Testicular Steroidogenesis

  • Christa E. FlückEmail author
  • Amit V. Pandey
Living reference work entry
Part of the Endocrinology book series (ENDOCR)


Testosterone is the major androgen in circulation in male humans, produced primarily in the Leydig cells of the testis. Biosynthesis of testosterone from cholesterol occurs via a series of enzymatic reactions. Testosterone may be further metabolized into a more potent androgen, dihydrotestosterone. In recent years an alternate pathway of dihydrotestosterone biosynthesis without using testosterone as a precursor has emerged. Majority of classically studied effects of androgens are thought to be mediated via nuclear receptor-dependent long-term transcriptional effects, but there also exist membrane receptor-based effects of androgens which are being uncovered from recent studies that may explain rapid effects of androgens in many cases. In this chapter we are describing the biosynthesis, mechanism of action, and therapeutic effects of testosterone and related androgens.


Androgens Anabolic steroids Androgen receptor Testosterone CYP17A1 SRD5A1 Dihydrotestosterone 

Introduction: Testosterone Synthesis Throughout Life

The human testis is a male specific organ, which is key to male sexual differentiation, function, and reproduction. It consists of two distinct functional units comprising of (1) the Leydig cells in the interstitial unit, which are responsible for androgen biosynthesis, and (2) the Sertoli cells of the seminiferous tubules, which are responsible for nourishing spermatogenesis. Hereby, seminiferous tubules make around 90% of the testis volume.

During human development the gonad is first formed as a neutral anlage, but with 46,XY genotype, it is genetically determined to form testes very early in fetal life by around 4–5 weeks of gestation. The testis then produces testosterone (T) and insulin-like 3 (INSL3) in Leydig cells, anti-Müllerian hormone (AMH) in Sertoli cells, and other sex differentiating paracrine factors for normal male sexual differentiation (Fig. 1). While AMH mainly helps to suppress the female Müllerian structures and also helps to maintain and develop male Wolffian duct structures through regulating T production, T and especially its highly potent metabolite 5α-dihydrotestosterone (DHT) virilize the neutral external genitalia by the end of the first trimester. Thus, T production in the male fetus will reach almost adult peak levels by the second trimester of pregnancy in order to promote full masculinization including testicular descent. Only by the third trimester T production is low until the end of pregnancy (Fig. 1).
Fig. 1

Testosterone (T) biosynthesis lifetime curve and associated developmental events. Prenatal period: Sex determination of the fetal gonad occurs at 6 weeks gestation. Then the fetal testis starts to produce T in the Leydig cells for sexual differentiation of the male sex organs, predominantly the external genitalia. Sertoli cells produce anti-Müllerian hormone (AMH) for the regression of the Müllerian ducts and for the differentiation of the Wolffian ducts. T production in the first and second trimesters may reach adult peak levels for fetal masculinization, but decreases during the third trimester of pregnancy. Neonatal period: In the first 72 hours after birth, T levels are high due to the surge of hormones (LH/FSH a.o.) at birth. Similarly, between days 30–100 of life, T levels are high, stimulated by LH during “minipuberty.” Pubertal and adult period: After 6–8 years of age, androgens (predominantly DHEA/S) rise slightly due to adrenarche, the functional activation of the zona reticularis of the adrenal cortex. However, the activation of the hypothalamic-pituitary-gonadal (HPG) axis only occurs after 12 years in boys (mean). At a testis volume of 6–8 ml, T production rises constantly to reach peak adult values between the second and fourth decade of life. With aging T production decreases, especially with chronic diseases influencing the HPG axis

At birth, with a general hormonal surge in the newborn in the first 24–48 hours, T levels are shortly high again before showing low levels until the event of “minipuberty” between days 30–100 of life. This event seems rather specific to boys, is centrally stimulated, and may serve for further (male) sexual development (Kuiri-Hanninen et al. 2014). Clinical observations during minipuberty may include descent of the testes and increase in its volume due to an underlying increase in seminiferous tubules, as well as phallic growth. Biochemically, androgens such as T, DHT, and androsterone rise significantly (Dhayat et al. 2015, 2016). Although the exact function of minipuberty remains unknown, it offers a diagnostic window of opportunity for functional testing of the testis, before it becomes hormonally quiescent until puberty.

At puberty, at around 12 years of age in boys, the hypothalamic-pituitary gonadal axis becomes reactivated. The hypothalamic GnRH pulse generator stimulates gonadotropin secretion (LH/FSH) from the pituitary gland and promotes the pubertal development of the testis (Fig. 2). The testicular volume increases during puberty, first predominantly through increase of the proportion of seminiferous tubules. Leydig cells will resume androgen biosynthesis, which will prompt the development of secondary male sex characteristics like further phallic growth, male-type hair growth, breaking of the voice, male-type growth spurt, and adult height as well as male-type body composition (more muscle and bone mass, less fat mass). The mature adult testis finally reaches a volume of 15–25 ml and produces about 5–7 mg testosterone per day, corresponding to serum concentrations of about 12–41.5 nmol/l (348–1197 ng/dl; measured by HPLC-MS/MS; total testosterone in adults 20–50 years of age. Spermatogenesis nurtured by Sertoli cells is also activated at puberty, and functioning sperms in spontaneously produced ejaculates are detected at a mean age of 13.5 years, long before completion of puberty.
Fig. 2

Schematic diagram of the hypothalamic-pituitary-gonadal axis and of testosterone (T) actions. T is produced upon central GnRH - LH stimulation in the testicular Leydig cell at a daily rate of 5–7 mg. Circulating T (>95% of testis, <5% of adrenal origin) reaches the target organs and exerts its effect via the androgen receptor (AR). 5–10% of T will be amplified to dihydrotestosterone (DHT) by 5α-reductase II (SRD5A2) in the sebaceous skin unit and in the prostate gland. DHT is 5–10 times more potent on the AR than T. By contrast, little amount of T (0.1%) is converted to estrogens by aromatase (CYP19A1) activity in peripheral tissues (e.g., fat, bone, brain, mammary gland). Eventually, all steroids (androgens and estrogens) are metabolized through oxidation and conjugation in the liver and excreted mainly through the kidneys into the urine (>95%). Only few inactive metabolites are excreted into the bile

After puberty, the mature testis plays a pivotal role for allowing full sexual functioning and reproduction. Generally, highest T production is observed in the third and fourth decades of life in males and starts to decrease slowly thereafter (Fig. 1). But sexual function and reproduction may be preserved until death. Aging of the testis function is poorly understood and may rather be secondary to general health problems related to metabolic or cardiovascular disorders (see related chapters).

Thus, given the important role of T in males throughout life, it is clear that abnormalities in androgen biosynthesis cause disorders of sex development (DSD) including puberty, as well as sexual function and reproduction. In the following section, we will summarize the current knowledge on androgen biosynthesis in health and disease, and provide insight into the biochemistry and action of T and some other biologic and synthetic androgens.

Steroidogenesis in the Testis

The biochemistry of T synthesis is a long known process, but the recent discovery of an alternative, so-called backdoor pathway for the production of DHT has brought some novel aspects to the field and evoked numerous unsolved questions.

The Classic Pathway of T Synthesis

T is an androgen , which is produced like all other steroid hormones from cholesterol. In males, more than 95% of circulating T is produced in the testes; only little is produced in the adrenal cortex and in peripheral organs through conversion of precursor steroids. The testicular Leydig cell is highly specialized for T production and expresses all required genes that are common to T-producing organs (Miller and Auchus 2011). Upon stimulation by LH, the Leydig cell enhances its steroid production. In a first step, cholesterol molecules are transported to the inner mitochondrial membrane by the steroidogenic acute regulatory protein (StAR). At the inner mitochondrial membrane, the side chain cleavage system comprised of enzyme CYP11A1 (P450scc) and its redox partners ferredoxin (FDX1) and ferredoxin-reductase (FDXR) convert cholesterol to pregnenolone (Fig. 3). Pregnenolone is then converted through the delta 5 pathway to 17α-hydroxypregnenolone (17OHPreg) and dehydroepiandrosterone (DHEA). Both conversions are supported by the enzyme CYP17A1 (P450c17). While the first reaction requires only the 17α-hydroxylase activity of the enzyme and electron donation by cofactor P450 oxidoreductase (POR), the second reaction requires not only POR support for 17,20-lyase activity, but also cytochrome b5 (CYB5) for facilitating optimal allosteric conformation. DHEA is then turned over step-wise to T through androstenedione or androstenediol catalyzed either first by 3β-hydroxysteroid dehydrogenase type II (HSD3B2/3βHSDII) or 17β-hydroxysteroid dehydrogenase 3 (HSD17B3/17βHSD3/AKR1C3). In the classic view, in genital skin and the prostate, T may be converted to DHT , which has about 10 times more affinity for the androgen receptor. This T to DHT conversion is catalyzed by 5α-reductase type II (SRD5A2/5αRed2). By contrast, very small amounts of androstenedione and T are converted to estrone and estradiol in the testis through aromatase activity (CYP19A1). In humans, only little conversion of pregnenolone to androstenedione occurs in the testis through the delta 4 pathway originating from progesterone (Prog) and 17-hydroxyprogesterone (17OHP) (Fig. 3), because 17,20-lyase activity is poor on the substrate 17OHP compared to 17OHPreg and because HSD3B2 activity is less abundant (Flück et al. 2003). By contrast, rodents produce T predominantly via the delta 4 pathway (Fevold et al. 1989).
Fig. 3

Androgen biosynthesis pathways. The classic testosterone biosynthesis pathway of the Leydig cell is shown in black, the more recently described alternative, backdoor pathway in red. Cholesterol is transported to the inner mitochondrial membrane by steroidogenic factor 1 (StAR). At the inner mitochondrial membrane cholesterol is cleaved to pregnenolone (Preg) by the side-chain cleavage catalytic unit consisting of the enzyme CYP11A1 and its redox partners adrenodoxin/adrenodoxin reductase Preg is then converted by CYP17A1 in two steps to 17-hydroxypregnenolone (17OHPreg) and dehydroepiandrosterone (DHEA), both supported by P450 oxidoreductase, the redox partner of cytochrome P450 proteins in the endoplasmic reticulum, the second step, which is the lyase reaction, also requires cytochrome b5 (CYB5). DHEA is then converted either through the intermediacy of androstenediol or androstenedione to testosterone (T) by the enzymes HSD3B2 and HSD17B3/AKR1C3. Small amounts of T or androstenedione are converted to estrogens in the testis by aromatase (CYP19A1) activity. Recent immunohistochemical findings also suggest that the Leydig cell expresses small amounts of 5α-reductase for the conversion of T to dihydrotestosterone (DHT) (see also Fig. 3). However, this conversion is mainly done by in peripheral tissues such as the skin and the prostate. The Leydig cell expresses only little HSD3B2, but 17-hydroxyprogesterone (17OHP) produced from 17OHPreg or from Preg through progesterone (Prog) may feed into the alternative, backdoor pathway for DHT synthesis. 5α-reductase type I (SRD5A1) is the gate-keeper to the backdoor. It will convert 17OHP to 17-hydroxy-dihydroprogesterone (17OH-DHP), and 3α-hydroxysteroid-dehydrogenase activity of AKR1C2/4 will yield 17-hydroxy-allopregnanlone (17OH-Allo); this is then converted to androsterone by CYP17. Further conversion by AKR1C3/HSD17B3 leads to androstanediol and by AKR1C2/4 or RoDH finally to DHT. Work from polycystic ovary disease and castration-resistant prostate cancer suggests that there is also a short loop to DHT via androstenedione and androstanedione (given in blue)

In comparison, the human adrenal cortex also expresses all enzymes needed for T production in the zona reticularis. However, because the expression of HSD3B2 is low and that of HSD17B5 extremely low, it produces predominantly DHEA and androstenedione as adrenal androgens. In addition, the adrenal cortex also produces mineralocorticoids and glucocorticoids from cholesterol. These steroid hormones regulate water, salt, and glucose homeostasis and are thus essential for life. The initial steps of biosynthesis are common to all steroid hormones. Therefore, genetic defects in genes involved in the initial steps of steroidogenesis shared by the adrenal cortex and the gonads affect steroid production of both organs and may lead to disorders of sexual development and function in both sexes as well as adrenal insufficiency, also known as congenital adrenal hyperplasia (CAH) in the medical literature (Table 1) (Miller and Auchus 2011; Miller and Flück 2014). Importantly, biosynthesis of the androgen precursors DHEA and androstenedione is also crucial for the production of all estrogens. Therefore, steroid biosynthetic defects such as CYP17A1 deficiency or HSD3B2 deficiency may not only lead to androgen deficiency, but also cause estrogen deficiency and thus lack of pubertal development and infertility in 46,XX females.
Table 1

Androgen biosynthetic defects causing disordered sexual development (DSD) and maturation with and without adrenal insufficiency




Adrenal insufficiency

46,XY DSD phenotype

(T deficiency)


Other features

Lipoid congenital adrenal hyperplasia (LCAH)




Classic form: 46,XY DSD, gonadal insufficiency

Nonclassic form: None



P450 side chain cleavage syndrome (CAH)




Classic form: 46,XY DSD, gonadal insufficiency

Nonclassic form: None



3β-hydroxysteroid dehydrogenase II deficiency (CAH)




46,XY DSD, gonadal insufficiency

Nonclassic form: No DSD, but premature adrenarche



Combined 17-hydroxylase, 17,20-lyase deficiency (CAH)




46,XY DSD, gonadal insufficiency


Hypertension and hypokalemic alkalosis (not seen with isolated lyase deficiency)

P450 oxidoreductase deficiency (CAH)





46,XY DSD, gonadal insufficiency


Maternal virilization during pregnancy; Antley-Bixler skeletal malformation syndrome; changes in drug metabolism

Cytochrome b5 deficiency







17β-hydroxysteroid dehydrogenase III deficiency / 17-ketosteroid reductase deficiency




46,XY DSD; progressive virilization and gynecomastia at puberty

Decreased or absent


5α-reductase II deficiency




46,XY DSD; progressive virilization and gynecomastia at puberty



3α-hydroxysteroid dehydrogenase deficiency





46,XY DSD; gonadal insufficiency



Steroidogenic factor 1




46,XY DSD; gonadal insufficiency – very variable



The Alternative, Backdoor Pathway for DHT Synthesis

About 15 years ago, studies of sex development in the tammar wallaby revealed predominance of an alternative pathway for DHT production within the testis of the pouch young (Fig. 3) (Auchus 2004). It was named “backdoor pathway” as it diverges off from common precursors of the classic pathway without using T as the intermediate to produce DHT. In detail, in the backdoor pathway 17OHP is 5α-reduced (SRD5A1) to 17OH-dihydroprogesterone (17OH-DHP) and 3α-reduced (AKR1C2/4) to 17OH-allopregnanolone (17OH-Allo). 17OH-Allo is an excellent substrate for 17,20-lyase (CYP17A1/POR) to form androsterone, which is further converted to androstanediol (by AKR1C3/HSD17B3), and finally oxidized (by AKR1C2/4 or RoDH) to DHT. Alternatively, androsterone may be first oxidized to androstenedione before being converted to DHT. The characteristic of this backdoor pathway is that the steroid flux bypasses conventional intermediates of the classic pathway (e.g., DHEA, androstenedione, T) and uses different enzymes (SRD5A1, AKR1C2/4, RoDH/HSD17B6) for DHT production.

As this pathway was also found in rodents, its existence in human (fetuses) was suspected (Auchus 2004). Urine steroid analysis of patients suffering from POR deficiency (PORD) showing variable signs of androgen excess and deficiency revealed first hints (Fukami et al. 2013). Intermediates of the backdoor pathway (17OH-Allo and androsterone) were disproportionately increased in PORD patients. Similar findings were seen in untreated patients with 21OH-ase deficiency (due to CYP21A2 mutations). Urine steroid profiling revealed an elevated androstanediol and an elevated androsterone/etiocholanolone ratio in the neonatal period, suggesting an increased steroid flux through the backdoor pathway in this disease state (Kamrath et al. 2012). It is therefore likely that in utero virilization of girls (46,XX DSD) with PORD or 21-hydroxylase deficiency might be promoted by increased 17OHP being processed to DHT through the backdoor pathway (Auchus 2004; Fukami et al. 2013). Final proof for a role of the backdoor pathway in human sexual development came from 46,XY patients suffering from moderate to severe undervirilization (DSD) with mutations in the genes for AKR1C2/4, but no mutations in genes comprised in the classic pathway (Biason-Lauber et al. 2013). However, a detailed biochemical analysis of such patients is still missing, and the question remains, why the classic pathway cannot compensate for a defect in the backdoor pathway. Similarly, severe undervirilization of the external genitalia at birth with 46,XY DSD due to 5α-reductase (SRD5A2) deficiency appears illogical when suggesting an alternative backdoor pathway that is able to produce DHT without depending on SRD5A2 during fetal life.

At present, the discovery of the backdoor pathway leaves us with more open than solved questions in the field of androgen biosynthesis. Recent findings from steroid metabolomics studies during fetal-neonatal transition and gene expression studies of fetal versus adult testis tissues suggest that the activity of the backdoor pathway changes from fetal to postnatal to adult life and varies with disease states. In addition, genes of the backdoor pathway were found to be expressed in a tissue- and developmental stage-specific manner (Flück et al. 2011; Marti et al. 2016), illustrating that there is a fetal to adult shift in the gene expression pattern of backdoor pathway genes in the human testis (Fig. 4).
Fig. 4

Chemical structures of natural and synthetic androgens. Several synthetic analogues of T and DHT have been available since 1930s and have been widely used in treatment of anemia, starvation, and bone disorders as well as in body building and sports doping. All known androgen analogues have adverse side effects that range from behavioral changes to acute cardiovascular defects, and precaution is needed in their usage after consideration of benefits and risks

Yet another alternate pathway for DHT production directly from androstenedione, through 5α-reductase activity (SRD5A1) which produces androstanedione as intermediate (Fig. 2), has recently been described in castration-resistant prostate cancers (Chang et al. 2011), and was also suggested from steroid profiling in PCOS patients (Fassnacht et al. 2003). Furthermore, 11β-hydroxyandrostenedione, an inactive C19 metabolite which is produced abundantly and secreted into circulation by the human adrenal cortex, may be further metabolized into potent, active androgens as shown again in prostate cancer studies (Swart and Storbeck 2015). Whether all these novel pathways and androgen metabolites are important for normal androgen physiology remains to be studied in detail. However, these novel findings illustrate that T and the classic T biosynthetic pathway alone may no longer suffice for understanding human androgen biology.


Biochemistry of T, DHT, and Their Physiologic Precursors

Similar to all androgens, the biosynthesis of T also starts from cholesterol (see section “The Classic Pathway of T Synthesis”). Pregnenolone derived from cholesterol metabolism is converted to DHEA in two steps by CYP17A1. In the next steps DHEA is converted to androstedione by HSD3B2 and then HSD17B enzymes produce testosterone. After the secretion from the testes, T gets distributed rapidly in peripheral tissues. Several steroid-binding proteins may dictate the amount of T availability, act as storage media, and slow down the degradation of T by liver enzymes. T is further metabolized into the highly potent dihydrotestosterone by 5α-reductase (SRD5A1) . Aromatase also metabolizes T and converts a small percentage of T into estradiol, but DHT is not metabolized by aromatase. Depending on tissue level bioavailability of T and enzymes present, different metabolites of T may be produced. Higher levels of SRD5A1 expression is directly associated with increased DHT concentration, while lower levels of 3α and 3β hydroxysteroid dehydrogenases reduce further metabolism of DHT. In the prostate HSD17B2 can convert T into androstenedione. Therefore, peripheral concentrations of T or DHT may not be used as an indicator of their bioavailability, which is heavily dependent on expression of other enzymes that may further metabolize these two androgens and impact the actual local concentrations in different tissues.

Many synthetic analogues of T have been produced to increase the potency of anabolic effects and bioavailability of T. Commonly used anabolic steroids include nandrolone (available as nandrolone phenylpropionate and nandrolone decanoate), methandienone, stanozolol, oxandrolone, methenolone, and trenbolone (Fig. 4). Among these nandrolone and stanozolol have been used frequently in sports doping. However, the main approved clinical uses of both nandrolone and stanozolol have been for treatment of anemia and osteoporosis. Nandrolone has been proposed to be less toxic than many T metabolites/derivatives and is also a poor substrate for aromatase, which reduces some of the undesirable side effects due to estrogen production from excess of T. However, many adverse side effects associated with use of anabolic steroids are still associated with nandrolone analogues including erectile dysfunction and cardiovascular damage. Stanozolol have also been associated with major adverse effects (see chapter by Handelsman on androgen use, misuse, and abuse).

Action of T, DHT, and Androgen Receptor

The majority of androgen effects are routed through androgen receptor (AR, NR3C4), which belongs to the nuclear receptor gene family reviewed in Matsumoto et al. (2013) and Davey and Grossmann (2016) (A detailed account of androgen is given in a separate chapter by Sutinen, Malinen, and Palvimo). After the binding of androgens, AR is activated and regulates the expression of multiple target genes and small RNAs at a tissue-specific level. The AR is a ligand-dependent transcription factor which upon being activated by binding of androgens forms complexes with androgen response elements located on target genes (Govindan 1990; Davey and Grossmann 2016). AR is expressed in numerous tissues including male reproductive organs. In the absence of a ligand AR protein is primarily located in the cytoplasm and is found in complex with heat-shock proteins (HSPs). Once the AR is occupied by androgens/androgen analogues a signaling cascade starts, where the first step is the dissociation of AR from HSPs, followed by translocation of the AR to the nucleus. This mechanism of relocation of AR into the nucleus, triggered by ligand binding, provides a target for the regulation of AR activity. Based on total AR and ligand availability, less than 50% of the individual AR units seem to form complexes with androgens for nuclear transport and start the process to transcriptional regulation (Davey and Grossmann 2016). A large amount of unbound AR molecules are indicated to participate in other metabolic events, and there are implications of ligand-free AR effecting the cell cycle (Ueda et al. 2002). Most of the studies on AR have focused on androgen binding and transcriptional regulation, and roles of unbound AR in alternate cellular processes are not well known.

The AR gene (NR3C4) is located on the X chromosome at the locus Xq11–12 (NC_000023, 67,644,032…67730619). The protein coding region of the AR gene contains 2763 nucleotides (NCBI# NM_000044) distributed among 8 exons, which encode a 99-kDa protein (NCBI# NP_000035) consisting of 920 amino acids (Fig. 5a). The AR structure can be divided into distinct domains. The distinct subdomains of the AR structure include the N-terminal domain (NTD, residues 1–530), the DNA-binding domain (DBD, residues 530–623), and a hinge region (residues 623–669) which joins the DBD to the C-terminus ligand-binding domain (LBD, residues 669–920) (Fig. 5a). All three domains of AR are required for its function. The NTD accounts for more than half the size of total AR protein and is encoded by a single exon. The structure of NTD is not known so far and based on predictions, it is proposed to have a highly disordered structure, which may alternate between several different conformations. The disordered nature of NTD may be crucial for its capacity to form complexes with a wide range to coactivators and transcription factors.
Fig. 5

Structure of AR protein. (a): The AR has distinct functional domains similar to other members of steroid-binding nuclear receptors. The NTD, DBD, and LBD have precise functional requirements and are needed for the function of AR. (b): Structure of DBD bound to androgen response elements. The AR functions as a dimer, and two zinc-binding motifs consisting of four cysteine residues each are located in each DBD unit. DNA is shown as green and pink ribbons, two distinct DBD from different AR proteins are shown as ribbons diagram in light blue and light orange. Zinc ions are shown as red spheres. Structure is from x-ray crystal data from rat AR DBD (PDB: 1R4I) (c): Ligand-binding domain of AR in complex with T based on x-ray crystal data (PDB: 2AM9). (d): A close-up of the ligand-binding pocket of AR

The DBD of AR is a cysteine-rich region and contains two zinc finger motifs which participate in direct binding of AR to DNA recognition elements (termed androgen response elements, ARE) in target genes. Based on crystal data the formation of these zinc-binding motifs comprises of four cysteine each that coordinate the zinc ions (Fig. 5b). AR proteins recognize AREs consisting of some variations of a repeat consensus palindromic sequence 5′-AGGTCA NNN TGACCT-3′. Towards the end of DBD there is a nuclear localization signal (NLS, residues 617–633), which sits between DBD and hinge region and is responsible for the transport of AR to nucleus. In addition, the hinge region has been found to be required for DNA and coactivator binding, emphasizing the fact that whole AR structure is required for its functional activities. The LBD of AR is located at the C-terminus and is a globular structure, which has been well characterized from x-ray crystallographic studies (Fig. 5c). The core structure of LBD is formed by 11 distinct helices which surround the hydrophobic ligand-binding pocket (Fig. 5d). Ligand binding to AR triggers conformational changes in protein structure, causing a shift of the terminal helix of LBD. This ligand-induced structural shift of the terminal helix has been proposed as necessary for complex formations with coregulators (van de Wijngaart et al. 2012). Therefore, binding of T/DHT/androgen analogues changes the conformation and induces the interaction of AR with its coregulatory proteins. The signaling cascades continue by the removal of repression elements and their replacement by coactivators. X-ray crystal structures of AR reveal that the changes in the terminal helix of LBD can be variable based on which ligand is bound with natural ligands, AR agonists, and antagonists showing different patterns of structural shifts, which may be responsible for the differences in the action of different ligands. Posttranslational modifications of the AR (e.g., phosphorylation) may also regulate its activities.

Nonclassical Actions of T/DHT

There are indications of some nuclear androgen receptor-independent pathways of T/DHT action (Foradori et al. 2008). In general, effects of androgens could be divided into two broad categories, the first being nuclear receptor-mediated long-term transcriptional effects and the second being the membrane receptor-mediated effects that resemble rapid signaling events (Walker 2010). Several reports showing rapid effects of androgens indicate that there may exist mechanisms of T/DHT action that do not use binding to AR as the first step. A key difference between these effects is the requirement for constant presence of androgens for membrane-mediated effects. Some examples of such rapid effect of androgens are effects on kinase phosphorylation/signaling, secretions of prostate-specific antigen, and GnRH and calcium flux (Walker 2010). T and its synthetic analogue nandrolone has been shown to cause rapid induction of calcium influx which may play a role in rapid power burst and recovery situations and suggests a major role in sports (Cavalari et al. 2012; de Castro et al. 2013). Constant high levels of androgens in female athletes produced by genetic mutations may have a role in aiding performance without the role of nuclear receptor-medicated long-term effects. Membrane receptors with much lower affinity for androgens than AR (GPRC6A, ZIP9) have been identified (Papakonstanti et al. 2003; Hatzoglou et al. 2005; Pi et al. 2010; Ko et al. 2014), and separate mechanisms requiring low or high concentrations of androgens may exist in different cell/tissues (Simoncini and Genazzani 2003). The major nonclassical effects of T seem to be mediated through increased intracellular calcium, inositol 1,4,5-triphosphate, and diacylglycerol (Lieberherr and Grosse 1994; Papakonstanti et al. 2003; Loss et al. 2004; Cavalari et al. 2012). Signaling pathways triggered by membrane effects of androgens could be further linked to nuclear receptor-mediated effects by a network of signaling pathways which seem to be linked. Therefore, some of these effects may not start with nuclear receptors but may still converge through intermediary networks (Walker 2010; Ko et al. 2014; Wang et al. 2014).

T Metabolism/Inactivation

Most of the T in plasma is protein bound and is found to be attached to gonadal steroid-binding globulin, sex steroid-binding globulin, or albumin. In the liver several cytochrome P450 enzymes including CYP2C9, CYP2C19, and CYP3A4 perform β-hydroxylations of testosterone to produce 2β-, 6β-, 11β-, and 16β-hydroxytestosterone (Choi et al. 2005). The major androgenic metabolite of T is DHT, which is formed by the action of SRD5A1 activity as described earlier. DHT is a much more potent androgen than T and binds to AR with even higher affinity than T. Conversion of T to DHT has been a target of drug development to produce inhibitors of DHT biosynthesis for treatment of prostate cancer, male-pattern baldness, and hirsutism. Among the compounds inhibiting 5α-reductase activities are finasteride, alfatradiol, and dutasteride. However, there are several adverse drug reactions associated with these inhibitors that include impotence, decreased libido, etc. DHT is inactivated by 3α/β hydroxysteroid dehydrogenases into 3α-androstanediol and 3β-androstanediol. Steroids are often found as sulfate conjugates, a transformation carried out by steroid sulfotransferases (Strott 1996). For example, most of the DHEA produced in adrenals is found in sulfonated form mediated by action of SULT2A1 (Neunzig et al. 2014). Several other cytochrome P450 proteins, including CYP3A5 and CYP3A7, can perform hydroxylation reactions on T/DHT. In addition, Uridine 5′-diphospho-glucuronosyltransferases (UGTs) can perform bioconjugation reactions on T/DHT to create inactive androgens.

Genetic Disorders of Testicular Steroidogenesis

Classic Androgen Biosynthesis Defects that May Also Cause Adrenal Insufficiency (StAR, CYP11A1, HSD3B2, CYP17A1)

These genetic defects affect early steps in steroid biosynthesis, which are essential for mineralocorticoid, glucocorticoid, and sex steroid biosynthesis. They cause adrenal insufficiency and sex hormone deficiency, and are therefore also known as congenital adrenal hyperplasias (CAH).

StAR and CYP11A1

Patients with severe defects in the genes for StAR or CYP11A1 have clinically indistinguishable features. They suffer from adrenal insufficiency including mineralocorticoid and glucocorticoid deficiency, as well as 46,XY DSD and gonadal insufficiency due to missing androgen production (Table 1) (Miller and Flück 2014). StAR facilitates the import of cholesterol from the outer to the inner mitochondrial membrane, where cholesterol is the essential initial substrate for the side chain cleavage system (CYP11A1/FDX1/FDXR) for all steroid biosynthesis (Miller and Auchus 2011). In 1955, Prader described a complete sex reversal 46,XY DSD patient who died from an adrenal crisis very early in life and was found to have grossly enlarged, fatty transformed adrenal glands (Prader and Gurtner 1955). The clinical findings prompted him to name this disorder lipoid CAH (LCAH). However, the underlying genetic defect of LCAH and the mechanism of the disease were only described years later, after the StAR gene had been cloned (Bose et al. 1996). Infants with severe StAR mutations (=classic lipoid CAH) manifest with adrenal insufficiency soon after birth, latest within the first year of life. Affected 46,XY babies present with female external genitalia (Miller and Flück 2014). Milder StAR mutations, in which activity is partially retained, cause nonclassic LCAH. They usually present after 4 years of age to adulthood with late-onset primary adrenal insufficiency only, and do not affect (male) sex development (Flück et al. 2011; Miller and Flück 2014). The mechanism of disease action of StAR deficiency has been described by a “two hit model” (Bose et al. 1996). In this model the first hit consists of the actual loss of StAR activity for cholesterol import into the mitochondrium of the affected cell. However, as about 10% of cholesterol import occurs StAR-independent, a second hit is necessary to explain a severe phenotype. Thus, the second hit consists of the destruction of the steroidogenic cell through accumulation of cholesterol and cholesterol esters. Consistent with this model, the testis Leydig cell, which produces androgens early in fetal life, will be damaged early, resulting in 46,XY DSD. By contrast, the ovary, which is basically inactive in steroidogenesis until puberty, might be only affected by StAR deficiency beyond puberty. Affected females may therefore present with normal pubertal development and menses for a certain time, until the second hit strikes (Miller and Flück 2014).

Human CYP11A1 mutations manifest clinically identical to StAR mutations (Miller and Flück 2014). They occur as a classic form with severe mutations and as a nonclassic form with partial loss of enzyme activity. In contrast to StAR mutations, which usually present with adrenal enlargement in adrenal imaging, this finding is not observed in CYP11A1 deficiency. The serum or urine steroid profile of patients with classic StAR or CYP11A1 deficiency is characterized by overall (very) low production of all steroids (mineralocorticoids, glucocorticoids, and sex steroids). However, the exact diagnosis is made by genetic testing.


The biochemical profile of HSD3B2 deficiency has been described more than 50 years ago (Bongiovanni 1962). Severe HSD3B2 deficiency causes mineralocorticoid and glucocorticoid deficiency as well as (partial) androgen deficiency and results in 46,XY undervirilization and 46,XX virilization (Miller and Flück 2014). This is due to the fact that in humans there are two functional HSD3B genes, which express enzymes with similar activity. While HSD3B2 is exclusively expressed in the gonads and the adrenals, HSD3B1 is more widely expressed in the placenta and in peripheral tissues including liver and skin. Mutations in the human HSD3B1 gene have not been described. But in case of HSD3B2 deficiency, peripheral enzyme activity of HSD3B1 may convert circulating androgen precursors secreted from the adrenals or gonads into more active androgens. This may cause virilization of 46,XX females. Likewise, in severe HSD3B2 deficiency massively increased 17OHPreg may be converted to 17OHP through peripheral HSD3B1 activity, and affected newborns may therefore be picked up in the neonatal screening for 21-hydroxylase deficiency. In general, 3β-hydroxysteroid dehydrogenases (3βHSDs/HSD3Bs) convert delta 5 steroids (Preg, 17OHPreg, DHEA, androstenediol) to delta 4 steroids (Prog, 17OHP, androstenedione, T) (Fig. 3). Thus high ratios of the delta 5 over the delta 4 steroids are characteristic for HSD3B2 deficiency. Although the steroid profile of HSD3B2 deficiency is very characteristic, genetic analysis of the HSD3B2 gene is recommended to confirm the diagnosis.


CYP17A1 is the qualitative regulator of steroidogenesis in humans and has two distinct enzyme activities. Therefore, CYP17A1 deficiency exists in two forms. The first, more frequent form consists of combined loss of both 17α-hydroxylase and 17,20-lyase activities and results in glucocorticoid and sex hormone deficiency. The second form consists of an isolated loss of 17,20-lyase activity, which is described only in few patients so far and affects androgen biosynthesis exclusively (Table 1, Fig. 3) (Miller 2012; Miller and Flück 2014). Thus both forms cause 46,XY DSD and gonadal insufficiency. Clinical presentation of 46,XY DSD due to CYP17 deficiency varies from apparently female to undervirilized male with absence of Müllerian structures, hypoplastic Wolffian structures, and intra-abdominal or maldescended testes. Pubertal development is missing. Gynecomastia is only seen in partial insufficiency. With severe 17-hydroxylase deficiency, the mineralocorticoid synthesis pathway of steroidogenesis is functional, and the CYP17 enzyme block will lead to increased production of corticosterone and 11-deoxycorticosterone (DOC), which both have mineralocorticoid properties. This will suppress renin and result in hypertension and hypokalemic alkalosis in the patient. As corticosterone also has glucocorticoid activity, patients do not generally suffer from clinically relevant adrenal insufficiency, although ACTH is mildly elevated. Thus the typical steroid profile of combined CYP17 deficiency consists of high DOC, but low(ish) cortisol and low androgens.

By contrast, isolated 17,20-lyase deficiency caused by rare CYP17A1 mutations is due to the loss of the 17,20-lyase activity of the enzyme. This activity is essential for androgen production of the classic and backdoor pathway (Fig. 3). Lyase activity requires POR and CYB5 for its full functionality. So far, only mutations at locations E305, R347, and R358 of the CYP17A1 protein have been shown to cause isolated lyase deficiency (Miller and Auchus 2011; Miller 2012). But specific mutations in POR and CYB5A may also cause isolated lyase deficiency.

Classic Androgen Biosynthesis Defects Without Adrenal Insufficiency (HSD17B3, SRD5A2)


There are maybe more than 14 isoforms of human 17β-hydroxysteroid dehydrogenases (17βHSDs), which have variable physiological functions. Some isoforms are preferentially reductases, others oxidases. Human mutations are only known for the HSD17B3 gene and cause 46,XY DSD due to 17-ketosteroid reductase/17βHSD3 deficiency. Type 3 17βHSD (HSD17B3) is exclusively expressed in the testes, where it reduces androstenedione to T, DHEA to androstenediol, androstanedione to DHT, and androsterone to androstanediol (Fig. 3). Thus HSD17B3 deficiency is a male sex-limited disorder causing 46,XY DSD with severe to complete undervirilization of the external genitalia with a blind vaginal pouch. Müllerian structures are absent while Wolffian structures are present. Testicular descent is disturbed, and testes are often located inguinally. When patients with HSD17B3 deficiency are raised as females, they virilize at puberty as redundant other 17βHSD isoform enzyme activities convert testicular androstenedione to T in the periphery. The diagnostic steroid pattern for HSD17B3 deficiency is a low ratio of T over androstenedione either basally or after hCG stimulation (T/AD <0.8 after hCG) (Faisal Ahmed et al. 2000). Genetic confirmation is recommended.


In humans, there are two functionally active 5α-reductases (5α-Red). Both convert T to more potent DHT (Fig. 3). The type I enzyme (5α-Red1) is encoded by the gene SRD5A1 located on chromosome 5p15 and expressed in peripheral tissues such as the skin. The type II enzyme (5α-Red2) is encoded by SRD5A2 on chromosome 2p23 and is predominantly expressed in male reproductive tissues (Miller and Flück 2014). The well-known syndrome of 5α-reductase deficiency is caused by numerous mutations in the SRD5A2 gene. Typically, affected 46,XY individuals manifest at birth with female-appearing external genitalia, as the virilization of the external genitalia seems to depend largely on DHT; although it remains unsolved why DHT production through the backdoor pathway, which depends on 5α-Red1, may not compensate. Apart from the severe under/nonvirilization of the external genitalia, patients with SRD5A2 deficiency show a rather normal male sex determination and differentiation during fetal development. At puberty, progressive virilization and gynecomastia occur spontaneously due to intact peripheral activity of 5α-Red1. This may prompt a change in gender role to male in individuals raised as female. Diagnosis of 5α-reductase deficiency may be suggested when a high serum T/DHT ratio (basal and hCG stimulated) is observed. However, assessment of a whole steroid profile (GCMS, LCMSMS) is even more informative as 5α-Reds are also important for the reduction of a variety of steroids (e.g., C21 steroids) in their metabolism, and this can be seen in the profile in addition (Miller and Auchus 2011). Human mutations in SRD5A1 have not been described. Overall, the type 1 and 2 genes show a complex pattern of developmental regulation of expression, which is thought to also play a role in fetal androgen biosynthesis through the classic versus the alternative backdoor pathway in the testis (Figs. 3 and 6) (Flück et al. 2011; Miller and Auchus 2011).
Fig. 6

Expression pattern of backdoor pathway genes in the human testis. (a) Fetal versus adult testis. Quantitative RT-PCR shows developmental changes in the expression of SRD5A1 and SRD5A2, as well as for AKR1C2 with the fetal testis expressing higher amounts of SRD5A1 and AKR1C2. (b) Immunohistochemical staining showing the expression and localization of backdoor pathway genes in an adult testis. Note that Leydig cells express CYP17, RoDH, and AKR1C3 abundantly, and SRD5A1, AKR1C2/4, and HSD3B2 at low levels

Redox Partner Defects (POR, CYB5)

Basic studies of isolated lyase deficiency have elucidated the important role of redox partner POR and CYB5 for the enzymatic reaction of CYP17A1 (Miller 2012). As a matter of fact, the same clinical phenotype seen with isolated 17,20-lyase deficiency due to specific CYP17A1 mutations may be mimicked by certain POR or CYB5 mutations (Hershkovitz et al. 2008; Idkowiak et al. 2012; Miller 2012).


POR is the obligate electron donor to all microsomal type 2 cytochrome P450s, which comprise many proteins involved in steroidogenesis and xenobiotic metabolism, heme catabolism, bile acid synthesis, as well as prostaglandin and retinoic acid synthesis (comprehensively reviewed in Pandey and Flück 2013; Pandey and Sproll 2014; Burkhard et al. 2017). For adrenal and gonadal steroidogenesis, reactions catalyzed by enzymes CYP21A2 (21-hydroxylase), CYP17A1, and CYP19A1 (aromatase) depend on POR for electron transfer from NADPH. The phenotype of POR deficiency (PORD) was first described 1985 in a 46,XY DSD patient with a steroid profile showing combined 21- and 17-hydroxylase deficiency (Peterson et al. 1985). Years later the underlying defect was then identified to lie within the human POR gene (Flück et al. 2004). POR mutations manifest clinically with a very broad phenotype ranging from 46,XY and 46,XX DSD, adrenal insufficiency and skeletal malformations (known as Antley Bixler syndrome) with severe mutations to a polycystic ovary syndrome-like phenotype with milder mutations. This broad phenotype may be explained by two facts: one, different mutations in the POR protein affect the electron transfer to its partners to different degrees; two, the same POR mutation affects the activity of different P450 partners also to different degrees as their interaction might differ (Pandey et al. 2007; Nicolo et al. 2010; Pandey and Flück 2013; Flück and Pandey 2017). For example, a severe POR mutation, which destroys electron transfer to all P450s (e.g., R457H), may be found in patients with either 46,XY or 46,XX DSD, adrenal insufficiency and skeletal malformations, and virilization of the mother during pregnancy may occur (Flück et al. 2004). By contrast, the POR mutation G539R for instance, which has been shown to affect predominantly 17,20-lyase activity, will have a milder phenotype and resemble isolated 17,20-lyase deficiency (Hershkovitz et al. 2008). PORD cannot be diagnosed from the clinical picture alone, but the steroid profile from urine or plasma (GCMS or LCMSMS) can be used for diagnosis. Interestingly, steroid metabolites of the backdoor pathway have been found elevated in patients with PORD, and it has been suggested that these steroids contribute towards the observed intrauterine virilization of an affected female fetus and the mother during pregnancy (Homma et al. 2006). Final proof of the diagnosis of PORD however requires genetic testing and laboratory analysis of mutations (Parween et al. 2016; Burkhard et al. 2017).


Human CYB5 mutations have been found in very rare cases of 46,XY patients with low androgens and gonadal insufficiency, but normal mineralocorticoid and glucocorticoid production due to apparent isolated 17,20-lyase deficiency, when genetic mutations in the CYP17A1 and POR genes were not found (Idkowiak et al. 2012). CYB5 was thought to support 17,20-lyase activity by facilitating the allosteric interaction between the proteins POR and CYP17A1 (Miller 2012; Miller and Flück 2014) but recently has also been shown to act as a redox partner (Duggal et al. 2016). Only two CYB5 mutations (W28X, H44L) have been described in 46,XY DSD individuals so far (Idkowiak et al. 2012).

Genetic Defects of the Backdoor Pathway: Human AKR1C2/4 Mutations

The backdoor pathway requires reductive and oxidative 3α-hydroxysteroid dehydrogenase (3αHSD) activities for androgen production (Fig. 3). The four major human 3αHSDs are aldoketoreductases of the AKR1C family and have in principle reductive activity (Miller and Auchus 2011). AKR1C1–4 are located on chromosome 10p14–15 and have specific tissue distribution and specific catalytic characteristics. AKR1C3 is also known as 17βHSD5 (HSD17B5) and catalyzes the conversion of androstenedione to T in the adrenals and ovaries, and in nonsteroidogenic tissues. Expression of AKR1C3 is higher in the human fetal adrenal and testis and may participate (together with 17βHSD3/HSD17B3) in the conversion of androsterone to androstanediol in the backdoor pathway (Fig. 6). Both AKR1C2 and AKR1C4 are able to convert 17OH-DHP to 17OH-Allo. They are both expressed in testes and adrenals, but the fetal testis expresses more AKR1C2 than AKR1C4 (Fig. 6) (Flück et al. 2011). By contrast, AKR1C2/4 harbor both only minimal oxidative activity. Therefore, in the fetal testis the conversion of androstanediol to DHT is unlikely supported by these enzymes, but rather by the oxidative HSD17B6/RoDH (retinol dehydrogenase), which is also abundantly expressed in the human prostate. Recently, we have identified first combined mutations in AKR1C2/4 in patients with a phenotype similar to isolated 17,20-lyase deficiency manifesting with moderate to severe forms of 46,XY DSD. In these patients mutations in the CYP17A1, POR, CYB5, and SRD5A2 genes were excluded (Zachmann 1996; Flück et al. 2011). The index family presented initially two 46,XY DSD patients, one with cryptorchidism and genital undervirilization (raised male) and the other with an apparent female phenotype and no uterus (raised female) (Zachmann 1996). Family history revealed an aunt with severe 46,XY DSD (female gender, tall stature, primary amenorrhea) and a cousin with 46,XY DSD with moderate undervirilization (raised male). Adrenal insufficiency was excluded in all. At first, genetic analysis of the genes of the backdoor pathway for androgen production revealed AKR1C2 mutations in the affected individuals and suggested an autosomal recessive, male sex-limited pattern of inheritance. Wide phenotypical variability and functional studies of identified AKR1C2 mutants showing only moderate activity loss of 20–80% of the enzymes did not satisfy to explain the disease and prompted to search for a second hit. Linkage analysis picked up the AKR1C locus, which contains five closely related AKR1C genes. Further investigations of these genes revealed a second hit. Affected patients were all found to harbor a splicing mutation in AKR1C4 together with the AKR1C2 (I79V) mutation. In another severely affected 46,XY DSD patient with female external genitalia and intra-abdominal testes, a complex chromosomal rearrangement in the AKR1C locus was found. This included an unequal crossing over between the AKR1C2 and the AKR1C1 genes, and an additional missense mutation (H222Q) in the AKR1C2 gene, which was inactive in functional tests (Flück et al. 2011). These multigenic defects found in the backdoor pathway of androgen biosynthesis in 46,XY DSD patients provide convincing evidence that the backdoor pathway plays a crucial role for human fetal male sex development, even though we do not understand the exact interplay with the classic pathway and do not know its impact in postnatal life yet. Also, in contrast to notes in current textbooks, the fetal testis expresses in the Leydig cells SRD5A1 and thus seems able to produce DHT (Figs. 3 and 4). However, why testicular or peripheral SRD5A1 may not compensate for SRD5A2 deficiency during fetal sex development remains to be explained.

Steroidogenic Factor 1 (SF1)/NR5A1 Deficiency

Steroidogenic factor 1 (SF1/NR5A1) was originally identified in 1991 as an important transcription factor regulating genes of steroidogenesis including StAR, CYP11A1, and CYP17A1 (Suntharalingham et al. 2015). The knockout mouse model revealed a phenotype of complete sex reversal and adrenal insufficiency (due to a lack of adrenal glands) in males. This phenotype was also found in a first patient with 46,XY DSD and cortisol deficiency harboring a heterozygote NR5A1 mutation. Meanwhile numerous patients are described, most of them with an isolated 46,XY DSD phenotype only, encompassing a wide spectrum from mild hypospadias to complete sex reversal, which remains poorly understood (Camats et al. 2012). Adrenal insufficiency with NR5A1 deficiency seems very rare. Genetically, most patients harbor heterozygous mutations, which manifest variably even within families. SF1 deficiency may also cause gonadotropins deficiency and asplenia (Suntharalingham et al. 2015). Affected females may present with primary ovarian insufficiency. Testicular steroidogenesis is mostly disturbed with SF1 mutations and T production is therefore low. However, as SF1 is also critically involved in early sex determination and differentiation, not only steroidogenesis of the Leydig cell may be disturbed, but the overall development of the gonad may show severe abnormalities, presenting as streak gonad in worst case. Müllerian structures are variably persistent reflecting variable AMH levels with SF1 mutations. Overall, a characteristic clinical or biochemical presentation of SF1 deficiency does not exist; therefore the diagnosis must be made by genetic analysis.

Similarly, the transcription factor GATA4, which regulates many steroidogenic genes in collaboration with SF1, may also cause 46,XY DSD with low T production (Lourenco et al. 2011). Characteristically, these patients also manifest with congenital heart defects.

Other Genetic Defects Affecting Testosterone Biosynthesis

In principal, steroidogenesis of the testis may be disturbed by any (genetic) disorder of testis development. These defects often result in severe forms of 46,XY DSD (e.g., SRY, SOX9 mutations; see chapter on DSD) with low or absent androgen biosynthesis. Additionally, genetic defects of the hypothalamic-pituitary gonadal (HPG) axis (e.g., GnRH receptor, KAL1, FGFR1/FGF8, ROKR2/PROK2, etc., mutations), which controls gonadal functions (Fig. 2), may also affect testicular steroidogenesis. However, these defects do not generally present with a DSD phenotype at birth, although with missing stimulation the male external genitalia may appear small at birth, qualifying as microgenitalia. Typically, these defects manifest after 10 years of age for failure of pubertal development due to T deficiency in boys and require T replacement therapy.

By contrast, mutations in the LH receptor (LHCGR gene), which do not transmit the HP stimulus to the gonad and result in Leydig cell aplasia or hypoplasia, lead to a moderate to severe DSD phenotype at birth, depending on the severity of LHCGR inactivation. LH levels are typically elevated, and T levels are low.

Furthermore, testicular steroidogenesis may also be impaired by any disorder affecting the overall cholesterol biosynthesis, as all steroid hormones are produced from cholesterol. A typical example for this is the Smith Lemli Opitz (SLO) syndrome, in which mild genital anomalies (bilateral cryptorchidism, hypospadia) are seen in 50% of 46,XY individuals. SLO is caused by dehydrocholesterol reductase (DHCR) deficiency and leads to a block in the last step of cholesterol biosynthesis. Cholesterol is not only essential for steroid hormone biosynthesis, but also for processing of sonic hedgehog. This is important for pattern formation of the limbs, face, and nervous system. DHCR deficiency therefore causes a multitude of syndromic features and impairs cognitive functions to variable degrees.

A transient T biosynthetic defect during fetal development is also suggested to explain the 46,XY DSD phenotype at birth with X-linked MAMLD1 (mastermind-like domain-containing 1) mutations. Observed genital anomalies are often milder (hypospadia), and postnatal testicular function is reported to be normal. Finally, it is important to note here briefly that androgen production may only be effective, if the hormonal receptor is intact (Fig. 2). Mutations in the androgen receptor (or thus far unidentified comodulators) lead to complete or partial androgen insensitive syndromes (also called 46,XY DSD of androgen action), which resemble phenotypically the androgen deficiency syndromes described above. However, with AR mutations T levels are usually elevated.

Effects of Disordered Steroidogenesis on Testis Histology, Fertility, and Spermatogenesis

Little is known on testis histology, fertility, and malignancy risk in 46,XY DSD due to androgen biosynthetic defects. A summary of the literature (mostly case reports) has been reported recently (Burckhardt et al. 2015). Overall, defects in T synthesis seem to affect predominantly Leydig cells, which might be absent or reduced and with vacuoles and fat accumulation (Fig. 7). However, anomalies in seminiferous tubules containing Sertoli cells and germ cells are also observed to different degrees. A characteristic picture for specific enzymatic defects is not observed. Spermatogenesis seems impaired or even absent in many cases and contributes towards infertility together with androgen deficiency. These findings confirm that androgens are needed in the testis itself for normal sperm production, and that there must be an indispensable interplay between Leydig and Sertoli cells. Concerning malignancy risk with 46,XY DSD due to androgen deficiency, data are even more scarce. Evaluation in very few cases suggested probably low risk, but further studies are needed to conclude on this question.
Fig. 7

Histology of testis tissues from patients diagnosed with congenital androgen biosynthetic disorders, NR5A1/SF1, StAR, and HSD3B2. General histology with hematoxylin and eosin stain (HE) shows well-developed seminiferous tubules with few germ cells and few Leydig cells. Leydig cells present with cytoplasmic vacuoles in SF1 and StAR deficiency. Overall, a Sertoli-cell-only pattern is characteristic

Outlook, What Do We Not Know?

Androgen metabolism has been a constantly growing field for a long time. There is large interest in use of androgens as antiaging agents, and novel synthetic androgens that do not possess undesirable side effects and toxicity of natural androgens are being synthesized since 1930s. Use of androgens to slow down or reverse “andropause” is actively being pursued in several large clinical studies. However, recent reports of severe side effects of T usage have dampened the enthusiasm, and less toxic, perhaps synthetic androgens (or selective androgen receptor modulators, SARMs) may be investigated in future to bypass the side effects associated with T/DHT usage. Despite heavy research on synthesis and metabolism of androgens, there are several aspects of androgen regulation and tissue level bioavailability that need further studies. A major area of research is the fate of androgens during aging. It is known that DHEA levels start rising around puberty and reach a peak during 20–25 years of age then slowly decline throughout life in a phenomenon known as adrenopause. A simplistic extrapolation would deduce that levels of T and DHT would also severely decline with age. However, not much is known about differential expression of different enzymes that activate T or deactivate T and DHT in different tissues. Since a wide range of combinations of activating and inactivating enzymes may influence the actual level of T/DHT in a particular cell/tissue, further studies are needed to have a systems biology approach towards androgen production and regulation through different stages of life. The 3α/3β hydroxysteroid dehydrogenases, 17β hydroxydehydrogenases, 5α/5β reductases, and catabolic cytochrome P450 enzymes that use T/DHT as a substrate (CYP3A4, CYP3A5, CYP3A7, etc.) dictate the actual availability of androgens. From the alternate pathway we have learned that DHT can also be produced from indirect sources and existence/activity of these enzymes, especially in aging and different physiological conditions requires further studies. It is possible that the alternate pathway exists to provide the more potent androgen, DHT, even under limiting T availability to keep the important androgen-dependent pathways functioning. Another aspect of androgen action that is not precisely clear is activation of AR by androgens. Very low quantities of DHT that are an order of magnitude lower than binding constants of androgens for AR have been known to cause AR activation. A partial activation theory and phosphorylation of AR have been proposed to be responsible for these observations, but exact mechanisms that may govern AR activation and whether different physiological conditions affect these mechanisms need further studies. Role of UGTs and sulfokinases in bioconjugation of androgens have not received much focus, and majority of androgen metabolism studies have focused on direct metabolism of androgens by steroid dehydrogenases and cytochrome P450 proteins. Recent works from Yuji Ishi and colleagues from Fukuoka has suggested that protein-protein interactions between P450s and UGTs can influence metabolic processes (Ishii et al. 2010, 2014). Here also a systemic look at such combinations and changes under different physiological conditions require further studies. Nonclassical roles of androgens that does not involve nuclear translocation upon the activation of AR as the first step of androgen signaling also need more investigation. Several novel compounds that may bind to AR but stop its relocation to nucleus, i.e., antiandrogens, have been developed.


T is a major circulating androgen, but there are more potent metabolites like DHT which have much higher affinity for AR. The classical pathway of T biosynthesis starts from import of cholesterol into mitochondria and then proceeds through several enzymes to form T/DHT. In addition to direct conversion of T to DHT by 5α-reductase, some alternate pathways of DHT production have recently been identified, especially in fetal life and under disease states like CAH, PCOS, and prostate cancer. Genetic disorders in any of the enzymes and their partner proteins like CYP11A1, CYP17A1, POR, CYB5A, HSD17B, and SRD5A1 may directly influence androgen levels and availability. Most of the T is bound to proteins, which may also control its bioavailability and degradation. The local expression of 3α/β hydroxysteroid dehydrogenases, 5α-reductase, and 17β-hydroxydehydrogenases, which can activate/inactivate T/DHT, may govern the actual amount of T/DHT in a particular cellular environment, and general estimations based on peripheral measurements in blood/saliva may not provide a full picture of androgen bioavailability. Impact of physiological conditions like aging and different disease states may create difference in cellular environments that could influence androgen production and bioavailability and require further investigation. In addition to long-term posttranslational effect of androgens through nuclear localization of AR upon androgen binding, there are also membrane receptor-based effects of androgens which result in rapid response to androgen exposure, an example being rapid calcium influx in muscle. Many synthetic analogues of T/DHT have been produced, and improvements to natural androgens are constantly being pursued to increase bioavailability and minimize the side effects associated with exposure to anabolic steroids . More research into the mechanism of action of androgens and bioavailability is required to potentially develop treatments to reverse the effects in men whose androgen levels drop upon aging.



This work has been supported by the Swiss National Science Foundation grant 320030-146127.


  1. Auchus RJ. The backdoor pathway to dihydrotestosterone. Trends Endoscrinol Metab. 2004;15:432–8.CrossRefGoogle Scholar
  2. Biason-Lauber A, Miller WL, et al. Of marsupials and men: "backdoor" dihydrotestosterone synthesis in male sexual differentiation. Mol Cell Endocrinol. 2013;371(1–2):124–32.CrossRefPubMedGoogle Scholar
  3. Bongiovanni AM. The adrenogenital syndrome with deficiency of 3 beta-hydroxysteroid dehydrogenase. J Clin Invest. 1962;41:2086–92.CrossRefPubMedPubMedCentralGoogle Scholar
  4. Bose HS, Sugawara T, et al. The pathophysiology and genetics of congenital lipoid adrenal hyperplasia. N Engl J Med. 1996;335(25):1870–8.CrossRefPubMedGoogle Scholar
  5. Burckhardt MA, Udhane S, et al. Human 3beta-hydroxysteroid-dehydrogenase deficiency seems to affect fertility but may not harbor a tumor risk: lesson from an experiment of nature. Eur J Endocrinol. 2015;173(5):K1–K12.CrossRefPubMedGoogle Scholar
  6. Burkhard FZ, Parween S, et al. P450 oxidoreductase deficiency: analysis of mutations and polymorphisms. J Steroid Biochem Mol Biol. 2017;165(Pt A):38–50.CrossRefPubMedGoogle Scholar
  7. Camats N, Pandey AV, et al. Ten novel mutations in the NR5A1 gene cause disordered sex development in 46,XY and ovarian insufficiency in 46,XX individuals. J Clin Endocrinol Metab. 2012;97(7):E1294–306.CrossRefPubMedGoogle Scholar
  8. de Castro AL, Cavalari FC, et al. Epitestosterone and testosterone have similar nonclassical actions on membrane of Sertoli cells in whole seminiferous tubules. Horm Metab Res. 2013;45(01):15–21.PubMedGoogle Scholar
  9. Cavalari FC, de Castro AL, et al. Non-classic androgen actions in JSertoli cell membrane in whole seminiferous tubules: effects of nandrolone decanoate and catechin. Steroids. 2012;77(1–2):118–25.CrossRefPubMedGoogle Scholar
  10. Chang KH, Li R, et al. Dihydrotestosterone synthesis bypasses testosterone to drive castration-resistant prostate cancer. Proc Natl Acad Sci USA. 2011;108(33):13728–33.CrossRefPubMedPubMedCentralGoogle Scholar
  11. Choi MH, Skipper PL, et al. Characterization of testosterone 11β-hydroxylation catalyzed by human liver microsomal cytochromes p450. Drug Metab Dispos. 2005;33(6):714–8.CrossRefPubMedGoogle Scholar
  12. Davey RA, Grossmann M. Androgen receptor structure, function and biology: from bench to bedside. Clin Biochem Rev. 2016;37(1):3–15.PubMedPubMedCentralGoogle Scholar
  13. Dhayat NA, Frey AC, et al. Estimation of reference curves for the urinary steroid metabolome in the first year of life in healthy children: tracing the complexity of human postnatal steroidogenesis. J Steroid Biochem Mol Biol. 2015;154:226–36.CrossRefPubMedGoogle Scholar
  14. Dhayat NA, Dick B, et al. Androgen biosynthesis during minipuberty favors the backdoor pathway over the classic pathway: insights into enzyme activities and steroid fluxes in healthy infants during the first year of life from the urinary steroid metabolome. J Steroid Biochem Mol Biol. 2017;165:312–322.Google Scholar
  15. Duggal R, Liu Y, et al. Evidence that cytochrome b5 acts as a redox donor in CYP17A1 mediated androgen synthesis. Biochem Biophys Res Commun. 2016;477(2):202–8.CrossRefPubMedGoogle Scholar
  16. Faisal Ahmed S, Iqbal A, et al. The testosterone:androstenedione ratio in male undermasculinization. Clin Endocrinol. 2000;53(6):697–702.CrossRefGoogle Scholar
  17. Fassnacht M, Schlenz N, et al. Beyond adrenal and ovarian androgen generation: increased peripheral 5 alpha-reductase activity in women with polycystic ovary syndrome. J Clin Endocrinol Metab. 2003;88(6):2760–6.CrossRefPubMedGoogle Scholar
  18. Fevold HR, Lorence MC, et al. Rat P450(17 alpha) from testis: characterization of a full-length cDNA encoding a unique steroid hydroxylase capable of catalyzing both delta 4- and delta 5-steroid-17,20-lyase reactions. Mol Endocrinol. 1989;3(6):968–75.CrossRefPubMedGoogle Scholar
  19. Flück CE, Pandey AV. Impact on CYP19A1 activity by mutations in NADPH cytochrome P450 oxidoreductase. J Steroid Biochem Mol Biol. 2017;165(Pt A):64–70.CrossRefPubMedGoogle Scholar
  20. Flück CE, Miller WL, et al. The 17, 20-lyase activity of cytochrome P450c17 from human fetal testis favors the Δ5 steroidogenic pathway. J Clin Endocrinol Metab. 2003;88(8):3762–6.CrossRefPubMedGoogle Scholar
  21. Flück CE, Tajima T, et al. Mutant P450 oxidoreductase causes disordered steroidogenesis with and without Antley-Bixler syndrome. Nat Genet. 2004;36(3):228–30.CrossRefPubMedGoogle Scholar
  22. Flück CE, Meyer-Boni M, et al. Why boys will be boys: two pathways of fetal testicular androgen biosynthesis are needed for male sexual differentiation. Am J Hum Genet. 2011;89:201–18.CrossRefPubMedPubMedCentralGoogle Scholar
  23. Flück CE, Pandey AV, et al. Characterization of novel StAR (steroidogenic acute regulatory protein) mutations causing non-classic lipoid adrenal hyperplasia. PLoS One. 2011;6(5):e20178.CrossRefPubMedPubMedCentralGoogle Scholar
  24. Foradori CD, Weiser MJ, et al. Non-genomic actions of androgens. Front Neuroendocrinol. 2008;29(2):169–81.CrossRefPubMedGoogle Scholar
  25. Fukami M, Homma K, et al. Backdoor pathway for dihydrotestosterone biosynthesis: implications for normal and abnormal human sex development. Dev Dyn. 2013;242(4):320–9.CrossRefPubMedGoogle Scholar
  26. Govindan MV. Specific region in hormone binding domain is essential for hormone binding and trans-activation by human androgen receptor. Mol Endocrinol. 1990;4(3):417–27.CrossRefPubMedGoogle Scholar
  27. Hatzoglou A, Kampa M, et al. Membrane androgen receptor activation induces apoptotic regression of human prostate cancer cells in vitro and in vivo. J Clin Endocrinol Metab. 2005;90(2):893–903.CrossRefPubMedGoogle Scholar
  28. Hershkovitz E, Parvari R, et al. Homozygous mutation G539R in the gene for P450 oxidoreductase in a family previously diagnosed as having 17,20-lyase deficiency. J Clin Endocrinol Metab. 2008;93(9):3584–8.CrossRefPubMedPubMedCentralGoogle Scholar
  29. Homma K, Hasegawa T, et al. Urine steroid hormone profile analysis in cytochrome P450 oxidoreductase deficiency: implication for the backdoor pathway to dihydrotestosterone. J Clin Endocrinol Metab. 2006;91(7):2643–9.CrossRefPubMedGoogle Scholar
  30. Idkowiak J, Randell T, et al. A missense mutation in the human cytochrome b5 gene causes 46,XY disorder of sex development due to true isolated 17,20 lyase deficiency. J Clin Endocrinol Metab. 2012;97(3):E465–75.CrossRefPubMedGoogle Scholar
  31. Ishii Y, Nurrochmad A, et al. Modulation of UDP-glucuronosyltransferase activity by endogenous compounds. Drug Metab Pharmacokinet. 2010;25(2):134–48.CrossRefPubMedGoogle Scholar
  32. Ishii Y, Koba H, et al. Alteration of the function of the UDP-glucuronosyltransferase 1A subfamily by cytochrome P450 3A4: different susceptibility for UGT isoforms and UGT1A1/7 variants. Drug Metab Dispos. 2014;42(2):229–38.CrossRefPubMedGoogle Scholar
  33. Kamrath C, Hochberg Z, et al. Increased activation of the alternative "backdoor" pathway in patients with 21-hydroxylase deficiency: evidence from urinary steroid hormone analysis. J Clin Endocrinol Metab. 2012;97(3):E367–75.CrossRefPubMedGoogle Scholar
  34. Ko E, Choi H, et al. Testosterone stimulates Duox1 activity through GPRC6A in skin keratinocytes. J Biol Chem. 2014;289(42):28835–45.CrossRefPubMedPubMedCentralGoogle Scholar
  35. Kuiri-Hanninen T, Sankilampi U, et al. Activation of the hypothalamic-pituitary-gonadal axis in infancy: minipuberty. Hormone research in paediatrics. 2014;82(2):73–80.CrossRefPubMedGoogle Scholar
  36. Lieberherr M, Grosse B. Androgens increase intracellular calcium concentration and inositol 1,4,5-trisphosphate and diacylglycerol formation via a pertussis toxin-sensitive G-protein. J Biol Chem. 1994;269(10):7217–23.PubMedGoogle Scholar
  37. Loss ES, Jacobsen M, et al. Testosterone modulates K+ATP channels in Sertoli cell membrane via the PLC-PIP2 pathway. Horm Metab Res. 2004;36(08):519–25.CrossRefPubMedGoogle Scholar
  38. Lourenco D, Brauner R, et al. Loss-of-function mutation in GATA4 causes anomalies of human testicular development. Proc Natl Acad Sci USA. 2011;108(4):1597–602.CrossRefPubMedPubMedCentralGoogle Scholar
  39. Marti N, Galvan JA, et al. Genes and proteins of the alternative steroid backdoor pathway for dihydrotestosterone synthesis are expressed in the human ovary and seem enhanced in the polycystic ovary syndrome. Mol Cell Endocrinol. 2016;441:116–23.CrossRefPubMedGoogle Scholar
  40. Matsumoto T, Sakari M, et al. The androgen receptor in health and disease. Annu Rev Physiol. 2013;75(1):201–24.CrossRefPubMedGoogle Scholar
  41. Miller WL. The syndrome of 17,20 lyase deficiency. J Clin Endocrinol Metab. 2012;97(1):59–67.CrossRefPubMedGoogle Scholar
  42. Miller WL, Auchus RJ. The molecular biology, biochemistry, and physiology of human steroidogenesis and its disorders. Endocr Rev. 2011;32(1):81–151.CrossRefPubMedGoogle Scholar
  43. Miller WL, Flück CE. Adrenal cortex and its disorders. In: Sperling MA, editor. Pediatric endocrinology. Philadelphia: Saunders; 2014.Google Scholar
  44. Neunzig J, Sánchez-Guijo A, et al. A steroidogenic pathway for sulfonated steroids: the metabolism of pregnenolone sulfate. J Steroid Biochem Mol Biol. 2014;144(Part B):324–33.CrossRefPubMedGoogle Scholar
  45. Nicolo C, Flück CE, et al. Restoration of mutant cytochrome P450 reductase activity by external flavin. Mol Cell Endocrinol. 2010;321(2):245–52.CrossRefPubMedGoogle Scholar
  46. Pandey AV, Flück CE. NADPH P450 oxidoreductase: structure, function, and pathology of diseases. Pharmacol Ther. 2013;138(2):229–54.CrossRefPubMedGoogle Scholar
  47. Pandey AV, Sproll P. Pharmacogenomics of human P450 oxidoreductase. Front Pharmacol. 2014;5:103.CrossRefPubMedPubMedCentralGoogle Scholar
  48. Pandey AV, Kempna P, et al. Modulation of human CYP19A1 activity by mutant NADPH P450 oxidoreductase. Mol Endocrinol. 2007;21(10):2579–95.CrossRefPubMedGoogle Scholar
  49. Papakonstanti EA, Kampa M, et al. A rapid, nongenomic, signaling pathway regulates the actin reorganization induced by activation of membrane testosterone receptors. Mol Endocrinol. 2003;17(5):870–81.CrossRefPubMedGoogle Scholar
  50. Parween S, Roucher-Boulez F, et al. P450 oxidoreductase deficiency: loss of activity caused by protein instability from a novel L374H mutation. J Clin Endocrinol Metab. 2016;101(12):4789–98.CrossRefPubMedGoogle Scholar
  51. Peterson RE, Imperato-McGinley J, et al. Male pseudohermaphroditism due to multiple defects in steroid-biosynthetic microsomal mixed-function oxidases. A new variant of congenital adrenal hyperplasia. N Engl J Med. 1985;313(19):1182–91.CrossRefPubMedGoogle Scholar
  52. Pi M, Parrill AL, et al. GPRC6A mediates the non-genomic effects of steroids. J Biol Chem. 2010;285(51):39953–64.CrossRefPubMedPubMedCentralGoogle Scholar
  53. Prader A, Gurtner HP. The syndrome of male pseudohermaphrodism in congenital adrenocortical hyperplasia without overproduction of androgens (adrenal male pseudohermaphrodism). Helv Paediatr Acta. 1955;10(4):397–412.PubMedGoogle Scholar
  54. Simoncini T, Genazzani A. Non-genomic actions of sex steroid hormones. Eur J Endocrinol. 2003;148(3):281–92.CrossRefPubMedGoogle Scholar
  55. Strott CA. Steroid sulfotransferases. Endocr Rev. 1996;17(6):670–97.CrossRefPubMedGoogle Scholar
  56. Suntharalingham JP, Buonocore F, et al. DAX-1 (NR0B1) and steroidogenic factor-1 (SF-1, NR5A1) in human disease. Best Pract Res Clin Endocrinol Metab. 2015;29(4):607–19.CrossRefPubMedPubMedCentralGoogle Scholar
  57. Swart AC, Storbeck KH. 11beta-Hydroxyandrostenedione: downstream metabolism by 11betaHSD, 17betaHSD and SRD5A produces novel substrates in familiar pathways. Mol Cell Endocrinol. 2015;408:114–23.CrossRefPubMedGoogle Scholar
  58. Ueda T, Mawji NR, et al. Ligand-independent activation of the androgen receptor by Interleukin-6 and the role of steroid receptor coactivator-1 in prostate cancer cells. J Biol Chem. 2002;277(41):38087–94.CrossRefPubMedGoogle Scholar
  59. Walker WH. Non-classical actions of testosterone and spermatogenesis. Philos Trans R Soc B: Biol Sci. 2010;365(1546):1557–69.CrossRefGoogle Scholar
  60. Wang C, Liu Y, et al. G protein-coupled receptors: extranuclear mediators for the non-genomic actions of steroids. Int J Mol Sci. 2014;15(9):15412.CrossRefPubMedPubMedCentralGoogle Scholar
  61. van de Wijngaart DJ, Dubbink HJ, et al. Androgen receptor coregulators: recruitment via the coactivator binding groove. Mol Cell Endocrinol. 2012;352(1–2):57–69.CrossRefPubMedGoogle Scholar
  62. Zachmann M. Prismatic cases: 17,20-desmolase (17,20-lyase) deficiency. J Clin Endocrinol Metab. 1996;81(2):457–9.PubMedGoogle Scholar

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  1. 1.Pediatric Endocrinology and Diabetology Department of Pediatrics, Bern University Hospital, and Department of Clinical ResearchUniversity of BernBernSwitzerland

Personalised recommendations