| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Commentary |
From the Division of Clinical Research, Fred Hutchinson Cancer Research Center, Seattle, Washington
Transgenic technology, despite its power to analyze and create animal models for gene function, sometimes comes up short of answers. In this issue of The American Journal of Pathology, Cai and colleagues1
demonstrate that astute analysis of mutant strains can also deliver the goods. Tuberous sclerosis complex (TSC) is inherited as an autosomal dominant trait, with two genes, TSC1 (hamartin) and TSC2 (tuberin), each responsible for 
50% of familial cases.2
The disease presents similarly, albeit with a high index of variability, with both mutations, excepting the occurrence of renal cystic disease.3
Renal cysts occur almost exclusively in TSC2 patients, with a subset developing a severe form similar to polycystic kidney disease (PKD), with early progression to renal failure. Many of these patients have deletions involving both TSC2 and PKD1, which are contiguous genes located at chromosome 16p13.3.4
Despite this convincing explanation, some questions remained. Autosomal dominant PKD differs in both natural history and pathology from the TSC2/PKD1 contiguous gene deletion patients, and although PKD involves somatic loss of the normal PKD1 allele, this has not been demonstrated to occur in renal cysts from TSC2 patients.
A dominant pattern of inheritance for a single gene trait does not indicate whether the mutant allele has dominant or recessive effects in the cell. For recessive mutants, inheritance of a mutant allele from each parent is only one option. Studies of cancer genetics focused attention on a second mechanism, somatic mutation or deletion (loss of heterozygosity) of the remaining normal allele of a tumor suppressor gene.5 Generation of cells with both copies inactivated by this second-hit mechanism provides a selective advantage for clonal expansion and evolution. Reduction to homozygosity may also occur by mitotic recombination, or chromosome nondisjunction, yielding two copies of the maternal or paternal allele bearing the mutation (uniparental disomy). As proposed by Hall,6 if postzygotic reduction to homozygosity occurs sufficiently early in development, mosaicism may result in an exaggerated phenotype with segmental patterns.
In this issue, Cai and colleagues1 perform genetic analysis on three Eker rats with variant, severe PKD phenotypes and demonstrate mosaicism attributed to uniparental disomy of a mutant TSC2 allele. Reidar Eker7 first described autosomal inheritance of renal adenomas and carcinomas as a spontaneous mutant in Wistar rats in 1954. Renal neoplasms in the Eker rat arise from proximal and collecting duct epithelial cells on a background of cystic tubular dilation. Yeung and colleagues8 in 1993 reported the mapping of the Eker renal tumor susceptibility gene to a region on rat chromosome 10 syntenic with human chromosome 16p13. In the same year, a human gene responsible for tuberous sclerosis, TSC2, was identified in the 16p13.3 region.9 Cloning of the rat TSC2 followed, and sequencing of the Eker TSC2 locus identified insertion of a retrotransposon at one allele, disrupting the open reading frame.10 Final confirmation of the role of TSC2 was obtained by demonstrating that a wild-type TSC2 transgene completely suppressed the Eker phenotype.11 The cellular recessive nature of the TSC2 phenotype was demonstrated by showing loss of heterozygosity or point mutations in the remaining normal TSC2 allele within Eker renal tumors, but not uninvolved tissues.12
The identification of germline mutations in TSC2 as responsible for the Eker phenotype was surprising because of the notable differences between human and rat disease phenotypes. Renal tumors are the only lesion in the Eker rat model observed with 100% penetrance. Uterine leiomyomas, splenic hemangiomas, and pituitary adenomas occur in decreasing frequency in the Eker rat.13 Although renal disease (most commonly renal angiomyolipomas) occurs in the majority of TSC patients, renal cell carcinoma is diagnosed in <1% of patients.14 Skin lesions (hypomelanotic macules and angiofibromas), cardiac rhabdomyomas, and pulmonary hamartomas occur commonly in TSC, but are not reported in Eker rats. Central nervous system tumors are the leading cause of morbidity and mortality in TSC. Cortical tubers (hence the name tuberous sclerosis) are distinctly uncommon in Eker rats. One possible explanation for the species-specific phenotypes may be that the loss of the wild-type copy through somatic deletion occurs at an earlier stage of development and/or in different cell lineages in the human disease. Attempts to analyze Tsc2 null animals have failed because of embryonic lethality in both gene-targeted mice and rats with two germline copies of the Eker TSC2 allele.15,16
Cai and colleagues1 collected three Eker rats with severe, early bilateral kidney disease from two different animal colonies. Pathological examination found diffuse PKD. These lesions resembled the typical Eker kidney lesions only in the histological appearance of papillary adenomas and carcinomas forming within a portion of the cysts. A single nucleotide polymorphism between Eker TSC2 and the allele from the congenic Long-Evans strain background was discriminated using polymerase chain reaction-denaturing high performance liquid chromatography (dHPLC) and confirmed by restriction fragment length polymorphism and allele-specific polymerase chain reaction. Using an epithelial cell line derived from one of the polycystic kidneys and DNA from normal and affected organs, the group demonstrated loss of the wild-type TSC2 allele and the adjacent PKD1 allele and duplication of the Eker allele. These results, combined with karyotypes showing two normal copies of chromosome 10 in affected cells, confirm mosaicism for uniparental isodisomy, probably because of postzygotic chromosomal nondisjunction. Somatic recombination is also a possible mechanism for the mosaicism described, although the reciprocal mosaic genotype, including two wild-type TSC2 alleles, was not identified.
The pattern of mosaicism in the Eker rats with PKD indicates that uniparental disomy occurred near the time of condensation of intermediate and lateral plate mesoderm in early organogenesis. Intriguingly, the loss of TSC2 function earlier in development did not recapitulate the TSC phenotype in humans. In particular, hamartomatous lesions such as angiomyolipomas and oncocytomas involving multiple organs are not observed. However, there is accumulating evidence that the cellular consequences of the loss of tuberin are conserved in rodents and humans. Multiple angiogenic growth factors are activated in angiofibromas from tuberous sclerosis patients and TSC2(-/-) rat embryonic fibroblasts support endothelial cell proliferation.17 Eker rat uterine leiomyomas and TSC pulmonary hamartomas and lymphangioleiomyomatosis express high levels of the chromosomal protein HMG2, normally restricted to embryonic tissues.18
A number of hypotheses for tuberin function have been put forward. Based on sequence similarity with a GTPase-activating protein for Rap1 (GAP3), two groups reported that tuberin has GAP activity for both Rap1a and Rab5.19,20 Identification of the Drosophila melanogaster locus, gigas, as a TSC2 homolog has led to important insights into intracellular signaling pathways involving tuberin. Mutant alleles for gigas cause enlarged eyes and wings because of cellular hypertrophy and hyperplasia.21 Crossing gigas flies with other mutant backgrounds established an epistasis relationship between dTsc2 and insulin-receptor signaling components, with cell growth phenotypes in insulin receptor, dPTEN and dPKB mutants dependent on dTsc2 function. Conversely, a downstream target of insulin receptor signaling, ribosomal protein S6 kinase, acted epistatically dominant to dTsc2. Studies in mammalian cells have determined that TSC2 is phosphorylated by PKB and that TSC2 function is required for regulation of S6 kinase by nutrients.22,23 Under conditions of amino acid deprivation, S6 kinase is inactivated, leading to suppression of ribosomal protein synthesis. In TSC2 mutant backgrounds, ribosomal protein synthesis is insensitive to amino acid supply. Understanding tuberin as an inhibitor of S6K, the loss of regulation of ribosomal protein synthesis may explain the cellular phenotypes of increased proliferation and growth in tuberin-deficient cells.
The clear implication of the phenotype of the mosaic Eker rats is that the consequences of tuberin deficiency are limited to renal tubular epithelium in the rat kidney, despite the opportunity for earlier developmental pathology affecting multiple cell lineages. The link between tuberous sclerosis and renal tubular cysts was first observed in patients with the contiguous gene deletion syndrome involving TSC2 and PKD1, presenting as early onset, severe renal cystic disease. A recent study by Kleymenova and colleagues24 uncovered another relationship between tuberin and polycystin-1. They demonstrated that functional localization of polycystin-1 at the basolateral membrane was defective in TSC2 mutant cells, thus presenting the possibility that tuberin deficiency can mimic a PKD1 null genotype, manifesting as PKD. The mosaic Eker rats described by Cai and colleagues1 with UPD for the TSC2 mutant and two normal PKD1 alleles are entirely consistent with this hypothesis.
Although mosaicism is rarely reported in inherited cancer predisposition syndromes, it seems to be a common mechanism in other types of autosomal dominant, cellular recessive diseases.25 Inherited skin diseases such as neurofibromatosis 1, cutaneous leiomyomatosis, and disseminated superficial actinic porokeratosis frequently present with a pattern of pronounced segmental lesions superimposed on the usual nonsegmental phenotype. This phenotype, termed type 2 segmental involvement by Happle,26 has been proposed to result from genetic mosaicism with reduction to homozygosity or hemizygosity of an inherited mutation. As shown in Cai and colleagues,1 recognition of additional examples of mosaicism in human disease and corresponding animal models may yield surprising and novel insights into gene function.
Footnotes
Address reprint requests to David M. Hockenberry, Division of Clinical Research, Fred Hutchinson Cancer Research Center, 1100 Fairview Avenue North, Seattle, Washington 98109-1024. E-mail: dhocken{at}fhcrc.org
Accepted for publication November 20, 2002.
References
Related articles in Am J Pathol:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
