Nucleolin mediated pro‐angiogenic role of Hydroxysafflor Yellow A in ischaemic cardiac dysfunction: Post‐transcriptional regulation of VEGF‐A and MMP‐9

Abstract Hydroxysafflor Yellow A (HSYA), a most representative ingredient of Carthamus tinctorius L., had long been used in treating ischaemic cardiovascular diseases in China and exhibited prominently anticoagulant and pro‐angiogenic activities, but the underlying mechanisms remained largely unknown. This study aimed to further elucidate the pro‐angiogenic effect and mechanism of HSYA on ischaemic cardiac dysfunction. A C57 mouse model of acute myocardial infarction (AMI) was firstly established, and 25 mg/kg HSYA was intraperitoneally injected immediately after operation and given once, respectively, each morning and evening for 2 weeks. It was found that HSYA significantly improved ischaemia‐induced cardiac haemodynamics, enhanced the survival rate, alleviated the myocardial injury and increased the expressions of CD31, vascular endothelial growth factor‐A (VEGF‐A) and nucleolin in the ischaemic myocardium. In addition, HSYA promoted the migration and tube formation of human umbilical vein endothelial cells (HUVECs), enhanced the expressions of nucleolin, VEGF‐A and matrix metalloproteinase‐9 (MMP‐9) in a dose‐ and time‐dependent manner. However, down‐regulation of nucleolin expression sharply abrogated the effect mentioned above of HSYA. Further protein‐RNA coimmunoprecipitation and immunoprecipitation‐RT‐PCR assay showed that nucleolin binded to VEGF‐A and MMP‐9 mRNA and overexpression of nucleolin up‐regulated the mRNA expressions of VEGF‐A and MMP‐9 in the HUVECs through enhancing the stability of VEGF‐A and MMP‐9 mRNA. Furthermore, HSYA increased the mRNA expressions of VEGF‐A and MMP‐9 in the extract of antinucleolin antibody‐precipitated protein from the heart of AMI mice. Our data revealed that nucleolin mediated the pro‐angiogenic effect of HSYA through post‐transcriptional regulation of VEGF‐A and MMP‐9 expression, which contributed to the protective effect of HSYA on ischaemic cardiac dysfunction.


| INTRODUCTION
Coronary artery heart disease has become the leading cause of death and disability globally in the last 15 years, 1 and it is mainly attributed to coronary artery stenosis or occlusion-induced myocardial hypoperfusion or ischaemia, which can result in acute myocardial infarction (AMI), myocardial fibrosis and further cardiac dysfunction.
Persistent myocardial ischaemia can subsequently result in irreversible myocardial injury, and the injured myocardial cells are very difficult to regenerate. Consequently, the myocardium supplied by occluded coronary artery becomes disable and even life-threatening. 2,3 Therefore, more and more attention has been focused on the therapeutic angiogenesis.
Angiogenesis, an essential event involved in various physiological and pathological processes, such as wound healing, formation of granulation tissue and embryogenesis, promotes the growth of new capillary blood vessels and restores the blood flow of ischaemic tissue. [4][5][6] It has been proved that neoangiogenesis can effectively restore the blood perfusion of coronary artery and further contribute to the myocardial regeneration. [4][5][6] Nowadays, therapeutic neovascularization or neoangiogenesis has been increasingly considered as a promising therapeutic strategy for ischaemic heart diseases due to its effectiveness and invasive property, and more and more attention has been focused on searching for drugs with strong pro-angiogenic activity and little toxicity. [7][8][9] Hydroxysafflor Yellow A (HSYA), a most representative and water-soluble ingredient of Carthamus tinctorius L., has long been used in the treatment of myocardial ischaemia in China and exhibits prominently antiplatelet, anti-inflammatory, antioxidant and proangiogenic activities. [10][11][12][13] What is worth to be mentioned is the proangiogenic effect of HSYA which has been reported in different ischaemic models recently. 12,14 However, the underlying mechanisms have not yet been fully elucidated.
Nucleolin is a ubiquitously expressed and multifunctional DNA-, RNA-and protein-binding protein in the nucleolus of eukaryotic cells. 15,16 It is conserved in animals, plants and yeast and plays an essential role in various pathophysiological processes such as assembly of ribosomes, DNA and RNA metabolism, chromatin structure, rDNA transcription, rRNA maturation, nucleogenesis, cell proliferation and apoptosis, tumour growth and angiogenesis. 15 The central region of nucleolin protein contains 4 tandem RNA-binding domains, of which the first 2 determine the RNA-binding specificity and affinity of nucleolin recognition element ((U/G)CCCG(A/G)). Nucleolin regulates the mRNA stability of target genes and consequently mediates their post-transcriptional control through binding the nucleolin recognition element to corresponding binding elements of different target genes. [17][18][19] It has been reported that nucleolin can mediate the expressions of angiogenesis-related genes, including vascular endothelial growth factor-A (VEGF-A) and matrix metalloproteinase-9 (MMP-9). 14,20,21 However, whether nucleolin mediates the proangiogenic effect of HSYA through post-transcriptional regulation of these angiogenesis-related genes is largely unknown. In this study, the cardioprotective and pro-angiogenic effect of HSYA on the ischaemic cardiac dysfunction were firstly determined, and then, the role as well as the underlying mechanisms of nucleolin in the proangiogenic effect of HSYA was further investigated.  Firstly, during the inhalational anaesthesia with isoflurane, the electrocardiogram of mice was recorded. Secondly, under sterile conditions, the heart was exposed through a left thoracotomy in the fourth intercostal space, then the LAD was ligated, the thoracic incision was closed when the elevation of ST segment was found in the electrocardiogram.
The mice were randomly divided into 4 groups as follows: (1) sham-operated group (sham); (2) sham-operated group with HSYA treatment (sham + HSYA); (3) LAD-occluded group (AMI group); (4) LAD-occluded group with HSYA treatment (AMI + HSYA). Mice in the sham group and sham + HSYA group underwent thoracotomy without the ligation of LAD. HSYA was administrated immediately after the operation through intraperitoneal injection at 25 mg/kg, while mice in the sham group and AMI group were given equal volume of normal saline to HSYA. Then, 25 mg/kg HSYA and the equal volume of saline were given to the mice once, respectively, each morning and evening.

| Measurement of cardiac haemodynamics
According to the protocol we previously reported, 22   were measured, and the maximal slope of systolic pressure increment (+dp/dt) and diastolic pressure decrement (Àdp/dt) were calculated.

| Measurement of myocardial infarct size
The myocardial infarct size was measured by 2% triphenyltetrazolium chloride (TTC) staining as described previously. 23 Positive TTC staining was in red colour, and the infracted area was pale. Images were analysed by Image-Pro Plus, and the infarct size was expressed as a percentage of ischaemic area at risk (% IAR).

| Pathological examination
The HE staining was performed as described previously. 23 Masson's trichrome staining was performed to display the fibration from viable myocardium in the peri-infarct zone. All the pictures were photographed under an optical microscope (Olympus). The extent of collagen deposition was calculated by Image-Pro Plus software (Media Cybernetics, Bethesda, MD, USA).

| Immunohistochemical staining of CD31 and microvessel density determination
The expression of CD31 in the myocardium was detected by immunohistochemistry as described previously. 24   air with 5% CO 2 to allow the cells to grow and form a monolayer in the flask. After confluence, cells were trypsinized using 0.25% trypsin in Hanks buffer for 2 minutes and resuspended in complete culture medium.

| Real-time quantitative polymerase chain reaction
The total RNA of myocardial tissues and HUVECs were extracted by TRIzol and reverse-transcribed to cDNA with PrimescriptTM RT reagent kit with gDNA; eraser according to the manufacturer's instructions (Takara shuzo Co., Kyoto, Japan). The concentration and purity of total RNA were determined by measuring the OD260 and OD260/OD280 ratio, respectively. The mRNA of VEGF-A, nucleolin and MMP-9 were measured by SYBR â Premix Ex Taq TM (Takara shuzo Co.) through an ABI 7500 real-time PCR system (Life Technology Corporation, Carlsbad, CA).
Each cDNA sample was carried out in triplicate. The relative quantitation of mRNA was analysed using the equation as follows: Ratio = 2 ÀMMCt and normalized by GAPDH. The following primers for mouse were used: VEGF-A forward: 5 0 -CTGCTGTAACGATGAAGCCCTG-3 0 and reverse:

| Immuoblotting
Myocardial tissues and HUVECs were homogenized or scraped with the lysis buffer as previously described. 25 Tissue homogenates and cell lysates were centrifuged at 4°C, 14 000 g for 15 minutes.
The concentrations of protein in supernatant were determined by BCA protein assay kit. 20-50 lg protein was loaded to 10% SDS polyacrylamide gel electrophoresis and then transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, USA).
Enhanced or super chemiluminescence (Invitrogen, USA) was used to detect specific proteins according to the manufacture's instruction. The relative band intensity was quantified by Quantity One software.

| Scratch wound-healing assay
As we reported previously, 20 the migration of HUVEC was determined by the scratch wound-healing assay. Six-well plates were seeded with cells to a final density of 2 9 10 5 cells per well, and a 10-lL sterile pipette tip was used to scratch at the mid of each well until the cells were adherent to the plate. Cells in each well were treated by 50 lg/mL HSYA or vehicle and photographed under a microscope at different time-points. The area of wound edges was measured and compared between different time-points. All scratch assays were performed in quadruplicate.

| In vitro angiogenesis assays: tube formation on Matrigel
The tube formation assay was performed according to the manufacture's protocols of BD Matrigel Matrix (BD Biosciences, USA).
For preparation, Matrigel Matrix was incubated and fully dissolved at 4°C overnight, and 50 ll Matrigel Matrix was applied to each well of 96-well plates and then incubated at 37°C for 30 minutes. Then, 150 ll cell culture medium containing 3-5 9 10 3 HUVECs was seeded on the matrigel in each well. Cells in each well were treated by HSYA or vehicle for different time.
The tubular formation of HUVECs was observed and photographed using an inverted phase-contrast microscope in 3 random fields. Image analysis was performed by Wimasis WimTube software to calculate the number of tubules, loops and branch points. 20 2.12 | Determination of the interaction between nucleolin and MMP-9, VEGF-A mRNA Protein-RNA coimmunoprecipitation was performed to determine the interaction between nucleolin and MMP-9, VEGF-A mRNA as we reported previously. 18,23 5 lg of nucleolin antibody was added into the pre-cleared protein extract, followed by a period of 1 hour incubation at 4°C. Then, the mixture of nucleolin antibody and protein extract was mixed with pre-washed 200 lL protein A/G magnetic beads at 4°C overnight. After being centrifugated at 4°C, 10 000 g for 30 seconds, the supernatant was removed, then total RNA was extracted from the magnetic beads and subjected to realtime quantitative PCR to detect the mRNA expression of MMP-9 and VEGF-A.

VEGF-A mRNA
Human umbilical vein endothelial cells transfected with nucleolinspecific siRNA (siRNA-NCL), pCMV-GFP-nucleolin (pCMV-NCL) and the corresponding siRNA-scramble or control plasmid for 48 hours were incubated with either 0.5% ethanol or 5 mg/mL actinomycin D in 0.5% ethanol for 1, 3 and 6 hours, respectively. Then, the total RNAs were isolated and reverse-transcribed into cDNA. Real-time quantitative PCR of the cDNAs was performed to detect the mRNA expressions of MMP-9 and VEGF-A.

| Enzyme-linked immunosorbent assay
Enzyme-linked immunosorbent assays were performed to detect the VEGF-A and MMP-9 contents according to the manufacture's protocols.

| HSYA improved ischaemia-induced cardiac haemodynamics and enhanced the survival rate of AMI mice
To investigate the protective effect of HSYA on AMI, the cardiac haemodynamics and survival of AMI mice were firstly analysed. As shown in Figure 1A, the survival rate of AMI mice was 37.5%, which was significantly lower than that of the sham-operated controls (100%) (P < .05). However, HSYA treatment obviously elevated the survival rate of AMI mice to 70.6% (P < .05). Moreover, the haemodynamic indexes such as LVSP, +dp/dt and Àdp/dt of saline-treated AMI mice showed significant decline in comparison with the sham-operated controls (P < .05, Figure 1B,D,E), while the LVEDP was evidently increased (P < .05, Figure 1C). However, HSYA treatment significantly abrogated the changes of LVSP, LVEDP, +dp/dt and Àdp/dt of AMI mice (P < .05,

| HSYA diminished the myocardial infarction size and alleviated the myocardial injury
To further determine the protective effect of HSYA on the myocardial ischaemic injury, TTC, Masson and HE staining were performed. It was shown that HSYA markedly attenuated LAD ligation-induced myocardial infarct size in mice (P < .05, Figure 2A). Masson staining showed more and more collagen production on the 7th and 14th day after LAD ligation, while HSYA treatment alleviated the collagen deposition of AMI mice by 50.05% and 59.98% (P < .05, Figure 2B). Furthermore, AMI mice showed more overt left ventricular wall thinning and more serious myocardial fibrosis as compared with the sham-operated controls, the myocardial cells were almost completely substituted by the fibroblasts, whereas a greater number of myocardial cells could be observed in HSYA-treated AMI mice, suggesting that HSYA treatment could alleviate LAD ligation-induced acute ischaemic myocardial injury ( Figure 2C).

| HSYA increased the angiogenesis and nucleolin expression in the ischaemic myocardium of AMI mice
Angiogenesis was proved to be beneficial to improve the ischaemic myocardial injury. 26 So the expression of angiogenic markers (CD31, VEGF-A) in the ischaemic myocardium was detected. It was shown that both the MVD (CD31+ endothelial cells) and VEGF-A expressions (mRNA and protein) of ischaemic myocardium in HSYA-treated AMI mice were substantially elevated than those of saline-treated AMI mice (P < .05, Figure 3A-D). In our previous study, nucleolin was proved to play a pro-angiogenic role during the recovery of heat-denatured dermis through enhancing the production of VEGF-A. 20 So in this study, the expression of nucleolin was also detected, and we ZOU ET AL.    To further identify that the regulation of nucleolin on the mRNA of VEGF-A and MMP-9 was involved in the pro-angiogenic effect of HSYA, the interaction between nucleolin and mRNA of VEGF-A and MMP-9 in HSYA-treated AMI mice was determined by IP-RT-PCR.
We found that nucleolin in the heart tissues of AMI mice and HSYA-treated AMI mice was successfully precipitated through protein A/G magnetic beads, and the expressions of MMP-9 and VEGF-A mRNA were detectable in the extracted from the magnetic beads, suggesting that nucleolin could bind to the MMP-9 and VEGF-A mRNA in the heart of saline-and HSYA-treated AMI mice (Figure 7I). Moreover, both the expressions of MMP-9 and VEGF-A mRNA were relatively weak in the extract of nucleolin antibody-precipitated protein from the heart tissue of AMI mice, which were prominently increased by HSYA treatment ( Figure 7J,K, P < .05).
These results suggested that HSYA treatment could enhance the interaction between nucleolin and mRNA of VEGF-A and MMP-9 in the myocardium of AMI mice.

| DISCUSSION
Angiogenesis is a vital and sequential process in the growth and development as well as wound healing. It is characterized by that successively smaller blood vessels sprout from existing ones to form networks of capillaries. 27  Angiogenesis is a pro-angiogenic factor (mainly VEGF, bFGF, MMP, etc) and antiangiogenic factors (mainly angiostatin, endostatin, thrombospondin-1, etc) coregulated and balanced process. Among these factors, VEGF-A and MMP-9 are considered as important regulators of blood vessel formation, which can promote the proliferation, migration and tube-like structure formation of vascular endothelial cells. As a result, they are considered as potential targets for angiogenesis-related diseases. 37,38 In the present study, we found that HSYA significantly increased the VEGF-A and MMP-9 mRNA and protein expressions in the ischaemic myocardium, which might contribute to the pro-angiogenic effect of HSYA. These results were partly similar to those Wei et al 14  These results suggested that nucleolin might be involved in the cardioprotective and pro-angiogenic effect of HSYA.
As a DNA-, RNA-and protein-binding protein, nucleolin regu- It is also worthy to be noted that HSYA exhibited obvious antiangiogenesis on different cancer models. Besides what we had mentioned above, the inhibitory effect of HSYA on the abnormal proliferation of HUVECs and angiogenesis was reported by Wang et al. 46 Yang et al. 47 found that HSYA inhibited the angiogenesis of hepatocellular carcinoma through blocking ERK/MAPK and NF-jB signalling pathway in H22 tumour-bearing mice. Furthermore, the antiangiogenic effect of HSYA was found in the transplantation tumour of gastric adenocarcinoma cell line BGC-823-induced nude mice. 48 We speculated that there may be a switch point for HSYA to balance angiogenesis. Nevertheless, the on-off switch for the effect of HSYA on the angiogenesis was still unknown, which could not be ignored and should be more focused on in the further study.
In summary, the present study showed that HSYA could ame-