Metabolic Diseases Branch National Institute of Diabetes, Digestive, and Kidney Diseases National Institutes of Health Bethesda, Maryland 20892
Address all correspondence and requests for reprints to: Lee S. Weinstein, M.D., Metabolic Diseases Branch, National Institute of Diabetes, Digestive, and Kidney Diseases, National Institutes of Health, Building 10, Room 8C101, Bethesda, Maryland 20892-1752. E-mail: leew{at}amb.niddk.nih.gov
GS is one of about 20
heterotrimeric guanine nucleotide-binding proteins (G proteins) that
transmit signals from cell-surface receptors to effector enzymes or ion
channels to produce intracellular "second messengers." Each G
protein is composed of three subunits: , ß, and
. The
-subunit binds guanine nucleotide and is important for receptor
coupling and effector activation. In the inactive state the
-subunit
has GDP bound in its guanine nucleotide-binding site and is associated
with a ß
dimer. On activation by a ligand-bound receptor, GDP is
released and replaced by GTP. On binding GTP, the
-subunit switches
to an active conformation and dissociates from ß
. The
"turn-off" mechanism is an intrinsic GTPase activity of the
-subunit that hydrolyzes bound GTP to GDP, allowing reassociation of
the
-subunit with a ß
dimer. The Gs
-subunit (Gs
) is ubiquitously expressed and
couples receptors for many peptide and glycoprotein hormones, biogenic
amines, and other neurotransmitters and circulating factors to the
enzyme adenylyl cyclase, and is, therefore, required for
hormone-stimulated intracellular cAMP generation (1). Its
single copy gene GNAS1 is located at chromosome 20q13.
Surprisingly, Gs is the only G protein
-subunit involved in hormone signaling, to date, that has been shown
to have genetic defects associated with human disease. Somatic missense
mutations affecting two residues (Arg201 and
Gln227) that are catalytically important for the
GTPase turn-off reaction lead to constitutively active forms of
Gs
protein, and are present in
40% of
GH-secreting pituitary adenomas (2). In these tumors,
the increased intracellular cAMP resulting from activating
Gs
mutations stimulates both proliferation and
GH secretion.
Similar mutations affecting Arg201 are also
present in the McCune-Albright syndrome (MAS), which in its most severe
presentation is associated with hyperpigmented skin
(café-au-lait) lesions, fibrous dysplasia of bone, and both
increased growth and hormone oversecretion from various endocrine
organs (including gonads, adrenal cortex, thyroid, and pituitary
somatotrophs) (3, 4). The diverse manifestations of MAS
result from increased intracellular cAMP in melanocytes, osteoblastic
precursor cells, and endocrine tissues, respectively. In MAS the
somatic mutation presumably occurs during early embryonic development,
leading to a widespread distribution of cells bearing the activating
Gs mutation. The disease is never inherited,
indicating that activating Gs
mutations are
embryonically lethal. Similar mutations are also present in bone
lesions of patients who have fibrous dysplasia in the absence of other
manifestations of MAS and in skeletal muscle myxomas. Recent studies
suggest that increased cAMP leads to fibrous dysplasia by altering the
differentiation program of bone marrow stromal cells (5).
Cholera toxin, an exotoxin of Vibrio cholerae, catalyzes a
reaction resulting in covalent modification of
Gs
Arg201. This
posttranslational modification leads to an activated form of
Gs
protein and increased intracellular cAMP in
intestinal epithelial cells, which underlies the severe secretory
diarrhea that is characteristic of intestinal cholera.
Heterozygous inactivating Gs mutations lead to
Albright hereditary osteodystrophy (AHO), a disorder characterized by
short stature, brachydactyly, sc ossifications, centripetal obesity,
and, in some cases, mental deficits (1). Consistent with
the presence of heterozygous mutations, Gs
expression and/or function is decreased by
50% in erythrocytes and
other tissues obtained from AHO patients. Similar mutations have also
been found in patients with more aggressive ossifications that invade
deeper tissues (progressive osseous heteroplasia) (6).
In this issue of the JCEM, Ahrens et al.
(7) report on the genetic analysis of GNAS1 in
29 unrelated AHO patients with decreased erythrocyte
Gs bioactivity and show that 21 of these
patients have a heterozygous mutation within 1 of the 13
Gs
coding exons that presumably affects
Gs
expression or function. These findings are
consistent with the results of prior studies showing that the majority
of AHO patients have a Gs
mutation
(8). Among the 15 mutations identified in this study, 11
have not been previously reported, providing further evidence for the
heterogeneity of inactivating Gs
mutations.
This and prior studies demonstrate that such mutations can occur in any
of the Gs
encoding exons (except perhaps exon
3, which can be spliced out and still encode a biologically active form
of Gs
).
On the other hand, the study by Ahrens et al.
(7) also provides evidence that there may be some mutation
"hot spots." A 4-bp deletion mutation in exon 7 previously reported
in several AHO patients was present in five of their unrelated
patients. They also identified four other missense mutations that are
present in at least two unrelated patients and a mutation of a residue
(Arg231 to cysteine) that was previously shown to
be mutated to histidine in another patient.
Arg231 is important for
Gs activation (7), but it is
unknown if the other residues that are mutated in more than one patient
are functionally important. Iiri et al. (9)
showed that mutation of Ala366 in two unrelated
males leads to AHO and precocious puberty, due to increased release of
GDP in the basal state. At core body temperature the major effect of
Ala366 substitution is thermolability of the
protein (leading to AHO), but at the slightly lower temperature of the
testes the major effect is increased Gs signaling
due to greater GDP-GTP exchange (leading to gonadotropin-independent
precocious puberty). Other Gs
residues that
are mutated in single AHO patients have also been shown to be
functionally important (1). The failure of the study by
Ahrens et al. (7) to identify mutations
in eight patients may be due to the fact that some patients have
Gs
mutations within regulatory regions such as
the promoter. There are also technical limitations of the study
(sensitivity of the mutation screening method or inability to amplify
and analyze all of exon 1) that may also account for the failure
to identify mutations in these patients. Overall this study
demonstrates the usefulness of GNAS1 genetic analysis for
confirming the diagnosis of AHO.
Unlike MAS, AHO is inherited in an autosomal dominant manner. The
genetics of this disorder are complicated by the fact that some AHO
patients also present with renal resistance to PTH and milder
resistance to TSH and the gonadotropins [a condition referred to as
pseudohypoparathyroidism type Ia (PHPIa)] whereas other affected
patients within the same kindred have AHO without evidence of hormone
resistance [referred to as pseudopseudohypoparathyroidism (PPHP)].
Because the receptors for PTH, TSH, and the gonadotropins all activate
Gs, one might predict that
Gs mutations might lead to target-tissue
resistance to these hormones. However, simple haploinsufficiency of
Gs
due to heterozygous inactivating mutations
that are present in both PHPIa and PPHP patients cannot fully explain
why some patients develop multihormone resistance (PHPIa) while others
do not (PPHP). It cannot also explain why PHPIa patients seem to be
resistant to these hormones, but do not show resistance to other
hormones that also activate Gs in their
target-tissues (e.g. ACTH, vasopressin effects in the renal
collecting ducts).
The first clue to understanding this apparent paradox was provided by
Davies and Hughes (10), who noted that all offspring who
inherit AHO from their mother also develop multihormone resistance
(PHPIa) whereas those who inherit AHO from their father do not develop
multihormone resistance (PPHP). This inheritance pattern has been
supported by subsequent studies and is consistent with the results now
published by Ahrens et al. (7), in which 12 of
12 PHPIa patients from 11 unrelated kindreds inherited the
Gs mutation from their mother, regardless of
whether the mother had PHPIa or PPHP.
One possible explanation for the effect of parental inheritance on
phenotype of the offspring is that the Gs gene
is imprinted. Genomic imprinting is an epigenetic phenomenon affecting
a small number of genes that results in partial or total loss of
expression from one parental allele (11). Imprinted genes
contain one or more regions where the two parental alleles are
differentially methylated. Often, although not always, the
differentially methylated region is the gene promoter, which is
methylated on the transcriptionally silent allele. The imprint is
presumably erased in primordial germ cells and is reestablished in
either the male or female germ line during gametogenesis or before
pronuclear fusion in the zygote (the only time during development when
the two parental genomes are physically separated).
In AHO patients, if Gs is primarily expressed
from the maternal allele in specific hormone target tissues
(e.g. renal proximal tubules, the primary renal target for
PTH), then mutations inherited on the active maternal allele would
markedly reduce Gs
expression and lead to PTH
resistance (PHPIa) whereas mutations inherited on the inactive paternal
allele would have little effect on Gs
expression or renal PTH sensitivity (PPHP). This model is consistent
with the observation that the acute urinary cAMP response to
administered PTH is markedly reduced in PHPIa patients but is totally
unaffected in PPHP patients (12).
Gs
imprinting would have be tissue specific,
because Gs
has been shown to be biallelically
expressed in human lymphocytes (13) and fetal tissues
(14), and its expression is equally reduced by
50% in
several tissues from both PHPIa and PPHP patients (12),
consistent with no parent-of-origin effect in these tissues.
This model of tissue-specific imprinting of
Gs has been confirmed in a
Gs
knockout mouse model (15).
Analogous to patients with AHO, maternal inheritance of the
Gs
mutation resulted in biochemical
hypoparathyroidism and renal PTH resistance whereas paternal
transmission of the same mutation had no effect on renal PTH action.
Consistent with maternal-specific expression of
Gs
, Gs
was poorly
expressed in proximal tubules isolated from mice with the maternal
Gs
defect but was expressed normally in
proximal tubules isolated from mice with the paternal
Gs
defect. Interestingly,
Gs
appeared not to be imprinted in many other
tissues, including other portions of the kidney (glomeruli, distal
nephron, collecting ducts). Variable imprinting between the proximal
and distal nephron might explain why PHPIa patients are resistant to
the effects of PTH on the proximal portion of the nephron
(phosphaturia, 1
-hydroxylation of vitamin D) but do not seem to be
resistant to the effects of PTH on the distal nephron (calcium
reabsorption) (16). Tissue-specific imprinting of
Gs
might also explain why PHPIa patients do
not show resistance to all hormones that activate
Gs (e.g. vasopressin in the renal
collecting ducts). In humans Gs
has been
recently shown to be imprinted in the pituitary with expression only
from the maternal allele (17).
The Gs imprinting story is further complicated
by the fact that the GNAS1 gene produces multiple other gene
products through the use of at least four alternative promoters and
first exons, which themselves are imprinted (see Fig. 1
; reviewed in Ref. 1). All
four alternative first exons splice onto a common set of downstream
exons (exons 213). The most downstream first exon (exon 1) generates
transcripts encoding Gs
. Its promoter is not
methylated. Alternative exons located 47 and 35 kb upstream of
Gs
exon 1 generate transcripts encoding the
chromogranin-like protein NESP55 and XL
s (a
Gs
isoform with a long amino-terminal
extension), respectively. Both are primarily expressed in
neuroendocrine tissues and are oppositely imprinted: NESP55 is only
expressed from the maternal allele and its promoter is methylated on
the paternal allele while XL
s is only expressed from the paternal
allele and its promoter is methylated on the maternal allele. Little is
known about their biological function. A fourth alternative first exon
(exon 1A) generates transcripts that are ubiquitously expressed but do
not appear to encode a functional protein. Like XL
s, this exon is
methylated on the maternal allele and is transcriptionally active on
the paternal allele. In mice maternal-specific methylation of the exon
1A region is established during oogenesis and is maintained
throughout development, suggesting that this region may be critical for
establishment of GNAS1 imprinting. Recently paternally
expressed antisense transcripts that traverse the NESP55 upstream exon
have also been identified.
|
The defect commonly found in PHPIb patients is the abnormal imprinting
of the exon 1A region, suggesting that this defect is important for the
pathogenesis of PHPIb. While it is possible that PHPIb is the direct
consequence of exon 1A-specific mRNA overexpression, this seems
unlikely given the fact that these transcripts do not encode a known
functional protein and the central role that
Gs plays in PTH signaling. More likely, the
exon 1A region is an important element for the tissue-specific
imprinting of Gs
, and, therefore, abnormal
imprinting of this region leads to loss of Gs
expression in specific tissues. One hypothetical model (shown in Fig. 2
) would predict that the exon 1A region
contains a silencer element which is a binding site for a
tissue-specific repressor protein that is expressed in only some
tissues, such as renal proximal tubule. In these tissues the repressor
binds to the paternal allele and inhibits transcription from the
Gs
promoter, but is unable to bind to the
maternal allele because its binding site is methylated, therefore,
allowing Gs
to be expressed from this allele.
In most other tissues the repressor is not expressed, and, therefore,
Gs
is expressed biallelically, even though the
exon 1A methylation pattern is the same. In PHPIb patients exon 1A is
not methylated on the maternal allele, allowing the repressor to bind
to both alleles in proximal tubules, leading to near total loss of
Gs
expression and PTH resistance. In contrast,
Gs
expression is unaffected in most other
tissues where the repressor is not expressed. Other potential
mechanisms for the tissue-specific imprinting of
Gs
have been proposed elsewhere
(1).
|
The models for the maternal inheritance of renal PTH resistance in AHO
(PHPIa) and PHPIb described above are all predicated on the assumption
that Gs is expressed primarily or exclusively
from the maternal allele in renal proximal tubules, the primary site of
PTH action in the kidney. Although this has been shown to be true in
mice (15) and Gs
has been shown
to be expressed only from the maternal allele in human pituitaries
(17), there is still no direct evidence that
Gs
is imprinted in renal proximal tubules in
humans. In this issue of JCEM, Zheng et al.
(23) try to address this question by examining the
allele-specific expression of Gs
and the other
GNAS1 gene products in renal cortex and other tissues
derived from human fetuses that were heterozygous for an informative
polymorphism within an exon common to all GNAS1 sense
transcripts. After isolating RNA from these tissues, they amplified
NESP55, XL
s, exon 1A, and Gs
-specific
transcripts by RT-PCR and determined whether or not each were mono- or
biallelically expressed. As expected, NESP55 was expressed from one
parental allele (presumably maternal) whereas XL
s and exon 1A
transcripts were expressed only from the opposite (presumably paternal)
allele in all tissues examined. This is consistent with previous
results and the known methylation patterns of their respective
promoters (see Fig. 1
). However, Gs
seemed to
be biallelically expressed in all kidney cortex samples examined. Based
on this observation, the authors conclude that
Gs
is not imprinted in kidney cortex and,
therefore, the PTH resistance in PHPIb (and presumably PHPIa) is not
due to loss of Gs
expression in renal proximal
tubules.
There are at least two other potential explanations besides lack of
Gs imprinting in renal proximal tubules for
the failure of Zheng et al. (23) to detect
Gs
imprinting in renal cortex. The first
possibility is that imprinting was not detected because their cortex
samples contained many elements in addition to proximal tubules,
including glomeruli, condensing mesenchyme, developing nephron
structures, and medullary tubular structures.
Gs
has been shown to be biallelically
expressed in glomeruli and more distal (including medullary) segments
of the nephron (16), and is likely to also be
biallelically expressed in condensing mesenchyme and developing nephron
structures. Our studies in mice showed that Gs
mRNA expression is very low in proximal tubules relative to neighboring
structures (15). If this is also true in humans, then only
a small fraction of Gs
mRNA in their samples
may have been derived from proximal tubules, even if proximal
tubules were highly represented in their samples.
The second possible explanation for the failure to detect imprinting in
the study by Zheng et al (23) relates to the
fact that fetal renal cortex is not the same as fully mature
(postnatal) renal cortex. Histologically, nephrons and glomeruli are
not fully developed during the fetal period and continue to mature even
during the postnatal period. This is consistent with the presence of
condensing mesenchyme and developing nephron structures in their cortex
samples. It is possible that Gs is not
imprinted in this immature fetal cortical tissue but becomes imprinted
as the proximal tubules fully mature in early postnatal development.
This is one possible explanation for why PHPIa patients do not show
evidence of renal PTH resistance at birth, but subsequently develop PTH
resistance over the first years of life (24, 25, 26). There
are other examples of genes whose imprinting is developmentally
regulated (27). In the repressor model described above,
the repressor may only be expressed in proximal tubules once they are
fully differentiated in the postnatal period. In my view, the large
weight of clinical and experimental evidence in both humans and mice
suggests that Gs
is imprinted in postnatal
renal proximal tubules. The definitive answer as to whether this is
true in humans awaits studies that examine allele-specific expression
of Gs
in proximal tubules that are isolated
from postnatal kidneys.
Recent evidence suggests that imprinting of Gs
may also have an effect on the clinical manifestations in patients with
activating Gs
mutations. In MAS patients, one
would expect that in tissues where Gs
is
imprinted the expression of the constitutively activated form of
Gs
would be much greater when the mutation is
present on the active maternal allele. Therefore, the manifestations
within an individual patient might be a function of both the
distribution of cells bearing the mutation and the parental allele that
has the mutation. Evidence for this is provided by Hayward et
al. (17), who recently showed that in 21 of 22
GH-secreting pituitary tumors with an activating
Gs
mutation, the mutation was on the maternal
allele.
In summary, our present knowledge suggests that tissue-specific
imprinting contributes to the multiple clinical manifestations that
result from genetic defects involving Gs.
Future studies in both mouse and humans will allow us to determine the
physiological mechanisms by which Gs
deficiency leads to the AHO phenotype, the mechanisms underlying the
complex imprinting of GNAS1, and how this mechanism goes
awry in PHPIb. Clinical genetic studies in patients with diseases such
as MAS, AHO, and PHPIb provide a good example of how important insights
into gene regulation and protein function can be derived from careful
examination of patients with relatively uncommon diseases.
Acknowledgments
Footnotes
Abbreviations: AHO, Albright hereditary osteodystrophy; MAS, McCune-Albright syndrome; PHPIa and Ib, pseudohypoparathyroidism types Ia and -Ib; PPHP, pseudopseudohypoparathyroidism.
Received July 30, 2001.
Accepted July 30, 2001.
References