Academic Commons

Theses Doctoral

Genetic Analysis of the "Levin Rat" - a Rodent Model of Diet-Sensitive Obesity

Goffer, Yossef

Obesity, or the presence of an excessive amount of body fat is a major public health problem in the United States and, increasingly, the rest of the world. The apparent drivers of the increased prevalence of obesity over the past several decades are environmental changes, e.g., dietary and lifestyle changes that interact with the individual’s genetic susceptibility for weight gain. In humans, obesity appears to be driven primarily by increases of energy intake relative to expenditure; that is, to uncompensated hyperphagia. The heritability of adiposity, i.e., the extent to which differences in adiposity among individuals living in the same environment can be attributed to genetic differences is estimated by twin and other studies to be about 50%. Large scale population-based association studies (e.g., GWAS) have suggested that genetic variants (e.g., SNPs) associated with susceptibility or resistance to obesity affect primarily the development and regulation of the central nervous system (CNS). In particular, SNPs in genes that play a role in brain cellular structures and molecular pathways known to regulate energy homeostasis, most notably, the leptin-melanocortin signaling pathway, are among the most highly associated with human obesity. For example, SNPs around the melanocortin receptor, MC4R, are associated with increased adiposity and mutations in MC4R represent the most prevalent genetic variations associated with monogenic obesity. Ultimately, however, relatively little is understood about the biological mechanisms by which an individual’s genetic sequence confers susceptibility or resistance to weight gain in a specific environment. Such understanding could open new avenues for the prevention and treatment of obesity and would advance our understating of genetic predisposition to other complex diseases.

The goal of this research is to identify genomic regions contributing to susceptibility and resistance to hyperphagic obesity by analysis of whole genome sequence and hypothalamic gene expression data from two genetically related cohorts of Sprague-Dawley rats – the ‘Levin Rat’. Dr. Levin developed these animals by successive generations of selective breeding for differences in adiposity resulting from exposure to a calorically dense, highly palatable diet (described in detail in Chapter 2). These selectively bred diet-induced obese (DIO) and diet-resistant (DR) Levin rats have been the topic of a large body of physiological research (reviewed in Chapter 1) showing potentially important similarities to the physiology of human obesity. In particular, implication of diet-sensitive hyperphagia as the primary driver for the differential susceptibility of DIO (diet-induced obese) animals to gain weight in response to palatable diet; neuroanatomical and functional differences between DIO and DR in hypothalamic nuclei (e.g., ARH, PVH) and leptin signaling, prior to the development of obesity; and, neurophysiological differences between DIO and DR (diet-resistant) in ‘reward circuit’ nuclei (e.g., NAc) and their differential responses to pharmacological stimuli, e.g., cocaine, as well as palatable diet. These findings established the Levin rat as an interesting model for aspects of the biology of human obesity. Importantly, the genetic bases for these Levin rat phenotypes have remained unknown. Our efforts to elucidate the underlying genetics of this model system are, therefore, of potential relevance to human obesity.

We obtained phenotypic, whole genome sequence (WGS) and hypothalamic gene expression (RNA-Seq) data from selected Levin rats and analyzed these data to identify several loci that are highly associated with the body weight phenotype in the Levin cohorts, as well as in a confirmation cohort of genetically related progeny being studied for phenotypes related to addictive behaviors. In Chapter 2, I describe our methods and approaches to collecting the relevant phenotypic and genetic data, and to selecting primary and confirmation cohorts for the WGS and RNA-Seq studies. In Chapter 3, I describe our bioanalytical and statistical approaches to the WGS data designed to detect genomic loci likely inherited Identical by Descent (IBD) from common ancestors of the DIO, or DR animals; such loci constitute candidate obesity susceptibility loci for the Levin rat. In Chapter 4, I present the phenotypic differences between the DIO and DR animals used for the study, the results of IBD analysis of the WGS, and indicate genes we found to be differentially expressed or spliced in their respective hypothalami. I also show, using confirmation groups (CG) of animals from the primary cohorts of Levin rats (not used for IBD mapping), that the identified susceptibility loci are highly associated with DIO/DR lineage. In Chapter 5 I show, using representative SNPs from the loci with the highest association to the obesity susceptibility phenotype, significant association between the ‘DIO’ genotype and increased body weight in an independent cohort of Levin rat progeny [designated Obesity Prone (OP) and Obesity Resistant (OR)] maintained by Dr. Ferrario (U. Michigan). Using gene set enrichment analyses (GSEA), I show that the candidate susceptibility loci are enriched in CNS genes and genes previously associated with obesity traits in human GWAS. I also analyzed the genes implicated by differential expression and/or splicing (RNA-Seq) in relation to the genetic susceptibility map (IBD analyses) and identified several genes whose hypothalamic differential regulation (e.g., expression, splicing) may be linked to genetic sequence variations between the DIO and DR animals. Interestingly, essentially, all of the identified genes have been previously implicated in body weight regulation in other species, including humans.

Finally, I propose future studies to build upon this work in order to further refine the genetic susceptibility map, test the roles of putative candidate genes, and ultimately elucidate the genetic bases for the differences in body weight in this genetically complex mammalian model of diet-sensitive obesity: the Levin DIO/DR rats.

In summary, the conclusions from these studies are that:
- The 15 susceptibility loci we identified, span in total ~35 Mb, or 1.15% of the rat genome, likely contain a significant portion of the causal alleles underlying the Levin rat phenotype.
- The majority of genetic variants in these susceptibility loci are ‘non-coding,’ e.g., intronic, in untranslated regions, or intergenic – similar to common obesity SNPs.
- Our susceptibility map is enriched for genes governing CNS function and development, and its human syntenic genes/loci are enriched in obesity SNPs, identified by GWAS. Therefore, Levin rat obesity risk may have a genetic architecture similar to that in humans.
- We identified several genes that are differentially expressed or spliced in the hypothalamus of DIO/DR animals, and are also implicated by our genetic map: Zfr, Slc24a2, Fhit, Adarb1, Lrp2. Interestingly, human orthologs of all of these genes have been implicated in obesity by GWAS or familial linkage studies, enhancing them as putative candidates for a role in the Levin rat obesity phenotypes.
- Our studies further establish the Levin rat as a model system for polygenic human obesity and lay the foundation for further studies to elucidate the genetic basis of this interesting and important complex trait of hyperphagic diet-sensitive obesity.


  • thumnail for Goffer_columbia_0054D_15736.pdf Goffer_columbia_0054D_15736.pdf application/pdf 4.68 MB Download File

More About This Work

Academic Units
Neurobiology and Behavior
Thesis Advisors
Leibel, Rudolph L.
Ph.D., Columbia University
Published Here
May 15, 2020