Extracellular vesicles‐encapsulated let‐7i shed from bone mesenchymal stem cells suppress lung cancer via KDM3A/DCLK1/FXYD3 axis

Abstract Accumulating evidence has suggested that extracellular vesicles (EVs) play a crucial role in lung cancer treatment. Thus, we aimed to investigate the modulatory role of bone marrow mesenchymal stem cell (BMSC)‐EV‐derived let‐7i and their molecular mechanism in lung cancer progression. Microarray‐based analysis was applied to predict lung cancer‐related miRNAs and their downstream genes. RT‐qPCR and Western blot analyses were conducted to determine Let‐7i, lysine demethylase 3A (KDM3A), doublecortin‐like kinase 1 (DCLK1) and FXYD domain‐containing ion transport regulator 3 (FXYD3) expressions, after which dual‐luciferase reporter gene assay and ChIP assay were used to identify the relationship among them. After loss‐ and gain‐of‐function assays, the effects of let‐7i, KDM3A, DCLK1 and FXYD3 on the biological characteristics of lung cancer cells were assessed. Finally, tumour growth in nude mice was assessed by xenograft tumours in nude mice. Bioinformatics analysis screened out the let‐7i and its downstream gene, that is KDM3A. The findings showed the presence of a high expression of KDM3A and DCLK1 and reduced expression of let‐7i and FXYD3 in lung cancer. KDM3A elevated DCLK1 by removing the methylation of H3K9me2. Moreover, DCLK1 suppressed the FXYD3 expression. BMSC‐EV‐derived let‐7i resulted in the down‐regulation of KDM3A expression and reversed its promoting role in lung cancer development. Consistently, in vivo experiments in nude mice also confirmed that tumour growth was suppressed by the BMSC‐EV‐derived let‐7i. In conclusion, our findings demonstrated that the BMSC‐EV‐derived let‐7i possesses an inhibitory role in lung cancer progression through the KDM3A/DCLK1/FXYD3 axis, suggesting a new molecular target for lung cancer treatment.


| INTRODUC TI ON
Lung cancer is the most prevalent cancer and the leading cause of cancer-related death, accounting for over 2 million cases and approximately 1.76 million deaths in 2018 globally and has been a significant burden on health care around the world. 1,2 Risk factors for lung cancer include tobacco consumption, air contamination, second-hand exposure to tobacco, genetic mutations and single nucleotide polymorphisms. 3 Current treatments for lung cancer include surgery, radiation, chemotherapy and molecular-targeted therapy; however, further understanding of the immune landscape of malignancies is required to enhance the therapeutic effects of molecular-targeted therapy. 4 Hence, this study was designed to explore the therapeutic target at the molecular level to alleviate lung cancer.
Extracellular vesicles (EVs) are tiny, subcellular sacs released in both physiological and pathological conditions, their function including carrying biological cargos derived from parent cells. 5,6 EVs derived from mesenchymal stem/stromal cells (MSCs) were identified to be involved in numerous lung pathologies, such as acute lung injury, acute respiratory distress syndrome and lung carcinoma. [7][8][9] Most importantly, EVs derived from bone marrow mesenchymal stem cells (BMSCs) have been implicated in the development of lung cancer. 10 The Let-7 family is a group of mi-croRNAs (miRNAs) that have been confirmed as key regulators of several physiological processes and immune responses of several cancers. [11][12][13] The Let-7i is a vital member of the let-7 family and characterized as the premier miRNAs found to have an aberrant expression in multiple malignant tumours. 14 For instance, according to a previous study, sera obtained from smokers and lung cancer patients was observed to have significantly down-regulated levels of let-7i-3p, which is indicative of its association with the pathogenesis of smoking and smoking-related lung cancer. 15 Moreover, the miRNAs from EVs have been proposed as potential biomarkers for several diseases. 16 Accordingly, let-7a in serum exosomes has been reported to be involved in the epithelial-to-mesenchymal transition process and could be implicated in the treatment of metastatic melanoma. 17 Notably, let-7i-5p within EVs has been illustrated as a tumour suppressor of head and neck squamous cell carcinoma. 18 Nevertheless, the histone lysine demethylase 3A (KDM3A) is a histone demethylase in the JmjC domain-containing protein family and has been associated with the development of tumours due to its ability to enhance gene transcription by demethylating H3K9me1 and H3K9Me2. [19][20][21] Moreover, KDM3A has been indicated as a crucial factor in lung adenocarcinoma. 22 On the other hand, doublecortin-like kinase 1 (DCLK1), a cancer stem cell marker, is accounted for pathogenesis, development and poor prognosis in numerous types of cancer including non-small cell lung cancer. 23,24 The specificity of DCLK1-42 mAb and DCLK1-87 mAb in NCM460, HCT116 and colorectal cancer tissues has been previously confirmed, and a high degree of overlap was observed between DCLK1 and microtubule protein expression, indicating that both DCLK1-42 mAb and DCLK1-87 mAb recognized DCLK1 in the cytoplasm. 25 Of note, the clinical significance of FXYD domain-containing ion transport regulator 3 (FXYD3), a sodium-potassium ATPase regulator, has been demonstrated in several types of cancer. 26 Particularly, the FXYD3 has been confirmed as a promising regulator in the progression of lung cancer. 27 With the aforementioned findings taken into consideration, we conducted the present study with the aims of elucidating the regulatory role of EV-let-7i in lung cancer with the involvement of KDM3A/DCLK1/FXYD3 for finding a novel target for lung cancer treatment.

| Ethics statement
The current study was performed in accordance with the
Following 10-minute incubation with 3,3'-diaminobenzidine tetrahydrochloride, the sections were re-stained with haematoxylin for 2 minutes. Finally, the pathology results were obtained under a microscope.

| BMSC treatment
Human BMSCs were purchased from American type culture collection (ATCC, Manassas, VA, USA) and cultured in Mesenchymal Stem Cells medium (HUXMA-03011-440, Cyagen Biosciences Inc) in an environment containing 5% CO 2 with 90% humidity. Oligonucleotides with suppressed or overexpressed let-7i were cloned into the lentiviral vector pLenti-U6-pgkpuro for BMSC infection for the purpose of suppressing or overexpressing let-7i.

| Isolation and identification of EVs
After 48 -72 hours of cell incubation, the culture medium was collected and the EVs were isolated by ultracentrifugation. An equal volume of plasma (1 mL) and filtered phosphate-buffered saline (PBS) was mixed to reduce the viscosity of the solution before centrifugation. Briefly, the cell culture medium was centrifuged at 300 g for 10 minutes, 2000 g for 15 minutes, and 12 000 g for 30 minutes to aid the removal of the floating cells and cell debris followed by filtration through a 0.22 μm filter. The supernatant was further ultracentrifuged at 1 × 10 6 g at 4℃ for 2 hours, washed in PBS, and subjected to the second ultracentrifugation under the same conditions. Finally, the pellet was resuspended in about 100 mL of PBS and stored at −80℃ for future use or immediate use.

| Transmission electron microscope (TEM) observation
EV pellet obtained by ultracentrifugation was fixed in 2% glutaraldehyde overnight at 4℃, washed with PBS, fixed with 1% OsO4 for 1 h, dehydrated in ethanol and embedded using epoxy resin. The embedded material was sectioned using a microtome and saturated with sodium periodate and 0.1N hydrochloric acid. After 10 minutes, the size and morphological characteristics of EV were examined with the use of a TEM (JEM-1010, JEOL, Tokyo, Japan). The EV suspension was mixed with an equal volume of 4% paraformaldehyde and deposited on a Formvar carbon-coated EM grid. Images were acquired using a TEM (Hitachi, Tokyo, Japan). Western blot analysis was applied to identify the EV surface marker proteins of rabbit anti-CD63 (ab134045, 1:1000, Abcam, Cambridge, UK), rabbit anti-CD81 (ab109201, 1:5000, Abcam), and rabbit anti-Calnexin (ab92573, 1:20 000, Abcam).

| Nanoparticle tracking analysis (NTA)
The size distribution and concentration of EVs were determined by the NTA (Zetasizer Nano ZS90 instrument, Malvern Panalytical) according to the characteristics of light scattering and Brownian motion. The EVs were resuspended and mixed in 1 ml of PBS and the diluted EVs were injected into a Zetasizer Nano ZS90 instrument followed by the determination of particle size according to Brownian motion and diffusion coefficients.
Filtered PBS was used as a control. All samples were measured with NP100 membrane using 44.5 mm and 0.64 V. Samples were diluted (1:1000) using CPC100 standard particles at the same settings. Collectively, five videos were recorded, usually lasting for 60 seconds. Subsequently, the data were analysed with Zetasizer software (Malvern Panalytical) and optimized to identify and track each particle frame by frame.

| EV labelling and immunofluorescence
EVs were resuspended in 400 μL of PBS at a concentration of 0.1-0.2 μg and stained with CellMask Deep Red (Thermo Fisher Scientific) with the excitation/emission wavelength of 649/666 nm. At the time of labelling, EVs were incubated with crimson dye (1:1000) at 37℃ for 20 minutes. Unbound dye was removed by PBS washing (1 to 10 000 v/v ratio), after which EVs were centrifuged at 100 000 × g for 1 hour and diluted in PBS.
The protein concentration was determined using a BCA protein detection kit.
Cells were stained with the CellTrace TM Carboxyfluorescein succinimidyl ester (CFSE, Life Technologies) with a maximum excitation/emission wavelength of 492/517 nm. The CFSE dyes diffuse into cells after being digested by the endoesterase and covalently bound to the intracellular amines to form stable fluorescent staining. About 3 to 5 × 10 5 , A549 cells in serum-free medium were stained with CFSE dye at a ratio of 1:1000 (5 μM working concentration) and then incubated at 37℃ for 20 minutes avoiding light exposure. The solution was settled, and the cells were washed with a serum-free medium at a ratio of 1:10 to remove the free dye.
Cells were inserted into an 8-well chamber slide (Millipore).
CFSE-stained cells were incubated and treated with EVs at different time-points. Cells were washed and fixed with 3.7% (w/v) formaldehyde for 5 minutes at room temperature for cell imaging. The observation was conducted using a fluorescence microscope.

| RT-qPCR
Total RNA was extracted using the TRIzol reagent (15596026, Invitrogen), RNA was reversely transcribed into cDNA according to the instructions of the PrimeScript RT reagent Kit (RR047A, Takara), and the synthesized cDNA was determined by RT-qPCR using Fast SYBR Green PCR reagent (Applied biosystems, Thermo Fisher Scientific) and ABI PRISM 7300 RT-PCR system (Applied Biosystems). Each well was set up with 3 replicates. Glyceraldehyde phosphate dehydrogenase (GAPDH) was used as an internal reference, and the relative expression of MALAT1 was analysed using the

| Western blot analysis
Cells were trypsinized and lysed with an enhanced radioimmunoprecipitation assay (RIPA) lysis buffer (Boster Biological Technology Co. Ltd.) containing a protease inhibitor, after which the protein concentration was determined using a bicinchoninic acid (BCA) protein quantification kit (Boster Biological Technology Co. Ltd.). Proteins were separated by 10% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). The separated proteins were electrotransferred to a polyvinylidene fluoride (PVDF) membrane (Millipore). The membrane was blocked with 5% bovine serum albumin (BSA) at room temperature for 2 hours to block non-specific binding and incubated with diluted primary antibodies as follows: mouse anti-KDM3A (ab91252), rabbit anti-DCLK1 (ab37994), rabbit anti-FXYD3 (ab205534), rabbit anti-vimentin (ab92547), rabbit anti-N-cadherin (ab18203), rabbit anti-E-cadherin (ab76319), rabbit anti-β-actin (ab8227) and mouse anti-β-actin (ab8226) overnight at 4℃ and then cultured with HRP-labelled goat anti-rabbit secondary antibody (ab205718) or goat antimouse secondary antibody (ab205719) for 1 hour at room temperature. All antibodies used above were obtained from Abcam. Subsequently, the membrane was developed with ECL working solution (Millipore). The ImageJ analysis software (Bio-Rad) was used to quantify the grey levels of each band in the Western blot analysis and β-actin was used as an internal reference.

| Cell counting kit 8 (CCK-8) assay
The CCK-8 kit (  cell suspension. After incubation at 37℃ for 24 hours, cells that did not invade the surface of the Matrigel membrane were gently removed with a cotton swab, whilst cells that invaded the membrane were fixed with 100% methanol and stained with 1% toluidine blue (Sigma-Aldrich). Stained invading cells were counted by hand using an inverted light microscope (CarlZeiss). Five fields of view were randomly selected for counting.

| Flow cytometry
According to the manufacturer's instructions, apoptosis of

| Chromatin immunoprecipitation (ChIP) assay
After reaching about 70%-80% of cell confluence, the cells were added with 1% formaldehyde and fixed at room temperature for 10 minutes to fix and cross-link the DNA and protein in cells. After

| Xenograft tumours in nude mice
Twenty-four healthy Balb/c nude mice ( the mice were killed on the 30th day following these procedures.
Haematoxylin-eosin staining was used to detect lung metastasis on paraffin sections.

| Statistical analysis
The SPSS 21.0 statistical software (IBM Corp.) was used for statistical analysis. Data were expressed as the mean ± standard derivation.
Data of cancer tissues and paracancerous tissues were compared using paired t test, and data between the other two groups were compared using an unpaired t test. Comparison among multiple groups was analysed by one-way analysis of variance (ANOVA). Comparison among groups at different time-points was analysed by the two-way ANOVA. Tumour volume was analysed by the repeated-measures ANOVA P < .05 was considered statistically significant.

| BMSC-EV-derived let-7i inhibits proliferative, migrative and invasive potentials of lung cancer cells
According to existing literature, mesenchymal stem cell (MSC)-derived exosomes are capable of inhibiting lung cancer progression. 28 To study its specific regulatory mechanism, miRNAs that were closely related to lung cancer were screened out. cgi) was used to analyse downstream pathways. Then, the survival curve illustrated that let-7i-5p was closely related to the prognosis of lung cancer ( Figure 1C, Table 2). Additionally, results from RT-qPCR showed an evident decrease in the expression of let-7i in lung cancer tissues when compared with that of paracancerous tissues (P < .05) ( Figure 1D). Subsequently, the expression of let-7i in lung cancer cells was determined and we observed that compared with that of human normal lung fibroblasts LL29, let-7i expression was considerably lowered in lung cancer cells A549 and H125, whereas the expression was the lowest in A549 cells ( Figure 1E). Therefore, lung cancer cell line A549 was selected for subsequent experiments.

GSE63805 GSE102286
Furthermore, the BMSC-EVs were extracted and then identified with the application of TEM, NTA and Western blot analysis. It was revealed that BMSC-EVs had an average diameter of 155 ± 2.8 nm and expressed CD63 and CD81 proteins but that did not express nanoparticles of calnexin protein (Figure 2A-C). Next, lung cancer cells A549 were stained by CFSE. Whilst the BMSC-EVs were stained with a deep red dye of lipophilic cell membrane. Then, 10 μg of the stained BMSC-EVs were added to the CFSE-stained lung cancer cells and cultured for a period of time. Our data from the fluorescence microscope revealed that lung cancer cells could adhere to and internalize/uptake EVs ( Figure 2D).
The expression of let-7i in BMSC-derived EVs was determined, and we found that compared to the supernatant, let-7i expression was remarkably higher in EVs ( Figure 2E). BMSC-EVs extracted from BMSCs were transfected with let-7i mimic/let-7i inhibitor and added into lung cancer cells in order to prove that let-7i in EVs are capable of regulating the pathogenesis and development of lung cancer.
The expression of let-7i was measured by the RT-qPCR. Our results

| BMSC-EV-derived let-7i inhibits the pathogenesis of lung cancer by repressing KDM3A
To further study the downstream regulation mechanism of let-7i, RAID, mirDIP, DIANA TOOLS, miRDB, starBase and miRWalk were applied and intersections were taken by plotting the Venus diagrams and acquired 14 key genes ( Figure 3A), among which the most important transcription factor KDM3A was obtained ( Figure 3B).
The binding site of let-7i and KDM3A was predicted by the StarBase ( Figure 3C). Results of the dual-luciferase reporter gene assay demonstrated that the luciferase activity of cells cotransfected with let-7i and KDM3A WT 3'UTR was significantly decreased, whereas the cells which received other treatment presented with no remarkable changes ( Figure 3D). The positive expression rate of the KDM3A in lung cancer tissues and paracancerous tissues was determined by the RT-qPCR, and KDM3A was found to have a high expression in lung cancer tissues ( Figure 3E). Moreover, IHC results further confirmed that KDM3A was overexpressed in lung cancer tissues ( Figure S1). The expression of KDM3A in lung cancer cell line A549 was also measured, and the results showed that compared with that in LL29 cells, KDM3A was highly expressed in lung cancer cell line A549 ( Figure 3F). Following treatment with EVs, the expression of KDM3A in lung cancer cells was measured. As a result, compared

| KDM3A promotes DCLK1 expression by removing H3K9me2 modification, thereby promoting proliferation, migration and invasion of lung cancer cells
According to a previous study, both KDM3A and DCLK1 can deteriorate the lung cancer, whilst KDM3A can promote the expression of DCLK1 through its demethylation modification. 21,29 Therefore, the underlying regulatory mechanism of demethylase KDM3A on DCLK1 was further evaluated in lung cancer. Our results from starBase analysis indicated no significant correlation between the expression of KDM3A and DCLK1 ( Figure 4A), whilst MEM analysis illustrated a remarkable co-expression relationship between KDM3A and DCLK1 ( Figure 4B). Then, the expression of DCLK1 in lung cancer tissues and paracancerous tissues was measured with the use of RT-qPCR, the results of which showed that DCLK1 was highly expressed in the lung cancer tissues ( Figure 4C). IHC results further confirmed that DCLK1 was overexpressed in lung cancer tissues ( Figure S1).
To determine the regulatory relationship between KDM3A and DCLK1, lung cancer cells were transfected with si-KDM3A and si-KDM3A-1 with high efficiency was selected to detect its effect on the expression of DCLK1 in lung cancer cells. It was found that compared with that of cells transfected with si-NC, the expression of DCLK1in cells transfected with si-KDM3A was notably reduced ( Figure 4D and

| DCLK1 promotes the proliferation, migration and invasion of lung cancer cells by inhibiting FXYD3
Previously reported research work has confirmed that DCLK1 is an inhibitor of FXYD3. 30 Thus, we hypothesize that DCLK1 regulates the development of lung cancer by mediating the expression of FXYD3.
For this purpose, StarBase was employed to predict the relationship between DCLK1 and FXYD3 and we found that they were negatively correlated ( Figure 5A). The expression of FXYD3 in lung cancer tissues and paracancerous tissues was measured with the application of RT-qPCR, the findings of which showed a low expression of FXYD3 in lung cancer tissues ( Figure 5B). IHC results further demonstrated that FXYD3 was overexpressed in lung cancer tissues ( Figure S1).
Thereafter, the lung cancer cells were transfected with sh-DCLK1 to detect the knockdown efficiency of DCLK1, whereas sh-DCLK1 with high transfection efficiency was selected for follow-up experiments ( Figure 5C). It was found that the expression of FXYD3 protein in the cells transfected with sh-DCLK1 was notably up-regulated ( Figure 5D).
To further determine that DCLK1 regulates the lung cancer cell proliferation, migration and invasion through FXYD3, lung cancer cells were transfected with oe-DCLK1 + oe-FXYD3 and the results illustrated that the overexpression of FXYD3 led to increased FXYD3 expression but inhibited proliferation, migration and invasion abilities of lung cancer cells; however, these changes could be reversed by the addition of oe-DCLK1 ( Figure 5E-G). Hence, DCLK1 could inhibit the FXYD3 and promote the proliferation, migration and invasion of lung cancer cells.

| BMSC-EV-derived let-7i suppresses the pathogenesis of lung cancer by suppressing DCLK1/ FXYD3 axis through KDM3A
To identify that the BMSC-EV-derived let-7i was involved in the occurrence and development of lung cancer by inhibiting the DCLK1/ were noticeably down-regulated, whereas the proliferation and invasion of lung cancer cells were inhibited remarkably, whereas the number of apoptotic cells was significantly increased (Figure 6A-E).
Hence, BMSC-EV-derived let-7i was identified as an inhibitor of the pathogenesis of lung cancer by suppressing the DCLK1/FXYD3 axis through KDM3A. whereas E-cadherin expression was remarkably increased (P < .05) ( Figure 7G). Therefore, let-7i in BSMC-EVs exerted a repressing effect on the progression of lung cancer in vivo.  domain-containing protein family, is known to be up-regulated in tumours and exerts a pro-tumorigenic function. 19,20,42 For instance, KDM3A expression was notably increased in colorectal cancer metastatic lesions and its high expression was correlated with poor prognosis and short overall survival of colorectal cancer patients. 43 Accordingly, previous studies have showed the presence of a high expression of KDM3A in lung cancer cells, and thus, it is considered to be responsible for the development of lung cancer. 22 Furthermore, our results illustrated that KDM3A promotes the DCLK1 expression by binding to DCLK1 promoter and removing H3K9me2 modification, thereby promoting proliferation, migration and invasion of lung cancer cells. DCLK1 has also been proposed as a novel cancer stem cell marker in a plethora of human cancers. 44 According to previous studies, the overall survival of lung cancer patients was remarkably reduced after methylation of DCLK1 promoter, which was consistent with our findings. 45 Particularly, KDM3A has been reported to enhance the expression of DCLK1 with the removal of di-and mono-methyl residues from H3K9me2/me1, and their high expression in pancreatic cancer tissues was correlated with shorter survival times of patients. 21 Lastly, we found that the DCLK1 promotes the proliferation, migration and invasion of lung cancer cells through the inhibition of FXYD3. Accordingly, FXYD3, being a sodium-potassium ATPase regulator, has been confirmed as a key mediator in a large number of cancer types. 26 Meanwhile, the poor expression of FXYD3 has been previously reported in lung cancer cells whereas its inactivation was identified as a potential player in lung cancer development. 27 Another study has demonstrated that the expression of FXYD3 was reduced in colorectal cancer cells with overexpression of DCLK1, thus representing a novel therapeutic target for colorectal cancer. 30 Collectively, these findings illustrated KDM3A as a tumour promoter in lung cancer through FXYD3 suppression by elevating DCLK1 via the removal of the methylation of H3K9me2 in the DCLK1 promoter region.

| D ISCUSS I ON
In summary, our study provided evidence in support of the suppressive effects of BMSC-EV-derived let-7i on the development of lung cancer cells through FXYD3 down-regulation by increasing DCLK1 via KDM3A inhibition. However, the underlying mechanism of let-7i in lung cancer requires additional investigations. Taken together, the aforementioned findings validated the antitumour effect of BMSC-EV-derived let-7i, which could be promising in its use as a clinically viable target in lung cancer treatment.

ACK N OWLED G EM ENTS
We would like to give our sincere appreciation to the reviewers for their helpful comments on this article.

CO N FLI C T S O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.