Department of Internal Medicine University of Texas Southwestern Medical Center, Dallas, Texas 75235-8857
Address correspondence and requests for reprints to: M.J. McPhaul, M.D., Department of Internal Medicine, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, Texas 75235-8857.
Androgens exert their effects in mediating the development of the normal male phenotype via a single receptor protein, the androgen receptor (AR), which is encoded on the X chromosome. Abnormalities that alter the function of this receptor result in a range of abnormalities of male phenotypic development. These phenotypes range from that of normal females (complete testicular feminization, complete androgen insensitivity) to those that are characterized by only minor degrees of undervirilization and/or infertility.
The defects of receptor function that have been characterized fall into two major categories. The first are those that disrupt the primary sequence of the AR. These mutations can be due to the introduction of premature termination codons, frameshift mutations, deletions or insertions, or alterations of RNA splicing. The ARs that are produced as a result of these genetic alterations are uniformly associated with complete androgen resistance.
The second, and most common type of AR mutation, is that which is caused by single amino acid substitutions within the AR protein. In contrast to the preceding category, mutations of this type may result in the full spectrum of androgen resistant phenotypes. These mutations cluster in two important domains of the receptor protein: the DNA- and the hormone- binding domains. Substitutions in the DNA-bindingdomain act to impair the ability of the AR to recognize target sequences within or adjacent to androgen-responsive genes. The degree of impairment of DNA binding is directly related to the degree of impairment of receptor function. Amino acid replacements within the hormone-binding domain can have a range of effects on the binding of ligand by the AR. In some instances, the capacity to bind hormone is completely abolished, while in other instances, only subtle qualitative abnormalities of ligand binding can be detected. In all mutants containing amino acid substitutions within the hormone-binding domain, the ability to form stable AR-hormone complexes determines the amount of receptor function that remains.
The phenotype that is observed does not appear to correlate with the identity of the residue that is mutated or its replacement. Instead, the phenotype appears to be a reflection of the degree to which androgen action is impaired. As the level of mutant AR protein expressed is similar to that observed in normal subjects, this usually reflects the degree to which AR function is altered.
The male hormones, androgens, control wide range of processes in
the male during embryonic development and in postnatal life. During
embryogenesis, the concerted actions of testosterone (T) and
5-dihydrotestosterone (DHT) are critical to the virilization of the
Wolffian duct structures and the external genitalia. Genetic defects
that impair androgen synthesis or action will result in abnormalities
of male phenotypic development in 46, XY individuals with testes (male
pseudohermaphroditism). Among the potential etiologies, defects of the
AR are among the most common.
AR Defects Cause A Spectrum of Male Phenotypic Abnormalities
In 1953, Morris (1) described a series of patients with the syndrome of "testicular feminization." According to his characterization, these individuals possessed a phenotype characterized by "a female habitus... , normal female breasts... , absent or scant axillary and pubic hair... , female external genitals... , absent internal genitals... ", and gonads "essentially... [those] encountered in cases of undescended testes." While considerable debate revolved for a time around the exact etiology of this syndrome, clinical experiments established that such conditions represent deficiencies of androgen action and not abnormalities of androgen production or metabolism (2, 3). In addition, although attention was initially focussed on phenotypes caused by the complete lack of response to androgen, subsequent studies suggested that other less severe defects of male development were caused by less than complete defects of AR function (4, 5).
In addition to the relative high frequency of patients exhibiting clinically apparent forms of androgen resistance, AR disorders are unusual in the wide range of phenotypic abnormalities. This high frequency appears to be the result of two characteristics of the AR gene defects. First, the AR is located on the X-chromosome. For this reason, in normal 46, XY males only a single AR gene will be present and any substantial defect of function will be evident. Second, although such defects of AR function may result in abnormalities of sexual development, these alterations do not appear to alter the viabilities of affected subjects.
The phenotypic spectrum caused by defects of the AR is most easily
explained as reflecting the degree to which the AR-dependent processes
that are critical to male sexual development have been disturbed (Table 1). As the development of the male
internal and external genitalia is an androgen dependent process, the
male internal and external structures fail to develop when the AR is
completely defective. This syndrome, complete testicular feminization
or complete androgen insensitivity, is characterized by a complete
absence of male development. Affected individuals have normal female
external genitalia and normal breast development. Physical or
ultrasound examinations will identify testes within the labia majora or
abdominal cavity. The uterus and fallopian tubes are absent, and the
vagina is blind ending in such subjects, as the production of
anti-Müllerian hormone by the testes is preserved.
|
In some instances virilization is normal, and the defects of AR function are more subtle. In such patients, although the virilization of the external genitalia is normal (normal phallic and scrotal development), gynecomastia and decreased facial and body hair are present as signs of mild androgen resistance. In a small subset of patients even these clinical signs are absent, and severe oligospermia or azoospermia are the only manifestations of the AR defect (8).
Structure and function of the AR
The AR is a member of the nuclear receptor family, and its
sequence is highly conserved compared to the predicted amino acid
sequences of other steroid receptors and more distantly related members
of the nuclear receptor family (9). The AR, like other members of the
steroid receptor family, contains conserved elements that mediate the
binding of ligand with high affinity (ligand-binding domain or LBD) and
that mediate the recognition of specific DNA sequence elements
(DNA-binding domain or DBD). In addition, the nucleotide sequence
predicts a large amino terminal segment that is required for activity
of the receptor in transcription-based assays and that contains a
number of repeated motifs (Fig. 1).
|
A remarkable number of mutations have now been described as causing various forms of androgen resistance (7, 10, 11). Using this information as backdrop, it is possible to illustrate features of the four main categories of mutation causing androgen resistance.
Interruptions of the AR open reading frame
Interruption of the primary sequence of the human AR can be caused
by a number of different mutation types. Deletions and insertions,
premature termination codons, and defects of AR mRNA splicing have all
been described as causing androgen resistance. As the critical
hormone-binding and DNA-binding domains are localized to the carboxyl
terminus of AR protein, each type of defect results in the production
of a receptor protein that is lacking at least one of these important
functional domains or in which the structure of the domain has been
interrupted. These mutations result in the synthesis of mutant ARs that
are unable bind ligand (Fig. 1). Such defects have been localized to
all eight exons of the AR gene (Fig. 2
).
|
In one series of patients studied using ligand-binding assays,
approximately 20% of patients did not display discernible
abnormalities of ligand binding by the AR (10). This observation
suggested that such individuals either carried subtle abnormalities of
AR function or harbored defects in genes (other than the AR) that were
required for normal AR function. The analyses of the AR genes from a
number of such pedigrees have now been reported. In the vast majority
of cases, amino acid substitutions are detected within the DBD of the
receptor (Figs. 1 and 2
).
In the studies of Zoppi et al. (13), the amino acid substitutions were localized to the DBD of receptor protein in four unrelated subjects with complete or incomplete testicular feminization. When these mutant ARs were expressed and analyzed in heterologous cells, each was found to bind ligand with normal or near normal affinity. Despite this, assays of AR function demonstrated that each was markedly impaired in terms of function. Experiments conducted using fusion proteins containing the normal or mutant AR DBDs demonstrated that each of the mutant receptors was unable to bind normally to target DNA sequences (13). Similar findings have been obtained in studies conducted in a number of different laboratories to analyze similar groups of patients with complete or partial forms of androgen resistance (14, 15, 16, 17, 18).
In many ways, this class of mutation is the easiest to conceptualize mechanistically and appears to comprise a relatively homogeneous group. Each amino acid substitution alters the structure of the DBD of the receptor in a way that impairs the binding of the receptor to specific target DNA sequences. Mutations that alter the structure of the DBD by means other than amino acid substitution appear to alter receptor function by similar mechanisms. The degree of clinical androgen resistance that is observed appears to exhibit a direct relationship to the extent receptor function is impaired (see below).
Amino acid substitutions in the hormone-binding domain of the AR
Single nucleotide substitutions that result in single amino acid
replacements in the HBD are the most frequent abnormalities that are
identified in patients with androgen resistance (Fig. 2). As noted
above for the substitutions in the DBD of the receptor, such
substitutions have been identified in individuals comprising the entire
spectrum of androgen resistance. These mutations can be divided into
two categories based on the effects on the binding of ligand: those
resulting in the absence of detectable ligand binding and those that
cause qualitative abnormalities of ligand binding (Fig. 1
).
Amino acid substitutions causing absent ligand binding
Amino acid replacements within the hormone-binding domain of the AR that abolish ligand binding in genital skin fibroblast (GSF) cultures constitute two general classes. In rare instances, the amino acid replacement occurs at a residue within the HBD that results in a major structural alteration of domain structure. In these instances, the resulting AR protein is unable to bind hormone. The best example of this type of mutant receptor is the W741R AR, described by Marcelli et al. (19). In this mutant AR, a hydrophobic residue at the amino terminus of helix 5 is replaced with a charged arginine residue. In addition to disrupting the hydrophobic nature of the ligand binding pocket, this residue is predicted to make important contacts with the C ring of testosterone by analogy to the structure of the progesterone receptor ligand binding domain (LBD) (20). Such mutant ARs are unable to bind hormone under any conditions.
Although studies performed in patient fibroblasts may indicate that the mutant ARs lack the capacity to bind ligand, mutations that completely abolish the binding of hormone by the AR are quite infrequent. Careful studies in which mutant receptors containing amino acid substitutions are expressed in heterologous cells often demonstrate that they are capable of binding hormone. An example of one such mutation of this type is the mutant R774C. When genital skin fibroblast monolayers were used to study ligand binding by the mutant AR, this mutation was classified as ligand-binding negative. When cDNA encoding this same mutant receptor was expressed in heterologous cells, however, ligand binding was measurable in whole cell assays, although the expressed AR exhibited altered kinetics and stability of ligand binding (21). In fact, a number of other mutant receptors with amino acid substitutions in the HBD have been described that exhibit these same three type characteristics: 1) relatively normal or near normal levels of immunoreactive receptor, 2) markedly decreased or absent levels of ligand binding in fibroblast ligand binding assays, and 3) detectable levels of ligand binding when expressed in heterologous cells. While such differences could be viewed as reflecting variations in the handling or processing of the AR in native compared to heterologous, transfected cells, it is more likely that this apparent discrepancy reflects differences in the level of expression in the two types of cells and the sensitivity of the assays that are employed to analyze the samples.
Amino acid substitutions that lead to qualitative abnormalities of ligand binding
The use of genital skin fibroblast cultures to study and classify patients with the various forms of androgen resistance led to the recognition that in some individuals although the levels of AR binding detected were quantitatively normal, qualitative defects of the receptor could be detected. Such abnormalities include alterations in the affinity of ligand binding by the receptor, increased lability of the receptor to thermal denaturation, or an increased instability of the hormone-receptor complex (increased rate of ligand dissociation). In one series, these qualitative abnormalities constituted over 40% of patients with androgen resistance (10).
The genetic bases of androgen resistance associated with such
alterations have been determined for a considerable number of patients.
The mutations that cause qualitative abnormalities of ligand binding
are single amino acid substitutions within the hormone-binding domain
of the receptor. Although a large number of different residues have
been shown to be mutated, an examination of the distribution of
residues substituted in samples that display qualitative abnormalities
of ligand binding has revealed a substitution pattern that is similar
to that identified for samples that exhibit absent ligand binding (22,
Fig. 1).
This observation suggested that the differences between these categories was one of degreemajor disruption of LBD structure causing absent binding and those causing lesser changes of structure causing qualitative abnormalities of binding. That this concept is fundamentally correct has been shown in experiments examining mutant receptors in which the same amino acid has been mutated to different residues. In such instances, substitution at a position within the AR protein is capable of causing absent ligand binding or a qualitative abnormality of ligand binding, depending on the nature of the amino acid substitution. For example, Prior et al. (23) reported that replacement of arginine 774 by a cysteine residue caused androgen resistance associated with undetectable levels of ligand binding in patient fibroblasts. By contrast, substitution of the same amino acid residue with histidine led to normal levels of androgen binding in fibroblasts, although the ligand binding that was present displayed marked thermal instability (23). A number of additional pedigrees have been described in which different amino acid substitution mutations lead to discernibly different effects on the binding of ligand by the receptor (24, 25, 26).
The assessment of the function of qualitatively abnormal ARs poses challenges not encountered in the characterization of ARs in other mutant categories. The ligand employed, the promoter used, and the target cells in which the assays are conducted are each capable of profoundly affecting the results of the functional assays. This impact of ligand and dosing was evident in the work of Marcelli et al. (19), who examined the responsiveness of a number of different mutant ARs caused by amino acid substitution mutations in the hormone binding domain. The mutant ARs chosen for study included those in which the amino acid replacements caused the mutant ARs to exhibit a range of distinct qualitative and quantitative abnormalities of ligand binding. In addition, the transfection experiments were performed in a cell line (CV1) in which hormone catabolism was rapid, as occurs in most androgen target tissues. The results were interesting in several respects. First, they demonstrated the markedly different results that can be obtained when assays conditions are varied. In most instances, testosterone was a less potent stimulus to the activation of reporter gene than was either DHT or mibolerone. Second, stimulation of the mutant ARs using a nonmetabolizable androgen, mibolerone, led to a marked enhancement of AR function, even for mutant receptors that were severely impaired when stimulated with DHT or T. This finding suggested the critical importance of the stability of the hormone-receptor complex and was reinforced by the finding that the same effect could be observed for several mutant ARs when metabolizable androgen was presented as several distinct pulses of hormone.
This observation suggests that mutations of the AR that diminish the stability of the hormone-receptor complex will have a major effect on the activity of the liganded receptor protein. In addition to raising caution as to the manner in which functional assays are conducted in heterologous cells, these findings also suggest that it should be possible to manipulate mutant receptors that display qualitative abnormalities of ligand binding pharmacologically, both in vitro and potentially in vivo. This has been attempted in a limited number of cases. In several instances, the response observed clinically has supported this concept (27, 28).
Mutations of the AR that result in reduced levels of ligand binding
The first mutation identified in this category was discovered in a subject with complete androgen insensitivity. Attention was focused on the amino terminus of the AR early in the analysis, as studies using antibodies directed at the amino terminus of the receptor failed to detect immunoreactive AR, although considerable levels of the receptor could be detected using ligand binding assays. Nucleotide sequence analysis of the AR identified a single nucleotide substitution within the amino terminus of the AR that predicted the insertion of a premature termination codon at codon 60 (Q601; ref. 29). Subsequent experiments revealed the residual binding detected was due to the synthesis of a shorter form of the AR was synthesized in normal and mutant fibroblasts. Transfection experiments examining the function of this shortened AR demonstrated subtle differences function on selected response elements. These experiments suggested that phenotype of complete androgen insensitivity was caused by reduced level of an AR that exhibits a diminished functional capacity (30).
The study of the partial androgen insensitivity pedigree analyzed by Choong et al. (31), implicates a different mechanism, but with an outcome that is quite similar. AR gene analysis of affected subjects within this pedigree identified a nucleotide substitution mutation that changed the amino acid sequence of the mutant AR at the second amino acid residue, predicting a lysine residue in place of the normal aspartate residue. Experiments examining the activity of the mutant AR protein, as well as examining the levels of expression suggested that the androgen resistance phenotype was due primarily to the reduced levels of AR that were expressed.
AR gene mutations and phenotype
Given the number of mutations that have been analyzed, it is now possible to draw general conclusions regarding the nature of the relationship between clinical phenotype and AR mutation. Alterations of the AR gene that result in interruption of the AR open reading frame are invariably associated with a phenotype of complete androgen insensitivity. From a mechanistic standpoint, this derives from the carboxyl terminal locations of the DBDs and HBDs. Mutations that interrupt the receptor protein at any point during its synthesis will invariably remove portions of one or both of these important functional domains. Such alterations of AR structure can result from the insertion of a premature termination codon, an alteration of AR mRNA splicing, insertions, or the deletion of partial or complete exon segments.
In contrast to mutations that interrupt the primary amino acid sequence of the AR, mutations that result in single amino acid substitutions within the AR protein may cause any of the androgen-resistant phenotypes. At this juncture, it is apparent that the degree of androgen resistance observed clinically does not correlate with the nature or location of specific mutations. Instead, the phenotype appears to be a direct reflection of the level of residual AR function that is present. Such decreases in the activity of the AR could be caused by alterations in the functional capacity of the AR, its level of expression, or combinations of both factors. In the vast majority of cases, however, the levels of immunoreactive AR that are synthesized are normal, and the observed impairment of androgen action reflects a decreased functional capacity of the receptor protein that is expressed.
Identical amino acid substitutions have been reported by different groups at a number of individual amino acid residues within the AR open reading frame. On the whole, a great deal of consistency is evident between the phenotypes reported for patients within the individual pedigrees. Complete agreement is observed between individual subjects in over 75% of the reported instances. Those mutations in which variations in phenotype have been noted account for less than a quarter of the amino acid replacements, and occur within the DNA- and hormone-binding domains of the receptor. While the genesis of these differences has not been established, it is tempting to speculate that the differences observed may reflect the influence of other genes or processes that serve to either accentuate or ameliorate the functional effects of the AR gene mutation.
Footnotes
1 The original work cited in this manuscript was supported by grants
from the National Institutes of Health (DK03892) and the Robert A.
Welch foundation (I-1090).
Received August 3, 1999.
Accepted August 12, 1999.
References