Genetic basis underlying connection between hyperglycemia and dyslipidemia in Apoe-deficient mice

Individuals with dyslipidemia often develop type 2 diabetes, and diabetic patients often have dyslipidemia. It remains to be determined whether there are genetic connections between the 2 disorders. A female F2 cohort, generated from BALB/cJ (BALB) and SM/J (SM) Apoe-deficient (Apoe−/−) strains, was fed a Western diet for 12 weeks. Fasting plasma glucose and lipid levels were measured before and after Western diet feeding. 144 genetic markers across the entire genome were used for analysis. One significant QTL on chromosome 9, named Bglu17 [26.4 cM, logarithm of odds ratio (LOD): 5.4], and 3 suggestive QTLs were identified for fasting glucose levels. The suggestive QTL near the proximal end of chromosome 9 (2.4 cM, LOD: 3.12) was detected when mice were fed chow or Western diet and named Bglu16. Bglu17 coincided with a significant QTL for HDL and a suggestive QTL for non-HDL cholesterol levels. Plasma glucose levels were inversely correlated with HDL but positively correlated with non-HDL cholesterol levels in F2 mice fed either diet. A significant correlation between fasting glucose and triglyceride levels was observed on the Western but not chow diet. Haplotype analysis revealed that “lipid genes” Sik3 and Apoc3 were probable candidates for Bglu17. We have identified multiple QTLs for fasting glucose and lipid levels. The colocalization of QTLs for both phenotypes and the sharing of potential causal genes suggest that dyslipidemia and type 2 diabetes are genetically connected. Article Summary Patients with dyslipidemia often develop type 2 diabetes, and diabetic patients often have dyslipidemia. It remains unknown whether there are genetic connections between the 2 disorders. Using a female F2 cohort derived from BALB/cJ and SM/J Apoe-deficient mice, we identified one significant QTL on chromosome 9, named Bglu17, and 3 suggestive QTLs were identified for fasting glucose levels. Bglu17 coincided with a significant QTL for HDL and a suggestive QTL for non-HDL levels. Plasma glucose levels were significantly correlated with HDL and non-HDL levels in F2 mice. Haplotype analysis revealed Sik3 and Apoc3 were probable candidates for both QTLs.


Introduction
Individuals with dyslipidemia have an increased risk of developing type 2 diabetes (T2D), and diabetic patients often have dyslipidemia, which includes elevations in plasma triglyceride and LDL cholesterol levels and reductions in HDL cholesterol levels , which is in contrary to the positive correlations observed at the clinical level. Furthermore, it is challenging to establish causality between genetic variants and complex traits in humans due to small gene effects, complex genetic structure, and environmental influences. . In the present study, we performed quantitative trait locus (QTL) analysis using a female cohort derived from an intercross between BALB-Apoe−/− and SM-Apoe−/− mice to explore potential genetic connections between dyslipidemia and T2D.

Methods
Mice: BALB and SM Apoe-/-mice were created using the classic congenic breeding strategy, as previously described . BALB-Apoe-/-mice were crossed with SM-Apoe-/-mice to generate F1s, which were intercrossed by brother-sister mating to generate a female F2 cohort.
Mice were weaned at 3 weeks of age onto a rodent chow diet.  . The expectation maximization (EM) algorithm was used to detect main effect loci for each trait. One thousand permutations of trait values were run to define the genome-wide LOD (logarithm of odds) score threshold needed for significant or suggestive linkage of each trait. Loci that exceeded the 95th percentile of the permutation distribution were defined as significant (P<0.05) and those exceeding the 37th percentile were suggestive (P<0.63).
Prioritization of positional candidate genes. The Sanger SNP database (http://www.sanger.ac. uk/sanger/Mouse_SnpViewer/rel-1410) was used to prioritize candidate genes for overlapping QTLs affecting plasma glucose and HDL cholesterol levels on chromosome (Chr) 9, which were mapped in two or more crosses derived from different parental strains for either phenotype. We converted the original mapping positions in cM for the confidence interval to physical positions in Mb and then examined SNPs within the confidence interval. Probable candidate genes were defined as those with one or more SNPs in coding or upstream promoter regions that were shared by the parental strains carrying the "high" allele but were different from the parental strains carrying the "low" allele at a QTL, as previously reported . . The suggestive QTL on distal Chr5 (101.24 cM, LOD 3.198) was novel. The BALB allele conferred an increased glucose level for both of the Chr9 QTLs while the SM allele conferred increased glucose levels for the 2 Chr5 QTLs (Table   2).
Coincident QTLs for fasting glucose and lipids. LOD score plots for Chr9 showed that the QTL for fasting glucose (Bglu17) coincided precisely with the QTLs for HDL (Hdlq17) and non-HDL (Nhdlq11) in the confidence interval (Fig. 6). F2 mice homozygous for the BB allele exhibited elevated levels of fasting glucose and non-HDL but decreased levels of HDL, compared to those homozygous for the SS allele (Table 2). These QTLs affected their respective trait values in an additive manner.
Correlations between plasma glucose and lipid levels. The correlations of fasting glucose levels with plasma levels of HDL, non-HDL cholesterol, or triglyceride were analyzed with the F2 population (Fig. 7). A significant inverse correlation between fasting glucose and HDL cholesterol levels was observed when the mice were fed a chow (R=-0.220; P=8. 1E-4)  .

Discussion
BALB and SM are two mouse strains that exhibit distinct differences in HDL, non-HDL cholesterol, and type 2 diabetes-related traits when deficient in Apoe . BALB-Apoe -/mice have higher HDL, lower non-HDL cholesterol, and lower glucose levels than SM-Apoe -/mice when fed a Western diet. To identify the genetic factors underlying these differences, we performed QTL analysis on a female cohort derived from an intercross between the two Apoe -/strains. We have identified four loci for fasting glucose levels, four loci for HDL cholesterol levels, nine loci for non-HDL cholesterol levels, and three loci for triglyceride levels.
Moreover, we have observed the colocalization of QTL for fasting glucose with the QTLs for HDL and non-HDL cholesterol on chromosome 9 and the sharing of potential candidate genes.
We identified a significant QTL near 26 cM on chromosome 9, which affected fasting plasma glucose levels of mice on either chow or Western diet. We named it Bglu17 to represent a novel locus regulating fasting glucose levels in the mouse. This locus is overlapping with a significant QTL (not named) for blood glucose levels on the intraperitoneal glucose tolerance test identified in a BKS.Cg-Leprdb+/+m x DBA/2 intercross and a suggestive QTL identified in a B6-Apoe -/x BALB-Apoe -/intercross   . These "lipid genes" might also be the causal gene ( .
Plasma triglyceride levels were strongly correlated with fasting glucose levels in this cross on the Western diet, although no significant correlation was found when mice were fed the chow diet. Despite the strong correlation, no overlapping QTLs were found for fasting glucose and triglyceride. The reason for the discrepancy between non-HDL cholesterol and triglyceride in terms of the presence or absence of colocalized QTLs is unclear.
A suggestive QTL for fasting glucose near the proximal end of chromosome 9 (2.37 cM) was detected in this cross, initially on the chow diet and then replicated on the Western diet. The LOD score plot for chromosome 9 has shown 2 distinct peaks, one with a suggestive LOD score at the proximal end and one with a significant LOD score at a more distal region, suggesting the existence of two loci for fasting glucose on the chromosome. , also indicated the existence of two QTLs for the trait on chromosome 9. We named the proximal one Bglu16 to represent a new QTL for fasting glucose in the mouse. Naming a suggestive locus is considered appropriate if it is repeatedly observed . One probable candidate gene for this QTL is Hnf1a, which encodes hepatocyte nuclear factor 1α. In humans, Hnf1a mutations are the most common cause of maturity-onset diabetes of the young (MODY) .

Sources of funding:
This work was supported by NIH grants DK097120 and HL112281.

Conflict of Interest Disclosures:
None.
glucose locus, and identification of apcs as an underlying candidate gene. Physiol. Genomics 44:      Two green horizontal lines represent genome-wide significance thresholds for suggestive or 27 significant linkage (P=0.63 and P=0.05, respectively). Black plots reflect the LOD score calculated at 1-cM intervals, the red plot represents the effect of the BALB allele, and the blue plot represents the effect of the SM allele. If BALB represents the high allele, then the red plot will be to the right of the graph; otherwise, it will be to the left.