Major reduction of atherosclerosis in fractalkine (CX3CL1)-deficient mice is at the brachiocephalic artery, not the aortic root

Abstract

Fractalkine (CX3CL1) is of particular interest in atherogenesis because it can serve as an adhesion molecule and a chemokine. Fractalkine and its receptor CX3CR1 are expressed in atherosclerotic lesions of humans and mice. However, the effect of fractalkine deficiency on atherosclerosis susceptibility is unknown. Fractalkine-deficient mice on the C57BL/6 (B6) background were bred to the atherosclerosis-sensitizing B6.ApoE^-/- and B6.LDLR^-/- backgrounds. Compared with controls, aortic-root lesion area was unchanged in fractalkine-deficient male and female B6.ApoE^-/- mice at 16 weeks of age and males at 12 weeks of age, but it was mildly reduced (30%, P = 0.005) in females at 12 weeks of age. In contrast, lesion area at the brachiocephalic artery (BCA) was reduced dramatically by ≈85% in fractalkine-deficient females [42,251 ± 26,136 μm² (n = 15) vs. 6,538 ± 11,320 μm²;(n = 24), P < 0.0001] and males [36,911 ± 32,504 μm² (n = 24) vs. 6,768 ± 8,595 μm² (n = 14); P = 0.001] at 16 weeks of age. Fractalkine-deficient B6.ApoE^-/- mice were comparable with controls in body weight, plasma cholesterol, plasma high-density lipoprotein cholesterol and white blood cell counts. On the B6.LDLR^-/- background, lesion areas were reduced by 35% at the aortic root (P < 0.01) and by 50% at the BCA (P < 0.05) in fractalkine-deficient females at 16 weeks of age. Lesions in fractalkine-deficient mice on the B6.ApoE^-/- and B6.LDLR^-/- backgrounds were less complex and contained significantly fewer macrophages than controls. In conclusion, the major reduction of atherosclerosis in fractalkine-deficient mice appears to be at the BCA rather than the aortic root.

Inflammatory processes play a critical role in atherogenesis (1). Monocytes adhere to the vessel wall and are recruited into the subendothelial space where they differentiate into macrophages and subsequently develop into foam cells (2). Adhesion molecules and chemokines play an important role in early lesion development (3-5). Fractalkine (CX3CL1) is particularly interesting in this regard, because it not only mediates strong binding of monocytes and T cells to the endothelium, but it can be cleaved from the cell membrane and serve as a chemokine (6, 7). Fractalkine consists of an N-terminal chemokine domain, a mucin-like stalk, and transmembrane and cytoplasmic domains. The metalloprotease ADAM-17 can cleave the mucin-like stalk and release the soluble-chemokine domain (8). A fractalkine receptor, CX3CR1, which is a G-protein-coupled receptor that can mediate leukocyte adhesion and migration and also triggers signaling, has been identified (7, 9).

Both fractalkine and CX3CR1 are expressed in atherosclerotic lesions of humans and mice (10). A CX3CR1 polymorphism that changes amino acid residue 249 has been associated with increased risk for acute coronary events (11) and prevalence and severity of coronary artery disease (12). Also, CX3CR1-deficient mice were shown to have diminished atherosclerotic lesions on the atherosclerosis-sensitizing apolipoprotein E-deficient background (13, 14).

However, the effect of fractalkine on atherosclerosis susceptibility has not yet been reported. Fractalkine knockout mice were shown to have normal gross anatomy and hematological profiles, except for a decrease in the number of circulating leukocytes expressing the cell surface marker F4/80 (15). Interestingly, fractalkine-deficient mice were less susceptible to cerebral ischemia-reperfusion injury (16).

The aim of the current study was to investigate the effect of fractalkine deficiency on atherosclerotic lesion formation at the aortic root and the brachiocephalic artery (BCA) in apolipoprotein E-deficient (ApoE^-/-) and low-density lipoprotein (LDL)-receptor (LDLR)-deficient (LDLR^-/-) mice.

Methods

Mice. Heterozygous fractalkine knockout mice were obtained from Millennium Pharmaceuticals (16) and backcrossed at The Rockefeller University to the C57BL/6J background (stock no. 000664, The Jackson Laboratory) by using the marker-assisted technique (17). These mice (B6.129SvEv-CX3CL1^tm, hereafter referred to as B6.CX3CL1^+/-) were then crossed with ApoE^-/- mice (B6.129P2-Apoe^tm1Unc/J, stock no. 002052, The Jackson Laboratory, hereafter referred to as B6.ApoE^-/-) and LDLR-deficient mice (B6.129S7-Ldlr^tm1Her/J, stock no. 002207, The Jackson Laboratory, hereafter referred to as B6.LDLR^-/-). These strains were then used to generate CX3CL1^+/+, CX3CL1^+/-, and CX3CL1^-/- mice on both the B6.ApoE^-/- and B6.LDLR^-/- backgrounds. For atherosclerosis studies, mice were weaned at 28 days of age and fed a semisynthetic modified AIN76 diet containing 0.02% cholesterol (18) until they were killed at 12 or 16 weeks of age. On the day on which the animals were killed, food was removed from the cage at 9 a.m. and mice allowed access to water. At the time at which the animals were killed (≈8 h later), the mice were exsanguinated by left-ventricular puncture, and the blood was collected into EDTA-containing syringes. The circulation was flushed with PBS, and the heart and BCA were removed and stored in buffered formalin or snap-frozen in Tissue-Tek OCT compound, respectively (18). The animals were housed in the The Rockefeller University Laboratory Animal Research Center in a specific pathogen-free environment in rooms with a 7 a.m. to 7 p.m. light/dark cycle. The Rockefeller University Institutional Animal Care and Use Committee approved all procedures involving mice.

Fractalkine Genotyping. Animals were genotyped for fractalkine deficiency by PCR of genomic DNA. For both the wild-type and mutant alleles, we used the following common forward primer: CX3CL1-com, GCGAAGGAGTCTGCGGGTAGC. The reverse primers for the wild-type and mutant alleles, respectively, were as follows: CX3CL1-wt, TCACTCCACATTGTGGGAAAGGAA; and CX3CL1-ko, CGTGGGATCATTGTTTTTCTCTTG. The PCR products were resolved by agarose gel electrophoresis. Allele sizes were 415 bp for the wild type and 563 bp for the fractalkine knockout allele.

Blood Analyses. At the time of killing, whole blood was drawn by left-ventricular puncture. For blood cell counting, 36 μl of whole blood was diluted in 144 μl of PBS containing 5% BSA and analyzed in an automated hematology analyzer (Advia 120, Bayer, Wuppertal, Germany) that was calibrated for mouse blood. Lipoproteins were isolated by sequential ultracentrifugation from 60 μl of plasma at densities (d) of <1.006 g/ml (very-low-density lipoprotein), 1.006 ≤ d ≤ 1.063 g/ml (intermediate-density lipoprotein and LDL), and d >1.063 g/ml (high-density lipoprotein) in a TL-100 ultracentrifuge (Beckman Coulter). Cholesterol was determined enzymatically by using a colorimetric method (Roche).

Quantification of Atherosclerosis. To quantify atherosclerosis at the aortic root, formalin-fixed hearts were processed as described (18). To quantify atherosclerosis at the BCA, the OCT-compound-embedded vessels were sectioned from distal to proximal at a 10-μm thickness. Atherosclerotic lesions luminal to the internal elastic lamina were quantified in three equidistant oil red O-stained sections 200, 400, and 600 μm from the branching point of the BCA into the carotid and subclavian arteries (18).

Immunostaining. Frozen sections of the BCA were fixed in ice-cold acetone (10 min) and washed with PBS. Peroxidases were quenched with 1% H₂O₂ (10 min). Sections were washed and blocked with 5% normal serum (goat serum for CD68 and sheep serum for α-actin and CX3CL1) for 20 min. Primary antibodies were incubated for 60 min (rat anti-mouse CD68, MAC1957GA, Serotec, 1:100 dilution; rabbit anti-mouse α-actin, BT-560, Biomedical Technologies, 1:80 dilution; and rabbit anti-mouse CX3CL1, Millennium Pharmaceuticals, 1:80 dilution). Sections were washed and incubated with horseradish peroxidase (HRP)-conjugated secondary antibody (1:50 dilution) for 30 min (goat anti-rat IgG/HRP, STAR 72, Serotec, for CD68; sheep anti-rabbit IgG/HRP STAR54, Serotec, for α-actin and CX3CL1). After washing, peroxidase was visualized by incubation with Nova Red (Vector Laboratories) and sections were counterstained with hematoxylin. Sections were dried and permanently mounted with VectaMount mounting medium (Vector Laboratories).

Statistical Analysis. All data are expressed as mean ± SD unless indicated otherwise. Statistically significant differences between two groups were analyzed by using the Mann-Whitney or χ² tests. Statistically significant differences between more than two groups were analyzed by using the Kruskal-Wallis test and a posttest to determine differences between any two groups (prism software, GraphPad). Analysis of covariance was calculated by using ncss software. A probability value of ≤0.05 was considered to be significant.

Results

Fractalkine-deficient mice were bred to a congenic B6 background and subsequently crossed to the atherosclerosis-sensitive B6.ApoE^-/- background. At 16 weeks of age, B6.ApoE^-/- CX3CL1^+/+, B6.ApoE^-/-CX3CL1^+/-, and B6.ApoE^-/- CX3CL1^-/- mice had comparable body weights and plasma total and high-density lipoprotein cholesterol levels (Table 1), and they did not differ in aortic-root lesion area (females: 289,591 ± 51,971 μm², 277,039 ± 76,145 μm², and 271,289 ± 79,646 μm², respectively; males: 169,911 ± 66,310 μm², 138,754 ± 66,448 μm², and 142,065 ± 65,484 μm², respectively) (Fig. 1 A and C). However, there were significant differences between the genotypes in BCA lesion area. Lesions at the BCA were quantified at three sections, which were located 200, 400, and 600 μm proximal from its branching site into the carotid and subclavian arteries. As shown in Fig. 1B, lesion area was dramatically reduced by ≈85% in female B6.ApoE^-/-CX3CL1^-/- mice at 200 μm and 400 μm from the bifurcation, compared with controls (42,251 ± 26,136 μm² vs. 6,538 ± 11,320 μm²; 24,778 ± 15,207 μm² vs. 3,609 ± 7,334 μm², respectively; P < 0.0001). Moreover, approximately one-half of the fractalkine-deficient mice had no detectable lesions in the 200- and 400-μm sections, whereas all of the controls had lesions (P ≤ 0.001 χ² test). Lesion area was not reduced significantly in B6.ApoE^-/-CX3CL1^-/- females at 600 μm from the bifurcation. As shown in Fig. 1D, in fractalkine-deficient males, BCA lesion area was reduced 82% at 200 μm (36,911 ± 32,504 μm² vs. 6,768 ± 8,595 μm²; P = 0.001), 58% at 400 μm (37,428 ± 32,177 μm² vs. 15,635 ± 20,840 μm²; P = 0.02) and 59% at 600 μm (34,600 ± 28,505 μm² vs. 14,185 ± 17,188 μm²; P = 0.02).

Table 1. Body weight, plasma cholesterol, and blood count in B6.ApoE^-/- mice deficient for CX3CL1 (16 weeks of age).

	Female B6.ApoE-/-			Male B6.ApoE-/-
Measurement	CX3CL1+/+	CX3CL1+/-	CX3CL1-/-	CX3CL1+/+	CX3CL1+/-	CX3CL1-/-
Body weight, g	21.2 ± 1.8 (15)	21.2 ± 3.1 (31)	21.0 ± 2.0 (26)	29.3 ± 3.0 (28)	30.0 ± 3.7 (21)	28.5 ± 3.1 (17)
Cholesterol, mmol liter	10.2 ± 2.5 (15)	9.2 ± 1.6 (29)	8.9 ± 1.6 (26)	11.8 ± 3.2 (28)	12.2 ± 2.4 (21)	10.9 ± 2.5 (17)
HDL-C, mmol liter	0.4 ± 0.1 (15)	0.4 ± 0.2 (29)	0.5 ± 0.3 (26)	0.6 ± 0.2 (28)	0.6 ± 0.3 (21)	0.7 ± 0.6 (17)
WBC, × 109 per liter	0.9 ± 0.6 (5)	0.7 ± 0.6 (8)	0.8 ± 0.4 (8)	0.8 ± 0.4 (8)	1.1 ± 0.6 (8)	0.8 ± 0.3 (8)
Neutrophils, %	23 ± 5 (5)	24 ± 10 (8)	25 ± 9 (8)	20 ± 6 (8)	21 ± 8 (8)	21 ± 5 (8)
Lymphocytes, %	65 ± 7 (5)	64 ± 13 (8)	54 ± 17 (8)	69 ± 9 (8)	69 ± 9 (8)	66 ± 6 (8)
Monocytes, %	2 ± 1 (5)	2 ± 1 (8)	1 ± 2 (8)	2 ± 1 (8)	2 ± 1 (8)	1 ± 2 (8)

Data are given as mean ± SD. The number of mice is given in parentheses. P values were calculated with the Kruskal-Wallis test, and no statistically significant differences were found. HDL-C, high-density lipoprotein cholesterol; WBC, white blood cells.

Fig. 1.

Atherosclerotic lesion area in CX3CL1-deficient B6.ApoE^-/- mice at 16 weeks of age at the aortic root and BCA. Sections at the BCA were quantified 200, 400, and 600 μm proximal to the branching point of the BCA into the subclavian and carotid arteries. (A) Aortic-root lesion area in female B6.ApoE^-/- mice. (B) BCA lesion area in female B6.ApoE^-/- mice. (C) Aortic-root lesion area in male B6.ApoE^-/- mice. (D) BCA lesion area in male B6.ApoE^-/- mice. Bars represent means ± SEM, and the numbers of animals in the individual groups are given on the bars.

The effect of fractalkine deficiency on atherosclerosis progression was examined also on the LDLR-deficient background. Fractalkine deficiency had no effect on body weight in females; however, there was a significant effect in males that was largely due to decreased weight in the fractalkine heterozygous knockout animals (Kruskal-Wallis test, P = 0.02; Dunn's posttest, P < 0.05) (Table 2). There was also a significant effect in female mice of fractalkine genotype on total plasma cholesterol concentrations, with a decrease observed in fractalkine homozygous knockout mice (Kruskal-Wallis test, P = 0.004; Dunn's posttest, P < 0.01). In males, total plasma cholesterol concentrations were also reduced significantly in fractalkine heterozygous and homozygous knockout mice (Kruskal-Wallis test, P = 0.0005; Dunn's posttest, P < 0.01; Table 2). In both sexes, the reduction of total plasma cholesterol was mainly due to significantly reduced LDL-cholesterol concentrations in fractalkine knockout mice (Table 2). Compared with B6.LDLR^-/-CX3CL1^+/+ mice, aortic-root lesion area was reduced significantly in females by 28% in B6.LDLR^-/-CX3CL1^+/- and 35% in B6.LDLR^-/-CX3CL1^-/- mice (101,663 ± 38,987 μm² vs. 73,468 ± 28,599 μm² and 65,870 ± 24,381 μm²; Kruskal-Wallis test, P = 0.001; Dunn's posttest, P < 0.01 and 0.05, respectively) (Fig. 2A). Analysis of covariance suggested that the reduction of aortic-root lesion area in female B6.LDLR^-/-CX3CL1-deficient mice was independent of their reduction in total- and LDL-cholesterol levels. Compared with B6LDLR^-/-CX3CL1^+/+ mice, there was no significant reduction of aortic-root lesion area in males in fractalkine heterozygous or homozygous knockout mice (35,936 ± 17,976 μm² vs. 29,976 ± 18,056 μm² and 32,360 ± 17,890 μm², respectively).

Table 2. Body weight, plasma cholesterol, and blood count in B6.LDLR^-/- mice deficient for CX3CL1 (16 weeks of age).

	Female B6.LDLR-/-				Male B6.LDLR-/-
Measurement	CX3CL1+/+	CX3CL1+/-	CX3CL1-/-	P	CX3CL1+/+	CX3CL1+/-	CX3CL1-/-	P
Body weight, g	19.9 ± 1.6 (24)	20.3 ± 1.1 (29)	19.5 ± 1.5 (29)	n.s.	28.4 ± 4.0 (38)	25.9 ± 2.5 (30)	26.4 ± 2.7 (33)	0.03
Cholesterol, mmol liter	12.6 ± 2.9 (24)	11.4 ± 2.2 (29)	10.7 ± 2.5 (29)	0.004	10.3 ± 2.5 (38)	8.4 ± 2.7 (30)	8.2 ± 2.0 (33)	0.0005
VLDL-C, mmol liter	2.0 ± 1.5 (23)	1.4 ± 0.7 (28)	1.3 ± 0.8 (26)	n.s.	0.7 ± 0.4 (29)	0.5 ± 0.4 (26)	0.5 ± 0.4 (33)	0.01
LDL-C, mmol liter	8.0 ± 1.6 (23)	7.5 ± 1.2 (28)	6.8 ± 1.0 (26)	0.006	6.8 ± 1.9 (29)	5.0 ± 2.1 (26)	5.2 ± 1.6 (33)	0.001
HDL-C, mmol liter	0.9 ± 0.4 (23)	1.0 ± 0.4 (28)	1.0 ± 0.4 (26)	n.s.	1.4 ± 0.3 (29)	1.7 ± 0.6 (26)	1.4 ± 0.3 (33)	n.s.
WBC, × 109 per liter	1.1 ± 0.4 (3)	0.9 ± 0.1 (3)	1.2 ± 0.7 (16)	n.s.	1.1 ± 0.3 (8)	—	1.3 ± 0.8 (18)	n.s.
Neutrophils, %	15 ± 3 (3)	19 ± 1 (3)	16 ± 4 (16)	n.s.	18 ± 6 (8)	—	19 ± 9 (18)	n.s.
Lymphocytes, %	78 ± 2 (3)	69 ± 3 (3)	72 ± 7 (16)	n.s.	73 ± 5 (8)	—	70 ± 11 (18)	n.s.
Monocytes, %	2 ± 1 (3)	2 ± 1 (3)	1 ± 1 (16)	n.s.	1 ± 1 (8)	—	1 ± 1 (18)	n.s.

Data are given as mean ± SD. The number of mice is given in parentheses. P values were calculated with the Kruskal-Wallis test. VLDL-C, very-low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; WBC, white blood cells; n.s., Not significant.

Fig. 2.

Atherosclerotic lesion area in CX3CL1-deficient B6.LDLR^-/- mice at 16 weeks of age at the aortic root and BCA. Sections at the BCA were quantified 200, 400, and 600 μm proximal to the branching point of the BCA into the subclavian and carotid arteries. (A) Aortic-root lesion area in female B6.LDLR^-/- mice. (B) BCA lesion area in female B6.LDLR^-/- mice. Bars represent means ± SEM, and the numbers of animals in the individual groups are given on the bars.

BCA lesion areas were considerably smaller on the B6.LDLR^-/- background than on the B6.ApoE^-/- background. Nevertheless, compared with B6.LDLR^-/-CX3CL1^+/+ mice, female B6.LDLR^-/-CX3CL1^-/- mice had a ≈50% decrease in BCA lesion area at 200 μm (11,500 ± 9,969 μm², and 4,752 ± 6,637 μm²; P = 0.02), but no significant difference was present in the 400- or 600-μm sections (Fig. 2B). BCA lesion areas were not quantified in male mice because the lesions are much smaller than those of female mice (18) and, under the conditions of this study, were not large enough for meaningful comparisons between the different genotypes to be made.

The observation that fractalkine deficiency decreased aortic-root lesion area at 16 weeks of age in female B6.LDLR^-/- but not female B6.ApoE^-/- mice suggests a gene interaction of fractalkine with the atherosclerosis-sensitizing background. However, aortic-root lesions were much larger in B6.ApoE^-/- mice, and it was possible that the lesions were too advanced in these mice for an effect of fractalkine to be detected. Therefore, aortic-root lesion areas were compared between wild-type and fractalkine-deficient B6.ApoE^-/- mice that were killed at an earlier stage of lesion development (i.e., at 12 instead of 16 weeks of age). In this experiment, there were no differences between the genotypes in weight, total cholesterol, or high-density lipoprotein cholesterol (data not shown). However, compared with controls, aortic-root lesion area was reduced in females by ≈30% in fractalkine-deficient mice (107,489 ± 42,431 μm² vs. 75,160 ± 33.617 μm², respectively; Mann-Whitney test, P = 0.005) (Fig. 3A). As observed when the mice were killed at 16 weeks of age, there was no effect of fractalkine deficiency on aortic-root lesion area in males (65,653 ± 35,519 μm² and 70,344 ± 31,072 μm², P value, not significant) (Fig. 3B). Thus, at the aortic root, fractalkine deficiency slows lesion progression independent of the atherosclerosis-sensitizing background (B6.ApoE^-/- vs. B6.LDLR^-/-). However, a sexual dimorphism exists because the effect is seen in females but not in males. This gender-specific effect contrasts with the effect of fractalkine deficiency on brachiocephalic lesion area, where the effect is stronger and is detected in both females and males.

Fig. 3.

Atherosclerotic lesion area in CX3CL1-deficient B6.ApoE^-/- mice at 12 weeks of age. (A) Aortic-root lesion area in female B6.ApoE^-/- mice. (B) Aortic-root lesion area in male B6.ApoE^-/- mice. Bars represent means ± SEM, and the numbers of animals in the individual groups are given on the bars. n.s., Not significant.

It is likely that fractalkine participates in atherosclerotic lesion formation as an adhesion molecule and chemokine, recruiting monocytes into the vessel wall, which subsequently transform into macrophages. Therefore, in control and fractalkine-deficient mice, macrophage accumulation in the BCA at 200 μm proximal to the bifurcation was quantified by CD68 staining. In B6.ApoE^-/- mice, fractalkine deficiency resulted in an 80% decrease in CD68 staining area (30,439 ± 19,729 μm² vs. 6,014 ± 9,591 μm², P < 0.0001) (Fig. 4A), whereas in B6.LDLR^-/- mice, fractalkine deficiency reduced the CD68 staining area by 57% (12,800 ± 10,605 μm² vs. 5,468 ± 7,016 μm², P = 0.01) (Fig. 4B).

Fig. 4.

CD68-stained area of atherosclerotic lesions at the BCA in CX3CL1-deficient B6.ApoE^-/- and B6.LDLR^-/- mice at 16 weeks of age. (A) Female B6.ApoE^-/- mice. (B) Female B6.LDLR^-/- mice.

To assess the influence of fractalkine deficiency on plaque morphology, serial sections of the BCA were stained and compared between B6.ApoE^-/-CX3CL1^+/+ and B6.ApoE^-/-CX3CL1^-/- mice (Fig. 5). Lesions were much less advanced in the fractalkine-deficient mice. CD68 staining was more uniform in the fractalkine-deficient mouse lesions, suggesting that lesions were composed almost entirely of foam cells (Fig. 5D), whereas CD68 staining of CX3CL1^+/+ lesions indicated greater heterogeneity (Fig. 5C). The less complex lesion morphology in fractalkine-deficient mice was confirmed by α-actin staining of lesions, which in CX3CL1^+/+ mice revealed many more smooth muscle cells in the core of the lesions as well as the presence of a fibrous cap (Fig. 5 G and H). The influence of fractalkine deficiency on plaque morphology was assessed also in B6.LDLR^-/- females (Fig. 6). On this background, CX3CL1^+/+ and CX3CL1^-/- animals had small uniform lesions consisting mainly of CD68-staining macrophages (Fig. 6 C and D), with very few α-actin-staining smooth muscle cells (Fig. 6 G and H). Whereas immunostaining revealed the presence of fractalkine in the fibrous cap, the core of the lesion, and medial smooth muscle cells (Fig. 5E) in the B6.ApoE^-/-CX3CL1^+/+ mice, fractalkine was visualized only beneath the internal elastic lamina in the media in the B6.LDLR^-/-CX3CL1^+/+ mice (Fig. 6E). The lack of intimal staining for fractalkine was probably due to the small size and lack of complexity of the lesions in this model. The medial cells that stained for both oil red O and fractalkine but not CD68 may be lipid-containing medial smooth muscle cells.

Fig. 5.

Representative serial sections through the BCA of CX3CL1-deficient B6.ApoE^-/- mice at 16 weeks of age. Oil red O staining in CX3CL1^+/+ (A) and CX3CL1^-/- (B) mice. CD68 staining (red) in CX3CL1^+/+ (C) and CX3CL1^-/- (D) mice. CX3CL1 staining (red) in CX3CL^+/+ (E) and CX3CL1^-/- (F) mice. α-Actin staining (red) in CX3CL1^+/+ (G) and CX3CL1^-/- (H) mice.

Fig. 6.

Representative serial sections through the BCA of CX3CL1-deficent B6.LDLR^-/- mice at 16 weeks of age. Oil red O staining in CX3CL1^+/+ (A) and CX3CL1^-/- (B) mice. CD68 staining (red) in CX3CL1^+/+ (C) and CX3CL1^-/- (D) mice. CX3CL1 staining (red; underneath internal elastic lamina marked by arrows) in CX3CL^+/+ (E) and CX3CL1^-/- (F) mice. α-Actin staining (red) in CX3CL1^+/+ (G) and CX3CL1^-/- (H) mice.

Discussion

This study shows a striking effect of fractalkine deficiency on atherosclerotic lesion area in the BCA, with only a minor effect at the aortic root. Fractalkine deficiency decreased lesion area in both female and male B6.ApoE^-/- mice by up to ≈85% in the portion of the BCA closest to its bifurcation into the carotid and subclavian arteries. In contrast, on the same genetic background at the same time point, fractalkine deficiency had no effect on aortic-root lesion area in female or male mice. Although fractalkine deficiency also decreased aortic-root lesion area at an earlier time point, this effect was relatively small (≈30%) and occurred only in females. On the B6.LDLR^-/--sensitizing background, both aortic root and BCA lesion areas were smaller to begin with, and fractalkine deficiency decreased aortic-root lesion area in female mice by ≈35% and in the portion of the BCA closest to its bifurcation by ≈50%. On this background, fractalkine deficiency had no effect on aortic-root lesion area in male mice and no effect on BCA lesion area in the portion of the vessel closest to the aorta in female mice. Thus, the effect of fractalkine on atherosclerotic lesion development is strongly influenced by lesion location and sensitizing background.

Although few studies have examined the effects of potential atherosclerosis modifiers at more than one location, there is some precedent for site-selective modulators of atherosclerosis akin to what we report for fractalkine (19). In male ApoE^-/- mice fed a Western-type diet with probucol (1%) for 24 weeks, cross-sectional lesion area was increased at the aortic root but decreased at the aortic arch, descending aorta, and proximal abdominal aorta (20). In another study, LDLR^-/- mice at 6 weeks after transplantation of IL-4 deficient bone marrow were fed a high-fat, high-cholesterol, cholic-acid-containing diet for 4 weeks. Compared with transplantation of control bone marrow, the IL-4 deficient bone marrow resulted in no change in aortic-root lesion area, but it reduced en face lesion area by approximately two-thirds at the aortic arch and thoracic aorta (21). Compared with immunocompetent ApoE^-/- mice, immunodeficient ApoE^-/-RAG2^-/- mice fed a Western-type diet for 27 weeks had an 81% decrease in aortic-root lesion area but no change in BCA cross-sectional lesion area (22). A similar finding was reported for mice fully backcrossed to the B6.LDLR^-/- background, although in partially (93%) backcrossed mice a reduction in lesion area was seen at both the aortic root and the BCA (23). Differences in lesion formation at different sites of the vasculature may be due to the specific flow conditions at these anatomical locations (19). Studies in human aortic endothelial cells have shown remarkable expression differences of a variety of genes, including cell-adhesion molecules, when exposed to disturbed flow or steady laminar flow (24). Although fractalkine has not been included in these studies, we speculate that differences in atherosclerosis susceptibility are due to differences in fractalkine expression in response to flow conditions or to the interaction of fractalkine with differentially regulated genes. Because the accumulation of macrophage foam cells, as shown by oil red O and CD68 staining, was markedly reduced at the BCA of fractalkine-deficient mice compared with controls, it appears that fractalkine exerts its effect on atherosclerotic lesion formation through the modulation of monocyte recruitment into the vessel wall of the BCA, which subsequently transform into macrophages.

The effect of fractalkine deficiency on BCA cross-sectional lesion area was more pronounced on the potent B6.ApoE^-/- than on the milder B6.LDLR^-/--sensitizing background, suggesting a fractalkine-ApoE^-/- background interaction. If there were no specific fractalkine-background interaction, then the strongest effect of fractalkine deficiency would be expected on the weaker rather than the stronger sensitizing background. Few studies have assessed genetic modifiers of atherosclerotic lesion development on different sensitizing backgrounds. In recent work, we have mapped aortic-root lesion area loci in crosses between C57BL/6J and FVB/N strains on both the ApoE^-/- and LDLR^-/- backgrounds. Both backgrounds revealed a locus on chromosome 10, but the ApoE^-/- background yielded a distinct locus on chromosome 19 and the LDLR^-/- background distinct loci on chromosomes 3, 12, and 18 (D.T. and J.L.B., unpublished data). Although both ApoE and the LDLR are important for lipoprotein clearance from the circulation, their functions are not identical. ApoE is a ligand for the entire LDLR family, not just the LDLR. Also, ApoE synthesis in macrophages is antiatherogenic, independent of the lowering of plasma cholesterol (25). Last, ApoE^-/- mice accumulate cholesterol-rich very-low-density lipoprotein and intermediate-density lipoprotein in plasma, whereas LDLR^-/- mice accumulate LDL.

In 2003, two studies were published in which the effect of the fractalkine receptor (CX3CR1) knockout on atherosclerosis was examined. Lesnik et al. (13) compared ApoE^-/- mice with double-knockout ApoE^-/-CX3CR1^-/- mice that were fed a Western-type diet for 5, 10, and 15 weeks. They reported that aortic root cross-sectional lesion area was significantly reduced by 32% at 10 weeks of diet (the only time point examined). Significant reductions of en face lesion area for the total aorta and the thoracic plus abdominal aorta were observed at 5, 10, and 15 weeks of diet (total aorta: reduction of 43%, 49%, and 36%, respectively; thoracic plus abdominal aorta: reduction of 56%, 60% and 72%, respectively), but the reduction was significant only at 10 weeks of diet for the aortic arch (40%). Combadiere et al. (14) compared the same two groups of mice fed a chow diet at 25 weeks of age. They found aortic root cross-sectional lesion area significantly reduced by 48% and thoracic (from the BCA to the renal arteries) aortic en face lesion area reduced by 55%. However, the Lesnik et al. and Combadiere et al. studies did not measure BCA cross-sectional lesion area, and the current study did not measure aortic en face lesion area. However, based on the examination of aortic root cross-sectional lesion area, the results published for the CX3CR1 knockout are at odds with those for the fractalkine knockout. In the current study, fractalkine-deficient B6.ApoE^-/- male and female mice at 16 weeks of age and male mice at 12 weeks of age had unchanged aortic-root lesion areas when fed the semisynthetic AIN76 diet containing 0.02% cholesterol. In fractalkine-deficient B6.ApoE^-/- females at 12 weeks of age, aortic-root lesion area was 30% smaller than in controls. The differences between the Lesnik et al. and Combadiere et al. studies and the current study may have several explanations. The Lesnik et al. and Combadiere et al. experiments were done on mice that had been backcrossed incompletely to the B6 background (in each case, ≈92% B6), whereas mice in the current study had been backcrossed completely. Also, the three studies used different diets. Last, it is not clear whether Lesnik used male or female mice and, although Combadiere used both male and female mice, their numbers were few and they appear to have been grouped together in their analysis. Because gender effects on aortic-root cross-sectional lesion area have been shown on the B6 background for both ApoE^-/- and LDLR^-/- mice, combining males and females may be a confounder. Thus, the issue of whether fractalkine and CX3CR1 act solely as ligand and receptor in proatherogenic pathways or whether they have at least some nonoverlapping functions remains unsettled. It would be interesting to study the CX3CR1 knockout mice on the B6.ApoE^-/- background for BCA cross-sectional lesion area.

In summary, we found that the major effect of fractalkine deficiency on atherosclerosis was at the BCA rather than the aortic root. We also found strong evidence for a specific fractalkine-background interaction. Our results differ from those reported for the fractalkine receptor, CX3CR1 knockout, but it cannot currently be determined whether this difference is due to differences in experimental conditions or some degree of nonoverlapping functions between fractalkine and its receptor.