Genetic Findings in a Cohort of Patients with Androgen Insensitivity Syndrome

Background: Androgen insensitivity syndrome (AIS) is a rare X-linked recessive inherited disorder caused by mutations in AR, a gene encoding androgen receptor. The aim of this study was to expand genetic spectrum of AIS. Methods: We performed a retrospective study on consecutive patients diagnosed as AIS from 2010 to 2020 in a single tertiary center. Variant analysis of AR gene was performed by PCR-Sanger sequencing. The pathogenicity of novel variants was evaluated by dual-luciferase reporter assay and immunouorescence of AR protein in vitro. Results: A total of 19 unrelated 46,XY patients were enrolled, 14 with complete insensitivity syndrome (CAIS) and 5 with partial insensitivity syndrome (PAIS). We identied 19 AR variants: 12 (63.2%) were missense variants and 7 variants (36.8%) resulted in premature stop codon. Eight AR variants were novel, including P15Afs*69, S258Efs*47, W435Gfs*44, C560F, C577W, C580Afs*46, K718X and V819G. Dual-luciferase reporter assay found residual transcription activity of approximately 1% in six novel variants, which may explain the CAIS phenotype. The AR mutant protein (V819G) showed transcription activity of 59%, consistent with mild clinical features in one PAIS patient. Interestingly, another AR mutant protein (K718X) related to CAIS showed increased transcription activity but impaired nuclear localization. Conclusions: We identied eight novel AR variants related to AIS. Both residual transcription activity of AR and nuclear localization of AR protein were associated with the severity of AIS. patients[22]. However, different from the previous studies which found that exon 5 and exon 7 were commonly involved exons in AR, most variants in our study located on exon 1, exon 2 and exon 5, each accounting for 21.1% (4/19). This may be due to the high proportion of CAIS enrolled, as Batista et al said, defects in NTD (mainly encoded by exon 1) are more frequent in CAIS patients[22]. Four variants (A749D, A844E, G569W, A597T) have been reported to be associated with partial AIS which is consistent with phenotype presentation in our study[23–26]. A novel AR variant V819G on ligand-binding domain was identied in case 18 with PAIS. It is reported that most missense mutations involved in LBD are related to partial AIS[27]. Ten variants (52.6%) resided in ligand-binding domain which has been reported to be the most frequently involved domain.


Nuclear localization of the AR mutant protein K718X
To further explore the underlying cause of CAIS phenotype in case 8, nuclear localization normality of the AR variant K718X was validated. COS-1 cells were transiently transfected with three different plasmids (pCMV-GFP-empty, pCMV-GFP-AR WT or pCMV-GFP-AR K718X ) separately. As seen in Fig. 3, under the stimulation of 100nM DHT, the pCMV-GFP-AR WT translocated into the nucleus from cytoplasm. However, the pCMV-GFP-AR K718X plasmid still existed in the cytoplasm and accumulated around the nucleus after DHT stimulation. Therefore, the AR mutant protein K718X resulted in impaired AR function through abnormal nucleus localization.

Discussion
In this study we reported 19 unrelated patients with androgen insensitivity syndrome whose phenotype was in relation to the nature of the speci c AR variant. A total of 19 AR variants were identi ed. Eight variants were rstly reported. Six novel AR variants showed seriously impaired transcription activity approximately of 1%. One AR mutant protein (V819G) presented mild AR-de cient activity of 59%. Interestingly, another AR mutant protein (K718X) showed increased transcription activity but impaired nuclear localization. The current study broadened the genetic spectrum of androgen insensitivity syndrome which may facilitate in the real clinical-setting for individualized medical approach.
During embryonic development, sex differentiation of male fetus required androgen secreted by Leydig cells to act on normal androgen receptor to exert its biological effects. Subsequently, the wol an duct developed, male external genitalia differentiated and testes gradually descended to the scrotum. Phenotype spectrum of AIS patients is broad and closely related to the severity of AR impairment, ranging from typical female genitalia at birth to male hypospadias, micropenis, cryptorchidism, oligospermia and gynecomastia. In our study, half of the CAIS patients (50%) came for consultation due to inguinal swelling before puberty which is easily overlooked but an important clue for early diagnosis of AIS. Like Costagliola's results [10], the median age of genetic diagnosis for CAIS patients in our study was delayed compared with the age of rst symptom occurrence (18.5 [16-22] vs 14 [IQR 1. [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22]). It is reported that CAIS patients accounted for 1-2% in infancies presented with inguinal hernia which is rare in normal female infants, and nearly 57% of the CAIS population presented with an inguinal hernia in the U.K [11,12]. Therefore, further examination in females with inguinal hernia, especially karyotype analysis, is needed for timely diagnosis of CAIS. Two female patients had regular sexual life without vagina creation and dilation. Although the vagina varies from dimple in the perineum to normal length, most CAIS women were satis ed with their psychosexual development and sexual function [13]. Another two CAIS patients presented pubic hair (Tanner P2-P3) which developed in sparse CAIS patients while axillary hair was reported generally absent [1,14].
Partial androgen insensitivity syndrome should be excluded in infants presented with hypospadias. All ve male patients in the current study showed varying degrees of severities of hypospadias. Urethral rupture was repaired and testicular descent was performed in three patients. The prevalence of DSD is estimated to be 5‰ in births, with 73% of them being boys with hypospadias [15].  [20]. Therefore, genetic counseling for the individual and family should be strongly encouraged.
More than 1000 AR mutations have been reported so far, and up to 30% were sporadic de novo mutations [14]. Most AR mutations were identi ed in androgen insensitivity syndrome, and a small number of variants were detected in prostate cancer, spinal and bulbar muscular atrophy [21]. Missense variants in the current study accounted for the most 63.2% (12/19), followed by small insertions and deletion 21 [22]. However, different from the previous studies which found that exon 5 and exon 7 were commonly involved exons in AR, most variants in our study located on exon 1, exon 2 and exon 5, each accounting for 21.1% (4/19). This may be due to the high proportion of CAIS enrolled, as Batista et al said, defects in NTD (mainly encoded by exon 1) are more frequent in CAIS patients [22]. Four variants (A749D, A844E, G569W, A597T) have been reported to be associated with partial AIS which is consistent with phenotype presentation in our study [23][24][25][26]. A novel AR variant V819G on ligand-binding domain was identi ed in case 18 with PAIS. It is reported that most missense mutations involved in LBD are related to partial AIS [27]. Ten variants (52.6%) resided in ligand-binding domain which has been reported to be the most frequently involved domain.
AR belongs to the nuclear receptor superfamily which needs to combine with androgen to form a complex and enter to nucleus, then activates transcription factors and promotes the expression of downstream genes. It is widely expressed in the body, such as liver, adipose tissue, endometrium, ovary, prostate, testis, skin, etc., and is closely related to the occurrence and development of many diseases. The residual activity of the affected AR underlies the phenotype of AIS and could be analyzed based on reporter assay in vitro. In our study, the transcription activity analysis demonstrated six out of eight AR mutant proteins (P15Afs*69, S258Efs*47, W435Gfs*44, C560F, C577W, C580Afs*46) have severely impaired residual activity in vitro (about 1%) which showed a strong correlation between genotype and phenotype. In terms of the two variants C560F and C577W which located in DNA-binding domain, different nucleotide substitution at the same position has been reported and all variants were associated with the phenotype of CAIS [28,29]. The four frameshift variants resulting in premature stop codon have never been reported and the transcriptional activity was severely impaired. In contrast to decreased transcription activity, the AR mutant protein (K718X) on ligand-binding domain showed activated transcription compared with wildtype under DHT stimulation at different concentrations.
The similar phenomenon has been reported in Bevan et al study [30]. The two variants D864N and L907F were identi ed in complete androgen insensitivity and presented with considerable binding and transactivation activity. Considering the mechanism of nuclear receptor action, we further studied nuclear localization function of the novel AR variant K718X. Under 100nM DHT, the AR variant K718X showed impaired transport function from the cytoplasm to nucleus in COS-1 cells whose normality is vital to exert androgen biological function. In fact, in terms of the correlation between phenotype and genotype of AR variant, it is heterogenous even with the same substitution. A small number of mutations, as L581R, R608Q, R609K, Q641X, L723F, R727L, W742C, W752X, Y764C, R787X, Q799E, V867L, were identi ed in patients with androgen insensitivity syndrome but also found to be gain of transcription activity in prostate cancer which may imply the complex mechanism of AR action [9]. The AR mutant protein (V819G) showed residual activity of 59%. The patient with the AR variant V819G presented with mild hypospadias and gynaecomastia at puberty. Reporter assay showed increased transcription activity with gradually elevated DHT concentration which further con rmed the proof of pathogenicity and severity. As Hellmann et al reported gynaecomastia may be ameliorated by androgen therapy, de nite molecular diagnosis of AIS may reveal the severity of the AR variant and contribute to the individual's clinical management.
Our ndings suggested the necessity of genetic test and functional study in patients with AIS, especially in PAIS patients. For female patients presented with inguinal swelling before puberty, karyotype analysis is greatly encouraged. Functional analysis of the AR variant may be in favor of assessment of the responsiveness to androgen treatment. For those variants with increased transcription activity, AR function may be severely affected through abnormal nuclear localization in androgen insensitivity syndrome.

Conclusions
We reported 19 unrelated patients with AIS and identi ed eight novel AR variants. Inguinal swelling in females before puberty was a key clue for timely diagnosis of CAIS. We also con rmed the phenotype of AIS patients was closely associated with the residual AR activity, and transcription activity assessment could indicate the severity of the disease, thus provide professionals and individuals with available medical approach. Besides, our results indicated that AR variants could affect its normal function in different ways which may be the underlying causes of considerable heterogenous phenotypes.

Patients
A series of 19 patients with the diagnosis of androgen insensitivity syndrome from unrelated families during 2010 to 2020 were collected in Department of Endocrine and Metabolic Diseases, Shanghai Jiao Tong University School of Medicine, Ruijin Hospital. The study was approved by Ruijin Hospital Ethics Committee, Shanghai Jiao Tong University School of Medicine. All patients and/or their parents signed an informed consent for the genetic study. The diagnosis of androgen insensitivity syndrome was made according to the Williams Textbook of Endocrinology 14th (including karyotype, physiologic ndings, and hormone pro le) and further con rmed by molecular analysis of the AR gene.

Serum hormone measurements
Blood samples were collected in the morning and immediately centrifuged at 4℃. Serum testosterone (T), luteinizing hormone (LH), follicle-stimulating hormone (FSH) were measured by chemiluminescence immunoassay (Abbott Laboratories, Abbott Park, IL).

Sequence analysis of AR gene
Genomic DNA of the subjects and their family members was isolated from peripheral leucocytes using QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany). Eight exons and the anking splicing site of AR (NM_000044.6) gene were ampli ed with primers. The PCR products were puri ed by QIAquick PCR puri cation kit (Qiagen, Hilden, Germany) and sequenced in both sense and antisense direction on ABI 3700 sequencer (Applied Biosystems Perkin-Elmer, Transactivation assay of eight AR mutant proteins The transactivation activity of wildtype androgen receptor and eight novel AR variants were measured in transiently transfected COS-1 cells. 5×10 4 COS-1 cells/per well were seeded on 48 well plates in complete medium and cultured overnight up to 60-70% con uence. The cells were transfected with 400ng pCMV-GFP-AR WT or pCMV-GFP-ARmut, 400ng pARE-LUC, 4ng pRL-TK using Lipofectamine 3000 (Invitrogen) as directed by the manufacturer. After 1h, DHT with concentration of 1nM, 10nM, 100nM was added into the medium. After incubation for 48h, the cells were lysed and assayed using the Dual-Luciferase® Reporter Assay System (Promega). The ratio of Fire y to Renilla luciferase units was measured using a Turner TD-20/20 luminometer (Turner Designs, Sunnyvale, CA). Each experiment was done triplicate in at least three independent experiments, the transcription activity of wildtype AR at 100nM DHT was taken as 100% and the residual activity of eight AR variants expressed relative to that.
Nuclear localization of the AR mutant protein K718X 5×10 4 COS-1 cells/per well were seeded on 8-well glass slide (Merck-Millipore) in medium without penicillin-streptomycin and cultured overnight up to 60-70% con uence. The COS-1 cells were transfected with 500ng pCMV-GFP-empty (NC), pCMV-GFP-AR WT or pCMV-GFP-AR K718X plasmid using Lipofectamine 3000 (Invitrogen). After 24h incubation, cells were treated with 100nM DHT or DMSO. One hour after DHT stimulation, the 8-well glass slide was washed with PBS three times and fixed in 4% paraformaldehyde for 20min. After washing the slide with PBS, the nuclei were stained with Antifade Mounting Medium with DAPI (Beyotime, China) and then observed and photographed under a laser confocal microscope (Fluoview FV1000; Olympus, Japan).

Statistical analysis
Continuous variables were presented as median (IQR) and frequency (%) for categorial variables. The transcription activity of AR wildtype and eight novel variants was performed with GraphPad Prism (version 8.0).