Co-Culture with Chorionic Villous Mesenchymal Stem Cells Promote Endothelial Cell Proliferation and Angiogenesis via ABCA9-AKT Pathway

Objective It is found that human chorionic villous mesenchymal stem cells (CV-MSCs) are a promising and effective treatment for tissue injury. Trophoblast dysfunction during pregnancies obviously entangles to the pathogenesis of preeclampsia (PE). This work was to make it clear how CV-MSCs regulated vascular endothelial cell function. In this study, we treated HUVEC with CV-MSC CM and RNA-seq analysis was used to understand the changes in HUVEC. We examined the levels of ABCA9 and AKT signaling in HUVEC by immunohistochemistry, western blot and qRT-PCR assays. CCK-8, colony formation and tube formation assays were used to understand the role of ABCA9 in the growth and angiogenesis of HUVEC mediated by CV-MSCs.


Introduction
Vascular endothelial cell is a type of epithelium which exists in blood vessels' the interior surface. It referred to various aspects of vascular biology, including blood clotting, angiogenesis and barrier function 1,2 . Therefore, endothelial dysfunction is often regarded as a hallmark for many vascular diseases, including hypertension, atherosclerotic diseases, and preeclampsia 3,4 . Preeclampsia is a pregnancy-speci c disorder affecting 2-8% of all pregnancies 5 . The poorly established placenta produces anti-angiogenetic factors inducing generalized systemic in ammation and endothelial cell dysfunction, resulting in clinical signs of preeclampsia, such as hypertension and proteinuria 6,7 . To date, although the potential mechanisms underlying the pathogenesis of preeclampsia remain unclear, endothelial cellular injury has become a major research focus of the pathogenetic processing 8 .
Derived from BM which adhered to formed and plastic broblast colonies, mesenchymal stem cells (MSCs) were rstly to be considered as spindle shaped cells, and they are able to turn into cell derivatives of the mesenchymal lineage, such as myocytes, osteocytes, adipocytes and chondrocytes 9,10 . MSCs have been derived from various adult tissues, such as amniotic uid, and umbilical cord blood, liver, adiposetissue, bone marrow, endometrium, dental pulp, muscle and placenta 11 . Because of their innumerable biological characteristics connected with the ease of their obtention and in vitro expansion, MSCs have played an important role in the cell therapy-based regenerative medicine eld [12][13][14] . With the increasing knowledge about MSC regenerative property, more and more attention to their clinical potential is paid.
For example, MSCs signi cantly reduce cell angiogenesis and proliferation 15,16 . Moreover, MSCs are found to have antimicrobial, antiapoptotic, and immunoregulatory functions 17,18 . The more important thing is, it has been discovered that placental cells play an important part in regulating peroxidation reactions and speci c damaged tissue recovery 18,19 . Placental MSCs are shown to have an effect on experimental animal models with cerebral ischemia, spinal cord injury, Parkinson's disease, ischemic brain disease and Alzheimer's disease by taking advantage of cytokines and/or neurotrophic factors secreted from placental MSCs and angiocrine growth factors [20][21][22] . Whereas, information about the chorionic villous mesenchymal stem cells (CV-MSCs) in vascular endothelial cells in placentalumbilical cord circulation, including cellular proliferation, migration, and formation of capillary-like tube structures, is limited. Therefore, the present study used CV-MSCs conditional medium treated with Human Umbilical Vein Endothelial Cells (HUVECs), and then investigate whether and how CV-MSCs would be affected endothelial cells.
Transcriptome analyses manifested considerable upregulation of ABCA9 in HUVEC by CV-MSCs CM. In addition, the expression of ABCA9 affected AKT signalling in HUVEC with corresponding changes of proliferation and tube formation capability. Our results con rmed that CV-MSCs could promote proliferation and tube formation of HUVEC through ABCA9 and AKT regulation.

Patients and samples
The use of human placentas and all experiments were supported by the Ethical Committee of the A liated Hospital of Qingdao University (Shandong, China). Moreover, to participate in the research, all of the volunteers signed the written informed consent. With parental permission, the age-matched placentas from severe preeclampsia after caesarean section (n = 10), while placentas were derived from full-term births (n = 10). All placentas were gathered through elective cesarean section with induced labor. Besides, placental tissues were collected by the way of getting rid of a vertical plane through the full thickness of an obviously normal and central area, containing not only the maternal but also fetal surfaces. However, tissues with clots or calci cation were not used.
Cell Culture and CV-MSC conditioned medium preparation The above-mentioned placentas tissues were was minced and removed blood. Then collagenase (0.1%) and trypsin (0.25%) (type I; Sigma-Aldrich, St. Louis, MO) were made use of to digest the tissues for 30 min. After ltrating through a 100-μm nylon lter and centrifugation, the cells were placed on culture plates in a stem cell culture medium (SCCM), and the culture medium consisted of 5% UltraGRO TM (Helios, USA) and Stem Cell Basic Medium (Dakewe Biotech Co., Guangzhou, China). Besides, an incubator, with 5% CO 2 and thirty-seven degrees Celsius, was applied to culture the primary cells. In this study, all CV-MSCs were mentioned at passages 3-6.
The medium was replaced with DMEM/F12 (Gibco, Carlsbad, CA) lacking FBS and the cells were cultured for another 24 h, after CV-MSCs isolated from the placentas reached 80% con uence. What's more, the medium went through a ltration through a lter (0.22 µm) (Millipore, Billerica, MA) after a twelve-minute centrifugation (1,200 × g). And the medium was used for later experiments. Besides, as control CM, the DMEM/F12 medium lacking FBS were adopted.
From the Type Culture Collection China Centre, HUVEC were purchased. P1 cells were human umbilical vein endothelial cells. To culture two cells in an incubator with 37°C and 5% CO2, DMEM/F12 including 10% FBS was adopted.

Transient transfection, lentivirus infection
Transient transfections for lentivirus infection were performed by Lipofectamine 2000 (Invitrogen, Carlsbad, USA). In order to set uo stable transfected cell lines , the transfected cells were isolated as single clones after puromycin treatment.
The sequences of the shRNAs used are the following: sh-Control, 5'-GCAAGCTGACCCTGAAGTT -3'; sh- Control lentivirus (Vector), lentiviral constructs expressing full-length ABCA9 were used to establish stably overexpressed cell lines.

Cell Proliferation Analysis
In the 96-well plates, cells (5×10 3 /well) were embedded. According to the manufacturer's instructions, the CCK-8 reagent (Thermo Fisher Scienti c) was applied to examine cell proliferation every day. All samples were tested at 450 nm with a microplate reader. Through three independent experiments, the growth curves were justi ed.

Immunohistochemistry
In order to x PE placenta (n=20) 4% and the normal human term placenta (n = 20), paraformaldehyde was adopted for 60 min. The tissues were embedded in para n, sliced into 4-μm sections, and then depara nized. Later, for antigen retrieval, the slides were boiled in 6.0-pH sodium citrate buffer (10 mM) for 7 min under one hundred and twenty degrees Celsius. At the same time, Hydrogen peroxide was adopted to block endogenous peroxidase for 10 min. Each of the slides was washe d with TBS three times for 5 min. The TBS contained 0.05% Tween 20 (TBS/T; Merck; Darmstadt, Germany). Next, these slides were incubated with anti-P62 antibodies (1:1,000) and monoclonal anti-STAT3 antibodies (1:200) for 12 h under four degrees Celsius. Moreover, under thirty-seven degrees Celsius, diluted biotinylated secondary antibodies were applied to incubate the sections for 20 min. The target proteins were visualized by fresh DAB solution; later, haematoxylin was used as a tissue counterstain. The expression of the target proteins was independently assessed by two observers with an optical microscope (Olympus FV500, Tokyo, Japan). The area was analyzed via image-Pro Plus 5.1, with the intensity of staining in ve random regions (200× magni cation); by doing this, the expression level of proteins was evaluated.
Colony forming assay A total of 500 -1000 HUVECs were incubated in 6-well plates for 14 days. Paraformaldehyde was then used to x the colonies which were afterwards stained with a crystal violet solution for visualization.

RNA sequencing
Samples were collected using TRIzol (1 ml) (Thermo Fisher Scienti c), and were placed at minus 80 °C.
According to the instructions for an Illumina TruSeq RNA Sample Prep Kit, the libraries were prepared , and on a MiSeq instrument, the sequencing was conducted. Analysis of data (from Annoroad Gene Technology, Beijing, China) was performed on RSEM software. With labeling of no. SRR11493647-no. SRR11493652, all RNA-seq data are available at Sequence Read Archive (SRA).

Tube formation assay
Tube formation assays were conducted according to previous description 24 . Matrigel was added to the pre-chilled culture plates (Corning; Midland, Michigan, USA) and they were put under 37 °C for 30 min.
Then with different treatments, 2 × 10 4 HUVEC cells were embedded in each well and incubated for 6 h.

Statistical Analysis
Statistical analyses were conducted with the use of unpaired two-tailed Student's t-test. And Data are given as mean ± SEM. All experiments were conducted with more than three replicates. To analyze those data, GraphPad Prism version 7.00 software program (GraphPad; La Jolla, USA) was adopted. After taking a P value < 0.05 into consideration, a signi cant difference was worked out.

CV-MSCs promote the proliferation and tube formation of HUVEC cells in vitro via ABCA9
First, we isolated CV-MSCs from healthy placentas, and then, the CCK-8, tube formation assays and colony forming assays was used to gure out the angiogenesis and proliferation of HUVEC dealt with or without CV-MSC conditional medium (CM). The HUVEC treated with CV-MSC CM displayed signi cantly higher proliferation rates and angiogenesis abilities than those untreated cells (Fig. 1a-c, all P<0.01). To further elucidate the possible molecular mechanisms of CV-MSCs CM, differentially expressed genes (DEGs) in CM-treated HUVECs were investigated using RNA sequencing, and the molecular mechanism of underlying functional changes in HUVEC was explored. We compared the transcriptomes between HUVEC cells co-cultured with or without CV-MSCs CM for 48 h, and 4006 upregulated and 4412 downregulated genes were found to change more than two-fold (Fig. 1 d). Based on the results, ABCA9 was found to be downregulated signi cantly in HUVECs treated with CV-MSCs CM. The result was clari ed by western blot and qRT-PCR assays (Fig. 1e-f).
These outcomes showed the CV-MSCs promoted HUVEC proliferation and angiogenesis, and ABCA9 was upregulated in this process.

ABCA9 is downregulated in the placenta of PE pregnancies
To better understand the role ABCA9 played in placenta of PE patients, we examined the protein and mRNA expression levels in placenta from normal and PE pregnancies by immunohistochemistry, western blot and qRT-PCR. The immunohistochemistry assays reveal a lower expression of ABCA9 in PE placenta (Fig. 2a). Our qRT-PCR results con rmed that mRNA levels of ABCA9 were signi cant decreased in PE placenta (Fig. 2b). In consistence with the changes of mRNA levels, the ABCA9 protein levels were also signi cantly decreased in PE placenta (Fig. 2c). Our results suggested that ABCA9 could be inhibited in placenta by PE pathological process, which suggested that ABCA9 could play an role in the pathophysiological process of PE.

ABCA9 promotes HUVEC proliferation and angiogenesis
To better understand the role ABCA9 played in HUVEC after CV-MSC CM treatment, HUVECs were transfected with ABCA9 siRNA and overexpression lentiviruses. The mRNA levels of ABCA9 were examined, and signi cant decrease and increase of ABCA9 mRNA levels were respectively observed in transfected HUVECs (Fig. 3a). In consistence with the changes of mRNA levels, the ABCA9 protein levels were also signi cantly changed in transfected HUVECs compared to the control cells (Fig. 3b). Moreover, the growth of HUVECs was signi cantly inhibited in the presence of ABCA9 siRNA, while the ABCA9 overexpression promoted HUVECs proliferation (Fig. 3c). In addition, we examined the colony and tube formation of transfected HUVECs, and found that the cellular colony formation and angiogenesis changes in the corresponding HUVECs shared the same curve as observed changes in proliferation (Fig.  2d-e). Our results suggested that ABCA9 could be promoted by CV-MSCs which regulated the proliferation and angiogenesis of HUVECs.

ABCA9 enhances cell proliferation and angiogenesis through AKT pathway in vivo and in vitro.
In a phosphatidylinositol-3 kinase-dependent manner, Akt is a serine/threonine protein kinase activated by cytokines and various growth factors 26 . It is well known that Akt signaling pathways regulates vascular homeostasis and angiogenesis through antiapoptotic activity in endothelial cells 27,28 , suggesting that it is involved in ABCA9 associated HUVEC regulation. To understand whether the akt signaling plays a part in promoting effects of ABCA9, protein levels of β-catenin, p-GSK3β, p-akt and akt in HUVEC treated with ABCA9 siRNA and primary cultured human umbilical vein endothelial cells treated with ABCA9 overexpressed plasmids were examined by western blot (Fig. 2g). Our results con rmed that the Akt signaling pathways could be regulated by ABCA9 in HUVECs.
The ABCA9 overexpressed HUVECs were treated with or without the PI3K-Akt pathway speci c inhibitor LY294002 (ab120243) in order to justify the role of akt signaling played in HUVECs proliferation and angiogenesis regulation of ABCA9 (Abcam Biotech, Shanghai, China). Therefore, the AKT and p-AKT expressions were examined with the help of western blot in ABCA9 overexpression primary cultured human umbilical vein endothelial cells with or without LY294002. The relative p-AKT expression levels were signi cantly upregulated in the ABCA9 overexpressed HUVECs, and LY294002 could signi cantly hinder the elevated p-AKT expression (Fig. 4a). Moreover, the CCK-8 assays revealed that the ABCA9 overexpression markedly promoted cell proliferation, which could be attenuated by LY294002 (Fig. 4b).
As our expected, ABCA9 overexpression resulted in the promotion of tube and colony formation of primary HUVECs, and impaired angiogenesis and colony formation ability was also observed in ABCA9 overexpressed HUVECs treated with LY294002 ( Fig. 4c-d).
The results suggested the possible involvement of ABCA9 in modulating the AKT pathway in CV-MSCs treated HUVEC. Then, we investigated if the CV-MSCs could regulate the proliferation and angiogenesis of HUVECs through ABCA9 regulation in vitro.
To further con rm whether akt signaling plays a role in ABCA9 mediated HUVEC promoting effect, we then investigated the angiogenesis and proliferation of HUVECs by Metrigel plug angiogenesis assay (Fig. 4e). HUVECs with ABCA9 overexpression showed faster proliferation and angiogenesis, and LY294002 could attenuate the effect of ABCA9 overexpression.
To summarize, these data indicate that CV-MSC-CM may activate the ABCA9-AKT pathway to promote the proliferation and angiogenesis of HUVECs.

Discussion
Preeclampsia and their complications during pregnancy are main causes of maternal morbidity and mortality, and endothelial cell dysfunction has been shown to be a central event in this pathophysiology 29 . When endothelium is in uenced by oxidation and in ammation, developing dysfunction includes increased cell adhesion, losing the vasodilation and vasoconstriction balance 30,31 . Due to aberrant placentation and shallow invasion of trophoblast cells in PE, multiple high levels of in ammatory cytokines were released, leading to excessive vascular endothelium injury 32 . Many researches have demonstrated that MSCs could secrete greater amounts of chemokines and various growth factors, and the therapeutic potential of MSCs can be applied to many diseases 33  Human Mesenchymal Stem Cells (hMSCs) are multipotent stem cells that are able to renew themselves and grow into tissues with different functions 35 . In vivo hMSCs are sources of trophic factors which are capable of inducing intrinsic stem cells to x damaged tissues and modulating the immune system 36 . For the present, a great number of clinical trials are applying hMSCs as therapeutic purposes in many immune diseases, and the results are promising 37 . The placenta is made up of villous and smooth chorion, the amnion, decidua basalis, umbilical cord and chorionic plate 38 . The chorionic plate has already been an available source of MSCs, which exhibiting undifferentiation marker genes expression and transdifferentiation, such as TERT, OCT3/4, KLF4, SOX2 and c-MYC 39 . The placenta is a readily reliable and available source of allogeneic cells apart from other sources such as cartilage, bone marrow, ligament and adipose tissue in various cases of genetic and degenerative diseases 40 . As for industrial scale-up for a sustainable supply of a large amount of quality-controlled and affordable cells, chorionic villous mesenchymal stem cells isolation and subsequent propagation are suitable 41 .
To understand the function of CV-MSCs in improving endothelial function, we performed the RNA-seq analysis in CV-MSCs treated HUVEC. Then, ABCA9 was found to be signi cantly increased in CV-MSC CM treated HUVECs, which also shows lower expression in placental tissue from PE pregnancies. Moreover, activation of Akt signaling induced by ABCA9 overexpression may contribute to the increased angiogenesis and proliferation of HUVEC mediated by CV-MSCs.
The A-subclass of ATP-binding cassette (ABC) transporters is a well-protected family member of potent lipid transporters 42 . The roles of the ABCA6-like subgroup transporters are still unknown 43 . And they have been identi ed as a unique gene cluster on human chromosome 17q24. Several research revealed that ABCA9 mRNAs could be isolated from vascular endothelial cells and placenta. This is possible to indicate the importance of ABCA9 in regulating vascular endothelial cells function [43][44][45] . Our results con rmed that ABCA9 could be regulated by CV-MSC and take part in the vascular endothelial cell proliferation promoting effects induced by CV-MSCs. ABCA9 overexpression in HUVEC signi cantly increased the proliferation and angiogenesis with an activation of AKT signaling.
Many studies have con rmed that mesenchymal stem cells could enhance endothelial cells function through AKT activation 46 . AKT and Wnt signaling pathways are of great signi cance for the survival and angiogenesis of cells through modulating endothelial cell migration, proliferation, remodeling, vascular system maturation and vascular sprouting [47][48][49] . In this study, the elevated protein levels of p-GSK3β and p-AKT are probably associated with the increased cell proliferation in ABCA9 overexpressed HUVECs, and the outcome is consistent with previous ndings in many functional endothelial cells 50 .
In summary, it is revealed in our study that ABCA9 activation induced by CV-MSCs promoted the HUVEC proliferation and angiogenesis in vitro, partially due to activation of AKT pathways in HUVECs. However, whether CV-MSCs affect other placenta-based cells was still unknown; therefore, further studies are needed on whether CV-MSCs have similar effects on them. Further studies should be performed to understand to how ABCA9 regulates AKT pathways, which will be important for a thorough understanding of the PE vessel pathogenesis.

Declarations
Ethics approval and consent to participate The study was approved by the Ethics Committee of the Qingdao University A liated Hospital. All participants provided a written informed consent.

Consent for publication
Not applicable.

Availability of supporting data
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare no potential con icts of interest.