Human bone marrow mesenchymal stem cell‐derived extracellular vesicles impede the progression of cervical cancer via the miR‐144‐3p/CEP55 pathway

Abstract Cervical cancer is the most common gynaecological malignancy, with a high incidence rate and mortality rate in middle‐aged women. Human bone marrow mesenchymal stem cells (hBMSCs) have been implicated in the initiation and subsequent development of cancer, along with the involvement of extracellular vesicles (EVs) mediating intracellular communication by delivering microRNAs (miRNAs or miRs). This study is aimed at investigating the physiological mechanisms by which EVs‐encapsulated miR‐144‐3p derived from hBMSCs might mediate the progression of cervical cancer. The expression profiles of centrosomal protein, 55 Kd (CEP55) and miR‐144‐3p in cervical cancer cell lines and tissues, were quantified by RT‐qPCR and Western blot analysis. The binding affinity between miR‐144‐3p and CEP55 was identified using in silico analysis and luciferase activity determination. Cervical cancer cells were co‐cultured with EVs derived from hBMSCs that were treated with either miR‐144‐3p mimic or miR‐144‐3p inhibitor. Cervical cancer cell proliferation, invasion, migration and apoptosis were detected in vitro. The effects of hBMSCs‐miR‐144‐3p on tumour growth were also investigated in vivo. miR‐144‐3p was down‐regulated, whereas CEP55 was up‐regulated in cervical cancer cell lines and tissues. CEP55 was targeted by miR‐144‐3p, which suppressed cervical cancer cell proliferation, invasion and migration and promoted apoptosis via CEP55. Furthermore, similar results were obtained by hBMSCs‐derived EVs carrying miR‐144‐3p. In vivo assays confirmed the tumour‐suppressive effects of miR‐144‐3p in hBMSCs‐derived EVs on cervical cancer. Collectively, hBMSCs‐derived EVs‐loaded miR‐144‐3p impedes the development and progression of cervical cancer through target inhibition of CEP55, therefore providing us with a potential therapeutic target for treating cervical cancer.


| INTRODUC TI ON
Cervical cancer, a malignancy that arises from chronic human papillomavirus (HPV) infections, ranks as the 4th most commonly diagnosed cancer among females and the 4th leading cause of cancer-related deaths with an estimated 570 000 cases and 311 000 deaths reported across the world in 2018. 1 Current available primary treatment approaches for patients with cervical cancer include surgery, or a concurrent chemoradiotherapy regimen supplemented with cisplatin-based chemotherapy, beam radiotherapy and brachytherapy. 2 The disease is not considered refractory yet the oncologic outcome of patients with recurrent and metastatic cervical cancer is still below par, despite recent achievements regarding therapeutic strategies, 3 thus highlighting the need for developing more effective approaches against the progression of cervical cancer.

Microarray and databases conducted by both Yi et al and Koch
et al have revealed that the centrosomal protein, 55 Kd (CEP55), is a clinically relevant biomarker for cervical cancer. 4,5 A functional report has demonstrated that CEP55 has the ability to reflect and indicate unfavourable clinical prognosis of patients suffering from cervical cancer, 6 whereas the specific mechanism governing the action of CEP55 still requires further study. Intriguingly, bioinformatics analysis prior to our investigation proved that microRNA-144-3p (miR-144-3p) was a putative upstream regulatory miRNA for CEP55.
Concordantly, miR-144-3p has been elucidated to inhibit cancer cell proliferation and promote apoptosis by targeting CEP55 in the context of prostate cancer. 7,8 miR-144-3p has been identified as one of the down-regulated miRNAs in serum of patients with negative HPV16. 9 The tumour-suppressive action of miR-144-3p in cervical cancer has also been reported, 10 whereas the underlying mechanism still remains enigmatic. Notably, miR-144-3p has been detected to be abundant in extracellular vesicles (EVs) derived from mesenchymal stem cells (MSCs) in association with cell growth regulation. 11 Multiple types of cancer cells constitute tumours where MSCs, a particular population of cancer stem cells, particularly exhibit proor antitumorigenic influences on cancerogenesis. 12,13 Bone marrow-derived MSCs (BMSCs) have been described as 'magic bullets' in the suppression of tumour progression, regarding their capabilities of differentiation. 14 The paracrine functions of MSCs have been found to be partially mediated by EVs, which can shuttle miRNAs, messenger RNAs (mRNAs) and proteins involved in cell-to-cell communication. All of this helps suggest the promising application of MSCs-derived EVs in mediation of cancer progression. 15,16 Although the role of miR-144-3p and CEP55 in cervical cancer has already been investigated, the mechanism by which EV communication affects cervical cancer cells involving the interplay between miR-144-3p and CEP55 is still poorly understood, highlighting a major gap in knowledge given that MSCs-derived EVs may be of significance to the development and progression of cervical cancer. Hence, we have been suggested that the transfer of miR-144-3p via BMSCsderived EVs might alter the biology of recipient cervical cancer cells in mediating the development and progression of cervical cancer.

| Ethics statement
The study was conducted with the approval of the Ethics Committee of Shandong Medical College and was performed in strict accordance with the Declaration of Helsinki. Each participant signed a written informed consent prior to the study. Animal experiments were strictly designed and conducted according to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health. Extensive efforts were made to ensure minimal suffering of the animals used during the study.

| Sample collection
We collected cancer and adjacent normal tissue samples from 60 patients (aged: 10-58 years, with a mean age of 21.63 ± 9.70 years) with cervical cancer at the Shandong Medical College, from April 2010 to April 2013. None of these patients received either chemotherapy or radiotherapy prior to their operation. Meanwhile, 15 mL of bone marrow samples was extracted from healthy volunteers.

| Isolation and identification of human BMSCs (hBMSCs)
hBMSCs were isolated from the donor bone marrow as previously

| Detection of hBMSC differentiation potential in vitro
The hBMSCs at passage 3 were detached with the concentration adjusted to 5 × 10 4 cells/mL, followed by seeding into a 6-well plate.
The cells were allowed to grow and adhere to the coverslip, after 24 hours of seeding. Subsequently, hBMSCs were cultured in the medium for OriCell™ MSC osteogenesis, lipogenesis or cartilage differentiation (all purchased from Cyagen Company, Guangzhou, China) according to the manufacturer's instructions, followed by identification of cells using Alizarin Red S staining, oil red O staining and Alcian blue staining, respectively. 18

| Plasmid transfection and lentiviral transduction
Plasmid transfection was performed according to steps described in the manual of Lipofectamine 2000 reagents (11668-019, Invitrogen, Carlsbad, WI, USA). Briefly, cells were added with miR-144-3p/ NC mimic, miR-144-3p/NC inhibitor and sh-CEP55/NC (a final concentration of 50 nmol\L) for incubation at room temperature for 5 minutes. Another 250 μL Opti-MEM was used to dilute 5 μL Lipofectamine 2000 and was left to incubate at room temperature for 5 minutes. The above two were mixed and allowed to settle at room temperature for another 20 minutes. The mixture was then added into the cell culture plate. After culture, subsequent experiments were conducted.
LV3-hsa-miR-144-3p and package plasmids were co-transfected into 293T cells using liposomes to construct experimental lentivirus, whereas LV3-NC and package plasmids were co-transfected into 293T cells using liposomes to construct NC lentivirus. The lentiviruses of LV3-hsa-miR-144-3p and LV3-NC were obtained, following incubation in a 5% CO 2 incubator at 37℃. The hBMSCs in the logarithmic growth phase were seeded into a 6-well plate at a density of 1 × 10 6 cells/mL and cultured for 48 hours in a 5% CO 2 incubator at 37℃. The cells were subsequently screened using puromycin for at least one week to select stably transfected cell lines.

| Dual-luciferase reporter gene assay
The target genes of miR-144-3p were predicted on microRNA.org, the results of which were further verified by dual-luciferase reporter gene assay. CEP55 3' untranslated region (3'UTR) gene fragment was artificially synthesized and introduced into pGL3-control (Promega, Madison, WI, USA) using enzyme sites. The mutant (MUT) sites of complementary sequence of seed sequence were designed on the CEP55-wild-type (WT). Target fragments were inserted into pGL3-control vector using T4 DNA ligase after restriction enzyme digestion. The luciferase reporter plasmids WT and MUT were cotransfected with miR-144-3p mimic into H293T cells (Cell Bank of Shanghai Institutes for Biological Sciences), respectively. The luciferase activity was then measured.

| Co-culture of hBMSCs and cervical cancer cells
The hBMSCs were detached with 0.25% trypsin and cultured in lowglucose DMEM to a cell concentration of 1.0 × 10 6 cells/mL. Cervical cancer cells were labelled with pCDNA3.1-GFP, whereas hBMSCs were transduced with miR-144-3p-Cy3 (GenePharma, Shanghai, China). They were subsequently collected and mixed at a ratio of 1:1 after 12 hours of transfection and then seeded in a 96-well plate (100 cells/well). Co-culture was maintained for 2 days before flow cytometry-based separation.

| Isolation and characterization of EVs
The EVs in the serum were isolated and removed through the ultracentrifugation of medium/serum at 100,000 × g overnight at 4℃.
The hBMSCs in viable growth conditions were cultured in a EVsdepleted serum. The supernatant of hBMSCs at passage 3 was centrifuged at 500 × g for 15 minutes at 4℃ to remove cellular debris.
The cellular debris or apoptotic bodies were further removed by centrifugation at 2000 × g for 15 minutes at 4℃, whereas large vesicles were removed by centrifugation at 10 000 × g for 20 minutes at 4℃. After the resulting supernatant was filtered through a 0.22-µm filter and spun at 110 000 × g for 70 minutes at 4℃, hBMSCs were resuspended in PBS and subjected to another round of centrifugation (110 000 × g, 70 minutes, 4℃), followed by resuspension in 100 µL sterile PBS for subsequent experiments. The size distribution of the isolated EVs was measured using a nanoparticle tracking analysis (NTA) instrument, following the steps described in the manual. Next, 10 µL of isolated and purified EVs was left to counterstain with 3% (w/v) sodium phosphotungstate solution for 1 minute at room temperature. EVs were then observed under a transmission electron microscope (TEM).

| Isolation and quantification of RNA
A reverse transcription quantitative polymerase chain reaction (RT-qPCR) was employed to quantify the expression of miR-144-3p and mRNA expression of CEP55. Based on the related sequences provided by GenBank, the primer sequences were designed by Primer 5 software as shown in Table 1. The primers of downstream target genes were analysed by a homology analysis using BLAST software. The cells were collected, and the total RNA was extracted using a Trizol kit (Invitrogen) after transfection.
Then, total RNA was reversely transcribed into complementary DNA (cDNA) according to the instructions provided by TaqMan MicroRNA Assays Reverse Transcription Primer (4427975, Applied Biosystems, Foster City, CA, USA), followed by PCR amplification.
The relative expression levels of miR-144-3p and CEP55 were analysed by the 2 -ΔΔCt method with U6 serving as the loading control for miR-144-3p and β-actin for CEP55.

| Protein preparation and Western blot analysis
The total protein content was extracted from cells, and the protein concentration was determined by a bicinchoninic acid kit (20201ES76, Yeasen Biotech Co., Ltd., Shanghai, China). Proteins were loaded into 20 µg-sized wells, allowed to separate by 8% SDS-PAGE for 1 hour and blocked in 5% skimmed milk. The resulting protein was subsequently probed with diluted primary antibodies (Abcam) to β-actin

F I G U R E 1
The significance of CEP55 in cervical cancer. A, B, Heat maps of significantly up-regulated genes in cervical cancer-related microarray data sets. The abscissa refers to the sample number, the ordinate refers to the genes, the left dendrogram represents the expression level, each cube represents the expression level of a gene in one sample, and the upper right histogram represents the colour scale. C, Venn analysis of significantly up-regulated genes in cervical cancer-related microarray data sets. Two circles show the significantly up-regulated genes in cervical cancer samples in two cervical cancer-related microarray data sets, and the middle part indicates the intersection of two sets of data. D, The interaction network of significantly up-regulated genes in cervical cancer samples. Each circle represents a gene and the line between circles indicates the presence of interaction between two genes. E, The expression of CEP55 in cervical cancer samples (the left box plot) and normal samples (the right box plot) according to TCGA database. The abscissa refers to the sample type, and the ordinate refers to the expression level. F, Expression of CEP55 was determined by RT-qPCR in cervical cancer and adjacent normal tissues, relative to GAPDH. The left image represents the detection results of CEP55 expression in cancer tissues, and the right represents the detection results of CEP55 expression in adjacent normal tissues. G, Representative Western blots of CEP55 protein and its quantitation (on the right side) in cervical cancer and adjacent normal tissues, relative to GAPDH. Data between cervical cancer tissues and adjacent normal tissues were compared with paired t test
The apical chambers were added with 200 µL cells, whereas the basolateral chambers were added with 300 μL DMEM (Invitrogen)

| 5-ethynyl-2'-deoxyuridine (EdU) assay
Cells were left to culture in 50 µmol\L of EdU (EdU labelling/detection kit, Ribobio Co., Ltd., Guangzhou, Guangdong, China) for 12 hours, fixed with paraformaldehyde and incubated with antibodies to EdU. Next, 5 µg/mL Hoechst 33 342 was applied to stain cells for 30 minutes. The results were observed under a fluorescence microscope.

| Colony formation assay
At 48 hours post-cell transfection, the 6-well plate was coated with Roswell Park Memorial Institute (RPMI) 1640 medium and allowed to stand. When solidified, the plate was added with cell suspension (0.5 mL, 2 × 10 3 cells/mL) and resuspended by a RPMI 1640 medium formulation (GE Health Care) containing 10% FBS and 0.3% agar.
The formation of cell colonies was observed after 21 days of culture at 37℃. The number of cell colonies was counted in 5 randomly selected fields in each group.

| Flow cytometry
After

| Xenograft tumour in nude mice
SiHa cells were dispersed into signal cell suspension after transfection. PBS was mixed with Matrigel (E1270, Sigma) at a ratio of

| Immunofluorescence
The paraffin samples were cut into sections, dewaxed and dehy-

| Terminal deoxynucleotidyl transferasemediated 2'-deoxyuridine 5'-triphosphate nick end labelling (TUNEL) staining
Five sections were dewaxed, hydrated and incubated with 50 μL 1% proteinase K dilution. The activity of endogenous peroxidase (POD) F I G U R E 2 High expression of CEP55 contributes to tumorigenesis of cervical cancer. A, mRNA and protein expression of CEP55 was determined by RT-qPCR and Western blot analysis in normal cervical epithelial cell line End1/E6E7, and cervical cancer cell lines HeLa, CaSki, SiHa and ME180, normalized to GAPDH. B, mRNA and protein expression of CEP55 was determined by RT-qPCR and Western blot analysis in SiHa cells, normalized to GAPDH. C, SiHa cell migration and invasion were detected by Transwell assay (scale bar = 50 µm). D, Colony formation of SiHa cells was detected by colony formation assay. E, SiHa cell proliferation was detected by EdU assay (scale bar = 25 µm). F, SiHa cell apoptosis was detected by flow cytometry. Data comparison was analysed by independent sample t test between two groups and by one-way ANOVA among multiple groups, followed by Tukey's test. *P < 0.05 versus the End1/E6E7 cell line or SiHa cells treated with NC. # P < 0.05 versus SiHa cells treated with sh-NC. Data are shown as the mean ± standard deviation of three technical replicates

| mRNA expression profiles in cervical cancer
At the initial stage of our study, cervical cancer-related microarray data sets GSE63678 and GSE9750 were retrieved from the GEO database, followed by differential analysis. The results revealed 105 and 100 significantly up-regulated genes in cervical cancer samples, respectively. A heat map displaying the top 50 significantly up-regulated genes in the two microarray data sets was plotted ( Figure 1A,   B). An intersection of prediction results available from two microarray data sets was obtained to further screen out cervical cancer-related genes ( Figure 1C). In total, 20 significantly up-regulated genes were detected in prediction results of both microarray data sets. An interaction network diagram was then developed for a correlation analysis ( Figure 1D). CEP55 was found at the core position in the interaction network diagram of the 20 significantly up-regulated genes, suggesting that CEP55 could probably interact with multiple up-regulated genes. Therefore, CEP55 might play an important role in the development of cervical cancer. The expression of CEP55 was subsequently quantified in cervical cancer samples and normal samples that were available from the TCGA database ( Figure 1E). The results indicated that CEP55 was significantly highly expressed in cervical cancer samples. For further verification, we collected cervical cancer tissues and adjacent normal tissues, followed by conducting RT-qPCR and Western blot analysis. The results revealed that the expression of CEP55 was much higher in cervical cancer tissues than that in adjacent normal tissues ( Figure 1F, G), which was just in line with bioinformatics analysis in microarray data sets related to cervical cancer.  Figure 2B).

| CEP55 was highly expressed in cervical cancer cell lines that contributed to the progression of cervical cancer
Subsequent gain-and loss-of-function assays were performed to evaluate cell migration and invasion by Transwell assay (Figure 2C), clone-forming ability by colony formation assay ( Figure 2D), proliferation by EdU assay ( Figure 2E) and apoptosis by flow cytometric analysis ( Figure 2F). The presence of shCEP55 corresponded to weakened cell migration, invasion, clone-forming ability and proliferation, along with strengthened cell apoptosis, whereas a contrary F I G U R E 3 High expression of miR-144-3p exerts inhibitory effects on cervical cancer tumorigenesis. A, Putative miRNAs that might regulate CEP55. The 4 ellipses indicate prediction results from 4 databases, and the middle part represents the intersection. B, Expression of miR-144-3p was determined by RT-qPCR in normal cervical epithelial cell line End1/E6E7, and cervical cancer cell lines, HeLa, CaSki, SiHa and ME180, (on the right side) as well as cervical cancer and adjacent normal tissues (on the left side), relative to U6. C, SiHa cell migration and invasion were detected by Transwell assay (scale bar = 50 µm). D, SiHa cell proliferation was detected by EdU assay (scale bar = 25 µm). E, Colony formation of SiHa cells was detected by colony formation assay. F, SiHa cell apoptosis was detected by flow cytometry. G, Representative Western blots of CDK4, Cyclin D1, E-cadherin, Vimentin, Bcl-2 and Bax proteins and their quantitation in SiHa cells, normalized to GAPDH. Data comparisons were analysed by an independent sample t test between two groups and by one-way ANOVA among multiple groups, followed by Tukey's test.

| miR-144-3p was poorly expressed in cervical cancer cell lines and tissues acting as an anti-oncomiR in vitro
To further explore the upstream regulatory mechanism of CEP55, miRNAs that could regulate CEP55 were predicted by databases, including DIANA. According to the intersection of prediction results of 4 databases, the only one miRNA found was miR-144-3p ( Figure 3A).

| hBMSCs were able to secret EVs
After hBMSCs were isolated at passage 3, the surface antigen was identified by flow cytometry using FITC-labelled mouse anti-human F I G U R E 6 miR-144-3p can be transferred into cervical cancer cells by EVs derived from hBMSCs. A, Expression of miR-144-3p was determined by RT-qPCR in cells, relative to U6. B, miR-144-3p-Cy3 in hBMSCs and cervical cancer cells under the fluorescence microscope (scale bar = 20 µm). C, Expressions of miR-144-3p and CEP55 were determined by RT-qPCR in cells, relative to U6 and GAPDH, respectively. D, Representative Western blots of CEP55 protein and its quantitation in cells, normalized to GAPDH. Data comparison was analysed by independent sample t test between two groups and by one-way ANOVA among multiple groups, followed by Tukey's test. *P < 0.05 versus hBMSCs treated with miR-NC or NC. # P < 0.05 versus EVs treated with NC. Data are shown as the mean ± standard deviation of three technical replicates Western blot analysis showed that the expression of ALIX and CD63 was detected in EVs, in the absence of GRP94 expression ( Figure 5F). These findings demonstrated that EVs were successfully isolated from hBMSCs, which harboured the great potential of multi-differentiation.

| miR-144-3p in hBMSCs-derived EVs exerted inhibitory effects in cervical cancer cells
EVs were isolated from hBMSCs-miR-NC and hBMSCs-miR-144-3p and added into cervical cancer cells, assigned as the SiHa + EVs-NC and SiHa + EVs-miR-144-3p groups. Transwell assay ( Figure 7A), EdU assay ( Figure 7B), colony formation assay ( Figure 7C) and flow cytometric analysis ( Figure 7D) were performed to evaluate the resulting cellu-    We found that CEP55 was up-regulated in cervical cancer cells and tissues, thus contributing to the development and progression of cervical cancer. Largely in agreement with our findings, the expression of CEP55 has been reported to be 9-fold higher when compared with adjacent normal tissues, which is associated with advanced tumour stage and a higher risk of lymph node metastasis in patients with cervical cancer. 6 A similar expression profile of CEP55 has also been indicated in multiple types of cancers, including non-small-cell lung cancer and breast cancer. 22,23 An analysis of the circular RNA-miRNA-mRNA network has revealed CEP55 as a potential mRNA implicated in the pathology of cervical cancer. 4 Furthermore, an inverse relation between CEP55 and miR-144-3p was identified in our study and miR-144-3p was found to be expressed at a low level in cervical cancer cells and tissues.

| D ISCUSS I ON
The targeting relationship between CEP55 and miR-144-3p has been validated in prostate cancer cells. 7  Nucleus circumstantial evidence of anti-cancer effects. 29 The function of overexpressed miR-144-3p in SiHa, following the delivery of miR-144-3p mimic, was observed to be counteracted by up-regulated CEP55.
Subsequently, miR-144-3p was found to be loaded in EVs isolated from hBMSCs and further transferred into recipient cells

ACK N OWLED G EM ENT
We offer our sincere gratitude to the viewers for their valuable suggestions and comments.

CO N FLI C T O F I NTE R E S T
None.

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