Human mesenchymal stem cell derived exosomes inhibit the survival of human melanoma cells through modulating miR-138-5p/SOX4 pathway

Abstract

Melanoma, a skin cancer derived from malignant melanocytes, is characterized by high aggressiveness and mortality. However, its exact etiology is unknown. Recently, the roles of exosomes and exosomal microRNAs (miRNAs) in the progression and therapy of various disorders, including melanoma, have gained attention. We investigated the impact of miR-138-5p from exosomes released by human mesenchymal stem cells (HMSCs) on the pathogenesis of melanoma. We isolated exosomes from HMSCs (HMSC-exos) by ultracentrifugation and verified them by specific biomarkers and transmission electron microscopy. We used CCK8, flow cytometry, quantitative real-time PCR (qRT-PCR), and Western blots to investigate cell proliferation, apoptosis, and mRNA and protein levels, respectively. Additionally, we used luciferase assays to examine the relationship between miR-138-5p and SOX4. Administration of HMSC-exos dramatically repressed the growth of melanoma cells. Elevated miR-138-5p levels in HMSC-exos were linked to increased cell apoptosis, and miR-138-5p downregulation had the opposite effects on cells. SOX4 was targeted by miR-138-5p through direct binding to the SOX4 3’UTR. In melanoma tissues, miR-138-5p was downregulated, and SOX4 was upregulated and was negatively correlated.

MiR-138-5p plays a crucial role in melanoma progression. The negative regulation of SOX4 transcription mediates the function of miR-138-5p. These findings provide a novel concept of melanoma pathogenesis and identify a valuable target (miR-138-5p/SOX4 axis) in treating this disease.

1. Introduction

Melanoma, which evolves from the malignant proliferation of melanocytes, is a severe skin cancer with high aggressiveness, metastasis, mortality, and recurrence [1, 2]. Furthermore, a rapid increase in melanoma incidence has occurred over the past few decades, reaching 287,723 cases worldwide in 2018 [1, 3, 4]. Early-stage melanoma can be removed surgically, while metastasized melanoma requires multiple treatment strategies, including chemotherapy, radiotherapy, targeted therapy, and immunotherapy [1]. However, the outcomes and prognoses of most current therapies are under satisfactory. Therefore, exploring the pathogenic mechanisms of melanoma and identifying novel drug targets for the treatment of melanoma are urgently needed.

Although melanoma is related to genetic variance and ultraviolet light exposure, its exact etiology and mechanisms are still unclear. Recent publications highlighted the importance of microRNAs (miRNAs) in the development and prognosis of melanoma [5, 6, 7]. Moreover, aberrant miRNA expression has been linked to melanoma progression [8, 9, 10]. For example, upregulated expression of the miR-17-92 cluster was observed in many tumors and is linked to the unconstrained proliferation of cancer cells [11]. Conversely, overexpression of miR-138 suppresses proliferation and invasion and promotes apoptosis of tumor cells [12, 13], suggesting its potential as a molecular target. However, the functional mechanism of miR-138 in melanoma is unclear.

Exosomes are bioactive lipid bilayer vesicles secreted by most cells. Exosomes mediate intercellular communication and communication of cells with the surrounding environment by transmitting various molecules like miRNAs [14, 15]. Previous report has illustrated that exosomes show great promise in cancer immunotherapy because of their immunogenicity and molecular transfer function [16]. Moreover, miR-200b in exosome from colorectal cancer cells promoted colorectal cancer proliferation upon TGF- $\beta$ 1 exposure [17]. Furthermore, exosome that derived from melanoma contributed to the progression of cancer-related fibroblasts [18]. Exosomes stemming from human mesenchymal stem cells (HMSCs) have exhibit beneficial effects in treating cancers. It has been confirmed that exosomes derived from miR-143-overexpressing MSCS inhibit cell migration and invasion in human prostate cancer cells [19]. Importantly, exosome that derived from HMSCs enhanced radiotherapy-induced cell death in melanoma [20]. Therefore, further investigation of MSCs-derived exosomes may provide a novel vision of melanocyte biology and pathogenesis, which will help identify new targets for the evolution of therapeutic strategies.

MiRNAs elicit responses in cells by targeting specific target genes/proteins. For instance, miR-14-5 and miR-30a-5p elicit changes in melanoma through targeting SOX4 [18, 20]. As an essential developmental transcription factor, SOX4 has versatile roles in many physiological processes and modulate multiple signaling pathways [18, 20]. Whether SOX4 mediates the action of miR-138-5p in the context of melanoma remains undetermined. This study aimed to determine the role of miR-138-5p from MSC-derived exosomes, in the melanocyte survival and investigate the molecular mechanisms involved.

2. Materials and methods

2.1 Human samples

The independent ethics committee of Fudan University Shanghai Cancer Center authorized this study, which was complied with the Declaration of Helsinki. All patients signed a consent agreement before the experiment. We collected 50 human primary melanoma tissues (I–III, $n=$ 25; IV–V, $n=$ 25, respectively) and 15 corresponding para-cancerous tissues (Tissue adjacent to the primary) from Fudan University Shanghai Cancer Center, Shanghai, China and used to inspect the relative mRNA levels of miR-138-5p and SOX4, using real-time-quantitative PCR (RT-qPCR). The stages of tumor tissues were charactered by the depth of invasion and lymph node metastasis according to criterion as indicated in a previous report [21].

2.2 Cell culture

We acquired the cells used in the present study from the cell bank of Shanghai Biology Institute (Shanghai, China), including human mesenchymal stem cells (HMSCs), melanoma (A2058, A375, SKMEL1 and WM115) and a corresponding normal human epidermal melanocyte HEMA-LP. Cells were grown in DMEM (Trueline, Kaukauna, WI, USA) supplied with a 10% FBS (Thermo Fisher Scientific), 2 mM L-glutamine, and 1% penicillin/streptomycin (Solarbio, Beijing, China) and were maintained in an incubator with 5% CO ${}_2$ at 37 ${}^{\circ }$ C.

2.3 Isolation and characterization of human mesenchymal stem cell-derived exosomes (HMSC-exos)

We prepared HMSC-exos as described previously, with some modifications [22]. We collected conditioned media from cultured HMSCs and centrifuged them at 1,000 g for 20 min to clear away cell debris, followed by sequential centrifugation at 2,000 g and 10,000 g for 20 min each. The supernatants were then gathered and condensed with a molecular weight cut-off (MWCO) of 100 kDa (Millipore) at 1,000 g for 30 min. The condensed supernatants were layered over 5 mL of 30% sucrose/D2O and subjected to ultracentrifugation at 100,000 g for 1 hour (optimal-90K; Beckman Coulter). The fraction with enriched exosomes was diluted with PBS and subjected to further centrifugation 3x at 1,000 g for 30 min using a 100 kDa MWCO. The purified exosomes were filtered through a 0.22 $\mu$ m pore filter (Millipore) and stored at $-$ 80 ${}^{\circ }$ C. The protein concentration of the condensed exosomes was determined using a BCA protein assay kit (Pierce). We examined the morphology of HMSC-exos with transmission electron microscopy (TEM) (FEI Tecnai 12; Philips) and the expression of HMSC specific positive biomarkers CD105 (12-1057-41, eBioscience ${}^{\text{TM}}$ , USA) and CD90 (11-0909-42, eBioscience ${}^{\text{TM}}$ , USA) and negative biomarkers CD34 (12-0349-42, eBioscience ${}^{\text{TM}}$ , USA) and CD45 (555482, BD, USA) biomarkers were measured using flow cytometry.

2.4 Validation of exosome uptake by fluorescent labeling

Green fluorescent linker PKH67 (UR52303, Umibio, Shanghai, China) was applied to stain exosomes according to manufacturer’s protocols. Briefly, we seeded SKMEL1 and WM115 cells into 24-well plates at a density of 50,000 cells/well and prepared 5 wells for each group. Normal culture media or culture media with PKH67 labeled exosomes (20 $\mu$ g/ml) was added into control or model group cells, respectively. We stained cell nuclei with DAPI (4’, 6’-diamidino-2-phenylindole; Vector Laboratories, Inc., Burlingame, CA) at room temperature for 5 min. Four hours later, we imaged the prepared samples using a fluorescent microscope (Olympus IX71, Tokyo, Japan).

2.5 Real-time quantitative PCR (RT-qPCR)

Total RNA was extracted from samples using TRIzol Reagent (Invitrogen, Waltham, MA, USA). Corresponding cDNAs were synthesized from isolated RNAs using a kit (Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer’s instructions. The conditions used for RT-qPCR were: 95 ${}^{\circ }$ C for 10 minutes, 40 cycles of 95 ${}^{\circ }$ C for 15 seconds, 60 ${}^{\circ }$ C for 45 seconds. The 2- $\mathrm{\Delta }\mathrm{\Delta }$ Ct method [23] and normalization with $\beta$ -actin levels were applied to calculate the relative mRNA expression, while RNU6-1 was used to that in determining microRNA levels. Data represent the average value of three replicates. The primer sequences were as follows: miR-138-5p, F: 5’-GCGAGCTGGTGTTGTGAATC-3’, R: 5’-AGTGCAGGGTCCGAGGTATT-3’. SOX4 F: 5’-CAAGATCCCTTTCATTCG-3’, R: 5’-CGACCTTG TCTCCCTTC-3’. GAPDH, F: 5’-AATCCCATCACCATCTTC-3’, R: 5’-AGGCTGTTGTC ATACTTC-3’. RNU6-1, F: 5’-CTCGCTTCGGCAGCACA-3’, R: 5’-AACGCTTCACGAATTTGCGT-3’.

2.6 Western blot

We prepared total protein from various samples using RIPA lysis buffer (JRDUN, Shanghai, China) plus EDTA-free Protease Inhibitor Cocktail (Roche, Heidelberg, Germany). Protein concentrations were determined using the BCA method (Thermo Fisher Scientific). Total protein (25 $\mu$ g/well) was fractionated on 10% SDS-PAGE and transferred overnight to nitrocellulose membranes (Millipore, Billerica, MA, USA). The membranes were incubated sequentially with 5% nonfat dry milk at room temperature for 1 hour, primary antibodies at 4 ${}^{\circ }$ C overnight, and secondary antibody anti-mouse IgG (1: 1,000; Beyotime, Shanghai, China) at 37 ${}^{\circ }$ C for 1 hour. An enhanced chemiluminescence system (Tanon, Shanghai, China) was applied to detect protein levels. The following primary antibodies were used: SOX4 (Ab90696, Abcam, UK), Bcl-2 (ab196495, Abcam, UK), Bax (Ab32503, Abcam, UK), CD9 (Ab92726, Abcam, UK), CD63(Ab134045, Abcam, UK), and GAPDH (60004-1-1G, Proteintech, USA).

2.7 Knockdown and overexpression of SOX4

Short interfering RNAs targeting human SOX4 (siSOX4-1: 5’-GGAAGCUGCUCAAAGACAGTT-3’; siSOX4-2, 5’-GACAGCGACAAGAUCCCUUTT-3’; and siSOX4-3 5’-GCGACAAGAUCCCUUUCAUTT-3’) and negative control siRNA (siNC, 5’-UUCUCCGAACGUGUCACGUTT-3’) were synthesized (Major Industrial Co., Ltd, Shanghai, China) and cloned into lentiviral plasmids (pLKO.1). For SOX4 overexpression, we prepared a pLVX-puro lentiviral plasmid containing SOX4 (NM_ 003107.3) cDNA (oeSOX4), along with a mock plasmid as a negative control (oeNC). Lipofectamine 2000 was used to transfect cells following the manufacturer’s instructions and cells were cultured for 48 hours before the subsequent analysis.

2.8 Cell proliferation

We measured cell proliferation was measured using a cell Counting Kit-8 (CCK-8) assay (Signalway Antibody, Maryland, USA). At 0, 24, 48, and 72 h, we incubated cells with CCK-8 solution (1: 10) for 1 h. Cell proliferation was quantified using a microplate reader (Pulangxin, Beijing, China) with optical densities at a wavelength of 450 nm. We performed each experiment in triplicates.

2.9 Flow cytometry

Cells were harvested 48 h post-transfection and stained with Annexin V-fluorescein isothiocyanate (Beyotime, Beijing, China). Apoptosis was measured by flow cytometry (BD, San Diego, CA, USA). Annexin V $-$ and PI $-$ populations indicated early apoptotic cells, whereas Annexin V $+$ and PI $+$ represented necrotic cells (post-apoptotic necrosis or late apoptosis). We conducted three independent repeats for each experiment.

2.10 Dual-luciferase reporter gene assay

Potential binding sites for miR-138-5p in the SOX4 3’-UTR were predicted using TargetScan and Starbase. According to the predictions, DNA fragments containing wild-type or mutant sequences for miR-138-5p binding were synthesized and cloned into luciferase reporter vectors (pGL3-Basic). We introduced these constructs (namely, wild-type (WT) and Mut 3’UTR), with an internal reporter plasmid and miR-138-5p mimic or inhibitor, into human 293T cells and measured luciferase activities 48 h after transfection using a dual-luciferase reporter gene kit (Beijing Yuanpinghao Biotechnology Co., Ltd.).

2.11 Statistical analysis

We conducted statistical analyses with GraphPad Prism Software Version 7.0 (La Jolla CA, USA). Data are presented as mean $\pm$ SD from at least three samples/repeats. We compared two groups using $T$ -tests and compared multiple groups using a one-way analysis of variance with Tukey’s post hoc test. A $p$ -value $<$ 0.05 denoted statistical significance.

3. Results

3.1 MiR-138-5p expression is reduced in human melanoma tumors

The relative levels of miR-138-5p in 50 human melanoma tissues were divided into two groups (I–III, $n=$ 25; IV–V, $n=$ 25), and 15 corresponding para-cancerous tissues were measured RT-qPCR. As indicated in Fig. 1, the levels of miR-138-5p in group one (stages I–III) were significantly decreased compared with the levels in the para-cancerous tissues ( $p<$ 0.001). Moreover, this downward trend was further aggravated in group two (stages IV–V, $p<$ 0.001). These results demonstrate that miR-138-5p expression was suppressed in melanoma tissues and corresponded to melanoma severity, indicating the involvement of miR-138-5p in the development of melanoma.

Figure 1.

miR-138-5p was downregulated in human melanoma tumors. ${}^{***}p<$ 0.001 vs control, $!!p<$ 0.001 vs I–III.

Figure 2.

Human mesenchymal stem cell (HMSC)-derived exosomes (HMSC-exo, HE) inhibited the proliferation of human melanoma cells in a dose-dependent manner. A and B. The uptake of control-exos or HMSC-exos labeled with PKH-67 by human SKMEL1 and WM115 cells was determined using immunofluorescence. C and D. CCK8 assays were used to examine the proliferation of SKMEL1 and WM115 cells after co-culturing with 0, 100, 200, and 500 $\mu$ g/ml HE for 0, 12, 24, and 48 h. ${}^{*}p<$ 0.05 vs HE 0 $\mu$ g/ml, ${}^{**}p<$ 0.01 vs HE 0 $\mu$ g/ml, ${}^{***}p<$ 0.001 vs HE 0 $\mu$ g/ml. E and F. Western blot was used to examine the protein levels of KI-67 in SKMEL1 and WM115 cells that co-cultured with HMSC-exo in different concentrations, 100, 200, 500 $\mu$ g/ml. G and H. qRT-PCR was used to examine the level of miR-138-5p in SKMEL1 and WM115 cells that co-cultured with HMSC-exo in different concentrations, 100, 200, 500 $\mu$ g/ml. ${}^{***}p<$ 0.001 vs vehicle.

Figure 3.

Dysregulated miR-138-5p affects the impact of HMSC-exo on human melanoma cell apoptosis. A and B. Flow cytometry was used to examine the apoptosis of SKMEL1 and WM115 cells after co-culturing with miNC-exo, mimic-exo, or inhibitor-exo. ${}^{***}p<$ 0.001 vs miNC-exo. C and D. Western blotting was used to examine the relative protein levels of Bcl-2 and Bax in SKMEL1 and WM115 cells after co-culturing with miNC-exo, mimic-exo, or inhibitor-exo.

Figure 4.

Has-miR-138-5p inhibited the transcription of SOX4 through binding the SOX4 3’UTR in SKML1 and WM115. A–D. The relative levels of miR-138-5p and SOX4 were examined in SKML1 and WM115 cells in the presence of miNC, inhibitor, or mimic. ${}^{***}p<$ 0.001 vs miNC. E. The binding sites of has-miR-138-5p in the 3’ UTR of SOX4 and the corresponding mutant sites. F–G. Human SKML1 and WM115 cells co-transfected with WT/Mut 3’UTR of SOX4 and has-miR-138-5p miNC, mimic, or inhibitor. Dual-luciferase reporter genes were used to verify the binding of has-miR-138-5p and SOX4. ${}^{***}p<$ 0.001 vs miNC.

Figure 5.

Negative correlation between miR-138-5p and SOX4 in human melanoma tissues. A. The relative mRNA level of SOX4 was upregulated in human melanoma tissues. ${}^{***}p<$ 0.001 vs control. B. Correlation between miR-138-5p and SOX4 in human melanoma tissues.

Figure 6.

Overexpression SOX4 disrupts the effects of HMSC-exo and miR-138-5p mimic on apoptosis in WM115 cells. A. Flow cytometry was used to examine the apoptosis of WM115 cells after transfecting with oeNC and oeSOX4 in the presence of HMSC-exo, ${}^{***}p<$ 0.001 vs Vehicle; $!!!p<$ 0.001 vs oeNC $+$ HMSC-exo. B. Protein levels of SOX4 in WM115 cells after transfection with oeNC and oeSOX4 in the presence of HMSC-exo were measured using Western blots. C. Apoptosis of WM115 cells after transfecting with oeNC and oeSOX4 in the presence of miR-138-5p mimic was measured using flow cytometry. ${}^{***}p<$ 0.001 vs miNC; $!!!p<$ 0.001 vs oeNC $+$ HMSC-exo. D. Protein levels of SOX4, Bcl-2, and Bax were measured by Western blots in WM115 cells after transfecting with oeNC or oeSOX4 in the presence of miR-138-5p mimic.

3.2 HMSC-exos was absorbed by human melanoma via endocytosis and suppressed its proliferation

We characterized HMSCs and HMSC-derived exosomes (HMSC-exos). Flow cytometry revealed that over 95% of prepared HMSCs were CD90 and CD105 positive and CD34 and CD45 negative (Fig. S1A). WE observed the expected morphology of isolated HMSC-exos by TEM (Fig. S1B). The CD9 and CD63 biomarkers were present on the exosomes (Fig. S1C).

Then, we traced the endocytosis of control or HMSC-exos by human SKMEL1 and WM115 cells by PKH-7 labeling. We observed immunofluorescent signals in SKMEL1 and WM115 cells incubated with HMSC-exos (200 $\mu$ g/ml) but not in cells incubated with control-exos (Fig. 2A and B). We incubated SKMEL1 and WM115 cells with the indicated concentrations of HMSC-exos and measured cell proliferation by CCK8 assay. Interestingly, the proliferation of SKMEL1 cells was inhibited by HMSC-exos, from 24 h post-incubation ( $p<$ 0.05); 48 h post-incubation, the suppressive effect was further enhanced ( $p<$ 0.01). HMSC-exos decreased the cell number of SKMEL1 cells in a dose-dependent fashion (Fig. 2C). A similar phenomenon was observed in WM115 cells (Fig. 2D). Meanwhile, the protein expression of KI-67 was also decreased by HMSC-exos in a dose-dependent manner in SKMEL1 and WM115 cells (Fig. 22-2F). Moreover, we have identified that miR-18-5p was upregulated in HMSC-exo ( $p<$ 0.001, Fig. S1D). Then, we examined the expression of miR-138-5p both in SKMEL1 and WM115 cells with culturing with HMSC-exos in different concentrations. Clearly the level of miR-138-5p was significantly upregulated in a dose-dependent manner after co-culturing with HMSC-exos ( $p<$ 0.001, Fig. 2E and F). These results demonstrate that HMSC-exos repress melanoma cell growth in a dose-dependent manner.

3.3 Overexpression of miR-138-5p in HMSC-exos promoted the apoptosis of melanoma cells

Gain and lose function assay were used to induce miR-138-5p overexpression or silencing in HMSCs by using corresponding mimic and inhibitor. When miR-138-5p mimic or inhibitor was introduced into HMSCs, we confirmed increased or decreased miR-138-5p levels relative to the control (miNC) ( $p<$ 0.001, Fig. S2A). Similarly, we observed upregulation or downregulation of miR-138-5p in corresponding exosomes, isolated from mimic or inhibitor-treated HMSCs, relative to the control ( $p<$ 0.001, Fig. S2B).

Apoptosis in SKMEL1 and WM115 cells treated with exosomes, including miNC-exos, mimic-exos, and inhibitor-exos, was visualized by flow cytometry. Compared to miNC-exo-treated cells (either SKMEL1 or WM115), apoptosis increased remarkably in cells treated with mimic-exos, whereas apoptosis significantly decreased in cells treated with inhibitor-exos ( $p<$ 0.001, Fig. 3A and B). Consistent with these results, we observed downregulation of anti-apoptotic protein Bcl-2 and upregulation of pro-apoptotic protein Bax in cells treated with mimic-exos. WE observed downregulation of Bax and upregulation of Bcl-2 in cells treated with inhibitor-exos relative to miNC-exo-treated cells (both in SKMEL1 and WM115 cells) (Fig. 3C and D). These results demonstrate that elevated miR-138-5p levels are associated with enhanced apoptosis, while decreased miR-138-5p levels are associated with decreased apoptosis in melanoma cells, suggesting a positive relationship between miR-138-5p and melanoma cell apoptosis.

3.4 miR-138-5p inhibits SOX4 transcription by binding to its 3’UTR in melanoma cells

MiRNAs elicit cellular responses by targeting genes coding for effector protein(s). We exploited TargetScan and Starbase to predict the target(s) of miR-138-5p. The prediction indicated that SOX4 was a potential target of miR-138-5p. To validate this prediction, we introduced miR-138-5p inhibitor, mimic, or control (miNC) into SKMEL1 and WM115 cells and measured the levels of miR-138-5p and SOX4 by qRT-PCR and/or Western blot. Interestingly, increased miR-138-5p content was linked to decreased SOX4 levels at both the mRNA and protein levels compared with levels in cells treated with the control ( $p<$ 0.001, Fig. 4A and D). To substantiate the direct relationship between miR-138-5p and SOX4, we analyzed the binding sequence of miR-138-5p in the SOX4 3’ UTR and created WT and mutant (Mut) reporter constructs (Fig. 4E). We introduced these reporter genes, along with the miR-138-5p inhibitor, mimic, or control (miNC), into SKMEL1 and WM115 cells. The presence of miR-138-5p inhibitor dramatically enhanced the activity of the WT construct but had no action on the Mut construct ( $p<$ 0.001, Fig. 4F and G). Conversely, the activities of the two constructs remained low in the presence of miR-138-5p mimic relative to the miNC. Collectively, these data suggest that miR-138-5p negatively regulates SOX4 expression by binding to the SOX4 3’ UTR.

3.5 Negative correlation between miRNA-138-5p and SOX4 in melanoma tissues

We measured the expressions of miR-138-5p and SOX4 in melanoma and para-cancerous tissues. As indicated in Fig. 5A, SOX4 expression in tumor tissues was significantly increased at early stages (I–III) and further enhanced at late stages (IV–V) compared to SOX expression in para-cancerous tissues ( $p<$ 0.001). More importantly, SOX4 expression negatively correlated with miR-138-5p ( $r=-$ 0.6274, $p<$ 0.001) (Fig. 5B). These results indicate that the severity of melanoma is associated with increased SOX4 and decreased miR-138-5p expression.

3.6 SOX4 overexpression partially abolished the action of HMSC-exos and the miRNA-138-5p mimic on apoptosis

We explored SOX4’s direct influence on the apoptosis of melanoma cells. For this purpose, we measured SOX4 expression in several melanoma cell lines. Intriguingly, WM115 cells exhibited the highest mRNA and protein levels of SOX4 among the cell lines, including normal human epidermal melanocyte HEMA-LP ( $p<$ 0.001, Fig. S3A and S3B). Therefore, we manipulated the expression of SOX4 in WM115 cells using siRNA-based knockdown and lentivirus-mediated overexpression and validated the downregulation and upregulation of SOX4 at both mRNA and protein levels ( $p<$ 0.001, Fig. S3C–S3F).

Then, we performed apoptotic analyses SOX4 lentiviruses-transduced and HMSC-exos-treated WM115 cells. Relative to the control, HMSC-exos treatment promoted apoptosis in WM115 cells, and SOX4 overexpression almost completely abrogated this effect ( $p<$ 0.001, Fig. 6A). SOX4 overexpression rectified HMSC-exo-induced SOX4 downregulation (Fig. 6B). Meanwhile, compared with the control, the presence of the miR-138-5p mimic robustly accelerated WM115 cell apoptosis, and SOX4 upregulation ameliorated these effects ( $p<$ 0.001, Fig. 6C). Furthermore, the miR-138-5p mimic induced downregulation of SOX4 and Bcl-2, while Bax2 levels were rescued by high levels of SOX4 (Fig. 6D).

The A2058 cell line was chosen for further investigation since these cells exhibited the lowest level of SOX4 among all examined cell lines ( $p<$ 0.001, Fig. S3A and S3B). Compared with the control, increased apoptosis induced by SOX4 knockdown was effectively suppressed by the miR-138-5p inhibitor, and similar effects were observed on SOX4 expression ( $p<$ 0.001, Fig. S4). These results demonstrated that knockdown of SOX4 rescued the function of miR-138-5p.

4. Discussion

The severity of melanoma, including high morbidity and mortality [1, 3, 4], together with limited outcomes and prognosis after most current therapies [1] indicate the urgent need for extensive investigations into the pathogenesis of melanoma and the identification of novel targets and therapies. In this regard, we focused on miR-138-5p, which was downregulated dramatically in melanoma tissues. This important miRNA in exosomes derived from HMSCs. Our results demonstrate that: a) altered expression of miR-138-5p HMSC-exos severely affects the apoptosis in melanoma cells; b) downregulation of miR-138-5p and upregulation of SOX4 were observed in melanoma tissues, indicating a negative correlation between the two signaling molecules; c) miR-138-5p modulates SOX4 through binding to the SOX4 3’ UTR; and d) altered levels of SOX4 influence the impact of HMSC-exos and miR-138-5p on melanocyte apoptosis. Therefore, this study provides novel insights into the actions of HMSC-exos and confirmed important functions of miR-138-5p and its target SOX4 in the pathogenesis of melanoma.

The occurrence of melanoma is often associated with abnormal proliferation and apoptosis of melanocytes [24, 25]. Thus, inhibiting proliferation or promoting apoptosis of melanocytes can be an effective way to treat melanoma. The beneficial effects of HMSC-exos were previously demonstrated [26, 27]. For instance, HMSC-exos ameliorated hypoxia-induced cardiomyocyte apoptosis [26]. Similarly, exosomes from MSCs protected rats from glucocorticoid-triggered osteonecrosis [27]. In this study, we demonstrate an important action of HMSC-exos on the survival of melanoma cells, thereby expanding the functional scope of the exosomes from MSCs.

The most abundant cargo molecules within exosomes are miRNAs [28]. Therefore, the content and composition of miRNAs have fundamental effects on the functions of exosomes. These miRNAs affect gene expression in a post-transcriptional fashion by targeting corresponding mRNAs. Furthermore, the crucial impact of miRNAs on the progression of melanoma has been established [6, 29, 30]. For example, miR-155, mi-193b, and miRNA-205 modulate melanoma cell proliferation, while miR-145 and miR-182 participate in the survival of melanoma cells [31]. Similarly, the impact of miR-138 on the proliferation of melanoma cells has been verified [13]. Previous reports have indicated that exosomes from highly metastatic melanomas increased the metastatic behavior of primary tumors by permanently ‘educating’ bone marrow progenitors and regulates the polarization of macrophage activation [32, 33]. Moreover, breast cancer-derived exosomes promoted the progression of cell invasion and metastasis [34].

Therefore, tumor-derived exosome not only have roles of control of malignant phenotype in vitro model of melanoma cells but also give contrary effects as to promote cell invasion and metastasis. They have dual possible effects, which is different to the function of HMSC-derived exosome. Here, we validated the impact of miR-138 from HMSC-exos, the current study showed that upregulating miR-138-5p promoted melanoma cell apoptosis, thereby highlighting the important influence of this miRNA on the pathogenesis of melanoma.

The multiple functions of miR-138-5p in physiological and pathological conditions are elicited by targeting effector molecules [35, 36, 37]. For instance, MSC-derived exosomal miR-138-5p from bone marrow protects astrocytes from ischemic stroke via suppression of lipocalin 2 [36]. MiR-138 inhibits the proliferation of melanoma cells via direct targeting of hypoxia-inducible factor-1 $\alpha$ [13]. Our results demonstrate that HMSC-derived exosomal miR-138-5p regulates melanoma cell growth and apoptosis by directly targeting SOX4. The vital role of SOX4 in the pathogenesis of melanoma has been verified by previous studies [38, 39]. In particular, miR-140-5p and miR30a5p [38, 39], together with miR-138-5p, exert their cell proliferation and/or apoptosis functions through modulation of SOX4, suggesting that SOX4 mediates the functions of multiple miRNAs and serve as a vital target for melanoma treatment.

5. In conclusion

Together, our findings indicated that HMSC-derived exosomes inhibited the apoptosis of melanoma cells through delivering miR-138-5p, then targets SOX4. Our results not only indicated the critical role of miR-138-5p/SOX4 in the progression of melanoma cells but also provided evidence to point out the potential value of stem cell-based therapy as a promising management option for melanoma.

Data availability statement

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

Funding

This research was Funded by Natural Science Foundation of the Tibet Autonomous Region (NO. XZ2017 ZRG-87 (Z)).

Author contributions

Conception: ZL.

Interpretation or analysis of data: XW and ZC.

Preparation of the manuscript: XW, ZC and BS.

Revision for important intellectual content: XW and ZC.

Supervision: ZL.

Supplementary data

The supplementary files are available to download from http://dx.doi.org/10.3233/CBM-210409.

sj-pdf-1-cbm-10.3233_CBM-210409.pdf - Supplemental material

Supplemental material, sj-pdf-1-cbm-10.3233_CBM-210409.pdf