The miR-101-3p mimics, inhibitor and control were obtained from Genepharma (Shanghai, China). Human recombinant VEGFA and VEGFA neutralizing antibody were purchased from R&D Systems (Minneapolis, MN). The antibodies against VEGFA, Vimentin, AKT, p-AKT, eNOS, p-eNOS, MMP-9, TWIST1, N-cadherin were purchased from Cell Signaling Technology (Beverly, MA). The antibody against β-actin was purchased from Sigma-Aldrich (St Louis, MO). The antibody against E-cadherin was purchased from BD Bioscience (San Jose, CA). The antibody against α-smooth muscle actin (α-SMA) was purchased from Abcam (Cambridge, United Kingdom).

Human lung cancer cell lines A549, H1299 and H661 were obtained from American Type Culture Collection (Manassas, VA). A549 cells were cultured in DMEM medium (GIBCO BRL, Grand Island, NY). H661 and H1299 cells were grown in RPMI1640 medium (GIBCO). Medium were supplemented with 10% fetal bovine serum (GIBCO). All cells were maintained at 37°C under 5% CO₂.

Lung cancer-associated fibroblasts (CAFs) and normal lung fibroblasts (NFs) were isolated and cultured, and conditioned medium (CM) were collected after 48 h according to the method previously described (Wang et al., 2017). The tumor tissues and adjacent normal tissues were obtained from the NSCLC patients underwent surgery at Tianjin Medical University General Hospital (TMUGH; Tianjin, China). The informed consents were obtained from patients. The study was approved by the Institutional Review Board of TMUGH.

Lung cancer cells were plated in a 96-wells plate at a density of 5 × 10³ cells/well. The cells were cultured with same amount of conditioned medium for 48 h and the cell viability was determined by the CCK-8 kit (Dojindo, Kumamoto, Japan).

Wound healing assays were performed to examine the ability of cell migration as previously described (Li X. et al., 2020). Cells were plated in 6-well plates and cultured with same amount of different CM. A plastic pipette tip was used to make a clean wound area across the well when cells grew to 100% confluence. Cells were allowed to migrate in the medium. Pictures were taken by a microscope (Nikon, Tokyo, Japan) after 24 h to estimate the wound closure.

Transwell assay can be a method to examine the cell invasion ability as previously described (Cao et al., 2020). Chambers (Corning, Tewksbury, MA) with 8.0 mm polycarbonate filter inserts were put in 24-well plates and coated with 40 μl Matrigel (BD Matrigel and serum-free DMEM were mixed in a ratio of 1:6) (BD Biosciences). 1-6 x 10⁴ cells suspended in serum-free DMEM medium were added to the upper compartment of the chamber. 700 μl CM were added to the lower compartment. After 48 h co-culture, the invaded cells were stained with 1% crystal violet and counted under a microscope.

Total RNA was extracted from cells using Trizol (Invitrogen). 2 μg RNA was used for reverse transcription by using TaKaRa kit (Dalian, China). For gene expression detection, Power SYBR Green Master Mix (ABI) was used and qPCR was performed on an ABI Prism 7900HT Sequence Detector System. The primers were designed by Primer Premier 5.0 and synthesized by BGI (Beijing, China). The sequences of the primers were listed in Table 1. The miRNAs Quantitation kit (GenePharma) was used to detect the expression of miR-101-3p, and U6 level was used for normalization.

Western blottings were performed as previously described (Zhang Q. et al., 2020). Cells were treated with RIPA lysis buffer (Beyotime Biotechnology, China) containing protease inhibitor (Sigma- Aldrich) for 30 min on ice. Cell lysate was separated by 12% SDS-PAGE and transferred onto nitrocellulose membranes. The membranes were blocked with 5% milk in Tris-buffered saline with 0.5% Tween 20 (TBST), then were probed with primary antibodies and HRP-conjugated secondary antibodies. The blots were visualized with the ECL Western Blotting System (ThermoFisher Scientific).

CAFs were transfected with miR-101-3p mimics, miR-NC, miR-101-3p inhibitor, inhibitor NC using lipofectamine 2000 (Invitrogen, Carlsbad, CA) following the manufacturer’s instruction. Briefly, the wild-type (wt) and mutant (mt) 3′-UTR of human VEGFA were cloned into pGL3 luciferase vector (Promega, Madison, WI). 293T cells were plated in 12-well plates and transfected with 100 nM miRNAs mimics or inhibitor, and 400 ng wt or mt VEGFA 3′-UTR plasmid, and 40 ng pRL-SV40 (Invitrogen) as an internal control. Luciferase activity were detected 24 h after transfection using Dual-luciferase Reporter Assay System (Promega).

CAFs were transfected with miR-101-3p mimics, miR-NC, miR-101-3p-inhibitor or inhibitor NC for 24 h. The supernatants were collected to detect VEGFA using an human VEGFA ELISA Kit (RayBiotech Life, GA) following the manufacturer’s instructions.

To establish stable cell lines, CAFs were transduced with lentiviruses expressing pre-miR-101-3p or negative control (CAFs/miR-101-3p, CAFs/miR-NC), and selected by puromycin. Female BALB/c nude mice (5-week-old) were obtained from the Cancer Institute of the Chinese Academy of Medical Science (Beijing, China). Mice were maintained under specific pathogen-free conditions, and randomly divided into 3 groups including 3 × 10⁶ A549-luc (expressing luciferase) alone, 3 × 10⁶ A549-luc cells and 6 × 10⁶ CAF-miR-NC (ratio 1:2) cells, or 3 × 10⁶ A549-luc and 6 × 10⁶ CAF-miR-101-3p cells (ratio 1:2). Cells were subcutaneously injected into the flank of mice. The experiment was conducted for 5 weeks. Tumor growth was measured by IVIS Imaging System (Xenogen Corporation, Alameda, CA). The animal experiments were conducted in accordance with the Tianjin Medical University Institutional Animal Care and Use Committee guidelines.

Data are presented as the means ± SEM. The differences between two groups were analyzed by student’s t-test, and the differences among more than two groups were analyzed by one-way analysis of variance. Analysis was performed using SPSS21.0. The P value less than 0.05 was considered statistically significant.

In order to investigate CAFs’ effect on NSCLC cells, we have performed experiments on NSCLC cell lines A549, H1299 and H661, which represent adenocarcinoma, carcinoma and large cell carcinoma, respectively. CAFs and NFs were isolated and characterized as our previous work (Wang et al., 2017). Both CAFs and NFs displayed spindle-like shape, and morphology showed no notable difference between CAFs and NFs (Figure 1A). We then detected the fibroblast markers by Western blotting analysis. Different from A549 cells which are epithelial cells, CAFs and NFs did not express E-cadherin, but expressed vimentin and α-SMA. Notably, CAFs possessed higher level of α-SMA than NFs (Figure 1B).

Next, NF-CM and CAF-CM were collected and cultured lung cancer cells, and cell proliferation was examined by CCK-8 kit. Figure 1C showed that both CAFs and NFs promoted lung cancer cell growth, and CAFs was more effective than NFs. To assess the effect of CAFs on migration and invasion of NSCLC cells, NSCLC cells were treated with NF-CM and CAF-CM, DMEM medium was used as a control, then wound healing assay and transwell assay was performed. The results of wound healing assay showed that the adhesive rates for A549, H1299 and H661 cells were increased to 1.80-, 1.24- and 1.56-fold of control by NFs, respectively. When cultured with CAF-CM, the migration levels were increased to 2.21-, 1.43- and 2.11-fold of control, respectively (Figures 1D,E). Consistent with the migration assay, the transwell assay indicated that both NFs and CAFs stimulated cell invasion, and CAFs were more effective (Figures 1F,G).

EMT is a key stage in tumor metastasis process (Welch and Hurst, 2019), we further investigated the EMT process when lung cancer cells were treated with NF-CM or CAF-CM. The expression of epithelial marker E-cadherin decreased whereas the expression of mesenchymal marker vimentin and N-cadherin increased (Figure 1H). Given that metastasis is mediated by metastasis-related genes, we then examined the effect of CAFs on metastasis-related genes MMP9 and Twist1. Figure 1I showed that NF-CM and CAF-CM upregulated the expression of MMP9 and Twist1, and CAFs had stronger effect. Taken together, our data suggested that CAFs enhanced the migration and invasion of NSCLC cells through regulating EMT process and metastasis-related genes.

Numerous studies have shown that miR-101-3p is dysregulated in various types of tumors and suppress tumor growth including NSCLC (Li Z. et al., 2020; Wang et al., 2020; Zeng et al., 2020). However, the role of miR-101-3p in CAFs is unclear. We first evaluated the miR-101-3p level in CAFs, and found that miR-101-3p was down-regulated in CAFs compared with NFs in NSCLC (Figure 2A). We then manipulated the miR-101-3p expression in CAFs by transfecting miR-101-3p mimics or inhibitor, and examined its effect on CAFs proliferation. MiR-101-3p had no remarkable effect after 24 h transfection, however, after 48 h transfection, miR-101-3p mimics suppressed CAFs growth whereas miR-101-3p inhibitor promoted CAFs growth (Figures 2B,C).

To determine whether miR-101-3p play a role in CAFs activation, we overexpressed or knocked down miR-101-3p, then detected the expression of CAFs markers including N-cadherin, vimentin and α-SMA. At transcription level, miR-101-3p repressed mRNA expression of N-cadherin, vimentin and α-SMA expression, respectively. On the contrary, miR-101-3p inhibition enhanced mRNA expression of these genes. In line with this finding, Western blotting results showed that miR-101-3p also affected CAFs markers expression at protein level (Figures 2D,E). Collectively, these data demonstrated that miR-101-3p repressed CAFs activation.

Given that miR-101-3p repressed CAFs activation, this provided rationale for further evaluating the role of miR-101-3p in CAFs’ effect on NSCLC cells. CAFs were transfected with miR-101-3p mimics or inhibitors, and CM was collected and cultured lung cancer cells. Figures 3A,B showed that miR-101-3p inhibited lung cancer cell proliferation whereas miR-101-3p down-regulation stimulated lung cancer cell proliferation. Furthermore, we verified the role of miR-101-3p in lung cancer cell invasion. When miR-101-3p was up-regulated in CAFs, the promoting effects of CAFs on lung cancer cell migration and invasion were mitigated significantly, on the contrary, when miR-101-3p was down-regulated in CAFs, the promoting effect of CAFs on lung cancer cell migration and invasion were enhanced (Figures 3C–F). Our results demonstrated that miR-101-3p mediated the effect of CAFs on migration and invasion of NSCLC cells.

Furthermore, we investigated the clinical implication of miR-101-3p level in NSCLC patients by using TCGA database. A cohort of 336 NSCLC specimens and 44 non-tumor specimens were analyzed. MiR-101-3p was significantly down-regulated in lung cancer tissues compared with non-cancerous tissues (Figure 3G). For Kaplan–Meier analysis, we found that low/medium-expression of miR-101-3p was associated with poor OS rate (Figure 3H, p = 0.0021). These findings suggested that miR-101-3p might serve as a prognostic marker for OS in NSCLC patients.

Since miR-101-3p played a role in CAFs’ effect on cell invasion, we further explored the underling mechanism. Given that CAFs regulated EMT process and metastasis-related genes in NSCLC cells (Figures 1H,I), we therefore investigated whether miR-101-3p was involved. MiR-101-3p was overexpressed in CAFs and conditioned medium was collected and treated lung cancer cells. qPCR and Western blot analysis showed that E-cadherin was up-regulated whereas vimentin and N-cadherin were down-regulated (Figures 4A,B). Figures 4C,D indicated that metastasis-related genes MMP9 and Twist1 were decreased at both mRNA and protein level. Moreover, we repressed miR-101-3p expression in CAFs by using miR-101-3p inhibitor, as expected, E-cadherin was down-regulated whereas vimentin and N-cadherin were up-regulated, and metastasis-related genes were stimulated (Figure 4E). These results demonstrated that miR-101-3p attenuated CAFs’ effect on EMT process and metastasis-related genes activation in lung cancer cells.

Cumulative studies have reported that AKT signaling pathway plays an important role in tumor metastasis (Wan et al., 2013). We next examined the activation of AKT signaling pathway in this setting. Compared with DMEM, CAF-CM elevated the phosphorylation of AKT and eNOS in lung cancer cells, meanwhile, the total level of AKT and eNOS remained unchanged. However, miR-101-3p diminished CAFs’ effect (Figure 4F). Interestingly, when miR-101-3p was repressed in CAFs, the CAFs’ effect on AKT/eNOS pathway activation was boosted (Figure 4G). Taken together, our data illustrated that miR-101-3p down-regulation in CAFs is responsible for facilitating EMT program and AKT/eNOS pathway activation in lung cancer cells.

MiRNA executes its function via modulating its target genes. To search for miR-101-3p targets, we applied bioinformatics algorithm including miRTarBase and TargetScan. Among many predicted targets, VEGFA drew our attention because VEGFA is reported to be involved in tumor progression (Matsumoto and Ema, 2014), we therefore selected VEGFA as a candidate target for further investigation.

To determine whether VEGFA is a direct target of miR-101-3p, we performed dual-luciferase assay. We first cloned the putative target sequence (wild type, wt) of VEGFA 3′UTR and mutant sequence (mt) into pGL3 luciferase vector (Figure 5A), then transfected these plasmids along with miR-101-3p mimics into 293T cells. The luciferase assay results showed that miR-101-3p mimics significantly reduced the luciferase activities of wt 3′UTR vector, but had no effect on mt 3′UTR vector. In contrast, miR-101-3p inhibitor increased luciferase activities of wt 3′UTR vector (Figure 5B).

To further validate VEGFA as a target of miR-101-3p, we investigated whether VEGFA expression was regulated by miR-101-3p. CAFs were transiently transfected with miR-101-3p mimics or inhibitor, and VEGFA expression was detected by qPCR and Western blot. Figures 5C,D indicated that miR-101-3p suppressed VEGFA expression at both mRNA and protein levels, and miR-101-3p inhibition enhanced VEGFA expression. Since VEGFA was a cytokine, we further examined whether CAFs secreted VEGFA or not by ELISA assay. Our data revealed that CAFs secreted VEGFA into CAF-CM, and miR-101-3p effectively reduced VEGFA secretion by CAFs (Figure 5E). Collectively, these results demonstrated that VEGFA was a direct target of miR-101-3p.

Since CAFs promote cancer progression via secreting cytokines and growth factors (Kalluri, 2016), and our study showed that CAFs secreted VEGFA (Figure 5E), we explored whether VEGFA play a role in CAFs mediated metastasis. We manipulated VEGFA level in CAF-CM by adding 20 ng/ml of human recombinant VEGFA or 5 μg/ml of VEGFA neutralizing antibody to cell cultures, and determined these effects on cell growth, migration and invasion.

Figure 6A showed that recombinant VEGFA boosted CAFs’ effect on cell proliferation, whereas VEGFA neutralizing antibody mitigated CAFs’ effect. Interestingly, when recombinant VEGFA was added to CAF-miR-101-3p-CM, the inhibitory effect of miR-101-3p on cell growth was reversed. We next examined the role of VEGFA in lung cancer cell migration and invasion. Both wound healing assay and transwell assay revealed that VEGFA enhanced CAFs’ effect on cell migration and invasion, and VEGFA reversed the inhibition of miR-101-3p on cell migration and invasion (Figures 6B–E). Moreover, the roles of VEGFA in EMT process and metastasis-related genes of lung cancer cells were also investigated. As expected, VEGFA promoted the enhancement of CAFs on EMT process and metastasis-related genes, and VEGFA attenuated miR-101-3p′s inhibitory effect (Figures 6F,G). Furthermore, Figure 6H showed that VEGFA also affected miR-101-3p′ effect on AKT/eNOS signaling pathway in lung cancer cells. Taken together, these results demonstrated that CAFs promoted metastasis potential of lung cancer cell via miR-101-3p regulated VEGFA secretion.

Since our study showed that CAFs stimulated lung cancer cell growth in vitro, we next evaluated their effect on tumor growth in vivo. Mice were randomly divided into 3 groups consisting of A549-Luc cells alone, A549-Luc cells combined with CAFs-miR-NC, and A549-Luc cells combined with CAFs-miR-101-3p. Cells were subcutaneously injected into the flank of mice. Tumor growth was observed in successive 5 weeks by IVIS Imaging System. CAFs stimulated tumor growth significantly, the average size of tumors in A549-Luc cells combined with CAFs-miR-NC group was 3.77-fold of A549-Luc cells alone group. Notably, the average size of tumors in A549-Luc cells combined with CAFs-miR-101-3p group was reduced to 1.87-fold of A549-Luc cells alone group, indicating that miR-101-3p mitigated CAFs’ effect (Figures 7A,B).

Next, we investigated the EMT status and AKT pathway in tumors. The tumor tissues were dissected and related molecules were examined by Western blottings. In line with our findings of in vitro study, CAFs promoted EMT progress, upregulated metastasis-related genes, and activated AKT pathway, meanwhile, miR-101-3p attenuated these effects (Figure 7C). Collectively, our results demonstrated that CAFs facilitated tumor growth through promoting EMT and AKT pathway by downregulation of miR-101-3p.