3.1. miRNAs may be involved in high fat induced endothelial inflammation.
Compared with normal chow diet, rats fed with high fat diet for 10 weeks showed significant increase in body weight. Consistent with the body weight gain, significant increase in both subcutaneous fat and visceral fat mass was observed in HF group. The fasting glucose was also impaired when rats challenged with high fat diet. Additionally, the lipid profile in serum was also determined. As expected, higher levels of TC, TG and LDL-C as well as lower HDL-C were observed in HF group (Supplemental Fig 1).
As we all known, fat deposition was always accompanied by chronic low-grade inflammation. We assessed inflammatory factors in aortas of rats, including NF-κB p65, monocyte chemotactic protein 1 (MCP-1), intercellular cell adhesion molecule-1 (ICAM-1) and interleukin 6 (IL-6). These inflammatory factors significantly increased in aortas of HFD rats (Fig 1A).The phosphorylation of NF-κB p65, an important indicator for inflammation, was markedly elevated in aortas of HFD group (Fig 1B). At the same time, the expressions of inflammatory factors and phosphorylation of NF-κB p65 were also up-regulated in HAECs treated with PA (Fig 1C, 1D).
Next, we examined the expressions of miRNAs related with inflammation in vivo and in vitro. Compared with NC group, the relative expressions of miR-146a, miR-155, and miR-21 were decreased in aortas of HF group (Fig 1E). Previously, our group demonstrated that miR-146a could directly target IRAK1 and TRAF6 by luciferase assay [9]. NF-κB p65 have been reported to be the direct targets of miR-155 [10]. As shown in Fig 1, the transcriptional expressions of IRAK1,TRAF6 and NF-κB p65 were significantly increased when exposure to HFD in aortas of rats (Figure 1F). Meanwhile, the protein levels of IRAK1, TRAF6 and NF-KB p65 were also up-regulated in HF group (Figure 1G). In HAECs, exposure to PA significantly reduced the expressions of miR-146a and miR-15 but not miR-21 in HAECs (Figure 1H), which indicates that miR-21 may not be implicated in the inflammation induced by PA in HAECs. Additionally, the mRNA and protein levels of IRAK1, TRAF6 and NF-κB p65 were both elevated in HAECs treated by PA (Figure 1I, 1J). Therefore, we speculate that miRNAs may be involved in endothelial inflammation induced by high fat.
3.2. The endothelial inflammation induced by high fat could be improved by regulating miRNA-146a and miRNA-155.
To further investigate whether high fat induced endothelial inflammation was mediated by miRNAs, we transfected mimics and inhibitors of miR-146a and miR-155 into HAECs. As shown in Figure 2, miR-146a and miR-155 mimics could decrease the up-regulation of inflammatory factors (NF-κB p65, IL-6, MCP-1 and ICAM-1) induced by PA in HAECs (Fig 2A, 2C). On the contrary, miR-146a inhibitors and miR-155 inhibitors exacerbated the inflammation in HAECs treated with PA (Fig 2B, 2D). Regulating miR-21 have no effects on the inflammation in HAECs (Supplemental Fig 2). These results revealed that miR-146a and miR-155 mediated endothelial inflammation induced by high fat and regulating miRNAs can reverse the endothelial inflammation.
3.3. AE and metformin prevent high fat-induced fat accumulation, dyslipidemia and endothelial inflammation
To explore the effects of AE and metformin on high fat induced endothelial inflammation both in vivo and in vitro, we fed high fat diet to Sprague Dawley rats for 10 weeks and then treated them with AE/metformin orally for 8 weeks. As shown in Figure 3A, HFD induced a significant increase in body weight. After treatment with AE and metformin for 8 weeks, the body weight in HF-Met group was decreased compared with HF group (Figure 3A). AE treatment led to a slight decrease of body weight compared with HF group without significant difference (Figure 3A). Meanwhile, AE and metformin reduced both subcutaneous fat and visceral fat contents (Figure 3B, 3C). AE and metformin could improve the impaired fasting glucose of rats that challenged with high fat diet (Figure 3D). Additionally, the lipid profile in serum was determined. As expected, higher levels of TC, TG and LDL-C as well as lower HDL-C were observed in HF group, AE and metformin significantly improve the lipid profile (Figure 3E, 3F, 3G and 3H). The low dose AE (HF-OB) exhibited more efficient in lowering LDL-C level than high dose AE (HF-OA) (Figure 3G). The effects of different doses of AE and metformin on metabolic parameters were also listed in Table 1.
Besides, treatment with AE and metformin could significantly decrease the inflammatory factors (NF-κB p65, MCP-1, ICAM-1, IL-6) (Figure 3I). The results revealed that metformin could reverse inflammatory factors to the levels of NC group, and moderate-dose AE (HF-OB group) was more effective to attenuate inflammation than high-dose AE (HF-OA group). As shown in Figure 3, the phosphorylation of NF-κB p65 was markedly elevated in aortas of HFD group, while AE and metformin could reverse it and metformin was more efficient (Figure 3J). Furthermore, we treated HAECs with PA for 19 hours to investigate the effects of AE and metformin on inflammation in vitro. We compared the effects of different concentrations of AE and metformin on inflammatory response in HAECs. The 150μg/ml AE and 2mmol/l metformin could reverse the inflammatory factors most effectively without influence on the status of HAECs, so we choose 150μg/ml AE and 2mM metformin in the following experiments (Supplemental Figure 3). Consistent with the results in rats, the inflammatory factors and the phosphorylation of NF-κB p65 were significantly up-regulated by PA compared with NC group. In comparison, AE and metformin obviously inhibited the PA-induced inflammatory response in HAECs (Figure 3K, 3L).
Therefore, AE and metformin could prevent high fat-induced fat accumulation, dyslipidemia and endothelial inflammation.
3.4. AE and metformin regulate miR-146a and miR-155 as well as their target genes in high fat treated endothelium
To determine whether the inhibitory effects of AE and metformin on inflammation is depended on the regulation of miRNAs, we examined the expressions of miRNAs related with inflammation both in vivo and in vitro. Compared with NC group, the relative expressions of miR-146a, miR-155, and miR-21 were decreased in aortas of HF group. Importantly, AE and metformin treatments resulted in increases in the expressions of miR-146a and miR-155 in HF-OB group and HF-Met group. However, neither AE nor metformin exhibited effects on the expression of miR-21 (Figure 4A). Moreover, the mRNA and protein levels of the target genes (IRAK1,TRAF6 and NF-κB p65) were significantly decreased when rats treated with HFD were orally administrated with AE or metformin compared with HF group (Figure 4B, 4C). Moreover, it showed that moderate-dose AE (HF-OB group) is more effective than high-dose AE (HF-OA group) in increasing the expressions of miR-146a and miR-155 as well as decreasing their target genes, which suggested that the concentration of AE may be involved in its effect on miRNAs. And there were no obvious differences between HF-OB group and HF-Met group.
Treatment with AE and metformin could reverse the expressions of miR-146a and miR-155 but not miR-21 to the normal level in PA-stimulated HAECs (Figure 4D). Additionally, the mRNA and protein expressions of IRAK1, TRAF6 and NF-κB p65 were up-regulated when exposure HAECs to PA, which showed decreases after treated with AE or metformin (Figure 4E, 4F). Thus, we assume that AE and metformin may improve the inflammatory response in endothelium via up-regulating miR-146a and miR-155 as well as inhibiting their target genes.
3.5. The inflammation-inhibitory roles of AE and metformin are depended on miR-146a in endothelial cells.
miR-146a was found to be down-regulated by high fat, while AE and metformin could modulate its expression both in vivo and in vitro. To examine whether miR-146a mediates the protective role of AE and metformin, the miR-146a mimics and inhibitors were transfected into HAECs. As shown in Supplemental Fig 4A, miR-146a mimics resulted in a significant increase in the expression of miR-146a compared with control mimics. The protein levels of IRAK1 and TRAF6, as the target genes of miR-146a, were both repressed by miR-146a mimics compared with control mimics in HAECs (Figure 5B). The expressions of inflammatory factors (NF-κB p65, IL-6, MCP-1 and ICAM-1) and the phosphorylation level of NF-κB p65 were reduced after miR-146a mimics transfection into HAECs (Figure 5A, 5B). In addition, the inflammatory factors and target genes were further inhibited by miR-146a mimics in PA-AE and PA-Met groups compared with control mimics transfected group (Figure 5A, 5B).
By comparison, the expression of miR-146a were effectively reduced in miR-146a inhibitors transfected group (Supplemental Fig 4B). The protein levels of IRAK1 and TRAF6 were elevated by miR-146a inhibitors when compared with control inhibitors (Figure 5D). Besides, miR-146a inhibitors induced up-regulation of inflammatory factors (NF-κB p65, IL-6, MCP-1 and ICAM-1) and the phosphorylation level of NF-κB p65 compared with corresponding groups transfected with control inhibitors (Figure 5C, 5D). Furthermore, inhibition of miR-146a deteriorated the endothelial inflammation in PA-treated group compared with control inhibitors. Meanwhile, AE and metformin mediated improved inflammation were weakened by miR-146a inhibitors transfection (Figure 5C, 5D). These results suggest that miR-146a participate in high fat-induced inflammation in endothelium through directly regulating IRAK1 and TRAF6, and miR-146a may be indispensable for AE and metformin mediated the improvement of inflammation.
3.6. The inflammation-inhibitory roles of AE and metformin are depended on miR-155 in endothelial cells.
To further investigate whether the inhibitory effects of AE and metformin on high fat-induced inflammation in aortas and HAECs is depended on its regulation on miR-155, we transfected miR-155 mimics or inhibitors into HAECs (Supplemental Figure 4C and 4D). As the target gene of miR-155, the mRNA and protein levels of NF-κB p65 were down-regulated by miR-155 mimics (Figure 6A, 6B). Meanwhile, the miR-155 mimics reduced the mRNA levels of inflammatory factors (IL-6, MCP-1 and ICAM-1) and the phosphorylation level of NF-κB p65 when compared with control mimics in HAECs (Figure 6A, 6B). In addition, miR-155 mimics further suppressed the inflammatory factors and target genes in PA-AE and PA-Met group compared with control mimics (Figure 6A, 6B).
In comparison, the mRNA and protein levels of NF-κB p65 were up-regulated by miR-155 inhibitors (Figure 6C, 6D). The expressions of inflammatory factors (IL-6, MCP-1 and ICAM-1) and the phosphorylation level of NF-κB p65 were increased by miR-155 inhibitors compared with control inhibitors (Figure 6C, 6D). Additionally, the elevated inflammatory factors and target genes induced by PA were exacerbated by miR-155 inhibitors in PA group, and miR-155 inhibitors abolished the benefit effects of AE and metformin on inflammation in PA-AE and PA-Met group (Figure 6C, 5D). Taken together, these results indicate that miR-155 plays an important role in inflammation activation induced by high fat via the direct regulation on NF-κB p65, AE and metformin could protect against high fat-induced inflammation though regulation of miR-155 in endothelium.
3.7. The inflammation-inhibitory roles of AE and metformin is independent of miR-21 in endothelial cells.
In this study, the expression of miR-21 was decreased in aortas from HF group, while it showed no difference in PA-exposed HAECs. In addition, neither AE nor metformin exhibited effect on the expression of miR-21 in aortas of HF rats and PA-stimulated HAECs, indicating that miR-21 may not be involved in high fat-induced endothelial inflammation. To validate this hypothesis, we transfected miR-21 mimics and inhibitors into HAECs (Supplemental Figure 4E, 4F). The results showed that inflammatory factors were activated in PA group as well as the phosphorylation of NF-κB p65. However, neither miR-21 mimics nor inhibitors exhibited effect on inflammatory factors and the phosphorylation level of NF-κB p65 compared with control mimics or inhibitors in HAECs (Figure 7). In brief, the results suggest that miR-21 may not participate in high fat-induced inflammation in HAECs. AE and metformin mediated inflammatory improvement is not dependent on miR-21.