Bone Marrow Mesenchymal Stem Cells-Derived Exosomes Upregulate MicroRNA-424-5p to Suppress Hepatocellular Carcinoma Cell Growth by Inhibiting Forkhead Box K1

Bone marrow mesenchymal stem cells (BM-MSCs) have been shown to exert a potential therapeutic effect during tumor treatment and it has been proved that exosomes derived from BM-MSCs play crucial roles in the progression of malignant tumors. The current study aims to investigate the effect of BM-MSC-derived exosomal microRNA-424-5p (miR-424-5p) on hepatocellular carcinoma (HCC) progression. The expression of miR-424-5p was compared between HCC and adjacent normal tissues, and its prognosis value was analyzed. Additionally, exosomes were extracted from the BM-MSCs and their identity was veried. Luciferase reporter assay was conducted to identify the putative binding sites between miR-424-5p and the 3’-UTR of forkhead box K1 (FOXK1). The BM-MSC-derived exosomes were co-cultured with HCC cells to assess the effect of the BM-MSC-derived exosomes, miR-424-5p, and FOXK1 on the proliferation, migration, invasion, and in vivo tumorigenesis of the HCC cells. Then, the expression of FOXK1 was also examined in HCC and normal tissues. miR-424-5p was downregulated and FOXK1 was upregulated in HCC tissues and cells. BM-MSC-derived exosomes upregulated miR-424-5p expression to suppress the proliferation, migration, invasion, and in vivo tumorigenesis of HCC cells. Knockdown of FOXK1 also repressed the malignant behavior of the HCC cells, and FOXK1 was veried as the target of miR-424-5p. The role of FOXK1 silencing in HCC cells was reversed by miR-424-5p downregulation. Our results suggested that BM-MSC-derived exosomes upregulated miR-424-5p expression to restrain HCC cell growth and invasion via inhibition of FOXK1 expression and as a result, decelerating HCC development.


Introduction
Liver cancer is the 6th most common type of cancer and approximately 841,000 new cases are diagnosed and 782,000 deaths occur each year 1,2 . HCC accounts for about 80% of primary liver cancers and is the main pathological type 2 . HCC is a leading cause of cancer-related death worldwide 3 . The prognosis of HCC patients is poor due to the limited number of therapeutic methods available for advanced HCC 4 . All therapeutic options, such as surgery with adjuvant chemotherapy or radiotherapy, are typically only available for those with early-stage HCC [5][6][7] . Therefore, there is an urgent need to explore potential biomarkers for HCC, elucidate mechanisms implicated in HCC carcinogenesis, and development for the evaluation of molecular therapeutic targets.
BM-MSCs are non-hematopoietic stem cells found in bone marrow that have the ability of self-renewal, pronounced proliferation, and differentiation into various other types of cells 8,9 . Exosomes are small (diameter of 30-100 nm) membrane-bound vesicles secreted by multiple cell types, and it has been reported that exosomes show tremendous potential for the effective treatment of HCC [10][11][12] . However, the role of BM-MSC-derived exosomes in HCC remains largely unknown. miRNAs are small non-coding RNAs that repress gene expression by binding to the 3'-untranslated region (3'-UTR) to in uence mRNA stability or protein translation 13,14 . Speci c miRNAs, such as miR-342-3p 15 , miR-612 16 , and miR-541 17 , have been con rmed to suppress HCC cell growth. Importantly, it has been veri ed that miR-424-5p acts as a potential biomarker of HCC 18 . Nevertheless, the role of exosomal miR-424-5p, especially BM-MSC-derived exosomal miR-424-5p in HCC remains scarcely investigated. The FOXK family is a subgroup of forkhead transcripts and is known to modulate various biological processes, such as metabolism, survival, apoptosis, differentiation, and senescence [19][20][21] . As a member of the FOXK family, FOXK1 has been con rmed to be correlated with a poor prognosis of HCC and regulates the stemness of HCC cells 22 . It has also been veri ed that FOXK1 knockdown was able to suppress HCC cell growth 23 . However, the relationship between miR-424-5p and FOXK1 is still unknown.
We aimed to investigate the effect of BM-MSC-derived exosomal miR-424-5p on the biological processes of HCC cells by targeting FOXK1, and we speculated that BM-MSC-derived exosomes may regulate miR-424-5p expression and thus affect HCC cell growth through the involvement of FOXK1.

Materials And Methods
Patients and tissue samples Ninety pairs of primary HCC tissue and adjacent non-tumor liver tissue (> 2 cm and < 5 cm from tumor edge) were collected from patients (56 males and 41 females; age, 41-73 years; mean age, 52.6 years) who underwent surgical resection at our hospital from October 2013 to January 2015. The tissue samples were immediately frozen and preserved at -80 0 C in liquid nitrogen. None of the patients suffered from hepatitis or had previously accepted anti-in ammatory treatment or oncotherapy.
Written informed consent was obtained from all patients before the study and the diagnosis of all tissue samples was made and independently con rmed by two pathologists. Overall survival was calculated during the duration between surgery and death or last follow-up. This study was approved by the Shenzhen Hospital, Peking University Institutional Ethics Committee and conducted following Ethical Principles for Medical Research Involving Human Subjects of the Helsinki Declaration.
Cell culture HCC cell lines (HL7702, HCCLM3, HepG2, and MHCC97-H) and human immortalized hepatocyte line L02, and human BM-MSCs were obtained from the American Type Culture Collection. The cells were all cultured with DMEM containing 10% fetal bovine serum and 1% penicillin-streptomycin.

Flow cytometry identi cation of the BM-MSCs
Speci c stem cell markers were detected using ow cytometry, as previously described 24,25 . In brief, BM-MSCs were harvested and washed. Then, Phycoerythrin (PE)-labeled anti-CD34, CD44, CD45, CD90, and the isotype control was incubated with the BM-MSCs for 20 min. After washing with PBS, the BM-MSCs were examined using ow cytometry.
Isolation and identi cation of the BM-MSC-derived exosomes Exosomes were extracted from the culture supernatant of the BM-MSCs using ultracentrifugation and their identity was con rmed using previously reported methods 26, 27 . In brief, the BM-MSCs were cultured in DMEM containing 10% FBS (FBS was pre-centrifuged at 16,000 ×g for 60 min to remove the exosomes). After 24 h of culture, the medium supernatant was centrifuged at 16,000 ×g for 60 min at 4°C and the exosomes were harvested.

Cell transfection
The cells were plated and cultured at 37°C for 24 h. Next, the cells were transfected with miR-424-5p mimic, miR-424-5p inhibitor, siRNA-FOXK1, or the NC, respectively, using a Lipofectamine 2000 system, following the manufacturer's instructions. Transfection e ciency was measured after 48 h.

PKH26-labeled exosomes
For exosome labeling, a PKH26 kit was used as instructed by the manufacturer. In brief, exosomes were suspended in Diluent C and then mixed with PKH26. After incubation for 20 minutes in the dark at room temperature, the labeling reaction was stopped and the labeled exosomes were ultra-centrifuged at 100,000 ×g for 60 min. After washing with PBS, the exosomes were again ultra-centrifuged at 100,000 ×g for 60 min and resuspended in a serum-free medium, and co-cultured with the HCC cells. Then, the labeled exosomes were observed under a confocal laser scanning microscope.

Transmission electron microscopy
The isolated exosomes were xed in 200 μL of 3% glutaraldehyde for 100 min. A drop (20 μL) of exosomes was transferred onto a formvar-carbon-coated grid. After rinsing, the grids were stained with 1% uranyl acetate solution for 5 min. Then, the grids were washed twice and examined under a transmission electron microscope at 80 kV.
Exosome uptake experiment MHCC97-H cells were incubated with the BM-MSC-derived exosomes and stained with 1 mL of 4',6diamidino-2-phenylindole 2 HCI for 10 min. BM-MSC uptake by the MHCC97-H cells was observed under a uorescence microscope.

qRT-PCR analysis
Total RNA was extracted from the tissues and cells using TRIzol reagent, and reverse transcription was performed using a TaqMan microRNA Reverse Transcription Kit. PCR ampli cation was conducted using SYBR Premix Ex Taq with U6 and GAPDH as the loading controls of miR-424-5p and FOXK1.

Western blotting analysis
Total protein in the cells and exosomes was extracted, sodium dodecyl sulfate-polyacrylamide gel electrophoresis was conducted, and then the samples were transferred onto membranes. The membranes were blocked with 5% skim milk powder and incubated with primary antibodies FOXK1 (1: 1000; Abcam, MA, USA), CD63 (1: 500), and CD81 (1: 500) at 4℃ overnight, and then with the respective secondary antibody (1: 3000, CA, USA). An enhanced chemiluminescent system was used for detection using GAPDH as the internal reference.

Cell apoptosis and viability analyses
Apoptotic BM-MSCs were detected using FCM with an Annexin-V FITC kit, according to the manufacturer's instructions. The apoptosis of the HCC cells was determined using ow cytometry, and an FC500 MCL ow cytometer was used to detect the apoptosis rate. Cell activity was analyzed using an MTT assay, as previously described and absorbance was assessed at 570 nm using a microplate reader (Bio-Rad Laboratories, CA, USA).

Transwell assay
The transfected cells were added into Transwell apical chambers coated with Matrigel (not used in the migration assay), while the basolateral chambers were appended with 500 μL of complete medium. After 24 h of incubation, the cells on the microporous membrane were removed and the cells in the chambers were xed and stained with crystal violet for 10 min, and were then observed under an inverted microscope.

Subcutaneous in vivo tumorigenesis in nude mice
Male BALB/c nude mice were subcutaneously injected with transfected HL-7702 cells. After tumors had developed, the tumors were measured and volume was calculated using the formula: ab 2 /2 (a: length diameter, b: width diameter). The mice were euthanized on the 35 th day and the tumors were isolated and weighed. Animal experiments were conducted in strict accordance with the Guide to the Management and Use of Laboratory Animals issued by the National Institutes of Health. The mice were housed as stipulated by the protocols of animal experiments and were approved by the Institutional Animal Care and Use Committee of our Hospital.

Statistical analysis
All data analyses were conducted using SPSS 21.0 software. Measurement data that conformed to normal distribution were expressed as mean ± standard deviation. The unpaired t-test was conducted for comparisons between two groups, while one-way analysis of variance (ANOVA) was used for comparisons among multiple groups and Tukey's post hoc test was used for pairwise comparisons after one-way ANOVA. A P value of < 0.05 was indicative of a statistically signi cant difference.

Results
MiR-424-5p was downregulated in HCC tissues and cells and was associated with poor survival in HCC patients miR-424-5p expression in HCC and adjacent non-tumor liver tissues was determined using qRT-PCR, and the results showed that miR-424-5p was downregulated in HCC tissues, compared with adjacent nontumor liver tissues ( Figure 1A). Subsequently, the expression of miR-424-5p in HCC patient tissues was determined to evaluate the prognostic value of miR-424-5p for HCC patients, and patients were divided into a high expression group (n = 49) and a low expression group (n = 48) based on their level of miR-424-5p expression. The results implied that HCC patients with higher miR-424-5p expression had a much longer overall survival time, compared with HCC patients with lower miR-424-5p expression ( Figure 1B). Meanwhile, miR-424-5p expression in the HCC cell lines was also detected and it was found that the HCC cell lines had lower miR-424-5p expression than the L02 cell line, while the MHCC97-H cell line had the lowest miR-424-5p expression ( Figure 1C).
BM-MSC derived exosomes transfer miR-424-5p into HCC cells BM-MSCs have been con rmed to be potential therapeutic tools for tumor treatment. BM-MSCs were isolated and subjected to ow cytometry for the detection of BM-MSC surface markers (CD44, CD90, CD45, and CD34). Our results showed that the cells expressed high levels of CD90 and CD44 but were negative for CD45 and CD34 ( Figure 2A). Then, the exosomes were extracted from BM-MSCs and were observed under a transmission electron microscope (TEM). We discovered that the mean diameter of the exosomes was 30-100 nm ( Figure 2B). Western blot analysis indicated that the exosomes expressed exosomal markers including CD63 and CD81, which were isolated from BM-MSCs ( Figure 2C). To verify whether exosomes could be taken-up by HCC cells, PKH26 was used for exosome labeling through coculture with the MHCC97H cells. Laser scanning confocal microscopy was used to demonstrate that the labeled exosomes were taken-up by the MHCC97H cells ( Figure 2E).

BM-MSC derived exosomes inhibit HCC cell proliferation and suppress motility in vitro and in vivo
To investigate whether the BM-MSC-derived exosomes could affect HCC cell function, the exosomes were extracted and added to the HCC cells. Then, MTT assay, ow cytometry, and transwell assay were performed to detect any proliferative, cell cycle, apoptosis, and motility ability changes that may take place under the effect of the BM-MSC-derived exosomes. The results show that the BM-MSC-derived exosomes inhibited HCC cell proliferation, migration, and invasion, and promoted apoptosis of MHCC97-H cells, and also arrested the cell cycle at the G0/G1 phase ( Figure 3A, B, C, and D). The in vivo assay results also showed that injection of BM-MSC-derived exosomes can suppress tumor volume and weight in a xenograft model ( Figure 3E and F). All these results suggested that BM-MSC-derived exosomes could ameliorate the malignant activity of the HCC cells.

BM-MSC-derived exosomes deliver MiR-424-5p to suppress HCC cell growth and motility in vitro and in vivo
The results mentioned above demonstrated that BM-MSC-derived exosomes can inhibit the malignant function of HCC cells both in vitro and in vivo. Meanwhile, the downregulation of miR-424-5p was found in both HCC cell lines and HCC tissues. To further elucidate the function of miR-424-5p, we upregulated and suppressed miR-424-5p expression in exosomes and delivered them into HCC cells, and then reevaluated cell function.
The results of the MTT assay, ow cytometry, and transwell assay also con rmed that the BM-MSCderived exosomes that transferred the miR-424-5p mimic restrained proliferation, migration, and invasion, and promoted the apoptosis of MHCC97-H cells and also arrested the cell cycle at the G0/G1 phase. The effects of the BM-MSC-derived exosomes that silenced miR-424-5p in HCC cells were contrary to that of the BM-MSC-derived exosomes that overexpressed miR-424-5p ( Figure 4A, B, C, and D). The in vivo assay also showed that the injection of BM-MSC-derived exosomes conveying the miR-424-5p mimic repressed tumor volume and the weight of the xenografts, while BM-MSC-derived exosomes that downregulated miR-424-5p exerted an opposite impact ( Figure 4E and F). Therefore, we can conclude that BM-MSCderived exosomal miR-424-5p can suppress the malignant activity of HCC cells.

FOXK1 is the target gene that is negatively regulated by miR-424-5p
The online miRNA prediction website Targetscan was used to predict the potential target of miR-424-5p. We found that miR-424-5p can bind to the 3'-UTR of FOXK1, which was con rmed through dual-luciferase reporter assay ( Figure 5A and B). Next, we examined FOXK1 expression in HCC tissues and adjacent nontumor liver tissues using quantitative real-time polymerase chain reaction and western blotting analysis. The results showed that FOXK1 was signi cantly upregulated in HCC tissues, compared with adjacent non-tumor liver tissues ( Figure 5C and D).
We detected further changes in FOXK1 expression after the MHCC97H cells were supplemented with BM-MSC-derived exosomes using quantitative real-time polymerase chain reaction and western blotting analysis. The results showed that BM-MSC-derived exosomes that overexpress miR-424-5p can downregulate FOXK1 expression, while exosomes that inhibit miR-424-5p can upregulate FOXK1 expression in MHCC97-H cells ( Figure 5E and G), demonstrating the negative regulatory relationship between miR-424-5p and FOXK1. The results of the Pearson analysis also implied the negative relationship between the expression of miR-424-5p and FOXK1 ( Figure 5F).

Knockdown of FOXK1 expression suppresses biological processes in HCC cells
Since FOXK1 was overexpressed in HCC tissues and HCC cells, the speci c siRNA that targeted FOXK1 was used to detect the function of FOXK1 in HCC cells in combination with the miR-424-5p inhibitor. It was found that si-FOXK1 suppressed FOXK1 gene and protein expression, inhibited the proliferation, migration, and invasion abilities of HCC cells, and promoted HCC cell apoptosis, and arrested the cell cycle at the G0/G1 phase. Furthermore, knockdown of FOXK1 simultaneously inhibited miR-424-5p expression, and this effect was reversed by the silencing of FOXK1 in HCC cells ( Figure 6A, B, C, D, and E).
Overall these results provide evidence indicating that miR-424 regulates HCC cell functions by targeting FOXK1.

Discussion
The incidence and severity of HCC are increasing globally, while HCC prognosis is still unsatisfactory due to high rates of recurrence and metastasis 28, 29 . Thus, improved methods of treatment are critically required to provide a better prognosis for HCC patients 5,30 . This study was designed to investigate the impacts of BM-MSCs-derived exosomal miR-424-5p on the development of HCC via targeting FOXK1. We found that BM-MSC-derived exosomes upregulated miR-424-5p expression to suppress malignant phenotypes of HCC cells by inhibiting FOXK1, thus restraining HCC progression.
First, we determined miR-424-5p and FOXK1 expression levels in tissues and cells and found that miR-424-5p expression was downregulated whereas FOXK1 expression was upregulated in HCC tissues and cells, compared with adjacent normal tissues and the human normal hepatocyte line, L02. Similarly, it was previously found that miR-424-5p expression is downregulated in HCC tissues, cell lines 31 , and detached anoikis-resistant HCC cells 32 . A recent study revealed that FOXK1 is upregulated in human HCC tissue and is positively associated with cancer progression 22,23 . Moreover, miR-424-5p expression in BM-MSC-derived exosomes was also detected. The results showed that BM-MSC-derived exosomes were able to upregulate miR-424-5p expression. Delivery of miR-424-5p via exosomes can promote triple-negative breast cancer cell apoptosis 33 . However, the expression of BM-MSC-derived exosomal miR-424-5p in HCC is still unexplored. Additionally, we evaluated the relationship between miR-424-5p and FOXK1 in HCC, and it was con rmed through a bioinformatics method and dual-luciferase reporter gene assay that FOXK1 is a target gene of miR-424-5p. Nevertheless, this relationship has rarely been studied.
The HCC cells were co-cultured with BM-MSC-derived exosomes to determine the effect of the exosomes on the biological processes of HCC cells, and we found that BM-MSC-derived exosomes suppressed the proliferation, migration, invasion, in vivo tumorigenesis, and promoted the apoptosis of HCC cells. Similar to our ndings, BM-MSC-derived exosomes delivered miR-124 to inhibit pancreatic cancer cell proliferation, epithelial-mesenchymal transition, and enhanced chemotherapy sensitivity 34 . BM-MSCderived exosomes alleviated radiation-induced bone loss by activating Wnt/β-catenin signaling 35 . In another study, it was demonstrated that BM-MSC-derived exosomes could inhibit the viability, migration, and invasion of bladder cancer cells 36 . It has been also been shown that hypoxia-elicited MSC-derived exosomes can facilitate cardiac repair through miR-125b-mediated prevention of cell death in myocardial infarction 37 .
In our study, HCC cells with abnormal miR-424-5p expression were treated with transfected BM-MSCderived exosomes to determine the effect of BM-MSCs-derived exosomal miR-424-5p on HCC cells. Our results indicated that miR-424-5p overexpression suppressed the malignant behaviors of HCC cells.