0021-972X/99/$03.00/0 The Journal of Clinical Endocrinology & Metabolism Copyright © 1999 by The Endocrine Society
Vol. 84, No. 5 Printed in U.S.A.
A Novel Mutation in the CAG Triplet Region of Exon 1 of Androgen Receptor Gene Causes Complete Androgen Insensitivity Syndrome in a Large Kindred* YUAN-SHAN ZHU, LI-QUN CAI, JUAN J. CORDERO, WILLIAM J. CANOVATCHEL, MELISSA D. KATZ, AND JULIANNE IMPERATO-MCGINLEY Division of Endocrinology, Diabetes, and Metabolism, Department of Medicine, Cornell University Medical College, New York, New York 10021. ABSTRACT Complete androgen insensitivity syndrome (CAIS) is an X-linked inherited disease caused by mutations in the androgen receptor (AR) gene. We have previously reported the largest kindred of CAIS, with 17 46,XY psychosexual and phenotypic females who lack secondary sexual hair. Analysis of AR binding indicated a receptor-negative form of complete androgen insensitivity, and DNA linkage analysis indicated that the absent binding was not caused by a large AR gene deletion. Using PCR-single-strand DNA conformational polymorphism, PCR-denaturing gradient gel electrophoresis, and DNA sequencing, we have identified a novel mutation in the polymorphic CAG trinucleotide region of exon 1 of the AR gene, where a single adenine is inserted, or equivalently, a GC-dinucleotide is deleted at
T
HE COMPLETE androgen insensitivity syndrome (CAIS), or testicular feminization, is a form of male pseudohermaphroditism caused by defects in the androgen receptor (AR) (1, 2). Subjects with CAIS, despite a 46,XY karyotype, testes, and normal-to-elevated plasma levels of testosterone, have female external genitalia and female psychosexual orientation (1– 4). At puberty, breast development occurs with scant-to-absent pubic and axillary hair. Partial forms of androgen insensitivity also occur with a broad spectrum of phenotypes, from females with clitoromegaly, to males with hypospadias and/or micropenis, to males with infertility or gynecomastia in its mildest forms (1, 2). The AR is a member of the nuclear steroid receptor superfamily (2, 5). It consists of 910 –919 amino acids and is encoded by a gene with 8 exons located in Xq11–12. Like other steroid receptors, the AR is a single polypeptide comprised of relatively distinct domains: an amino-terminal domain, a DNA binding domain, a hinge region, and a steroidbinding domain. The large amino-terminal domain encoded by exon 1 is the least conserved region among the steroid receptors and is involved in transcriptional activation of target genes. The DNA binding domain encoded by exon 2 Received August 31, 1998. Revision received February 3, 1999. Accepted February 9, 1999. Address all correspondence and requests for reprints to: Dr. Julianne Imperato-McGinley, Division of Endocrinology, Diabetes, and Metabolism, Department of Medicine, New York Hospital-Cornell Medical Center, 1300 York Avenue, Box-149, Room F-263, New York, New York 10021. E-mail:
[email protected]. * This work was supported, in part, by NIH Grants HD-09421–15 and M01-RR-00047 and by the Merck Foundation.
this region of the gene. The mutation results in a frameshift at amino acid 60 and a premature termination of the receptor downstream of the mutation. This predicts a mutant AR with only 79 amino acids in the amino-terminal of AR protein, prohibiting binding to the ligand, as well as the cognate DNA. The rest of the encoding regions of the AR gene in the affected subjects are normal. These results are consistent with previous ligand binding and DNA linkage analysis studies. This new mutation in the CAG trinucleotide area of exon 1 of the AR gene represents the first example of a defect in a CAG repeat causing CAIS in this large kindred. All previous reported variants in this region are changes in the number of triplet repeats. (J Clin Endocrinol Metab 84: 1590 –1594, 1999)
and 3 contains two zinc finger motifs (6) and is the most highly conserved region and is responsible for specific binding to its cognate DNA, i.e. androgen response element of target genes. The carboxyl-terminal of the AR contains the steroid-binding domain, encoded by the 39 portion of exon 4, and exons 5– 8. It is responsible for the specific high-affinity ligand binding. The carboxyl-terminal region also contains the subdomains involved in dimerization and transcriptional activation (7–9). Between the DNA-binding domain and the steroid-binding domain is the hinge region, which is encoded by the 59 portion of exon 4 and which contains the nuclear translocation signal (7, 8). We have previously reported the largest known kindred with complete androgen insensitivity (1, 3, 10), consisting of 17 affected subjects (Fig. 1). The clinical phenotype and the hormone profiles of the affected subjects from this kindred were previously published (3). Pedigree analysis (Fig. 1) demonstrated a maternal transmission, with the defect manifested in the fourth, fifth and sixth generations. Plasma testosterone levels were high normal or elevated. Dihydrotestosterone (DHT) levels were decreased and the ratio of testosterone/DHT elevated, as a consequence of a secondary 5a-reductase deficiency caused by androgen resistance (3). Analysis of AR binding in genital skin fibroblasts from CAIS subjects demonstrated a complete absence of receptor binding (3). DNA linkage analysis indicated that the absence of receptor binding was not caused by an entire AR gene deletion (10). In the present study, we identified the genetic defect responsible for complete androgen insensitivity in this large kindred. It is a novel mutation in the polymorphic CAG
1590
AR MUTATION IN CAIS
trinucleotide region of exon 1 of the AR gene, resulting in a frameshift and consequently premature termination of the AR in exon 1. Materials and Methods PCR amplification of the AR gene Peripheral blood was drawn from two affected subjects (indicated by an asterisk in Fig. 1), into EDTA-containing tubes; and genomic DNA from white blood cells was isolated by a genomic DNA isolation kit (Qiagen, Chatsworth, CA), according to the manufacture’s instruction. The concentrations of DNA were determined by ultraviolet absorbance. This study was approved by the Institutional Review Board of Cornell University Medical College. Exons 1– 8 of the AR gene (11, 12) were amplified by PCR using primers and conditions shown in Table 1. Six sets of inner primers were used for exon 1. The reaction mixture contained 0.18 mg genomic DNA, 200 mmol/L of each of four deoxyribo-
1591
nucleotide triphosphates, 10 mmol/L Tris-HCl (pH 8.3), 50 mmol/L KCl, 0.1% Triton X-100, 2.5 U thermostable DNA polymerase (Promega Corp., Madison, WI), and the amounts of primers and MgCl indicated in Table 1. For hot PCR, 10 mCi [a-32P]deoxy-ATP was added. The samples were denatured at 95 C for 2 min and then sequentially denatured at 95 C for 15 sec, annealed at the temperature shown in Table 1 for 30 sec, and extended at 72 C for 30 sec for a total of 35 cycles. A final extension cycle consisted of 72 C for 10 min.
Single-strand DNA conformational polymorphism (SSCP) analysis and denaturing gradient gel electrophoresis (DGGE) SSCP analysis was performed as described earlier (13, 14). Briefly, exon DNA was amplified and radiolabeled as described above. One microliter of the PCR product was added to 9 mL formamide denaturing dye (98% formamide, 20 mmol/L EDTA, 10 mmol/L NaOH, and 0.05%
FIG. 1. Pedigree of a large kindred with CAIS. Circles and squares indicate females and males, respectively. The Roman numerals indicate the generation. Solid circles indicate the affected CAIS subjects. The asterisks indicate the affected CAIS subjects analyzed in the present study.
TABLE 1. Primer sequences and PCR conditions for the amplification of AR gene Exon
1-a 1-b 1-c 1-d 1-e 1-f 2 3 4 5 6 7 8
Primer (mM)
Primer
Sequences
Size (bp)
Anneal T (C)
MgCl2 (mM)
AR1N AR1C1 AR1N2 AR1C2 AR1N3 AR1C3 AR1N3 AR1C3 AR1N4 AR1C4 AR1N5 AR1C AR2N AR2C AR3N AR3C AR4N AR4C AR5N AR5C AR6N AR6C AR7N AR7C AR8N AR8C
GGAAGTAGGTGGAAGATTCAGC CGAAAGCGACATTTCTGGAAGG TCCCCAAGCCCATCGTAGAG CTCCACACCCAGGCCCATG GGAGTTGTGTAAGGCAGTGTCG GTAGACGGCAGTTCAAGTGTCC ACCAAAGGGCTAGAAGGCGA GTGTGCCAGGATGAGGAAGC GACCTGGCGAGCCTGCATG TGCGGTGAAGTCGCTTTCCTG GGAGCTGTAGCCCCCTACGG CGAAAGCGACATTTCTGGAAG TTCAGTGACATGTGTTGCATTGG GGTTAGTGTCTCTCTCTGGAAG GTTTGGTGCCATACTCTGTCCAC TCTGGTCTAAAGAGAGACTAG CCACTGATGATAAATTCAAGTCTCTC CTAAATATGATCCCCTTATC CCAACAGGGAGTCAGACTTAGC AGGTCTGGCCAAGCTGCTG CCCTCATTCCTTTTTCCTCTG GGCATTCCCTGCACTTCTAG TCTAATGCTCCTTCGTGGGC CTCTTATCAGGCTGTTCTCCCTG GAGGCCACCTCCTTGTCAAC AGTTATAACAGGCAGAAGACATCTG
360
60
1.5
1
469
65
1.5
1
320
60
1.5
1
364
54
1.5
1
270
60
2
1
221
54
1.5
1
262
54
1.5
1
300
50
1.5
1
376
60
2
1
277
65
1.5
0.5
195
65
1.5
0.5
265
65
1.5
0.5
277
70
1.5
1
Notes: Exon 1-a to 1-f indicates exon 1 fragment a and fragment f, respectively. T, temperature.
1592
JCE & M • 1999 Vol 84 • No 5
ZHU ET AL.
each of xylene cyanol and bromophenol blue), denatured at 100 C for 6 min and immediately cooled on ice. Three microliters of this solution were loaded onto a 0.5 3 Hydrolink MDE gel (J. T. Baker Inc., Phillipsburg, NJ), containing 10% glycerol, and electrophoresed at 350 V at room temperature overnight in 0.6 3 TBE buffer (54 mmol/L Tris-borate (pH 8.3) and 2.4 mmol/L EDTA). An aliquot of hot PCR sample, without denaturation, was loaded in an adjacent lane to determine the position of migration of the double-stranded DNA fragment. After electrophoresis, the gel was dried and exposed to Kodak BioMAX film (Eastman Kodak Co., Rochester, NY) at room temperature. For DGGE, GC-clamp primers were synthesized for exons 4 – 8 of the AR gene, and PCR amplification was carried out as described above. Amplified DNA fragments were denatured at 95 C for 10 min and reannealed by cooling down slowly to room temperature, and electrophoresed in an 8% denaturing gradient polyacrylamide (19:1 acrylamide:bisacrylamide) gel in 0.5 3 TBE buffer. The 100% denaturant was 7 mol/L urea plus formamide (60:40 by vol). Intermediate denaturants were prepared by diluting 100% denaturant with 8% acrylamide in 0.5 3 TBE. Gradients were prepared bottom-up by gravity flow in a gradient maker. A Miniprotein apparatus (Bio-Rad Laboratories, Hercules, CA.) with 1.5 mm spacers was used for all of the experiments. The results were visualized by ethidium bromide staining, as described previously (15, 16).
DNA sequencing DNA fragments of the AR gene were amplified by PCR, as described above. The PCR product was purified by PAGE (17). The concentration of purified DNA fragment was estimated using DNA QuikSTRIP (Eastman Kodak Co.). DNA sequencing was carried out using fmol DNA sequencing kit (Promega Corp.) with 32P end-labeled primer (13, 14). Both strands of DNA were sequenced.
Results
To identify the genetic defect in these CAIS subjects, genomic DNA was isolated from the white blood cells of 2 affected CAIS subjects (indicated by an asterisk in Fig. 1). The entire encoding region of the AR gene was amplified by PCR using the primers and conditions shown in Table 1. The exon 1 coding sequence was amplified with 6 pairs of overlapping primers. Using SSCP analysis, all 8 exons of the AR gene were screened for a mutation. Figure 2 shows a representative SSCP analysis of the 59 portion of exon 1 of the AR gene, which demonstrates a differential migration pattern of the single-stranded DNA in the affected CAIS subject, compared with the normal control. This differential migration pattern results from either a potential mutation in this region or a difference in the size of CAG trinucleotide repeats of the AR gene. DNA sequencing (Fig. 3A) of this putative mutant fragment identified a novel mutation in the polymorphic CAG trinucleotide repeats of exon 1 of the AR gene in the affected CAIS subjects (see Fig. 3), where either a single adenine (A) is inserted between nucleotide position 179 and 180 (. . . 178CAAGCAG183. . . ), or alternatively, a GC-dinucleotide is deleted at nucleotide positions 180 and 181 (. . . 178 CAGCAG183. . . ) of the AR gene (base 1 starts at the first nucleotide of the translation start codon of the AR gene, and all nucleotide and amino acid position numbers referred to in this paper are according to human AR complementary DNA in GenBank accession no. M20132). Either mutation would result in a frameshift at amino acid position 60 and an introduction of premature stop codons in exon 1 of the AR gene. Thus, the predicted mutant AR is a truncated protein with only 79 amino acids at the amino-terminal (see Fig. 3B). The remainder of the encoding region of the AR gene in
FIG. 2. A representative SSCP analysis of a 59 portion of exon 1 of the AR gene. Hot PCR amplification was done using primers AR1N and AR1C1 and then analyzed by SSCP, as described in Materials and Methods. The bands on the top represent the mobility of various single-strand DNA conformations, and the band on the bottom is the double-stranded DNA, as indicated by an arrowhead.
the affected subjects was normal by SSCP analysis and DNA sequencing. Exons 4 – 8 were also screened by DGGE, and no mutation was detected (data not shown). The identified mutation in the CAG triplet repeats of exon 1 of the AR gene was not detected in a normal 46,XY brother, suggesting that this new mutation is responsible for complete androgen insensitivity in this large pedigree. Discussion
We have previously reported and characterized the largest known kindred with CAIS in the world (3, 10). In the present study, a new and novel mutation of the AR gene is identified
AR MUTATION IN CAIS
1593
as the genetic defect in this kindred (see Fig. 3). A single base, adenine (A), is inserted between nucleotide position 179 and 180 (nucleotide position no. 1 starts at the translation start codon) in the polymorphic CAG trinucleotide region of exon 1. Alternatively, because of the presence of the polymorphic CAG trinucleotide repeats, the defect could also be attributable to a GC-dinucleotide deletion at positions 180 –181 of exon 1 of the AR gene (see Fig. 3). The precise mutation in the sequence is ambiguous. However, either an adenine insertion or a GC-dinucleotide deletion yields a frameshift at amino acid 60, introducing a premature termination codon downstream from the mutation, predicting the synthesis of a truncated protein with only 79 amino acids in the aminoterminal and all functional domains absent (see Fig. 3, B and C). The rest of the exon sequences and splicing junctions of the AR gene in the affected subjects were normal, by DNA sequencing, SSCP, and DGGE analysis. The identified mutation was not detected in a normal 46,XY sibling. Our previous studies of affected subjects from this pedigree demonstrated absent DHT-binding in genital skin (3), consistent with the fact that a ligand-binding domain is absent in the mutant receptor. Also, a large deletion of the AR gene was not detected by previous studies of DNA linkage analysis (10), because of the fact that only a single base insertion is present in exon 1. Because this mutation leads to the synthesis of a nonfunctional truncated AR fragment in the affected subjects, it is not surprising that affected subjects from this kindred present with the phenotype of complete androgen insensitivity. Different frameshift mutations in exon 1 of the AR gene have been previously reported by others, in affected CAIS subjects (2, 18, 19) and in the Tfm mouse (20). It has been shown that in the Tfm mouse, reinitiation of protein translation from an internal methionine codon downstream of the abnormal premature stop codon produces a low level of truncated receptor (20), containing both the DNA binding domain and the ligand binding domain. This explains why a small amount of DNA binding and androgen binding activity is detected in Tfm mice (21). A similar phenomenon has been reported in CAIS affected subjects with a nonsense mutation (CAG3TAG) at amino acid position 60 (Q60stop) of exon 1 of the AR gene (22). Reinitiation of AR translation distal to the abnormal stop codon results in synthesis of a small amount of truncated AR and a low level of androgen binding activity. This truncated AR is named as an A form of AR, and has recently been shown to possess similar functional activities as the wild-type, B form (23). Theoretically,
FIG. 3. The identification of a novel new mutation in exon 1 of AR gene in the affected CAIS subjects from a large kindred. Panel A shows a representative DNA sequencing of a 59 portion of exon 1 of the AR gene. The fragment was amplified by PCR using primers AR1N and AR1C1, purified by polyacrylamide gel and sequenced using 32P end-labeled AR1C1 primer, as described in Materials and Methods. The solid arrowheads indicate the insertion of an adenine (A) in the normal sequence. The sequences on both sides are read from
top to bottom as 59 to 39. Panel B shows partial nucleotide and amino acid sequences of exon 1 of human AR to illustrate the single base (adenine) insertion (indicated by a bold and underlined A), or a GCdinucleotide deletion in the exon 1 of AR gene. This mutation results in a frameshift at amino acid position 60 and an abnormal premature termination at codon 80 (TAG). The sequences after the mutation are indicated as italic letters. Panel C shows the genomic structure and the corresponding encoded functional domains of the human AR gene. The exons are indicated by boxes and introns by dashed lines. The shaded boxes indicate the coding region and the open boxes indicate the untranslated regions. The currently identified mutation within the polymorphic CAG trinucleotide repeats is indicated by a big bolded A and an arrow for a single base (adenine) insertion or by the underlined GC for the alternative GC-dinucleotide deletion.
1594
ZHU ET AL.
reinitiation of AR translation downstream of the mutation could also occur in our currently identified mutation. However, because we did not detect any AR binding activity in the genital skin fibroblasts from the CAIS affected subjects of this kindred (3), it is likely that the reinitiation of truncated AR translation is either absent or negligible. Although reinitiation of truncated AR translation has been reported (20 –22), most AR exon 1 mutations identified in CAIS subjects did not report downstream reinitiation. The reason for this difference is unclear. Translation efficiency may be one explanation, as McPhaul’s group reported that the replacement of the upstream sequences of the translation reinitiation site by the upstream sequences of the normal translation initiation site could significantly increase mutant receptor expression (23). To date, more than 245 different mutations, involving all 8 exons in the AR gene, have been reported (2, 19). These mutations range from a single point mutation to an entire gene deletion and result in various degrees of functional impairment of the AR and a wide phenotypic spectrum of the syndrome of androgen insensitivity. However, there is no consistent relationship between the clinical phenotypes and the molecular defects in the AR gene. Most of the reported mutations are single point missense mutations located in exons 2– 8 of the AR gene. Interestingly, mutations in exon 1 of the AR gene are found less frequently than found in other exons. Seventeen mutations of exon 1 have been reported in androgen insensitivity. Six are attributable to small deletions or insertions (2, 19), resulting in CAIS. The other mutations include 8 nonsense and 3 missense mutations. Our reported mutation has never been previously reported. It is unique in terms of its location within the region of polymorphic CAG trinucleotide repeats of exon 1 and its feature as either a single adenine insertion or a GC-dinucleotide deletion (see Fig. 3). The region of CAG triplet repeats in exon 1 of the AR gene is genetically unstable, and this instability is greater in male than in female meiosis (2). An increase in CAG repeats in this region is associated with Kennedy disease, i.e. X-linked spinal and bulbar muscular atrophy (2, 24). A shortening of the number of CAG repeats may also be related to prostate cancer risk (25). Thus, all previously reported variants in this region are variations in the numbers of triplet repeats. Our currently reported mutation represents the first example of small insertion or deletion in this CAG triplet region. The mechanism of this genetic instability is unclear, although slippage of DNA polymerase during DNA replication is possible (2). In summary, we have identified a novel new mutation in exon 1 of the AR gene, responsible for the CAIS in the largest known kindred with this syndrome in the world. References 1. Imperato-McGinley J, Canovatchel WJ. 1992 Complete androgen insensitivity: pathophysiology, diagnosis and management. Trends Endocrinol Metab. 3:75– 81.
JCE & M • 1999 Vol 84 • No 5
2. Quigley CA, De Bellis A, Marschke KB, el-Awady MK, Wilson EM, French FS. 1995 Androgen receptor defects: historical, clinical, and molecular perspectives. Endocr Rev. 16:271–321. 3. Imperato-McGinley J, Peterson RE, Gautier T, et al. 1982 Hormonal evaluation of a large kindred with complete androgen insensitivity: evidence for secondary 5a-reductase deficiency. J Clin Endocrinol Metab. 54:931–941. 4. Imperato-McGinley J, Peterson RE, Gautier T, Sturla E. 1985 The impact of androgens on the evolution of male gender identity. In: DeFries Z, Friedman R, Corn R, eds. Sexuality: new perspectives. Westport, CT: Greenwood Press; 126 –140. 5. Evans RM. 1988 The steroid and thyroid hormone receptor superfamily. Science. 240:889 – 895. 6. Freedman LP. 1992 Anatomy of the steroid receptor zinc finger region. Endocr Rev. 13:129 –145. 7. Jenster G, van der Korput HA, van Vroonhoven C, Van der Kwast TH, Trapman J, Brinkmann AO. 1991 Domains of the human androgen receptor involved in steroid binding, transcriptional activation, and subcellular localization. Mol Endocrinol. 5:1396 –1404. 8. Simental JA, Sar M, Lane MV, French FS, Wilson EM. 1991 Transcriptional activation and nuclear targeting signals of the human androgen receptor. J Biol Chem. 266:510 –518. 9. Wong CI, Zhou ZX, Sar M, Wilson EM. 1993 Steroid requirement for androgen receptor dimerization and DNA binding. Modulation by intramolecular interactions between the NH2-terminal and steroid-binding domains. J Biol Chem. 268:19004 –19012. 10. Imperato-McGinley J, Ip NY, Gautier T, et al. 1990 DNA linkage analysis and studies of the androgen receptor gene in a large kindred with complete androgen insensitivity. Am J Med Genet. 36:104 –108. 11. Chang C, Kokontis J, Liao S. 1988 Molecular cloning of human and rat complementary DNA encoding androgen receptors. Science. 240:324 –326. 12. Lubahn DB, Brown TR, Simental JA, et al. 1989 Sequence of the intron/exon junctions of the coding region of the human androgen receptor gene and identification of a point mutation in a family with complete androgen insensitivity. Proc Natl Acad Sci USA. 86:9534 –9538. 13. Cai LQ, Zhu YS, Katz MD, et al. 1996 5a-reductase-2 gene mutation in the Dominican Republic. J Clin Endocrinol Metab. 81:1730 –1735. 14. Can S, Zhu YS, Cai LQ, et al. 1998 The identification of 5a-reductase-2 and 17b -hydroxysteroid dehydrogenase-3 gene defects in male pseudohermaphrodites from a Turkish kindred. J Clin Endocrinol Metab. 83:560 –569. 15. Sheffield VC, Cox DR, Lerman LS, Myers RM. 1989 Attachment of a 40base-pair G 1 C-rich sequence (GC-clamp) to genomic DNA fragments by the polymerase chain reaction results in improved detection of single-base changes. Proc Natl Acad Sci USA. 86:232–236. 16. Stavrou SS, Zhu YS, Cai LQ, et al. 1998 A novel mutation of the human Luteinizing hormone receptor in 46XY and 46XX sisters. J Clin Endocrinol Metab. 83:2091–2098. 17. Sambrook J, Fritsch EF, Maniatis T. 1989 Molecular cloning: a laboratory manual. 2nd ed. Plainview, NY: Cold Spring Harbor Laboratory Press. 6. 46 – 6.48 18. Batch JA, Willams DM, Davies HR, et al. 1992 Androgen receptor gene mutations identified by SSCP in fourteen subjects with androgen insensitivity syndrome. Hum Mol Genet. 1:497–503. 19. Gottlieb B, Lehvaslaiho H, Beitel LK, et al. 1998 The androgen receptor gene mutations database. Nucleic Acids Res. 26:234 –238. 20. He WW, Kumar MV, Tindall DJ. 1991 A frame-shift mutation in the androgen receptor gene causes complete androgen insensitivity in the testicular-feminized mouse. Nucleic Acids Res. 19:2373–2378. 21. Young CY, Johnson MP, Prescott JL, Tindall DJ. 1989 The androgen receptor of the testicular-feminized (Tfm) mutant mouse is smaller than the wild-type receptor. Endocrinology. 124:771–775. 22. Zoppi S, Wilson CM, Harbison MD, et al. 1993 Complete testicular feminization caused by an amino-terminal truncation of the androgen receptor with downstream initiation. J Clin Invest. 91:1105–1112. 23. Guo T, McPhaul MJ. 1998 Functional activities of the A and B forms of the human androgen receptor in response to androgen receptor agonists and antagonists. Mol Endocrinol. 12:654 – 663. 24. La Spada AR, Wilson EM, Lubahn DB, Harding AE, Fischbeck KH. 1991 Androgen receptor gene mutations in X-linked spinal and bulbar muscular atrophy. Nature. 352:77–79. 25. Stanford JL, Just JJ, Gibbs M, et al. 1997 Polymorphic repeats in the androgen receptor gene: molecular markers of prostate cancer risk. Cancer Res. 57:1194 – 1198.
