Determination of 17OHPreg and DHEAS by LC-MS/MS: Impact of Age, Sex, Pubertal Stage, and BMI on the Δ5 Steroid Pathway

Abstract

Background:

Dehydroepiandrosterone sulfate (DHEAS) and 17-hydroxypregnenolone (17OHPreg) are important for understanding the Δ5 pathway (e.g., in adrenarche and obesity). Although mass spectrometry has become the state-of-the-art method for quantifying steroids, there are few comprehensive age-, sex-, and pubertal stage–specific reference ranges for children.

Aims:

To develop a sensitive and reliable ultra-performance liquid chromatography tandem mass spectrometry (UPLC-MS/MS) method for simultaneous quantification of DHEAS and 17OHPreg and to establish entire age-, sex- and pubertal stage–specific reference ranges in children.

Methods:

A total of 684 children, 453 (243 female, 210 male) with normal body mass index (BMI; <90th) and 231 (132 female, 99 male) obese subjects (>97th), were categorized into 11 age groups, and age- and Tanner stage (PH)–specific reference ranges were determined.

Results:

The limit of detection was 0.05 nmol/L for 17OHPreg and 0.5 nmol/L for DHEAS. Levels of both steroids declined after the neonatal period. Comparisons with RIA assays (Siemens, Munich, Germany) (DHEAS) and an in-house kit (17OHPreg) revealed 0.95 and 0.93, respectively, as coefficients of determination. Although DHEAS—generally higher in boys—increased continuously starting at 3 to 6 years, 17OHPreg remained largely constant. In obese patients, both were significantly elevated, also in part after alignment to Tanner stages (PH).

Conclusions:

UPLC-MS/MS is sensitive and reliable for quantifying DHEAS and 17OHPreg. Our data support differential maturation of CYP17 during adrenarche with successively increasing 17,20-lyase activity but largely constant 17α−hydroxylation activity. Endocrine interpretation of 17OHPreg and DHEAS must consider differential patterns for age, sex, pubertal stage, and BMI.

Dehydroepiandrosterone sulfate (DHEAS), which is produced in the adrenal cortex, is the steroid hormone with the highest serum concentrations in humans. Starting with high production in the fetal adrenal, DHEAS declines in the first months of life and remains low until adrenarche (1). Adrenarche is an ongoing process of functional and morphological maturation of the adrenal cortex (2, 3) observed only in humans and higher primates (4, 5). During this process, expression of 3β-hydroxysteroid dehydrogenase type 2 (HSD3B2) is suppressed, whereas the cofactor cytochrome b5 (enhancer of 17,20-lyase activity of CYP17) and steroid sulfotransferase (SULT2A1) are expressed at higher levels (6, 7). These mechanisms contribute to increased androgen production (8). In humans, adrenal steroidogenesis primarily follows the Δ5 pathway. After cholesterol cleavage, pregnenolone is hydroxylated to 17-hydroxypregnenolone (17OHPreg) via 17α−hydroxylation activity of CYP17 (P450c17), followed by conversion to dehydroepiandrosterone (DHEA) through 17,20-lyase activity of CYP17. A substantial portion of DHEA is sulfated to DHEAS through SULT2A1 (9). In most children, circulating DHEAS is sufficient to stimulate pubic and axillary hair growth (7). All enzymes required for the formation of active androgens from DHEAS are expressed in hair follicles as well as in dermal endocrine glands (10, 11).

In humans, sulfated steroids such as DHEAS can be converted directly into other sulfated steroids without cleaving the sulfate group (8). Recent reports (12) have revealed that sulfation is the critical step in regulating the conversion of DHEA to active androgens. Hence, DHEAS is regarded as an important biochemical marker of adrenarche, starting at serum concentrations of ∼400 ng/mL (8, 13). Elevated concentrations of DHEAS can be an indication of premature adrenarche (14, 15) underlying premature pubarche (8) and are associated with obesity (16–18). In addition, DHEAS has been associated with important brain functions, including neuronal survival, memory, and behavior, showing therapeutic potential in various neuropsychiatric and cognitive disorders (19, 20).

Although 17OHPreg is one of the most important serum biomarkers for diagnosing 3β-hydroxysteroid dehydrogenase deficiency (3βHSDIID), a rare monogenic form of congenital adrenal hyperplasia (21, 22), its biochemical role in adrenarche and obesity and the interrelationship with DHEAS are less well understood. To date, very few reference ranges for 17OHPreg have been documented, for example, for the first year of life (23, 24) and for ages 4 to 9 years (25). Although liquid chromatography tandem mass spectrometry (LC-MS/MS)–based reference data for DHEAS have previously been published (26), data on 17OHPreg are incomplete and do not cover the entire pediatric age range. This makes functional understanding of the role of 17OHPreg more difficult.

The objectives of our study were to develop a sensitive and reliable method for simultaneous quantification of 17OHPreg and DHEAS by ultra-performance liquid chromatography tandem mass spectrometry (UPLC-MS/MS). Furthermore, we aimed to document age-, sex-, and pubertal stage–specific reference ranges from birth to adulthood for both steroids in children with normal weight control and in children with obesity. For validation purposes, we included 3 patients with molecular proven 3βHSDIID.

Subjects and Methods

Hormone analysis

Plasma concentrations of DHEAS and 17OHPreg were simultaneously determined using LC-MS/MS as previously described (18). All standards, the deuterium-labeled standard DHEAS-d6 and the standards DHEAS and 17OHPreg, were purchased from Sigma-Aldrich (Steinheim, Germany). Oasis WAX μ-elution plates were obtained from Waters (Milford, MA). LC-MS–grade solvents such as methanol, waters, formic acid, and ammonium hydroxide were also obtained from Sigma-Aldrich. A stock solution was prepared in methanol for hormones and deuterium-labeled substances to optimize the LC-MS/MS method. The quantifier and qualifier transitions (m/z) were 331 > 287, 331 > 303 for 17OHPreg and 367 > 96, 367 > 79 for DHEAS. For daily use, dilutions were carried out in methanol to prepare a calibration curve in the range of 0.03 to 66 ng/mL (0.1 to 200 nmol/L) for 17OHPreg; a tenfold higher concentration range was used for DHEAS. Calibration curves and quality controls at concentrations of 0.22 ng/mL (0.75 nmol/L), 1.2 ng/mL (4 nmol/L), and 25 ng/mL (75 nmol/L) (for DHEAS, tenfold higher concentrations) were prepared daily by spiking steroid-free plasma into the mixture solution. Steroid-free plasma was generated by stirring human plasma from healthy blood donors with activated charcoal 3 times and subsequent filtration with Whatman Vacuflo, 0.2 μm (Th. Geyer, Renningen, Germany). Patient samples were prepared by using solid-phase extraction (SPE). Aliquots of serum samples, calibrator, and controls with a volume of 0.1 mL were combined with the internal standard to monitor recovery. All samples were extracted using Oasis WAX SPE system plates (Waters). Two washing steps followed after loading the samples on the SPE plate: first with 200-µL formic acid (2%) and then with methanol (7%). Hormones were eluted using 60 µL of 5% ammoniumhydroxid in methanol followed by 100 µL of water.

Hormone analysis was performed using a Waters Quattro Premier/Xe triple quadrupole mass spectrometer connected to a Waters Acquity. The Waters Acquity UPLC BEH C18 column (1.7 μm, 100 × 2.1) was used at a flow rate of 0.3 mL/min at 50°C. Water (A) and methanol (B) with 0.01% ammoniumhydroxid were used as a mobile phase. Separation was delivered with the following gradient: from 35% to 78% methanol in 2 minutes, a step gradient to 100% methanol, and re-equilibration to initial conditions in 2.5 minutes. Total running time was 3.2 minutes, and the injection volume was 5 µL. Electrospray was used in the positive mode. Two mass transitions were monitored for each hormone. Data were acquired with MassLynx 4.1 software, and quantification was performed by TargetLynx software (Waters). During all analyses, the ambient temperature was kept at 21°C with air conditioning.

Assay performance characteristics

These characteristics are presented in the supplemental data.

Plasma samples of control children and obese children

The study was approved by the ethical committee of the Christian-Albrechts University of Kiel, Germany (file number D531/16). Leftover samples from blood checks before minor surgery (e.g., inguinal hernia repair, circumcision, tonsillectomy) or for exclusion of endocrine and other diseases (e.g., presumed growth disorder or presumed hypothyroidism) were used to set up reference data. The subjects showed no active signs of endocrine or systemic disease and were not on steroid medication. No additional venous punctures were performed for the study. All children or their caregivers gave their informed consent for use of routine clinical data. Before biochemical analysis of 17OHPreg and DHEAS, the data set was completely anonymized.

Blood was collected between 0800 and 1000 am from children and adults between 0 and 40 years of age and in the follicular phase of the menstrual cycle if applicable. In total, 684 subjects (309 males [m] and 375 females [f]) were included. On the basis of body mass index (BMI) score, the subjects were divided into 2 cohorts, 453 with a normal BMI [<90th percentile according to German reference values (27); 192 m and 261 f “control children”] and 231 with a BMI >97th percentile (99 m and 132 f “obese children”). All subjects were further subdivided by sex and were allocated to 8 age groups: AG1 (<1 year), AG2 (1 to 3 years), AG3 (3 to 6 years), AG4 (6 to 8 years [girls]), AG4 (6 to 9 years [boys]), AG5 (9 to 11 years), AG6 (11 to 13 years), AG7 (13 to 15 years), and AG8 (16 years and older), according to a published statistical method (28). AG4 has been defined differently in boys than in girls because of differences in regular puberty onset. The pubertal developmental stage was determined according to Marshall and Tanner (29, 30). In a subgroup, pubic hair according to Tanner stage (PH) was recorded. In total, these data were available for 400 reference subjects (174 m, 226 f) and 222 obese subjects (96 m, 126 f).

Plasma samples of patients with 3βHSDIID

For validation purposes, plasma was analyzed from 3 patients with 3βHSDIID due to documented mutations in the HSD3B2 gene. Two patients had a 46,XY karyotype and 1 patient had a 46,XX karyotype. All 3 patients with 3βHSDIID were in AG1 (<1 year) at the time of diagnosis, and all samples were measured before first initiation of treatment. These patients carried the following mutations: P1 (46,XY) was compound heterozygous for c.299A>G in exon 3 (p.N100S) (31) and for c.721delG in exon 4 (pD241fsX); P2 (46,XY) was homozygous for c1022C>T in exon 4 (P341L); and P3 (46,XX) was homozygous for c.245C>A in exon 3 (p.Ala82Asp). The missense mutation N100S was the first mutation in the HSD3B2 gene, which affected the affinity and specific activity of the HSD3B2 enzyme described by Mébarki et al. (31). The missense mutation P341L was examined by Welzel et al. (32), causing a residual conversion activity of 6% of the wildtype activity. The mutation in P3 has not been described in the literature. Therefore, the PolyPhen-2 (Polymorphism Phenotyping v2.1) (33) tool was applied, predicting that the p.Ala82Asp mutation is probably damaging. Moreover, Rabbani et al. (34) showed that alanine is a conserved amino acid in the membrane-binding domain of the enzyme.

Statistical analysis

Each group was tested for normal distribution with a Kolmogorov-Smirnov test with Lilliefors correction for each hormone. Nonnormally distributed variables were compared by Mann-Whitney test and a P value <0.05 was recorded as significant.

All statistical analyses were performed with SigmaStat (Systat Software, Erkrath, Germany). Medians and reference ranges were calculated using the Harrell-Davis estimator (35).

Results

Assay performance characteristics

These characteristics are presented in the supplemental data.

DHEAS

In control children, median DHEAS peaked in AG1 in both sexes and then sharply declined after the first year of life (Table 1; Fig. 1). Subsequently, there was a continuous increase in both sexes, with a maximum in AG8 (Table 1; Fig. 1). As expected, median DHEAS also increased with higher Tanner stages (PH) in both sexes (Table 2; Fig. 2). After the first year of life, DHEAS concentrations were always higher in control boys than in control girls and were statistically significant in AG4, AG5, and AG8 (Table 1). Median DHEAS in control boys was significantly higher than in control girls only in Tanner stage (PH) 5 (Table 2). It must be noted that the number of control children with Tanner stage (PH) 2 or higher was comparably low in each group, which may have compromised detection of potentially significant differences.

Obese children had significantly higher median DHEAS concentrations than did control children in both sexes at the same age (Table 1; Fig. 1). When we displayed DHEAS concentrations against the Tanner stages (PH), the difference between obese children and control children was still very noticeable; however, this was statistically significant only in the largest group with Tanner stage (PH) 1 (Table 2; Fig. 2). As with the control children, the number of obese children in each Tanner stage (PH) was comparably low, which may have compromised the detection of statistically significant differences.

We also calculated the median ages of both control and obese children in each Tanner stage (PH) and for each sex separately whenever these data were available. Table 3 demonstrates that in all Tanner stages (PH), the median ages of obese girls and boys were significantly younger than the median ages of the girls and boys in the control groups (Table 3). Thus, in this UPLC-MS/MS data set, after the first year of life, median DHEAS appears to reflect primarily not only increasing pubertal maturation but also the sex difference between boys and girls per se, as well as BMI.

17-Hydroxypregnenolone

Median 17OHPreg concentration also peaked in AG1 (Table 1; Fig. 1). As with DHEAS, there was a remarkable decrease in serum concentration in the first year of life compared with AG1 and AG2 in both sexes. In contrast to DHEAS, median 17OHPreg plasma concentrations did not subsequently increase, either with increasing age (Table 1; Fig. 1) or with higher Tanner stage (PH) (Table 2; Fig. 2). Table 1 shows only a slight increase in 17OHPreg concentration after AG2. Also in contrast to DHEAS values, control girls mostly tended to have higher 17OHPreg concentrations compared with control boys; this was statistically significant in age groups 1, 2, and 5 (Table 1). However, when we displayed the data against Tanner stages (PH), this sex difference partly disappeared, with the exception of prepubertal control children, for whom the girls still had statistically significantly higher plasma concentrations (Table 2).

The obese children showed higher median 17OHPreg concentrations than the control children in both sexes at the same age, starting with AG2 (Table 1; Fig. 1). When we displayed median 17OHPreg concentrations in obese children against the Tanner stages (PH), this difference between obese and control children was still present (Table 2; Fig. 2) and was statistically significant in most cases despite the smaller number of individuals per Tanner stage (PH) (Table 2; Fig. 2). Interestingly, elevation of median 17OHPreg levels in the obese children was more prominent in obese boys than in obese girls at the same Tanner stage (PH) (Table 2). In essence, the sex influence on 17OHPreg was reversed in obese children compared with normal-weight control children, with higher 17OHPreg levels in obese boys (Table 2). Therefore, in our UPLC-MS/MS data set, after the first year of life, the median 17OHPreg plasma concentrations appear to reflect primarily the BMI, a BMI-influenced sex difference between girls and boys, and to a limited extent only, pubertal maturation, which contrasts with the DHEAS data.

Four different patterns of Δ5 pathway steroids in obese children

Of the 231 obese children, 46 (20%) showed elevated concentrations of both DHEAS and 17OHPreg above the reference range as presented here for control children with normal BMI. However, a second subset of obese children showed elevation of DHEAS level only in the presence of a normal 17OHPreg concentration (35; 15%). A third subset of obese patients showed an increased 17OHPreg concentration but a normal DHEAS level (49; 21%). A fourth subset showed concentrations for both steroids within the reference range (Supplemental Table 1). We have expressed the four different patterns in the form of ratios: 17OHPreg/DHEAS × 100 (Fig. 3). The age-dependent decline of curves is due to the continuous increase of DHEAS concentration with age. Interestingly, obese children with an elevation of 17OHPreg—regardless of whether they had elevated DHEAS—showed significantly higher serum concentrations of 17OH-progesterone and cortisol than obese children with only isolated DHEAS elevation (Supplemental Fig. 1).

3βHSDII deficiency

The 3 patients with proven 3βHSDIID had extremely elevated serum concentrations of 17OHPreg compared with the respective age group. In 2 cases, DHEAS was in the top third of the reference range, whereas patient 3 had a clearly elevated DHEAS concentration: P1(46,XY): 17OHPreg, 178 ng/mL (ref, 1.76 to 6.86); DHEAS, 530 ng/mL (ref, 15.47 to 708.1); P2(46,XY): 17OHPreg, 130 ng/mL (ref, 1.76 to 6.86); DHEAS, 500 ng/mL (ref, 15.47 to 708.1); and P3(46,XX): 17OHPreg, 247 ng/mL (ref, 2.72 to 15.86); DHEAS, 2947 ng/mL (ref, 10.59 to 933).

Discussion

We have presented a combined data set of DHEAS and 17OHPreg, 2 important Δ5 steroids, measured in parallel by a UPLC-MS/MS method and covering the whole pediatric age range. We found age, pubertal maturation, and sex effects, as well as a significant difference in both 17OHPreg and DHEAS levels between children with normal weight and obese children. This suggests that different normal values regarding age, pubertal maturation, sex, and weight status should be used for meaningful hormone data interpretation in pediatric endocrine patient care and research. Extreme elevation of 17OHPreg levels and high concentrations of DHEAS in patients with molecular proven 3βHSDIID represent an independent validation of this method.

General comparisons of our data on DHEAS with data previously published by other groups revealed differing aspects. On one hand, in a previous study by Soeborg et al. (36), DHEAS concentrations were considerably higher than our data. One explanation might be that the Soeborg study did not separate individuals according to normal BMI and increased BMI, as we did in our study. On the other hand, our data are in concordance with findings of several other studies that measured DHEAS in obese children (16–18, 37). In contrast to our study, the studies mentioned focused only on the age range of 5 to 9 years, the time of adrenarche, whereas our study included the entire pediatric age range. We confirmed the finding of a Spanish study that showed higher DHEAS concentrations in obese boys than in obese girls (17).

Reference data on 17OHPreg in children are extremely sparse. Several years ago, Riepe et al. (38) published reference ranges for the first year of life in which 17OHPreg was measured by an RIA assay prior to extraction. In contrast, we looked specifically at 17OHPreg in plasma in obese children in the whole age range.

Interestingly, after the neonatal period, median DHEAS and median 17OHPreg concentrations showed very different patterns across age groups and Tanner stages (PH) in nonobese control children. Increasing DHEAS concentrations in the presence of largely constant 17OHPreg concentrations suggest differential maturation and regulating mechanisms of 17α−hydroxylase vs 17,20-lyase. This is in line with published data associating higher expression of cytochrome b5 during adrenarche with activating 17,20-lyase (7). The persistence of generally higher median DHEAS concentrations in control boys vs control girls after alignment of data to Tanner stage (PH) supports the notable influence of sex background on adrenarche. The absence of a clear relationship of median 17OHPreg concentration with increasing age and increasing Tanner stage (PH) further suggests that adrenal delivery of this steroid precursor is not modulated primarily by ongoing adrenarche in the normal-weight control children. Rather, our data indicate that the 17α−hydroxylase activity of the CYP17 enzyme appears to deliver 17OHPreg independent of age and pubertal stage, similar to a continuous “flat rate.” This pattern is physiologically rational because 17α−hydroxylase is crucial for vital glucocorticoid supply, which has to function independently during adrenarche and puberty.

This situation becomes modified in obesity. Our data confirmed earlier reports by ourselves and others showing that median DHEAS plasma concentrations were elevated in obese children compared with control children of the same age (16, 18, 37). However, in our cohort, the obese children were significantly younger than the normal-weight control children within the same Tanner stage (PH) groups, suggesting that obese children have reached the same pubertal stage earlier (Table 3). This could support the hypothesis that earlier adrenarche is the starting factor behind higher DHEAS concentrations in obese children. This view is in line with a cross-sectional study in which no direct association between BMI and higher DHEAS concentrations were shown (39) and with a longitudinal study demonstrating no decrease in DHEAS after weight loss (18). However, this view is not fully supported by our current data. Rather, the persistence of generally higher DHEAS values, even after alignment of hormone concentrations to Tanner stages (PH), supports the supposition that in obesity BMI must have independent modifying effects on DHEAS biosynthesis, in addition to adrenarche, that do not work primarily through pushing adrenarche. This is supported by the presence of sulfatase in fat tissue (40). Our interpretation is also substantiated by the concentration pattern of 17OHPreg observed in this study. The largely age-independent and Tanner stage (PH)–independent elevation of 17OHPreg in obese children, particularly in boys, suggests that this steroid is an important driver of enhanced adrenal activity in obesity, eventually contributing, at least in part, to higher DHEAS concentrations in obese children regardless of the state of adrenarche.

A more detailed look at our data revealed the existence of qualitatively different signatures of elevated Δ5 steroids in obese children. In particular, not all obese children with an elevated DHEAS concentration showed an elevated 17OHPreg level, and not all children with an elevated 17OHPreg concentration showed an elevated DHEAS level. Interestingly, median plasma 17OH-progesterone and cortisol concentrations were significantly higher in those obese children with elevated 17OHPreg levels (Supplemental Fig. 1), supporting enhanced glucocorticoid synthesis. In conclusion, 17α−hydroxylase converting pregnenolone to 17OHPreg is probably an important biological target in a relevant subset of obese children, contributing to the higher glucocorticoid levels known in obesity (18). In some patients, 17OHPreg may also contribute to hyperandrogenism via feeding the Δ5 pathway. This is supported by a study showing activation of 17α−hydroxylase by leptin and insulin (41), which are elevated in obesity (42). Furthermore, we have previously found that higher concentrations of mineralocorticoids in prepubertal obese children may contribute to elevated blood pressure levels (18). Prospective studies correlating the individual clinical manifestation of obesity phenotypes (e.g., the degree of metabolic syndrome and hirsutism) to different steroid signatures would be interesting.

Weaknesses of our study are the cross-sectional design, the relatively small subgroup size despite the large number of study individuals, the possibility that hormone data may have been influenced by stress and circadian rhythm, and the fact that the p.Ala82Asp mutation in P3 has not yet been studied in vitro. Another weakness is the methodological focus on only 2 steroid hormones. Moreover, DHEAS is already the sulfated form of DHEA, which we did not measure directly. However, there are solid data showing that DHEA and DHEAS increase in parallel with increasing age and increasing puberty (43). It is also important to bear in mind that other endocrine mechanisms that are not covered by measuring 17OHPreg and DHEAS influence the complex process of adrenal maturation, both in control children and in obese children. Of these, the inhibition of HSD3B2 through high cortisol and corticosterone levels in obesity, stimulating DHEA production (44) and higher expression of SULTA1 on the zona reticularis with increasing adrenarche (7), are important.

In summary, we present a reliable UPLC-MS/MS method for parallel determination of 17OHPreg and DHEAS concentrations. This method needs only very small sample volumes of 100 μL, which is an advantage for pediatric patients. Because of the low injection volume of 5 μL, a rerun of the sample should always be possible if required. The method has a very low limit of quantification of 3 ng/dL (0.1 nmol/L) and a short analysis time of 3.2 minutes. We present reference data covering the entire pediatric age range, sex, and pubertal development, thus enabling the diagnosis of various endocrine disorders including 3βHSDIID. Moreover, our UPLC-MS/MS data set supports the separation of different mechanisms in the complex physiological concert of adenarche contributing to regulation of the Δ5 pathway. Although 17,20-lyase appears to be largely regulated by factors underlying adrenarche, 17α−hydroxylase and 17OHPreg are predominantly age independent but BMI dependent and may contribute to enhanced activation of the Δ5 steroid pathway in obesity. Therefore, not only normal age-, sex- and pubertal stage–specific values should be used, but also weight status must be considered.

Abbreviations:

 
• 3βHSDIID

  3β-hydroxysteroid dehydrogenase deficiency

 
• 17OHPreg

  17-hydroxypregnenolone

 
• BMI

  body mass index

 
• DHEA

  dehydroepiandrosterone

 
• DHEAS

  dehydroepiandrosterone sulfate

 
• HSD3B2

  3β-hydroxysteroid dehydrogenase type 2

 
• LC-MS/MS

  liquid chromatography tandem mass spectrometry

 
• SULT2A1

  steroid sulfotransferase

 
• UPLC-MS/MS

  ultra-performance liquid chromatography tandem mass spectrometry

Acknowledgments

The authors thank Susanne Olin, Brigitte Karwelies, Tanja Stampe, Gisela Hohmann, Sabine Stein, and Silke Struve for excellent technical assistance.

Disclosure Summary: The authors have nothing to disclose.

References

Author notes