miR-144-3p Regulates Vascular Smooth Muscle Cell Phenotypic Switch via FBN1

Background: Carotid artery dissection (CAD) represents a commonly reported factor causing stroke in young and middle-aged adults. Vascular wall remodeling is one of its important pathogenetic mechanisms. FBN1 is a common pathogenic gene leading to Marfan syndrome, whose mutation can cause the formation of aneurysm and arterial dissection. It was recently demonstrated multiple miRNAs contribute to the development of arterial dissection, while miR-144-3p’s function is undened. Methods: In the current study, vascular smooth muscle cells (VSMCs) were transfected with miR-144-3p mimic and inhibitor, as well as siFBN1 and miR-144-3p + siFBN1, to determine vascular smooth muscle’s contractile genes, extracellular matrix-associated proteins. In addition, miR-144-3p’s effects on cell proliferation, migration, adhesion, invasion and apoptosis were evaluated. Results: The results revealed miR-144-3p had elevated amounts, while the brillin-1 protein showed reduced expression in arterial dissection tissues. Meanwhile, FBN1 was shown to be a miR-144-3p target by dual-luciferase gene reporter assay. In response to miR-144-3p mimic transfection, decreased expression of VSMC contractile gene markers, increased apoptosis, and decreased proliferation, migration, and invasion were found. Conclusions: Overall, miR-144-3p affects the biological function of VSMCs by targeting and regulating FBN1, decreases the expression of contractile genes(cid:0)transforms the phenotype and leads to vascular wall remodeling.


Introduction
Carotid artery dissection (CAD) represents a frequent factor causing stroke in young and middle-aged adults. Studies have found that 10-25% of patients aged 19 to 45 years with rst-episode stroke have undergone arterial dissection [1]. Dissection of the arterial wall, due to a tear in the artery's intimal layer or arterial wall bleeding, can be serious and life-threatening. Although the exact pathogenetic mechanism of arterial dissection remains uncertain, vascular smooth muscle cells (VSMC), as the major constituent of the arterial wall media, are currently considered the main cells leading to arterial dissection formation. [2,3] Fibrillin-1, a product of the FBN1 gene, is a 2871 amino-acid glycoprotein with a molecular weight of 350 kDa, and the micro ber component containing it is the backbone of elastic bers in the vessel wall, which, together with other extracellular matrix proteins, determines the mechanical stability and histophysical properties of the extracellular matrix. [4] The FBN1 gene is an important pathogenic gene in Marfan syndrome, and consists of 65 exons. The RGD sequence of the TB4 domain of the brillin-1 protein encoded by this gene can bind to a variety of integrin receptors, which in turn are involved in cytoskeletal regulation, affecting vascular smooth muscle function and extracellular matrix homeostasis. When brillin-1 protein content in the vascular wall is reduced, extracellular matrix stability is affected and intercellular adhesion is decreased, which can cause morphological remodeling and phenotypic changes in vascular smooth muscle cells (VSMC). [5] Although previous studies have found that arterial dissection can be largely attributed to polymorphic defects in the FBN1 gene, other disruptive factors, such as brillin-1 protein amounts and translational regulation, might also have an important function in arterial dissection's pathogenesis.
MicroRNAs (miRNAs) are non-coding RNAs of about 20 ~ 22 nucleotides in length. They are also the shortest eukaryotic RNAs.
This study indicated miR-144-3p targets the FBN1 gene in both humans and rats. Bioinformatic analysis and dual-luciferase reporter assay demonstrated miR-144-3p downregulated FBN1 at the gene and protein levels by interacting with the 3' -UTR of FBN1 mRNA. Additionally, in the animal model, higher miR-144-3p levels and lower brillin-1 were detected in carotid artery specimens on the modeling side compared with the control side. Transfection of vascular smooth muscle cells with miR-144-3p mimics reduced both FBN1 mRNA and protein amounts, altering the levels of contractile genes, extracellular matrix genes, and matrix metalloproteinases associated with smooth muscle cell phenotype. Cell proliferation, scratch and adhesion assays also revealed miR-144-3p suppressed migration, proliferation, and adhesion in vascular smooth muscle, but promoted apoptosis.

Modeling
Male SD rats (4 W, 60-80 g), provided by Shanghai Jasper Laboratory Animal Co., Ltd. and housed under speci c pathogen-free (SPF) conditions, received standard chow, with free access to water. They were randomly divided into the normal group (drinking tap water, n=8) and model animals (n=8), which were administered β-aminopropionitrile monofumarate (BAPN, Sigma, USA) in drinking water at 0.4 g/100 g diet for 28 days as previously proposed. [16] Establishment of the animal model of carotid artery dissection (a) Anesthesia: Before sampling, the animals were fasted for 4 hours and intraperitoneally administered pentobarbital sodium (3 ml/Kg body weight) for anesthesia. The experimental design and anesthesia had approval from the Animal Room Ethics Committee of Fudan University.

Page 4/26
(b) Modeling: After local disinfection of the neck, the neck's skin along the midline was longitudinally incised, and the left and right carotid arteries were carefully separated. A carotid artery was randomly selected, and two small edentulous microhemostatic clips were used to clamp the carotid artery from both sides of the head and tail in the opposite direction of the handle end (the distance between the two hemostatic clips was 1 cm). Then, both hemostatic clips were rotated at 180° in the opposite direction and xed ( Figure 2A); the other internal carotid artery was left untreated and used as a control.
(c) Specimen collection and storage: After 20 min, the microvascular clamp was gently removed, and the skin was sutured and disinfected. Carotid arteries were taken 24 hours later for subsequent experiments.
(d)All rats dying before the expected study end time (4 weeks) were necropsied right away.
Animal experiments were performed at the Institute of Neurology, Huashan Hospital, Fudan University. The animal experiments followed the Guide for the Care and Use of Laboratory Animals, and had approval from the Institutional Animal Care and Use Committee of Huashan Hospital, Fudan University.
Cell culture, grouping and transfection A7R5 smooth muscle cells (SMC), provided by Shanghai Cell Bank, Chinese Academy of Sciences, were maintained in Dulbecco's modi ed Eagle's medium (DMEM, 190040; Gibco) with 10% fetal bovine serum (FBS) in a 5% CO 2 incubator at 37°C. . Cells were then seeded in 6-well plates at 10 5 /well and underwent incubation at 37°C in a 5% CO 2 incubator, with the medium refreshed at 2 to 3-day intervals. At 70-80% con uency, cell sub-culturing was carried out. This was followed by two PBS washes, a 1-2-minute incubation with 0.25% trypsin, and resuspension and subculture in DMEM with 10% FBS.
Next, 250 μL of Opti-MEM (Gibco) without serum was added to 5 μL of RNAimax reagent and incubated for 5 min at ambient. The two abovementioned solutions were added to cell culture wells following a 20minute incubation at ambient. After 6-8 hours of culture at 37°C in 5% CO 2 incubator, the medium was changed to complete medium for further culture for 24-48 hours. A uorescence microscope (Wanheng Precision Instrument, Shanghai, China) was used for analysis. RNA and protein extractions from cells were carried out for subsequent assays.
Hematoxylin -eosin staining and EVG staining Carotid artery tissue specimens from normal animals and carotid artery dissection tissue samples from model animals were collected. Upon xation with 4% formalin for 24 h, the tissue specimens underwent dehydration with graded ethanol and N-butanol and para n embedding at 60°C, followed by 5-μm serial sectioning. Dewaxing was carried out at 60°C for 1 h with xylene.
For H&E staining, incubation was rst performed with hematoxylin (3-5 min). After incubation with acid and ammonia solutions for 40 s each, the samples underwent staining with eosin (2 min) and were soaked in xylene (5 min). An inverted microscope (Olympus, Tokyo, Japan) was utilized for analysis.
In EVG staining, a 30-min incubation was carried out with EVG solution (hematoxylin, iodine solution and ferric chloride at 5:2:2) followed by washing. Ferric chloride differentiation solution was utilized for background. These steps were repeated until a grey white background was obtained. This was followed by staining with VG solution (saturated picric acid and Fuchsin solution at 9:1), multiple washes and dehydration with 100% ethanol. After neutral gum sealing, an inverted microscope (Olympus) was utilized for analysis.

Immunohistochemistry
Carotid artery tissue specimens from normal animals and carotid artery dissection tissue samples from model animals underwent xation with 10% formalin, para n embedding and sectioning 3-4 μm. After treatment with 3% H 2 O 2 , xylene dewaxing (10 min) and dehydration with graded ethanol were performed. This was followed by antigen retrieval by boiling for 90 seconds. The samples were blocked with 5% bovine serum albumin for 30 min at 37°C, followed by incubation with rabbit anti-rat FBN1 antibodies (ab523076, Abcam; 1:300) overnight at 4°C. Next, the specimens underwent incubation with horseradish peroxidase (HRP)-linked goat anti-rabbit IgG-secondary antibodies (ab6721, Abcam) for 30 min at 37°C. The sections were counterstained with hematoxylin for 30 s and developed with diaminobenzidine (P0202; Beyotime Biotechnology). Next, hydrochloric acid solution was utilized for dehydration, which was followed by mounting with neutral balsam and microscopic analysis. In this study, positivity was considered as >25% cells stained for brillin-1 (brown or tan granules in the cytosol). Totally ve highpower visual elds (40×) were randomly examined, and 200 cells were enumerated per eld. Positive cells per eld were counted and averaged.

Migration assay
Upon transfection for 24 h, cells were seeded in 6-well plates at 5×10 5 /well. At 90% con uency, a 200-µl tip was utilized for scratching, which was followed by a PBS wash. Serum-free medium was next supplemented for another 0.5-1 h. At 0 h, 24 h and 48 h, respectively, cell imaging was performed, followed by migration distance measurement with Image-Pro Plus Analysis (Media Cybernetics, MD). Assays were performed three times.

Cell Adhesion Assessment
Totally 50 mg/L Matrigel (356234; BD Diagnostics) was diluted 1:10 with FBS-free culture medium, and 60 μl/well was used to coat 96-well plates, followed by incubation in a 37°C/5% CO 2 incubator for 4 hours. The supernatant was aspirated and discarded. Then, the transfected cells underwent seeding in 96-well plates at 5×10 4 /well, with 5 duplicate wells per group. Meanwhile, a control group was set up, that is, a group with the supernatant not discarded after incubation. After 4 h, the culture medium was aspirated, and non-adherent cells were washed off. Totally 100 μl of fresh medium was added to each well, and 5 μl of CCK8 solution was added to each well for incubation in the dark. In the control group, 5 μl of CCK8 was directly added (counted as 0 h). After 4 hours of culture (counted as 4 h), OD at 450 nm was obtained on a Synergy H1 BioTEK (Ptotem Instruments Co., Ltd. USA).
Transwell assay Matrigel (356234; BD Diagnostics) diluted with serum-free medium at 1:3 was used to coat the upper transwell chambers (50 μL/well) for 30 min at 37ºC. Totally, 10 5 cells/ml were added to upper chambers in medium without serum, while 10% FBS-containing medium was placed into lower chambers. Based on the amounts of cells passing through the matrigel, cell invasion was quantitated. The assays were run thrice.
Statistical analysis SPSS 22.0 and GraphPad Prism 8 were utilized for data analysis. Data are mean ± standard deviation (SD). The t test and one-way analysis of variance were performed to assess group pairs and multiple groups, respectively. P<0.05 re ected statistical signi cance.

Results
Pathological alterations of vessel wall tissues after dissection An animal model of spontaneous arterial dissection was established with mechanical clip rotation as previously described ( Figure 1A). H&E staining showed that compared with the carotid artery tissue of completely normal animals and the carotid artery tissues of the non-rotation side of model animals, the carotid artery tissue of the rotation side was disorganized and disordered, accompanied by surrounding in ammatory cell in ltration. EVG elastic ber staining showed disorganized, disordered tissue, and multiple breaks in the elastic ber component in the arterial dissection tissue ( Figure 1B and 1C).

Fibrillin-1 protein expression is lower in the tissues of arterial dissection
After immunohistochemical staining (Figure 2A and 2B), the stained artery dissection tissues had a lighter color with reduced amounts of tan particles. Meanwhile, the normal tissues were dark upon staining, with overtly more than particles in comparison with the artery dissection tissue. In comparison with normal tissues, the positivity rate of brillin-1 was higher in the artery dissection tissues (P<0.05).
Elevated miR-144-3p expression, reduced FBN1 and extracellular matrix protein amounts, and enhanced MMPs expression in arterial dissection tissues.

FBN1 is a miR-144-3p target
Online analysis demonstrated miR144-3p's binding sites at the 3'-UTRs of human and rat FBN1 mRNAs were from 2611 to 2617 ( Figure 4A) and 2585 to 2591 base pairs ( Figure 4B), respectively. After comparison, the binding sequences were very similar in human and rat species. Therefore, this study in the rat model might be applied to humans.

Upregulated miR-144-3p or downregulated FBN1 inhibits smooth muscle cell proliferation
The CCK8 assay showed that cell proliferation ability in the other four groups was signi cantly changed after 48 h compared with the blank and NC group (all P<0.05). In comparison with the blank and NC group, cell proliferation in the miR-144-3p mimic and siRNA-FBN1 groups was reduced, while the miR-144-3p inhibitor group showed enhanced proliferation (all P<0.05); the miR-144-3p inhibitor + siRNA-FBN1 group showed no signi cant change (P>0.05).

Upregulated miR-144-3p or downregulated FBN1 inhibits cell adhesion
The adhesion assay demonstrated the blank and NC groups were comparable in adhesion ability (P>0.05). In comparison with the blank and NC groups, signi cantly reduced cell adhesion ability was found in the miR-144-3p mimic and siRNA-FBN1 groups (P<0.05); cells transfected with miR-144-3p inhibitor had signi cantly elevated adhesion ability (P<0.05), which was unchanged in the miR-144-3p inhibitor + siRNA-FBN1 inhibitor group (P>0.05).

Discussion
Carotid artery dissection represents an important cause of stroke in young and middle-aged adults. Many risk factors, e.g., hypertension, dyslipidemia, and genetic diseases, all elevate the incidence of arterial dissection. Previous studies related to aortic dissection have focused on genetic susceptibility and genetic impact on the pathogenesis of arterial dissection. [3] The FBN1 gene is the main causative gene of Marfan syndrome, and defects in the latter gene can cause abnormalities in the encoded protein brillin-1, resulting in a large group of connective tissue diseases, collectively known as type I brillinopathies. Among them, Marfan syndrome (MFS) represents the commonest, and the most important cause of death is cardiovascular accidents, including arterial dissection and aneurysms. Weill-Marchesani syndrome, acromicric and geleophysic dysplasia, ectopia lentis and scleroderma syndrome are rare. [18] Our team found in a study performed a few years ago that the brillin-1 protein also has an important function in the occurrence and development of cerebral artery dissection, with brillin-1 protein level closely related to the severity of the disease. In addition to plasma brillin-1 levels being signi cantly higher in patients with craniocerebral carotid dissection compared with healthy individuals, they are also higher than in patients with ischemic stroke caused by other etiologies; patients with craniocerebral carotid dissection are also characterized by dynamic changes of increased plasma brillin-1 in the acute phase and decreased amounts in the chronic phase. [19] In the latter study, we made a new attempt based on the classical method in previous aortic dissections. An animal model of spontaneous carotid dissection was successfully established using mechanical rotation of the carotid artery. The levels of the brillin-1 protein were decreased in the dissected tissue after rotation compared with both the nonrotation side and normal carotid tissues.
The brillin-1 protein is the main structural component of micro bers, which surround the outermost layer of elastic bers and constitute the skeleton formed by elastic bers. This structural composition of micro bers and elastic bers plays a crucial role in VSMC in billions of stretching and recoil cycles. [20] Multiple miRNAs are considered to contribute to the pathogenesis of vascular remodeling and aneurysm. [21,22] The latter ndings overtly reveal a role for miRNAs in the development of arterial dissection and aneurysms. In the present study, it was con rmed that miR-144-3p functionally targets FBN1 and negatively regulates the function of VSMC as well as secreted brillin-1 protein. The miR-144-3p/FBN1 axis has not been reported previously for its regulatory function. We also rstly examined the biological functional changes of VSMC via regulation of the FBN1 gene by miRNAs in carotid dissection disease. It has been previously shown the miR-29 family of miRNAs contribute to elastin downregulation in adult aorta and are closely related to aortic dissection. Merk et al. further demonstrated miR-29b is important in early aneurysm development in a mouse model of MFS. [23] In the latter work, miR-29b amounts were elevated in thoracic aortic aneurysms of MFS mice, alongside enhanced apoptosis and MMP-2 activity as well as reduced amounts of anti-apoptotic proteins (Mcl-1 and Bcl-2) and elastin. [24] This regulatory mechanism was further con rmed by other studies. In addition, LNA-anti-miR-29b administration suppresses AA development, aortic wall apoptosis and ECM degeneration. [25] However, miR-29 has multiple target genes that directly interact with ≥16 ECM genes, including collagen isoforms (COL1A1, COL1A2, and COL3A1), brillin-1 (FBN1) and elastin (ELN), and contribute to extracellular matrix remodeling in many organs.
In the pathogenesis of arterial dissection, vascular remodeling is considered an important link in the development of arterial dissection. [29] Vascular remodeling involves cell growth, death, migration, and extracellular matrix production and degradation. It is not only an adaptive physiological process to maintain vascular homeostasis but also a key pathological link common to many important vascular diseases. VSMC, as the main cellular components of the vascular wall, are critical for maintaining vascular wall tension and structural stability based on their number, distribution and function. [30,31] The proliferation rate of VSMC in adult blood vessels is very low, and the synthetic function is not active, mainly expressing contractile proteins such as α-SMA, CNN1, SM22α and MYH11. [32] Although the proliferation rate of VSMC in adult blood vessels is different from that of end-differentiated cells of myocardial and skeletal muscle cells, VSMC show certain differentiation ability in case of changes in the vascular environment, regardless of physiological or pathological changes, e.g., a shift to proliferation or synthesis ability to adapt to changes in the internal environment of the body. They also secrete a variety of cytokines such as MCP1, MMPs and ADAMTS, and subsequently induces in ammatory cells to release in ammatory factors; in addition, elastic bers and collagen bers are broken, and cell matrix degradation further destroys the structural stability of the vascular wall and participates in the regulation of vascular wall remodeling. [33] In this work, miR-144-3p upregulation signi cantly suppressed the proliferative function of VSMC by regulating FBN1. The balance of cell proliferation and apoptosis is an important event in tissue homeostasis. [34] The proliferation of VSMC might have a crucial function in the occurrence and progression of vascular remodeling, while in vasodilatation and hemorrhagic diseases, including aortic dissection and aneurysm, apoptosis of VSMC is more signi cant in vascular remodeling than cell proliferation in the weak link of the wall. [35] As shown above, miR-144-3p upregulation signi cantly induced apoptosis in vascular smooth muscle cells, and cell injury and late apoptosis were predominant. Combined with the suppressed effects of miR-144-3p on SMC proliferation in previous studies, it is suggested that in the vascular remodeling of arterial dissection, miR-144-3p upregulation may contribute to the vascular remodeling of the arterial wall by promoting apoptosis in SMC to weaken the arterial medial wall, which in turn affects the occurrence of arterial dissection. Moreover, this study further demonstrated that the expression of the transcription factor Srf was decreased after miR-144-3p overexpression or FBN1 gene knockdown, suggesting that FBN1 dysfunction may also affect the expression of smooth muscle cell's contractile genes through other mechanisms, e.g., affecting the formation of transcriptional complexes or their binding to SREs.
In addition, the amounts of related extracellular matrix proteins, including elastin and collagen, were also decreased in this work after miR-144-3p upregulation. The expression levels of MMPs, including MMP-2 and MMP-9, which are highly associated with the development of arterial dissection, were elevated in cells transfected with miR-144-3p mimic and siFBN1. Elevated MMPs can further degrade extracellular matrix proteins such as brillin-1, produce more brillin-1 fragments, forming a vicious cycle and ultimately leading to increased destruction of the vessel wall and poorer stability.
[36] It has also been shown that MMPs induce proliferation and migration in VSMC and aggravate vascular wall remodeling. [37]

Conclusions
Overall, this work provides evidence that miR-144-3p promotes smooth muscle cell apoptosis by regulating its target gene FBN1, thereby inhibiting smooth muscle cell invasion, migration, adhesion and proliferation. Smooth muscle cells were also found to transform from contractile to synthetic, with changes in the expression of related extracellular matrix proteins and MMPs. However, there are some major points requiring further analysis, e.g., the detailed mechanism underpinning FBN1 inhibition by miR-144-3p. Therefore, further studies are warranted to identify a safer and more effective approach for targeted therapy and diagnosis.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. AuthorChecklistFull.pdf