CYP21A2 Gene Expression in a Humanized 21-Hydroxylase Mouse Model Does Not Affect Adrenocortical Morphology and Function.

Steroid 21-hydroxylase is an enzyme of the steroid pathway that is involved in the biosynthesis of cortisol and aldosterone by hydroxylation of 17α-hydroxyprogesterone and progesterone at the C21 position. Mutations in CYP21A2, the gene encoding 21-hydroxylase, cause the most frequent form of the autosomal recessive disorder congenital adrenal hyperplasia (CAH). In this study, we generated a humanized 21-hydroxylase mouse model as the first step to the generation of mutant mice with different CAH-causing mutations. We replaced the mouse Cyp21a1 gene with the human CYP21A2 gene using homologous recombination in combination with CRISPR/Cas9 technique. The aim of this study was to characterize the new humanized mouse model. All results described are related to the homozygous animals in comparison with wild-type mice. We show analogous expression patterns of human 21-hydroxylase by the murine promoter and regulatory elements in comparison to murine 21-hydroxylase in wild-type animals. As expected, no Cyp21a1 transcript was detected in homozygous CYP21A2 adrenal glands. Alterations in adrenal gene expression were observed for Cyp11a1, Star, and Cyb11b1. These differences, however, were not pathological. Outward appearance, viability, growth, and fertility were not affected in the humanized CYP21A2 mice. Plasma steroid levels of corticosterone and aldosterone showed no pathological reduction. In addition, adrenal gland morphology and zonation were similar in both the humanized and the wild-type mice. In conclusion, humanized homozygous CYP21A2 mice developed normally and showed no differences in histological analyses, no reduction in adrenal and gonadal gene expression, or in plasma steroids in comparison with wild-type littermates.

Steroid 21-hydroxylase is a cytochrome P450 enzyme that catalyzes the hydroxylation of the adrenal steroid hormones progesterone and 17-hydroxyprogesterone at the C21 position of the steroid biomolecules [1]. It is involved in the biosynthesis of 11-deoxycortisol and deoxycorticosterone, the precursors of the glucocorticoid cortisol and the mineralocorticoid aldo-sterone, respectively. Steroid 21-hydroxylase is localized in endoplasmic reticulum membranes of the adrenal cortex. The human steroid 21‐hydroxylase consists of 495 amino acids with a molecular weight of 52 kDa [2, 3]. The encoding human CYP21A2 gene (MIM*613815) is located at chromosome 6p21.3, spans 3.35 kb, and consists of 10 exons [2, 4]. Homozygous or compound heterozygous mutations in CYP21A2 are the most common cause of congenital adrenal hyperplasia (CAH). A blockade in the steroid biosynthesis pathway leads to an excessive accumulation of steroid precursors, which are converted into adrenal androgens, resulting in an excess of androgens. In female patients with the classic (severe) form of CAH, adrenal androgen excess causes prenatal virilization of the external genitalia. Patients of both sexes have an increased risk for life-threatening adrenal crises throughout their entire life [5]. The lack of cortisol leads to increased adrenocorticotropic hormone (ACTH) secretion by the pituitary gland due to the lack of negative feedback regulation of the hypothalamic-pituitary-adrenal (HPA) axis. Elevated ACTH levels stimulate adrenal growth resulting in adrenal hyperplasia.

The pseudogene (CYP21A1P) is located 30 kb downstream from the functional CYP21A2 gene and has 98% exonic sequence identity and 96% intronic homology to functional CYP21A2 [2, 4]. Although the pseudogene is not functional, it plays a crucial role in the frequency of changes in CYP21A2. Mutations in the pseudogene are often transferred to the functional gene by microconversion events. Intergenic recombinations with the pseudogene are responsible for more than 90% of CYP21A2 mutations associated with 21-hydroxylase deficiency [6, 7]. In the mouse genome, Cyp21a1 is the functional gene and Cyp21a2-p is the pseudogene [8, 9]. The murine and human 21-hydroxylase genes show a high level of homology, with 77.5% DNA identity and 72.5% identity on the protein level (HomoloGene—NCBI, 2021). In addition, both genes are located in the same region within the class III region of the major histocompatibility complex in both species (Fig. 1) [10].

Figure 1.

Schematic diagram of the gene locus of human and mouse 21-hydroxylase. The CYP21A2/Cyp21a1 gene and its related pseudogene CYP21A1P/Cyp21a2-p are located in the major histocompatibility (MHC) locus neighboring 3 other genes which are serine/threonine kinase 19 (STK19/Stk19), complement C4A/C4a and C4B/C4b as well as tenascin-X A and B (TNXA/B, Tnxa/b). The mouse locus on chromosome 17 have the same gene duplication cassette as the human locus on chromosome 6 except the extra pseudogene superkiller viralicidic activity 2-like (Skiv2l-ps1). Pseudogenes are framed in red. Arrows mark the reading direction of the genes.

The first rodent model for 21-hydroxylase deficiency was the H-2aw18 mouse strain [11]. Extensive genetic analysis verified complex gene rearrangement due to an unequal crossing over, which generates a mutant Cyp21a2-p-Cyp21a1 hybrid gene [9]. Unfortunately, these mice die at an early postnatal stage [9, 11, 12]. This lethality can be mitigated by intensive dexamethasone treatment of dams and pups, but the mouse model remains difficult to handle [12]. Another animal model for human CAH is the Cyp11b1 knock-out mouse. These mice exhibit glucocorticoid deficiency, mineralo-corticoide excess, adrenal hyperplasia, mild hypertension, and hypokalemia [13]. Although useful in studying CAH, none of the existing mouse models can be used to test the efficacy of molecules interacting with human CYP21A2 protein for CAH drug development. Therefore, the development of a mouse model with the human CYP21A2 gene is of particular importance.

In this study, we generated and characterized a huma-nized CYP21A2 mouse model with the aim of investigating the effects of CAH-causing mutations in the future. We show that the human 21-hydroxylase enzyme completely replaces the function of the mouse enzyme without reducing the steroid hormone levels, viability, growth, and fertility. This mouse model is therefore an excellent animal model for further studies and for introducing human point mutations to develop novel humanized CAH mouse models.

Materials and Methods

Experimental Animals

All animal experiments were conducted in accordance with accepted standards of animal care as outlined in the Ethical Guidelines of the German Animal Welfare Act and the Directive 2010/63/EU for the protection of animals used for scientific purposes. C57Bl/6NCrl mice were purchased from Charles River (Sulzfeld, Germany). Chimeras and founder CYP21A2 heterozygous mice were generated and bred on a C57Bl/6NCrl genetic background. Breeding and maintenance of CYP21A2 homozygous, heterozygous, and wild-type animals were done at the mouse facility of the Center for Regenerative Therapies Dresden (CRTD), Technische Universität Dresden, Germany. All mice were housed under specific pathogen-free conditions in individually ventilated cages on a 12-hour day/night cycle with free access to food and water. All animal procedures were performed according to institutional guidelines and in accordance with the state authorities (Landesdirektion Sachsen). The stress assessment of the genetically altered line was carried out according to the protocols of the Federal Institute for Risk Assessment (BfR) (https://www.bfr.bund.de/cm/343/beurteilung-der-belastung-genetisch-veraenderter-maeuse-und-ratten-version-2.pdf). The animal experiments were approved by the ethical and research board of the Regierungspräsidium Dresden (reference AZ: 25-5131/449/37—TVV 35/2018).

Generation of Transgenic CYP21A2 Mice C57Bl/6NCrl-Cyp21a1tg(CYP21A2)Koe

Humanized CYP21A2 mice were generated by replacing the 2620-bp Cyp21a1 mouse gDNA sequence with the orthologous 2713-bp CYP21A2 human gDNA sequence using CRISPR/Cas9-mediated gene targeting. RNA guides were designed using Geneious software and the CRISPOR online tool (http://crispor.tefor.net/) and purchased as crRNA (Integrated DNA Technologies Germany GmbH). Gene targeting was performed in mouse embryonic stem (mES) cells JM8A1.N3 (from C57Bl/6NTacBom, Taconic) using the guide crRNAs 5′-TGTGGGGCCAATGGAAGCTC-3′ (5′ side of Cyp21a1) and 5′-GTGAGCGTCCTTGACAGGAT-3′ (3′ side of Cyp21a1) and a homology-directed repair (HDR) donor plasmid pUC57 (GenScript) containing fully humanized CYP21A2 sequence (both exons and introns) flanked by 1000 bp 5′ and 3′ Cyp21a1 mouse homology arms. The construct contains several silent point mutations in CYP21A2 to create protospacer adjacent motif (PAM) sequences at the sites where specific point mutations will be inserted in the future. A BamHI restriction site was added in intron 7 to allow restriction-ligation cloning of the loxP-site flanked PGK-NeoR selection cassette. The Cas9, guide RNAs, and HDR donor were electroporated into mES cells using a Neon Transfection System (Thermo Fischer). Forty-eight hours after electro-poration, mES cells were selected with G418. Genomic DNA from resistant colonies was tested for correct replacement of the mouse Cyp21a1 by the human CYP21A2 gene with 2 polymerase chain reaction (PCR) using primers located upstream of the 5′ and downstream of the 3′ homologous region in combination with one internal primer each (5′-CAGGTAGAGCAGATGACTAG-3′ with 5′-CAATGGTCCTCTTGGAGTTC-3′ and 5′-GTGTCATCCTCAAGATGCAG-3′ with 5′-CTCTGGGGTTTGAATTGGG-3′). Positive clones were PCR-verified for exclusion of random integration of the donor plasmid backbone with ampicillin sequence-specific primers. Correct clones were introduced into C57Bl/6NCrl 8-cell stage embryos by laser-assisted ES-cell injection followed by transfer into pseudopregnant donor mice. Germ line transmission was obtained by mating male chimeras with C57Bl/6NCrl wild-type females. Genotypes were determined by multiplex PCR using the following primers: for the wild-type allele, forward primer (5′-CCAAGACCAGGGTGAGCGT-3′); for the CYP21A2 allele, forward primer (5′-GTGTCATCCTCAAGATGCAG-3′) in combination with reverse primer (5′-CTCACACCCCAGTAGAGAAG-3′) for both alleles. Animals carrying the CYP21A2 allele were crossbred with Cre-deleter mice (PGK-Cre [N]). Resulting offspring that no longer carried the selection cassette in intron 7 were then intercrossed to generate homozygous CYP21A2 mice (see Fig. 2).

Figure 2.

Generation of the transgenic CYP21A2 mouse model. Schematic presentation of homologous recombination of the targeting vector carrying the human CYP21A2 sequence (dark blue) flanked by 1 kb mouse homology arms (light blue), with the endogenous Cyp21a1 gene locus (light blue): wild-type Cyp21a1 gene locus (top), targeting HDR donor plasmid containing fully humanized CYP21A2 sequence and a PGK-Neo selection cassette flanked by LoxP sites (middle with twice zoom-in), and transgenic gene locus (bottom) after homologous recombination and LoxP sites excision by Cre recombinase, leaving a single LoxP as a genomic scar.

Blood Pressure Measurement

Under light isoflurane (0.5%) anesthesia, blood pressure was recorded by the tail-cuff method as described previously [14], using the NIBP System for mice (ADInstruments, Oxford, UK) connected to a PowerLab 4/16 system.

Mouse Specimen Collection

100 µL blood was collected from 4-week and 8-week-old mice from the retrobulbar venous plexus into a heparin tube and the plasma separated and stored at −80 °C for further analysis. At an age of 10 weeks, mice were anesthetized intraperitoneally with ketamine-xylazine solution followed by a final cardiac puncture for blood collection. Cervical dislocation was used to ensure euthanasia after bleeding out. Adrenal glands, testis, and ovaries were harvested and grossly examined. All organs were snap-frozen in liquid nitrogen for gene expression and kept at −80 °C before analysis or fixed in 4% neutral-buffered formalin for histology.

Plasma Steroid and ACTH Measurement

Plasma steroid levels were determined by liquid chromato-graphy–tandem mass spectrometry (LC-MS/MS) as previously described [15]. ACTH levels were evaluated in mouse plasma by luminescent immunoassay kit (Siemens Healthcare Diagnostics Cat# LKAC1, RRID:AB_2909441, https://scicrunch.org/scicrunch/data/source/nif-0000-07730-1/search?q=AB_2909441&l=AB_2909441) at IMMULITE 1000 system according to the manufacturer’s protocol.

Histological Analysis

Two pairs of male and female wild-type (n = 4: 2 male, 2 female) and CYP21A2 homozygous mice (n = 4: 2 male, 2 female) were sacrificed to obtain different organs for tissue sections. After dissection, all tissues were fixed in 4% form-aldehyde overnight and embedded in paraffin at the Histology Facility of the BIOTEC Dresden, Germany. Four-micrometer sections of paraffin-embedded mouse adrenal, ovary, and testis specimens were placed on glass slides and stained with hematoxylin and eosin (H&E). Images were captured on a Keyence BZ-X700 microscope (Keyence Corporation of America, Itasca, USA) using identical camera settings and analyzed for histological appearance.

Immunohistochemical Analysis

For immunohistochemistry (IHC), 4-µm sections of paraffin-embedded tissues were dewaxed and treated with citrate buffer for 30 minutes in a microwave oven to demask the binding sites. After washing with phosphate-buffered saline (PBS), tissues were treated with 3% H2O2 in methanol for 15 minutes to quench endogenous oxidases. Tissue sections were blocked with 3% normal goat serum, 0.1% bovine serum albumin, and 0.3% Triton X in PBS for 1 hour and probed with the primary antibody rabbit anti-human CYP21A2 (Sigma-Aldrich Cat# HPA053371, RRID:AB_2682131, 1:2500, https://antibodyregistry.org/search.php?q=AB_2682131 in blocking buffer) at 4 °C overnight. After washing with PBS, the sections were incubated with the peroxidase-labeled anti-rabbit secondary antibody using Histofine Simple Stain Mouse MAX PO (Nichirei Biosciences Inc., Tokyo, Japan, Cat# 414341F, RRID:AB_2819094, https://antibodyregistry.org/search?q=AB_2819094) for 30 minutes followed by HistoGreen detection, counterstaining with nuclear fast red and mounted with Pertex mounting medium. The procedure was repeated 2 times with a negative control to confirm the specificity of the staining. Images were made with the Keyence BZ-X700 microscope (Keyence corporation of America, Itasca, USA) followed by colorimetric detection of the antibody signal.

Immunofluorescence Staining

Adrenal glands were fixed (4% PFA, 4 hours), cryoprotected (30% sucrose in PBS, 4 °C overnight), embedded in Tissue-Tek Medium (OCT; Sakura Finetek), and stored at −80 °C. Cryosections were prepared using a CryoStar NX70 Cryostat (Thermo Fisher Scientific) and sliced to 10 µm thickness (Superfrost slides; Thermo Scientific). Sections were then immunostained using specific antibodies against steroidogenic factor-1 (SF-1, rabbit, TransGenic Inc, Cat# KO611, RRID:AB_2861370, https://antibodyregistry.org/search?q=AB_2861370), steroidogenic acute regulatory protein (StAR, rabbit, SantaCruz Biotechnology Cat# sc-25806, RRID:AB_2115937, https://antibodyregistry.org/search?q=AB_2115937), and tyrosine hydroxylase (TH, rabbit, Millipore Cat# AB152, RRID:AB_390204, https://antibodyregistry.org/search?q=AB_390204). Primary antibodies were incubated on sections overnight at 4 °C. Sections were then washed and incubated for 2 hours at room temperature with the secondary antibody (Cy3-goat-anti-rabbit; Jackson ImmunoResearch Labs Cat# 111-165-144, RRID:AB_2338006, https://antibodyregistry.org/search?q=AB_2338006). Nuclei were stained with 4′-6-diamidino-2-phenylindole (DAPI; Thermo Fisher Scientific) and slides were mounted with fluorescent mounting medium (Aqua-Poly/Mount; Polysciences). Fluorescence microscopy was done either with a Zeiss Axio Scan.Z1 widefield slide scanner for whole adrenal scan or with the confocal laser scanning microscope Zeiss LSM 780 (Zeiss, Oberkochen, Germany) and ZEN 3.1 (Zeiss) software.

Total RNA Extraction and Quantitative Reverse Transcription PCR

Total RNA was extracted from mouse adrenal glands, ovaries, and testis (n = 7 of each group) using a NucleoSpin RNA II kit (Macherey-Nagel, Düren, Germany) according to the manufacturer’s protocol. Purity of the RNA was assessed using Nanodrop Spectrophotometer (ND-1000) (NanoDrop Technologies, Wilmington, DE, USA). 500 ng of total RNA were reversely transcribed into cDNA using GoScript Reverse Transcription System (Promega, Mannheim, Germany) according to the manufacturer’s protocol. Primers for amplification of the specific genes encoding steroidogenic enzymes (hCYP21A2, Star, Cyp11a1, Cyp17a1, Cyp11b1, Cyp11b2, Hsd3b2) and of the gene encoding steroidogenic factor-1 (Nr5a1) were designed with Primer Express 3.0 and obtained from Eurofins Genomics. A TaqMan Gene Expression Assay was purchased (Mm00487230_g1, ThermoFisher) for analysis of mouse Cyp21a1. Real-time PCR was performed using the GoTaq Probe qPCR MasterMix (Promega, Mannheim, Germany) according to the manufacturer’s reaction parameters. All samples were assayed in triplicates and quantitative reverse transcriptase PCR was performed using the Quantstudio 5 (Thermo Fisher Scientific Inc.). For sample normalization, the housekeeping gene β-Actin was selected as a reference gene. The delta-delta threshold cycle (ΔΔCt) method was used to determine the fold changes in mRNA expression levels [16].

Statistical Analysis

All data are presented as mean ± SD. All statistical analyses were performed using GraphPad Prism version 6 for Windows (GraphPad Software, La Jolla, CA, USA). Data were analyzed using the Mann-Whitney U test or Student’s t test with P values ≤ 0.05 considered to be statistically significant for comparison between male and female animals and wild-type control vs transgenic homozygous (h-Hom) line.

Results

Generation of the Humanized CYP21A2 Mouse C57Bl/6NCrl-Cyp21a1tg(CYP21A2)Koe

Cas9-mediated genome editing via HDR in mES cells was used to generate humanized mice in which the murine Cyp21a1 gene has been replaced with the human CYP21A2 gene (Fig. 2). The mES cells containing CYP21A2 were screened for correct integration of the human gene and the exclusion of random off-target integrations and subsequently introduced into 8-cell embryo stages by laser-assisted ES-cell injection. Resulting chimeric male mice were mated with C57Bl/6NCrl wild-type (WT) females to produce heterozygous hCYP21A2 pups. Heterozygous hCYP21A2 mice were intercrossed with a PGK-Cre deleter strain in order to remove the PGK-Neo-resistance cassette and thus to yield homozygous hCYP21A2 mice. Hence, the established mouse strain, designated according to the international nomenclature as C57Bl/6NCrl-Cyp21a1tg(CYP21A2)Koe and also referred to as h-Hom, is a humanized CYP21A2 transgenic mouse line.

h-Hom Mice Show No Alteration in Viability, Growth, and Fertility

The genotype frequencies correspond to Mendelian expectations and are in line with Hardy-Weinberg equilibrium for this mouse population. Intercross of heterozygotes resulted in an approximately 1:2:1 distribution of the genotypes (1.2 WT:2 Het:0.8 Hom; n = 18 litter; litter size 6.7). Homozygous hCYP21A2 mice (h-Hom) were generally characterized regarding their viability, growth, and fertility and then phenotypically characterized by biochemical and histological analyses. Offspring of h-Hom matings showed no abnormalities in severity assessment of the genetically altered line in the first weeks after birth and had normal litter sizes of an average of 6 pups per litter (range, 4-8 [n = 9]) with a normal sex distribution (46% male, 54% female) and 96.5% survival. The body weight of 10-week-old male and female h-Hom mice were compared with the body weight of 10-week-old WT mice and no significant differences were observed between the 2 animal groups (Fig. 3). Male h-Hom mice had a mean body weight of 26.90 g (± 2.51 g) and WT mice of 25.25 g (± 1.42 g). As expected, male mice were heavier than female mice, which had an average weight of 20.3 g (± 1.83 g) for h-Hom and 21.51 g (± 1.21 g) for WT mice. In the blood pressure measurements of h-Hom and WT females, no significant differences were observed. However, there was noticeable a trend toward higher blood pressure values in h-Hom females compared with WT animals (see Supplementary Figure S1 in [17]).

Figure 3.

Body weights of 10-weeks old male and female mice. There is no significant difference between WT and h-Hom mice using Student’s t test. Values represent each data point for one mouse with mean; n = 10.

Effect of Humanization of Cyp21a1 on Steroidogenic Enzyme Expression Levels

The expression of steroidogenic enzymes in the adrenals of 10-week-old mice was analyzed to verify that the expression pattern is retained in h-Hom mice. As expected, h-Hom mice showed no expression of the Cyp21a1 gene at all (Fig. 4A), while showing CYP21A2 expression (Fig. 4B). As anticipated, the expression in WT mice was inverse. They showed an adequate Cyp21a1 expression level, females even higher than males (P = 0.007), whereas CYP21A2 was not expressed. We also analyzed the integrity of the full-length cDNA of CYP21A2 by reverse transcriptase PCR and proved that the integration of silent point mutations for generation of optimal PAM sites used later for insertion of specific point mutations did not affect the splicing (see Supplementary Figure S2 in [17]). Cyp11a1 and Star were significantly higher expressed in male h-Hom mice compared to male WT animals (p = 0.02) but did not differ in females (Fig. 4C + D). In female mice, only Cyp11b1 (P = 0.007) was higher expressed in h-Hom animals than in WT animals (Fig. 4E). Compared with the other adrenal enzymes, a low gene expression for Cyp11b2 was observed, which did not differ between groups (Fig. 4F). The same low expression was shown for the gene Nr5a1, and the only significant difference was between WT mice, where males had a lower expression than females (Fig. 4H). Gender differences were also seen in the expression of Cyp11a1. Male mice, both h-Hom and WT, had a higher expression than female mice. Star expression in the h-Hom group was 2-fold higher in males than in females. In contrast, gene expressions of Cyp11b1 and Hsd3b2 were significant lower in h-Hom males in comparison with h-Hom females.

Figure 4.

Adrenal gene expression in h-Hom CYP21A2 mice compared to WT mice. Expression of Cyp21a1 (A), CYP21A2 (B), Cyp11a1 (C), Star (D), Cyp11b1 (E), Cyp11b2 (F), Hsd3b2 (G) and Nr5a1 (H) in adrenal glands from male and female 10-wk-old WT and h-Hom mice was determined by qRT-PCR. Relative mRNA expression levels were normalized to β-actin expression (Actb). Experiments were performed in triplicates, and data are expressed as means with SEM. Statistical significance was determined by Mann-Whitney U test, *P < 0.05, **P < 0.01 and ***P < 0.001; n = 7.

Ovary gene expression showed no significant differences in female animals of the h-Hom group in comparison with the WT group. The h-Hom and WT mice presented a similar expression pattern for Cyp17a1, Cyp11a1, Star, and Hsd3b2 (see Supplementary Figure S3 in [17]). The same was seen for the expression of these 4 genes in the male testis (see Supplementary Figure S4 in [17]). As expected, Cyp17a1 was only detectable in the ovaries and testis, not in the adrenals of mice, while Cyp21a1, Cyp11b1, and Cyp11b2 were only detectable in the adrenals, not in the gonads. In general, the investigated gene expressions in the gonads were very low. In addition, CYP21A2, similar to Cyp21a1, was scarcely expressed neither in h-Hom mouse ovaries nor testis.

Effect of Humanization of Cyp21a1 on Blood Plasma Steroids and ACTH Levels

Steroid plasma measurements were performed on 4-, 8- and 10-week-old male and female mice in both the h-Hom and WT groups (see Supplementary Figure S5 in [17]and Fig. 5). Aldosterone levels at the ages of 4 and 8 weeks showed no significant difference between the 2 groups examined, neither between h-Hom and WT mice nor between male and female (see Supplementary Figure S5 in [17]). However, 12 of the 4-week-old examined mice (n = 31) and 10 of the 8-week-old mice (n = 33) had aldosterone levels below the detection limit because plasma volumes were too low (see Supplementary Figure S5A in [17]). The aldosterone levels of 10-week-old mice were higher than in young animals, with no difference between h-Hom and WT (Fig. 5F). Although the concentrations in WT animals did not differ significantly between male and female (0.38 ± 0.17 ng/mL vs 0.37 ± 0.14 ng/mL), female h-Hom mice had significantly higher aldosterone concentrations than males (0.48 ± 0.18 ng/mL vs 0.29 ± 0.1 ng/mL).

Figure 5.

Steroid hormone levels in 10-week-old h-Hom CYP21A2 mice compared with WT mice. Diagram of steroidogenesis pathway in mice (A). Plasma progesterone (B), 11-deoxycorticosterone (C), corticosterone (D), 18OH-corticosterone (E), aldosterone (F), and testosterone (G) levels were determined by LC-MS/MS. Values represent each data point for one mouse (male, blue dots; female, red dots) with mean (black crossbar). Statistical significance was determined by Mann-Whitney U test, *P < 0.05, **P < 0.01 and ***P < 0.001; n = 10.

In 4- and 8-week-old mice, h-Hom mice did not differ significantly from the WT group regarding corticosterone levels (see Supplementary Figure S5B in [17]). Compared with younger mice, corticosterone levels were increased with age. In 10-week-old mice, there was a significant difference between female h-Hom and female WT animals (P = 0.02) (Fig. 5D). In addition, in both groups (h-Hom and WT), female mice had also significantly higher corticosterone concentrations than male mice.

Blood plasma testosterone concentrations in male mice increased with maturity at the age of 8 weeks in both h-Hom and WT mice. In plasma, 8- and 10-week-old mice had a mean concentration of 3 to 5 ng/mL testosterone, while 4-week-old animals had a mean level of 0.8 ng/mL. No significant difference could be observed between h-Hom and WT groups (see Supplementary Figure S5C in [17] and Fig. 5G).

No additional steroids could be measured in 4- and 8-week-old mice because plasma sample volumes were low, and concentrations were below detection level. However, in the 10-week-old mice, progesterone, 18OH-corticosterone, and 11-deoxycorticosterone levels could be analyzed. The greatest disparity was observed for progesterone levels (Fig. 5B). Both h-Hom males and females showed higher proge-sterone levels in comparison with WT mice (males P < 0.001 and females P = 0.004). Moreover, there were significant differences between male and female mice in both the WT and h-Hom groups. Females had significantly lower progesterone concentrations than males.

For 18OH-corticosterone, a significant difference (P = 0.01) was seen between female h-Hom group and female WT group (Fig. 5E). However, no difference was found between the 2 male groups. Blood plasma 18OH-corticosterone concentrations of both sexes of h-Hom mice were also significantly different, in that male mice (1.2 ± 0.1 ng/mL) had lower concentrations than female mice (2.4 ± 0.2 ng/mL). Female h-Hom mice had substantially higher 18OH-corticosterone concentrations than all the other groups. Similarly, female h-Hom mice had the highest 11-deoxycorticosterone concentrations (8.2 ± 1.2 ng/mL), which was significantly higher than in female WT mice (P = 0.009) and also much higher than in male h-Hom mice (P = 0.0003) (Fig. 5C). However, no significant difference was seen between the male groups.

ACTH levels of h-Hom mice (n = 8) varied from 15.1 to 121 pmol/L (mean 61.7 ± 38.4 pmol/L). WT mice had ACTH levels from 16.7 to 59.2 pmol/L (mean 34.9 ± 17.0 pmol/L; n = 8). Compared with the WT mice, the mean ACTH value of the h-Hom was increased nonsignificantly by a factor of 1.77.

Replacement of Cyp21a1 by CYP21A2 Has No Influence on Organ Morphology

Histological investigations were performed to identify abnormalities in morphology and in cell structure of the inner organs of 10-week-old h-Hom CYP21A2 mice compared with WT mice.

Adrenal gland morphology and zonation of h-Hom mice were similar to WT mice as shown by H&E staining (Fig. 6A-6D). The medulla and cortex of the adrenal could be identified, and no gross histological differences were observed between genotypes. Investigation of medulla specific marker on adrenal cryosections confirmed that the medulla is formed normally in h-Hom mice and that there are no abnormalities and that a sharp border exists between cortex and medulla (Fig. 7A). In addition, immunofluorescence studies were performed to demonstrate protein expression of the steroidogenic enzyme Star and factor Sf-1 in the adrenal glands of genetically engineered mice. No differences were visible between h-Hom and WT sections (Fig. 7B-D).

Figure 6.

Histological and immunohistochemical staining of organ sections of h-Hom CYP21A2 mice compared with WT mice at 10 weeks of age. Representative H&E-stained sections of adrenal (A-D), testis (E + F) and ovary (G + H) from h-Hom and WT mice. Representative images of adrenals were obtained from female animals. No gross anatomical differences were observed at low (40×) or high (200×) magnification in h-Hom animals. For specific staining of human CYP21A2, adrenal sections from h-Hom (I) and WT female mice (K) were incubated with anti-human CYP21A2 antibody and visualized by Universal Immuno-peroxidase Polymer method (Nichirei Biosciences Inc.). Note the positively stained cells in the cortex (cyan staining) in the h-Hom mice (I). Cells were counterstained with nuclear fast red staining the nuclei in pink. Scale bar: 50 μm (A,B,E-K) and 200 µM (C-D).

Figure 7.

Immunofluorescence staining of cryosections of h-Hom CYP21A2 mice compared to WT mice. For specific staining adrenal cryosections from h-Hom and WT mice were incubated with anti-tyrosine hydroxylase antibody (A), anti-Sf-1 antibody (B + C) or anti-Star antibody (D) and visualized by a secondary Cy3-fluorescent labeled antibody (red). Nuclei were stained with 4′-6-diamidino-2-phenylindole (DAPI; blue). Scale bar: 200 µM (A, B) and 50 μm (C, D).

Morphological comparison of testis and ovaries from h-Hom and WT mice revealed no irregularities assuming a normal development of the humanized mice. Testis from h-Hom and WT males were histologically similar, showing normal seminiferous tubules with spermatozoa in the lumen (Fig. 6E-6F). Ovaries from female h-Hom mice appeared normal, containing numerous follicles in various stages of development as seen in wild-type tissue sections (Fig. 6G-6H).

Immunohistochemical techniques were used to distinguish the humanized adrenals from wild-type adrenals and to verify the correct integration and expression of the human CYP21A2 gene into the mouse genome. Using an anti-human CYP21A2 antibody, the human 21-hydroxylase was visualized within the adrenal cortex of h-Hom mice (Fig. 6I). As expected, no 21-hydroxylase staining was detectable in the adrenal tissue of WT animals (Fig. 6K). Nuclei and background were counterstained with nuclear fast red, and they had a pink appearance in both h-Hom and WT sections.

Discussion

We established the mouse strain C57Bl/6NCrl-Cyp21a1tg(CYP21A2)Koe, which is a humanized CYP21A2 transgenic mouse line with a targeted replacement of the mouse Cyp21a1 gene by the human CYP21A2 gene in mice with a C57Bl/6NCrl genetic background. The aim was to create a model to study the effects of human CAH-causing mutations in the future. In the absence of Cyp21a1 expression the human CYP21A2 gene exhibits full gene function. CYP21A2 expression leads to a humanized 21-hydroxylase and does not grossly affect the synthesis of adrenal steroid hormones. ACTH levels in h-Hom mice were not significantly higher than in WT animals. However, the mean values of WT and h-Hom are in the range of the wild-type sample from the study of Markmann et al, which showed a 4-fold increase of ACTH in 21OH-/- mice [18]. Minor differences in steroid hormone levels are not pathological and range from significant increase of progesterone concentrations in both sexes of the humanized CYP21A2 transgenic mouse line and significant increase of 11-deoxycorticosterone, corticosterone, and 18-hydroxycorticosterone only in the female mice to nonsignificant changes of aldosterone in comparison with wild-type animals. Consistent with the 7-fold progesterone increase in male h-Hom mice, the enzyme expression of cholesterol side-chain cleavage enzyme (Cyp11a1) and Star is significantly increased in male h-Hom compared with male wild-type animals, whereas all other enzyme expressions measured (Hsd3b2, Cyp11b1, and Cyp11b2) did not change. Star regulates the rate-limiting step in the production of steroid hormones, which is the transfer of cholesterol from the outer to the inner mitochondrial membrane. Cyp11a1 is a mitochondrial enzyme associated with the inner mitochondrial membrane and catalyzes the first reaction in the process of steroidogenesis, the conversion of cholesterol to pregnenolone. The next step from pregnenolone to progesterone is catalyzed by 3beta-hydroxysteroid dehydrogenase (Hsd3b2), which showed no changes in enzyme expression between the groups. However, there was a significant decrease in Hsd3b2 enzyme expression in males compared to females of the h-Hom group, which is in contrast to the higher progesterone levels in the h-Hom males compared to h-Hom females. These findings suggest that not only the level of enzyme expression plays a role for steroid production, but also enzyme activity and the presence of sufficient precursor molecules.

The most prominent difference between h-Hom and WT mice was observed in progesterone levels, in females as well as in males, as plasma progesterone was significantly increased in h-Hom mice. The disparity between h-Hom and WT females could be explained by different estrus cycle phases of the females. In mice, the estrus cycle is divided into 4 phases, proestrus, estrus, metestrus, and diestrus, and repeats every 4 to 5 days [19]. For hormonal analyses, it is substantial to examine female mice being in the same cycle as different phases of the estrus cycle have different impact on the progesterone level in the blood. In a previous study, it was demonstrated that the serum levels of progesterone were highest in metestrus and lowest in diestrus in 3- to 4-month-old female C57BL6 mice [20]. These cyclical changes in serum levels of progesterone could be one explanation for the significant differences between h-Hom and WT females. Therefore, the estrogen levels are important to determine in advance and to identify the stage of estrus cycle should be considered for future studies on these mice.

There is also a sex dimorphism with regard to the effect of the CYP21A2 transgene on circulating corticosterone and the precursors 11-deoxycorticosterone and 18OH-corticosterone with significant increases only in female h-Hom animals compared with wild-type females, but not in males, which we cannot explain yet. The kinetics of the 21-hydroxylase may play a role. At least the end product corticosterone is not lowered, as would be expected in the case of an enzyme defect. Therefore, no impact on the usefulness of the model for studying the effects of CYP21A2 mutations is expected.

Corticosterone is the major glucocorticoid in rodents and has the highest blood plasma concentration of all steroids produced by the adrenal gland. It regulates metabolism and has roles in regulating the immune system and stress response. Aldosterone is the major mineralocorticoid; it plays a central role in the regulation of sodium and potassium levels in plasma and is involved in the homeostatic regulation of blood pressure. In our study, most prominent differences in steroid hormone concentrations were measured not between WT and the humanized hom group, but between the male and female mice. This phenomenon is already known. In rodent studies, it has been shown that females have higher circulating corticosterone and ACTH levels than males [21]. Sex hormones in rodents affect the HPA axis at different levels. This includes the activation of gene expression at the hypothalamic paraventricular nucleus, the regulation of gene expression of the ACTH precursor protein proopiomelanocortin in the anterior pituitary, and a negative feedback role via circulating glucocorticoids in these central organs [22]. Androgen deficiency due to orchiectomy increases basal and stress-induced corticosterone and ACTH levels in the blood, while ovarian resection has the opposite effect [23]. Conversely, sex hormone replacement can reverse these effects [24]. This points to the conclusion that estrogens stimulate the HPA axis while androgens inhibit its activation.

For its catalytic activity, 21-hydroxylase depends on electron transfer from its redox partner cytochrome P450 oxidoreductase (POR). In the presented mouse model human 21-hydroxylase has to interact with mouse POR which could have significant impact on the catalytic activity. However, the conserved domains for interaction with the cytochrome P450 enzymes like NADPH-dependent FMN reductase domain and ferredoxin reductase (FNR) domain are highly conserved between human and mouse POR (Supplementary Figure S6 [17]). Mus musculus POR (NP_032924) shows 92% identity (624/678 AA) and 96% positives (653/678 AA) to Homo sapiens POR (NP_000932) (Supplementary Figure S6C [17]). Whether the mouse POR influences the catalytic activity of the human 21-hydroxylase in the mouse organism has to be further investigated.

Humanized mice have become important preclinical tools for biomedical research. In the early 2000s, the development of immunodeficient mice bearing mutations in the IL-2 receptor common gamma chain (IL2rgnull) was a key breakthrough in drug research. Since then, immunodeficient mice engrafted with functional human cells and tissues have become increasingly important as small preclinical animal models for the study of human diseases [25]. Moreover, transgenic mice expressing antibody-coding human gene sequences have proven to be useful for generation of high-affinity human sequence-specific monoclonal antibodies against a wide variety of clinical indications in cancer, autoimmune or inflammatory diseases, and infectious diseases [26]. Despite existing CAH models, the importance of a humanized mouse model for the research of human specific treatment options in CAH is evident, and we aim to support this research with our mouse line.

All in all, the generated humanized CYP21A2 knock-in mouse model can be utilized to provide additional information on the functional significance of the CYP21A2 gene. On the basis of clinical information from patients with CAH, these humanized mice can also be useful for translation into the clinic [25, 27]. The future goal is to integrate naturally occurring and clinically relevant mutations that show residual enzyme activity and a phenotype with the simple virilization form of CAH in humans [28]. We therefore will focus in the future on 2 point mutations that cause different severity of symptoms in patients with CAH: (i) c.518T>A p.Ile173Asn, causing a simple virilizing form with lower than 2% enzymatic activity; and (ii) c.1451G>A p.Arg484Gln, which results in a simple virilizing form with 1% to 4% enzymatic activity [29, 30]. With the integration of one of these different mutations into each mouse strain, a viable mouse model is expected based on the residual enzyme activity. Targeted insertion of point mutations can better mimic the clinical phenotypes of affected CAH patients than traditional transgenic mice. Moreover, a careful phenotypic characterization of these different mouse strains, in which an adrenal phenotype might be discovered, can serve as a starting point for further functional analysis [10]. One fact to consider of the future CAH mouse models is that they will not be able to replicate the excessive secretion of androgens observed in CAH patients. In contrast to the human adrenal gland, the mouse adrenal cortex cannot produce androgens (DHEA) because it lacks the enzyme steroid 17α-hydroxylase (Cyp17a1). Thus, androgen production in mice is restricted to the gonads. After the establishment and characterization of the new transgenic mouse models, it is intended to use them further as an in vivo test system for potential treatments of CAH, for example for newly developed antibodies, peptides, and biological active chaperones. Overall, humanized homozygous CYP21A2 mice are not only a very powerful model to study 21-hydroxylase function and CAH metabolic features, but clinical applications could also be easier to anticipate in the future.

With the characterization of the humanized 21-hydroxylase mouse model, we show that the human 21-hydroxylase replaces the function of the mouse 21-hydroxylase enzyme and catalyzes the sequential step from 11-deoxycorticosterone to corticosterone in the steroidogenic pathway without reducing the steroid hormone levels. Homozygous C57Bl/6NCrl-Cyp21a1tg(CYP21A2)Koe mice did not show any impairment of viability, growth, and fertility. Therefore, they are an excellent model for introducing human point mutations and studying human 21-hydroxylase enzyme activity.