Obesity Is Associated with Increased 11-Oxyandrogen Serum Concentrations during Puberty

While the effects of classical androgens – such as testosterone (T), androstenedione (A4), and dehydroepiandrosterone (DHEAS) – on puberty have been widely investigated, the role of adrenal-derived 11-oxygenated androgen (11-OA) levels has received less attention. In the adrenal cortex, 11-β-hydroxyandrostenedione (11-OHA4) and 11-β-hydroxytestosterone (11-OHT) are synthesized from adrenal A4 and T, whereas in the periphery, primarily in the kidneys, further oxidation produces 11-ketoandrostenedione (11-KA4) and 11-ketotestosterone (11-KT). In addition, e.g., in the adipose tissue, 11-KA4 can also be activated to 11-KT, similar to the activation from A4 to T. Thus, the previously known pool of potent androgens is enlarged by 11-KT, which is equally as potent as T [1, 2]. Interestingly, this newly identified pool of androgens accounts for approximately half of all circulating androgens in women, highlighting the importance of its impact on sexual development [3].

Among various disorders of androgen excess – such as congenital adrenal hyperplasia, polycystic ovary syndrome (PCOS), and androgen-dependent tumors – 11-OA levels also seem to be associated with the timing of adrenarche and pubarche [4]: As a stage of puberty, adrenarche refers to the onset of androgen synthesis in the adrenal cortex, typically occurring in children between ages 6 and 8, and usually leading to pubarche, which is defined as the growth of pubic hair [5, 6]. As girls with premature adrenarche and pubarche (appearance of pubic and axillary hair before the age of 8) have a higher prevalence of metabolic syndrome and elevated insulin concentrations [6‒8], a high body mass index (BMI) appears to be a risk factor for an early onset of puberty.

In both sexes, 11-OA levels typically increase throughout puberty, with the upper limits of reference intervals being very similar between boys and girls, both pre- and postpubertally [9]. However, the trajectories of 11-OA concentrations in children and adolescents have typically been studied in the context of premature adrenarche. Information is lacking for children who experience puberty at an average onset, especially for boys; and, despite associations between premature adrenarche and metabolic syndrome, results concerning the association of overweight and obesity and 11-OA levels are contradictory.

AKR1C3 efficiently catalyzes the activation of 11-KA4 into 11-KT. AKR1C3 expression levels in subcutaneous adipose tissue correlate with BMI and are reduced after weight loss [1]. In addition to the conversion of 11-KA4 to 11-KT by AKR1C3, obesity-associated changes in glucocorticoid metabolism may influence 11-OA levels. HSD11B Type 1 and 2 activate and inactivate both glucocorticoids and 11-OAs and are expressed in adipose tissue [2, 3]. Altogether, this suggests a complex interplay of adipose tissue-derived enzymes where the effects of one enzyme may be superimposed by others.

The aim of this study was to examine 11-OA serum concentrations in the context of the physiological onset and progression of puberty in a large cohort of healthy children from the LIFE Child study Leipzig [10]. Additionally, we investigated how 11-OA levels might be associated with overweight and obesity by using measurements of BMI and skinfold thickness and examining their associations with the timing and magnitudes of 11-OA levels. Steroid hormone concentrations, including 11-OAs, were simultaneously determined from single serum samples by employing a recently developed liquid chromatography-tandem mass spectrometry (LC-MS/MS) method [11].

The data for this study were derived from the LIFE Child study, a large population-based cohort study conducted in Leipzig, Germany, as part of the Leipzig Research Center for Civilization Diseases (LIFE) [10]. Since it began in 2011, the study has examined over 5,500 children and has stored biological samples from more than 18,000 visits for long-term analysis. Comprehensive assessments – including physical examinations, standardized tests, questionnaires, and interviews – have been conducted. One of the main focuses of LIFE Child lies on obesity-related measurements, such as the assessment of the Body Mass Index Standard Deviation Score (BMI-SDS) and the measurement of skinfold thickness at various body locations.

Children with anthropometric measurements who were in the pubertal stage and were available to provide serum samples without thyroid disorders or epilepsy and who were not taking hormone-based medication, antidiabetics, neuroleptics, or immunosuppressants were selected as the base sample. Serum samples from 1,784 children across a total of 5,649 visits met the inclusion criteria. Subsequently, serum samples from 1,054 children (506 boys and 548 girls) aged 6 to 19 years across 1,734 visits were randomly selected (stratified by Tanner stage) (shown in Fig. 1), including 403 serum samples in Tanner stage 1; 385 serum samples in stage 2; 283 serum samples in stage 3; 295 serum samples in stage 4; and 368 serum samples in stage 5 (shown in Table 1).

In addition, we analyzed data from a longitudinal subcohort. This included 420 measurement pairs that were not more than 1.5 years apart from each other. The mean difference was 1 year, reflecting the yearly follow-up visits.

The pubertal stage for both sexes was assessed by qualified personnel in accordance with Tanner [12]. To determine the total Tanner stage, the testicular size was the main criterion in boys and the breast size in girls. Pubic hair status and age were taken in account additionally. If not all parameters were recorded, the Tanner stage of pubic hair was used instead (applied to 162 serum samples from boys and 31 serum samples from girls). The standard deviation score (SDS) for BMI was determined in accordance with Kromeyer-Hauschild [13]. Skinfold thickness (iliac, subscapular, biceps, and triceps) was measured using calipometry (Holtain caliper, range 0–40 mm; Harpenden skinfold caliper, >40 mm) to estimate body fat.

Serum samples were obtained between 7 and 10 a.m., clotted for 20 min, centrifuged, aliquoted, and stored at < −80°C in the LIFE-Biobank. Steroid hormones in serum, which include T, A4, DHEAS, and the 11-OA subtypes 11-KT, 11-KA4, 11-OHT, and 11-OHA4, were simultaneously quantified using LC-MS/MS as described in detail by Zeidler et al. [11]. Inter- and intra-assay coefficients of variation for all four 11-OAs were between 2% and 15%, and mean recovery ranges were between 85% and 117% [11]. The serum samples that were used had undergone no more than two freeze-thaw cycles. In freeze-thaw stability experiments, the concentrations of 11-OAs were found to be stable across five cycles. For 11-OHT, there were 344 values below the detection limit of 0.064. We generated random values between 0 and the lower detection limit using a beta distribution with shape parameters 2 and 1.

Descriptive statistics are given as means and standard deviations for continuous measures and counts and percentages for categorical measures. BMI measures were transformed into age- and sex-adjusted SDSs in accordance with the evidence-based therapy guidelines of the German working group on obesity in childhood and adolescence [14]. Subsequently, weight groups were defined on the basis of these guidelines as follows: BMI-SDS ≤1.28: normal weight; 1.28 ≤BMI-SDS ≤1.88: overweight; BMI-SDS >1.88 adipose (shown in Table 2, 3). Skinfold measures and 11-OA values were also transformed into age- and sex-adjusted SDSs using LIFE-Child-based references (skinfolds: [15], 11-OA: publication in preparation). Linear mixed-effects models were used to analyze relationships between hormone levels, Tanner stage groups, and BMI-SDSs. Trend values of hormone levels across individual Tanner stages were reported as mean values. Because associations between hormone levels and BMI-SDSs differed significantly by age (cutoffs were identified by a stepwise approximation using regression results and visual inspection), two age groups were defined for each sex (males: age ≤13 and age >13; females: age ≤11 and age >11). All models were adjusted for multiple measurements per participant by including a random intercept. We estimated the associations between the change in 11-OA-SDS from the first measurement (t1) to the second measurement (t2) and (a) the change in BMI-SDS and (b) the association between the change in BMI-SDS adjusted for the value at t1. Furthermore, we estimated the association between the 11-OA-SDS at t1 and t2.

All models were adjusted for age and sex if necessary. The significance level α was set to α = 0.05. All analyses were carried out using R version 4 [16].

In boys, we found a continuous increase in 11-KT levels from 0.79 nmol/L in TS1 to 1.76 nmol/L in TS4 and in 11-KA4 levels from 0.34 nmol/L in TS1 to 0.63 nmol/L in TS5. 11-OHA4 showed an increase across all Tanner stages from 1.84 nmol/L to 4.38 nmol/L, whereas 11-OHT increased in boys only from 0.11 nmol/L in TS1 to 0.14 nmol/L in TS2 and from 0.16 nmol/L in TS3 to 0.40 nmol/L in TS5 continuously (shown in Fig. 2).

In girls, 11-KT levels and 11-KA4 levels increased continuously only from TS1 to TS3 (11-KT: 0.82 nmol/L to 1.41 nmol/L; 11-KA4: 0.35 nmol/L to 0.53 nmol/L). The levels of 11-OHT and 11-OHA4 increased across all Tanner stages from 0.11 nmol/L to 0.35 nmol/L (11-OHT) and from 1.77 nmol/L to 3.87 nmol/L (11-OHA4) (shown in Fig. 2).

For the classic androgens in boys, a continuous increase was observed between TS1 and TS5 for T (0.40 nmol/L to 18.9 nmol/L), A4 (0.70 nmol/L to 2.87 nmol/L), and DHEAS (1,585 nmol/L to 5,691 nmol/L) (shown in online suppl. Fig. 1; for all online suppl. material, see https://doi.org/10.1159/000540433). In girls, there was a continuous increase from Tanner stages 1 to 5 for T (0.22 nmol/L to 1.07 nmol/L) and DHEAS (1,201 nmol/L to 4,148 nmol/L), whereas, for A4, a continuous increase was observed only from Tanner stages 1–4 (0.77 nmol/L to 4.09 nmol/L) (shown in online suppl. Fig. 1).

Among boys aged 13 years and under, positive correlations were observed between all 11-OA measures and BMI (11-KT: β = 0.2, p < 0.001; 11-KA4: β = 0.16, p < 0.001; 11-OHT: β = 0.17, p < 0.001; 11-OHA4: β = 0.12, p = 0.003, shown in Fig. 3). In boys older than 13, the positive correlations remained only for 11-OHT (β = 0.14 p = 0.001). In terms of Tanner stages, 11-KA4 (β = 0.17, p = 0.008) and 11-OHA (β = 0.31, p = 0.031) exhibited correlations only in TS1, whereas 11-KT and 11-OHT showed correlations in TS1 (11-KT: β = 0.18, p = 0.004, 11-OHT: p = 0.17, p = 0.006) and higher Tanner stages (11-KT: β = 0.09, p = 0.015, 11-OHT: β = 0.16, p < 0.001) (shown in online suppl. Fig. 6).

Among girls, only 11-KT and 11-OHT showed a positive association with BMI. Whereas 11-KT exhibited positive correlations with BMI only in girls aged 11 years and under (β = 0.1, p = 0.047), 11-OHT displayed positive correlations in both age groups (11 and under: β = 0.17, p < 0.001; over 11 years: β = 0.18, p < 0.001, shown in Fig. 3). 11-OHT correlated in TS1 (β = 0.17, p = 0.005) and higher Tanner stages (β = 0.19, p < 0.001), and 11-OHA (β = 0.07, p = 0.049) was correlated with BMI in TS2 and higher (shown in online suppl Fig. 6).

In boys, associations between skinfold thickness and 11-OA exhibited similar patterns to those found for BMI, with correlations between all 11-OA and triceps, biceps, subscapular, and iliac skinfold measurements: We found similar positive associations in boys aged 13 years and under and no correlations above 13 years, except for the correlations between 11-OHT and triceps, biceps, and subscapular skinfold thickness observed in boys older than age 13 years (shown in Table 4, online suppl. Fig. 2–5).

In girls, also as observed with BMI, 11-OHT was significantly associated with all skinfold measurements in both the 11 and under group and the over 11 group. In girls aged 11 years and under, 11-KT was correlated only with subscapular and iliac skinfold thickness. Different from BMI, in girls older than 11, 11-OHA4 was positively associated with triceps and subscapular skinfold thickness (shown in Table 5, online suppl. Figs. 2–5).

The follow-up measurements’ SDSs were strongly associated with the respective SDS 1 year before (all betas ∼0.6, p < 0.001). The percentage of explained variance varied between 25% and 35%. Further, the change rates Δ11-KT-SDS (β = 0.46, p < 0.001), Δ11-KA4-SDS (β = 0.44, p < 0.001), and Δ11-OHA4 (0.35, p = 0.008) were significantly associated with the respective change in BMI-SDS. The effect sizes of the associations got slightly smaller but persisted after adjusting for the initial 11-OA-SDS (shown in online suppl. Table 1). The effect for Δ11-OHT was considerably smaller and did not reach statistical significance. It vanished completely after adjustment for the initial value (shown in online suppl. Table 1).

Adrenal-derived 11-OAs contribute to the phenotypic features of puberty having the same androgenic potency as T; thus, premature adrenarche is characterized by early increases in circulating 11-OAs [17]. Our study demonstrates a continuous increase in the levels of 11-OHA4, 11-OHT, 11-KA4, and 11-KT throughout puberty. The diagnosis of the beginning of adrenarche has traditionally relied on an increase in DHEAS serum concentration as an indicator of adrenal maturation [5]. 11-OAs are less diagnostically discriminant (sensitivity 67–81%) than DHEAS or DHEA (sensitivity 100%) in distinguishing premature adrenarche from normal adrenarche [18]. However, cases of premature pubarche have been documented in the absence of DHEAS levels above the adrenarchal cutoff of 50 μg/dL. Also, DHEA and DHEAS are not bioactive androgens and studies have shown that serum DHEAS concentrations do not correlate well with clinical signs of adrenarche [19]. These findings suggest that elevated 11-OA levels – rather than DHEA, DHEAS, and A4 – may be of particular diagnostic and clinical significance in cases of premature pubarche and may thus mediate the phenotypic changes in hair growth in pubarche [1, 7, 20]. Furthermore, to address the partially conflicting results of previous studies regarding the influence of BMI on 11-OA levels, we examined a large cohort of healthy children and adolescents and included skinfold thickness measurements and longitudinal data to validate the results.

Positive correlations between 11-KT, 11-KA4, 11-OHA4, and BMI were observed in boys aged 13 years and under, but no correlations were found in older age groups. Only in girls aged 11 years and under, 11-KT was correlated with BMI. For 11-OHT in boys and girls, a positive correlation with BMI was observed regardless of age, but the results were not significant in the longitudinal data analysis (likely due to the low concentration levels in serum).

The use of a statistically approximated age cutoff provided more precise results than the use of Tanner stages. One possible explanation for this discrepancy is the visual assessment of Tanner stages, which can be challenging to interpret, particularly in overweight and obese children. Additionally, 11-OAs are suspected to primarily influence the timing of adrenarche and pubarche, rather than gonadarche, menarche, and thelarche [19]. Consequently, the conventional method of determining Tanner stage based on breast and testicular size may be less precise. However, calculations based on the Tanner staging system yield similar results to those obtained using age cutoffs. It is noteworthy that our exploratory age thresholds reflect the average age of puberty onset for each gender, but further investigations are needed to determine the underlying cause for the observed changes, specifically at the age of 11 years in girls and 13 years in boys.

The age-dependent associations between 11-OA levels and BMI were confirmed through the use of skinfold thickness measurements. Skinfold thickness has been shown to provide more information than using BMI alone as a parameter for overweight or obesity, as demonstrated by Rönnecke et al. [15]. Marković-Jovanović [21] observed that the use of BMI as a sole parameter showed lower sensitivity and specificity in the diagnosis of obesity in children than BMI combined with skinfold measurements. Subscapular skinfold thickness, waist circumference, and waist/height ratio showed stronger correlations with serum insulin levels, low-density lipoprotein cholesterol, and family history than BMI itself. Freedman et al. [22] demonstrated that information on the skinfold sum significantly (p < 0.001) improved the prediction of obesity beyond that obtained with BMI-for-age and reduced the overall prediction errors for percentage of body fat by 20–30%.

In previous studies, the role that 11-OA levels play in the onset of puberty has been studied primarily in children with premature adrenarche, where serum concentrations of 11-KT and 11-OHT were all significantly higher than in children with a physiological onset of adrenarche [17]. Despite associations between premature adrenarche and metabolic syndrome [6‒8], studies examining girls with premature adrenarche found no associations between BMI and 11-OAs; however, Burt Solorzano et al. [23] examined only 11 premenarcheal girls versus 24 postmenarcheal girls with and without overweight, whereas Rege et al. [17] tested the correlation between 11-KT and BMI exclusively in 91 girls with and without premature adrenarche.

By contrast, in one of the few studies involving 249 boys and girls with an average onset of adrenarche, Breslow et al. [24] described the modest increase in 11-OA levels in younger obese girls, similar to our results. In the study by Breslow et al., data were lacking, particularly for young boys; thus, Tanner stage 1 was not included, and Tanner stages 2 and 3 had to be grouped together and involved only 26 boys. To fill this gap in data, the results of the current study indicated a correlation between BMI and all 11-OAs for boys, a finding that, similar to girls, predominantly manifests at a young age. In accordance with this finding, Marakaki et al. [25] demonstrated that children with premature adrenarche were significantly larger and more obese than the control group not only in puberty, but already in the first few years of life. High 11-OA levels have been discussed in numerous studies as a potential trigger for premature adrenarche and PCOS, and both pathologies are associated with obesity [26]. Studies on premature adrenarche and pubarche outcomes have revealed associations with central obesity, insulin resistance, ovarian hyper-responsiveness to GnRH stimulation, oligo/amenorrhea, and hirsutism [19, 27]. Our results suggest that, in addition to the adiposity-associated imbalance of classical androgens in childhood [28], an adiposity-associated imbalance of 11-OA in children and adolescents may also play a role in the timing of adrenarche and the development of PCOS in adulthood. Measurement of 11-OA levels in childhood could therefore serve for the early detection of an increased risk of metabolic syndrome, adult PCOS, and other adrenal disorders as congenital adrenal hyperplasia. Corresponding reference values for clinical application are being developed in the aforementioned study by Zeidler et al. [11].

Our study’s main strength lies in its large and balanced cohort, comprising both boys and girls and exclusively consisting of healthy children, qualities that distinguish it from previous studies. In addition to BMI, skinfold thickness as a marker of amount of body fat was utilized to evaluate the associations of overweight and serum 11-OA levels. The utilization of LC-MS/MS allows for the precise and simultaneous quantification of steroid hormone levels. Further research directions in this field include investigating the impact of underweight on 11-OA levels and exploring the effects of hormone and glucocorticoid therapies. Future research is also needed to determine which factors are responsible for the termination of the associations between 11-OAs and BMI-SDS at the age of 11 years in girls and 13 years in boys. It would be of interest to conduct investigating the hormonal influences on enzymes located within adipose tissue, as ARK1C3 and HSD11B Type 1 and 2. Particularly, the investigation of the interrelationship between corticosteroids and 11-OAs may yield novel insights.

In conclusion, our results show that 11-OA levels increase during the course of puberty and are associated with BMI-SDSs mainly in younger children. Overweight and obesity may potentially influence the onset and progression of puberty by affecting 11-OA levels.

The results presented here are part of the LIFE Child study. We wish to thank everybody in the team performing the examinations and, of course, the children who participated in the study and their parents for their collaboration.

The study was designed pursuant to the declaration of Helsinki and was approved by the Ethical Committee of the University of Leipzig (Reference No. Reg. No. 264-10-19042010). LIFE Child is registered by the trial number: NCT02550236. Written informed consent was obtained from all participants’ parents or legal guardians prior to their participation (at every visit) in the study. In addition, the participants themselves gave their informed consent from the age of 12 years.

The authors have no conflicts of interest to declare.

This publication is supported by LIFE – Leipzig Research Center for Civilization Diseases, University of Leipzig. LIFE is funded by the European Union via the European Social Fund (ESF), by the European Regional Development Fund (ERDF), and by the Free State of Saxony within the framework of the excellence initiative of the Saxonian Ministry of Science and Arts (SMWK), Free State of Saxony, Germany. We acknowledge support from the German Research Foundation (DFG) and Universität Leipzig within the program of Open Access Publishing.

Wieland Kiess and Jürgen Kratzsch designed the research. Friederike Wagner and Robert Zeidler conducted the research. Robert Zeidler and Alexander Gaudl performed measurements. Mandy Vogel and Friederike Wagner analyzed the data and performed statistical analysis. Friederike Wagner and Ronald Biemann wrote the main manuscript. Wieland Kiess, Jürgen Kratzsch, Mandy Vogel, Alexander Gaudl, and Uta Ceglarek discussed the results, contributed valuable advice, and edited the manuscript. The manuscript was critically revised by all authors.

Friederike Wagner and Robert Zeidler share first authorship.Mandy Vogel and Ronald Biemann share last authorship.

The dataset presented in this article cannot be shared publicly because of ethical restrictions. The LIFE Child study is a study collecting potentially sensitive information. Publishing data are not covered by the informed consent provided by the study participants. Furthermore, the data protection concept of LIFE requires all (external as well as internal) researchers interested in accessing data to sign a project agreement. Researchers interested in accessing data from the LIFE Child study may contact the study by writing to forschungsdaten@medizin.uni-leipzig.de.