Fig 1.
Canonical and alternative (backdoor) pathways of DHT synthesis.
The canonical pathway has potential Δ4 and Δ5 subpathways. The enzymes that catalyze each step are indicated within the arrows. Enzymes written in black on a gray arrow are essential components of the canonical pathway, and some appear in both canonical and backdoor pathways (e.g., CYP17A1). Enzymes written in black on a green background are specific to the backdoor pathway. Enzymes in gray text will carry out the described conversion, but they may not be the principal enzyme involved. Other enzymes, not shown, may also be involved in components of the backdoor pathway [17]. Human CYP17A1 can convert progesterone to 17α-hydroxyprogesterone, but 17–20 lyase activity is very low with 17α-hydroxyprogesterone as substrate, and significant androstenedione is not produced by the Δ4 pathway in humans [18]. Similarly, 17OHDHP is a poor substrate for 17–20 lyase activity [19]. androstanediol, 5α-androstan-3α, 17β-diol; androstanedione, 5α-androstane-3,17-dione; androstenediol, androst-5-ene-3β,17β-diol; DHEA, dehydroepiandrosterone; DHT, dihydrotestosterone; 5αDHP, 5α-dihydroprogesterone; 17OHDHP, 17α-hydroxydihydroprogesterone.
Fig 2.
Concentrations of both canonical and backdoor steroids in the plasma of human male and female fetuses during the second trimester.
Steroid levels from 39–42 individual male fetuses and 16 female fetuses (measured by GC-MS/MS) are shown in each graph, with the mean level indicated with a black line. The number of samples that were ND for each steroid are shown and, where appropriate, the LOD is shown as a red dotted line. Data shown in gray were above the LOD but below the formal LOQ, which means that the quantified data shown for these samples are less reliable than for data shown in blue or pink, which were above the LOQ. The mean for male androsterone does not include the outlier. Differences between male and female steroid levels were measured by t test (using the Cohen correction when appropriate), and the significance of the difference is shown on each graph. The significance value shown for androsterone was determined without inclusion of the male outlier; if the outlier is included, P = 0.0001 after log transformation of the data. The plasma used in these studies was from fetuses aged between 12 and 19 weeks. Steroid abbreviations used are the same as those in Fig 1. Raw data are shown in S1 Data (Sheet 1). F, female; GC-MS/MS, gas chromatography–tandem mass spectrometry; LOD, limit of detection; LOQ, limit of quantification; M, male; ND, nondetectable.
Fig 3.
Levels of steroid intermediates involved in the canonical and backdoor synthesis of DHT in male fetal tissues.
Tissue levels of steroids (measured by LC-HRMS) from the placenta and fetal liver (n = 20; placentas and fetal livers were from the same pregnancies), fetal adrenal (n = 30), and fetal testis (n = 10 [for DHEA and androsterone] or 25 [for all other steroids]) are shown as individual points in each graph and arranged in the pathways shown in Fig 1. Levels of 17α-hydroxylated intermediates were not measured in this part of the study. The number of samples that were ND for each steroid are shown and, where appropriate, the LOD is shown as a red dotted line. Data shown in gray were above the LOD but below the formal LOQ, which means that the quantified data shown for these samples are less reliable. The LOD for each sample (in ng/mg tissue) depended on the mass of tissue extracted, and the lines drawn are based on the average mass of each tissue used. Green arrows indicate that the relevant enzymes are detectable (as mRNA transcripts) in that tissue, while red arrows indicate that the presumed enzyme is not detectable (based on data in Fig 4). Raw data are shown in S1 Data (Sheet 2). androstanediol, 5α-androstan-3α, 17β-diol; androstanedione, 5α-androstane-3,17-dione; androstenediol, androst-5-ene-3β,17β-diol; DHEA, dehydroepiandrosterone; DHT, dihydrotestosterone; LC-HRMS, liquid chromatography–high-resolution mass spectrometry; LOD, limit of detection; LOQ, limit of quantification; ND, nondetectable; 5αDHP, 5α-dihydroprogesterone; 17OHDHP, 17α-hydroxydihydroprogesterone.
Fig 4.
Levels of steroid intermediates involved in the canonical and backdoor synthesis of DHT in the testis and ovary as measured by GC-MS/MS.
Intra-gonadal steroid levels from 6 male and 4 female fetuses (aged 15–19 weeks) are shown as individual points in each graph and arranged in the pathways shown in Fig 1. Levels of 5αDHP were not measured in this part of the study. The number of samples which were ND for each steroid are shown. Data shown in gray were above the LOD but below the formal LOQ, which means that the quantified data shown for these samples are less reliable. Raw data are shown in S1 Data (Sheet 3). androstanediol, 5α-androstan-3α, 17β-diol; androstanedione, 5α-androstane-3,17-dione; androstenediol, androst-5-ene-3β,17β-diol; DHEA, dehydroepiandrosterone; DHT, dihydrotestosterone; F, female; GC-MS/MS, gas chromatography–tandem mass spectrometry; LC-HRMS, liquid chromatography–high-resolution mass spectrometry; LOD, limit of detection; LOQ, limit of quantification; M, male; ND, nondetectable; 5αDHP, 5α-dihydroprogesterone; 17OHDHP, 17α-hydroxydihydroprogesterone.
Fig 5.
Levels of mRNA transcripts encoding enzymes involved in the synthesis of DHT through the backdoor pathway in the fetal male.
Data show levels of transcripts in testis (n = 22), adrenal (n = 21), liver (n = 44–50), genital tubercle (n = 10), and placenta (n = 20) from individual male fetuses during the second trimester. Transcript levels have been measured relative to the housekeeping gene TBP. For each transcript, the y-axis has been maintained constant for all 5 tissues so that direct comparison of transcript levels can be made. The number of ND samples is shown on each graph. The horizontal black bar indicates mean expression and, in cases in which levels are very low, the mean is also provided in text on the graph. Raw data are shown in S1 Data (Sheet 4). DHT, dihydrotestosterone; GT, genital tubercle; ND, nondetectable; TBP, TATA box–binding protein.
Table 1.
Fetal tissue weights during the second trimester.
Fig 6.
Proposed steroidogenic pathways leading to androsterone synthesis and masculinization in the second trimester human male fetus.
Steroid hormone conversion is shown by wide green arrows, with the converting enzymes written within the arrow. Red arrows show potential transport between organs in the fetal circulation. The blue double-headed arrow indicates that exchange is also taking place between the placenta and the maternal circulation. Most circulating progesterone in the fetal circulation is likely to come from the placenta, and this will be reduced to 5αDHP by SRD5A1 in the placenta, fetal liver, and fetal testis, with the fetal liver likely to be the major site. Allopregnanolone (AlloP5) production by AKR1C2 is also most likely to occur in the placenta and fetal liver because the substrate is present in those tissues, and they express the highest total levels of enzyme transcript. Some conversion may also occur in the testis. Significant levels of androsterone are only detectable in the placenta and adrenal, and thus they are a likely source of the circulating steroid, although, given sex differences, other tissues are probably involved. The adrenal lacks other intermediates in the backdoor pathway, and thus AlloP5 must come from other tissues. The placenta lacks CYP17A1, so androsterone production is likely to depend on adrenal DHEA as substrate. Testosterone from the fetal testes also acts as an essential substrate for DHT synthesis at the external genitalia. AlloP5, allopregnanolone; An, androsterone; DHEA, dehydroepiandrosterone; DHT, 5α-dihydrotestosterone; P4, progesterone; P5, pregnenolone; T, testosterone; 5αDHP, 5α-dihydroprogesterone.