GDF11 prevents the formation of thoracic aortic dissection in mice: Promotion of contractile transition of aortic SMCs

Abstract Thoracic aortic dissection (TAD) is an aortic disease associated with dysregulated extracellular matrix composition and de‐differentiation of vascular smooth muscle cells (SMCs). Growth Differentiation Factor 11 (GDF11) is a member of transforming growth factor β (TGF‐β) superfamily associated with cardiovascular diseases. The present study attempted to investigate the expression of GDF11 in TAD and its effects on aortic SMC phenotype transition. GDF11 level was found lower in the ascending thoracic aortas of TAD patients than healthy aortas. The mouse model of TAD was established by β‐aminopropionitrile monofumarate (BAPN) combined with angiotensin II (Ang II). The expression of GDF11 was also decreased in thoracic aortic tissues accompanied with increased inflammation, arteriectasis and elastin degradation in TAD mice. Administration of GDF11 mitigated these aortic lesions and improved the survival rate of mice. Exogenous GDF11 and adeno‐associated virus type 2 (AAV‐2)‐mediated GDF11 overexpression increased the expression of contractile proteins including ACTA2, SM22α and myosin heavy chain 11 (MYH11) and decreased synthetic markers including osteopontin and fibronectin 1 (FN1), indicating that GDF11 might inhibit SMC phenotype transition and maintain its contractile state. Moreover, GDF11 inhibited the production of matrix metalloproteinase (MMP)‐2, 3, 9 in aortic SMCs. The canonical TGF‐β (Smad2/3) signalling was enhanced by GDF11, while its inhibition suppressed the inhibitory effects of GDF11 on SMC de‐differentiation and MMP production in vitro. Therefore, we demonstrate that GDF11 may contribute to TAD alleviation via inhibiting inflammation and MMP activity, and promoting the transition of aortic SMCs towards a contractile phenotype, which provides a therapeutic target for TAD.


| INTRODUC TI ON
Thoracic aortic dissection (TAD) is one of the most fatal aortic diseases with high morbidity and mortality rates. 1 The incidence of thoracic aortic aneurysms and dissection in the world has increased year by year and occurs at a rate of 4-6 cases per 100 000 person-years. 2,3 Although the advances are made in computed tomography imaging, surgical repair and endovascular techniques, there are neither specific biomarkers for prompt diagnosis or alternative therapies for treating this disease. The underlying pathological mechanism of TAD remains unclear. Previous studies have shown that the phenotype of vascular smooth muscle cells (SMCs) changed from contractile to synthetic, 4 and the dysfunction of vascular SMCs interacting with extracellular matrix (ECM) could affect the behaviour of vessel. 5 Besides, the inflammatory cytokines and matrix metalloproteinases (MMPs) released by activated macrophages can be detected in the aortic tunica media. Del Porto reported that acute aortic dissection caused a significant increase of serum IL-6. 6 High expression levels of IL-6 and MMP-2 were also found in the aortic tissues of rats with aortic dissection. 7 These findings suggest that TAD is the result of artery remodelling which is associated with aneurismal phenotypic transition of vascular SMCs, inflammation and extracellular matrix degradation.
Previous studies have indicated an involvement of TGFβ signalling in TAD formation, however, its exact role is still equivocal and controversial. Increasing evidence showed that the transforming growth factor β (TGFβ) signalling pathway plays an important role in the formation of aortic aneurysm and dissection. 8 Both over-activation and over-inhibition of this pathway was reported to induce the formation of aortic aneurysm and dissection.
Meester and co-workers found that activation of TGFβ pathway promotes the genesis and development of aortic aneurysms. 9 However, Wang et al 10 reported that TGFβ suppressed angiotensin II (Ang II) -induced aortic aneurysm in mice via controlling excessive monocyte and macrophage activation, inhibiting matrix degradation and preserving medial smooth muscle cell survival.
Growth Differentiation Factor 11 (GDF11) is a member of TGFβ superfamily and broadly expresses in embryonic tissues, spinal cord, skeletal muscle, brain, heart, etc. 11,12 Higher GDF11 levels were found to be closely related to lower risk of cardiovascular disorders and death. 13 GDF11 could reduce atherosclerosis and protect against endothelial injury. 14 It is noteworthy that GDF11 counteracts sclerotic arterial disease through preventing the phenotypic transition of carotid arterial SMCs induced by autophagy deficiency. 15 Therefore, we hypothesize that GDF11 participates in the formation of TAD.
In this study, the ascending thoracic aortas from patients with TAD and healthy individuals were collected to detect the expression of GDF11. Furthermore, β-aminopropionitrile monofumarate (BAPN) and angiotensin II (Ang II) were used to establish TAD mouse model. Exogenous GDF11 and adeno-associated virus type 2 (AAV-2) mediated GDF11 overexpression were used to investigate its effect in the phenotypic switching of vascular SMCs. Our data indicated that BAPN/Ang II-induced GDF11 down-regulation as a contributor for synthetic switching of aortic SMCs. Normal control aortic tissues and serum were derived from heart donors (n = 8), aortic valve replacement for aortic valve insufficiency with normal aorta (n = 2). Some medial layer tissues were paraffinembedded for serial histological sections and subsequent staining, and the rest were used to extract tissue proteins and RNAs.

| Content of GDF11 and inflammatory factors
The content of GDF11 in the serum samples and supernatant of cultured mouse SMCs and endothelial cells (ECs) was measured by a GDF11 ELISA kit (Uscn, Wuhan, China) according to the manufacturer's instructions. Serum levels of inflammatory cytokines (TNFα and IL-6) were assessed by TNFα (Multi Sciences, Hangzhou, China) and IL-6 (Roche, Switzerland) ELISA kit respectively according to the manufacturers' instructions.

| Expression and production of GDF11
GDF11 CDS fragment (GenBank Accession No. NM_005811) was synthesized by GENEWIZ (Jiangsu, China) and cloned into the pET30a prokaryotic protein expression vector (Novagen, MerckEurolab, Fontenay-sous-Bois, France). This expression vector was then introduced into Escherichia coli BL21 for protein production. BL21/pET30a-GDF11 cells were grown in LB medium (NaCl 10 g/L, tryptone 10 g/L, yeast extract 5 g/L) with shaking at 37°C for 3 hours. Isopropylβd-thiogalactoside was then added to the medium (final concentration, 1 mmol/L) with shaking overnight at 16°C, followed by centrifugation at 5000 g. Proteins were purified   16 Briefly, mice were fed on a normal diet and administered BAPN solutions dissolved in the drinking water (1 g/kg/d) for 4 weeks. Subsequently, osmotic mini pumps (Alzet, Cupertino, CA, USA) filled with Ang II were implanted subcutaneously in mice for 24 hours (1 μg/kg).

| Animal model of TAD
Intraperitoneal injection of GDF11 into mice (0.1 mg/kg/d) was performed from the beginning of BAPN administration. The survival of mice was recorded daily. Mice were euthanized by excess pentobarbital sodium and thoracic aortic tissues were collected.

| Immunofluorescence
Human and mouse aorta samples were taken from the obviously thickened areas of the thoracic aorta. After fixation, paraffinembedded sections (5μm-thick) were treated with xylene, followed by antigen retrieval for 10 minutes. Subsequently, the sections were

| Histological assessments
The aortic tissues were stained with haematoxylin and eosin (H&E).
After deparaffinization and rehydration, sections from aortic tissues were stained with haematoxylin for 5 minutes and eosin for

| Isolation and culture of mouse aortic SMCs and ECs
SMCs and ECs were isolated from aortas of C57BL/6 mice for in vitro experiments according to previous studies. 15,17 Briefly, aortic medial tissues free of fat tissues were sliced into pieces (1 mm 3 ) and pasted on the bottom of culturing bottle using micro-dissecting scissors. For isolation of ECs, aorta was placed with lumen-side-down onto the culturing bottle. The dissected tissues were incubated in DMEM medium (Hyclone, Utah, Logan, USA) with 20% foetal bovine serum (FBS, Biological Industries, Kibbutz, Israel) at 37°C in a humidified incubator with 5% CO 2 . The medium was replaced every 2 days for 1 week. SMCs or ECs were isolated when they grew out from the dissected tissue. Subsequently, ACTA2 immunofluorescence was performed to identify the SMCs. Primary SMCs and ECs of passage 3 were used for following study. SMCs were incubated with Ang II (500 nmol/L), with GDF11 (50 ng/mL) and/or SB-431542 (10 μmol/L) post-infection with AAV-2 containing GDF11 or GDF11-shRNA (Wanleibio, Shenyang, China).

| Real-time quantitative PCR (RT-qPCR)
Total RNAs were isolated from SMCs through TRIpure reagent (BioTeke, Beijing, China), and then reversely transcribed into cDNA using Reverse Transcriptase M-MLV (Takara, Dalian, China) in the presence of random hexamers and oligo (dT). RT-qPCR was performed by using SYBR Green (BioTeke) and the primer sequences were listed in Table S1. The 2 −ΔΔct method was used to determine relative gene expression.

| Statistics
Data were analysed by Graphpad Prism 8 (Graphpad Prism Software, inc.) and presented as means ± SD. Statistical analysis was done through unpaired t-test between two groups and one-way analysis of variance (ANOVA) with Tukey's multiple comparison among multiple groups. The threshold of statistical significance was set at P < .05.

| Patient characteristics
Demographics and clinical characteristics of all participants are shown in Table 1. The average age of TAD patients was older than healthy individuals (54.2 ± 8.5 years vs 46.6 ± 9.2 years). The average aortic diameter of TAD patients was significantly bigger than those of healthy individuals (56.8 ± 6.5 mm vs 32.1 ± 3.2 mm). TAD patients had significantly higher percentage of aortic valve insufficiency, hypertension, smoking and chest pain than the control (40% vs 20%; 75% vs 10%; 30% vs 10%; 80% vs 10%, respectively). Uniform structures and strong elastin laminae were presented in the control tissues, while TAD samples showed disorderly arrangements, collagen over-deposition and elastin degradation. Serum levels of inflammatory cytokines (TNFα and IL-6) were higher in TAD patients ( Figure 1E). Protein expression levels of TNFα, IL-6, MMP-2, MMP-3 and MMP-9 in TAD thoracic aortic tissues were also elevated ( Figure 1F). ELISA results showed a significant decrease of serum GDF11 in TAD patients ( Figure 1G), which were consistent with the results from western blotting ( Figure 1H). Furthermore, the expression of ACTA2 in aortic medial tissues was decreased significantly ( Figure 1H), which was confirmed by immunofluorescence staining ( Figure 1I). GDF11 found to be co-localized with ACTA2 in thoracic aortic tissues and had a positive correlation with the expression of ACTA2 ( Figure 1J,K).

| GDF11 expression in the mouse model of TAD
We established mouse model of TAD by BAPN/Ang II to verify the above results. Similar to human specimens, the expression of GDF11 and ACTA2 in the aortic tissues from TAD mice also decreased ( Figure 2).

| GDF11 inhibited the formation of experimental TAD
After verifying the size of recombinant GDF11 with SDS-PAGE ( Figure 3A), GDF11 was then injected into TAD mice in order to detect its influence on the formation of TAD. We found that GDF11 treatment improved the survival of TAD mice ( Figure 3B). While 55.56% of mice treated with BAPN/Ang II developed TAD, only 33.33% developed TAD when treated with GDF11 ( Figure 3C).
Mice treated with BAPN/Ang II developed TAD, and their thoracic aortas were significantly dilated ( Figure 3D,E). The tunica  Figure 3F).

TA B L E 1 Clinical characteristics of patients
Significant adventitial thickening and collagen over-deposition were shown ( Figure 3G). GDF11 treatment prevented the pathological damage induce by BAPN/Ang II ( Figure 3E-G). The average thoracic aortic diameter in TAD mice was reduced from 3.001 mm to 1.724 mm after GDF11 treatment ( Figure 3D,H). In addition, quantification analysis of elastin breaks was performed to assess the medial degeneration, and the results showed that TAD mice had a remarkably higher elastin degradation score than the control. GDF11 treatment significantly prevented elastin degradation ( Figure 3I).
The development of TAD is associated with loss of SMC contractile markers, secretion of MMPs and inflammatory infiltration. The expression of ACTA2 and elastin was enhanced in the thoracic aorta tissues of TAD mice treated with GDF11 ( Figure 4A). Further, the expression levels of MMPs, IL-6 and TNFα were also confirmed to be significantly down-regulated in aortas after GDF11 treatment in TAD mice ( Figure 4B,C).

| GDF11 prevented Ang II-induced phenotypic transition of aortic SMCs
The expression levels of GDF11 and ACTA2 was determined by immunofluorescence and western blotting ( Figure S1A,B). GDF11 also reduced their MMP expression ( Figure S1F). Moreover, SB-431542 was used to block TGFβ/Smad-2/3 signalling pathway.
We noted that SB-431542 reversed the effects of GDF11 on the expression of contractile/synthetic markers and MMPs ( Figure S1E,F).
Besides, forced overexpression of GDF11 showed similar effects as GDF11 recombinant protein in SMCs ( Figure 5).
Next, to simulate the in vivo condition, primary SMCs were further stimulated with Ang II. As indicated in Figure 6A, GDF11  Figure S2). We also found that GDF11 expressed in SMCs and ECs, and was released by these cells ( Figure S3). These results implied that GDF11 may be involved in the formation of TAD through both autocrine and paracrine pathways.

| D ISCUSS I ON
In this study, our data firstly revealed that GDF11 level was lower in the human thoracic aorta tissues with TAD than healthy aorta tissues. We noticed that GDF11 could also decline during aging. 18 The ages were different in our human groups, which may be a of TGFβ signalling reduced these alterations. 59 It is suggested that increased contractile SMCs and elevated stiffness of aorta and ECM could be a result rather than a cause of aortic wall dilation, and these could represent disease progression rather than initiation. 60 Further study are required to unveil how aortic dilation and its relation with TGFβ signalling during TAD formation and development.
To analyse potential mutations in genes involved in the biosynthesis or processing of connective tissue proteins (eg PLOD1), genes encoding the ECM proteins (eg FBN1, FBN2 and COL3A1) or genes encoding cytoskeleton components (eg MYH11 and SM22α) is indeed import because the genetic state of these genes affects TAD progress. [61][62][63] It would be interesting to determine whether the down-regulation of GDF11 is correlated to genetic mutations of particular genes that participate in TAD progress in larger-sized clinical samples.
In conclusion, our data revealed that GDF11 alleviated BAPN/ Ang II-induced aortic injury by inhibiting MMP production and ECM remodelling and maintaining contractile phenotype of SMCs possibly via TGFβ signalling pathway. These results suggest that GDF11 may be a therapeutic target for TAD.

CO N FLI C T O F I NTE R E S T
The authors declare that there are no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data generated or analysed during this study are available from the corresponding author on reasonable request.