Reduction of miR-744 delivered by NSCLC cell-derived extracellular vesicles upregulates SUV39H1 to promote NSCLC progression via activation of the Smad9/BMP9 axis

Non-small cell lung cancer (NSCLC) is a common type of lung cancer. Extracellular vehicles (EVs) are nano-sized particles containing proteins, lipids, and miRNAs secreted by various cells, which play important roles in the development of lung cancer by regulating a wide range of signaling pathways. This study focused on elucidating a potential mechanism by which EVs promote the development of NSCLC. Expression levels of miR-744, SUV39H1, Smad9, and BMP4 in clinical tissue samples of NSCLC patients and cell lines were quantified by RT-qPCR and/or western blot analysis. The interaction between SUV39H1 and miR-744 was identified by dual-luciferase reporter assay. miR-744, SUV39H1, and BMP4 expression levels were modulated in A549 and H460 cells, in order to evaluate their effects on cell proliferation, colony formation and cell cycle. A NSCLC xenograft mouse model was used to verify the in vitro findings. NSCLC cell-derived EVs and normal bronchial epithelial cell-derived EVs were isolated and their roles in NSCLC development were evaluated in vivo and in vitro. miR-744 expression was lower in cancer cell-derived derived EVs than in normal lung epithelial cell-derived EVs. Reduced miR-744 expression in EVs upregulated SUV39H1 in NSCLC cells and further increased BMP4 levels to promote NSCLC development. BMP4 was upregulated in NSCLC cells upon suppression of Smad9 mediated by SUV39H1. Reduced miR-744 expression transferred from cancer cell-derived EVs into NSCLC cells enhanced cancer development. Overall, our findings unveiled a mechanism whereby miR-744 delivered by NSCLC-derived EVs upregulated SUV39H1, activating the Smad9/BMP9 axis and thus promoted cancer development.


Page 2 of 19
Gao et al. J Transl Med (2021) 19:37 Extracellular vesicles (EVs) comprise of nano-sized particles secreted by different types of cells into the microenvironment. Carrying a variety of bioactive molecules such as proteins, lipids, and nucleic acids including microRNAs (miRNAs), EVs are actively involved in both physiological and pathological processes [3][4][5]. For instance, cancer cell-derived EVs have been shown to regulate the metastasis, angiogenesis, proliferation, and chemosensitivity of NSCLC [6][7][8][9]. Though accumulating evidence points to the importance of EVs in NSCLC development and progression, the underlying molecular mechanisms remain elusive. miRNAs and EVs are closely involved in cancer pathogenesis. miRNAs can be packaged in EVs and exert regulatory functions on tumorigenesis. For instance, breast cancer cells secrete EVs, which can deliver miRNAs to normal mammary epithelial cells and induce neoplastic transformation [10]. Further evidence has demonstrated that EV-delivered miR-222-3p promotes NSCLC by increasing cancer cell proliferation and migration [11]. EV-derived miR-96 has also been shown to promote lung cancer cell proliferation and migration by targeting LIMdomain-only protein 7 (LMO7). Interestingly, targeting miR-96 resulted in reduced cell proliferation and migration in lung cancer cells [12]. Together, these findings highlight the importance of EV-derived miRNAs in cancer development. In another instance, miR-744, which has been shown to inhibit NSCLC proliferation [13], has also been shown to be downregulated in NSCLC cellderived EVs and has been considered an early diagnostic biomarker for NSCLC [14]. However, the mechanisms underpinning the involvement of downregulated miR-744 in cancer cell-derived EVs in the development of NSCLC has not been studied.
Suppressor of variegation 3-9 homolog 1 (SUV39H1), which is known as a histone methyltransferase, has been implicated in the development of NSCLC. A recent study has shown that silencing SUV39H1 led to elevated NSCLC cell proliferation, migration, and survival [15]. On the other hand, reduced expression of SUV39H1 is associated with upregulation of Smad9 [16], a negative regulator of bone morphogenetic protein 4 (BMP4) [17]. Notably, recent evidence has suggested that BMP4 promotes the tumorigenesis of NSCLC [18]. Based on these reports, it appears reasonable to speculate that SUV39H1 is associated with NSCLC tumorigenesis in a Smad9/ BMP4 dependent manner.
In the current study, we have confirmed that miR-744 expression in NSCLC cell derived EVs was reduced compared with that in normal non-tumor cell derived EVs. Target gene screening and validation demonstrated that the histone methyltransferase SUV39H1 is the target gene of miR-744. Moreover, our data showed that reduced miR-744 in cancer cell-derived EVs contributes to upregulation of SUV39H1. Furthermore, our findings showed that upregulated SUV39H1 positively regulates the expression of BMP4 to promote NSCLC tumorigenesis by suppressing Smad9 in NSCLC cells. Thus, our study has connected the gap between downregulated miR-744 in tumor cell-derived EVs and upregulated SUV39H1 promoted NSCLC development. This study will improve our understanding of the pathogenesis of NSCLC and will also contribute to the identification of novel therapeutic target toward NSCLC.

Ethics statement
The study protocol was ratified by the Ethics Committee of the First Hospital of Qinhuangdao and conformed to the Declaration of Helsinki. All participants provided signed written informed consent documentation. Experiments involving animals were implemented under ratification from the Animal Ethics Committee of the First Hospital of Qinhuangdao strictly following the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health. Best efforts were made to minimize the numbers and suffering of the included animals.

EV isolation
Extraction of EVs was conducted based on a previously described method [19]. Briefly, NSCLC cells (H1299 and H522 cells) and BEAS-2B cells were cultured in medium without serum for 72 h and the supernatants were attained following 10-min centrifugation at 3000 g and 4 ℃. Thereafter, large EVs were removed by 30-min centrifugation of the supernatants at 10,000 g and 4 ℃. Then, microvesicles and apoptotic bodies were discarded by passing the supernatants through a 0.22 μm filter (Millipore, Burlington, MA, USA). The 2-h ultracentrifugation was performed subsequently (100,000 g, 4 ℃) (Optima L-80XP, Beckman Coulter, Brea, CA, USA) and the supernatants were carefully discarded. The obtained pellets were suspended in ice-cold phosphate-buffered saline (PBS) and ultracentrifugation was repeated. After centrifugation, PBS was removed and the EVs were suspended in 200 μL ice-cold PBS.

Immunofluorescence assay
Isolated EVs were tagged with PKH67 green fluorescence dye (PKH67, Sigma-Aldrich, St Louis, MO, USA) in the light of the manufacturer's directions. Cells were seeded into 8-well chamber slides (Thermo Fisher Scientific, Rockford, IL, USA) at a concentration of 8000 cells/well, then added with 5 μL PKH67-labeled EVs, and incubated for 4 h for internalization. Next, the slides were fixed with 4% paraformaldehyde (Beijing Leagene Biotech. Co., Ltd., Beijing, China) for 15 min. EV specific marker protein CD63 (ab217345, 1:100, Abcam) was supplemented for overnight incubation of slides at 4 ℃. The next day, the red fluorescence-labeled secondary antibody was supplemented for 1-h incubation of slides in the dark at ambient temperature. Subsequent to nuclei staining with 4′,6-diamidino-2-phenylindole, the slides were observed, and images were obtained using a Zeiss LSM 780 (Zeiss, Germany) confocal microscope.

Clinical samples
NSCLC and normal adjacent tissues were harvested from 20 patients receiving digestive tract surgery at the First Hospital of Qinhuangdao between 2015 and 2019. All the included patients had not undergo any anti-tumor therapies prior to diagnosis.

Cell counting kit-8 (CCK-8) assay for cell proliferation [21]
Cells were seeded into 96-well plates at 5,000 cells/well. Thereafter, 20 μg/mL EVs isolated from the BEAS-2B, H1299 and H522 cells were incubated with A549 and H460 cells, and at 0, 24, 48 and 72 h, 10 μL CCK-8 solution (Dojindo, Kyushu Island, Japan) and 100 μL fresh medium were supplemented into each well. After 1 h of incubation at 37 ℃, the optical density (OD) value at 450 nm was measured using a microplate reader (Bio-Rad 680, Bio-Rad). The average OD values from each group were used to reflect the cell proliferation.

Flow cytometry-based cell cycle analysis [20]
Following 48-h EV treatment, cells were attained. Then Annexin V/propidium iodide (PI) assay kit (BD Biosciences, Franklin Lakes, NJ, USA) was utilized in the assay. In short, 0.5 μg/mL fluorescein isothiocyanate (FITC)-Annexin V and 5 μg/mL PI were applied for double staining at ambient temperature for 30 min. A flow cytometer (FACScan, BD Biosciences) equipped with Cell Quest 3.0 software was employed to analyze and calculate the percentage of cells at different stages. For cell cycle analysis, PI staining of the treated cells was implemented using CycleTEST TM PLUS DNA kit (BD Biosciences), and according to the provided instructions, the percentage of the cells at the G0/G1, G2/M and S phases was calculated and compared.

Flow cytometry-based cell apoptosis analysis
After treatment, the cells were detached with 0.25% trypsin without EDTA (YB15050057, Shanghai Yubo Biotechnology Co., Ltd., Shanghai, China), positioned in flow tubes and centrifuged, after which the supernatant was discarded. Subsequent to centrifugation, the supernatant was removed. The Annexin-V-FITC cell apoptosis detection kit (K201-100, BioVision Milpitas, CA, USA) was used, and Annexin-V-FITC, PI and 4-(2-hydroxyethyl)-1-piperazineëthanesulfonic acid (HEPES) buffer solution were prepared into Annexin-V-FITC/PI staining solution at the ratio of 1:2:50, based on the manufacturer's instructions. Every 100 μL staining solution was adopted to re-suspend 1 × 10 6 cells, which were mixed by shaking. Subsequent to 15-min incubation at ambient temperature, cells were supplemented with 1 mL HEPES (PB180325, Procell Life Science & Technology Co., Ltd., Wuhan, Hubei, China) and mixed by shaking. Subsequent to initiation by excitation at 488 nm, fluorescence measurement was implemented with emission filters at 525 nm (FITC) and 620 nm (PI) to determine cell apoptosis.

Scratch test
The 24-h cell incubation was conducted in a 6-well plate at 2.5 × 10 4 cells/cm 2 . Thereafter, the medium was removed, and a 10 μL sterile disposable straw was used to scratch. After 2 washes with PBS, the cells were subjected to additional culture in the FBS-free medium, and the images of each well were captured at 0 and 48 h post scratch under an inverted microscope, with 3 replicates set in each group. Image J software was adopted to determine the width of each scratch, and the cell migration ability was evaluated by comparing the scratch width of each group. The ratio of scratch healing rate = (scratch width at 0 h-scratch width at 48 h)/scratch width at 0 h × 100%.

Colony formation assay
A549 and H460 cells under different treatments were cultured in 6-well plates at 1000 cells/well for 2 weeks. Cells underwent 20-min 100% absolute methanol fixing. After fixation, 10-min cell staining was implemented with 0.1% crystal violet (G1064, Solarbio). The number of colonies with more than 50 cells was counted under a microscope (Leica, Germany). Counting was repeated three times for each group.

Human NSCLC Xenografts in nude mice
Establishment of a xenograft model was performed as previously described [22]. A total of 56 male BALB/c nude mice (aged 6-8 weeks old; Shanghai SLAC Laboratory Animal Co., Ltd., Shanghai, China) were selected for this study. Each mouse received a subcutaneous injection of 150 μL PBS encompassing 5 × 10 6 NSCLC cells into the unilateral dorsal flank area (n = 8/group). Seven days after transplantation, EVs (10 μg) were injected into the tumor every other day. Measurement of the subcutaneous tumors was performed every three days. Tumor volume was calculated: V = LW 2 /2, where L and W represent tumor length and width (in mm), respectively. The mice were euthanized after 4 weeks whereupon the tumors were harvested and weighed. The harvested tumor tissues were subjected to histologic analysis and immunohistochemical staining.

Histological analysis
Tumor tissues from the xenografts were fixed using formalin, paraffin-embedded, and then sliced. The slices were subjected to hematoxylin-eosin (HE) staining for histological evaluation and immunohistochemistry (IHC) staining for Ki-67 detection. HE staining was conducted using HE Staining Kit (Beyotime, Shanghai, China) as per the manufacturer's manuals. For Ki-67 detection, the tissue sections were stained in the light of the protocols of streptavidin biotin peroxidase complex-POD IHC kit with primary rabbit antibody against Ki-67 (A2094, abclonal, Woburn, MA, USA, 1:100).

Dual-luciferase reporter assay
Wild type (WT) or mutant type (MUT) 3′ untranslated region (3′UTR) of SUV39H1 encompassing the predicted target sequence of miR-744 were engineered into the pGL3 promoter vector (Genscript, Nanjing, China) and a dual-luciferase reporter assay was performed as previously reported [23]. Briefly, HEK293T cells (American Type Culture Collection, Manassas, VA, USA) were seeded into 24-well plates at 5 × 10 5 cells/well one day prior to transfection. The following day, 0.12 μg luciferase reporter vectors were co-transfected with miR-744 mimic or mimic-NC into HEK293T cells using Lipofectamine 3000 reagents. Forty-eight hours post-transfection, luciferase activity determination was implemented on a luminometer as per the provided protocol. Luciferase activity was calculated: luciferase activity = luciferase activity of firefly/luciferase activity of Renilla.

Reverse transcription quantitative polymerase chain reaction (RT-qPCR)
TRIzol reagents (Invitrogen) were applied for total RNA isolation from NSCLC cells. For mRNA quantification, cDNA was generated using Revert Aid first-strand cDNA synthesis kit (Fermentas) from 1 μg total RNA. RT-qPCR was implemented in an ABI PRISM 7900 System using SYBR Premix ExTaq ™ II (Takara, Kyoto, Japan) with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as a normalizer. miRNAs from EVs were extracted using SeraMir EVssome RNA Purification Kit (System Biosciences, San Francisco, CA, USA) underwent reverse transcription to generate cDNA using TaqMan micro-RNA assay kit (Applied Biosystems, Carlsbad, CA, USA). The samples were analyzed using FastStart Universal SYBR Green Master Mix (Roche, Indianapolis, IN, USA) with miRNA-specific forward primer (Sangon, Shanghai, China) and universal reverse primer provided by the TaqMan microRNA assay kit. U6 was adopted as a normalizer. All primers used in this study are listed in Table 1.

Statistical analysis
Data analysis was conducted using SPSS 21.0 software (IBM Corp. Armonk, NY, USA). Measurement data were summarized as mean ± standard deviation. Data comparisons between two groups were made using unpaired t-test, while data from multiple groups was compared using one-way analysis of variance (ANOVA) with Tukey's post hoc tests. Comparisons of data obtained at multiple time points were conducted using Bonferroni-corrected repeated measures ANOVA. A value of p < 0.05 was considered statistically significant.

Cancer cell-derived EVs promoted the proliferation of NSCLC in vivo and in vitro
EVs were isolated from the medium of normal bronchial epithelial cells (BEAS-2B) and NSCLC cells (H1299 and H522) and characterized. TEM was used to observe the morphology of the isolated EVs, which revealed that the EVs displayed round-shaped structures (Fig. 1a). The diameter of the isolated EVs analyzed using NTA was found to range from 40 to 120 nm (Fig. 1b). Further, western blot analysis showed that EVs expressed high levels of CD63, CD81, and TSG101, which are typical surface markers of EVs, but no Calnexin, which is a marker of endoplasmic reticulum (Fig. 1c). These results indicated the successful isolation of EVs. To confirm that these EVs could be delivered into host cells, we tracked the PKH67 labeled EVs (green fluorescence) by incubating them with A549 and H460 cells. After 4 h of incubation, green fluorescence could be observed in the cells (Fig. 1d), indicating that EVs had entered the host cells. We hypothesized that cancer cell-derived EVs could promote the proliferation of NSCLC cells and thus, we performed CCK-8 proliferation and colony formation assays. The viability of untreated A549 and H460 cells was comparable to that of A549 and H460 cells treated with normal bronchial epithelial cell-derived EVs (BEAS-2B-EVs). Compared with BEAS-2B-EVs, cancer cell-derived EVs (H1299-EVs or H522-EVs) increased NSCLC cell (A549 and H460) viability (Fig. 1e). In agreement, the results of the colony formation assay also showed that no obvious difference was observed in colony formation between untreated A549 and H460 cells and BEAS-2B-EV-treated A549 and H460 cells. Upon treatment with H1299-EVs or H522-EVs,   NSCLC cells (A549 and H460) formed more colonies than upon treatment with BEAS-2B-EVs (Fig. 1f ). As reflected by flow cytometry, the cell cycle distribution of untreated A549 and H460 cells was almost equivalent to that of BEAS-2B-EV-treated A549 and H460 cells. Moreover, H1299-EVs or H522-EVs led to greater number of cells arrested in the S phase and fewer cells arrested in the G1 phase (Fig. 1g), which explained, at least partially, why cancer cell-derived EVs promoted the cancer cell proliferation.
We further confirmed these findings in vivo using an established human NSCLC xenograft model. Seven days after the mice were injected with A549 cells, BEAS-2B-EVs or cancer cell-derived EVs (H1299-EVs or H522-EVs) were injected into the tumors, with the tumor growth recorded. The tumor growth curves showed that tumor growth of mice injected with untreated A549 cells were similar to that of mice injected with BEAS-2B-EVtreated A549 cells. The tumor growth was enhanced in mice treated with H1299-EVs or H522-EVs versus those in mice treated with BEAS-2B-EVs (Fig. 1h). The tumor volume and weight of mice treated with H1299-EVs or H522-EVs were higher than those in mice treated with BEAS-2B-EVs (Fig. 1i, j). In addition, we evaluated the proliferation of tumor cells by measuring the expression of proliferation marker, Ki-67, via IHC staining. The results showed that treatment with H1299-EVs or H522-EVs led to an increase in the positive expression of Ki-67 in cancer cells compared with treatment with BEAS-2B-EVs (Fig. 1k). Therefore, these findings suggested that cancer cell-derived EVs facilitated the proliferation of NSCLC cells in vitro and in vivo.

Expression of miR-744 is decreased in cancer cell-derived EVs
According to the most recent studies, miR-744, which exerts an inhibitory effect on NSCLC proliferation and migration [13], is poorly expressed in cancer cellderived EVs [14]. Here, we quantified the expression of miR-744 in BEAS-2B-EVs, H1299-EVs, and H522-EVs by RT-qPCR and found that the expression of miR-744 was lower in NSCLC cell (H1299 and H522)-derived EVs than that in normal bronchial epithelial cell (BEAS-2B)-derived EVs (Fig. 2a). These findings supported the notion that the different effects of normal cell-derived EVs and cancer cell-derived EVs on cancer cell biology are associated with the expression of miR-744. We then treated A549 and H460 cells with BEAS-2B-EVs, H1299-EVs, or H522-EVs and evaluated the expression of miR-744 in the cells. The result showed that as compared to treatment with BEAS-2B-EVs, treatment with H1299-EVs or H522-EVs led to lower levels of miR-744 in NSCLC cells (Fig. 2b). Furthermore, RT-qPCR revealed that miR-744 expression was reduced in NSCLC cell lines (H1299, A549, PC-9, H358, H522, and H460) when compared with BEAS-2B cells (Fig. 2c). The average expression value of miR-744-5p in lung cancer and normal samples included in TCGA was determined using a webbased database (http://bioin fo.life.hust.edu.cn/miR_path/ downl oad.html), and it was found that the expression of miR-744 was lower in lung cancer tissues than in normal tissues (Fig. 2d). For further verification, NSCLC clinical samples were collected and the miR-744 expression was quantified in 20 pairs of lung cancer tissues and normal adjacent tissues. The results verified that the expression of miR-744 was indeed decreased in tumor tissues (Fig. 2e). Collectively, these data indicated that miR-744 was expressed poorly in cancer cell-derived EVs.

Overexpression of cancer cell-derived EV-delivered miR-744 inhibits NSCLC cell proliferation in vitro
In order to confirm the hypothesis that aberrantly expressed miR-744 in cancer cell-derived EVs results in their different effects on NSCLC cell biology, the expression of miR-744 was modulated in NSCLC cells (H1299 and H522). We transfected H1299 and H522 cells with miR-744-mimic or mimic-NC and EVs were extracted from these transfcted cells (H1299-EVs-miR-744, H522-EVs-miR-744, H1299-EVs-NC, or H522-EVs-NC), after which the expression of miR-744 in these cells was measured by RT-qPCR. The resutls depicted that as compared to mimic-NC, miR-744-mimic increased the expression of miR-744 in both cells and EVs (Fig. 3a, b). A549 and H460 cells were treated with H1299-EVs-miR-744, H522-EVs-miR-744, H1299-EVs-NC, or H522-EVs-NC, which increased the expression of miR-744 in A549 and H460 cells (Fig. 3c). CCK-8 proliferation assay manifested that the viability of A549 and H460 cells treated with EVs from untreated H1299 and H522 cells did not differ from that of A549 and H460 cells treated with H1299-EVs-NC, or H522-EVs-NC, respectively (Fig. 3d). Besides, H1299-EVs-miR-744 and H522-EVs-miR-744 inhibited the viability (Fig. 3d) and colony formation (Fig. 3e) of A549 and H460 cells. Moreover, treatment of H1299-EVs-miR-744 or H522-EVs-miR-744 reduced the number of cells in the S phase and retained the cells in the G1 phase (Fig. 3f ). In addition, we investigated whether EVs from BEAS-2B cells transfected with miR-744 inhibitor could have an effect on the phenotype of A549 and H460 cells. The results showed that BEAS-2B-EVs-miR-744 inhibitor could significantly increase the viability, migration, and invasion but decrease apoptosis of A549 and H460 cells (Additional file 1: Figure S1A-G). Overall, these findings indicated that overexpressed miR-744 delivered by cancer cell-derived EVs contributed to decreased cancer cell proliferaiton in vitro.

Downregulation of miR-744 derived from cancer cell-derived EVs increased the expression of SUV39H1 in vitro
Having demonstrated that upregulation of miR-744 in NSCLC cell-derived EVs repressed the proliferation of cancer cells, our focus shifted to investigate the underlying molecular mechanisms. Two genes, SUV39H1 and BIN1, were predicted as targets of miR-744 by three different online miRNA databases (miRSearch, TargetScan, and miRWalk) (Fig. 4a). We investigated their expression levels in lung cancer tissues using the GEPIA website which analyzes the data from the TCGA database (Fig. 4b, c). The results indicated that the expression of SUV39H1, a histone methyltransferase, was higher in lung cancer tissue than that in normal adjacent tissues, which opposed the expression trend of miR-744. In agreement, the expression of SUV39H1 has been shown to be upregulated in NSCLC tissues [15]. These findings highlighted the possibility of SUV39H1 as one of the targets of miR-744. To confirm this, miR-744 targeting sequence of SUV39H1 was predicted using the TargetScan website, which revealed the binding sites between miR-744 and the 3′UTR of SUV39H1 mRNA (Fig. 4d). Thereafter, a dual-luciferase reporter gene assay showed that the luciferase activity of SUV39H1-3′UTR-WT was decreased in HEK293T cells transfected with miR-744 mimic, while that of SUV39H1-3′UTR-MUT remained unaffected (Fig. 4e). These data indicated miR-744 might target the SUV39H1 gene. To further confirm this finding, we modulated the expression levels of miR-744 in NSCLC cells (A549 and H460) (Fig. 4f ) and evaluated the expression of SUV39H1 at the transcriptional and translational levels in these cells (Fig. 4g, h). The results showed that miR-744 mimic resulted in decreased SUV39H1 mRNA (Fig. 4g) and protein expression (Fig. 4h), which was the opposite upon miR-744 inhibitor transfection. Importantly, we also noticed that the expression of miR-744 was lower in A549 and H460 cells treated with H1299-EVs or H522-EVs than that in cells with BEAS-2B-EV treatment (Fig. 4i). Consequently, the mRNA and protein expression levels of SUV39H1 were higher in A549 and H460 cells treated with H1299-EVs or H522-EVs than in cells following BEAS-2B-EV treatment (Fig. 4j, k). In addition, the isolated EVs from the BEAS-2B cells transfected with miR-744 inhibitor were used to incubate A549 and H460 cells, followed by the determination of SUV39H1 protein expression by western blot analysis. The results showed a significant increase of SUV39H1 protein expression in the A549 and H460 cells treated with EVs from miR-744 inhibitor-transfected BEAS-2B cells (Fig. 4l). These results suggested that miR-744 targeted SUV39H1, while downregulated miR-744 in cancer cell-derived EVs could lead to upregulated the expression of SUV39H1 in vitro in NSCLC cells.

miR-744 encapsulated by cancer cell-derived EVs suppresses the development of NSCLC by targeting SUV39H1 in vitro
Our aforementioned findings suggested that EV-derived miR-744 targeted SUV39H1 and downregulated its expression in NSCLC cells. Thus, it was deemed possible that EV-derived miR-744 regulates the development of NSCLC by targeting SUV39H1. To validate this hypothesis, we first investigated the expression of SUV39H1 in clinical samples of lung tumor and normal adjacent tissues by RT-qPCR. The results showed markedly higher expression of SUV39H1 in lung tumor tissues than normal adjacent tissues (Fig. 5a). To explore the role of SUV39H1 in NSCLC development, we silenced SUV39H1 in NSCLC cells (A549 and H460) using siRNAs. The silencing efficiency was evaluated by RT-qPCR (Fig. 5b) and Western blot analysis (Fig. 5c), which revealed that siRNAs against SUV39H1 diminished SUV39H1 expression in A549 and H460 cells.
Here, si-SUV39H1-1 exhibited superior silencing efficiency, and, was therefore, used in further experiments. Then A549 and H460 cells were treated with EVs from BEAS-2B or H1299 cells and then transfected with si-SUV39H1-1. A lower level of miR-744 was noted in A549 and H460 cells treated with H1299-EVs + si-SUV39H1-1 compared to cells treated with BEAS-2B-EVs + si-SUV39H1-1 (Fig. 5d). Moreover, compared to treatment with BEAS-2B-EVs + si-NC, SUV39H1 expression was elevated in A549 and H460 cells treated with H1299-EVs + si-NC, which was negated by treatment EVs. e and f RT-qPCR (e) and western blot analysis (f) measuring the expression of SUV39H1 in si-SUV39H1-1-transfected A549 and H460 cells treated with BEAS-2B-and H1299-derived EVs. g-i CCK-8 assay (g), colony formation assay (h), and cell cycle analysis (i) in si-SUV39H1-1-transfected A549 and H460 cells treated with BEAS-2B-, and H1299-derived EVs (* p < 0.05, vs. BEAS-2B-derived EVs, si-NC-transfected A549 and H460 cells treated with BEAS-2B-derived EVs; # p < 0.05, vs. si-NC-transfected A549 and H460 cells treated with H1299-derived EVs). j Western blots of SUV39H1 protein and its quantitation in tumor tissues of mice injected with A549 cells treated with BEAS-2B-, H522-and H1299-derived EVs (* p < 0.05, vs. mice injected with A549 cells treated with BEAS-2B-derived EVs). Data are summarized as mean ± standard deviation. Data from two groups were compared using unpaired t-test and data from multiple groups were compared by one-way ANOVA with Tukey's post hoc tests. Data obtained at different time points were compared by Bonferroni-corrected repeated measures ANOVA. Mice: n = 8/group. Each experiment was repeated three times with H1299-EVs + si-SUV39H1-1. In contrast to treatment with H1299-EVs + si-NC, treatment with si-SUV39H1-1 + BEAS-2B-EVs resulted in the decline of SUV39H1 expression in A549 and H460 cells (Fig. 5e, f ). Next, we investigated the effect of SUV39H1 silencing on EV-derived miR-744 mediated NSCLC cell proliferation using CCK-8 assay and colony formation assay. The results showed that the proliferation and colony formation in BEAS-2B-EV-treated A549 and H460 cells were retarded by si-SUV39H1-1 treatment. At the same time, A549 and H460 cells displayed enhanced proliferation and colony formation following H1299-EVs + si-NC treatment, when compared with that in BEAS-2B-EVs + si-NC treatment. However, this promotive effect on cell proliferation and colony formation was countered by further si-SUV39H1-1 treatment (Fig. 5g, h). Cell cycle analysis revealed that si-SUV39H1-1 reduced the A549 and H460 cells arrested in the S phase and increased the number of cells arrested in the G1 phase in the presence of BEAS-2B-EVs. Moreover, upon treatment with H1299-EVs + si-NC, greater numbers of A549 and H460 cells were retained in the S phase while cells in the G1 phase were reduced, when compared to treatment with BEAS-2B-EVs + si-NC. However, additional treatment of si-SUV39H1-1 diminished this effect (Fig. 5i). We also investigated SUV39H1 expression levels in A549 xenografts and found that upon treatment with H1299-EVs or H522-EVs, the expression of SUV39H1 was enhanced in tumor tissues of mice as compared to that upon treatment with BEAS-2B-EVs (Fig. 5j). In order to exclude the out-off target effect of the selected siRNA, A549 and H460 cells were treated with EVs from BEAS-2B or H1299 cells and then transfected with si-SUV39H1-2 to observe functional changes in A549 and H460 cells.
The results were consistent with those obtained following si-SUV39H1-1 transfection (Additional file 2: Figure  S2A-F). Altogether, these findings suggested that cancer cell-derived EV-encapsulated miR-744 reduced the development of NSCLC by targeting SUV39H1.

SUV39H1 regulates the expression of Smad9/BMP4 in vitro
SUV39H1 has been reported to inhibit the expression of Smad9 [16], which has been shown to suppress the expression of BMP4 [17]. Moreover, BMP4 has been documented to promote the initiation of NSCLC [18] and is highly expressed in NSCLC tumor tissues [24]. Thus, we speculated that reduced expression of miR-744 in cancer cell-derived EVs enhanced SUV39H1, which further suppressed the expression of Smad9 and consequently elevated the BMP4 expression, promoting the NSCLC progression. By analyzing the expression of Smad9 in NSCLC tumor tissues using data obtained from the TCGA database, we found that Smad9 was poorly expressed in NSCLC tumor tissue samples (Fig. 6a). We then evaluated the expression of SUV39H1, Smad9, and BMP4 inA549 and H460 cells treated with BEAS-2B-EVs or H1299-EVs by RT-qPCR and western blot analysis. The findings showed that cells treated with H1299-EVs expressed higher levels of SUV39H1 and BMP4 but lower Smad9 levels and extent of Smad9 phosphorylation (Fig. 6b, c, and Additional file 3: Figure S3A, B). We also observed that silencing SUV39H1 decreased the expression of BMP4 but increased Smad9 expression and the extent of Smad9 phosphorylation in A549 and H460 cells (Fig. 6d, e and Additional file 3: Figure S3C, D). This indicated that reduction of miR-744 resulted in upregulated SUV39H1 expression, which, in turn, could upregulate BMP4 expression by suppressing the expression of Smad9 in NSCLC cells. In order to validate our hypothesis that SUV39H1 regulates the expression of BMP4 in NSCLC cells by suppressing Smad9 expression, Smad9 was silenced in A549 and H460 cells and the knockdown efficiency was evaluated by RT-qPCR (Fig. 6f, Additional file 3: Fig. 3e) and western blot analysis (Fig. 6g, Additional file 3: Figure S3F). The results manifested that siRNAs against Smad9 diminished Smad9 expression in A549 and H460 cells, and that si-Smad9-1 exhibited the highest silencing efficiency and was therefore used in further experiments. We then investigated the expression of BMP4 in A549 and H460 cells in which SUV39H1 and/or Smad9 silencing was performed. We found that, SUV39H1 knockdown (si-SUV39H1) resulted in upregulated Smad9 expression and extent of Smad9 phosphorylation, while it downregulated the expression of SUV39H1 and BMP4 in A549 and H460 cells. However, silencing of Smad9 led to the opposite effects except for unchanged SUV39H1 expression in A549 and H460 cells. Moreover, concomitant silencing of SUV39H1 and Smad9 (si-SUV39H1 + si-Smad9) reduced the expression levels of Smad9 and extent of Smad9 phosphorylation but increased BMP4 expression in A549 and H460 cells in contrast to treatment with si-SUV39H1 (Fig. 6h, i and Additional file 3: Fig. S3G, H). These results suggested that the regulatory role of SUV39H1 on BMP4 depended upon Smad9 in vitro. by RT-qPCR (Fig. 7a, Additional file 4: Figure S4A) and western blot analysis (Fig. 7b, Additional file 4: Figure  S4B). We then treated OE-Smad9-transfected A549 and H460 cells with BEAS-2B-or H1299-derived EVs, and examined the expression of miR-744, SUV39H1, Smad9, and BMP4 by RT-qPCR (Fig. 7c, Additional file 4: Figure S4C) or western blot analysis (Fig. 7d, Additional file 4: Figure S4D). In the presence of BEAS-2B-EVs, OE-Smad9 treatment increased Smad9 and extent of Smad9 phosphorylation but decreased BMP4 expression in A549 and H460 cells. H1299-EVs + OE-NC-treated A549 and H460 cells displayed higher expression of SUV39H1 and BMP4, and lower expression of Smad9 and extent of Smad9 phosphorylation, compared to BEAS-2B-EVs + OE-NC-treated A549 and H460 cells. Interestingly, H1299-EVs-mediated upregulation of BMP4 could be diminished upon overexpression of Smad9. Thus, cancer cell-derived EVs could upregulate BMP4 expression by suppressing Smad9 expression in NSCLC cells. We then addressed whether cancer cell-derived EVs promoted NSCLC cell growth by regulating the Smad9/ BMP4 pathway and thus, CCK-8 proliferation assay (Fig. 7e, Additional file 4: Figure S4E) and colony formation assay (Fig. 7f, Additional file 4: Figure S4F) were performed. The results showed a reduction of viability and colony formation of A549 and H460 cells in response to   . c Expression of miR-744, SUV39H1, Smad9, and BMP4 determined by RT-qPCR in OE-Smad9-transfected A549 cells treated with EVs from BEAS-2B and H1299 cells. d Representative Western blots of SUV39H1, Smad9, and BMP4 and p-Smad9 and their quantitation in OE-Smad9-transfected A549 cells treated with EVs from BEAS-2B and H1299 cells. e Viability of OE-Smad9-transfected A549 cells treated with EVs from BEAS-2B and H1299 cells measured using CCK-8 assay. f Colony formation of OE-Smad9-transfected A549 cells treated with EVs from BEAS-2B and H1299 cells measured using colony formation assay (* p < 0.05, vs. si-NC-transfected A549 cells treated with EVs from BEAS-2B cells; # p < 0.05, vs. si-NC-transfected A549 cells treated with EVs from H1299 cells). g-j Results of in vivo tumor growth. A549 cells with or without Smad9 overexpression were injected into mice to establish the xenograft model and the EVs were injected into the tumors 7 days after transplantation. Tumor volume (g) was measured every 3 days; and final tumor volume (h) and weight (i) were measured, averaged, and compared after 4 weeks. The expression of proliferation marker, Ki-67 (j), was measured by IHC staining images were taken (* p < 0.05, vs. mice injected with A549 cells treated with BEAS-2B-EVs; # p < 0.05, vs. mice injected with A549 cells treated with H1299-EVs). k Knockdown efficiencies of 3 siRNAs targeting BMP4 determined by RT-qPCR in A549 cells. l Western blot analysis validating the knockdown efficiencies of 3 siRNAs targeting BMP4 in A549 cells (* p < 0.05, vs. A549 cells transfected with si-NC). m Expression of SUV39H1, Smad9, and BMP4 in si-BMP4-transfected A549 cells treated with EVs from BEAS-2B and H1299 cells. n Western blots of SUV39H1, Smad9, BMP4 and p-Smad9 and their quantitation in si-BMP4-transfected A549 cells treated with EVs from BEAS-2B and H1299 cells. o Viability of si-BMP4-transfected A549 cells treated with EVs from BEAS-2B and H1299 cells measured using CCK-8 assay. p Colony formation of si-BMP4-transfected A549 cells treated with EVs from BEAS-2B and H1299 cells measured using colony formation assay (* p < 0.05, vs. si-NC-transfected A549 cells treated with EVs from BEAS-2B cells; # p < 0.05, vs. si-NC-transfected A549 cells treated with EVs from H1299 cells). Data are summarized as mean ± standard deviation. Data from two groups were compared using unpaired t-test and data from multiple groups were assessed by one-way ANOVA with Tukey's tests. Data obtained at different time points were assessed by Bonferroni-corrected repeated measures ANOVA. Mice: n = 8/group. Each experiment was repeated three times thyroid cancer and liver cancer [31,32], which is consistent with our findings. Therefore, the effect of miR-744 on the function of NSCLC cells is worthy of further exploration.

Downregulated miR-744 in cancer cell-derived
A previous study has demonstrated that upregulated SUV39H1 facilitates cancer cell proliferation, enhances migratory potency of cancer cells, and suppresses cancer cell apoptosis [15]. In our study, we uncovered upregulated SUV39H1 expression in NSCLC tissues and cells, and showed that SUV39H1 was a direct target of miR-744. More importantly, we observed that normal lung epithelial cell-derived EVs treated cells express lower levels of SUV39H1 compared with cells treated by NSCLC cell-derived EVs. Our data also demonstrated that miR-744 shuttled by NSCLC cell-derived EVs arrested the proliferation of cancer cells by inhibiting SUV39H1 expression. Similarly, emerging evidence demonstrates that miRNAs play an important role in regulating cancer cell growth, invasion and metastasis by inhibiting the expression of their targets [33]. Interestingly, accruing evidence has shown that SUV39H1 negatively regulates the expression and activity of BMP4, which is closely related to NSCLC development and progression [18]. This evidence indicated the possibility that in NSCLC, upregulated SUV39H1 may contribute to tumor progression by regulating BMP4. In this study, we noticed that SUV39H1 silencing downregulated the expression of BMP4, which validated the aforementioned assumption. BMP4 expression has been shown to be negatively regulated by Smad9 [17]. Further evidence has suggested that SUV39H1 exerts a suppressive effect on Smad9 expression [16]. Moreover, the expression of BMP4 is higher in NSCLC tumors than that in paired adjacent non-tumor tissues [24]. In alignment with these studies, our data showed that when SUV39H1 was silenced in NSCLC cells, the expression of Smad9 was upregulated and the expression of BMP4 was downregulated. Consistently, NSCLC cell-derived EVs resulted in increased SUV39H1 and BMP4 expression levels, and decreased Smad9