Monocyte Chemoattractant Protein-1 Expression Is Enhanced by Granulocyte-Macrophage Colony-stimulating Factor via Jak2-Stat5 Signaling and Inhibited by Atorvastatin in Human Monocytic U937 Cells*

The proinflammatory cytokine granulocyte-macrophage colony-stimulating factor (GM-CSF) is expressed in inflammatory and atherosclerotic lesions. GM-CSF is known to enhance monocytic expression of monocyte chemoattractant protein-1 (MCP-1). However, the molecular mechanism(s) by which GM-CSF up-regulates the MCP-1 expression remains to be clarified. Thus, in this study, we examined our hypothesis that GM-CSF up-regulates the MCP-1 expression via Jak2-Stat5 signaling pathway. In human monocytic cell line U937, GM-CSF increased MCP-1 expression in protein and mRNA levels. Furthermore, analysis of the GM-CSF promoter element revealed that the STAT5 (signal transducer and activator of transcription-5) transcription factor binding site, located between –152 and –144 upstream of the transcription start site, as well as Janus kinase-2-mediated Stat5 activation were necessary for the GM-CSF-induced transcriptional up-regulation of the MCP-1 gene. This GM-CSF-induced MCP-1 expression, measured as both protein and mRNA levels, was down-regulated by atorvastatin, a 3-hydroxy-3-methylglutaryl-CoA reductase inhibitor. However, this decrease in MCP-1 expression was not at the transcriptional level of MCP-1 gene but rather at the level of the stability of MCP-1 mRNA. These results indicate that GM-CSF regulates MCP-1 expression via Janus kinase-2-Stat5 pathway and by a novel regulatory mechanism of statins to reduce inflammatory reactions by down-regulating the expression of monocytic MCP-1, which promotes atherogenesis.

Inflammatory events are involved in the pathogenesis of atherosclerosis (1,2), and T-lymphocytes and monocytes/macrophages are abundant in the atherosclerotic lesion (3). Monocytes/macrophages are present in all stages of atherosclerosis, playing a central role in atherogenesis; their multiple functions include migration and production of growth factors, cytokines, and matrix-degrading enzymes as well as uptake of modified lipoproteins (4). Specifically, the adhesion and migration of circulating monocytes to endothelial cells and the subendothelial microenvironment are important for the initiation of atherogenesis (5).
An association between matrix turnover and migration of many types of cells may be involved in atherogenesis, and the production of matrix-degrading enzymes, such as matrix metalloproteinase (MMP), 2 from monocytes is one of the key events to enhance the migratory action of the monocytes (6). Granulocyte-macrophage colony-stimulating factor (GM-CSF) has been reported to enhance MMP-1, MMP-9, and MMP-12 expression as well as stimulate the activation of RhoA and integrin clustering in monocytes (7)(8)(9). Thus, GM-CSF is thought to play pivotal roles in various monocyte functions, including migration in atherogenesis (7,9). On the other hand, studies of blockage of the monocyte chemoattractant protein-1 (MCP-1) signaling pathway by MCP-1 knock-out and MCP-1 receptor knock-out have provided direct evidence of a critical role for MCP-1 and its receptor in monocyte migration during atherogenesis (10 -12). The atherogenic effect of MCP-1 is mainly explained by its potent chemoattractive effect on monocytes. Since our previous study demonstrated that monocytic MCP-1 expression is up-regulated by GM-CSF, we speculated that there was a synergistic effect of MCP-1 and GM-CSF on monocytic migration (9). However, the molecular mechanism(s) by which GM-CSF enhances MCP-1 expression in monocytes is not clearly understood.
There is growing evidence that some beneficial effects of 3-hydroxy-3-methylglutaryl-CoA reductase inhibitors (statins) are independent of their effects on lowering lipid levels. Statins have pleiotropic effects, such as improvement of endothelial dysfunction, increased nitric oxide bioavailability, anti-inflammatory activity, and antioxidant properties (13,14). In particu-lar, anti-inflammatory effects would be beneficial for prevention of atherosclerosis. In fact, atorvastatin therapy reduces blood levels of proinflammatory cytokines, such as tumor necrosis factor-␣, interleukin (IL)-1, IL-6, and MCP-1 in hypercholesterolemic patients and in those with acute coronary syndrome (15,16).
The aim of the present study was to investigate how GM-CSF up-regulates the expression of MCP-1 in human monocytic U937 cells. Furthermore, we investigated whether atorvastatin would inhibit GM-CSF-induced MCP-1 expression as a potentially novel anti-inflammatory effect.
Cell Culture-Human monocytic U937 cells were obtained from the American Type Culture Collection (ATCC CRL-1593) and maintained in RPMI1640 medium (Invitrogen) with 5% fetal calf serum (ICN, Costa Mesa, CA).
MCP-1 in U937 Conditioned Medium-The U937 cells were seeded into 24-well plates at a density of 1 ϫ 10 6 cells/1 ml of RPMI1640 medium containing 0.5% fetal calf serum and were then stimulated with 1-10 ng/ml of GM-CSF for 24 h. MCP-1 in the culture supernatant was subsequently measured by an enzyme-linked immunosorbent assay according to the manufacturer's instructions (R&D, Minneapolis, MN).
Northern Blot Analysis-U937 cells were incubated with GM-CSF (1-50 ng/ml) in the RPMI1640 medium with 0.5% fetal calf serum for 16 h. Then total RNAs were extracted by using Trizol reagent (Invitrogen) and fractionated on 1% agarose gel containing 6% formaldehyde (10 g of total RNA/lane). After the RNAs were transferred onto a nylon membrane, the expression of MCP-1 mRNA was detected by hybridization with 32 P-labeled MCP-1 probe, as previously described (9). Equal loading of RNA was confirmed by detecting ␤-actin as an internal control. The expression levels were measured by density analysis using the imagel program, free software available at the National Institutes of Health Web site.
Plasmid Construction-We cloned the human MCP-1 promoter region from the genomic DNA of U937 cells by PCR using the primer pairs 5Ј-GCTGGAGGCGAGAGTGC-GAG-3Ј and 5Ј-TCTAGATTCTCCTTTAGCTGT-3Ј, corresponding to nucleotide positions Ϫ932 to ϩ60 relative to the transcription start site. The PCR products were cloned into pGL 3 -basic vector (Promega, Madison, WI), and the sequence was confirmed. The 5Ј-serial and internal deletion mutants were generated by appropriate enzymatic digestion and ligation. The point mutations in the activator protein-1 distal site (AP-1d at Ϫ97 to Ϫ91) and the AP-1-proximal site (AP-1p at Ϫ69 to Ϫ63), NF-B (at Ϫ89 to Ϫ80), and signal transducer and activator of transcription 5 (STAT5p; at Ϫ152 to Ϫ144; see the map in Fig. 2A) were introduced by site-directed mutagenesis as follows: TCACTCA to caACTCA for AP-1d, TGACTCC to caACTCC for AP-1p, GGAAGATCCC to GGAAGATggg for NF-B, and TTCCTGGAA to TTCCTGGtt for STAT5p. Jak2 (Janus kinase-2)-and Stat5a-expressing vectors were used for co-transfection studies (17).
To investigate the effects of the 3Ј-untranslated region (3Ј-UTR) of the MCP-1 gene (ϩ361 to ϩ732) on mRNA stability, the 3Ј-UTR was cloned into the MCP-1 luciferase construct (Ϫ538 construct in Fig. 2A) at the end of the luciferase gene using the XbaI site. In this construct, thus, the luciferase gene is driven by the MCP-1 promoter, and the mRNA stability is regulated by the 3Ј-UTR. The AUUUA sequence in the 3Ј-UTR located at ϩ496 to ϩ500 was deleted by SpeI/AflII digestion, followed by fill-in reaction by Klenow and ligation. This clone had the deletion from ϩ477 to ϩ538.
Transfection and Luciferase Assay-Luciferase assays were performed as previously described (18). Briefly, U937 cells (1.4 ϫ 10 7 /cuvette) were transfected with 20 g of reporter constructs along with 0.5 g of ␤-galactosidase-expressing plasmid by electroporation (Gene Pulser II; Bio-Rad). After electroporation, the cells were seeded into a 24-well plate and incubated with or without GM-CSF (1-20 ng/ml) for 8 h. Then the cells were lysed, and luciferase activity was measured with a luminometer. For the co-transfection experiments, cells were electroporated with Jak2-and Stat5a-expressing plasmids (5 g each per electroporation cuvette). To minimize variations in the transfection efficiency, we transfected cells in a single batch for each reporter plasmid and then divided the desired number of transfected cells to each well. Furthermore, co-transfection of ␤-galactosidase expression plasmid followed by measurement of ␤-galactosidase activity was used as internal controls to monitor the transfection efficiency. All luciferase assays were repeated at least three times, each in triplicate wells.
Effect of Stat5a Knockdown by siRNA-The effect of Stat5a knockdown on the MCP-1 transcriptional activity was investigated by transfection of siRNA (Custom SMARTpool, Dharmacon, Amersham Biosciences). The siRNA for Stat5a knockdown was a mixture of four different synthetic siRNAs targeting the Stat5a mRNA (J-005169-11, -12, -13, and -14). For a negative control, nontargeting siRNA 5 (D-001210-05-05) was used. The U937 cells were transfected with 20 g of luciferase reporter constructs along with 0.75 g of siRNA by electroporation. The knockdown efficiency of the siRNA was monitored by the expression of Stat5a protein by Western blotting described below.
Electrophoretic Mobility Shift Assay (EMSA)-U937 cells were incubated with GM-CSF (20 ng/ml, 15 min), and nuclear extracts were prepared as previously described (19). Doublestranded oligonucleotide, including the STAT5p (Ϫ152 to Ϫ144 bp; 5Ј-CTTTCCTACTTCCTGGAAATCCACAG-3Ј) site of the MCP-1 promoter region and the mutant STAT5 (5Ј-CTTTCCTACTTCCTGGttATCCACAG-3Ј) were used. The probes were biotin-labeled at their 3Ј-end by terminal deoxynucleotidyltransferase (Pierce). Nuclear extract (5 g) and labeled probe (20 fmol) were incubated for 30 min at room temperature in a binding buffer (Lightshift Chemiluminescent EMSA kit; Pierce). The DNA-protein complexes were separated on a native 4% polyacrylamide gel and then electrophoretically transferred onto a positively charged nylon membrane. The biotin-labeled DNA was detected by a chemi-luminescence method (Lightshift Chemiluminescent EMSA kit; Pierce). For antibody supershift experiments, 2 g of antibodies specific for Stat5a and Stat5b (Chemicon) were incubated with the nuclear extract at 4°C for 1 h prior to incubation with the labeled probes.
Analysis of MCP-1 mRNA Stability-MCP-1 mRNA stability was analyzed as follows. The U937 cells stimulated with GM-CSF (10 ng/ml) for 16 h were treated with AcD (10 g/ml) in the absence or presence of Atv (5 mol/ml). The total RNA was extracted at 0, 2, 3, and 6 h after the addition of AcD and subjected to Northern blotting to evaluate mRNA degradation.

Effects of 3Ј-UTR of MCP-1 Gene on Luciferase mRNA
Stability-The MCP-1 luciferase construct including the 3Ј-UTR was transfected as described above. Afterward, the cells were incubated with GM-CSF (10 ng/ml) for 8 h and then incubated with AcD (10 g/ml) in the absence or presence of Atv (5 mol/ml). The cells were harvested at 0 and 6 h after AcD treatment, and then luciferase activity, which would indirectly reflect the stability of luciferase mRNA, was measured.
Statistical Analysis-All quantitative data were presented as the mean Ϯ S.D. Student's t test was applied, and differences at p Ͻ 0.05 were considered significant.

MCP-1 Expression by GM-CSF in the U937 Cells-GM-CSF
stimulated the U937 cells to produce and secrete MCP-1 into the culture medium (Fig. 1A). The GM-CSF-induced MCP-1 production was accompanied by a dose-dependent increase in MCP-1 mRNA expression (Fig. 1B).
Transcriptional Assay of MCP-1 Gene-The vector constructs used were the human MCP-1 promoter region up to Ϫ932 bp, including AP-1, NF-B, and STAT5 motifs, and its deletion mutants ( Fig. 2A). GM-CSF increased the MCP-1 transcriptional activity from the Ϫ932 bp construct in a dosedependent manner (1-20 ng/ml for 8 h) (Fig. 1C). When the Ϫ932 bp construct was deleted up to Ϫ538 bp, the basal activity was increased, thus indicating the presence of an unknown repressive region at Ϫ932 to Ϫ538 bp. The deletion constructs up to Ϫ415 and Ϫ292 showed decreased basal activity but still maintained the GM-CSF response. When the 5Ј-deletion construct up to Ϫ145 was used for transfection, the GM-CSF response was decreased, but the basal activity was still preserved (Fig. 2A). The Ϫ314/Ϫ292 internal deletion construct showed a similar basal activity and GM-CSF response in comparison with the parental Ϫ538 construct. In contrast, the Ϫ314/Ϫ145 and Ϫ292/Ϫ145 internal deletion constructs showed a decreased GM-CSF response (Fig. 2B). These results together indicate the pres- ence of a GM-CSF-responsive motif(s) in the region spanning from Ϫ292 to Ϫ145. Involvement of STAT5, but not AP-1 and NF-B, in MCP-1 Transcription-Since the Ϫ292 to Ϫ145 region included a putative STAT5 motif at Ϫ152/Ϫ144, which is known as one of the GM-CSF-responsive motifs, a point mutation at the STAT5 (STAT5p) motif was introduced in the Ϫ538 construct. This point mutation reduced the GM-CSF response, whereas mutations in the AP-1 or NF-B motifs still resulted in a 2-fold activation by GM-CSF (Fig. 3A).
GM-CSF activated the Jak2 and Stat5a of both constitutively expressed and exogenously overexpressed Jak2 and Stat5a which was demonstrated by the detection of phosphorylated Jak2 and Stat5 (since, however, the phosphospecific anti-Stat5 antibody recognizes both Stat5a and Stat5b proteins, the bands include both isoforms) (Fig. 3B). Co-expression of Jak2 and Stat5a with the Ϫ538 reporter construct exhibited increased transcriptional activity induced by GM-CSF (Fig. 3C). Furthermore, co-transfection of siRNA, which leads to knockdown of Stat5a but not Stat5b expression (Fig. 3D) resulted in reduced GM-CSF-induced transcriptional activity (Fig. 3E). Together with the results obtained from the luciferase assay using mutations of the STAT5 motif, these results suggest that the STAT5 located at Ϫ152/Ϫ144 (STAT5p) is a potential GM-CSF-responsive motif for Jak2-and Stat5a-mediated MCP-1 gene transcription in the U937 cells.
EMSA for GM-CSF-induced Stat5 Binding-To confirm that the MCP-1 gene expression was mediated through the Stat5 binding to the STAT5 motif, we subjected the nuclear extracts from GM-CSF-stimulated U937 cells to EMSA (Fig. 5).  in lane 9). Therefore, the supershift assay indicated that the specific complex contained, at least, the Stat5a transcriptional factor as a major component.
Atorvastatin Did Not Affect the MCP-1 Promoter Activity but Did Affect mRNA Stability-Atv treatment, however, did not inhibit the GM-CSF-induced MCP-1 transcription even at 10 mol/liter concentration when the Ϫ538 construct were transfected (Fig. 7A). Furthermore, a nuclear run-on assay showed that Atv (5 mol/ liter) did not reduce the de novo synthesis of MCP-1 mRNA (114 Ϯ 19% of GM-CSF treatment alone; not significant) (Fig.  7B). These results indicated that some post-transcriptional mechanism(s) for the reduction of MCP-1 mRNA expression is present, so we next investigated MCP-1 mRNA stability by using AcD. After a 16-h stimulation with GM-CSF, U937 cells were incubated with AcD to stop mRNA synthesis. Then total RNAs were harvested at 2, 3, and 6 h after AcD treatment to investigate the RNA stability in either the presence or absence of Atv (5 mol/liter). When the cells were incubated with Atv, the MCP-1 mRNA half-life was significantly decreased compared with that with AcD treatment alone (half-life of 11.5 versus 8.1 h; Fig. 7B), indicating that Atv increased the destabilization of MCP-1 mRNA.
Since the AUUUA Sequence in the 3Ј-UTR of the MCP-1 gene is suggested to regulate the mRNA stability of cytokines and chemokines (20), next we addressed the 3Ј-UTR effects on the stability of the luciferase gene by measuring the luciferase  FEBRUARY 22, 2008 • VOLUME 283 • NUMBER 8 activity (Fig. 8). When the 3Ј-UTR was absent, the activity was decreased to ϳ60% of initial activity after 6 h of AcD treatment (Fig. 8A). The addition of 3Ј-UTR at the end of the luciferase gene increased the activity (Fig. 8B), which was significantly decreased by the treatment with Atv. When the AUUUA sequence was deleted from the 3Ј-UTR, the activity was decreased in comparison with AUUUA(ϩ) constructs (Fig. 8B) and Atv-mediated suppression of the luciferase activity was no longer detected (Fig. 8C). However, in comparison with the 3Ј-UTR(Ϫ) construct (Fig. 8A), the AUUUA-deleted constructs still showed higher activities. These suggested that the 3Ј-UTR increased the mRNA stability in a AUUUA sequencedependent manner and that Atv decreased the stability also via a AUUUA-dependent mechanism(s), at least in part.

DISCUSSION
GM-CSF is a hematopoietic growth factor and a proinflammatory cytokine produced by inflammatory cells, including T-lymphocytes, monocytes/macrophages, and vascular endothelial and smooth muscle cells (21,22). Previously, we demonstrated that GM-CSF has multiple effects, enhancing monocytic migration via RhoA/integrin activation and via expression of MMPs and MCP-1 (7,9). In the present study, we show a molecular mechanism by which GM-CSF induced MCP-1 expression in U937 monocytic cells. GM-CSF increased MCP-1 transcription, mRNA expression, and MCP-1 production mediated through the Jak2-Stat5 signaling pathway and STAT5 motif but not via the AP-1 or NF-B pathway. Furthermore, Atv inhibited GM-CSF-induced MCP-1 expression via destabilization of MCP-1 mRNA. Therefore, the findings presented here provide evidence that GM-CSF is an important regulator of monocytic MCP-1 expression and, moreover, suggest that the inhibitory effects of Atv on MCP-1 expression, which are independent of its lipid-lowering effect, may decrease the inflammatory reaction in atherosclerotic lesions.
Among the cells constituting the vascular system, human monocytic cell lines (THP-1 and U937), human umbilical vein endothelial cells, and vascular smooth muscle cells have been used for investigating the regulation of MCP-1 expression. For example, bacterial lipopolysaccharide has been reported to stimulate THP-1 cells to enhance MCP-1 expression, whereas IL-1␤, IL-4, IL-6, IFN-␥, and tumor necrosis factor-␣ were shown to increase MCP-1 expression in U937 cells (23,24). It has also been reported that vascular endothelial growth factor and IL-1␤ enhance MCP-1 expression in human umbilical vein endothelial cells (25,26) and that tumor necrosis factor-␣ increases MCP-1 in rat vascular smooth muscle cells (27). Some of these studies have described the molecular mechanisms involved in the MCP-1 expression. Namely, lipopolysaccharide-induced MCP-1 up-regulation is mediated by NF-B activation in THP-1 cells (23). In human umbilical vein endothelial cells, vascular endothelial growth factor-and IL-1␤-induced expression of MCP-1 is up-regulated by the AP-1 and NF-B signaling pathways, respectively (25,26). Thus, many proinflammatory cytokines are able to up-regulate the MCP-1 expression in the vascular cells, at least in part, via AP-1 or NF-B signaling pathways. In the present study, we studied the effects of GM-CSF, one of the proinflammatory cytokines, and   lanes 8 and 9, respectively). wt, wild type probe; mt, mutant probe; Comp., cold competitor; NRS, normal rabbit serum.
found that GM-CSF up-regulated MCP-1 expression in human monocytic U937 cells. However, neither AP-1 nor NF-B played a critical role in the GM-CSF-induced MCP-1 expression. Previous reports, including our own, demonstrated GM-CSF stimulation of MCP-1 expression in monocyte/macrophage (8,9,28). These studies, however, did not address the molecular mechanism(s) by which GM-CSF enhances monocytic MCP-1 expression. In general, GM-CSF binding to its receptor leads to activation of several kinases and transcriptional factors, such as Jak2, Ras, Raf, Erk, Stat5, and c-Fos, which consequently regulate gene expression (29). In the case of monocytes/ macrophages, Jak2-Stat5 activation is essential for GM-CSF signaling, and GM-CSF-induced Jak2-Stat5 activation is required for gene expression, monocyte/macrophage differentiation, phagocytosis, and macrophage proliferation (30 -32).
In addition, our current findings clearly demonstrated that GM-CSF stimulation resulted in the activation of the Jak2-Stat5 signaling pathway and that the STAT5 motif was critical for MCP-1 gene transcription in U937 monocytic cells. Therefore, the activation of Jak2-Stat5 signaling induced by GM-CSF might be associated with the critical functions of monocytes/macrophages. The knockdown experiments by siRNA targeting Stat5a (Fig. 3, D and E), however, did not show a complete suppression of the MCP-1 transcriptional activity, indicating the presence of further unknown mechanism(s). Since the phosphospecific anti-Stat5 antibody used in this study recognizes both Stat5a and Stat5b proteins, GM-CSF-mediated activation of Stat5 includes both isoforms (Figs. 3B and 4C). Thus, together with the results obtained from EMSA (Fig.  5), the Stat5b protein might be involved in the GM-CSF-induced MCP-1 expression.
The adhesion of monocytes to the endothelium and consequent  inflammatory response are an initial process in the early stage of atherogenesis, and the expression of proinflammatory cytokines by vascular cells modifies the vascular microenvironment (5,33,34). For example, it has been reported that monocyteendothelial cell interaction induces GM-CSF and MCP-1 synthesis in both cells (35,36). Since both GM-CSF and MCP-1 activate lymphocyte function-associated antigen-1 and very late antigen-4 on U937 monocytes via RhoA activation (9), monocyte-endothelial interaction might be enhanced by GM-CSF and further induce the cytokine expression. Although the molecular mechanisms underlying the adhesive interaction enhancing GM-CSF and MCP-1 expression are currently unknown, expression of GM-CSF and MCP-1 might be closely related to the initiation of inflammatory reactions. Therefore, GM-CSF-induced transcriptional up-regulation of the MCP-1 gene by monocytes in the vascular microenvironment is quite possible for the initiation of atherosclerosis. Interestingly, it has been reported that cervastatin and atorvastatin reduce U937 monocyte adhesion to endothelial cells via down-regulation of lymphocyte function-associated antigen-1 and very late antigen-4 expression and via RhoA inactivation (37,38). In addition, recent studies reported that GM-CSF-induced RhoA activation is inhibited by statins (cervastatin and simvastatin) (39). Thus, statins could inhibit both GM-CSF-induced monocytic MCP-1 expression and integrin/RhoA activation, which action would consequently result in down-regulation of monocytic recruitment in the subendothelial microenvironment. Just as simvastatin reduces tumor necrosis factor-␣-induced MCP-1 expression in human macrophages (40), we demonstrated that Atv treatment decreased GM-CSF-induced MCP-1 mRNA expression in U937 monocytes. The luciferase assay using 3Ј-UTR-including constructs suggests a possibility that the 3Ј-UTR increases mRNA stability and that AUUUA sequence is responsive to Atv. In rat vascular smooth muscle cell, platelet-derived growth factor increases MCP-1 mRNA expression via both increased transcription and mRNA stability (41,42), and dexamethasone decreases the mRNA stability via the AUUUA-independent pathway (43).
Since the Atv effects were partially reversed by mevalonate and mimicked by GGTI or FTI, Atv may regulate the MCP-1 mRNA half-life by inhibiting geranylgeranylation or farnesylation of proteins. In agreement with the present study, Atv has also been reported to decrease mRNA stability of angiotensin AT1 receptor (half-life of 6 versus 2.5 h) in vascular smooth muscle cells via inhibition of geranylgeranylation (44). In contrast, statins up-regulate endothelial nitric-oxide synthase expression by causing an increase in mRNA stability (half-life of 14 versus 27 h) via blocking of the geranylgeranylation of RhoA (45). The exact mechanism(s) of Atv-mediated mRNA stabilization and destabilization is still unknown.
In conclusion, our results indicate that GM-CSF enhanced monocytic MCP-1 expression at the transcription level via activation of the Jak2-Stat5 signaling pathway and that Atv decreased the stability of MCP-1 mRNA, potentially via an isoprenoid pathway. Thus, these results provide further evidence that statins have anti-inflammatory effects independent of their cholesterol-lowering effects. In addition to the inhibitory effects of statins on RhoA and integrins (36 -38), the novel reg-ulatory mechanism by which statins reduce monocytic MCP-1 expression, thereby negatively affecting atherogenesis, would be a new therapeutic target for cardiovascular diseases.