Han-Jung Lee¹, Irene Mowszowicz², and Chawnshang Chang¹

¹Department of Medicine, and Endocrinology-Reproductive Physiology Program, University of Wisconsin, Comprehensive Cancer Center, Madison, WI 53792

²Biochemistry Laboratory B, Hospital Necker, Paris 75743, France

Received 04/19/96; Accepted 06/10/96; On-line 07/01/96

Patients 8044 and 8047 are affected siblings aged 17 and 18, respectively. They were first seen at the endocrine clinic for primary amenorrhea. Both presented the typically clinical phenotype of CAI with normal female genitalia, harmonious breast development, and the absence of body hair. Examination revealed a short and blunt vagina and the absence of uterus. Both karyotypes were 46,XY and hormonal profiles were summarized in Table 1. Plasma testosterone was 5.5 and 8.2 ng/ml in patients 8044 and 8047, respectively (normal adult male level, 5.8 ± 2.4 ng/ml). Testosterone/DHT ratio was in the normal range for adult males. Plasma delta 4-androstenedione level was 1.1 and 1.2 ng/ml (normal adult male range, 0.90-1.20 ng/ml). Plasma estradiol was 30 and 53 pg/ml, which is in the high range value for adult males that may due to the estrogen replacement therapy. LH and FSH levels were high due to negative feedback from lacking of normal AR, 15 and 20 mU/ml for patient 8044, and 12 and 18 mU/ml for patient 8047, respectively (normal adult male range, LH: 1-5 and FSH: 2-9 mU/ml). Urinary androstanediol glucuronide excretion was 74 and 50µg/24 h in patients 8044 and 8047, respectively (normal adult male level, 193 ± 77mg/24 h). Both patients were castrated and treated with estro-progestative replacement therapy. Skin biopsies were obtained with informed consent at the time of surgery including genital skin (labia majora) for patient 8044 who also had a vaginoplasty, and pubic skin for patient 8047.

However, 5alpha-reductase activity was normal, 68±6 fmol/mg DNA/h, in genital skin from patient 8044 (normal value, >2.1 fmol/g DNA/h). In pubic skin fibroblasts from patient 8047, 5alpha-reductase activity was low, 2.5 fmol/µg DNA/h (normal male value, 3-8 fmol/µg DNA/h). These data confirmed that 5alpha-reductase is neither androgen dependent in genital skin nor involved in the mechanism of androgen insensitivity in these patients.

Northern blot analysis of the expression of the AR mRNA in both patients exhibited a 10-kilobase band characteristic of the normal size of the AR mRNA. This result eliminates the possibility that a gross deletion or gene rearrangement occurred (6). To further examine if there was any mutation within the AR cDNA, we cloned and sequenced the entire AR cDNA (Fig. 1).Complementary DNAs were generated by reverse transcription (RT) from total RNA isolated from skin fibroblasts of both CAI patients, and amplified by the PCR using three pairs of primers (5 and G, 2 and 4, as well as A and B) that flank the entire AR cDNA (6, 7). The amplified DNA fragments corresponding to the N-terminal, DNA-,and ligand-binding domains were then subcloned into either pCR1000 or pT7Blue plasmid, and sequenced. Surprisingly, no mutation was found in the entire coding sequence of the AR gene from either patient (1, 15).

Fig. 1. (A) Schematic structure of the human AR cDNA and plasmids containing each domain amplified from RT-PCR. The entire human AR (hAR) cDNA can be basically divided into three parts, the N-terminal (NTD), DNA- (DBD) and ligand-binding (LBD) domains, flanked by three pairs of primers, 5 and G, 2 and 4, as well as A and B, respectively. (B) RT-PCR analysis. Each domain of the AR was amplified by RT-PCR and directly subcloned into either pCR1000 or pT7Blue vector. Sequence was analyzed by the dideoxynucleotide chain termination method.

These results reveal an unusual insight into the AR action at the molecular level of androgen resistance. In recent years, well-defined mutations in the AR gene have been identified in patients with androgen insensitivity syndrome (3-7). However, some evidence suggested that mutations of the AR gene could be responsible for most, but not all, cases of AIS (16). Interestingly, the data described here show no mutation in the coding region of the AR gene in two siblings (8044 and 8047) with AIS. To our knowledge thus far, this may represent the first case of a normal wild-type coding sequence within the AR gene in patients with AIS. These results may, therefore, highlight the importance of an unidentified mechanism for the androgen action, apart from the AR coding region itself. Thus, translational or post-translational control required for the expression of functional AR may be a contributing factor in AIS. In addition, it could be that cells of these patients contain an inhibitor that could bind to AR and prevent hormone binding, DNA binding, and/or receptor activation. Alternatively, other AR accessory proteins (co-activators or co-factors) may be needed for the fulfillment of normal AR function in these patients (17, 18). Therefore, defects at the translational level or of accessory factors of the AR may increase the complexity of the molecular heterogeneity in this clinical spectrum.