5. RESULTS
5.1. Effects of EPA on aneurysmal tissue remodeling
Baseline mice characteristics were mostly not significantly different amongst the
experimental groups
In order to investigate the effects of EPA on murine CaCl2-induced AAAs, I designed the study with three experimental groups: (1) Sham group that received periaortic NaCl application (instead of CaCl2) and were fed the control diet, (2) Control diet group that received periaortic CaCl2 application and were fed the control diet, and (3) EPA diet group that received periaortic CaCl2 application and were fed the EPA-supplemented diet. The preparation of the diets has been described in detail in the Materials and Methods section. The experimental protocol is outlined in Figure 1.
Prior to performing study analyses, I first had to (1) confirm the baseline characteristics of mice after they have received control or EPA-supplemented diets and (2) confirm the surgical procedure for inducing AAA formation in mice with periaortic CaCl2 application.
As EPA is a fatty acid, there was a possibility that the EPA-supplemented diet could cause differences in the body weight of the mice between the Control diet and
33
EPA diet groups after AAA surgery, which may affect the interpretation of results. In addition, whether there were any differences in the amount of chow consumed between the two groups also needed to be confirmed. To this end, I assessed these baseline characteristics in the mice by recording their body weights on a weekly basis for 6 weeks as well as the amount of chow consumed daily. Interestingly, the results showed that mice fed with an EPA diet had gradual increases in body weight over 6 weeks despite having received AAA surgery, whereas the Sham and Control diet groups had an acute decrease in body weight at 1 week after AAA surgery but then recovered from Week 2 onwards to a level that was not significantly different to the EPA diet group (Figure 2A).
Consistent with the body weight data, there was also no significant difference in the mean daily amount of chow consumed between the two experimental groups (Figure 2B).
34
Figure 1. Study protocol. Mice in the Sham, Control diet, and EPA diet groups began receiving the indicated study diets 4 days prior to AAA surgery. Sham group received sham surgery with NaCl periaortic application, and served as a baseline group for future comparisons and analyses. Control diet and EPA diet groups received CaCl2 periaortic application to induce AAA formation. The respective study diets were continued for the duration of the study. Mice were kept for a maximum of 6 weeks after surgery until sacrifice for analysis.
35
Figure 2. Baseline effects of EPA on mice in the CaCl2-induced AAA model.
A. Body weights of mice in the Sham, Control diet, and EPA diet groups were recorded weekly after AAA surgery (NaCl or CaCl2 application) was performed, as indicated by the red arrow. Sham and Control diet groups received control-diet while EPA diet group received an EPA-supplemented diet. *P < 0.05, EPA diet group versus the Control diet group. B. Mean daily amounts of chow consumed by mice at baseline (prior to AAA surgery) in the Control diet and EPA diet groups. No significant differences were detected.
A
B
36
EPA treatment attenuates CaCl2-induced AAA formation and elastic lamina
destruction
Next, I investigated the effects of EPA on AAA formation. Marked dilatation and calcification of the aorta in the Control diet group was clearly visible macroscopically 6-weeks after CaCl2 was applied to the infra-renal abdominal aorta; in contrast, the aortas of the mice on the EPA-supplemented diet were dilated significantly less than those of mice in the Control diet group (Figure 3A). The aortic diameters in the Control diet group were shown to have increased to approximately 1.6 times that of the aortic diameter of Sham group mice, therefore meeting the definition for aneurysm formation (≥1.5 time increase in aortic diameter [34]). In contrast, the diameter of aortas in the EPA group was only increased by approximately 1.3 times, and furthermore this increase was not statistically significant. This therefore indicates that EPA treatment attenuated the formation of CaCl2-induced AAA (Figure 3B).
In order to assess the condition of the elastic fibers in the aortic wall, I performed histological staining using the Elastic van Gieson (EVG) stain. EVG staining is a well-established method for the visualization of arterial wall elastic lamina, and is one of the most commonly used stains in the assessment of AAA histology. Histological examination of EVG-stained AAAs demonstrated that the extensive matrix and elastic
37
lamina destruction seen in Control diet group AAAs was greatly suppressed in aortas from the EPA diet group (Figure 4). Higher magnification views showed that elastic lamina strand breaks, a hall-mark feature of AAAs, are clearly seen in AAAs of the Control diet group but were relatively absent in the EPA diet group. Taken together, these results support the notion that EPA attenuates aortic dilatation via the suppression of elastic lamina degradation, leading to the attenuation of vascular wall tissue remodeling.
38
Figure 3. EPA attenuated aortic dilatation after CaCl2-induced AAA surgery.
A. Macroscopic appearances of in situ infra-renal aortas (demarcated by the black broken lines) at 6-weeks after AAA surgery, showing a much less dilated infra-renal aorta in mice that received an EPA-supplemented diet compared to the Control diet
Sham Control diet EPA diet
+AAA A
B
39
group. Mice in the Sham group received sham AAA surgery (periaortic application of NaCl) and served as a baseline for calculations of fold-change in aortic diameter for the other two groups. Representative images of at least three independent experiments are shown. B. Quantitative analysis of the maximal external aortic diameters of aortas at 6-weeks after AAA surgery. *P < 0.05, NS, non-significant.
40
Figure 4. Administration of EPA preserves vascular wall structure in AAAs.
Histological analysis by Elastica van Gieson staining, showing preserved aortic wall structure and less elastic lamina strand breaks in the aorta of mice from the EPA diet group compared to the Control diet group. Scale bars: 200 μm (upper panels) and 50 μm (lower panels). Representative images of at least three independent experiments are shown.
41
EPA attenuated the CaCl2-induced up-regulation of MMPs but did not affect the
expression levels of TIMPs or other extracellular matrix components
Given this phenotype, I subsequently began to elucidate the molecular mechanism underlying how EPA suppressed AAA formation. I first focused on examining the mRNA expression of a set of genes related to tissue remodeling such as MMPs and TIMPs. Among the genes analyzed by real-time PCR, the expression levels of Mmp2 and Mmp9 were significantly increased in the aortas of mice in the Control diet group at 1- and 3-weeks after CaCl2 application, consistent with previous reports [19,21,25]. In contrast, mice in the EPA diet group had significantly lower levels of Mmp2 and Mmp9 expression (Table 2). This suggests that because of the lower
expression levels of the MMPs critical to AAA formation, Mmp2 and Mmp9, the tissue milieu in AAAs of mice in the EPA diet group may have been less proteolytic compared to that of the Control diet group, thereby leading to less tissue destruction.
However, considering that the balance between proteolysis and anti-proteolysis is determined by the levels of MMPs versus TIMPs, the levels of TIMPs also needed to be evaluated so as to conclude that EPA does indeed reduce the proteolytic environment of AAAs via the suppression of MMP up-regulation. To this end, the expression levels of the major TIMPs, Timp1 and Timp2, was also assessed by real-time PCR. The results
42
showed that while the levels of TIMPs were also upregulated by the CaCl2 treatment, EPA did not affect their expressions (Table 2).
Major changes in the components of the aortic wall extracellular matrix (ECM) occur as a result of the significant tissue remodeling that is invariably associated with AAA development. Therefore, I investigated whether or not EPA also had some effects on these ECM components. Since collagen I, III, and fibronectin are known to be major constituents of the aortic wall ECM [3,53], the expression levels of these factors in CaCl2-induced AAAs at 1- and 3-weeks after surgery were assessed by real-time PCR.
The results showed that while the expression of all three ECM components did indeed increase during AAA development compared to the Sham group, there was no significant difference between the Control diet and EPA diet groups (Table 2). These results indicate that in terms of tissue remodeling, the effects of EPA in attenuating AAA formation is most likely exerted through its suppression of MMP up-regulation rather than modulation of ECM components, resulting in a less proteolytic AAA tissue environment and less vascular wall degradation.
43
Table 2. Gene expression profile in 1-week AAAs analyzed by quantitative real-time PCR.
Gene
1-week 3-weeks
Sham (n=5)
Control diet (n=11)
EPA diet (n=10)
Sham (n=6)
Control diet (n=11)
EPA diet (n=13)
Ccl2 1.0±0.51 8.6±1.06 17.8±3.69 #* 1.2±0.14 4.2±0.32 # 5.8±0.72 # Col1a1 1.0±0.10 1.5±0.07 # 1.3±0.11 0.7±0.09 2.0±0.14 # 1.8±0.10 # Col3a1 1.0±0.11 2.0±0.11 # 1.7±0.15 # 0.8±0.08 2.1±0.15 # 2.2±0.10 # Fn1 1.0±0.06 8.4±0.96 # 10.1±0.98 # 1.0±0.13 2.1±0.27 # 2.9±0.70 Mmp2 1.0±0.11 1.8±0.07 # 1.3±0.10 * 1.6±0.18 4.2±0.27 # 3.5±0.17 #*
Mmp9 1.0±0.30 36.3±8.84 # 10.6±2.11 #* 2.3±0.55 10.0±1.31 # 4.6±0.47 *
Timp1 1.0±0.14 9.1±0.98 # 8.5±0.98 # 0.8±0.12 4.2±0.38 # 3.4±0.44 # Timp2 1.0±0.04 0.9±0.04 0.9±0.03 1.2±0.10 1.6±0.07 # 1.8±0.07 # Opg 1.0±0.05 1.0±0.12 1.5±0.11 #* 1.2±0.16 1.6±0.14 2.2±0.33 # Rankl 1.0±0.41 29.2±4.38 # 14.1±2.64 * 2.5±0.55 14.2±1.12 # 7.6±0.96 #*
Runx2 1.0±0.07 4.3±0.29 # 3.9±0.41 # 1.0±0.18 5.8±0.43 # 4.5±0.43 #
Messenger RNA levels of major MMPs associated with AAA formation, ECM components, and vascular calcification factors in infra-renal aortas at 1- and 3-weeks after AAA surgery were analyzed using real-time PCR. All expression levels were first normalized to 18s rRNA levels (house-keeping gene) and then presented as fold change over the Sham group value at 1-week. Results are mean ± SEM. #P < 0.05 vs. Sham group of the same time-point; *P < 0.05 vs. Control diet group of the same time-point (further indicated in bold-type), one-way ANOVA with Tukey’s post-hoc test.
.
44
In order to confirm the decrease in MMP levels in AAAs by another method, I performed zymography using 1-week AAA samples. Gelatin zymography is a well-established technique to detect the functional activity of MMPs present in tissues or cells [50]. To detect the activities of MMP-2 and MMP-9, the two MMPs that appeared to be affected by EPA based on the real-time PCR results (Table 2), a gelatin gel was chosen because gelatin is a substrate that can be degraded by these two MMPs.
Consistent with the results of real-time PCR, zymography showed that the functional activities of pro-MMP-2, cleaved MMP-2, and MMP-9 were indeed all markedly decreased in the EPA diet group compared to the Control diet group, suggesting that the reduced mRNA levels of Mmp2 and Mmp9 translated to a significant difference in their functional activities at the protein level in the AAA tissues as well (Figure 5A, B).
45
Figure 5. Reduced functional activities of MMP-2 and MMP-9 in AAAs after EPA treatment. A. Representative gelatin zymography gel showing reduced activities of the proform of MMP-2 (pro-MMP2), cleaved form of MMP-2 (cleaved-MMP2), and MMP-9 in 1-week AAA samples after EPA-feeding compared to the AAAs of the Control diet group. Each lane represents a separate AAA sample within the same treatment group. Equal amount of protein (20 μg) was loaded per AAA sample. kDa, kilodalton. B. Quantitative analysis of zymographic MMP activities. Data are mean ± SEM of three independent experiments. *P < 0.05 versus Control diet group.
A
B
46
EPA suppresses Mmp9 expression in AAA macrophages
Previous reports have demonstrated that Mmp9-deficient mice are resistant to experimental AAA formation [16,25]. Given that EPA seemed to impart a greater effect on Mmp9 expression than on Mmp2, I decided to analyze specifically how EPA suppresses MMP-9 activity in AAAs. Macrophages have been reported to be the major producer of MMP-9 in AAA tissues [22,25,54]. Therefore, there were at least two possible mechanisms by which EPA could have suppressed MMP-9 levels in the AAA:
(1) reducing the number of macrophages recruited to the AAAs, and (2) suppressing the ability of macrophages to produce MMP-9.
Flow cytometry was used to test these two hypotheses. Since the number of macrophages recruited to the AAA could be affected by the number of available circulating monocytes as well as the level of chemoattractant cytokine, i.e. CCL2, expressed by the AAA to recruit monocytes, I proceeded to investigate the number of circulating monocytes, the number of macrophages in the AAAs, and the expression level of Ccl2 in the AAAs of the Control diet and EPA diet groups. However, before performing these analyses, the gating strategy for isolating macrophages needed to be confirmed from a technical perspective. The cell surface markers of circulating monocytes are well established and can be easily identified by gating for Ly-6G- cells
47
(to exclude granulocytes such as neutrophils) that are CD11b+ and Ly-6Chi in the peripheral blood (Figure 6) [29,55]. Meanwhile in the AAA tissue, by first gating for Ly-6G- cells from the dissociated AAA cells, macrophages could subsequently be identified as Ly-6ClowCD11b+F4/80+ cells (Figure 7A, B) [32]. These macrophages were isolated by fluorescence-activated cell sorting (FACS), and their cellular appearance was shown to be consistent with that of typical macrophages (Figure 7C).
48
Figure 6. Gating strategy for the flow cytometric analysis of peripheral circulating monocytes. Representative flow cytometric plots of peripheral blood analysis are shown. Living cells isolated from the peripheral blood of mice at 1-week after AAA surgery were first gated on Ly-6G (granulocyte marker), and Ly-6G- cells were further analyzed for expression of the myeloid markers Ly-6C and CD11b. Ly-6ChiCD11b+ cells were taken to be monocytes according to previous reports [29,55].
49
Figure 7. Gating strategy for the flow cytometric analysis of AAA macrophages.
Representative flow cytometric plots of AAA analysis are shown. Similar to the gating strategy for the peripheral blood analysis, living cells isolated from AAA tissues 1-week after the AAA surgery were first gated on Ly-6G, and Ly-6G- cells were further analyzed for expression of Ly-6C and CD11b (A); Ly-6ClowCD11b+ cells were shown to be positive for F4/80, a macrophage marker (B), and together Ly-6ClowCD11b+F4/80+ cells
were taken to be aneurysmal macrophages and used in all subsequent analyses.
C. Giemsa staining of sorted Ly-6ClowCD11b+F4/80+ cells from the aorta shows cells with the characteristic macrophage appearance. Scale bar, 10 μm.
A
B C
50
Using these gating strategies to test my first hypothesis, I proceeded to assess the number of monocytes in peripheral blood and the number of macrophages in AAAs at 1-week after the surgery in the Control diet and EPA diet groups. Surprisingly, while the difference in the number of circulating Ly-6ChiCD11b+ monocytes was not statistically significant, mice treated with EPA tended to have higher numbers of circulating monocytes (Figure 8). In addition, when I examined the tissue mRNA expression levels of Ccl2 in the AAAs at the same time-point of 1-week after AAA surgery, aortas of mice in the EPA diet group had significantly higher Ccl2 expression than the aortas of mice in the Control diet group (Table 2). These two results together should have suggested more potent recruitment of circulating monocytes to AAAs by CCL2 in the EPA diet group, leading to the presence of more macrophages in the AAA and which would be completely contrary to the initial hypothesis. However, upon examining the actual number of macrophages in the AAA amongst the Control diet and EPA diet groups, there was interestingly no statistically significant difference in the number of aneurysmal Ly-6ClowCD11b+F4/80+ macrophages between the Control diet and EPA diet groups (Figure 9). This indicates that despite the higher potential for monocyte recruitment to AAAs in the EPA diet group, EPA suppressed the actual recruitment and/or infiltration of monocytes.
51
To test the second hypothesis, I sorted the AAA macrophages from both groups at 1-week after AAA surgery and examined their Mmp9 mRNA expression by real-time PCR. The results showed that there was significantly less Mmp9 expressed by macrophages sorted from the AAAs of mice in the EPA diet group (Figure 10), suggesting that EPA directly affected macrophage function, such as MMP production, within the AAA tissue while AAA macrophage numbers were unaffected. The combination of no difference in macrophage numbers and an absolute decrease in macrophage-derived MMP-9 levels in the AAAs of the EPA diet group resulted in a net fall in total MMP-9 levels and activity, thereby helping to explain the reduced MMP-9 activity and gene expression in whole AAA samples as well as the attenuation of AAA formation.
52
Figure 8. EPA tends to increase circulating monocyte numbers at 1-week after AAA surgery. Representative flow cytometric plots of circulating monocytes in peripheral blood of mice in the Control diet and EPA diet groups at 1-week after the AAA surgery, gated by the myeloid cell markers Ly-6C and CD11b. After the blood was sampled via cardiac puncture, the mice were sacrificed and aortas were harvested for subsequent flow cytometric and mRNA expression analyses. While the difference in the number of CD11b+Ly-6Chi circulating monocytes was not statistically significant between the two groups (P = 0.1213 by unpaired Student’s t-test), EPA tended to increase the number of monocytes.
53
Figure 9. AAA macrophage numbers are not significantly different between the Control diet and EPA diet groups. Representative flow cytometric plots of AAA macrophages from mice in the Control diet and EPA diet groups at 1-week after the AAA surgery, gated by the myeloid cell markers Ly-6C and CD11b. Three AAAs from each group were pooled into one sample in each experiment. Quantifying the mean number of Ly-6ClowCD11b+F4/80+ aneurysmal macrophages per AAA sample showed that there was no statistically significant difference between the two groups. Data are mean ± SEM of six independent experiments.
54
Figure 10. mRNA levels of Mmp9 in sorted AAA macrophages. Macrophages from AAAs in the Control diet and EPA diet groups at 1-week after AAA surgery were sorted and their Mmp9 expression levels were analyzed by real-time PCR. Expression levels were first normalized to 18s rRNA levels and then expressed as the relative expression to the level of Control diet group. Three AAAs from each group were pooled into one sample in each experiment. Data are mean ± SEM of five independent experiments. *P
< 0.05 compared to the Control diet group.
55
Macrophage expression of Mmp9 is directly suppressed by EPA
Given the in vivo results, the next question was whether EPA would have the same effect in vitro and suppress Mmp9 expression in cultured macrophages. To address this possibility, I used the well-established RAW264.7 macrophage cell line and treated these cells with EPA in vitro with or without TNF-α stimulation to induce Mmp9 expression.
Firstly, the results showed that TNF-α effectively induced Mmp9 expression in RAW264.7 macrophages, as can be seen by the more than two-fold increase in Mmp9 expression in the vehicle control group after TNF-α stimulation (Figure 11). When RAW264.7 macrophages were treated with increasing concentrations of EPA, there was a dose-dependent reduction in the expression of Mmp9 at both baseline and after TNF-α stimulation, although the effects at baseline were small and non-significant (Figure 11).
This lends strong support to the direct effects of EPA on macrophage Mmp9 expression.
Taken together with the previous in vivo results, it appears that EPA directly affects macrophages to reduce their Mmp9 expression in both in vivo and in vitro conditions.
56
Figure 11. EPA dose-dependently suppresses macrophage Mmp9 expression in vitro.
RAW264.7 macrophages were cultured with either vehicle (10% BSA) or EPA (10, 25, or 50 μmol/L) for 48 hours. The cells were then stimulated with TNF-α (20 ng/mL) for a further 6 hours and harvested for analysis by real-time PCR. Expression levels were normalized to 18s rRNA levels. n=3 per condition. Results are mean ± SEM. *P < 0.05 compared to the vehicle control after TNF-α stimulation.
57