Edinburgh Research Explorer Characterization of human adrenal steroidogenesis during fetal development

: 35 Context: The endocrine function of human fetal adrenals (HFA) is activated already during first trimester, but 36 changes in adrenal steroidogenesis during fetal life are not well characterized. 37 Objective: This study aimed to investigate HFA steroidogenesis by analyzing adrenal glands from 1 st and 2 nd 38 trimester. (GW) 40 including a total This study investigated the adrenal steroidogenic expression pattern during the 1 st and 2 nd trimester of human fetal development. HFA glands function as active endocrine organs from early fetal development, which was 249 confirmed by this study demonstrating expression of all investigated steroidogenic enzymes both at the gene 250 and protein level in all investigated samples from GW 8-19. The distinct morphological expression pattern of 251 the investigated enzymes is in accordance with previous studies focusing on either 1 st or 2 nd trimester human 252 adrenals (4,7,8,10,11,13,14). Thus, our data support the prominent hypothesis of the FZ, and later in 253 development also the TZ, being the main site of de novo steroid synthesis in the HFA. Interestingly, the 254 transcription of steroidogenic genes appeared to be regulated differently with unaltered expression of ARK1C3 255 and 3β-HSD2 during the investigated period while the expression of the other steroidogenic enzymes ( StAR, 256 CYP11A1, CYP17A1, CYP21A2, CYP11B1/2, and SULT2A1) and the ACTH receptor MC2R were significantly 257 increased in 2 nd trimester. The observed increases in gene expression levels in 2 nd trimester was not evident 258 at the protein level; presumably due to the lack of sensitivity of immunohistochemistry to detect differences in 259 expression level. However, the genes upregulated during 2 nd trimester were also the steroidogenic enzymes 260 that showed the most abundant immunostaining in the FZ while the low constitutively expressed 3β-HSD1/2 261 only in a small subpopulation of cells mainly located in the DZ. Furthermore, the up-regulation 262 in gene steroidogenic enzymes not 2 Thus, our results demonstrate that 3β-HSD1/2 is highly expressed in a sub-population of DZ cells from GW 8- 271 19, which is in accordance with recent published data (13). Even though the percentage of 3β-HSD1/2 positive 272 cells in DZ and FZ appeared to decrease during 2 nd trimester, this study is the first to report a small number of 273 3β-HSD1/2 positive FZ cells in 2 nd trimester fetuses up to GW 19. In general, the small number of 3β-HSD1/2 274 positive cells compared with the abundant expression of the majority of the adrenal steroidogenic enzymes is 275 consistent with the reported low 3β-HSD2 transcripts levels in this study as well as previous reports 276 (3,4,11,13). The immediate imbalance between the low 3β-HSD1/2 expression and high tissue concentrations 277 of the 3β-HSD2 catalyzed Δ 4 steroids: progesterone, 17OH-progesterone, and androstenedione, has 278 previously been suggested to be the result of either circulating placental progesterone or the transfer of 279 intermediates of the steroid pathways between the adrenal zones (13) . Since it has been shown by double 280 immunofluorescence staining in 2 nd trimester HFAs, CYP11A1 and CYP21A2 only co-localize in a limited 281 number of FZ cells (13) this could support the hypothesis of transport of steroidogenic intermediates between 282 the adrenal zones. Moreover, the presence of high levels of intra adrenal androstenedione in 2 nd trimester 283 determined in this study favors 3β-HSD2 dependent de novo synthesis, as 17OH-progesterone is a poor 284 substrate for human CYP17A1 (C17,20 lyase) indicating that adrenal virilization of CAH fetuses during nd trimester other human This study provides a detailed characterization of both male and female adrenal endocrine function during 1 st 328 and 2 nd trimester of human fetal development. It is evident from the gene and protein expression pattern of 329 steroidogenic enzymes as well as the steroid measurements in tissue that the HFA functions as a 330 steroidogenic active organ from early development by producing high levels of mineralocorticoids, 331 glucocorticoids and androgens. Even from GW 8 we report a distinct expression pattern for the investigated 332 adrenal steroidogenic enzymes with a significant increase of gene expression in 2 nd trimester samples for the 333 majority of enzymes, with exception of the unaltered expression of 3β-HSD2 and ARK1C3 . Based on the intra- 334 adrenal steroid hormone concentrations we found that the androstenedione is the most abundant Δ 4 adrenal 335 androgen synthesized via the classical steroidogenic pathway throughout 1 st and 2 nd trimester. Additionally, 336 this study confirms that cortisol is produced throughout 1 st and 2 nd trimester, suggesting continued regulation 337 of the HPA axis during this entire period.

1 Total RNA was extracted from one frozen HFA glands in samples from 114 collected from 22 fetuses), while half of a frozen HFA gland was used in samples from GW 14-19 (9 male and 115 8 female, collected from 17 fetuses and isolated using the NucleoSpin RNA II purification kit, according to the 116 manufactures instructions (Macherey-Nagel, Düren, Germany). cDNA was synthesized using a dT20 primer 117 and random hexamers. Real time polymerase chain reaction (RT-PCR) was performed using specific primers 118 targeting preselected mRNAs. All primers were designed to span intron-exon boundaries with optimal 119 annealing temperatures ~62 °C, comparable primer length and CG contents ( Table 1). All amplicons were 120 initially verified by sequencing (Eurofins MWG GmbH, Germany) while primer amplification efficiency and 121 detectable dynamic range of all primer-sets was validated prior to the analysis of the HFA samples.  cycle conditions: one cycle of 3 min at 95°C; 40 cycles of 30 sec at 95°C, 1 min at 62°C, 1 min at 72°C, and 123 one cycle of 5 min at 72°C. Quantitative RT-PCR analysis was performed in triplicates using Brilliant II SYBER 124 Green qPCR Master mix (Agilent technologies). Changes in gene expression were quantified using the 2 -ΔΔCt 125 method (19). The expression levels were normalized to the reference gene, RSP20, and calculated as a ratio 126 with male or female GW 8-9 set to 1 in respective samples. 127 Immunohistochemistry 128 Adrenal tissue from GW 8-12 (13 male and 13 female, collected from 26 fetuses) and GW 14-19 (10 male and 129 6 female, collected from 16 fetuses) were used for immunohistochemistry. Formalin fixed adrenal tissue were 130 dehydrated, paraffin embedded and sectioned (4 µm) using standard procedures. Primary antibodies, dilutions 131 and retrieval buffers are listed in Table 2. Immunohistochemistry (IHC) staining was initially conducted 132 according to a standard protocol as described previously (20). Subsequently, the protocol was modified to 133 include antigen retrieval by pressure cooker as previous described (21). In brief, tissue sections were 134 subjected to heat-induced antigen retrieval buffer in a pressure cooker and endogenous peroxidase was 135 blocked with 3% (v/v) H2O2 in MeOH for 30 min. Between each step sections were washed in Tris-buffered 136 Saline (TBS). Sections were incubated in 5% bovine serum albumin (BSA w/v) in Horse serum (20% v/v 137 ImmPRESS) and TBS (80% v/v) or 0.5% milk powder for 30 min depending on optimization for each antibody. 138 chamber followed by 1h at room temperature. Sections were then incubated for 30 min with the appropriate 140 ImmPRESS HRP (peroxidase) secondary antibody diluted in normal serum. Visualization was performed using 141 ImmPACT AEC peroxidase substrate (Vector Laboratories, Burlingame, CA, USA). Included negative controls 142 replaced the primary antibody with dilution buffer only, none of which showed any staining. All sections were 143 counterstained with Meyer's hematoxylin before mounting with Aquatex (Merck, Damstad, Germany). 144

Quantification of stained cells 145
Two independent investigators evaluated all stainings. Sections were first investigated manually on a Nikon 146 Microphot-FXA microscope, subsequently slides were scanned on a Nano-Zoomer 2.0 HT (Hamamatsu 147 Photonics, Herrsching am Ammersee, Germany) and analyzed using the software NDPview version 1.2.36 148 (Hamamatsu Photonics). The intensity of immunoreactivity was classified according to a pre-defined scoring 149 system: ++, strong staining in all cells of a given type in the sample; ++/+, strong staining prevalent, but some 150 weakly stained cells also visible; ++/-, strong staining present, but negative cells also present; +/++, majority 151 of cells weakly stained, but some strong staining present; +/++/-, heterogeneous pattern with a mixture of 152 strongly positive, weakly stained, and negative cells; +, (1) weak staining overall or (2) strong staining in a 153 small number of cells; +/-, weak staining in limited areas; -/+, weak staining in single cells. 154 Quantification of steroid hormones 155 Tissue samples from intact HFA glands GW 8-12 (10 male and 10 female, collected from 20 fetuses) and 156 halved HFA glands GW 14-19 (9 male and 8 female, collected from 17 fetuses) were included for the intra-157 adrenal steroid analysis by LC-MS/MS measurements. Tissue samples were weighed (1.7 mg -43 mg wet 158 weight) and ground in 1 mL 80% (v/v) MeOH. The homogenate was transferred to glass tubes evaporated to 159 almost dryness under a N2 stream. Internal standard stock solution (5-500 ng/mL in 100 µL) was added to 160 each sample pellet as previously described (22) and subsequently 375 µL 1M ammonium acetate buffer (pH 161 5.5) was added. Next, 2 mL Heptan: Ethylacetat 3:2 (v/v) was added and samples were transferred to 162 eppendorf tubes. Samples were then shaken for 15 min followed by 10 min of centrifugation at 2000 x g (4°C). 163 Each tube was transferred to a dry-ice bath (dry-ice pills in ethanol (99%) for a few minutes to freeze the 7 aqueous phase followed by decantation of the organic phase to a new glass tube. The organic phase was 165 evaporated to dryness under a stream of N2 and finally, the steroids were resolved in an appropriate amount 166 of 50% (v/v) MeOH (tissue GW 8-12: 100 µL, tissue GW 14-19; 200 µL) for LC-MS/MS analysis as previously 167 described (22). All samples were measured in one single batch which includes standards for calibration curves, 168 unknown samples, two blanks and for method control; three un-spiked human serum pool samples and three 169 serum pool samples spiked with low and high levels, respectively. 170

Statistical analysis
171 qPCR and LC-MS/MS data were statistically analyzed for age and sex specific differences. Age differences 172 were tested by the nonparametric Mann-Whitney test in which HFA age groups GW: 10-12, GW: 14-16 and 173 GW:17-19 were compared with male GW 8-9. Sex differences were also tested by nonparametric Mann-174 Whitney test within each age group. P<0.05 was considered statistically significant. 175

176
Gene expression patterns of adrenal steroidogenic enzymes 177 The selected steroidogenic enzymes were expressed in all investigated HFA glands at the transcriptional level. 178 Gene expression patterns were investigated separately in male and female samples, which were divided into 179 four age groups: GW 8-9, GW 10-12, GW 14-16, and GW 17-19. Expression levels were calculated as a ratio 180 of levels relative to male GW 8-9 (reference value set to 1). No sex differences were observed in the The expression profiles of steroidogenic proteins were analyzed using IHC on fixed adrenal samples. No 198 differences between male and female samples were observed regarding the level of expression and the cell-199 type specific enzyme localization. Therefore, only male samples are shown for the age groups GW 8-9, GW 200 10-12, GW 14-16, and GW 17-19 in Figure 3. Adrenal zone-specific protein expression, localization, and the 201 level of expression were evaluated based on a predefined scale and are summarized in Table 3. was also detected throughout the investigated period of fetal development. More specifically, weak 214 cytoplasmic MC2R expression was detected throughout 1 st and 2 nd trimester, but nuclear MC2R expression 215 was also occasionally found in late 1 st trimester and throughout 2 nd trimester. Therefore, the MC2R expression 216 (23), which is most likely an artifact of the antibody. 218 Human fetal adrenal Steroid tissue concentrations 219 The endocrine activity of the HFA was further evaluated by determination of the intra-adrenal steroid 220 concentrations. Thus, steroids were extracted from adrenal gland tissue followed by LC-MS/MS determination 221 of mineralocorticoid, glucocorticoid and androgen levels. Samples were divided according to sex and age (as 222 described previously). All investigated mineralocorticoids, glucocorticoids and androgen metabolites were 223 detected in HFA tissue from GW 8-19. 224

225
The measured concentrations of mineralocorticoid metabolites progesterone and corticosterone were in 226 general equivalent with no significant difference in relation to neither gestational age nor sex. The only 227 exception was a significant age-related increase in female corticosterone levels at GW 17-19 compared with 228 male tissue concentrations at GW 8-9 (Figure 4a). The intra-adrenal levels of glucocorticoids 17OH-229 progesterone was unaltered over the investigated developmental period with no differences between male 230 and female adrenal tissue samples were detected. In contrast, both 11-deoxycortisol, cortisol and cortisone, 231 a significant increase in intra-adrenal steroid concentrations was found at GW 17-19 compared with male GW 232 8-9 (Figure 4b). For 11-deoxycortisol a significant increase in intra-adrenal steroid concentrations was further 233 observed at GW 10-12 and GW 14-16, the latest only in male samples. The only observed sex-specific 234 difference was found in the intra-adrenal cortisone concentrations which were significantly higher in female 235 samples at GW 8-9 compared with age matched males (Figure 4b). Interestingly, cortisol levels were the 236 highest of all measured intra-adrenal steroid hormones throughout the investigated developmental period. 237 Hence, cortisol tissue concentrations were approximately 2-fold higher than those of 17OH-progesterone and 238 11-deoxycortisol and approximately 10-fold higher than cortisone (Figure 4b). adrenals (4,7,8,10,11,13,14). Thus, our data support the prominent hypothesis of the FZ, and later in 253 development also the TZ, being the main site of de novo steroid synthesis in the HFA. Interestingly, the 254 transcription of steroidogenic genes appeared to be regulated differently with unaltered expression of ARK1C3 255 and 3β-HSD2 during the investigated period while the expression of the other steroidogenic enzymes (StAR, 256 CYP11A1, CYP17A1, CYP21A2, CYP11B1/2, and SULT2A1) and the ACTH receptor MC2R were significantly 257 increased in 2 nd trimester. The observed increases in gene expression levels in 2 nd trimester was not evident 258 at the protein level; presumably due to the lack of sensitivity of immunohistochemistry to detect differences in 259 expression level. However, the genes upregulated during 2 nd trimester were also the steroidogenic enzymes 260 that showed the most abundant immunostaining in the FZ while the low constitutively expressed 3β-HSD1/2 261 was only detected in a small subpopulation of cells mainly located in the DZ. Furthermore, the up-regulation 262 in gene expression for the majority of the investigated steroidogenic enzymes was not reflected in higher tissue 263 concentrations of steroidogenic hormones which overall did not differ between 1 st and 2 nd trimester ( Figure  264 5b). It remains to be examined whether the 2 nd trimester transcriptional upregulation of steroidogenic enzymes 265 are reflected in steroid levels secreted by the HFA to the fetal circulation, which were not possible to determine 266 in this study. 267

11
The constitutive expression of 3β-HSD1/2 throughout 1 st and 2 nd trimester of human fetal development in male 269 and female HFA samples is in contrast with the previous reported transient expression profile (3,4,7,8,11). 270 Thus, our results demonstrate that 3β-HSD1/2 is highly expressed in a sub-population of DZ cells from GW 8-271 19, which is in accordance with recent published data (13). Even though the percentage of 3β-HSD1/2 positive 272 cells in DZ and FZ appeared to decrease during 2 nd trimester, this study is the first to report a small number of 273 3β-HSD1/2 positive FZ cells in 2 nd trimester fetuses up to GW 19. In general, the small number of 3β-HSD1/2 274 positive cells compared with the abundant expression of the majority of the adrenal steroidogenic enzymes is 275 consistent with the reported low 3β-HSD2 transcripts levels in this study as well as previous reports 276 (3,4,11,13). The immediate imbalance between the low 3β-HSD1/2 expression and high tissue concentrations 277 of the 3β-HSD2 catalyzed Δ 4 steroids: progesterone, 17OH-progesterone, and androstenedione, has 278 previously been suggested to be the result of either circulating placental progesterone or the transfer of 279 intermediates of the steroid pathways between the adrenal zones (13). Since it has been shown by double 280 immunofluorescence staining in 2 nd trimester HFAs, CYP11A1 and CYP21A2 only co-localize in a limited 281 number of FZ cells (13) this could support the hypothesis of transport of steroidogenic intermediates between 282 the adrenal zones. Moreover, the presence of high levels of intra adrenal androstenedione in 2 nd trimester 283 determined in this study favors 3β-HSD2 dependent de novo synthesis, as 17OH-progesterone is a poor 284 substrate for human CYP17A1 (C17,20 lyase) (24) indicating that adrenal androstenedione is not likely 285 generated from placental progesterone. 286

287
The HFA's produce androgens throughout 1 st and 2 nd trimester in both male and female samples. The 288 detection of high HFA tissue concentrations of androstenedione and low levels of testosterone indicate that 289 this step-in steroidogenesis is tightly regulated. The low intra-adrenal testosterone levels detected from GW 290 8-19 are in accordance with previous studies that have examined both fetal adrenal tissue concentrations and 291 secreted levels in organ culture (3,4,13). Furthermore, the low tissue concentrations of testosterone are also 292 in accordance with the relatively low transcriptional expression level of ARK1C3 which mediates adrenal de 293 novo testosterone synthesis. The expression level of ARK1C3 in 1 st trimester samples detected in this study 294 supports previous investigations (3,4) while ARK1C3 expression levels, to our knowledge, have not been 295 investigated previously in 2 nd trimester HFA tissue. Consequently, the significant increase in male testosterone 296 levels at GW 14-16, compared with male GW 8-9, is inconsistent with the observed overall unaltered 297 expression level of ARK1C3 suggesting that testosterone levels are not only regulated via gene transcription. 298 The observed peak in male testosterone around GW 14-16 is not significantly different from the female GW 299 14-16 age group and it is likely that the adrenal testosterone production is considerably lower than the 300 contribution from fetal testis (25). Moreover, this study is the first to report androstenedione as the most 301 abundant Δ 4 adrenal androgen synthesized by the classical steroidogenic pathway throughout 1 st and 2 nd 302 trimester in HFA, which is in contrast to previous observations of higher adrenal androgen production during 303 the 1 st trimester (13). 304

305
The HFA produces cortisol throughout 1 st and 2 nd trimester indicating continued regulation of the HPA axis 306 during this entire period. In line with this finding is the detection of cytoplasmic MC2R expression from GW 8, 307 which is in accordance to previous studies (3,4). However, due to the unexpected nuclear expression pattern 308 of MC2R observed in this study these results should be interpreted with caution as MC2R is known to be 309 expressed in the cytoplasm (at the endoplasmic reticulum) until its translocated with MRAP to the plasma 310 membrane to mediate ACTH signaling (23).The abundant HFA tissue concentrations of cortisol detected in 311 this study further supports the recent hypothesis that cortisol-mediated negative feedback of the HPA axis is 312 not restricted to the 1 st trimester, but extends through the 2 nd trimester (13) including GW 14-19. Detection of 313 intra-adrenal cortisol in 2 nd trimester is consistent with the observed expression of 3β-HSD1/2 and in 314 accordance with a previous study detecting cortisol in HFA tissue from 2 nd trimester (13) although the cortisol 315 concentrations measured in the present study are approximately 10-fold higher than the concentrations 316 previously reported. During GW 8-14 cortisol mediated negative feedback on the HPA axis is thought to be 317 crucial since this period is the window of development in which the external genitalia differentiates (15,16). 318 Hence, dysregulated adrenal steroidogenesis resulting in imbalance with excess of adrenal androgens and 319 reduced cortisol in this time window can thus cause virilization of female genitalia in fetuses with CAH (17). 320 Since adrenal androgens can be converted to estrogens by the placental aromatase (CYP19A1) later in 321 pregnancy, the adrenal secretion of androgens during 2 nd trimester is thought to have less effect on genital 322 differentiation (15,16)  Antigen retrieval was conducted by pressure cooking of the sections in indicated retrieval buffer for 473 30 min. in a decloaking chamber. TEG buffer: 10 mM Tris, 0.5 mM EGTA, pH 9.0; Citrate (CIT) 474 buffer: 10 mM, pH 6.0. TZ staining is indicated when the TZ protein expression differed from the FZ expression in the 2 nd trimester tissue (GW: 14-19). 480

DZ
Scale: ++, strong staining in all cells of a given type in the sample; ++/+, strong staining prevalent, but some weakly stained cells also visible; 481 ++/-, strong staining present, but negative cells also present; +/++, majority of cells weakly stained, but some strong staining present; +/++/-482 , heterogeneous pattern with a mixture of strongly positive, weakly stained, and negative cells; +, (1) weak staining overall or (2) strong 483 staining in a small number of cells; +/-, weak staining in limited areas; -/+, weak staining in single cells.