA_A receptor, i.e., 5β-dihydroprogesterone and epipregnanolone had higher levels in the late pregnancy compared to the positive regulator of GABA_A receptor 5α-dihydroprogesterone (Fig. 8), indicating that the negative regulation of GABA_A might be important for pregnancy maintenance.

20α-hydroxy progesterone is a naturally occurring, endogenous progesterone⁸⁹. It has a very low affinity for the progesterone receptor, and is much less potent in comparison to progesterone⁹⁰. It has also been found to act as an aromatase inhibitor⁹¹, and its role in pregnancy is not yet known. The level of 20α-hydroxy progesterone is much higher than 5α-dihydroprogesterone and 5β-dihydroprogesterone at the late pregnancy (Figs. 8 and 9), indicating that 20α-HSD may play an important role in metabolizing progesterone in late pregnancy.

Conclusion

In this study, we used urine samples in the pregnant women for LC/MS/MS detection of estrogen, progesterone and their metabolites, which is less invasive for sample collection. Profound changes in maternal steroidogenesis during pregnancy were found in this study. The main contribution of this study is that we quantified up to 23 steroid hormone metabolites, including estrogen, progesterone, and their metabolites, across the entire gestational period. This dynamic hormone profile provides valuable insights into hormonal changes during pregnancy and lays the foundation for future studies. Monitoring the changes of the levels of these hormones is important for ensuring maternal and fetal health during pregnancy. In addition, we have revealed varied physiological changes in the levels of a number of estrogen and progesterone metabolites during pregnancy, which lays the foundation for further functional mechanistic study of the roles of these metabolites in pregnancy.

However, the study’s sample size and the fact that samples were not collected from each subject for the entire pregnancy period limited our ability to perform stratified analyses on the influence of factors such as the number and sex of fetus, and lifestyle of the pregnant women (for example, smoking and alcohol consumption) on estrogen and progesterone metabolites levels, which merits future investigation.

Abbreviations

UPLC-MS/MS

ultrahigh performance liquid chromatography-tandem mass spectrometry

E₁

estrone

E₂

estradiol

E₃

estriol

16-epiE₃

16-epiestriol

17-epiE₃

17-epiestriol

2-MeO-E₁

2-methoxyestrone

4-MeO-E₁

4-methoxyestrone

2-OH-3-MeO-E₁

2-hydroxyestrone-3-methyl ether

2-MeO-E₂

2-methoxyestradiol

4-MeO-E₂

4-methoxyestradiol

2-OH-E₁

2-hydroxyestrone

4-OH-E₁

4-hydroxyestrone

2-OH-E₂

2-hydroxyestradiol

4-OH-E₂

4-hydroxyestradiol

SPE

Solid-phase extraction

MRM

Multiple reaction monitoring

LLOQ

Lower limit of quantification

RSD

Relative standard deviation

COMT

Catechol-O-methyltransferase

VCAM-1

Vascular cell adhesion molecule 1

Author contributions

P.W. and J.Y. conceived the project. C.J. contributed to the experimental design, conducted experiments, analyzed data, and edited the manuscript. Y.P., X.L., and Q.Z. coordinated and collected the clinical samples. L.L. helped conduct experiments. P. W. acquired the fund, contributed to experimental design, supervised the experiments and analysis, provided scientific oversight and resources. C. J. and P. W. wrote and edited the manuscript.

Funding

This study is supported by Ganghong Young Scholars Development Fund (No.2021E0009, 2022E0032), Shenzhen Basic Science Project (No. JCYJ20190813170607348), Shenzhen Key Laboratory of Steroid Drug Discovery and Development (No. ZDSYS20190902093417963), Shenzhen-Hong Kong Cooperation Zone for Technology and Innovation (No. HZQB-KCZYB-2020056) and Shenzhen Medical Academy of Research and Translation (No. A2401015).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study followed the principles in the Declaration of Helsinki for medical research involving human subjects. This study is approved by the Research Ethics Committee of Longgang District Maternity & Child Healthcare Hospital of Shenzhen City (Approval No. KYXMLL-01-CZGC-14-1-0). All procedures were carried out in accordance with the approved guidelines. Informed consent was obtained from all participants.

Competing interests

The authors declare no competing interests.

Author information

Shenzhen Key Laboratory of Steroid Drug Discovery and Development, School of Medicine, The Chinese University of Hong Kong, Shenzhen, Guangdong 518172, China.

Jin Chen, Langqi Lin & Pan Wang.

Longgang Maternity and Child Institute, Shantou University Medical College (Longgang District Maternity & Child Healthcare Hospital of Shenzhen City), Shenzhen, Guangdong, 518172, China.

Ying Peng, Xiaoyan Luo, Qi Zhu & Jinying Yang.