Genetic Analysis of Atherosclerosis and Glucose Homeostasis in an Intercross Between C57BL/6 and BALB/cJ Apolipoprotein E-Deficient Mice

Abstract

Background

Diabetic patients have an increased risk of developing atherosclerosis and related complications compared to non-diabetic individuals. The increased cardiovascular risk associated with diabetes is due in part to genetic variations that influence both glucose homeostasis and atherosclerotic lesion growth. Mouse strains C57BL/6J (B6) and BALB/cJ (BALB) exhibit distinct differences in fasting plasma glucose and atherosclerotic lesion size when deficient in apolipoprotein E (Apoe^−/− . Quantitative trait locus (QTL) analysis was performed to determine genetic factors influencing the two phenotypes.

Methods and Results

266 female F₂ mice were generated from an intercross between B6.Apoe^−/− and BALB.Apoe^−/− mice and fed a Western diet for 12 weeks. Atherosclerotic lesions in the aortic root, fasting plasma glucose, and body weight were measured. 130 microsatellite markers across the entire genome were genotyped. Four significant QTLs, Ath1 on chromosome (Chr) 1, Ath41 on Chr2, Ath42 on Chr5, and Ath29 on Chr9, and one suggestive QTL on Chr4, were identified for atherosclerotic lesion size. Four significant QTLs, Bglu3 and Bglu12 on Chr1, Bglu13 on Chr5, Bglu15 on Chr12, and two suggestive QTLs on Chr9 and Chr15 were identified for fasting glucose levels on the chow diet. Two significant QTLs, Bglu3 and Bglu13, and one suggestive locus on Chr8 were identified for fasting glucose on the Western diet. One significant locus on Chr1 and two suggestive loci on Chr9 and Chr19 were identified for body weight. Ath1 and Ath42 coincided with Bglu3 and Bglu13, respectively, in the confidence interval.

Conclusions

We have identified novel QTLs that have major influences on atherosclerotic lesion size and glucose homeostasis. The colocalization of QTLs for atherosclerosis and diabetes suggests possible genetic connections between the two diseases.

Introduction

Type 2 diabetes mellitus (T2DM) is a major risk factor for atherosclerotic cardiovascular disease. Diabetic patients have a two- to four-fold higher risk of developing atherosclerosis and its complications compared with non-diabetic individuals ¹. Part of the increased cardiovascular risk associated with diabetes is due to genetic variations that influence both glucose homeostasis and the development of atherosclerosis. A few rare gene mutations result in both early coronary heart disease and T2DM that are observable as Mendelian traits segregatingin families, which include LRP6², ABCA1 ^3,4, and APOB ⁵. Recent genome-wide association studies (GWAS) have identified dozens of common genetic variants for both atherosclerosis and T2DM (http://www.genome.gov/GWAStudies/). Several of the variants for coronary heart disease are within genes that are involved in glucose metabolism or are associated with T2DM, such as MTHFD1L and HNF1A. Thus far few studies have been conducted to examine such genetic variants and in those that have the findings are inconsistent ^{6, 7, 8, 9,10}. One major challenge for such studies is the difficulties inherent in establishing causality between genetic variants and complex disease in humans due to small gene effects, complex genetic structure, and environmental influences.

A complementary approach to the identification of genetic components in human disease is to use animal models. One commonly usedrodent model of atherosclerosis is the apolipoprotein E-deficient (Apoe^−/−) mouse, which develops all phases of atherosclerotic lesions seen in humans, progressing from the early foam cell stage to the advanced stage with a fibrous cap and necrotic lipid core ¹¹. We have observed that atherosclerosis susceptible C57BL/6 (B6) Apoe^−/− mice develop significant hyperglycemia when fed a Western-type diet ¹². In contrast, atherosclerosis-resistant BALB/cJ (BALB) Apoe^−/− mice are highly resistant to hyperglycemia ¹³. The concordant differences between the two Apoe^−/− strains in susceptibility to atherosclerosis and to hyperglycemia provide an ideal model for investigating genetic connections between the phenotypes. In the present study, we performed quantitative trait locus (QTL) analysis on female mice from an intercross between B6. Apoe^−/− and BALB. Apoe^−/− mouse strains to investigate the genetic control of atherosclerosis and glucose homeostasis.

Methods

Mice

B6.Apoe^−/− mice were purchased from the Jackson Laboratories. BALB. Apoe^−/− mice at the N10 generation were generated in our laboratory, as previously described ¹⁴. B6. Apoe^−/− mice were crossed with BALB. Apoe^−/− mice to generate F₁s, which were intercrossed by brother-sister mating to generate a large F₂ population. Female F₂ mice were weaned at 3 weeks of age onto a rodent chow diet, and male F₂ mice were euthanized at the time of weaning. At 6 weeks of age, F₂ mice were started with a Western diet containing 21% fat, 34.1% sucrose, 0.15% cholesterol, and 19.5% casein (Harlan Laboratories, TD 88137) and maintained on the diet for 12 weeks. All this procedures were in accordance with current National Institutes of Health guidelines and approved by the University Animal Care and Use Committees.

Measurements of plasma glucose

Mice were bled twice: once before initiation of the Western diet and once at the end of the 12 weeks’ high-fat feeding period. Mice were fasted overnight before blood was drawn from the retro-orbital venous plexus with the animals under isoflurane anesthesia. Plasma glucose was measured with a Sigma glucose (HK) assay kit, which was adapted for a microplate assay. Briefly, 10 μl of plasma samples (plasma from high-fat diet fed mice was diluted 1:2 in distill water) were mixed with 90 μl of reagent in a 96-well plate. After a 15-min incubation at room temperature, the absorbance at 340 nm was read on a Molecular Devices (Menlo Park, CA) plate reader.

Aortic lesion analysis

Atherosclerotic lesions in aortic root were measured as previously reported ¹⁵. Briefly, the aortic root and adjacent heart were excised en bloc and embedded in optimal cutting temperature compound. 10-μm thick cross sections of the vessel were collected, stained with oil red O and hematoxylin, and counterstained with fast green. Atherosclerotic lesion areas were quantified using an ocular lens with a square-micrometer grid on a light microscope. The lesion areas of five sections with the largest readings were averaged for each mouse and this average was used for statistical analysis.

Genotyping

Genomic DNA was isolated from the tails of mice by using the phenol/chloroform extraction and ethanol precipitation method. A total of 130 microsatellite markers covering all 19 autosomes and the X chromosome at an average interval of 12 cM were typed. Parental and F₁ DNA served as controls for each marker.

Statistical analysis

QTL analysis was performed using J/qtl and Map Manager QTX software as previously described ^{12,15, 16}. One thousand permutations of trait values were run to define the genome-wide LOD (logarithm of odds) score threshold required to be significant or suggestive for each specific 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) according to the criteria recommended by the genetics community in 2003 ¹⁷.

Results

Trait value distributions

Fasting plasma glucose levels of 266 F₂ mice before and after 12 weeks on the Western diet, atherosclerotic lesions in the aortic root, and body weight were measured. As shown in figure 1, values of fasting plasma glucose levels on both chow and Western diets, log-transformed atherosclerotic lesion sizes, and body weight approach normal distributions. These data were analyzed to identify chromosomal regions segregating with the traits. Those loci exhibiting significant linkage and suggestive linkage are presented in Table 1.

Figure 1.

Distributions of fasting plasma glucose before (A) and after (B) 12 weeks on the Western diet, LN (natural log)-transformed atherosclerotic lesion sizes (C), and body weight (D) in 266 female F₂ mice derived from B6. Apoe^−/− and BALB. Apoe^−/− mice.

Table 1.

Significant and suggestive QTLs for atherosclerotic lesion size and fasting plasma glucose levels l in F₂ mice derived from B6. Apoe^−/− and BALB. Apoe^−/− mice.

Locus namea	Chr	Trait	LOD scoreb	95% CIc	Peakd	P valuee	High allele	Mode of inheritancef
Ath1	1	Atherosclerosis	3.82	82.3–92.3	89	0.019	B6	Additive
Ath41	2	Atherosclerosis	3.77	26.2–64.2	52.2	0.026	B6	Additive
Athsq1	4	Atherosclerosis	2.8	55.6–69.1	63.3	0.191	BALB	Additive
Ath42	5	Atherosclerosis	5.69	48.5–60.7	54.7	0.001	B6	Dominant
Ath29	9	Atherosclerosis	3.88	17.8–31.8	19.8	0.016	B6	Additive
Bglu12	1	Glucose, chow	3.94	48.3–78.3	60.3	0.022	BALB	Additive
Bglu3	1	Glucose, chow	3.45	64.3–84.3	74.3	<0.05	BALB	Additive
Bglu13	5	Glucose, chow	6.72	42.7–56.7	47.3	0.001	B6	Dominant
Bglu14	9	Glucose, chow	3.06	17.8–61.8	47.8	0.11	BALB	Recessive
Bglu15	12	Glucose, chow	3.43	5.5–21.5	10.0	0.05	B6	Additive
Fbg-2	15	Glucose, chow	2.14	18.5–33.4	26.5	0.579	B6	Heterosis
Bglu3	1	Glucose, West	4.93	76.3–86.3	82.3	0.001	BALB	Additive
Bgl13	5	Glucose, West	4.81	38.7–58.7	48.7	0.001	B6	Dominant
Giq1	8	Glucose, West	2.90	17.7–57.7	33.7	0.16	B6	Additive
BW8q1	1	Body weight	5.42	76.3–94.3	80.3	0.0001	BALB	Dominant
W10q13	9	Body weight	2.785	55.8–69.8	67.8	0.188	B6	Recessive
Wtmq9	19	Body weight	2.898	18.3–46.3	40.3	0.154	BALB	Dominant

Plasma glucose levels were measured before (chow) mice were started on the Western diet and at the end of the 12 weeks’ Western diet (West) feeding period. Blood was drawn from overnight-fasted mice.

^aQTLs were named if they were significant or if they overlapped with previously reported suggestive ones. The nomenclature of Ath was for atherosclerosis QTLs, and Bglu for blood glucose QTLs. The newly identified QTLs were underlined to easily distinguish them from known ones.

^bLOD scores were obtained from QTL analysis. The significant LOD scores were highlighted in bold. The suggestive and significant LOD score thresholds were determined by 1,000 permutation tests for each trait. Suggestive and significant LOD scores were 2.087 and 3.458, respectively, for atherosclerosis; 2.071 and 3.428 for plasma glucose on the chow diet; 2.084 and 3.423 for glucose on the Western diet, and 2.099 and 3.486 for body weight.

^c95% Confidence interval in cM obtained from a whole genome QTL scan.

^dQTL peak position in cM.

^eThe p-values reported represent the level of genome-wide significance as they were generated based on genome-wide permutation tests

^fMode of inheritance was defined according to allelic effect at the nearest marker of a QTL.

Atherosclerotic lesion size

Four significant QTLs, located on chromosomes (Chr) 1, 2, 5, and 9, and one suggestive QTL on Chr4, were identified for atherosclerotic lesion sizes (Figure 2). Details of the QTLs detected, including locus name, LOD score, 95% confidence interval (CI), peak location, genome-wide significance P value, high allele, and mode of inheritance are presented in Table 1. The two significant QTLs on Chr1 and Chr9 replicated the previously reported QTLs, Ath1 and Ath29, respectively ^15,18. The other two significant QTLs were novel. The Chr2 locus had a significant LOD score of 3.77 and a genome-wide significant P value of 0.026. It peaked at 52.2 cM and did not overlap in the confidence interval with known mouse atherosclerosis QTLs. We named it Ath41 according to the QTL nomenclature for mouse atherosclerotic lesions. The Chr5 locus had a highly significant LOD score of 5.69 and a genome-wide P value of <0.001. Its peak appeared at 54.7 cM. We named it Ath42. The Chr4 locus had a suggestive LOD score of 2.8 and peaked at 63.3 cM. This QTL was partially overlapping with Athsq1, an atherosclerotic lesion locus identified in a (MOLF/Ei x B6.Ldlr^−/−) x B6.Ldlr^−/− backcross ¹⁹. Paradoxically, the BALB allele was associated with increased lesion size while the B6 allele was associated with decreased lesion size (Table 2). In contrast, for the 4 significant QTLs, the B6 allele was the high allele that increased lesion size and the BALB allele was the low one that reduced lesion size. Ath42 affected atherosclerotic lesion size in a dominant mode of inheritance while other QTLs exhibited an additive effect on the trait.

Figure 2.

A genome-wide scan to search for loci influencing atherosclerotic lesion sizes of the F₂ mice. Chromosomes 1 through 20 are represented numerically on the X-axis. The relative width of the space allotted for each chromosome reflects the relative length of each chromosome. The Y-axis represents the LOD score. Atherosclerotic lesion sizes were determined by averaging the lesion areas of 5 cross-sections with the largest readings for each F₂ mouse. Two horizontal dashed lines denote genome-wide empirical thresholds for suggestive (P=0.63) and significant (P=0.05) linkage.

Table 2.

Effects of B6 (B) and BALB (C) alleles in different QTLs on atherosclerosis, plasma lipids, and body weight in the B6. Apoe^−/− and BALB. Apoe^−/− intercross

Locus name	Chr	Trait	Peak Marker	BB	CC	BC	P-value
Ath1	1	Atherosclerosis	D1mit406	249571±123560 (55)	167588±83413 (64)	226485±106846 (137)	5.51E-05
Ath41	2	Atherosclerosis	D2mit328	243816±109941 (61)	186641±116201 (59)	223779±110268 (135)	1.76E-02
Athsq1	4	Atherosclerosis	D4mit203	189830±108061 (60)	253744±115878 (72)	212235±109735 (126)	3.37E-03
Ath42	5	Atherosclerosis	D5mit41	243106±106054 (68)	165375±99213 (55)	231499±115933 (127)	1.68E-04
Ath29	9	Atherosclerosis	D9mit247	253591±126554 (68)	173685±87719 (54)	217985±109284 (137)	4.59E-04
Bglu12	1	Glucose, chow	D1mit495	149.2 ± 25.8 (60)	170.7 ± 28.8 (55)	161.6 ± 29.1 (111)	2.91E-04
Bglu3	1	Glucose, chow	D1mit354	149.7 ± 25.1 (53)	169.1 ± 29.7 (61)	160.7 ± 29.7 (112)	1.78E-3
Bglu13	5	Glucose, chow	D5mit312	160.5 ± 28.6 (62)	141.6 ± 19.5 (49)	168.4 ± 29.6 (119)	2.12E-07
Bglu14	9	Glucose, chow	D9mit274	159.0 ± 29.9 (60)	172.8 ± 23.7 (57)	154.3 ± 28.0 (115)	2.27E-04
Bglu15	12	Glucose, chow	D12mit84	171.7 ± 34.2 (47)	149.5 ± 26.0 (60)	160.9 ± 26.8 (125)	3.41E-04
Fbg-2	15	Glucose, chow	D15mit123	166.9 ± 27.4 (45)	166.7 ± 32.0 (64)	154.6 ± 27.4 (123)	6.51E-03
Bglu3	1	Glucose, West	D1mit354	252.0 ± 54.2 (57)	301.5 ± 70.8 (68)	288.6 ± 61.5 (128)	4.44E-05
Bgl13 Giq1	5 8	Glucose, West Glucose, West	D5mit155 D8mit68	284.3 ± 56.1 (71) 305.0 ± 60.7 (73)	249.6 ± 71.3 (51) 265.3 ± 54.9 (63)	299.5 ± 64.0 (135) 283.4 ± 70.3 (126)	1.74E-05 1.69E-03
BW8q1	1	Body weight	D1mit354	22.1 ± 2.3 (57)	25.0 ± 4.1 (68)	24.3 ± 3.4 (132)	5.45E-06
W10q13	9	Body weight	D9mit279	25.3 ± 3.5 (62)	23.2 ± 3.4 (59)	23.7 ± 3.5 (142)	2.48E-03
Wtmq9	19	Body weight	D19mit90	22.7 ± 3.2 (64)	24.8 ± 4.1 (69)	24.2 ± 3.3 (131)	1.94E-03

Measurements are expressed as means ± SD. The units for these measurements are: μm2/section for atherosclerotic lesions; mg/dl for glucose; and g for body weight. BB, homozygous for B6 alleles at the linked peak marker; CC, homozygous for BALB alleles; BC, heterozygous for B6 and BALB alleles at the peak marker. The number in the bracket represents the number of progeny with a specific genotype at a specific marker. ANOVA was used to determine the significance level (p value) of differences for a specific phenotype among progeny with three different genotypes at a specific marker.

Fasting glucose levels

For fasting glucose on the chow diet, 4 significant QTLs on Chr1, Chr5, and Chr12, and two suggestive QTLs on Chr9 and Chr15 were identified (Figure 3). The significant locus on distal Chr1 and the 2 suggestive loci on Chr9 and Chr15 replicated the previously reported QTLs, Bglu3, a suggestive locus on Chr9 (now named Bglu14), and Fbg2, respectively (Table 1). The significant QTLs on the middle portion of Chr1, Chr5, and Chr12 were novel. The LOD score plot for Chr1 displayed two distinct peaks, located approximately 14 cM apart (Figure 3). The distal QTL peaked at 74.3 CM, overlapping with Bglu3, identified in a B6 x C3H Apoe^−/− F₂ cross ¹². The proximal peak occurred at 60.3 cM with a LOD score reaching 3.94. We named this QTL Bglu12 to represent a significant mouse QTL for fasting glucose. The Chr5 locus had a highly significant LOD score of 6.72 and peaked at 47.3 cM. We named it Bglu13. The QTL on Chr9 had a suggestive LOD score of 3.06, and overlapped with a suggestive locus near D9mit229 (26.8 cM) for fasting glucose levels identified in a B6 x KK-Ay F2 cross ²⁰. We designated this QTL as Bglu14 as it has not been named. The QTL on Chr12, named Bglu15, had a significant LOD score of 3.43 and peaked at 10 cM. Bglu15 is close to the centromere compared to Fbg-1, which is located in the middle portion of Chr12 between D12mit4 (35.5 cM) and D12mit227 (38.4 cM) ²¹. The suggestive QTL on Chr15 overlapped with Fbg-2 near D15Mit87 (16.7 cM), identified in (BALB x KK/Ta) x KK/Ta backcross ²¹. For fasting glucose on the Western diet, 2 significant QTLs on Chr1 and Chr5 and 1 suggestive QTL on Chr8 were identified (Figure 4). The Chr1 QTL replicated Bglu3, and the Chr5 replicated Bglu13. The Chr8 locus overlapped with Giq1, a locus with a strong influence on the late phase of glucose tolerance test identified in a B6×KK-A cross ²². Bglu13 affected fasting glucose levels on both chow and Western diets in a dominant mode of inheritance while all other QTLs exhibited an additive effect on the trait except for the QTL on Chr9 and Chr15 that affected glucose levels in a recessive and a heterosis mode, respectively (Table 2).

Figure 3.

A genome-wide scan to search for loci for fasting plasma glucose levels when the F₂ mice were fed the chow diet. Blood was collected from overnight fasted F₂ mice the day before being started on the Western diet.

Figure 4.

A genome-wide scan to search for loci for fasting plasma glucose levels on the Western diet. Mice were fed the high-fat diet for 12 weeks. Blood was collected from overnight fasted F₂ mice the day before being euthanized.

Body weight

One significant QTL on distal Chr1 and 2 suggestive QTLs on Chr9 and Chr19 were identified for body weight (Figure 5). The Chr1 QTL overlaps with Wt3q2 and Wt6q2 for body weight mapped in two F₂ populations created from a selection and an inbred mouse lines ²³, Bw8q1 originally mapped in a B6×A/J intercross ²⁴ and then replicated in a BxH Apoe^−/− F₂ cross¹², and Nob3 for body fat body weight and blood glucose mapped in NZO×B6 F₂ females ²⁵. The Chr9 locus overlaps with W10q13 mapped in an M16i x L6 intercross ²⁶, Do2 for dietary obese mapped in two crosses derived from AKR/J and SWR/J mice ²⁷, and Obq18 for obesity mapped in B6 x 129 F₂ females ²⁸. The Chr19 locus corresponds to Wtmq9 mapped in a B6 x C3H Apoe^−/− intercross ²⁹, W3q14 from an M16i x L6 F₂ intercross ²⁶, and Abfw4 for abdominal fat mapped in DU6i x DBA/2 intercross ³⁰.

Figure 5.

A genome-wide scan to search for loci for body weight. Mice were weighed the day being euthanized.

Coincident QTLs for atherosclerosis and fasting glucose

LOD score plots for chromosome 5 show that the QTLs for atherosclerosis (Ath42) coincided with the QTL for fasting glucose (Bglu13) in the confidence interval (Figure 6). Both loci exhibited a dominant effect from the B6 allele on atherosclerotic lesions or fasting glucose levels (Table 2). LOD score plots for chromosome 1 show partial overlapping of the QTL for atherosclerosis (Ath1) with the QTLs for fasting glucose (Bglu3) and body weight in the confidence interval (Figure 7). The B6 allele was associated with increased atherosclerotic lesions and decreased glucose levels and body weight while the BALB allele was associated with decreased lesions and increased glucose levels and body weight (Table 2). Ath29 was also partially overlapping with Bglu14 in the confidence interval on the Chr9 (Table 1).

Figure 6.

LOD score plots for atherosclerotic lesion size (A), fasting plasma glucose levels on the chow diet (B), and fasting plasma glucose levels on the Western diet (C) on chromosome 5. Plots were created with the interval mapping function of Map Manager QTX, including a bootstrap test shown as a histogram estimating the confidence interval for the QTL. Two green horizontal lines represent genome-wide significance thresholds for suggestive or significant peaks (P=0.63 and P=0.05, respectively). Black plots reflect the LOD calculated at 1-cM intervals. The blue plot represents the effect of the B6 allele, and the red plot represents the effect of the BALB allele.

Figure 7.

LOD score plots for atherosclerotic lesion size (A), fasting plasma glucose levels on the chow diet (B), fasting plasma glucose levels on the Western diet (C), and body weight (D) on chromosome 1. Plots were created with the interval mapping function of Map Manager QTXb20, as stated above. The histogram denotes the confidence interval of the QTL.

Discussion

In this study, we have identified five loci contributing to the development of atherosclerosis, six loci to fasting glucose levels on the chow diet, three loci contributing to fasting glucose levels on the Western diet, and three loci for body weight in an intercross between B6 and BALB Apoe^−/− mouse strains. Moreover, we have observed the colocalization of QTLs for atherosclerotic lesions and for plasma glucose levels on chromosomes 1, 5, and 9.

B6 and BALB are prototype mouse strains for genetic studies of atherosclerosis. In pioneering studies of recombinant inbred strains derived from the two strains, as well as from B6 and C3H/HeJ, Paigen et al ¹⁸ identified the first atherosclerosis susceptibility locus, Ath-1. Subsequent studies of female F₂ and N₂ progeny derived from the B6 and BALB strains demonstrates the segregation of Ath-1 with HDL cholesterol levels ³¹. However, there are several limitations in those studies: First, the number rather than the size of atherosclerotic lesions was measured. Thus, atherosclerotic lesions were treated as a “qualitative trait” rather than a “quantitative trait”. Second, the mapping was performed using a rather small numbers of animals, especially the recombinant inbred strains, thus the power for detecting susceptibility loci was low. Third, there were fewer polymorphic markers available at the time when the studies were conducted. Lastly, the diet-induced mouse model of atherosclerosis develops only small fatty streak lesions that are largely limited to the aortic root ³². In contrast, Apoe-deficient mice develop all phases of atherosclerotic lesions in large and medium sized arteries seen in humans ¹¹. Our present work has extended the prior studies by finding five atherosclerosis QTLs, including Ath1. Among the five QTLs, Ath1, Athsq1, and Ath29, have been previously reported ^15,18,33. Tnfsf4 has been identified to be the causal gene of Ath1 ³⁴. We recently have identified Rcn2, a calcium-binding protein in the endoplasmic reticulum, as a key regulator in oxidized phospholipid-induced cytokine production and a probable candidate gene of Ath29 ³⁵. The confidence interval of Athsq1 is corresponding to human chromosome 9p21, a region that is associated with coronary heart disease ^{36, 37,38}.

The QTL on chromosome 2, named Ath41, is close to Athla1, an atherosclerosis susceptibility locus mapped in a (PERA×B6-Ldlr^−/−)×B6-Ldlr^−/− N₂ backcross mice ³⁹. Athla1 is located in a more distal region (69 cM), and it increases lesion size only when homozygous for the B6 allele. Candidate genes for Ath41 include Dab2ip, Tfpi, and Slc38a11, which have been shown to be associated with coronary heart disease in humans ^40,41,42.

We identified a major locus on chromosome 5, approximately between 40 and 60 cM, which affected both atherosclerotic lesion size and fasting plasma glucose levels. We named it Ath42 for atherosclerotic lesions and Bglu13 for fasting glucose. As the two loci overlap significantly in the confidence interval, it is plausible to postulate that they share the same underlying causal gene. The present observation that both QTLs exhibited the same dominant B6 allele effect on the two different traits supports this speculation. Nevertheless, it is also likely that the two phenotypes are affected by two linked but unique genes residing in the QTL interval. The region from 40 to 60 cM on chromosome 5 in the mouse corresponds to chromosomal regions of 4q13, 4q21, and 12q24 in humans. The 4q13 region has been shown to be associated with variations in metabolic traits, including blood glucose ^43,44, and the 12q24 region is associated with coronary heart disease ^45,46, metabolic syndrome ^47,48, type 1 and type 2 diabetes ^49,50,51. One promising candidate gene in the region is Hnf1a, which encodes hepatocyte nuclear factor 1α. One A/G SNP in exon 9 between B6 and BALB leads to amino acid substitution (P580R) in the Hnf1a protein. In humans, Hnf1a mutations are the most common cause of maturity-onset diabetes of the young (MODY) ⁵². Polymorphisms in the Hnf1a gene are associated with risk for T2DM and coronary heart disease ^{49, 53}.

The QTLs for plasma glucose and body weight on distal chromosome 1 have been reported previously in three separate mouse intercrosses, including two B6×C3H crosses deficient in Apoe ^{12, 29}, and a cross between New Zealand obese (NZO) and B6 mice ²⁵. The confidence interval of the QTLs overlaps with a region of linkage to type 2 diabetes found in multiple human populations that has been extensively examined by the International Chromosome 1q Type 2 Diabetes Consortium ⁵⁴. In the current cross, we have observed two distinct peaks of the linkage curve for plasma glucose on the chow diet with the distal peak at 74.3 cM and the proximal peak at 60.3 cM. The bootstrap test, an effective statistical method for defining the confidence interval of QTLs using simulation ⁵⁵, also indicated the existence of two QTLs for the trait on chromosome 1. We have named the proximal QTL Bglu12 to represent a new locus for fasting glucose in the mouse.

The QTLs for atherosclerosis (Ath1), fasting glucose (Bglu3), and body weight (BW8q1) overlap in the confidence interval on the distal chromosome 1 region. The B6 allele was associated with increased atherosclerosis but decreased glucose levels and body weight. Apoa2 is a major gene in the region that may contribute to variations in the traits. The QTL effect on body weight disappeared when the influence from the Apoa2 allele was eliminated ⁵⁶. On the other hand, transgene expression of Apoa2 in mice results in several phenotypes observed in T2DM, including glucose intolerance, insulin resistance, hypertriglyceridemia, and obesity ^57,58. Apoa2 is also a major gene in the mouse that has a dramatic influence on plasma HDL cholesterol levels ⁵⁹. BALB mice have an Apoa2^b allele that elevates HDL cholesterol levels and B6 mice have an Apoa2^a allele that decreases HDL cholesterol levels ⁶⁰. High HDL cholesterol levels protect against atherosclerosis. Apcs, encoding serum amyloid P (SAP), is another candidate in the distal chromosome 1 region that may contribute to T2DM and atherosclerosis. Plasma SAP levels, which are primarily regulated by the Apcs gene, are correlated with blood glucose and body weight in a segregating F₂ population derived from B6 and C3H Apoe^−/− mice¹². In humans SAP is significantly correlated with obesity, blood pressure, lipids, common and internal carotid wall thickness, and ankle-brachial index ⁶¹. Cxcr4, Pask, Cntnap5a, Lct, and Pik3c2b are positional candidate genes for Bglu12. Variants of these genes have been found to be associated with susceptibility to T2DM, fasting glucose, or insulin resistance in humans ^{62, 63,64, 65, 66, 67}.

The QTL for fasting glucose on chromosome 9 was partially overlapping with the QTL for atherosclerosis. Sorl1, Rcn2, and Apoc3 are potential candidate genes in the region that may affect both atherosclerosis and T2DM ^{35, 68, 69}. For Bglu15, Adam17, encoding a disintegrin and metallopeptidase domain 17, and Ahr, encoding aryl-hydrocarbon receptor, are two likely candidate genes. Adam17 is involved in the shedding of the extracellular domain of cytokines, growth factors, receptors or adhesion molecules ⁷⁰. Ahr signaling affects molecular clock genes associated with glucose metabolism, and Ahr deficiency enhances insulin sensitivity and reduces PPAR-α pathway activity ⁷¹. In the present study, we have found some QTLs, such as Bglu3 and Bglu13, that influenced glucose levels when mice had normoglycemia on a chow diet also affected glucose levels when mice developed hyperglycemia on a high-fat diet. However, some other QTLs, such as Bglu12, Bglu14, and Bglu15 only exerted effect under a specific condition (the chow diet). Six QTLs were found for fasting plasma glucose when mice were fed the chow diet while only three QTLs were detected when fed the Western diet. The reasons for the discrepancy in the results are unknown. One probable explanation is that the Western diet has a significant influence on plasma glucose levels, which might overwhelm the influence from genetic factors on the trait. In addition, the western diet induces some metabolic changes that suppress gene expression. We recently have found that there are much more differentially expressed genes in the aorta of two strains when fed a chow diet than a Western diet ⁷².

In summary, we have identified multiple QTLs contributing to the development of atherosclerosis and glucose homeostasis in a segregating F₂ population. The finding on the colocalization of QTLs for atherosclerosis and glucose has laid the basis for further study to determine whether they are controlled by the same genes or different unique genes in the QTL intervals.