BoneKEy-Osteovision | Commentary
Because your mother told you so
DOI:10.1138/2001011
Resistance to the actions of parathyroid hormone (PTH) comes in three flavors. Deletion of the PTH/PTHrP receptor is lethal, because the developmental effects of PTHrP in cartilage do not occur. Loss of one allele of the gene for the coupling protein Gsα produces a milder phenotype, pseudohypoparathyroidism type 1A (PHP1A), with resistance to PTH and a characteristic set of somatic features - short stature, brachydactyly, and subcutaneous ossification - collectively known as Albright's hereditary osteodystrophy (AHO). Isolated PTH resistance without a somatic phenotype is called pseudohypoparathyroidism type 1B (PHP1B), and its cause is substantially worked out by the paper of Liu et al ().
In families with PHP1A there are members who have the somatic phenotype AHO without PTH resistance. Albright coined the tongue-twister pseudopseudohypoparathyroidism for this paradoxical condition, which may turn out to be the key to understanding the PTH resistant syndromes. Here's how: Nearly every PHP1A patient inherits PTH resistance and AHO from the mother; paternal transmission results in the somatic phenotype AHO only, i.e. pseudopseudo-hypoparathyroidism (). This suggests that imprinting of Gsα gene expression is involved. The argument goes like this: If Gsα is expressed only from the maternal allele in the renal proximal tubule of humans, as it is in the mouse (), then neither allele would be expressed in this target tissue in PHP1A patients, who inherited a null allele from their mothers, producing severe renal resistance to PTH. The imprinting of Gsα would have to be relatively tissue-specific, because both normal and mutant alleles are expressed in other tissues, e.g. erythrocytes, giving 50% of normal Gsα levels in PHP1A erythrocytes.
Although tissue-specific imprinting of Gsα gene expression has yet to be shown in humans, this explanation is very appealing - not only would it explain the exclusively maternal inheritance of PHP1A and the paternal inheritance of pseudopseudohypoparathyroidism, but it offers an explanation for another perplexing aspect of PHP1A, the comparatively normal response to hormones other than PTH whose receptors are coupled to Gs: receptors in tissues where imprinting did not occur could be coupled to adenylyl cyclase relatively normally by the expression of 50% of normal Gsα levels.
Juppner et al have shown that PHP1B maps to the same region as the Gsα locus on human chromosome 20q13.3 (), and the Weinstein group now provides strong, albeit indirect evidence that PHP1B is also caused by an imprinting defect affecting Gsα. The GNAS1 gene, which encodes Gsα, also has three additional upstream promoters. In the mouse, the exon 1A promoter, which encodes a transcript of unknown function, lies within a differentially methylated region. This region is methylated on the maternal allele, and exon 1A is expressed only from the unmethylated paternal allele () (Fig. 1).
Liu et al now report that the exon 1A region of the human GNAS1 gene is differentially methylated in humans as in the mouse. However, both alleles of GNAS1 are unmethylated (the paternal pattern) in all 13 patients with PHP1B who were studied. In some patients the disorder was inherited; in most it was sporadic, but the methylation pattern is similar in both groups. Although there are two other differentially methylated regions in this complex gene, methylation of neither of them was consistently abnormal in patients. This strongly suggests that abnormal imprinting of gene expression resulting from abnormal methylation of exon 1A underlies PTH resistance in PHP1B.
How would upstream imprinting of the GNAS1 gene affect expression of Gsα? One possibility involves activation of the Gsα promoter by an upstream enhancer. The Igf2 gene is also imprinted, and methylation of a boundary element in the upstream region of the Igf2 promoter has recently been shown to cause imprinting of that gene. Methylation prevents the binding of CTCF, a protein that insulates the Igf2 promoter from the action of an upstream enhancer element by binding to boundary elements (6;7). In the case of the GNAS1 gene, the proposal is that action of an enhancer would be blocked by a methylation-sensitive “boundary element” in the differentially methylated region around exon 1A. Methylation of this region would silence the boundary element in the maternal allele, permitting expression of Gsα(Fig 1). This explanation would require that expression of Gsα be enhancer-dependent only in kidney and other imprinted tissues, and independent of the enhancer elsewhere. A formally equivalent argument can also be made for methylation-sensitive repression by an tissue-specific upstream repressor.
How does abnormal methylation occur in patients? It is highly likely that sporadic cases result from failure to switch from the paternal to the maternal imprint in the oocytes, as normally occurs during gametogenesis. Inherited cases would presumably have a mutation affecting exon 1A methylation. The mechanism for imprinting proposed here would also account for severe renal resistance to PTH in PHP1A, in which the maternal allele would have a null mutation and the unmethylated paternal allele would not be expressed.
The mechanism for renal imprinting of Gsα and its derangements in PHP1B remains to be worked out in detail. The issue of imprinting in bone also remains for future study. Is renal resistance to PTH sufficient to cause hypocalcemia, or is resistance in bone also required? If resistance to the direct actions of PTH in bone is required for hypocalcemia, is the expression of Gsα in bone imprinted? Could differential imprinting be involved in the paradoxical presence of hyperparathyroid bone disease in some PHP1A and PHP1B patients? In any case, the article by Liu et al makes it clear that methylation-dependent imprinting will be a large part of the PHP story.
This work is licensed under a
Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License.