https://orcid.org/0000-0003-1669-6383
Abstract
For many decades, the prevailing paradigm in endocrinology was that testosterone and 5α-dihydrotestosterone are the only potent androgens in the context of human physiology. The more recent identification of adrenal derived 11-oxygenated androgens and particularly 11-ketotestosterone have challenged these established norms, prompting a revaluation of the androgen pool, particularly in women. Since being recognized as bone fide androgens in humans, numerous studies have focused their attention on understanding the role of 11-oxygenated androgens in human health and disease and have implicated them as role players in conditions such as castration resistant prostate cancer, congenital adrenal hyperplasia, polycystic ovary syndrome, Cushing's syndrome, and premature adrenarche. This review therefore provides an overview of our current knowledge on the biosynthesis and activity of 11-oxygenated androgens with a focus on their role in disease states. We also highlight important analytical considerations for measuring this unique class of steroid hormone.
Introduction
The term 11-oxygenated androgens refers to a group of adrenal derived C19 steroids that include both inactive precursors and potent androgens. Although the precursor 11β-hydroxyandrostenedione (11OHA4) was identified in the 1950s, it was considered a dead-end product of adrenal steroidogenesis and a way to prevent excessive androgen biosynthesis.1,2 Studies in the last decade making use of mass spectrometry approaches have however shown that 11OHA4 is in fact a precursor to the potent androgen 11-ketotestosterone (11KT), that 11-oxygenated androgens contribute to the androgen pool in humans, and that they are important role players in several disease states. This review therefore provides an overview of the biosynthesis, circulating levels, and activity of 11-oxygenated androgens, as well as providing an up-to-date summary of studies which have investigated the role of 11-oxygenated androgens in human health and disease.
Origins and biosynthesis
The 11-oxygenated androgens owe their origins to the adrenal cortex due to the adrenal specific expression of cytochrome P450 11β-hydroxylase (CYP11B1), which catalyzes the commitment step during their biosynthesis, namely the 11β-hydroxylation of androstenedione to 11OHA4.3 One study has suggested that 11-oxygenated androgens are produced in human gonads; however, testicular and ovarian expression of CYP11B1 is negligible in comparison to that of adrenal expression.4 Moreover, Polson et al.5 were unable to detect 11OHA4 in adrenalectomized women or a woman receiving chronic dexamethasone therapy, while Turcu et al.6 detected only trace levels of 11-oxygenated androgens in patients with 11β-hydroxylase deficiency and adrenal insufficiency. Using ovarian vein sampling, Auer et al.7 confirmed that 11-oxygenated androgens are not produced by the human ovary. Furthermore, the biosynthesis of 11-oxygenated androgens is not regulated by the hypothalamic-pituitary-gonadal axis (HPG axis) but is instead under the control of the hypothalamic-pituitary-adrenal axis (HPA axis) via adrenocorticotropic hormone (ACTH) and can therefore be suppressed by glucocorticoids, thereby confirming their adrenal origin.5,6,8-12
As with other adrenal derived androgens and androgen precursors, biosynthesis of 11-oxygenated androgens is reliant on the Δ5-pathway in which pregnenolone is converted to dehydroepiandrosterone (DHEA) by the sequential 17α-hydroxylase and 17,20-lyase reactions catalyzed by cytochrome P450 17α-hydroxylase/17,20-lyase (CYP17A1).13 A portion of the resulting DHEA is converted to androstenedione by 3β-hydroxysteroid dehydrogenase type 2 (HSD3B2) (Figure 1). CYP11B1 subsequently catalyzes the 11β-hydroxylation of androstenedione to produce 11OHA4, the most abundant 11-oxygenated androgen produced by the human adrenal cortex.9,14 CYP11B1 can also 11β-hydroxylate adrenal-derived testosterone; however, adrenal testosterone biosynthesis is relatively low and as such the vast majority of circulating 11-oxygenated androgens owe their origins to adrenal derived 11OHA4.15 Although the adrenal does appear to produce low levels of 11-ketoandrostenedione (11KA4) and 11KT directly, the majority of these 11-oxygenated androgens are produced by peripheral tissues due to the high tissue specific expression of the required enzymes.15 Circulating 11OHA4 serves as a substrate for 11β-hydroxysteroid dehydrogenase type 2 (HSD11B2), which is abundantly expressed in mineralocorticoid target tissues such as the kidney.16,17 The resulting 11KA4 is released into circulation and can subsequently be converted to 11KT by peripheral tissue such as adipose, which express the key androgen activating enzyme aldo-keto reductase 1C3 (AKR1C3).18-20 Importantly, the conversion of 11OHA4 to 11KA4 by HSD11B2 is a prerequisite for the biosynthesis of 11KT by AKR1C3, as 11OHA4 is itself not a substrate for AKR1C3.17,19
Schematic overview of 11-oxygenated androgen biosynthesis. CYP11B1 catalyzes the 11β-hydroxylation of androstenedione in the adrenal cortex yielding 11OHA4, which is released into circulation. 11OHA4 can be converted to 11KA4 in mineralocorticoid tissues expressing HSD11B2. The resulting 11KA4 can subsequently be converted to 11KT in peripheral tissue (such as adipose) which expresses AKR1C3. Additional expression of HSD11B1 in some peripheral tissues, including adipose, converts 11KA4 and 11KT to 11OHA4 and 11OHT, respectively. These can in turn be converted back to 11KA4 and 11KT in mineralocorticoid target tissue. Therefore, there is a constant interconversion of 11β-hydroxy and 11-keto as indicated by the green and purple arrows. Interconversion between the androstenedione and testosterone forms also occurs due to conversion of 11KA4 to 11KT by AKC1C3 and the conversion of 11KT and 11OHT to 11KA4 and 11OHA4 by oxidative 17β-hydroxysteroid dehydrogenases (eg, HSD17B2); however, only the AKR1C3 catalyzed activation is shown for the sake of simplicity. Steroids are color-coded according to their bioactivity: general precursors (yellow); androgen precursors (light blue); active androgens (dark blue). Arrows are labeled with the catalyzing enzyme and isoform where appropriate. Essential cofactor proteins are also indicated: AKR1C3, aldo-keto reductase 1C3; ADX, adrenodoxin; b5, cytochrome b5; CYP11B1, cytochrome P450 11β-hydroxylase; HSD11B1, 11β-hydroxysteroid dehydrogenase type 1; HSD11B2, 11β-hydroxysteroid dehydrogenase type 2; POR, cytochrome P450 oxidoreductase; PAPSS2, PAPS synthase 2; StAR, steroidogenic acute regulatory protein; 11OHA4, 11β-hydroxyandrostenedione; 11KA4, 11-ketoandrostenedione; 11KT, 11-ketotestosterone; 11OHT, 11β-hydroxytestosterone.
Adipose and other glucocorticoid target tissues express 11β-hydroxysteroid dehydrogenase type 1 (HSD11B1), which can convert 11KA4 and 11KT to 11OHA4 and 11β-hydroxytestosterone (11OHT), respectively.16,17 Like 11OHA4, 11OHT is a substrate for HSD11B2 and can thus be converted to 11KT in mineralocorticoid target tissue.16,17 The low levels of 11OHT produced in the adrenal cortex can thus be converted directly to 11KT; however, an examination of the enzyme expression levels and activities suggest that this makes only a minor contribution toward circulating levels of 11KT in healthy individuals, with the majority of 11KT owing its origins to the peripheral metabolism of 11OHA4 as outlined above.15 Moreover, as 11KT can be converted to 11OHT by peripherally expressed HSD11B1, much of the circulating 11OHT must originate from peripheral derived 11KT and not directly from adrenal biosynthesis of 11OHT.15 A further complexity to the 11-oxygenated androgen pathway is that oxidative 17β-hydroxysteroid dehydrogenases such as HSD17B2 can catalyze the conversion of 11KT and 11OHT to 11KA4 and 11OHA4, respectively.17,19 Therefore, although circulating levels of the 11-oxygenated androgens are relatively constant, these are steady state concentrations achieved by the continuous peripheral interconversion of the individual 11-oxygenated androgens as well as the constant adrenal biosynthesis and peripheral and hepatic metabolism.13
Circulating levels
Growing interest in 11-oxygenated androgens has resulted in their measurement in numerous clinical studies, covering a number of heterogeneous disease states as well as healthy controls. 11OHA4 is universally the most abundant 11-oxygenated androgen in circulation with levels reported to be between 4 and 9 nM. 11KA4 and 11KT circulate at levels closer to 1 nM, though there are variations between studies that may reflect differences in methodology and/or cohort demographics. The majority of studies report levels of 11KT equal to or higher than those of testosterone in women21-24 (Table 1). Notably, unlike with the classical androgens, 11-oxygenated androgen levels do not decline substantially during aging, with the result that 11KT is the most abundant potent androgen in postmenopausal women.23-25 In men, the circulating levels of 11KT are similar to those reported in women, however, testosterone levels are ≥15-fold higher due to testicular biosynthesis24 (Table 2).
Circulating concentrations of classic and 11-oxygenated androgens in healthy women as reported in selected studies.
| DHEA | A4 | T | 11OHA4 | 11KA4 | 11KT | 11OHT | ||
|---|---|---|---|---|---|---|---|---|
| Schiffer et al. 202325 | 84 women aged <50a | 9.6 (4.2–20.0) | 2.3 (.7–5.3) | 0.61 (<0.4–1.2) | 7.5 (3.4–14.0) | 3.2 (1.6–5.4) | 0.70 (<0.3–1.6) | 0.39 (<0.3–0.8) |
| 81 women aged ≥50a | 5.0 (1.3–16.0) | 1.1 (<0.7–2.9) | 0.45 (<0.4–1.2) | 8.4 (3.9–16.0) | 1.8 (0.6–4.2) | 0.94 (0.4–1.7) | 0.63 (<0.3–1.2) | |
| Caron et al. 202126 | 10 women, follicular phase, aged 20–40 yearsb | 11.2 ± 0.8 | 2.8 ± 0.1 | 0.6 ± 0.03 | 5.2 ± 0.4 | 0.8 ± 0.1 | 0.9 ± 0.1 | 0.34 ± 0.04 |
| 10 women, luteal phase, aged 20–40 yearsb | 9.4 ± 0.5 | 3.8 ± 0.1 | 0.6 ± 0.03 | 4.6 ± 0.4 | 0.8 ± 0.1 | 0.9 ± 0.1 | 0.3 ± 0.04 | |
| 10 postmenopausal women aged 51–70 yearsb | 14.5 ± 0.5 | 3.0 ± 0.2 | 1.1 ± 0.1 | 11.5 ± 0.6 | 1.3 ± 0.03 | 1.8 ± 0.1 | 0.8 ± 0.03 | |
| Davio et al. 202024 | 72 women aged 18–39c | n.m. | 3.1 (2.3–4.1) | 1.1 (0.8–1.4) | 3.9 (2.8–5.0) | 0.5 (0.4–0.7) | 0.7 (0.5–1.1) | 0.4 (0.2–0.5) |
| 78 women aged 40–59c | n.m. | 1.8 (1.4–2.3) | 0.8 (0.7–1.1) | 4.0 (2.8–5.7) | 0.5 (0.4–0.7) | 0.9 (0.6–1.0) | 0.4 (0.3–0.6) | |
| 81 women aged 60–79c | n.m. | 1.2 (0.8–1.5) | 0.7 (0.5–1.2) | 4.1 (3.4–6.1) | 0.5 (0.4–0.7) | 0.9 (0.6–1.2) | 0.5 (0.4–0.8) | |
| 40 women aged ≥ 8 °c | n.m. | 0.9 (0.6–1.1) | 0.7 (0.5–1.0) | 4.2 (2.9–5.7) | 0.4 (0.3–0.6) | 0.7 (0.5–1.0) | 0.4 (0.3–0.6) | |
| Nanba et al. 201923 | 100 premenopausal women aged 20–40c | 8.0 (5.4–13.9) | 3.5 (2.6–4.9) | 1.0 (0.8–1.4) | 5.7 (3.9–8.7) | 1.2 (0.9–1.8) | 0.9 (0.6–1.3) | 0.5 (0.3–0.8) |
| 100 postmenopausal women aged ≥ 60c | 2.7 (1.8–4.6) | 1.8 (0.8–1.8) | 0.7 (0.5–1.1) | 6.5 (4.7–9.6) | 1.1 (0.8–1.5) | 0.9 (0.7–1.2) | 0.7 (0.4–0.9) | |
| Skiba et al. 201927 | 163 women, follicular phase, aged 18–40d | 5.3 (1.0–21.3) | 1.8 (0.6–5.4) | 0.3 (0.04–0.9) | n.m. | 8.8 (0.1–28.8) | 1.3 (0.3–7.6) | n.m. |
| 184 women, mid-cycle, aged 18–40d | 4.9 (0.4–23.5) | 2.1 (0.6–7.9) | 0.4 (0.1–1.0) | n.m. | 7.7 (1.1–31.7) | 1.3 (0.03–5.8) | n.m. | |
| 241; women, luteal phase, aged 18–40d | 4.8 (0.1–17.9) | 2.0 (0.5–6.8) | 0.4 (0.1–0.9) | n.m. | 7.7 (0.6–28.0) | 1.2 (0.1–4.6) | n.m. |
| DHEA | A4 | T | 11OHA4 | 11KA4 | 11KT | 11OHT | ||
|---|---|---|---|---|---|---|---|---|
| Schiffer et al. 202325 | 84 women aged <50a | 9.6 (4.2–20.0) | 2.3 (.7–5.3) | 0.61 (<0.4–1.2) | 7.5 (3.4–14.0) | 3.2 (1.6–5.4) | 0.70 (<0.3–1.6) | 0.39 (<0.3–0.8) |
| 81 women aged ≥50a | 5.0 (1.3–16.0) | 1.1 (<0.7–2.9) | 0.45 (<0.4–1.2) | 8.4 (3.9–16.0) | 1.8 (0.6–4.2) | 0.94 (0.4–1.7) | 0.63 (<0.3–1.2) | |
| Caron et al. 202126 | 10 women, follicular phase, aged 20–40 yearsb | 11.2 ± 0.8 | 2.8 ± 0.1 | 0.6 ± 0.03 | 5.2 ± 0.4 | 0.8 ± 0.1 | 0.9 ± 0.1 | 0.34 ± 0.04 |
| 10 women, luteal phase, aged 20–40 yearsb | 9.4 ± 0.5 | 3.8 ± 0.1 | 0.6 ± 0.03 | 4.6 ± 0.4 | 0.8 ± 0.1 | 0.9 ± 0.1 | 0.3 ± 0.04 | |
| 10 postmenopausal women aged 51–70 yearsb | 14.5 ± 0.5 | 3.0 ± 0.2 | 1.1 ± 0.1 | 11.5 ± 0.6 | 1.3 ± 0.03 | 1.8 ± 0.1 | 0.8 ± 0.03 | |
| Davio et al. 202024 | 72 women aged 18–39c | n.m. | 3.1 (2.3–4.1) | 1.1 (0.8–1.4) | 3.9 (2.8–5.0) | 0.5 (0.4–0.7) | 0.7 (0.5–1.1) | 0.4 (0.2–0.5) |
| 78 women aged 40–59c | n.m. | 1.8 (1.4–2.3) | 0.8 (0.7–1.1) | 4.0 (2.8–5.7) | 0.5 (0.4–0.7) | 0.9 (0.6–1.0) | 0.4 (0.3–0.6) | |
| 81 women aged 60–79c | n.m. | 1.2 (0.8–1.5) | 0.7 (0.5–1.2) | 4.1 (3.4–6.1) | 0.5 (0.4–0.7) | 0.9 (0.6–1.2) | 0.5 (0.4–0.8) | |
| 40 women aged ≥ 8 °c | n.m. | 0.9 (0.6–1.1) | 0.7 (0.5–1.0) | 4.2 (2.9–5.7) | 0.4 (0.3–0.6) | 0.7 (0.5–1.0) | 0.4 (0.3–0.6) | |
| Nanba et al. 201923 | 100 premenopausal women aged 20–40c | 8.0 (5.4–13.9) | 3.5 (2.6–4.9) | 1.0 (0.8–1.4) | 5.7 (3.9–8.7) | 1.2 (0.9–1.8) | 0.9 (0.6–1.3) | 0.5 (0.3–0.8) |
| 100 postmenopausal women aged ≥ 60c | 2.7 (1.8–4.6) | 1.8 (0.8–1.8) | 0.7 (0.5–1.1) | 6.5 (4.7–9.6) | 1.1 (0.8–1.5) | 0.9 (0.7–1.2) | 0.7 (0.4–0.9) | |
| Skiba et al. 201927 | 163 women, follicular phase, aged 18–40d | 5.3 (1.0–21.3) | 1.8 (0.6–5.4) | 0.3 (0.04–0.9) | n.m. | 8.8 (0.1–28.8) | 1.3 (0.3–7.6) | n.m. |
| 184 women, mid-cycle, aged 18–40d | 4.9 (0.4–23.5) | 2.1 (0.6–7.9) | 0.4 (0.1–1.0) | n.m. | 7.7 (1.1–31.7) | 1.3 (0.03–5.8) | n.m. | |
| 241; women, luteal phase, aged 18–40d | 4.8 (0.1–17.9) | 2.0 (0.5–6.8) | 0.4 (0.1–0.9) | n.m. | 7.7 (0.6–28.0) | 1.2 (0.1–4.6) | n.m. |
All measurements were by tandem mass spectrometry and are shown as nmol/L.
Median (5th–95th centile range).
Mean ± SEM.
Median (IQR).
Median (range).
Abbreviations: n.m., not measured; A4, androstenedione; T, testosterone; 11OHA4, 11β-hydroxyandrostenedione; 11KA4, 11-ketoandrostenedione; 11KT, 11-ketotestosterone; 11OHT, 11β-hydroxytestosterone.
Circulating concentrations of classic and 11-oxygenated androgens in healthy women as reported in selected studies.
| DHEA | A4 | T | 11OHA4 | 11KA4 | 11KT | 11OHT | ||
|---|---|---|---|---|---|---|---|---|
| Schiffer et al. 202325 | 84 women aged <50a | 9.6 (4.2–20.0) | 2.3 (.7–5.3) | 0.61 (<0.4–1.2) | 7.5 (3.4–14.0) | 3.2 (1.6–5.4) | 0.70 (<0.3–1.6) | 0.39 (<0.3–0.8) |
| 81 women aged ≥50a | 5.0 (1.3–16.0) | 1.1 (<0.7–2.9) | 0.45 (<0.4–1.2) | 8.4 (3.9–16.0) | 1.8 (0.6–4.2) | 0.94 (0.4–1.7) | 0.63 (<0.3–1.2) | |
| Caron et al. 202126 | 10 women, follicular phase, aged 20–40 yearsb | 11.2 ± 0.8 | 2.8 ± 0.1 | 0.6 ± 0.03 | 5.2 ± 0.4 | 0.8 ± 0.1 | 0.9 ± 0.1 | 0.34 ± 0.04 |
| 10 women, luteal phase, aged 20–40 yearsb | 9.4 ± 0.5 | 3.8 ± 0.1 | 0.6 ± 0.03 | 4.6 ± 0.4 | 0.8 ± 0.1 | 0.9 ± 0.1 | 0.3 ± 0.04 | |
| 10 postmenopausal women aged 51–70 yearsb | 14.5 ± 0.5 | 3.0 ± 0.2 | 1.1 ± 0.1 | 11.5 ± 0.6 | 1.3 ± 0.03 | 1.8 ± 0.1 | 0.8 ± 0.03 | |
| Davio et al. 202024 | 72 women aged 18–39c | n.m. | 3.1 (2.3–4.1) | 1.1 (0.8–1.4) | 3.9 (2.8–5.0) | 0.5 (0.4–0.7) | 0.7 (0.5–1.1) | 0.4 (0.2–0.5) |
| 78 women aged 40–59c | n.m. | 1.8 (1.4–2.3) | 0.8 (0.7–1.1) | 4.0 (2.8–5.7) | 0.5 (0.4–0.7) | 0.9 (0.6–1.0) | 0.4 (0.3–0.6) | |
| 81 women aged 60–79c | n.m. | 1.2 (0.8–1.5) | 0.7 (0.5–1.2) | 4.1 (3.4–6.1) | 0.5 (0.4–0.7) | 0.9 (0.6–1.2) | 0.5 (0.4–0.8) | |
| 40 women aged ≥ 8 °c | n.m. | 0.9 (0.6–1.1) | 0.7 (0.5–1.0) | 4.2 (2.9–5.7) | 0.4 (0.3–0.6) | 0.7 (0.5–1.0) | 0.4 (0.3–0.6) | |
| Nanba et al. 201923 | 100 premenopausal women aged 20–40c | 8.0 (5.4–13.9) | 3.5 (2.6–4.9) | 1.0 (0.8–1.4) | 5.7 (3.9–8.7) | 1.2 (0.9–1.8) | 0.9 (0.6–1.3) | 0.5 (0.3–0.8) |
| 100 postmenopausal women aged ≥ 60c | 2.7 (1.8–4.6) | 1.8 (0.8–1.8) | 0.7 (0.5–1.1) | 6.5 (4.7–9.6) | 1.1 (0.8–1.5) | 0.9 (0.7–1.2) | 0.7 (0.4–0.9) | |
| Skiba et al. 201927 | 163 women, follicular phase, aged 18–40d | 5.3 (1.0–21.3) | 1.8 (0.6–5.4) | 0.3 (0.04–0.9) | n.m. | 8.8 (0.1–28.8) | 1.3 (0.3–7.6) | n.m. |
| 184 women, mid-cycle, aged 18–40d | 4.9 (0.4–23.5) | 2.1 (0.6–7.9) | 0.4 (0.1–1.0) | n.m. | 7.7 (1.1–31.7) | 1.3 (0.03–5.8) | n.m. | |
| 241; women, luteal phase, aged 18–40d | 4.8 (0.1–17.9) | 2.0 (0.5–6.8) | 0.4 (0.1–0.9) | n.m. | 7.7 (0.6–28.0) | 1.2 (0.1–4.6) | n.m. |
| DHEA | A4 | T | 11OHA4 | 11KA4 | 11KT | 11OHT | ||
|---|---|---|---|---|---|---|---|---|
| Schiffer et al. 202325 | 84 women aged <50a | 9.6 (4.2–20.0) | 2.3 (.7–5.3) | 0.61 (<0.4–1.2) | 7.5 (3.4–14.0) | 3.2 (1.6–5.4) | 0.70 (<0.3–1.6) | 0.39 (<0.3–0.8) |
| 81 women aged ≥50a | 5.0 (1.3–16.0) | 1.1 (<0.7–2.9) | 0.45 (<0.4–1.2) | 8.4 (3.9–16.0) | 1.8 (0.6–4.2) | 0.94 (0.4–1.7) | 0.63 (<0.3–1.2) | |
| Caron et al. 202126 | 10 women, follicular phase, aged 20–40 yearsb | 11.2 ± 0.8 | 2.8 ± 0.1 | 0.6 ± 0.03 | 5.2 ± 0.4 | 0.8 ± 0.1 | 0.9 ± 0.1 | 0.34 ± 0.04 |
| 10 women, luteal phase, aged 20–40 yearsb | 9.4 ± 0.5 | 3.8 ± 0.1 | 0.6 ± 0.03 | 4.6 ± 0.4 | 0.8 ± 0.1 | 0.9 ± 0.1 | 0.3 ± 0.04 | |
| 10 postmenopausal women aged 51–70 yearsb | 14.5 ± 0.5 | 3.0 ± 0.2 | 1.1 ± 0.1 | 11.5 ± 0.6 | 1.3 ± 0.03 | 1.8 ± 0.1 | 0.8 ± 0.03 | |
| Davio et al. 202024 | 72 women aged 18–39c | n.m. | 3.1 (2.3–4.1) | 1.1 (0.8–1.4) | 3.9 (2.8–5.0) | 0.5 (0.4–0.7) | 0.7 (0.5–1.1) | 0.4 (0.2–0.5) |
| 78 women aged 40–59c | n.m. | 1.8 (1.4–2.3) | 0.8 (0.7–1.1) | 4.0 (2.8–5.7) | 0.5 (0.4–0.7) | 0.9 (0.6–1.0) | 0.4 (0.3–0.6) | |
| 81 women aged 60–79c | n.m. | 1.2 (0.8–1.5) | 0.7 (0.5–1.2) | 4.1 (3.4–6.1) | 0.5 (0.4–0.7) | 0.9 (0.6–1.2) | 0.5 (0.4–0.8) | |
| 40 women aged ≥ 8 °c | n.m. | 0.9 (0.6–1.1) | 0.7 (0.5–1.0) | 4.2 (2.9–5.7) | 0.4 (0.3–0.6) | 0.7 (0.5–1.0) | 0.4 (0.3–0.6) | |
| Nanba et al. 201923 | 100 premenopausal women aged 20–40c | 8.0 (5.4–13.9) | 3.5 (2.6–4.9) | 1.0 (0.8–1.4) | 5.7 (3.9–8.7) | 1.2 (0.9–1.8) | 0.9 (0.6–1.3) | 0.5 (0.3–0.8) |
| 100 postmenopausal women aged ≥ 60c | 2.7 (1.8–4.6) | 1.8 (0.8–1.8) | 0.7 (0.5–1.1) | 6.5 (4.7–9.6) | 1.1 (0.8–1.5) | 0.9 (0.7–1.2) | 0.7 (0.4–0.9) | |
| Skiba et al. 201927 | 163 women, follicular phase, aged 18–40d | 5.3 (1.0–21.3) | 1.8 (0.6–5.4) | 0.3 (0.04–0.9) | n.m. | 8.8 (0.1–28.8) | 1.3 (0.3–7.6) | n.m. |
| 184 women, mid-cycle, aged 18–40d | 4.9 (0.4–23.5) | 2.1 (0.6–7.9) | 0.4 (0.1–1.0) | n.m. | 7.7 (1.1–31.7) | 1.3 (0.03–5.8) | n.m. | |
| 241; women, luteal phase, aged 18–40d | 4.8 (0.1–17.9) | 2.0 (0.5–6.8) | 0.4 (0.1–0.9) | n.m. | 7.7 (0.6–28.0) | 1.2 (0.1–4.6) | n.m. |
All measurements were by tandem mass spectrometry and are shown as nmol/L.
Median (5th–95th centile range).
Mean ± SEM.
Median (IQR).
Median (range).
Abbreviations: n.m., not measured; A4, androstenedione; T, testosterone; 11OHA4, 11β-hydroxyandrostenedione; 11KA4, 11-ketoandrostenedione; 11KT, 11-ketotestosterone; 11OHT, 11β-hydroxytestosterone.
Circulating concentrations of classic and 11-oxygenated androgens in healthy men as reported in selected studies.
| DHEA | A4 | T | 11OHA4 | 11KA4 | 11KT | 11OHT | ||
|---|---|---|---|---|---|---|---|---|
| Schiffer et al. 202325 | 49 men aged < 50a | 8.1 (2.5–27.0) | 1.5 (<0.7–5.0) | 13.0 (8.3–19.0) | 7.6 (2.9–14.0) | 3.5 (1.4–6.2) | 0.9 (<0.3–2.2) | 0.5 (<0.3–1.1) |
| 76 men aged ≥50a | 5.0 (1.5–12.0) | 1.7 (<0.7–3.2) | 12.0 (6.5–22.0) | 8.5 (4.6–16.0) | 2.1 (0.6–5.2) | 0.8 (0.36–1.8) | 0.7 (<0.3–1.5) | |
| Caron et al. 202126 | 9 men aged 21–62 yearsb | 8.2 ± 0.4 | 2.4 ± 0.1 | 18.7 ± 0.6 | 4.4 ± 0.2 | 0.5 ± 0.02 | 1.2 ± 0.03 | 0.6 ± 0.04 |
| 10 men aged 60–72 yearsb | 4.9 ± 0.2 | 2.4 ± 0.1 | 19.6 ± 0.8 | 6.8 ± 0.3 | 0.8 ± 0.04 | 1.5 ± 0.07 | 0.5 ± 0.02 | |
| Turcu et al. 202128 | 10 men aged 19–29c (circadian peak)c | n.m. | 2.7 (2.7–2.8) | 13.5 (13.0–14.0) | 8.6 (8.2–9.0) | 2.1 (2.0–2.1) | 1.4 (1.4–1.4) | 0.6 (0.6–0.6) |
| 10 men aged 19–29c (circadian nadir)c | n.m. | 1.1 (1.0–1.2) | 9.4 (8.7–10.1) | 0.7 (0.4–1.1) | 0.4 (0.3–0.5) | 0.3 (0.3–0.3) | 0.04 (0.01–0.06) | |
| 10 men aged 61–75c (circadian peak)c | n.m. | 2.1 (2.0–2.1) | 12.7 (12.2–13.3) | 7.3 (6.9–7.7) | 1.5 (1.4–1.6) | 1.0 (1.0–1.0) | 0.5 (0.5–0.5) | |
| 10 men aged 61–75c (circadian nadir)c | n.m. | 1.0 (0.9–1.0) | 9.6 (8.9–10.2) | 1.7 (1.4–2.1) | 0.6 (0.5–0.7) | 0.4 (0.4–0.4) | 0.2 (0.2–0.2) | |
| Davio et al. 202024 | 69 men aged 18–39c | n.m. | 1.8 (1.4–2.1) | 16.4 (11.8–20.8) | 4.1 (3.1–5.6) | 0.6 (0.4–0.7) | 1.0 (0.7–1.4) | 0.4 (0.4–0.6) |
| 99 men aged 40–59c | n.m. | 1.7 (1.3–2.2) | 15.0 (11.7–20.0) | 4.9 (3.6–6.8) | 0.6 (0.5–0.8) | 1.0 (0.6–1.3) | 0.5 (0.4–0.8) | |
| 108 men aged 60–79c | n.m. | 1.4 (1.1–1.8) | 16.2 (11.6–20.6) | 4.7 (3.6–5.9) | 0.5 (0.4–0.7) | 0.8 (0.6–1.1) | 0.5 (0.4–0.7) | |
| 43 men aged ≥ 80c | n.m. | 1.1 (0.9–1.3) | 10.6 (6.2–15.7) | 4.7 (3.0–6.4) | 0.4 (0.3–0.7) | 0.6 (0.4–1.1) | 0.5 (0.3–0.7) |
| DHEA | A4 | T | 11OHA4 | 11KA4 | 11KT | 11OHT | ||
|---|---|---|---|---|---|---|---|---|
| Schiffer et al. 202325 | 49 men aged < 50a | 8.1 (2.5–27.0) | 1.5 (<0.7–5.0) | 13.0 (8.3–19.0) | 7.6 (2.9–14.0) | 3.5 (1.4–6.2) | 0.9 (<0.3–2.2) | 0.5 (<0.3–1.1) |
| 76 men aged ≥50a | 5.0 (1.5–12.0) | 1.7 (<0.7–3.2) | 12.0 (6.5–22.0) | 8.5 (4.6–16.0) | 2.1 (0.6–5.2) | 0.8 (0.36–1.8) | 0.7 (<0.3–1.5) | |
| Caron et al. 202126 | 9 men aged 21–62 yearsb | 8.2 ± 0.4 | 2.4 ± 0.1 | 18.7 ± 0.6 | 4.4 ± 0.2 | 0.5 ± 0.02 | 1.2 ± 0.03 | 0.6 ± 0.04 |
| 10 men aged 60–72 yearsb | 4.9 ± 0.2 | 2.4 ± 0.1 | 19.6 ± 0.8 | 6.8 ± 0.3 | 0.8 ± 0.04 | 1.5 ± 0.07 | 0.5 ± 0.02 | |
| Turcu et al. 202128 | 10 men aged 19–29c (circadian peak)c | n.m. | 2.7 (2.7–2.8) | 13.5 (13.0–14.0) | 8.6 (8.2–9.0) | 2.1 (2.0–2.1) | 1.4 (1.4–1.4) | 0.6 (0.6–0.6) |
| 10 men aged 19–29c (circadian nadir)c | n.m. | 1.1 (1.0–1.2) | 9.4 (8.7–10.1) | 0.7 (0.4–1.1) | 0.4 (0.3–0.5) | 0.3 (0.3–0.3) | 0.04 (0.01–0.06) | |
| 10 men aged 61–75c (circadian peak)c | n.m. | 2.1 (2.0–2.1) | 12.7 (12.2–13.3) | 7.3 (6.9–7.7) | 1.5 (1.4–1.6) | 1.0 (1.0–1.0) | 0.5 (0.5–0.5) | |
| 10 men aged 61–75c (circadian nadir)c | n.m. | 1.0 (0.9–1.0) | 9.6 (8.9–10.2) | 1.7 (1.4–2.1) | 0.6 (0.5–0.7) | 0.4 (0.4–0.4) | 0.2 (0.2–0.2) | |
| Davio et al. 202024 | 69 men aged 18–39c | n.m. | 1.8 (1.4–2.1) | 16.4 (11.8–20.8) | 4.1 (3.1–5.6) | 0.6 (0.4–0.7) | 1.0 (0.7–1.4) | 0.4 (0.4–0.6) |
| 99 men aged 40–59c | n.m. | 1.7 (1.3–2.2) | 15.0 (11.7–20.0) | 4.9 (3.6–6.8) | 0.6 (0.5–0.8) | 1.0 (0.6–1.3) | 0.5 (0.4–0.8) | |
| 108 men aged 60–79c | n.m. | 1.4 (1.1–1.8) | 16.2 (11.6–20.6) | 4.7 (3.6–5.9) | 0.5 (0.4–0.7) | 0.8 (0.6–1.1) | 0.5 (0.4–0.7) | |
| 43 men aged ≥ 80c | n.m. | 1.1 (0.9–1.3) | 10.6 (6.2–15.7) | 4.7 (3.0–6.4) | 0.4 (0.3–0.7) | 0.6 (0.4–1.1) | 0.5 (0.3–0.7) |
All measurements were by tandem mass spectrometry and are shown as nmol/L.
Median (5th–95th centile range).
Mean ± SEM.
Median (IQR).
Abbreviations: n.m., not measured; A4, androstenedione; T, testosterone; 11OHA4, 11β-hydroxyandrostenedione; 11KA4, 11-ketoandrostenedione; 11KT, 11-ketotestosterone; 11OHT, 11β-hydroxytestosterone.
Circulating concentrations of classic and 11-oxygenated androgens in healthy men as reported in selected studies.
| DHEA | A4 | T | 11OHA4 | 11KA4 | 11KT | 11OHT | ||
|---|---|---|---|---|---|---|---|---|
| Schiffer et al. 202325 | 49 men aged < 50a | 8.1 (2.5–27.0) | 1.5 (<0.7–5.0) | 13.0 (8.3–19.0) | 7.6 (2.9–14.0) | 3.5 (1.4–6.2) | 0.9 (<0.3–2.2) | 0.5 (<0.3–1.1) |
| 76 men aged ≥50a | 5.0 (1.5–12.0) | 1.7 (<0.7–3.2) | 12.0 (6.5–22.0) | 8.5 (4.6–16.0) | 2.1 (0.6–5.2) | 0.8 (0.36–1.8) | 0.7 (<0.3–1.5) | |
| Caron et al. 202126 | 9 men aged 21–62 yearsb | 8.2 ± 0.4 | 2.4 ± 0.1 | 18.7 ± 0.6 | 4.4 ± 0.2 | 0.5 ± 0.02 | 1.2 ± 0.03 | 0.6 ± 0.04 |
| 10 men aged 60–72 yearsb | 4.9 ± 0.2 | 2.4 ± 0.1 | 19.6 ± 0.8 | 6.8 ± 0.3 | 0.8 ± 0.04 | 1.5 ± 0.07 | 0.5 ± 0.02 | |
| Turcu et al. 202128 | 10 men aged 19–29c (circadian peak)c | n.m. | 2.7 (2.7–2.8) | 13.5 (13.0–14.0) | 8.6 (8.2–9.0) | 2.1 (2.0–2.1) | 1.4 (1.4–1.4) | 0.6 (0.6–0.6) |
| 10 men aged 19–29c (circadian nadir)c | n.m. | 1.1 (1.0–1.2) | 9.4 (8.7–10.1) | 0.7 (0.4–1.1) | 0.4 (0.3–0.5) | 0.3 (0.3–0.3) | 0.04 (0.01–0.06) | |
| 10 men aged 61–75c (circadian peak)c | n.m. | 2.1 (2.0–2.1) | 12.7 (12.2–13.3) | 7.3 (6.9–7.7) | 1.5 (1.4–1.6) | 1.0 (1.0–1.0) | 0.5 (0.5–0.5) | |
| 10 men aged 61–75c (circadian nadir)c | n.m. | 1.0 (0.9–1.0) | 9.6 (8.9–10.2) | 1.7 (1.4–2.1) | 0.6 (0.5–0.7) | 0.4 (0.4–0.4) | 0.2 (0.2–0.2) | |
| Davio et al. 202024 | 69 men aged 18–39c | n.m. | 1.8 (1.4–2.1) | 16.4 (11.8–20.8) | 4.1 (3.1–5.6) | 0.6 (0.4–0.7) | 1.0 (0.7–1.4) | 0.4 (0.4–0.6) |
| 99 men aged 40–59c | n.m. | 1.7 (1.3–2.2) | 15.0 (11.7–20.0) | 4.9 (3.6–6.8) | 0.6 (0.5–0.8) | 1.0 (0.6–1.3) | 0.5 (0.4–0.8) | |
| 108 men aged 60–79c | n.m. | 1.4 (1.1–1.8) | 16.2 (11.6–20.6) | 4.7 (3.6–5.9) | 0.5 (0.4–0.7) | 0.8 (0.6–1.1) | 0.5 (0.4–0.7) | |
| 43 men aged ≥ 80c | n.m. | 1.1 (0.9–1.3) | 10.6 (6.2–15.7) | 4.7 (3.0–6.4) | 0.4 (0.3–0.7) | 0.6 (0.4–1.1) | 0.5 (0.3–0.7) |
| DHEA | A4 | T | 11OHA4 | 11KA4 | 11KT | 11OHT | ||
|---|---|---|---|---|---|---|---|---|
| Schiffer et al. 202325 | 49 men aged < 50a | 8.1 (2.5–27.0) | 1.5 (<0.7–5.0) | 13.0 (8.3–19.0) | 7.6 (2.9–14.0) | 3.5 (1.4–6.2) | 0.9 (<0.3–2.2) | 0.5 (<0.3–1.1) |
| 76 men aged ≥50a | 5.0 (1.5–12.0) | 1.7 (<0.7–3.2) | 12.0 (6.5–22.0) | 8.5 (4.6–16.0) | 2.1 (0.6–5.2) | 0.8 (0.36–1.8) | 0.7 (<0.3–1.5) | |
| Caron et al. 202126 | 9 men aged 21–62 yearsb | 8.2 ± 0.4 | 2.4 ± 0.1 | 18.7 ± 0.6 | 4.4 ± 0.2 | 0.5 ± 0.02 | 1.2 ± 0.03 | 0.6 ± 0.04 |
| 10 men aged 60–72 yearsb | 4.9 ± 0.2 | 2.4 ± 0.1 | 19.6 ± 0.8 | 6.8 ± 0.3 | 0.8 ± 0.04 | 1.5 ± 0.07 | 0.5 ± 0.02 | |
| Turcu et al. 202128 | 10 men aged 19–29c (circadian peak)c | n.m. | 2.7 (2.7–2.8) | 13.5 (13.0–14.0) | 8.6 (8.2–9.0) | 2.1 (2.0–2.1) | 1.4 (1.4–1.4) | 0.6 (0.6–0.6) |
| 10 men aged 19–29c (circadian nadir)c | n.m. | 1.1 (1.0–1.2) | 9.4 (8.7–10.1) | 0.7 (0.4–1.1) | 0.4 (0.3–0.5) | 0.3 (0.3–0.3) | 0.04 (0.01–0.06) | |
| 10 men aged 61–75c (circadian peak)c | n.m. | 2.1 (2.0–2.1) | 12.7 (12.2–13.3) | 7.3 (6.9–7.7) | 1.5 (1.4–1.6) | 1.0 (1.0–1.0) | 0.5 (0.5–0.5) | |
| 10 men aged 61–75c (circadian nadir)c | n.m. | 1.0 (0.9–1.0) | 9.6 (8.9–10.2) | 1.7 (1.4–2.1) | 0.6 (0.5–0.7) | 0.4 (0.4–0.4) | 0.2 (0.2–0.2) | |
| Davio et al. 202024 | 69 men aged 18–39c | n.m. | 1.8 (1.4–2.1) | 16.4 (11.8–20.8) | 4.1 (3.1–5.6) | 0.6 (0.4–0.7) | 1.0 (0.7–1.4) | 0.4 (0.4–0.6) |
| 99 men aged 40–59c | n.m. | 1.7 (1.3–2.2) | 15.0 (11.7–20.0) | 4.9 (3.6–6.8) | 0.6 (0.5–0.8) | 1.0 (0.6–1.3) | 0.5 (0.4–0.8) | |
| 108 men aged 60–79c | n.m. | 1.4 (1.1–1.8) | 16.2 (11.6–20.6) | 4.7 (3.6–5.9) | 0.5 (0.4–0.7) | 0.8 (0.6–1.1) | 0.5 (0.4–0.7) | |
| 43 men aged ≥ 80c | n.m. | 1.1 (0.9–1.3) | 10.6 (6.2–15.7) | 4.7 (3.0–6.4) | 0.4 (0.3–0.7) | 0.6 (0.4–1.1) | 0.5 (0.3–0.7) |
All measurements were by tandem mass spectrometry and are shown as nmol/L.
Median (5th–95th centile range).
Mean ± SEM.
Median (IQR).
Abbreviations: n.m., not measured; A4, androstenedione; T, testosterone; 11OHA4, 11β-hydroxyandrostenedione; 11KA4, 11-ketoandrostenedione; 11KT, 11-ketotestosterone; 11OHT, 11β-hydroxytestosterone.
The least abundant of the main four 11-oxygenated androgens is 11OHT, which circulates at levels <1 nM. Circulating levels of the potent 5α-reduced form of 11KT, namely 11-keto-5α-dihydrotestosterone (11KDHT), have been reported to be 20- to 89-fold lower than those of 5α-dihydrotestosterone (DHT), which may reflect the relatively inefficient 5α-reduction of 11KT by steroid 5α-reductase type 1 (SRD5A1).19,26,29 Importantly, due to their adrenal origin, 11-oxygenated androgens are not subject to cyclic changes during the course of the menstrual cycle, and do not appear to be affected by hormonal contraceptives.25,27 They are however subject to circadian rhythm with morning peaks and nocturnal nadirs comparable to what is observed for cortisol.25,28,30,31
Bioactivity
Several studies have confirmed that 11KT is a potent androgen with an androgenic activity equal to or similar to that of testosterone.9,21,32-36 11KDHT is even more potent with an androgenic activity similar to that of the most potent natural androgen, DHT.21,34 Currently the abundance and contribution of 11KDHT to the androgen pool is unclear as circulating levels are substantially lower than that of DHT.26,29 One preliminary study by Laforest et al.37 did however report 2-fold higher 11KDHT levels than DHT in adipose tissue from individuals with severe obesity, thus demonstrating that 11KDHT may be formed locally in target tissues, with levels not reflected in circulation.
Although initially reported as a partial agonist of the androgen receptor9,17,32,33 recent studies have questioned the androgenic activity of 11OHT.34,35 Variations in reported potencies appear to be due to the test system employed with the endogenous expression of HSD11B2 in several mammalian cell lines potentially leading to the formation of 11KT and therefore an overestimation of 11OHT activity in early studies. Nonetheless 11OHT does appear to be a partial agonist,34 though its physiological relevance remains to be determined. It is however clear that the precursors, 11OHA4 and 11KA4, have little to no androgenic activity.17,34,35 The prevailing picture is therefore that 11KT is the primary androgenic 11-oxygenated androgen in circulation with potential intracrine contribution from 11KDHT.
11-oxygenated androgens and body mass index
Given that AKR1C3 efficiently catalyzes the activation of 11KA4 to 11KT and that the expression levels of AKR1C3 in subcutaneous adipose tissue correlate with body mass index (BMI) and are reduced following weight loss; it stands to reason that there may be an association between circulating 11KT levels and BMI.18-20 Skiba et al.27 reported an inverse correlation of 11KA4 and BMI in women, which supports an increased utilization of 11KA4 by AKR1C3 with increasing BMI. Circulating levels of 11KT were not, however, associated with BMI, while 11OHT was not measured. Although AKR1C3 does not catalyze the conversion of 11OHA4 to 11OHT directly, some 11KT is converted to 11OHT due to the co-expression of AKR1C3 and HSD11B1 in adipose tissue.19,20 11OHT levels may therefore also be reflective of peripheral AKR1C3 expression and activity, although low levels of 11OHT are derived directly from the adrenal cortex.15 Indeed, Davio et al.24 reported a positive correlation between 11OHT and BMI in their cohort of women in addition to an inverse correlation between 11KA4 levels and BMI. Interestingly, in men, only 11KT levels were positively correlated to BMI. Nonetheless, both findings point toward increased AKR1C3 activity in obese individuals but also suggest underlying sex specific differences. More recently, Schiffer et al.25 reported that both 11KT and 11OHT tended to increase with BMI in their cohort of women, while 11OHA4 and 11KA4 levels were increased together with those of 11KT and 11OHT in men. Taken together, there is increasing evidence that higher BMI is associated with increased peripheral activation of 11-oxygenated androgens, but that there may also be divergent sex-specific effects. Given that both 11KT and 11OHT can be considered products of AKR1C3, it may be worth future studies considering the sum of 11KT and 11OHT or considering the ratio of 11KT/11KA4 or (11KT + 11OHT)/11KA4 as a marker of AKR1C3 activity.
11-oxygenated androgens and insulin resistance
In vitro studies have shown that AKR1C3 expression in adipocytes is stimulated by insulin.20,38,39 It follows that the higher expression levels of AKR1C3 in the subcutaneous adipose tissue of individuals with obesity may at least be in part be associated with the higher risk of insulin resistance and hyperinsulinemia during obesity.18,40 Total AKR1C3 and androgen activation may therefore be higher in individuals with obesity due to an increase in the total amount of adipose expressing AKR1C3 in combination with higher concentrations of AKR1C3 in the adipose tissue. In support of this, Yoshida et al.41 noted significantly higher levels of 11KT and 11OHT in women with obesity and polycystic ovary syndrome (PCOS) when compared to normal weighted women with PCOS. The ratio of 11KT/11KA4 in overweight women with PCOS was significantly higher than in the normal weight women with PCOS or healthy control groups, indicating enhanced AKR1C3 expression that may have been driven by insulin resistance in these women. We have previously proposed that AKR1C3-mediated androgen generation in adipose stimulates de novo lipogenesis, lipid accumulation and fatty acid overspill, which in turn results in systemic insulin resistance and hyperinsulinemia that further stimulates AKR1C3 expression, thus creating a vicious circle of hyperandrogenism and hyperinsulinemia38 (Figure 2). This was supported in the same human in vivo study by nontargeted metabolomic profiling; serum samples from women with PCOS revealed a prolipogenic signature 150 minutes after oral DHEA administration. Therefore, given that 11KA4 is a preferred substrate for AKR1C3 and that AKR1C3-mediated androgen activation has been linked to androgen-driven metabolic dysfunction in women, it is probable that 11KT makes an important contribution to the hyperandrogenism in these women.
Schematic representation of the proposed vicious circle between insulin resistance and hyperandrogenism. AKR1C3 catalyzes the conversion of androstenedione and 11-ketoandrostenedione (11KA4) to the potent androgens testosterone and 11-ketotestosterone (11KT) in adipose tissue. Androgens activate the androgen receptor leading to an increase in de novo lipogenesis and a decrease in β-oxidation and lipolysis, cumulatively leading to lipid accumulation. Fatty acid overspill in turn leads to insulin resistance and hyperinsulinemia. Increased insulin levels (1) putatively stimulate adrenal biosynthesis of androgen precursors as proposed by Walzer et al.42 and (2) stimulate the increased expression and activity of aldo-keto reductase 1C3 (AKR1C3), with both mechanisms leading to local androgen activation, thereby completing the vicious circle as originally proposed by O’Reilly et al.38 Androgen precursors are shown in light blue, while active androgens are shown in dark blue.
In addition to the potential role played by insulin in regulating AKR1C3 expression and thus the activation of 11-oxygenated androgens, Walzer et al.42 have shown that excess insulin directly stimulates 11-oxygenated androgen biosynthesis in the adrenal cortex, likely through adrenal insulin receptor (INSR) signaling promoting the upregulation of adrenal steroidogenic enzymes such as CYP11B1. Walzer et al.42 proposed that insulin-induced upregulation of adrenal 11-oxygenated androgen biosynthesis provides a mechanistic explanation for the increased 11-oxygenated androgen levels observed in conditions of mild insulin resistance, such as in PCOS.22 Intriguingly, Walzer et al. demonstrated that patients with monogenic loss of function mutations of the INSR did not have 11-oxygenated androgen excess, providing strong evidence that intact INSR signaling is required to drive adrenal hyperandrogenemia. A clear picture is therefore emerging that insulin and in particular hyperinsulinemia has the potential to drive both adrenal 11-oxygenated androgen biosynthesis through activation of adrenal INSR signaling, as well as subsequent AKR1C3 mediated peripheral activation (Figure 2).
Aromatization to estrogens
The classic androgens are obligatory precursors for the cytochrome P450 aromatase (CYP19A1) mediated biosynthesis of estrogens. Conversely, while 11-oxygenated androgens can be aromatized in vitro, they are poor substrates for CYP19A1, with the result that 11-oxygenated estrogens have not been detected in healthy individuals.43 This suggest that unlike the classic androgens, 11-oxygenated androgens may function exclusively as androgens in vivo, though intracrine biosynthesis of 11-oxygenated estrogens cannot yet be ruled out. One study reported fetal virilization due to overproduction of 11-oxygenated androgens by a maternal tumor and suggested that a lack of aromatization in the placenta was responsible for elevated exposure of the fetus to 11-oxygenated androgens.44 He et al. later demonstrated that in healthy women, placental HSD17B2 protects the fetus from exposure to maternal 11KT, by converting it to 11KA4.45
Analytical considerations
Serum
Blood samples for quantification of 11-oxygenated androgens in human subjects should be drawn at the same time each day (morning recommended) as 11-oxygenated androgens are subject to circadian rhythm.28,30,31 Samples should also be kept on ice and centrifuged to remove peripheral blood mononuclear cells (PBMCs), which have been shown to catalyze the conversion of 11KA4 to 11KT due to the expression of AKR1C3 and leading to increase levels of 11KT over time in unprocessed samples.46,47 It has also been reported that the glucocorticoids cortisol and cortisone decompose to 11OHA4 and 11KA4 when dry extracts are subjected to temperatures ≥45 °C.24 Care should therefore be taken as to minimize the heat exposure during sample extraction and methods should be thoroughly validated.
Saliva
To date only 11OHA4 and 11KT levels have been measured in saliva.25,30,48,49 Salivary 11OHA4 and 11KT correlate closely with their serum levels and can therefore be a useful noninvasive method of measuring these androgens.25,48 It should however be noted that the salivary glands express high levels of HSD11B2 which convert 11β-hydroxysteroids to their 11-keto counterparts.50 Therefore 11OHA4 levels are expected to be proportionally lower due to conversion to 11KA4, while 11KT levels are proportionally higher due to conversion from 11OHT.25 As with serum, diurnal variation of 11-oxygenated androgen levels have been demonstrated in saliva.25,30
Urine
The metabolites of 11-oxygenated androgens overlap substantially with those of the more abundant glucocorticoids, thus masking their detection in urine.51-54 An exception is 11β-hydroxyandrosterone (11OHAn) a metabolite that has primarily been attributed (≥90%) to 11-oxygenated androgen metabolism in healthy individuals with only a small glucocorticoid contribution (≤10%).51 Currently 11OHAn is therefore the only marker routinely included in urinary analysis.55 Our recent study revealed that 11-ketoandrosterone (11KAn) is a metabolite derived from 11KA4 and 11KT, with no overlap with glucocorticoid metabolism, and should therefore be included as an additional urinary marker of 11-oxygenated androgens.53
11-oxygenated androgens in disease
Castration resistant prostate cancer
Prostate cancer (PCa) is an androgen dependent disease and as such the primary treatment of metastatic and advanced PCa is androgen deprivation therapy, achieved by chemical or physical castration. Despite initially being effective in many cases the cancer frequently recurs and is then termed castration resistant prostate cancer (CRPC).56 A characteristic of CRPC is that the expression profiles of key steroid metabolizing enzymes are altered in such a way that the cancer is able to generate potent androgens from adrenal androgen precursors, which remain in circulation following ADT.12,56-61 It was within this context that much of the work delineating the peripheral conversion of 11OHA4 to the potent androgens 11KT and 11KDHT was carried out.17,62 After showing that PCa cell lines had the ability to convert 11OHA4 to 11KT and 11KDHT,17 we showed that 11KT and 11KDHT bind to and activate the human androgen receptor with affinities and potencies similar to that of testosterone and DHT, respectively.21 We also showed that 11KT and 11KDHT induce androgen-dependent gene expression and stimulate androgen-dependent cell growth in PCa cell lines, thereby confirming their status as potent androgens.21 Interestingly, Snaterse et al.34 recently showed that mutations in the AR, as occur in up to 20% of CRPC cases, modulate the AR's sensitivity toward 11-oxygenated androgens.
We went on to show that the keto forms of 11-oxygenated androgens are the preferred substrate for AKR1C3, which is upregulated in CRPC,19 and that potent 11-oxygenated androgens are inactivated by glucuronidation at a lower rate than classic androgens in PCa cell lines.21,63 It is therefore feasible that 11KT has the potential to accumulate in CRPC tissue at higher levels than classic androgens. To our knowledge, only a single study to date has measure 11-oxygenated androgens in PCa tissue. Despite a small sample number (n = 2), potent 11-oxygenated androgens were detected at levels equal to or higher than that of the classic androgens thus highlighting the need to include 11-oxygenated androgens when considering the intratumoral androgen pool.63 The presence of 11-oxygenated androgens in PCa tissue is however not surprising given that these androgens are abundant in the circulation of men both before and after castration.12,24,28,64,65 Although the biological significance of 11KT in eugonadal men is unclear, especially when considering the 10- to 20-fold higher circulating concentrations of testosterone, a study by Dahmani et al.65 reported that increased preoperative circulating levels of 11KT were associated with improved metastatic free survival in men with localized PCa, while 11-oxygenated adrenal androgen precursors were associated with progressive disease (n = 1793). Further studies investigating the role of 11-oxygenated androgens in PCa development and disease progression are therefore warranted.
Snaterse et al.12 showed that 11KT is the predominant circulating active androgen in men with CRPC (n = 29) with median concentrations 2.8-fold higher than those of testosterone, thus suggesting a role for 11KT in driving the reactivation of the androgen signaling axis in men with CRPC. Moreover, they confirmed the abundant expression of HSD11B2 and AKR1C3 in tumor biopsies from patients with CRPC, thereby showing that CRPC tissue would not only be exposed to 11KT from circulation, but it would have the ability to produce 11KT from circulating precursors (11OHA4 and 11KA4) as has been shown in PCa cell models.17,62 Wright et al.66 have shown that the CYP17A1 inhibitor Abiraterone, inhibits the biosynthesis of both classic adrenal derived androgens and 11-oxygenated androgens in patients with CRPC, thus confirming how abiraterone is initially effective in the treatment of CRPC.67,68
Interestingly, several studies of late have shown that the gut microbiome is able to convert both endogenous and exogenous steroids to androgenic compounds, thus providing an alternative source of androgens in patients with CRPC.69-72 The group of Jason Ridlon have confirmed early observations that gut bacteria are able to catalyze the side-chain cleavage of cortisol, yielding 11OHA4, thus potentially contributing to the circulating pool of 11-oxygenated androgens that can drive CRPC.72,73
Congenital adrenal hyperplasia
The most common cause of congenital adrenal hyperplasia (CAH) is 21-hydroxylase deficiency (21OHD) where homozygous or compound heterozygous mutations in cytochrome 21-hydroxylase (CYP21A2) impede the production of glucocorticoids, and in many cases mineralocorticoids, leading to the build-up of steroid precursors and subsequent androgen excess driven by high ACTH levels.55 Turcu et al.6 were the first to show that androgen excess in classic 21OHD includes significantly (> 3-fold) increased levels of 11-oxygenated androgens. Notably in female patients (n = 19), testosterone and 11KT were positively correlated due to increased adrenal output of both androgens. However, in male patients (n = 19) testosterone and 11KT were inversely correlated, suggesting that the increase in adrenal derived 11KT suppressed gonadal testosterone biosynthesis, most likely via suppression of pituitary gonadotropin secretion. However, a subsequent study by Auer et al.74 failed to detect any correlation between testosterone and 11KT levels in men (n = 42).
Increased 11-oxygenated androgen biosynthesis by 21OHD adrenals appears to be due to areas between the zona fasciculata and zona reticularis co-expressing HSD3B2 and CYB5A—proteins that are normally expressed in a zone specific manner.6 Notably, increased testosterone biosynthesis in 21OHD adrenals likely result in the direct 11β-hydroxylation of testosterone, yielding higher levels of adrenal derived 11OHT than occurs in healthy individuals where this contribution is minor.9,15
Following the finding that 11-oxygenated androgen levels are significantly increased in patients with 21OHD, several studies have investigated their use as biomarkers for the condition. Turcu et al.75 reported that 11-oxygenated androgen levels positively correlate with total adrenal volume as well as with the presence of testicular adrenal rest tumors (TARTs) in male patients with 21OHD (n = 70). The 11-oxygenated androgens have also been shown to be good candidate biomarkers for the monitoring of disease control in 21OHD.31 Notably, 11-oxygenated androgens can help to assess disease control in the case of discrepancies in the classic serum biomarkers 17α-hydroxyprogesterone (17OHP) and androstenedione, which are not uncommon.76 Moreover, salivary steroid analysis offers a noninvasive means of measuring disease control due to strong correlations between circulating steroid concentrations and those excreted in saliva.49
By retrospectively analyzing the urinary steroid profiles of children with 21OHD on glucocorticoid treatment, Kamrath et al.77 demonstrated that androgen excess in treated children with 21OHD (n = 99) is exclusively due to elevated 11-oxygenated androgens. Interestingly analysis of the urinary steroid profiles in their prepubertal cohort suggested that in addition to the 11β-hydroxylation of androstenedione, 11OHA4 is formed from the CYP17A1 catalyzed 17,20-lyase of 21-deoxycortisol, another biomarker of 21OHD.77 To date one study has indirectly shown that the CYP17A1 is capable of catalyzing the conversion of 21-deoxycortisol to 11OHA4,78 though to our knowledge comprehensive kinetic analyses are lacking. Moreover, it is not yet clear whether this proposed reaction plays a role in the biosynthesis of 11-oxygenated androgens in adults with 21OHD.
Testicular adrenal rest tumors
TARTs are benign intratesticular masses that can occur in male patients with CAH. Typically they are comprised of adrenal tissue which is present microscopically in healthy normal male testicular tissue due to the common embryonic journey of adrenal and gonadal tissue along the urogenital ridge. In 21OHD and forms of CAH, nests of adrenal tissue with the testes may enlarge significantly due to ACTH hypersecretion. Turcu et al.75 reported that 11-oxygenated androgen levels were significantly higher in male 21OHD patients with TART (n = 14), when compared to those without TART (n = 17) suggesting direct biosynthesis of 11-oxygenated androgens in the tumor. Auer et al.,74 however did not detect differences in the levels of 11-oxygenated androgen between men with 21OHD with (n = 11) and without (n = 28) TART. Using spermatic vein sampling, Schröder et al.79 however confirmed that 11OHT and 11KT are produced by TART. Cultured TART cells produced higher levels of testosterone and lower levels of A4 following ACTH stimulation in comparison to cultured adrenal cells, suggesting the availability of substrate for the direct 11β-hydroxylation of testosterone in TARTs. Indeed, TARTs are known to express both adrenal (CYP11B1) and testicular enzymes (HSD17B3), thus facilitating biosynthesis of testosterone and subsequently 11OHT.80 Although, 11KT was elevated in spermatic vein samples, it was not produced by cultured TART cells, thus suggesting that 11OHT may be converted to 11KT by surrounding testicular tissues expressing HSD11B2.
Polycystic ovary syndrome
Androgen excess is a defining feature of PCOS, the most common endocrine condition in women.81 Although originally considered a reproductive disorder due to anovulation or polycystic ovarian morphology on imaging, it is now apparent that PCOS is also a lifelong metabolic disorder, characterized by increased risks of type 2 diabetes, hypertension, cardiovascular disease and nonalcoholic fatty liver disease (NAFLD).82-90 Androgen excess is increasingly regarded as a major player in the etiology of many metabolic and other health complications of PCOS.38,90-94 Biochemical profiling of serum androgens in PCOS has traditionally been limited to serum total testosterone measurement, but our understanding of the androgen excess phenotype of women with PCOS continues to evolve. We previously demonstrated that circulating androstenedione, the precursor to testosterone, was preferentially elevated in many women with PCOS, and that biochemical hyperandrogenism would be missed in up to 25% of women who had measurement of serum testosterone in isolation.91 This finding clearly illustrated the heterogeneous nature of PCOS and that a comprehensive androgen profile to include adrenal and ovarian androgen precursors should be considered in the biochemical work up of these patients.
While it is apparent that both the ovaries and adrenals may contribute to the circulating androgen pool in women with PCOS, relative contributions may vary. Several early studies measured circulating levels of the adrenal 11-oxygenated androgen precursor, 11OHA4, in women with PCOS. However, these studies obtained mixed results, with some showing no change in circulating 11OHA4,5,95 while others observed elevated levels.96,97 In 2017, our group was the first to measure a panel of 11-oxygenated androgens in women with PCOS using liquid chromatography tandem mass spectrometry (LC-MS/MS).22 We found that all four 11-oxygenated androgens measured (11OHA4, 11KA4, 11OHT, and 11KT) were significantly elevated in our PCOS cohort (n = 114) when compared to our age- and BMI-matched healthy controls (n = 49). This mirrored the increase in serum levels we observed for the classic androgens testosterone, androstenedione and DHEA in the PCOS group. Notably, the adrenal androgen precursors androstenedione, 11OHA4 and 11KA4 all correlated with the homeostatic model assessment of insulin resistance (HOMA-IR), a marker of insulin resistance. This suggested that 11-oxygenated androgens may directly contribute to the adverse metabolic phenotype in PCOS, or that their generation may be driven by hyperinsulinemia, as supported by the work of others (Figure 2).38,42 We also observed that 11KT was the predominant active androgen in circulation, with median levels 3-fold higher than that of testosterone in women with PCOS. A limitation of our study was that only increases in median levels of each androgen were reported and inter-individual variations were not described.
A subsequent study by Yoshida et al.41 further highlighted the potential complexity in the androgen excess phenotype in women with PCOS. They raised the possibility of distinct subgroups with different androgenic signatures; in their cohort, they observed subsets of women with increases in only classic (12/28) or 11-oxygenated androgens (4/28). The third group was found to have both elevated classic and 11-oxygenated androgen levels (8/28). The heterogeneity of hyperandrogenism in women with PCOS was again demonstrated by Tosi et al.98 who found that 11OHT and 11KT were elevated in 28.5% and 30.1% of women with PCOS (n = 123). Free testosterone, measured following direct equilibrium dialysis, was elevated in 61.0% of the affected women, while total testosterone, androstenedione, and DHEAS were elevated in 48.0%, 20.3%, and 10.6% of women, respectively. Interestingly, 11KT and 11OHT weakly correlated with improved insulin sensitivity as measured by hyperinsulinemic euglycemic clamps, while free testosterone significantly correlated with reduced insulin sensitivity. A study by Taylor et al.,99 who reported increased 11-oxygenated androgens in a cohort of untreated adolescents with PCOS (n = 115), observed a significant correlation between free testosterone and HOMA-IR, but no association between any of the measured 11-oxygenated androgens and HOMA-IR. Both of these studies are in contrast to our earlier study which found 11OHA4 and 11KA4 to be correlated with HOMA-IR.22 However, 11KA4 was not measured by Tosi et al.98 or Taylor et al.99 with the former also not measuring 11OHA4. Notably, Taylor et al.99 did find a significant correlation between the three measured 11-oxygenated androgen levels and the Ferriman-Gallwey score used to evaluate hirsutism.
Interestingly, the studies by Tosi et al.98 and Taylor et al.99 reported that 11-oxygenated androgens may be inadequate markers of biochemical androgen excess in women with PCOS compared to their classical pathway counterparts. However, this is perhaps not a fair evaluation as the diagnosis of biochemical hyperandrogenism is defined as elevated testosterone; from the study by Yoshida et al.,41 it is clear that there may be cases of PCOS with elevated 11-oxygenated androgens, but normal classic androgens. The Tosi et al. study also found an inverse relationship between circulating 11-oxygenated androgens and markers of insulin resistance on hyperinsulinemic-euglycemic clamp testing, a somewhat unexpected observation. Clearly further research is required to understand these observations and to explain the discrepancies between studies. It is however clear that biochemical androgen excess in women with PCOS is heterogeneous and complex, and larger deeply phenotyped patient cohorts are required.
Hirsutism
Hudson et al.95 were the first to propose that 11OHA4 could be used as a marker to assess the adrenal contribution to hirsutism in 1990. More recently, Skiba et al.100 did not find any associations between 11-oxygenated androgens and self-assessed hirsutism in a cohort of 710 nonhealth-care-seeking women in the community. Similarly, Tosi et al.98 only noted a weak inverse relationship between 11OHT levels and the hirsutism score in women with PCOS, but none with 11KT (n = 123). Conversely, Taylor et al.99 found that 11OHA4, 11OHT, and 11KT all correlated with hirsutism severity in untreated adolescent women diagnosed with PCOS (n = 115). Local SRD5A1-catalyzed conversion of testosterone to the more potent androgen DHT is a contributing factor to androgen driven hirsutism.101-105 We have shown that 11KT and its precursor 11KA4, are not efficiently 5α-reduced by SRD5A1.53 Kinetics for the 5α-reduction of 11OHA4 and 11OHT by SRD5A1 have not yet been published to our knowledge, but unpublished results from our laboratory suggest that while these are better substrates than 11KA4 or 11KT, they are still not comparable to androstenedione or testosterone. Attenuated 5α-reduction of 11-oxygenated androgens by SRD5A1 may therefore account for the weak associations between 11-oxygenated androgen levels and hirsutism in comparison to that observed for testosterone and androstenedione, the preferred substrate for SRD5A1.106 Indeed, Skiba et al.100 found that only androstenedione was significantly associated with modified Ferriman–Gallwey scores when controlling for age and BMI.100
Cushing's syndrome
Cushing's syndrome is a constellation of clinical signs and symptoms driven by chronic exposure to excess glucocorticoids. Cushing's disease (CD) results in glucocorticoid excess specifically as a consequence of ACTH hypersecretion from corticotroph pituitary adenomas.55,107 Patients with CD often present with signs of hyperandrogenism including menstrual irregularities, acne and hirsutism.107 Recently, Nowotny et al.107 demonstrated that the 11-oxygenated androgens 11OHA4 and 11KT were significantly elevated in female patients with CD (n = 23), while the levels of the classical androgens androstenedione and testosterone were comparable to those in a control group (n = 26). This led to the conclusion that hyperandrogenism in CD is predominantly caused by excess 11-oxygenated androgens. Notably, transsphenoidal surgery or treatment with CYP11B1 inhibitors (osilodrostat and metyrapone) successfully reduced the elevated 11-oxygenated androgens levels. These observations highlight the important role of ACTH in 11-oxygenated androgen metabolism.
Premature adrenarche
Adrenarche refers to the development of a functional adrenal zona reticularis and the associated initiation of adrenal C19 steroid biosynthesis, which triggers the initial stages of axillary and/or pubic hair growth (pubarche) and adult body odor. Premature adrenarche is defined as the precocious onset of pubarche before the age of 8 in girls and 9 in boys. Notably, premature adrenarche has been associated with the development of metabolic disturbance, obesity, insulin resistance and PCOS in adulthood.36,108 Rege et al.36 used LC-MS/MS to quantify a panel of C19 steroids in girls with premature adrenarche (n = 37) and age-matched controls (n = 83) and found that both 11KT and testosterone increased significantly during adrenarche. Notably, 11KT levels were 3.5-fold higher than those of testosterone in girls with premature adrenarche, suggesting that 11KT is the predominant active androgen in circulation during premature adrenarche, which may drive the precocious androgenic phenotype. This finding was subsequently supported by Wise-Oringer et al.108 who reported 4-fold higher levels of 11KT than testosterone in their cohort of children with premature adrenarche (n = 11). In addition, they found that 11-oxygenated androgens are elevated during both premature adrenarche and premature pubarche (n = 9) compared to their control group (n = 11), but that there was no differences in the levels of the 11-oxygenated androgens between the two groups.108
Conclusions
The discovery of the 11-oxygenated androgens as role players in human health and disease has prompted a reevaluation of the androgen pool, which is clearly more complex than initially thought. Although 11KT and its precursors circulate at physiologically relevant levels, their physiological role in healthy individuals is not yet clear and requires further study. There is however little doubt that 11-oxygenated androgens are associated with several disease states and that studies measuring androgens can no longer afford to ignore this androgen class, especially in an era where steroid profiling by LC-MS/MS is more readily accessible.
Acknowledgments
The authors with to thank Shona Murray and Imken Oestlund for assisting with the data capture for Tables 1 and 2 as well as for proof reading the manuscript.
Funding
This work is based on the research supported in part by the National Research Foundation of South Africa (Grant Number 132503, to K.H.S.) and the Academy of Medical Sciences (Newton Advanced Fellowship NAF004\1002, to K.H.S.). MWOR is funded by a Health Research Board (HRB) Emerging Clinician Scientist Award (ECSA-2020-001).
Conflicts of interest: The author declares that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
