Mattel Childrens Hospital at UCLA and Department of Pediatrics, UCLA School of Medicine, Los Angeles, California 90095-1752
Address all correspondence and requests for reprints to: Edward R.B. McCabe, M.D., Ph.D., Department of Pediatrics, 22-412 MDCC, UCLA School of Medicine, 10833 Le Conte Avenue, Los Angeles, California 90095-1752. E-mail: emccabe{at}mednet.ucla.edu
The development of the steroidogenic axis represents a complex system involving multiple, intricately related steps that function in an exquisitely well-timed manner (1). The high frequency with which the networked steps produce a normally formed and functioning hypothalamic-pituitary-adrenal/gonadal axis is a tribute to the robust nature of this complex system. Diseases, such as the adrenal hypoplasia congenita (AHC) and hypogonadotropic hypogonadism (HHG) associated with DAX-1 mutations (2, 3, 4, 5), demonstrate specific sites of vulnerability within the robust network. The complexity of this system also explains the difficulty in precisely predicting phenotype from genotype among DAX-1 mutations, as is true for many other disorders (6, 7, 8, 9).
DAX-1 mutations disrupt steroidogenic axis development
AHC may be caused by deletion of the Xp21 DAX-1 gene as part of a contiguous gene syndrome or by intragenic mutations (1, 10). Mutations within the DAX-1 gene showed that AHC and HHG are both caused by alterations in this same gene (2), consistent with the observed expression of DAX-1 throughout the steroidogenic axis (11, 12). Subsequent investigations demonstrated that HHG was due to defects in both hypothalamic and pituitary function (13). DAX-1 may also have a role in spermatogenesis (14, 15, 16).
We have recently reviewed the 86 reported intragenic DAX-1 mutations (17). Frameshift and nonsense mutations are distributed throughout the gene. However, amino acid-altering mutations are observed almost always in the region encoding the C-terminal portion of DAX-1 with amino acid sequence similarity to the hydrophobic core of the ligand binding domain in other nuclear receptor family members (18).
DAX-1 has a critical role in adrenal cortical development, and most individuals with DAX-1 mutations present with adrenal insufficiency in the first months or years of life (1). However, occasional patients may develop signs and symptoms later in childhood and others may not be symptomatic until adulthood (1, 5, 16, 19). Although these later presentations are frequently due to adrenal insufficiency (1), symptoms associated with HHG may represent the patients chief complaint (5, 19).
An issue that remains to be determined is whether these differences in clinical presentation are consequent to distinct functional activities caused by specific DAX-1 mutations or to additional genetic and environmental influences on mutant DAX-1 function. For other simple Mendelian disorders it seems that the individual patients phenotype results from the interaction of the primary genetic abnormality with modifying influences from the individuals genetic background and environmental experiences (6, 7, 8, 9). The complexity of the dynamic transcriptional network within which DAX-1 is embedded provides ample opportunities for modifier genes.
DAX-1 functions within a complex network
DAX-1 is critical for normal development of the steroidogenic axis as evidenced by AHC and HHG in patients with DAX-1 mutations. DAX-1 also seems to be involved in sex determination. DAX-1 is within a 160-kb critical region of Xp21.3 that when duplicated leads to dosage-sensitive sex reversal and genotypic 46,XY individuals who have phenotypically female external genitalia or sexual ambiguity (20, 21, 22). Transgenic overexpression of Dax-1 on a murine background carrying the Y chromosome from the Poschiavanus mouse strain, which expresses a reduced Sry functional level, results in phenotypically female XY animals, an observation consistent with DAX-1 duplication being responsible for dosage-sensitive sex reversal (23). DAX-1, therefore, seems to be involved in the early events leading to sex determination, as well as in the development of the steroidogenic axis (24, 25).
DAX-1 is also expressed in murine embryonic stem cells (26). Whereas isolated expression of DAX-1 would not indicate a functional role in embryonic stem cells, the expression of other genes thought to network with DAX-1 (27) suggests participation of DAX-1 and related proteins in this early developmental cell type before initiation of sex determination or the development of the steroidogenic axis.
DAX-1 participates in a highly complex transcriptional network. The DAX-1 promoter contains a functional steroidogenic factor-1 (SF-1) response element (28, 29). DAX-1 contains a C-terminal transcriptional silencing domain that inhibits SF-1-mediated transactivation (30, 31, 32). DAX-1 recruits the corepressor N-CoR to SF-1 during transcriptional silencing (31), and also interacts with the corepressor, alien (33). The synergistic interaction of SF-1 and Wilms tumor-1 proteins, to promote Müllerian inhibiting substance expression during male sexual development, is antagonized by DAX-1 (34). FSH down-regulates DAX-1 expression in cultured Sertoli cells (35). DAX-1 has an LXXLL motif and acts as a corepressor for activated estrogen receptors (36). DAX-1 interacts with DNA hairpins in promoters for the DAX-1 and steroidogenic acute regulatory protein (STAR) genes to regulate their expression in vitro (37), and also binds mRNA (38).
Analysis of complex networks reveals robust properties and vulnerabilities
DAX-1 is embedded in a complex network involved in a variety of developmental processes. Analysis of such complex biological networks indicates that they are generally scale-free networks with hub-and-spoke structures (9, 39, 40). Such scale-free networks are extremely robust and tolerate high degrees of component failure without network compromise (9). These networks are vulnerable, however, to genetic mutations in highly connected nodes or in trunk lines that provide pathways between highly connected nodes (9). DAX-1 clearly represents such a highly connected node, interacting with numerous network components and influencing a variety of developing systems.
Roles for DAX-1 and its partners as highly connected nodes critical for mammalian development would be consistent with the observation that human SF-1 mutations also result in abnormal adrenal cortical and male sex determination. A heterozygous missense (G35E) SF-1 mutation results in partial loss of SF-1 function in a 45,XY phenotypically female patient with adrenal insufficiency (41, 42). A second individual who is a prepubertal 45,XX female with a transcriptionally inactive heterozygous mutation (R255L), has adrenal insufficiency and phenotypically normal female external genitilia (43). The phenotypes of these patients support the importance of dosage among the key components in this network and are consistent with the critical importance of DAX-1SF-1 interactions in adrenal development and sex determination.
Synergistic heterozygosity (44) is a concept that is valuable in our consideration of the pathogenic roles of DAX-1 and its network partners. This concept evolved to improve the understanding of the pathogenesis of biochemical genetic disorders and this evolution was facilitated by the accumulation of human genomic databases. Patients with phenotypes characteristic of specific autosomal recessive inborn errors of metabolism initially showed only single heterozygous alleles when the genes identified with the individual patients phenotypes were sequenced. As additional genes in those biochemical pathways were cloned and sequenced, the investigators examined these additional genes in their patients. They found that heterozygosity at two steps in a pathway could result in a phenotype equivalent to an individual homozygous at a single step. Their observations are consistent with metabolic control analysis, which argues that there is no single rate-limiting step in most pathways (7). The behavior of complex systems suggests that compromise of two steps in a network may impact cumulatively if these changes perturb activity through a trunk line or alter flux at a critical node (9). Therefore, as we consider patients with phenotypes involving developmental abnormalities, for example, in the steroidogenic axis and/or sex determination, it is important to recognize that partial defects in two or more proteins in a network may result in phenotypes similar to those observed among individuals with a complete defect involving a single protein in the network.
Conclusions regarding phenotypic variability
Mantovani et al. (5) have shown that HHG and a compensated, covert adrenal insufficiency in adulthood may be associated with a partial loss-of-function DAX-1 mutation. The late presentation of this patient may be due to the functional activity of the mutant protein above a threshold level, below which an earlier presentation with more obvious AHC would be anticipated (7). However, such a clear-cut genotype-phenotype correlation is observed only with extreme rarity among genetic disorders (6, 7, 8, 9).
The phenotypes of patients with simple Mendelian disorders are, in fact, complex traits, and their complexity is due to protein activity thresholds, modifier genes and environmental influences, and the systems dynamics within complex proteomic networks (7, 8, 9). One consequence of this complexity is that an individuals genotype at a single locus, such as DAX-1, is unlikely to be an accurate predictor of that patients phenotype. The capacity to predict accurately an individual patients clinical presentation and course will require a much more thorough understanding of that individuals genomic sequence variations and environmental experiences. In addition, such a predictive ability will require a detailed knowledge of the structural composition and functional dynamics of the relevant proteomic networks. Systematic accumulation of information on individuals and their biology will permit eventually an improved understanding of phenotypic variability.
Acknowledgments
Footnotes
This work was supported in part by funding from the NIH (Grants RO1-HD39322 and R01-HD22563).
Abbreviations: AHC, Adrenal hypoplasia congenita; HHG, hypogonadotropic hypogonadism; SF-1, steroidogenic factor-1.
Received November 5, 2001.
Accepted November 7, 2001.
References