Department of Pediatrics Oregon Health & Science University Portland, Oregon 97201
As is frequently the case with exciting discoveries, the identification of the SHOX gene has raised more questions than it has answered. Thus, recent observations concerning SHOX (short stature homeobox-containing gene) and, in particular, genotype-phenotype correlations (see Ref. 1) have identified it as a gene critical to normal growth, but have also resulted in uncertainty concerning the contribution of SHOX abnormalities to the short stature of Turner syndrome (TS) and idiopathic short stature (ISS).
Experience with deletions of the short arm of either the human X or Y chromosome suggested that a gene(s) relevant to normal growth resided within the pseudoautosomal region 1 (PAR1) of the sex chromosomes, a region of the X chromosome that normally escapes X inactivation. Ogata and Matsuo (2) demonstrated that this gene was located in the most distal 700 kb of PAR1, and two groups subsequently identified a single homeobox gene within this region, which was named SHOX by Rao et al. (3) and PHOG (pseudoautosomal homeobox-containing osteogenic gene) by Ellison et al. (4). The gene consists of seven exons over a genomic region of 40 kb, encodes two transcripts resulting from alternative splicing, and has been shown to be highly expressed in human embryos in osteogenic tissue, as well as in the first and second pharyngeal pouches (5).
Given its potential role in skeletal growth, Rao et al. (3) searched for SHOX mutations in 91 children with ISS, defined by the following criteria: 1) height below the 3rd percentile for chronological age; 2) no identifiable skeletal, endocrine, or psychiatric disorders; 3) normal food intake; and 4) normal karyotype. One subject was found to have a point mutation of SHOX resulting in a premature stop codon. On this basis, the authors speculated that 1% of children diagnosed with ISS could have abnormalities of the SHOX gene. Subsequently, two groups identified a linkage between SHOX abnormalities and Leri-Weill dyschondrosteosis (LWD), an autosomal dominant form of mesomelic dysplasia characterized by short stature, short arms and legs, and bowing of the radius resulting in Madelung deformity (6, 7). In a total of 14 affected families, gene deletions were identified in 12 and nonsense mutations in 2. Schiller et al. (8), on the other hand, identified SHOX deletions in only 10 of 18 LWD families, and no mutations. Their failure to identify SHOX abnormalities in 8 of 18 families may reflect, however, a lack of sensitivity of single-strand conformation polymorphisms, unidentified mutations affecting SHOX expression, different etiologies for the LWD phenotype, or phenotypic heterogeneity, as is occasionally observed in haploinsufficiency syndromes. It is of note that in the current report by Ross et al. (1), abnormalities of SHOX were identified in all 21 LWD families, with 37.5% of families having mutations and the remaining 62.5% having gene deletions.
In addition to their detailed and systematic analysis of genetic abnormalities, Ross et al. (1) have provided an extensive assessment of phenotypic features of LWD, as well as evaluated genotype-phenotype correlations. In this regard, it is important to point out that their study is characterized, inevitably, by ascertainment bias, because the study was limited to families that had been diagnosed, on clinical criteria, to have LWD. Consequently, the possibility that SHOX abnormalities might be found in subclinical patients with LWD, although acknowledged, could not receive proper evaluation, outside of the 21 families evaluated. Despite this caveat, the study has yielded important new insights concerning SHOX. Thus, although the mean height was -2.2 SD, the range was from -4.6 to 0.6 SD, with 49% of the subjects having heights within the normal range. Arms and legs, in general, were more affected than trunk, indicating some degree of mesomelia. No correlation of statural characteristics was found with chronological age, nor with various mutations or gene deletions. Other features identified included Madelung deformities (74% of subjects), high-arched palate (59%), cubitus valgus (71%), and scoliosis (22%).
It is clear that LWD has significant similarities and differences from TS, and that comparison of these two disorders provides important insights on the genetic regulation of growth. Madelung deformities, palatal abnormalities, wide carrying angle, and scoliosis are all characteristic of TS, although Madelung deformities appear much more common in LWD than in TS, possibly reflecting a role for estrogen in accentuating this skeletal abnormality. This hypothesis has been suggested by Kosho et al. (9), who also noted the female-dominant and age-dominant nature of Madelung deformities in LWD.
Inevitably, analysis of LWD engenders questions concerning its relationship with TS and ISS. Although it is certain that SHOX haploinsufficiency is involved in the short stature and skeletal abnormalities of TS, it is unlikely that it explains fully the growth abnormalities of TS. As noted by Ross et al. (1), the height deficit in LWD, reflected in a mean height of -2.2 SD, is considerably lower than that observed in TS (-3.2 SD). Indeed, these numbers probably reflect an overestimation of the contribution of SHOX haploinsufficiency to the growth failure of TS, since patients with this condition, because of ovarian failure, have delayed puberty, a prolonged growth phase during adolescence, and late epiphyseal fusion. Thus, although LWD patients show no correlation between height z-score and chronological age, the typical growth pattern of TS is a progressive deviation away from normal growth curves, with a partial catch-up during the prolonged adolescent period. Nevertheless, while it is highly likely that loss of other X-chromosome genes contributes to the growth failure of TS, a clinical trial of GH therapy in LWD appears indicated, given its success in girls with TS (10, 11).
The relationship between LWD and ISS is also of interest. Ross et al. (1) have reported a 3:1 female to male preponderance in LWD and have suggested that male patients may be underdiagnosed, since their growth is as affected as in females. Part of the female to male preponderance can be explained, however, by the fact that females can obtain an abnormal SHOX gene from either their mother or father, whereas males can only receive an abnormal SHOX gene from their mothers. Nevertheless, it would seem likely that some cases of LWD may be diagnosed as ISS or as familial short stature (FSS), especially given the phenotypic heterogeneity of these conditions. The initial report by Rao et al. (3) described one patient with a SHOX mutation in 91 individuals with ISS. Even if only 1% of ISS/FSS patients prove to have SHOX abnormalities, this still represents a relatively large number of patients. Even more important, however, is that the identification of SHOX abnormalities in some individuals with ISS/FSS underscores the fact that all cases of short stature must have, inevitably, a molecular and/or environmental explanation. Heterozygosity for certain mutations of the GH receptor gene, for example, have been proposed as an etiology of some cases of ISS, and it is likely that subtle abnormalities of the GH receptor, GH signaling, or IGF generation may play a role in other individuals with growth failure. Thus, the elucidation of the role of SHOX in the growth failure of LWD and TS adds to our understanding of the multifactorial nature of normal growth, and should serve to stimulate a search for other molecular abnormalities as potential etiologies of short stature.
Acknowledgments
Footnotes
Address all correspondence and requests for reprints to: Ron G. Rosenfeld, M.D., Department of Pediatrics, Oregon Health & Science University, 707 SW Gaines Road, CDRC-P, Portland, Oregon 97201.
Abbreviations: FSS, Familial short stature; ISS, idiopathic short stature; LWD, Leri-Weill dyschondrosteosis; TS, Turner syndrome.
Received October 23, 2001.
Accepted October 23, 2001.
References