Chorionic Villous Mesenchymal Stem Cells Derived Exosomes Deliver miR-135b-5p to Trophoblasts, Promoting their Proliferation and Invasion by Targeting TXNIP via β-Catenin Pathway

Background: Human chorionic villous mesenchymal stem cells (CV-MSCs) are found to be a promising and effective treatment for tissue injury. Trophoblast dysfunction during pregnancies is signicantly involved in the pathogenesis of preeclampsia (PE). This work was to understand how CV-MSCs regulated trophoblast function. Methods: In this study, we treated trophoblasts with CV-MSC-derived exosomes and RNA-seq analysis was used to understand the changes in trophoblasts. We examined the levels of TXNIP and β-catenin in trophoblasts by immunohistochemistry, western blot and qRT-PCR assays. Luciferase reporter assays and qRT-PCR assays were used to understand the role of miR135b-5p in the effects of CV-MSC-derived exosomes. The growth and invasion of trophoblasts was evaluated with the CCK-8 and transwell assays. Results: The treatment markedly enhanced the trophoblast proliferation and invasion. Furthermore, a signicant decrease of TXNIP expression and inactivation of the β-catenin pathway in CV-MSCs exosomes-treated trophoblasts was observed. Consistent with these ndings, TXNIP inhibition exhibited the same effect of promoting trophoblast proliferation and invasion as induced by CV-MSC-derived exosomes, also with the accompaniment of inactivation of β-catenin pathway. In addition, overexpression of TXNIP activated the β-catenin pathway in trophoblasts, and reduced the proliferation and invasion of trophoblasts. Importantly, miR135b-5p was found to be highly expressed in CV-MSC exosomes and interact with TXNIP. The miR-135b-5p overexpression signicantly elevated the proliferation and invasion of trophoblasts, which could be attenuated by TXNIP overexpression. Conclusion: Our results suggest that TXNIP-dependent β-catenin pathway inactivation mediated by miR135b-5p which is delivered by CV-MSC-derived exosomes could promote the proliferation and invasion of trophoblasts. migration/differentiation, . non-coding are for various cellular . Multiple miRNA placentas abnormality placental trophoblasts hypoxic


Introduction
In humans, aberrant development of placenta could result into pregnancy complications, including preeclampsia (PE) 1 . Trophoblast invasion into the decidua for accessing maternal blood supply is crucial for fetal development 2 . The placenta in PE patients is subjected to an excess of oxidative stress and in ammation, inducing anomalous differentiation of trophoblasts and a signi cant increase of antiangiogenic protein secretion 3 . PE proceeds from products released from the placenta, with inadequate perfusion and high oxidative stress, into the maternal blood, stimulating a systemic disorder in endothelia 4 . Furthermore, the clinical outcome of PE depends on the extent and the number of arteries invaded by the defective trophoblasts 5 . Since the exact etiology of PE is unclear, understanding the mechanisms in the proliferation and invasion of trophoblasts are important for PE treatment.
Mesenchymal stem cells (MSCs) feature immunomodulatory, anti-in ammatory, angiogenic, and antiapoptotic effects in tissue repairing 6 . Fetus or placenta-derived MSCs exhibit ontological characteristics that could consign more advantages, in comparison with their adult equivalents 7,8 . CV-MSCs provide a distinctive range of mediators which can impose regulation on the main functions in other cells 9 . These characteristics render CV-MSCs promising candidates for cell-based treatment 10 . Exosomes are homogeneous membrane vesicles secreted by various cell types which confer different attributes on exosomes 11 . Exosomes carry complicated and characteristic components, including proteins, lipids, mRNAs, and miRNAs 12 . Several studies reveal that exosomes participate in the cellular communication, immune response, antigen presentation as well as cellular migration/differentiation, affecting processes both in physiology and pathology 13 . miRNAs are a fragment of endogenous noncoding RNAs which are responsible for various cellular regulations 14 . Multiple miRNA expressions were observed in human placentas 15 and abnormality of miRNA expression was perceived in the placentas of patients with pregnancy complications, including PE 16 . Changes in the expression of various miRNAs were also demonstrated in placental trophoblasts in hypoxic conditions 17 .
Thioredoxin-interacting protein (TXNIP) endogenously inhibits the thioredoxin system and plays a key role in redox homeostasis 18,19 . It can regulate different biological processes, as cell growth and angiogenesis in nucleus, and modulation of apoptosis signal-regulating kinase-1 (ASK-1) and cell membrane receptor signaling in cytosol 20,21 . TXNIP elevates the level of reactive oxygen species (ROS) and contributes to oxidative stress and apoptosis 22 . TXNIP can also regulate redox/glucose-mediated stress and in ammation, and is reported to overexpress in ischemia-hypoxia/reoxygenation and irradiation injuries in many organs, thus providing a promising therapeutic target [23][24][25] .
In this study, we studied the effects of CV-MSC-derived exosomes on the proliferative and invasive capacities of the trophoblast cell lines, JAR and HTR-8/SVneoc. Transcriptome analyses manifested considerable downregulation of TXNIP and upregulation of WNT signalling in trophoblasts. In addition, the expression of microRNA 135b-5p was increased in trophoblasts treated with CV-MSC-derived exosomes, and affected the TXNIP expression. MiR-135b-5p putatively served as an upstream regulator of proliferation and invasion of trophoblasts through TXNIP and β-catenin regulation.

Methods And Materials
Cell culture Placentas were acquired from pregnancies at full term through caesarean delivery with parental permission. Every procedure was performed according to ethical requirement by the A liated Hospital of Qingdao University, China. CV-MSCs used were between passages 3 to 6. Brie y, the placental fetal part was sliced into roughly 1 mm 3 pieces and washed by PBS. Tissue were digested by trypsin (0.25%) and collagenase I (0.1%) (Sigma-Aldrich, St. Louis, MO), followed by incubation at 37°C for half an hour. Following nylon lter ltration and centrifugation, the cells were seeded in a stem cell culture medium (SCCM), supplemented with Stem Cell Basic Medium (Dakewe Biotech Co., Guangzhou, China) and 5% UltraGRO TM (Helios, USA). JAR and HTR-8/SVneo cell lines were procured from Type Culture Collection China Centre. DMEM/F12 (10% FBS) was used for cell incubation in a 5% CO 2 atmosphere at 37°C.
Since CV-MSCs from placentas of normal pregnancies or placentas with severe PE could differentiate into osteoblasts as well as adipocytes, we assessed their differentiation capability. CV-MSCs were cultured in 6-well plates until reaching a con uency between 70% and 80%. Subsequently, the CV-MSCs were transferred into osteogenic or adipogenic differentiation medium (Gibco, Carlsbad, CA) for 3 weeks of incubation. Then CV-MSCs were stained by Alizarin red S for verifying differentiation of osteoblasts, followed by oil red O staining for adipocyte differentiation.

Exosome isolation
Epidural CV-MSC supernatants were differentially centrifuged to obtain exosomes. The medium was disposed of after attaining 70% con uency. DMEM/F12 (serum-free) was then adopted to culture cells for an additional day. The supernatants were pooled and sequentially centrifuged at 15,000 x g for half an hour, followed by ltration through lters (0.22 mm) (Millipore, Billerica, MA) and ultracentrifuge at 120,000 x g. The exosomes were rinsed by sterile PBS and collected. Pierce BCA Assay Kit (Thermo Fisher) was utilized to detect exosome concentration. Nanoparticle tracking analysis was used for determining the pro le of size distribution and measurement of concentration through a particle tracker (ParticleMetrix, Germany).

Electron microscopy
During roughly ten mins, nearly 50 μl of exosome preparation was taken up by absorption in 300-mesh formvar/carbon coated grids. The grids were then dried at ambient temperature for half an hour and transferred to 3% phosphotungstic acid solution for negative staining. The exosomes were visualized by using an Olympus transmission electron microscope at 120.0 kV. Exosome images were taken using a CCD camera.
Exosome uptake CV-MSC-derived exosomes were tagged by the PKH67 dye (Sigma-Aldrich) in accordance with the slightly modi ed protocol provided by the manufacturer. In brief, exosomes were subjected to suspension with 250 μL PBS and blended with 750 μL diluted PKH67 (1/50) (for cytomembrane labeling, in Diluent C), followed by 5 min of incubation at ambient temperature. After that, 2 mL BSA was used to end the labeling process. The isolation of labeled exosomes was performed as previously speci ed, followed by suspension in 1×PBS.
Protein isolation, western blot analysis and qRT-PCR Protein isolation, western blot analysis and qRT-PCR were carried out as previously described 26 .
Cell proliferation analysis Cells (5×10 3 /well) were seeded in the 96-well plates. The CCK-8 (Thermo Fisher Scienti c, MA, USA) was added to evaluate cell viability each day as per the protocol. The OD value at 450nm was detected through a microplate reader. Cell growth curves were carried out in triplicate.

Trophoblast invasion assay
Two trophoblast lines (5×10 5 cells) were plated into transwell membrane (Corning, NY, USA). The pore membranes were covered by Matrigel™ matrix dilution (1:10) . The trophoblasts were then incubated for 12 h. Medium (with 10% FBS) treated with or without CV-MSCs-exosomes (control) was added into the lower layer. Transwell assays were conducted while trophoblasts were cultured without FBS for 24 h. Subsequently, cell xation was performed using 4% paraformaldehyde. Cells having traversed the membrane were quanti ed. Each experiment was conducted for at least 3 independent times.
The reporter plasmids containing wild-type (TOP ash) or mutated (FOP ash) TCF/LEF DNA binding sites were purchased from Addgene (Addgene, Cambridge, USA) and Renilla reporter constructs were from Promega (Promega; Madison, USA). Luciferase reporters and Renilla reporters were co-transfected into cells with Lipofectamine 3000 (Thermo Fisher Scienti c). Luciferase reporter activity was normalized against Renilla activity for 24 h.
Dual-Luciferase Reporter Assay Kit (Promega) was used to detect luciferase activity.

Immuno uorescence
Trophoblasts were treated with or without puri ed CV-MSCs-exosomes (10 μg/ml), and then cultured for 24 h. They were then collected, separated and settled in 4% paraformaldehyde for 60 min. The pellets were desected into 4-μm thick sections and placed in para n. Later, the sections were bathed thrice in PBS, closed with 10% goat serum for 1 h and subsequently bathed twice in 0.2% Triton X-100. Then, primary antibodies (TXNIP and β-catenin from Abcam, MA, USA), secondary antibodies (Invitrogen) and DAPI (Guangzhou RiboBio, Guangzhou, China) were used for cell incubation. And the images were take through a uorescence microscope.

Placental explant culture
After being collected within 10 mins, all placentas were then processed in 30 min and observed carefully for abnormalities. Following thorough rinse in PBS thrice to remove remaining blood, the villous tissue was cut into small bits of tissue (8 mm3). DMEM/F12 (4 ml per well) supplemented with 1% penicillin/streptomycin and amphotericin B (Gibco, Carlsbad, CA) was adopted for culturing the placental explants in an incubator at 37 °C at an oxygen concentration of 2% for 2 days. Following treatment with CV-MSC-exosomes for 24 h, the explants were rinsed with PBS and then placed in liquid nitrogen.

Statistical Analysis
Two-tailed Student's t-test was adopted for data analysis. Each experiment was conducted independently in triplicate and all data were given as mean ± SEM. Prism (Ver. 7) (GraphPad; La Jolla, USA) was used to analyze data. P values less than 0.05 suggested statistical signi cance and were denoted by "*"; P values between 0.05 and 0.01 were denoted by "**"; P values between 0.01 and 0.001 were denoted by "***"; P > 0.05 denoted no signi cance (n.s.).

Cell proliferation and invasion were promoted and TXNIP expression was inhibited in trophoblasts by CV-MSCexosomes
First, we performed the isolation of CV-MSCs from healthy placentas, and determined their multidirectional differentiation ability and surface marker expression (Fig. S1a-b). Then, we identi ed the exosomes isolated from CV-MSCs and observed the exosome endocytosis in trophoblasts (Fig. S2a-d). Next, we used the CCK-8 and transwell assays to examine the proliferation and invasion of trophoblasts treated with or without CV-MSC-derived exosomes. The trophoblasts treated with CV-MSC-derived exosomes displayed signi cantly higher proliferation rates and invasion abilities than those untreated cells ( Fig. 1a-b, all P < 0.01). To further verify the role of TXNIP in effects of CV-MSCs-exosomes on promoting trophoblasts proliferation and invasion, we examined the expression of TXNIP in placental explant cultures by immunohistochemical analysis. The CV-MSCs-exosome treated placental explants tissues exhibited signi cantly higher expression of Ki67 than the untreated explants, and the TXNIP expression was decreased after exosome treatment (Fig. 1c). Furthermore, to elucidate the possible molecular mechanisms of CV-MSC-derived exosomes, differentially expressed genes (DEGs) in exosometreated JAR cells were investigated using RNA sequencing, and the molecular mechanism of underlying functional changes in JAR was explored. We compared the transcriptomes between JAR cells co-cultured with or without CV-MSCs-exosomes for 48 h, and 2397 upregulated and 2014 downregulated genes were found to change more than two-fold (Fig. 1d). Based on the results, TXNIP was found to be downregulated signi cantly in JAR treated with CV-MSCs-derived exosomes, which was veri ed by qRT-PCR and western blot assays (Fig. 1e-f).
These results indicated that the CV-MSCs-exosomes promoted trophoblast proliferation and invasion, and TXNIP was downregulated in this process.

Txnip Inhibited The Proliferation And Invasion Of Trophoblasts
To better understand the role TXNIP played in trophoblasts after CV-MSC-derived exosome treatment, JAR cells were transfected with TXNIP shRNA and HTR-8/SVneoc cells were transfected with TXNIP overexpression lentiviruses. The mRNA levels of TXNIP were examined, and signi cant decrease and increase of TXNIP mRNA levels were respectively observed in transfected JAR and HTR-8/SVneoc cells (Fig. 2a). In consistence with the changes of mRNA levels, the TXNIP protein levels were also signi cantly changed in transfected trophoblasts compared to the control trophoblasts (Fig. 2b). Moreover, the growth of JAR cells was signi cantly promoted in the presence of TXNIP shRNA, while the TXNIP overexpression hindered HTR-8/SVneoc cells proliferation (Fig. 2c). In addition, we examined the invasion of transfected trophoblasts, and found that the cellular invasion changes in the corresponding trophoblasts shared the same curve as observed changes in proliferation (Fig. 2d). Our results suggested that TXNIP could be inhibited by CV-MSCs-exosomes which regulated the proliferation and invasion of trophoblasts.

Txnip Suppressed β-catenin Pathway Activation In Trophoblasts
Many researches have revealed the WNT/β-catenin pathway contributes vital roles in intracellular metabolism regulation, suggesting that it is involved in TXNIP associated trophoblasts regulation 27,28 . Moreover, Gene Set Enrichment Analysis (GSEA) of trophoblasts revealed the WNT/β-catenin pathway was among the most signi cantly upregulated pathways in trophoblasts treated with CV-MSCsexosomes (Fig. 3a). As a primary member in the typical WNT pathway, β-catenin is coordinated by a complex mesh of factors during WNT signal activation in the nucleus 29 . Therefore, we used TOP/FOP ash assays to investigate the effect of TXNIP on β-catenin and TCF-dependent transcriptional activity to identify the role of TXNIP in the typical WNT signaling. The TOP/FOP ash reporter activities were more signi cantly increased in shTXNIP JAR cells, than in the scrambled control cells. The reporter activity changes in TXNIP overexpressed HTR-8/SVneoc cells coincided with the change pattern of JAR cells (Fig. 3b). Moreover, the β-catenin and TXNIP expressions were evaluated by immuno uorescence and western blot. The relative β-catenin expression levels were signi cantly upregulated in the TXNIP-silenced JAR cells and downregulated in the TXNIP overexpressed HTR-8/SVneoc cells, compared to the scrambled control (Fig. 3c-d). The results suggested the possible involvement of TXNIP in modulating the β-catenin pathway by interacting directly with β-catenin in trophoblasts. Then, we investigated if the βcatenin pathway could be regulated by the TXNIP in trophoblasts.
MiR-135b-5p was highly expressed in exosomes and targets TXNIP in trophoblast cells On the premise of above results, we sought to understand the upstream regulator of TXNIP in CV-MSCsexosomes. Since small non-coding RNAs could regulate expressions of genes and contribute critical roles in various process of physiological developmental, such as placental growth and function, potential regulatory miRNAs at the upstream of TXNIP were screened using bioinformatics tools. We selected miR-135b for analysis over other miRNAs that might have interaction with TXNIP because miR-135b had been found to promote proliferation and invasion in many types of cancers 30,31 . The region of TXNIP bound with miR-135b-5p was found to be located in the TXNIP 3' UTR (Fig. 4a). Subsequently, RT-PCR assays revealed a higher miR-135b-5p level in CV-MSCs compared with CV-MSCs-exosomes (Fig. 4b).
Furthermore, the miR-135b-5p level was signi cantly increased in trophoblasts treated with CV-MSCsexosomes in comparison with those treated with PBS (Fig. 4c).
To explore the role miR-135b-5p played in TXNIP regulation of trophoblasts, the trophoblasts were treated with miR-135b-5p mimics or control microRNA. The miR-135b-5p levels were then evaluated by RT-PCR assays, and high levels of miR-135b-5p were con rmed in JAR and HTR8/SVneo cells (Fig. 4d). Subsequently, luciferase assay was used to identify the interplay between the miR-135b and TXNIP 3' UTR. The entire length of TXNIP 3' UTR was synthesized and was then inserted into a luciferase reporter plasmid, and miRNA mimics (miR-mimics) or scrambled negative control RNAs (miR-NC) were cotransfected. It revealed a signi cant reduction of luciferase activity in cells with miR-135b mimics (miRmimics) transfection, in comparison with those with control RNA transfection (Fig. 4e). In addition, the western blot assays con rmed the miR-135b-5p mimics-treated trophoblasts expressed higher levels of βcatenin protein and lower levels of TXNIP protein (Fig. 4f). These results indicate that miR-135b-5p may be delivered by CV-MSCs-exosomes and can directly regulate TXNIP expression in trophoblasts.

Mir-135b-5p Advances Trophoblasts Proliferation And Invasion Via Txnip-β-catenin
Next, to determine the role miR-135b-5p played in trophoblasts proliferation and invasion, lentiviral vector carrying miR-135b-5p were constructed and trophoblasts with stable miR-135b-5p expression were transfected with TXNIP-overexpression lentiviruses or vectors. Then the TXNIP overexpression was con rmed by the western blot (Fig. 5a). Furthermore, the luciferase activity raised by miR-135b increment was signi cantly decreased in cells transfected with TXNIP-overexpressed lentiviruses compared to cells transfected with vectors (Fig. 5b), indicating the effectiveness of our stably transfected constructs. Therefore, we examined the effect of TXNIP overexpression on miR-135b-mediated trophoblasts proliferation and invasion in vitro. The proliferation and invasion of trophoblasts were then determined by CCK-8 and transwell assays. The miR-135b-5p overexpression markedly promoted proliferation and invasion of trophoblasts, which could be attenuated by TXNIP overexpression (Fig. 5c-d). These data con rmed that miR-135b-5p promoted trophoblasts proliferation and invasion through TXNIP inhibition.

Discussion
The placenta maintains its functions mainly through trophoblasts. Dysfunctional trophoblast invasion can ultimately cause the PE, a great obstetric syndrome. MSCs are progenitor cells with capability if selfrenewal, and numerous studies have revealed the proangiogenic potentiality of MSCs deriving from their regenerating and paracrine activities. MSC-derived treatment turns up as one of most promising therapies for ischemic diseases. Moreover, it has recently been revealed that the therapeutic effects of MSC-derived exosomes on ischemic cardiac disease, renal impairment, and wound are analogous to the functional effect of their parental cells, indicating active roles of MSC-derived exosomes. Nevertheless, the underlying mechanisms in MSC-derived exosomes mediating these therapeutic effects remain unclear. In this study, we evaluated the roles of foetal CV-MSCs in the proliferative and invasive capacity of trophoblasts, and provide evidences that CV-MSC-derived exosomes mediate inhibition of TXNIP expression and β-catenin pathway, which is also regulated by the miR-135b-5p in exosomes.
Exosomes, containing various subtypes, both participate in regular biological processes and in pathological disorders. Notwithstanding that much remains unknown regarding their functions, more and more studies have revealed that exosomes function as crucial intercellular message transmitters. Exosomes can directly interact with their target cells, which accounts for exosomes' involvement in mediating intercellular signaling. The capability for packaging and transporting miRNAs enables exosomes to participate in intercellular communication. Exosomes are carriers of various miRNAs, are bound to particular targeted cytomembrane receptors, and participate in certain regulatory pathways to regulate diverse physiological and pathological processes. Recently, effects of exosomes on immunotherapy have increasingly been uncovered.
Many researches have demonstrated miRNAs show differential expression between normal placentas and placentas of PE patients, suggesting a signi cant role by miRNAs in physiological processes in placentas. Despite limited information on miRNA regulating functional development in placentas, multiple studies indicate that PE-related differentially expressed miRNAs contribute vital roles in the regulation of placental development and function through targeting functional genes. It is known that miRNA inhibits protein expression mainly through inhibiting translation. The miRNA carried by CV-MSCsexosomes could be an upstream regulator of trophoblast function. Shunsuke et al. found that miR-135b contributed vital roles in regulating the trophoblast invasion in hypoxic conditions through direct regulation of CXCL12 expression 32 . Meanwhile, miR-135b was also identi ed as being abnormally expressed in PE pregnancies, indicating its possible involvement in PE pathology 33 . In our study, this wellknown tumor-associated microRNA was found to interact with TXNIP and overexpressed in CV-MSCsexosomes, implying an intersection between CV-MSCs-exosomes and TXNIP in trophoblasts through miR-135b-5p.
TXNIP has been reported to be capable of binding to and activating the NLR family pyrin domain containing 3 (NLRP3) in ammasome, and contribute a key role in immuno-in ammatory modulation. TXNIP is signi cantly increased in oxygen-de cient trophoblasts, which can be noticed in placental tissue from early-onset PE pregnancies. Moreover, abnormal activation of Wnt/β-catenin may contribute to the development of PE by dysregulating the invasion and proliferation of trophoblast 34 . In our CV-MSCs-exosomes treated trophoblasts RNA-seq analysis, TXNIP and WNT/β-catenin pathways were substantially downregulated. Based on these ndings, we propose that TXNIP/β-catenin pathways could be massively involved in mediating the in ammatory dysregulation in PE, which could be reversed by CV-MSCs-exosomes miRNA.

Conclusion
Our study revealed that miR-135b-5p carried by CV-MSCs derived-exosomes promoted the trophoblast proliferation and invasion in vitro, partially due to inactivation of β-catenin pathways induced by the TXNIP downregulation. However, it remains undetermined how CV-MSCs affect other placenta-based cells in vitro or even in vivo; therefore, further studies are needed on whether CV-MSCs have similar effects on them. More studies with respect to how miR-315b-5p regulates TXNIP and β-catenin pathways will be important for a thorough understanding of the PE 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.

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