Research interests

Genetic Interactions in Mouse Models of Human Diseases

In humans, mice, and other model organisms, the segregation of mutations within single genes has been associated with both Mendelian as well as complex diseases. The identification of these disease genes and the causative mutations within them has greatly enhanced our understanding of normal function in the affected organs. At the same time it has become clear that these single gene mutations must reside in a permissive genetic background in order to manifest a disease phenotype. The background genes that interact with disease-causing mutations and that are responsible for specific phenotypic alterations are commonly called genetic modifier loci. The identification of these modifier genes is a powerful tool for defining biological pathways that lead from the primary genetic defect to the disease phenotype. Although phenotypic variations (e.g., age of onset, rate of disease progression, severity) in heritable forms of many diseases have been reported both within families and in individuals carrying the same genetic mutations, few modifier loci have been identified in human families to date. Presumably, this is due to the fact that most families are far too small to achieve sufficient power to chromosomally localize modifiers.

In mice, mapping of single gene mutations from which candidate genes can be identified and quantitative trait locus (QTL) analysis are powerful tools for identifying loci associated with specific disease phenotypes and for dissecting genetic components of complex traits. In addition, inclusion of a mutation that interacts epistatically with QTLs in genetic crosses is a useful method for revealing the function of novel genes and pathways. While identification of primary mutations is essential for understanding the role of molecules in maintaining normal cellular function, identification of modifiers, particularly those that suppress phenotypes, may lead to a more generalized blueprint upon which therapies can be designed.

Central Regulation of Energy Homeostasis

Obesity, type 2 diabetes (non-insulin-dependent diabetes mellitus) and heart disease are highly prevalent metabolic diseases that afflict a large proportion of the aging population in the United States. Nearly 55% of adults are overweight or obese, about 10% of individuals over 65 have type 2 diabetes, and 50% of deaths are caused by coronary artery disease. These diseases should be viewed as aspects of a metabolic syndrome that is produced by the interaction of many genes, rather than as separate entities. For example, only ~7% of obese patients develop type 2 diabetes, whereas ~90% of type 2 diabetes patients are obese. This suggests that in type 2 diabetes patients, obesity genes interact with diabetes susceptibility genes to produce the obese/diabetic phenotype. Obesity genes alone would only lead to obesity, and diabetes susceptibility genes alone may not cause an overt phenotype. To further complicate the issue, one can estimate from the number of transgenic and targeted mouse models that develop an obese phenotype that there are on the order of 500 to 1,000 genes that will lead to obesity when mutated. Therefore, we believe that a broader knowledge of obesity pathways is necessary for a rational design of interventions with minimal side effects. To obtain additional entry points into obesity pathways, our program focuses on identification of new obesity and type 2 diabetes mutations and their genetic modifiers, and on how these mutations influence energy homeostasis to lead to the obese-diabetic phenotype.

TALLYHO, a new model for type 2 diabetes

Because type 2 diabetes occurs in the context of obesity, and insulin resistance genes have to interact with pancreas insufficiency genes to create the hyperglycemic phenotype, few mouse models of type 2 diabetes exist. We have established a new inbred strain of diabetic mice, TALLYHO (TH), through a backcross/intercross strategy with selection for male hyperglycemia. This new model of type 2 diabetes is characterized by moderate obesity, hyperinsulinemia, glucose intolerance and enlargement of the pancreatic islets of Langerhans. Male TH mice become overtly diabetic by 8 weeks of age. Breeding experiments suggested that the hyperglycemic trait was caused by a newly acquired autosomal recessive, single gene mutation that occurred on a permissive genetic background. In mapping crosses, we localized Tanidd1 (TallyHo associated non-insulin-dependent diabetes mellitus 1), the major diabetes locus in the cross, to Chromosome 19 and constructed congenic lines to aid in fine mapping and cloning of the gene. Additional diabetes and obesity susceptibility genes segregate in a (TH x B6)F1 x TH backcross. Furthermore, the Tanidd1 mutation epistatically interacts with loci on Chromosomes 6, 8, and 18. Physiological experiments indicate that insulin secretion is normal in male TH mice, but that TH mice show decreased glucose uptake in adipose tissue and muscle. The decreased glucose uptake is at least in part due to an abnormal cellular distribution of the facilitative glucose transporter SLC2A4 in adipocytes, and a failure of insulin to recruit SLC2A4 to the cell surface. The abnormal intracellular localization of SLC2A4 is also observed histochemically and by subcellular fractionation in the TH strain as well as in a Chromosome 19 congenic line B6.TH-Tanidd1^TALLYHO, but not in the reverse congenic TH.CAST-Tanidd1^CAST carrying the castaneus derived wild-type allele of Tanidd1. This suggests that Tanidd1 is involved in insulin-stimulated recruitment of SLC2A4.

Obesity-sensory loss syndromes

Several syndromes exist in the human population that are characterized by obesity, diabetes, and loss of vision and hearing. Alstr?m syndrome is a rare, recessive human disease of childhood obesity, retinal and cochlear degeneration, and type 2 diabetes and heart disease. Alstr?m syndrome and the phenotypically similar but genetically distinct Bardet-Biedl syndrome show a remarkable phenotypic similarity with the tubby mouse. Because of this similarity and co-localization of the gene products in specific tissues (e.g., TUB and ALMS1 co-localize in the pancreatic alpha cells), as well as evidence that the underlying genes may function in intracellular transport (abnormal accumulation of vesicles is found in the inner segments of photoreceptor cells in Tub, Tulp1 and Alms1 mutant mice), we hypothesize that ALMS1, TUB, and the BBS proteins act in the same biochemical pathway.

Functional genomics of the tubby gene family

Tubby is an autosomal recessive mutation leading to a tripartite phenotype of maturity onset obesity, blindness, and deafness in B6.Cg-Tub^tub homozygotes. We have shown that the obesity in tubby mice is not associated with hyperphagia or hypercorticism but is associated with progressive insulin resistance. The progressive retinal degeneration in tubby mice is characterized by abnormal electroretinograms detected as early as 3 weeks of age and is caused by apoptotic loss of photoreceptor cells. Hearing loss is also apparent by 3 weeks of age and is characterized histologically by accelerated loss of outer hair cells and by progressive loss of inner hair cells. The obesity coupled with the retinal degeneration and hearing loss makes tubby mice a good model for rare human monogenic disorders described above.

The Tub gene encodes a novel protein present in the cytoplasm and nucleoli of neuronal cell nuclei in the retina and brain. We have identified the genes for three tubby-like proteins (TULPs) from mouse and human, respectively. Tulp1, when mutated in humans, causes retinitis pigmentosa 14 and leads to retinal degeneration in mice. Mutations in Tulp3 in the mouse lead to embryonic lethality and neural tube defects. An underlying abnormality in the knockout mice appears to be the apoptotic loss of ?III-tubulin positive cells in the ventral neuroepithelium of the hindbrain.

To search for biochemical pathways in which TUB plays a role, we carried out genetic modifier screens. We identified moth1, the modifier of tubby hearing 1, as the microtubule associated protein 1A (Mtap1a). Mutations in the C57BL6/J (B6) allele of this gene lead to the hearing loss observed in B6.Cg-Tub^tub homozygotes. We also showed that these mutations reduce the binding of MTAP1A to members of the post-synaptic density (PSD) family of proteins, specifically DLGH4 (formerly PSD95). PSDs are synaptic scaffolding proteins that link signaling components and the neuronal transport machinery. Our finding establishes a role for TUB in synapse function and suggests an interaction with the intracellular transport machinery.

Previously, little was known about the physiology of tubby mice. During testing of B6.Cg-Tub^tub mice using the Comprehensive Lab Animal Monitoring System (CLAMS), we found tubby mice to have a lower respiratory quotient (RQ) compared to B6 controls before the onset of obesity in both the light and the dark period (p<0.001). This indicates that tubby mice fail to activate carbohydrate metabolism and instead rely on fat metabolism for energy needs. In concordance with this data, tubby mice show a higher excretion of ketone bodies (7 weeks age, p<0.04, and 3 months age, p<0.0005) and accumulation of glycogen in the liver in the non-fed state. Quantification of mRNA levels using Real Time PCR showed that tubby mice fail to induce H6pd during the transition from the light to the dark period (p<0.0001), leading to reduced H6PD protein and enzymatic activity in the liver. In addition, several citric acid cycle genes are misregulated in tubby mice. The lipolytic enzymes acetyl-coA synthetase and carnitine palmitoyl transferase are increased during the dark cycle (p<0.005), and decreased during the light period (p<0.005). This abnormality in liver carbohydrate metabolism must be a secondary effect of the tub mutation, since tub is not expressed in liver. Examination of hypothalamic gene expression showed high levels of prepro-orexin mRNA leading to accumulation of orexin peptide in the lateral hypothalamus. Based on our study and published reports on orexin action, we hypothesize that abnormal hypothalamic orexin expression leads to changes in liver carbohydrate metabolism and may contribute to the obesity observed in tubby mice.

Alstr?m syndrome

Through ascertainment and recombinational analysis of families, we mapped the Alstr?m syndrome gene (ALMS1) to human chromosome 2p12, narrowed the minimal interval for ALMS1 to ~ 1.2 megabases, and through candidate gene testing identified a novel gene in which multiple mutations occurred in affected patients but not in unaffected controls. Mutation screening to date has identified major mutations such as nonsense mutations and insertions and deletions causing a frameshift and premature translation termination, suggesting that missense mutations do not cause disease or lead to a milder phenotype that is not recognized as Alstr?m syndrome. Although Alstr?m syndrome is a rare disease and mutations in the same gene may not play a role in more common forms of obesity, the gene still must be part of a pathway leading to obesity since an obese phenotype is observed. Identification of its functional role will contribute to our overall goal of describing the network of obesity pathways.

Analysis of Alms1 gene expression and protein abundance showed that ALMS1 is present in all tissues affected in the disease. However, there is specific expression in the hypothalamus–for example, in the parvicellular neurons of the paraventricular nucleus–and in the pancreatic glucagon-producing alpha cells but not the insulin-producing beta cells. This same pancreatic cell expression pattern is also found for the TUB and BBS gene products. This may indicate that the type 2 diabetes in Alstr?m is driven by abnormalities in glucagon secretion. We have also found that human Alstr?m’s patients, whether diabetic or not, have higher serum glucagon levels than either normal or type 2 diabetic controls.

We have recently obtained homozygous offspring derived from gene-trapped mice that recapitulate common ALMS1 mutations in humans. Gratifyingly, the mutant mice show all the phenotypes observed in human patients and will be an invaluable model system for elucidating the biochemical function of the ALMS protein and studying the pathological consequences of its loss of function. To address the question as to whether the type 2 diabetes in Alstr?m could be driven by glucagons (as it is in conditions of pseudo-glucagonoma), we are currently measuring the glucagon and insulin response to glucose in isolated islets from wild-type and mutant mice.

Novel single gene mouse obesity mutations

For the past several years, we have collected phenodeviants with increased body weight from the production and research animal colonies at The Jackson Laboratory. We are also mapping and cloning selected mutants from the Neuroscience Mutagenesis Facility and the Mouse Heart, Lung, Blood and Sleep Disorders Center, and are currently actively pursuing five mutants.

Molecular Genetics of Retinal Development and Maintenance

An estimated 50 million people worldwide are blind and approximately 150 million are significantly vision-impaired. Except for trauma and infections, the majority of human eye diseases are genetic in nature. A query of the Online Mendelian Inheritance in Man database using “retinal disease” as an identifier yields 353 entries. It is further estimated that the total number of retinal disease genes is 2-3 times the number of presently known or mapped genes (Wright & van Heyningen, 2001). Using the rich and unique genetic resources available for the mouse, our program has focused on identifying new entry points into pathways important in retinal function and finding common pathways involved in photoreceptor degeneration.

Elucidation of the function of molecules and pathways important in retinal development and maintenance

During the past five years, we discovered that the mutations retinal degeneration (rd) 6, rd7, rd8, and cornea1 (corn1), are alterations in membrane-type frizzled-related protein (Mfrp), photoreceptor nuclear receptor (Nr2e3), crumbs homolog 1 (Crb1), and destrin (Dstn), respectively. We also identified a novel gene responsible for Alstr?m syndrome in humans (ALMS1), which when mutated leads to cone dystrophy and early vision loss, hearing impairment, obesity and cardiomyopathy. In unpublished studies we have identified the mutations underlying the phenotypes in three more mutant strains: veils, rhigoletto, and smallies. Finally, a modifier of the hearing phenotype in tubby mice was identified as mutations within Map1a. As a result of these and previous efforts, and the use of database mining and transgenic approaches, we continue to work toward understanding the function of the tubby and crumbs gene families, and genes important in glycosylation in the retina.

 

The role of the tubby gene family and Alms1 in cone development and vesicle transport

After identifying Tub as a member of a small gene family, we found that Tub and Tulps are all expressed in the retina. Initially, to determine the normal function for Tulp1 and Tulp3, we made mouse models by homologous recombination. We determined that Tulp1, like Tub, is important in maintaining photoreceptors, with loss of functional TULP1 in mice and humans leading to rapid photoreceptor degeneration. Loss of functional TULP3 led to embryonic lethality, and a conditional allele of Tulp3 is being prepared to examine its role in the adult retina. We are particularly interested in its function because, unlike TUB and TULP1, it is solely expressed in the inner nuclear layer.

We hypothesize that tubby and Alstr?m syndrome (and Bardet-Biedl syndrome), which share many phenotypic features, function in the same pathway and have both early developmental and adult functions. Both TUB and TULP1 are first expressed in the nucleus, then switch to a cytosolic localization between P6 and P10, and P2 and P6, respectively. This switching of cellular localization has been shown to be a feature of gene products important in cell differentiation. Tubby is also found in the actively proliferating zone of the peripheral retina at P6. TULP1, at least, appears important in cone development because it is one of the earliest markers for differentiated cones in humans (Milam et al., 2000) and is severely downregulated in mice homozygous for a targeted mutation of Thrb (T. Reh, personal communication). These Thrb^tm2Dr homozygotes do not express m-cone opsin. Alms1 may also have a similar expression pattern because one of the first clinical functional deficits observed in patients with Alstr?m syndrome is cone dystrophy. Preliminary data suggest the near absence of m-cone opsin in the retina of adult Alms1-/- mice. We are currently breeding compound heterozygyotes and doubly and triply null mutant mice to examine the role of the three genes in retinal development. Intriguingly, photomicrographs of a P14 retina from a doubly homozygous Tub^tub/tub Tulp1^{tm1Pjn/tm1Pjn} mouse in a paper by Hagstrom et al. (2001), appear to show delay in retinal development by 2 to 3 days; at P17 photoreceptors were absent.

Determining the functional components of the external limiting membrane (ELM) and their role in maintaining retinal integrity

Prior to discovering the molecular basis for rd6, we knew of the existence of a strong genetic modifier that caused a more severe, earlier onset phenotype in rd6 mice. After identifying the basis for rd6 as a mutation within membrane-type frizzled-related protein, and subsequently, its modifier rd8 as a mutation in Crb1, we are still left with the question of how these two factors interact. Based on our hypothesis and work by A. Wright’s group, we know that MFRP directly interacts with C1QTNF5, a second gene product on the same transcript as MFRP. We have observed that C1QTNF5 appears to migrate from the retinal pigmented epithelium where it is expressed to the ELM during retinal development. We hypothesize that C1QTNF5 may be a direct or indirect bridging molecule to CRB1 that is expressed in photoreceptors and M?ller glial cells and localizes to the ELM. Using functional genomic approaches such as yeast-2-hybrid analysis and co-immunoprecipitation, we will determine whether C1QTNF5 and CRB1, or some intermediary factor, is involved in the genetic interaction observed. We will also intercross a B6.C3H-Crb1^C3H congenic line (N10) that does not show a dysplastic phenotype with other strains. We will score F2 progeny for retinal dysplasia and carry out genome-wide scans to identify factors that genetically interact with CRB1 to cause a more severe retinal phenotype. We know that crossing mice homozygous for Crb1^rd8 with DBA/2J or 129/SvJ strains produces different dysplastic phenotypes. By identifying factors that interact with CRB1, we may begin to unravel the complex and highly variable phenotypic presentation of the retinal disease observed among human patients with CRB1 mutations.

Identifying glycoproteins important in the maintenance of the dystrophin glycoprotein complex (DGC)

Glycosylation of alpha-dystroglycan is important in the muscle pathology observed in a number of mutations causing neurosensory, brain, and muscle pathology, including muscle-eye-brain disease and Fukuyama’s muscular dystrophy. In the muscle of mice with the myodystrophy mutation (myd) within the Large glycosyltransferase gene, alpha-dystroglycan is not glycosylated but the DGC is intact. While glycosylation of alpha-dystroglycan is important in the retina, it is likely that other glycoproteins play a role. This is true because in the retina, in contrast to the muscle, the DGC in the inner limiting membrane does not appear to assemble, while in the outer plexiform layer, the beta dystroglycan moiety is dramatically diminished. Using both single- and two-dimensional gel analysis, we have identified bands or spots that are evident in the wild-type that are not present in Large^myd homozygotes. Using mass spectrometry, we will identify these proteins as potential substrates of the glycosyltransferase in the retina. Furthermore, genes expressed in the retina that possess glycosyltransferase motifs and are important in glycosylation in the neuroretina are being mutagenized by various targeting approaches. We will determine if other retinally expressed glycosyltransferases cause similar retinal phenotypes and if they have similar substrate specificities.

Novel mutations and model development

We will continue our quest to identify molecules important in retinal development and function by characterizing and identifying the underlying molecular bases for new spontaneous and chemically induced mutations. It should be noted that despite the availability of a fair number of retinal mutants, they represent a small fraction of the total retinal disease genes identified in humans. Also, in our mutagenesis facility, more than half of the eye mutants evaluated represent novel loci. These models may allow us to flesh out the different pathways we described above or provide additional entry points into pathways important in the maintenance of retinal function and integrity. We are currently using in vitro fertilization approaches to produce large numbers of backcross or intercross mice in a short time frame to generate fine structure maps and to hasten the positional cloning process. Using this approach, we are currently cloning four novel mutants that lead to no b-wave (EMR-ERG), photoreceptor degeneration (nmf5 and nmf12) or retinal lamination (nmf67) defects. We have also recently mapped four new mutations, two of which we are pursuing. One, JR3288, leads to cone ERG dysfunction and the other, nmf192, may be a candidate for two previously mapped, but not cloned, loci involving light-induced retinal degeneration in mouse and LCA in humans. Finally, we are screening ENU mutagenized mice at generation G3 for late onset retinal diseases. We have identified 21 mutants and are in the process of testing for heritability.

Common genetic suppressors of photoreceptor degeneration

Considering the large number of mutant genes that lead to retinal degeneration, therapeutics that affect pathways or mechanisms that are shared or “common” to all types of primary mutations may be the best targets. One pathway that appears to be associated with retinal and potentially macular degeneration is programmed cell death, or apoptosis. Previous studies have shown that mutations in genes such as Pde6b, Rds, and Rho lead to apoptosis of photoreceptor cells. We have also shown that the degeneration in Tub^tub and Tulp1^tm1Pjn homozygotes is due to apoptosis. Little is known about what triggers PCD in retinal degeneration, what the actual pathway is in the eye or even if all mutations identified utilize the same apoptotic pathway.

A current study under way in our lab is designed to address these issues. We have chosen three genetically determined mutants that exhibit different clinical phenotypes and/or rates of retinal degeneration. rd3 homozygotes develop pan-retinal degeneration by 8 weeks, Tub^tu homozygotes by 12 weeks and Nr2e3^rd7 homozygotes by 16 weeks of age. rd3 and Tub,sup>tub homozygotes each exhibit similar loss of pigment epithelium and patches of pigment deposits, while Nr2e3^rd7 homozygotes present with small dots across the retina. Each of these mutations are being intercrossed with three genetically diverse strains, CAST/EiJ (a wild-derived strain), and two standard inbred strains, AKR/J (a stock originating from F?rth) and NOD.NON-H2^nb1/LtJ (a Swiss-derived stock originating in Japan). Photoreceptor function in F2 progeny from each cross is being assessed by histological examination. Genome-wide scans with simple sequence length polymorphic markers are being carried out and quantitative trait loci identified that delay or suppress retinal degeneration. Our results to date show that each strain is able to delay or suppress retinal degeneration of one or more of the mutants. We are identifying QTL traits that are overlapping between mutations within the same (e.g., a modifier for both Tub^tub and Nr2e3^rd7 exists on Chromosome 11 in the B6 x AKR intercross) or different crosses (a locus on Chromosome 9 prevents Tub^tub induced photoreceptor loss in a B6 x CAST intercross, and Nr2e3^rd7 induced photoreceptor loss in a B6 x NOD.NON-H2^nb1 intercross). Multiple loci have been identified within each intercross, indicating the complex modification of the retinal degenerative phenotype. Results from the crosses described will allow us to dissect and separate the genetic factors that lead to phenotypic variability and may provide access into the pathways through which photoreceptor degeneration proceeds.

In the near future, we will complete the crosses and genome scans to identify the photoreceptor degeneration suppressors. Common loci will be isolated in the form of congenics and these will be used as reagents to identify the genes underlying the suppression. In the long term, coupling QTL analysis of these crosses with other molecular techniques–such as gene expression profiling to identify candidate genes–will be powerful complementary approaches. The QTL analysis will give us loci that are important in delaying or suppressing retinal degeneration, and the gene expression profiling will provide potential candidate genes.

 

Publications

• Ikeda S, Cunningham LA, Boggess D, Hobson CD, Sundberg JP, Naggert JK, Smith RS, Nishina PM. 2003. Aberrant actin cytoskeleton leads to accelerated proliferation of corneal epithelial cells in mice deficient for destrin (actin depolymerizing factor). Hum Mol Genet 12:1029-1037.
• Kim J, Taylor PN, Young D, Karst S, Nishina PM, Naggert JK. 2003. New leptin receptor mutations in mice: Leprdb-rtnd, Leprdb-dmpg, and Leprdb-rlpy. J Nutr 133:1265-1271.
• Mehalow AK, Kameya S, Smith RS, Hawes NL, Denegre JM, Young JA, Bechtold L, Haider NB, Tepass U, Heckenlively JR, Chang B, Naggert JK, Nishina PM. 2003. CRB1 is essential for external limiting membrane integrity and photoreceptor morphogenesis in the mammalian retina. Hum Mol Genet 12:2179-2189.
• Coe NR, Naggert JK. 2003. Genetics of Obesity. In: Encyclopedia of Life Sciences, Atkins D, [ed], Nature Publishing Group.
• Ikeda S, Ikeda A, Nishina PM, Naggert JK. 2003. Towards Understanding the Function of the Tubby Gene Family in the Retina. In: Retinal Degenerations: Mechanisms and Experimental Therapy. Proceedings of the Xth International Symposium on Retinal Degenerations, LaVail MM, Hollyfield JG, Anderson RE, [eds], Kluwer Academic/Plenum Publishers, N.Y., pp. 309-314.
• Nishina PM, Naggert JK. 2003. Mouse Genetic Approaches to Access Pathways Important in Retinal Function. In: Retinal Degenerations: Mechanisms and Experimental Therapy. Proceedings of the Xth International Symposium on Retinal Degenerations, LaVail MM, Hollyfield JG, Anderson RE, [eds], Kluwer Academic/Plenum Publishers, N.Y., pp. 29-34.

[Edit my profile]