Primary Adrenal Insufficiency: New Genetic Causes and Their long-Term Consequences

Federica Buonocore; John C. Achermann


Clin Endocrinol. 2020;91(1):11-20. 

In This Article

Nuclear Receptors in Adrenal Development: DAX-1 (NR0B1) and SF-1 (NR5A1)

DAX-1 (officially termed NR0B1) and steroidogenic factor 1(SF-1, officially termed NR5A1) belong to the nuclear receptor superfamily. Humans have 41 different nuclear receptors, 20 of which currently have associated clinical conditions.[4] DAX-1 and SF-1 are the two nuclear receptors that regulate both adrenal and reproductive development and function.[5]

DAX-1 (NR0B1)

Pathogenic variants in NR0B1/DAX-1 were first reported as a cause of X-linked adrenal hypoplasia congenita (AHC) in 1994.[6] Efforts to localize the gene were helped by the co-existence of adrenal hypoplasia with Duchenne muscular dystrophy due to a contiguous gene deletion syndrome on the short arm of the X chromosome (Xp21). More than 300 individuals and families with X-linked AHC due to loss of NR0B1/DAX-1 have been reported.[5,7]

The classic clinical features of X-linked AHC include primary salt-losing adrenal insufficiency, hypogonadotropic hypogonadism (HH) and infertility. Boys may also present with predominantly either mineralocorticoid or glucocorticoid insufficiency, or may have paradoxical features such as macrophallia or early puberty.[8,9] One report of fertility in a man with X-linked AHC using testicular sperm extraction-intracytoplasmic sperm injection (TESE-ICSI) has been published.[10] Making the specific diagnosis is important so that associated features can be monitored and treated. The risk of presymptomatic adrenal insufficiency in brothers and males in the maternal family needs to be considered.[11]

Since 2000, several reports of late-onset X-linked AHC in men with PAI have emerged.[12–16] Usually, this condition is associated with partial HH but infertility might be the main feature. Often the genetic change involves a 5' (aminoterminal) stop variant in the gene and translation of a shorter protein with partial function, or a partial loss-of-function variant in the ligand-like binding domain of NR0B1 (Figure 1A, 1).[12] One recent review of adult men with PAI in a single UK centre identified two patients with X-linked AHC due to partial loss-of-function variants in DAX-1, suggesting that late-onset X-linked AHC may be underdiagnosed in the adult population (Figure 1B & 1).[16]

Figure 1.

Late-onset X-linked adrenal hypoplasia congenita (AHC) due to variations in DAX-1/NR0B1. (A) Cartoon showing the protein structure with selected variants associated with late-onset adrenal insufficiency highlighted. Black arrowheads represent repeat motifs. (B) Model of DAX-1 bound to LRH-1, a homologue of steroidogenic factor-1 (SF-1/NR5A1), with Pro279 indicated. The amino acid change p.P279L alters a hydrogen bond at the periphery of a key interaction domain and is associated with a partial phenotype. Disruption of codon L278 at the core of the interaction domain is associated with classic early onset X-linked AHC. (C) Selected testicular histology from men with late-onset phenotypes includes oligospermia (not shown), maturational arrest at a primary spermatid stage (left panel); or more severe atrophic, hyalinized tubules, Sertoli cell only features and Leydig cell pseudohyperplasia (right panel). LBD, ligand binding domain; HH, hypogonadotropic hypogonadism. Panels (B) and (C) modified from Kyriakakis N, Shonibare T, Kyaw-Tun J, et al Late-onset X-linked adrenal hypoplasia (DAX-1, NR0B1): two new adult-onset cases from a single centre. Pituitary. 2017;20(5):585–593 © The Authors (

SF-1 (NR5A1)

Steroidogenic factor-1 (NR5A1) (located on 9q33) is another nuclear receptor that regulates adrenal and reproductive development. Targeted deletion of the gene encoding Nr5a1 in the mouse causes adrenal and gonadal dysgenesis; children with a similar phenotype of adrenogonadal dysfunction were first reported in 1999 and 2002.[17–19] These individuals had variants affecting key DNA-binding regions of SF-1 (P-box and A-box).[20]

Since these first publications, more than 250 individuals with pathogenic variants in NR5A1/SF-1 have been reported.[5] These changes are usually heterozygous de novo variants but can occur in a "sex-limited dominant" pattern; in this situation, an unaffected woman carries a heterozygous change and passes it to affected 46,XY children, thereby resembling an X-linked condition. Pathogenic SF-1 variants are associated with a spectrum of phenotypes in 46,XY subjects including testicular dysgenesis/dysfunction (46,XY differences/disorders in sex development), severe hypospadias (accounting for approximately 5%-7% of cases) and male factor infertility (1%-2%).[21–28] Although data are limited, it has been proposed that this subset of infertile men could develop hypogonadism and low testosterone with time, so this may represent a group who need longer term endocrine follow-up.[25]

Analysis of larger pedigrees where individuals with 46,XY testicular dysfunction were found together with 46,XX women with ovarian insufficiency (POI) has revealed that defects in SF-1 can affect human ovary function too.[29] Loss-of-function variants in NR5A1 are now well-established in familial POI, but occur less commonly in sporadic (nonfamilial) POI or secondary amenorrhoea (1%-2%).[22,28–32] Of note, variants in a specific amino acid in the A-box of SF-1 (p.R92) are found in 46,XX ovotesticular DSD (ovotestes or testis), suggesting that very localized alterations in this key transcriptional regulator can "switch" ovary development into a testis development pathway in humans.[33–35]

Despite these wide-ranging effects on reproductive function, SF-1 variants causing adrenal dysfunction are comparatively rare. Only six children have been reported to date to have SF-1-associated adrenal dysfunction, usually with variants in p.G35 or p.R92.[18,19,36–38] It remains to be seen whether adrenal dysfunction will occur progressively in individuals with reproductive dysfunction due to defects in SF-1; available insights currently suggest that this is not the case, although longer term systematic follow-up studies are needed. Therefore, it seems that human gonadal function is more sensitive to haplo-insufficiency or partial loss of SF-1 activity than adrenal function.