Vascular progenitor cell senescence in patients with Marfan syndrome

Abstract Vascular progenitor cells (VPCs) present in the adventitia of the vessel wall play a critical role in the regulation of vascular repair following injury. This study aimed to assess the function of VPCs isolated from patients with Marfan syndrome (MFS). VPCs were isolated from control and MFS donors and characterized. Compared with control‐VPCs, MFS‐VPCs exhibited cellular senescence as demonstrated by increased cell size, higher SA‐β‐gal activity and elevated levels of p53 and p21. RNA sequencing showed that several cellular process‐related pathways including cell cycle and cellular senescence were significantly enriched in MFP‐VPCs. Notably, the expression level of TGF‐β1 was much higher in MFS‐VPCs than control‐VPCs. Treatment of control‐VPCs with TGF‐β1 significantly enhanced mitochondrial reactive oxidative species (ROS) and induced cellular senescence whereas inhibition of ROS reversed these effects. MFS‐VPCs displayed increased mitochondrial fusion and decreased mitochondrial fission. Treatment of control‐VPCs with TGF‐β1 increased mitochondrial fusion and reduced mitochondrial fission. Nonetheless, treatment of mitofusin2 (Mfn2)‐siRNA inhibited TGF‐β1‐induced mitochondrial fusion and cellular senescence. Furthermore, TGF‐β1‐induced mitochondrial fusion was mediated by the AMPK signalling pathway. Our study shows that TGF‐β1 induces VPC senescence in patients with MFS by mediating mitochondrial dynamics via the AMPK signalling pathway.


| INTRODUC TI ON
Marfan syndrome (MFS) is mainly caused by the fibrillin-1 (FBN1) gene mutation and is a hereditary disorder of connective tissue with effects on multiple systems including cardiovascular, skeletal and ocular. 1 FBN1 encodes the extracellular matrix protein FBN1 and is involved in the formation of complex extracellular structures in the arteries. FBN1 mutation can lead to contractile dysfunction of smooth muscle cells (SMCs) with consequent reduced tensile strength of aortic tissue. 2 Aortic complications such as aortic dilatation and dissection are the main cause of morbidity and mortality in patients with MFS. 3 Although considerable progress has been made over the past decades in the treatment of MFS-induced cardiovascular injury, including medical and surgical interventions, efficacy has been limited due to the unclear molecular aetiologies. An understanding of the fundamental molecular mechanisms that underlie MFS will provide a novel strategy for MFS management.
The wall of the aorta is composed of three layers: intima, media and adventitia. Recent research has revealed a range of vascular progenitor cells (VPCs) in the adventitia of the vessel wall that are positive for stem cell antigen-1(Sca-1) or c-kit. [4][5][6] These VPCs can give rise to many cell lineages, including endothelial cells and SMCs. 7,8 It has been well documented that VPCs contribute to cardiovascular regeneration therapies. 5 Transplantation of VPCs isolated from the adventitia of patients with coronary artery bypass graft has been shown to greatly enhance blood perfusion recovery and neovascularization in a mouse model of hindlimb ischaemia, 9 demonstrating their great potential in vascular disease treatment.
Nevertheless, the function of stem cells usually declines in a disease milieu. 10 It remains unclear whether the function of VPCs declines in patients with MFS.
A FBN1 defect activates transforming growth factor beta (TGF-β) signalling, leading to vascular injury. [11][12][13] Accumulating evidence has shown that TGF-β is involved in regulation of stem cell senescence. Treatment with TGF-β can induce cellular senescence of young MSCs but treatment with anti-TGF-β antibodies can reduce cellular senescence of old MSCs. 14 Inhibition of TGF-β1 signalling significantly inhibits serum-free-induced endothelial cell senescence and thereby improves endothelial function. 15 Disrupted mitochondrial dynamics that are regulated by fusion and fission are strongly associated with cellular senescence. 16,17 Mitochondrial fission is mainly regulated by dynamin-related protein 1 (Drp1) and mitochondrial fission factor (Mff) whereas mitochondrial fusion is mediated by mitofusion (Mfn) 1, Mfn2 and optic atrophy 1 (Opa1) proteins. Whether TGF-β1 induces VPC senescence via regulation of mitochondrial dynamics and the underlying mechanisms have not been determined. In this study, we revealed that VPCs isolated from MFS exhibited cellular senescence. Importantly, we found that TGF-β1 disrupted mitochondrial dynamics via regulation of the adenosine monophosphate-activated protein kinase (AMPK) signalling pathway, leading to VPC senescence in MFS patients.

| C-kit VPC isolation, culture and characterization
Vascular progenitor cells were isolated from control donors and patients with MFS at Guangdong Provincial People's Hospital, China.
Written informed consent was obtained from all study patients and detailed information is summarized in Table 1. This study was approved by the research ethics board of Guangdong Provincial People's Hospital.
Human VPCs were isolated, sorted and cultured as previously described. 18 Briefly, aorta specimens were taken from the ascending aorta of control donors and MFS patients. After removing the adipose tissue, the adventitial tissue was separated from the media and intima. The adventitia was cut into 1-2 mm 3 pieces and di- The differentiation capacity of VPCs into SMCs, osteocytes and adipocytes was evaluated as previously described. 19

| Scratch-wound assay
Control-VPCs and MFS-VPCs were seeded in a 12-well plate and cultured with complete culture media. When cells reached 90% confluence, scratches of the same width were made on the bottom of the plate using a 1 mL pipette tip. Cells were carefully washed with PBS to remove cell debris and then incubated with serum-free medium in an incubator with 5% CO 2 at 37°C. After 24 hours incubation, the migration of VPCs into the 'wound' area was evaluated by a phase contrast microscope.
After washing with PBS three times, VPCs were fixed with fixative solution for 15 minutes and then incubated overnight with SA-β-gal staining solution at 37°C (without CO 2 ). The percentage of senescent VPCs stained blue was assessed from five different view fields of each sample in three independent experiments.   at room temperature for 1 hour and then exposed in a dark room.

| RNA sequencing and RNA-seq data processing
RNA-seq analysis was performed on control-VPCs and MFS-VPCs using the Illumina sequencing platform and RNA-seq data processed as described previously. 21,22 In brief, differentially expressed genes (DEG) were defined by an absolute value of log1.5 (fold change) > 1 and adjusted P-value <0.05. Pathway enrichment analysis was performed with significant DEGs from the sequencing results. Pathways with a P-value <0.01 were considered significant results. For each pathway, we then calculated the ratio of the number of significant DEGs to the number of total sequenced genes involved in the corresponding pathway. We selected significant DEGs that were enriched in the cellular senescence pathway to plot the heatmap. Unknown genes and genes with missing expression in more than two samples were discarded.

| Immunofluorescence staining
Immunofluorescence staining was carried out according to the protocol as previously described. 23 Briefly, control-VPCs and MFS-VPCs were fixed with formaldehyde for half an hour. Following permeation with 0.1% Triton X-100 in PBS for 30 minutes, cells were stained with ki-67 antibody (Abcam, ab15580), c-kit antibody (Abcam, ab32363), γH2AX antibody (Abcam, ab81299) and incubated overnight at 4°C with a 1:100 dilution. After washing with PBS three times, cells were incubated with the secondary antibodies. Finally, the sample was mounted with DAPI and photographed. Images of five different view fields for each slide were captured randomly by a motorized inverted microscope and analysed using AxioVision (Zeiss). The percentage of positive cells was calculated in three independent experiments.

| Statistical analysis
All data are expressed as the mean ± SEM. Statistical analyses were performed with Prism 5.04 Software ( groups by one-way ANOVA followed by the Bonferroni test. A Pvalue <0.05 was considered statistically significant.

| Characterization of VPCs
Previous studies have shown the existence of VPCs that express Sca-1, CD34 and C-kit in the adventitia of the vessel wall. 24,25 We  Figure 1E). Moreover, both control-VPCs and MFS-VPCs were able to differentiate into adipocytes and osteocytes in vitro ( Figure 1F).
Importantly, compared with control-VPCs, the differentiated capacity of MFS-VPCs into adipocytes and osteocytes was significantly reduced, indicating that the function of MFS-VPCs was impaired ( Figure 1G).

| VPCs isolated from patients with MFS exhibit cellular senescence
Cellular senescence is closely associated with reduced cell function including reduced differentiation capacity and decreased proliferation. 26 Therefore, we first conducted immunofluorescent assay to identify senescent VPCs in the ascending aorta of control donors and MFS patients. As shown in Figure 2A, some c-kit positive cells were co-stained with p53, a senescence-associated marker, indicating the senescent VPCs ( Figure 2A). Notably, more c-kit and p53 double positive cells were observed in the ascending aorta of MFS patients compared with control donors ( Figure 2B). Next, we isolated the c-kit positive cells from the donors and cultured, and then immediately examined the cellular senescence of VPCs without expansion. We found that compared with control-VPCs, the senescence of MFS-VPCs was greatly enhanced as determined by SA-β-gal staining ( Figure S1). Notably, the senescence of VPCs was not increased when expanded to passage 4, indicating that no obvious replicative senescence of VPCs occurs less than passage 4 ( Figure S1)  Figure S2A,B). These findings suggest that VPCs isolated from patients with MFS exhibit cellular senescence.

| Transcriptomic comparison of control-VPCs and MFS-VPCs
To further verify the senescence of MFS-VPCs, we performed genome-wide RNA sequencing (RNA-seq). A total of 1724 up-regulated genes and 2555 down-regulated genes were obtained in MFS-VPCs relative to control-VPCs ( Figure 3A). Gene ontology term enrichment analysis showed that cellular process contains the highest number of significant differentially expressed genes (both up-and downregulated genes) ( Figure 3B). Then, we checked the enrichment analysis on the cellular process-related pathways and found that cell cycle, cellular senescence and signalling pathways regulating the pluripotency of stem cells were significantly enriched ( Figure 3C).
We further examined these pathways and established that 81 differentially expressed genes between control-VPCs and MF-VPCs belong to the cellular senescence pathway, one of the top-ranked cellular process-related pathways (P-value = 3.84e-5), accounting for more than 25% of the total detected cellular senescence-related genes. After removing novel DEGs, the samples could be correctly clustered according to the remaining known genes enriched in the cellular senescence pathway ( Figure 3D). These data suggested a critical role of cellular senescence in the MFS-VPCs.   Figure S4A) and SA-β-gal activity in MFS-VPCs ( Figure S4B,C). These data suggest that TGF-β1 regulates the cellular senescence of VPCs.

| TGF-β1 regulates VPC senescence via mitochondrial ROS generation
Accumulating evidence shows that elevation of reactive oxidative stress (ROS) generation contributes to cellular senescence. 27 Figure 5G). These results suggest that TGF-β1 regulates VPC senescence via ROS generation.

| TGF-β1 induces mitochondrial fusion in VPCs
It has been reported that abnormal mitochondrial dynamics are closely associated with mitochondrial ROS generation. 29 We exam-  Figure 6D). Importantly, Mfn2-siRNA administration also reduced cellular senescence ( Figure 6E,F) and TGF-β1-induced ROS generation ( Figure 6G,H). Collectively, these data show that TGF-β1 induces mitochondrial ROS generation via regulation of mitochondrial dynamics.

| AMPK signalling is involved in TGF-β1 mediation of mitochondrial dynamics
The AMPK signalling pathway plays a critical role in regulation of mitochondrial dynamics. 30 We therefore examined AMPK activation in control-VPCs and MFS-VPCs. We found that p-AMPK was markedly

| D ISCUSS I ON
Vascular stem cells/progenitor cells in the adventitia are essential for maintaining vessel integrity and function. The function of these cells isolated from a diseased donor declines although the underlying mechanisms remain unclear. In this study, we demonstrated that VPCs isolated from patients with MFS exhibited cellular senescence.
We also found that TGF-β1, via regulation of mitochondrial dynamics, induced VPC cellular senescence by elevation of mitochondrial ROS generation. More importantly, AMPK signalling was involved in TGF-β1 mediation of mitochondrial dynamics.
Over the past decade, accumulating evidence has highlighted that a population of stem/progenitor cells exists in the sub-endothelial zone and the adventitial zone of the vessel wall. [31][32][33] These adventitial cells including the stem/progenitor cells play an essential role in maintaining physiological function of the vessel. These stem/ progenitor cells express a panel of markers that define the progenitor cells such as Sca-1, CD34, Flk-1 or c-kit. 18 These cells usually are in a quiescent state in the vessel but can be activated by a pathophysiological condition to mediate tissue repair. In an injured vessel, these VPCs can migrate into the media or the intima to differentiate into SMCs or endothelial cells, participating in repair of the vessel to restore function. 18,33 In this study, we also found a population of VPCs expressing c-kit in the adventitia of the aortic root. We revealed that as well as c-kit, these cells expressed similar surface markers to mesenchymal stem cells such as CD90 and CD105. Importantly, these VPCs also can differentiate into adipocytes and osteocytes. activate TGF-β1, forming a positive TGF-β1-ROS autocrine loop. 37 In this study, we showed that administration of TGF-β1 could induce VPC cellular senescence via ROS generation and treatment of Mito-Tempo attenuated this process, suggesting that TGF-β1-induced ROS accumulation is responsible for VPC senescence. Nevertheless, the relationship between TGF-β1 and ROS generation has not been fully elucidated.
It is well known that mitochondria are the major source of ROS production. There is increasing recognition that altered mitochondrial dynamics are responsible for ROS overproduction. Depletion of protein disulfide isomerase A1 (PDIA1) promotes mitochondrial fission and elevates ROS generation, driving the senescence of endothelial cells. 38 In contrast, abnormally elongated mitochondria induce ROS production, thereby triggering mammalian senescence. 39 Furthermore, depletion of OPA1 leads to mitochondrial fission and rescues cells from F I G U R E 7 Adenosine monophosphate-activated protein kinase (AMPK) signalling is involved in TGF-β1 mediation of mitochondrial dynamics in vascular progenitor cells. A, Western blotting and quantitative analysis of the level of p-AMPK protein in control-vascular progenitor cells (VPCs) and Marfan syndrome (MFS)-VPCs. B, Western blotting and quantitative analysis of the level of p-AMPK, p-Drp1 and Mfn2 protein in control-VPCs that received TGF-β1 or TGF-β1 + AICAR treatment. C, Representative images of SA-β-gal staining in control-VPCs that received TGF-β1 or TGF-β1 + AICAR treatment. D, The SA-β-gal positive cells in control-VPCs that received TGF-β1 or TGF-β1 + AICAR treatment were calculated and are presented as a percentage of the total cells. E, Representative images of Mito-Sox staining in control-VPCs that received TGF-β1 or TGF-β1 + AICAR treatment. F, Quantitative analysis of ROS generation in control-VPCs that received TGF-β1 or TGF-β1 + AICAR treatment. Data are expressed as mean ± SEM (n = 3). *P < 0.05, **P < 0.01, ***P < 0.001. Scale bar = 100 μm senescence-associated phenotypic changes. 39 Senescent human adipose-derived mesenchymal stem cells exhibit increased mitochondrial elongation and ROS generation, suggesting that mitochondrial fusion contributes to cellular senescence. 40 30 We found that p-AMPK was significantly reduced in MFS-VPCs, combined with mitochondrial fusion in VPCs. We presumed that downregulation of p-AMPK could induce mitochondrial dysfunction. It has been reported that AMPK activation induces mitochondrial dysfunction, leading to human fibroblasts senescence. 43 In contrast, in this study, we found that treating MFS-VPCs with AICAR reversed mitochondrial fusion and attenuated cellular senescence, suggesting that AMPK activation inhibits MFS-VPCs senescence. This contradictory phenomenon may be cell type-and stimuli-dependent. It has been demonstrated that TGF-β1 activity impairs mitochondrial function in human skeletal muscle cells via suppression of AMPK activation. 44 Furthermore, the activation of AMPK can inhibit TGF-β1 release. 45 Collectively, there seems to be a positive feedforward mechanism by which TGF-β1 inhibits AMPK activation, that in turn down-regulation of AMPK activation increases TGF-β1 release. Consistent with this finding, in this study we showed that administration of TGF-β1 significantly inhibited AMPK activation, leading to mitochondrial fusion. AICAR treatment abrogated TGF-β1induced mitochondrial fusion and senescence in control-VPCs.
There are also some limitations in this study. First, in addition to oxidative stress, telomere shortening or abnormal autophagy contributes to cellular senescence. Whether TGF-β1 induces VPC senescence via regulation of telomere or autophagy requires further investigation. Second, whether targeting TGF-β1 can enhance the therapeutic effects of VPCs isolated from MFS needs to be examined in a mouse model of MFS. Third, excessive production of proinflammatory cytokines is the major characteristic of vascular injury in MFS. Whether senescent VPCs contribute to inflammation in the aortic wall requires investigation. Last but not least, to support the role of FBN1 defect in VPCs senescence of MFS patients, whether reversion of fibrillin mutation can rescue the VPCs senescence needs to be examined.
Our results demonstrate that TGF-β1 induced mitochondrial fusion, via suppression of AMPK signalling, leading to cellular senescence of VPCs from MFS patients. These findings pave the way to restoring the function of VPCs and provide a novel strategy for treatment of MFS.