Our original Gnas knockout mouse model was an insertion in Gnas exon 2, which we used to establish that Gs-alpha is imprinted in a tissue-specific manner. This explains why hormone resistance in these mice and in Albright hereditary osteodystrophy patients only develops after maternal inheritance of the Gs-alpha mutation. Homozygotes were embryonically lethal while heterozygotes had distinct phenotypes depending on whether the mutation was on the maternal or paternal allele. Maternal mutants had reduced survival, perinatal subcutaneous edema, and developed obesity in association with reduced metabolic rates and activity levels. In contrast paternal mutants failed to suckle after birth, had a high rate of lethality soon after birth, and developed a severely lean, insulin-sensitive phenotype associated with increased activity levels and metabolic rates. However interpretation of these initial results was limited by the fact that the exon 2 mutation is common to all Gnas gene product transcripts and therefore we could not determine which gene product deficiency caused the observed phenotypes. Therefore we and others have subsequently developed models in which specific Gs-alpha gene products have been disrupted to determine the effects of knocking out individual gene products. We studied XL-alpha-s deficient mice (provided by Kelsey and Plagge, Babraham Institute, UK) in detail and showed that virtually all the features observed in paternal exon 2 mice (with the exception of the increased activity levels) were the result of XL-alpha-s deficiency. Further we showed that these mice had increased metabolic activation of brown adipose tissue due to increased sympathetic nervous system activity. These changes were not due to cell-autonomous effects in adipocytes as these mice did not show increased metabolic responsiveness to a beta-3 adrenergic receptor agonist which specifically activates adipose tissue. Moreover, we showed that XL-alpha-s is normally only expressed in adipose tissue during the early postnatal period. We conclude that XL-alpha-s deficiency normally acts in the central nervous system to inhibit sympathetic activity. Whether XL-alpha-s plays a similar role in humans is unclear. We have recently generated mice with specific deficiency of Gs-alpha by placing loxP recombination sites around Gs-alpha exon 1 (floxed Gs-alpha mice). Using these mice we generated mice with germline deletion of Gs-alpha exon 1 (E1 mice). As with exon 2 mice, homozygotes were embryonically lethal. Mice with the E1 deletion only on the maternal allele (E1m-) had perinatal subcutaneous edema, reduced survival (although not to the same extent as E2m- mice), and developed severe obesity, insulin resistance, and hypertriglyceridemia. In contrast, mice with paternal E1 deletion (E1p-) had normal survival, no perinatal subcutaneous edema, and developed very mild insulin resistance and increase in adiposity. The reason that E1p- mice do not develop the severe phenotype observed in E2p- mice is that XL-alpha-s expression is not disrupted in these mice. We hypothesize that features specifically observed in E1m- mice and not in E1p- mice (eg. perinatal subcutaneous edema, lethality, severe obesity and insulin resistance) are due to severe Gs-alpha deficiency in specific tissues due to the combined effects of maternal Gs-alpha mutation and tissue-specific imprinting leading to tissue-specific loss of Gs-alpha expression on the paternal allele. Gs-alpha imprinting in the placenta is a likely explanation for the perinatal subcutaneous edema, and preliminary data examining the placentas of E1m- and E1p- offspring are consistent with this possibility. The presence of severe obesity only in E1m- mice is similar to what is observed in humans with Albright hereditary osteodystrophy, in which obesity only develops when the disease is inherited from the mother. Studies in brain-specific knockout mice showthat imprinting in one or more regions of the central nervous system underlie the differences in the metabolic phenotypes observed in E1m- and E1p- mice (project ZO1-DK043315-01). We also generated mice with deletion of the exon 1A differentially methylated region which we had proposed as an imprint control region for Gs-alpha. These mice confirmed this hypothesis, as mice with deletion of the 1A region on the maternal allele (1Am-) had no changes in Gs-alpha expression while paternal 1A deletion (1Ap-) led to loss of Gs-alpha imprinting and Gs-alpha overexpression in tissues where Gs-alpha normally is imprinted, such as renal proximal tubules. This increased Gs-alpha expression in renal proximal tubules was associated with lower circulating parathyroid hormone levels due to increased sensitivity to this hormone. Moreover, cross-mating of E1- females to 1A- males showed that all of the major features resulting from loss of Gs-alpha expression from the maternal allele (edema, lethality, obesity, insulin resistance, lower sympathetic activity and metabolic rates) are reversed by the presence of the paternal 1A deletion and the loss of Gs-alpha imprinting. These results confirm that the features specifically present in E1m- mice are a direct consequence of the effects of Gs-alpha imprinting. Ongoing studies are examining the molecular mechanisms by which the 1A region is responsible for tissue-specific Gs-alpha imprinting, including DNAse hypersensitivity assays and DNA pull down assays looking for proteins associated with the 1A region. We have also using floxed Gs-alpha mice to generate a number of tissue-specific Gs-alpha knockout models in various tissues in our laboratory and through collaborations within and outside NIH. Some of the models studied in our lab are outlined in other projects. We have previously reported on mice with liver-specific Gs-alpha deficiency and have more recently shown that these mice eventually develop islet cell carcinomas and hepatocellular carcinoma. These may be direct effects on liver cells or secondary to the development of chronic steatosis and inflammation in the liver. We are also examining mechanisms by which pancreatic islet cells are stimulated in this mouse model. Our findings are consistent with these changes in islets not being mediated by neural circuits or by the incretin glucagon-like peptide 1, even though the level of this hormone is extremely high in these mice. Recently these mice were used to show that Gs-alpha pathways are critical for renal function, in particular the induction of renin production and urinary concentrating ability in the distal nephron, and involved in the ectopic bone formation which is observed in pseudohypoparathyroidism. Moreover, it has been shown that Gs-alpha is critical for homing of lymphocytes to the bone marrow niche. We are now in the process of generating Gs-alpha knockin mice (wild type and constitutively active forms) that will allow us to reintroduce Gs-alpha into specific tissues using various transgenic cre lines in order to see if we can rescue phenotypes in the germline mice. The constitutively active mice will also be very useful to model features in the disease McCune-Albright syndrome/fibrous dysplasia and to look at the effects of constitutively activated Gs-alpha in specific tissues on metabolic regulation.
Showing the most recent 10 out of 41 publications
