Exosomal miR-141 promotes tumor angiogenesis via KLF12 in small cell lung cancer

Both plasma (n = 77) and serum samples (n = 101) from SCLC patients and normal volunteers were collected from the National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital. All SCLC patients were pathologically confirmed to have SCLC by surgery or tissue aspiration biopsy, and all blood samples were collected before any antitumor therapies. Informed consent was acquired from every participant, and approval of the experimental protocol was obtained from the medical ethics committee of the National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital. Detailed information about the clinical and pathological features of all individuals is listed in Supplementary Table 1.

Exosomes from the SCLC cell culture medium (CM) were isolated by ultracentrifugation. Briefly, cells were refreshed with CM containing 10% exosome-removed fetal bovine serum (FBS) (ultracentrifuged at 110,000×g at 4 °C overnight) after washing with phosphate-buffered saline (PBS) twice. The CM was collected after the cells were cultured for 48 h. The obtained CM was processed for differential centrifugation at 500×g for 10 min, 2000×g for 10 min and 10,000×g for 30 min successively to remove dead cells, cell debris and large vesicles. The supernatant was then ultracentrifuged at 110,000×g for 70 min to collect exosome pellet followed by washing pellet with PBS once. All centrifugation procedures were conducted under the condition of 4 °C. The obtained exosomes were finally resuspended and preserved in PBS. The BCA assay was used to detect the protein concentration of exosomes. 15 μg of exosomes were used to stimulate 1 × 10⁵ HUVECs. To isolate exosomes from plasma, we used the Total Exosome Isolation (from plasma) Kit (Invitrogen, cat. No. 4484450) to precipitate exosomes according to the manufacturer’s instructions.

A total of 10 μl of exosomes were used for sample preparation to observe exosomes under a transmission electron microscope (JEOL-JEM1400, Tokyo, Japan). Briefly, after incubation at room temperature for 10 min, exosomes on a copper mesh were washed with sterile distilled water followed by incubation with uranyl-oxalate solution for 1 min. The copper mesh was then dried for 2 min under incandescent light and observed under a microscope. The exosome suspension was diluted by PBS to examine the size and quantity of exosomes under a ZetaView PMX 110 (Particle Metrix, Meerbusch, Germany). Exosome-specific proteins, namely, CD9 and TSG101, were detected by western blot.

The following were synthesized by GenePharma (Shanghai, China): both the mimics and inhibitors of miR-141, their respective negative controls and Cy3-labeled mimics of miR-141 (miR-141, 5′-UAACACUGUCUGGUAAAGAUGG-3′; negative control (NC), 5′-UUCUCCGAACGUGUCACGUTT-3′; Anti-miR-141, 5′-CCAUCUUUACCAGACAGUGUUA-3′; Anti-NC, 5′-CAGUACUUUUGUGUAGUACAA-3′). Small interfering RNAs (siRNAs) of KLF12 together with their negative control (siRNA-1, 5′-GGACUCGUUAUCUGUAGAUTT-3′; siRNA-2, 5′-GCACAUUAUCCAUCCCGUATT-3′; siRNA-NC, 5′-UUCUCCGAACGUGUCACGUTT-3′) were synthesized by Sangon Biotech (Shanghai, China). Plasmids for KLF12 overexpression were purchased from Vigene Biosciences (Shandong, China).

To construct the dual-luciferase reporter vectors of KLF12, three 3′ UTR fragments of KLF12 containing the predicted seeding sequences or the mutant ones were inserted into PGL3.0 vectors that expressed firefly luciferase. The plasmid expressing Renilla luciferase was used as the internal reference.

The lentiviral vectors expressing pri-miR-141 (OE-miR-141) or the control (OE-NC), together with three lentiviral packaging vectors were cotransfected into 293 T cells, and the lentivirus was collected after 48–72 h to infect SCLC cells. All transfection assays were carried out using Lipofectamine RNAiMAX Transfection Reagent (Invitrogen, cat. No. 13778150) or Lipofectamine 3000 Transfection Reagent (Invitrogen, cat. No. L3000015) following the manufacturer’s instructions.

SCLC cell lines (H446 and H1048) obtained from the American Type Culture Collection (ATCC) were maintained in the recommended CM in an incubator containing 5% CO₂ at 37 °C. Dulbecco’s modified Eagle’s medium (DMEM, Corning) containing 10% FBS (Gibco) and 1% penicillin and streptomycin (Gibco) was used to culture HUVECs and EAhy.926 cells. To develop SCLC sublines that stably overexpressed miR-141, SCLC cells were screened with puromycin after lentivirus infection for half a month and then maintained in CM with low-dose puromycin.

To visualize the internalization of exosomal miR-141 in HUVECs, exosomes isolated from H446 cells, which were transfected with Cy3-labeled mimics of miR-141 for 48 h, were labeled with PKH67 and incubated with HUVECs. After 12 h, exosome-treated HUVECs were photographed under a fluorescence microscope.

TRIzol reagent was utilized to extract total RNA from cells, patient serum and plasma exosomes, with cel-miR-39 added to serum or exosome solutions before chloroform. Cel-miR-39 was used as the external reference for the PCR analysis of miRNA in serum or exosomes, while RNU6 was used as the internal reference for the detection of miRNA in cells. The primers of miRNAs or mRNAs used in the present study were listed as follows: miR-141 (Forward, 5′-CGCTAACACTGTCTGGTAAAGATGG-3′); cel-miR-39 (Forward, 5′-TCACCGGGTGTAAATCAGCTTG-3′); RNU6 (Forward, 5′-CTCGCTTCGGCAGCACA-3′); KLF12 (Forward, 5′-CCTTTCCATAGCCAGAGCAGTAC-3′; Reverse, 5′-CTGGCGTCTTGTGCTCTCAATAC-3′); and GAPDH (Forward, 5′-GTCTCCTCTGACTTCAACAGCG-3′; Reverse, 5′-ACCACCCTGTTGCTGTAGCCAA-3′).

RIPA lysis reagent was used to isolate proteins from cells or exosomes, and the concentrations of proteins were measured using BCA assay method. Proteins with different molecular weights were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) followed by transferring onto PVDF membranes. The membranes containing separated proteins were washed and then incubated with the objective primary antibodies overnight at 4 °C. The appropriate HRP-conjugated secondary antibodies were incubated with the membranes the next day to detect the chemiluminescence signals. The primary antibodies used in the present study included KLF12 (Abcam, ab129459), GAPDH (CST, #5174), CD9 (Abcam, ab92726), and TSG101 (Abcam, ab83).

For cell proliferation assay, 1500 HUVECs or 2500 EAhy.926 cells were added into 96-well plates per well, Cell Counting Kit 8 (CCK8, Dojindo) was used to detect the cell viability every day for 3–4 days.

Transwell inserts (Corning) were utilized to evaluate the invasion and migration of vascular endothelial cells. Cells (6 × 10⁴ or 3 × 10⁴) were resuspended in FBS-depleted CM and placed in the upper chambers of 24-well inserts coated with or without Matrigel. After twenty hours, cells remaining on the insert upper surfaces, namely, that did not pass through the pores, were wiped off. Then, cells on the insert lower surfaces were carefully fixed with methyl alcohol followed by staining with Giemsa, and the invaded or migrated cells were counted under a microscope.

For tube formation assays, 50 μl of Matrigel was added to each well of a precooled 96-well plate and incubated at 37 °C for 1 h. HUVECs or EAhy.926 cells were trypsinized, and their concentrations were adjusted to 2 × 10⁵/ml or 5 × 10⁵/ml, respectively. A total of 100 μl of cell suspension was added to each well of the Matrigel-precoated 96-well plate, and the tubes that formed were imaged and counted under a microscope.

For aortic ring assays, the thoracic aortas of 8-week-old C57BL/6 mice were processed according to a previously published protocol [32]. Briefly, the dissected thoracic aorta was first flushed gently with Opti-MEM (Gibco) through its lumen to remove residual blood, and the aorta was then cut into 0.5-mm rings. The obtained aortic rings were incubated in a starvation medium, namely, Opti-MEM, at 37 °C overnight to equilibrate their response to stimulation. Nucleotide transfections or exosome treatments were conducted at the same time as starvation of the aortic rings. Afterward, aortic rings were embedded in collagen (collagen type I, rat tail, Millipore) and supplemented with Opti-MEM CM containing 2.5% FBS and 30 ng/ml VEGF (Peprotech, cat. No. 450–32-2). Microvessels that sprouted from the aortic rings were observed under a microscope.

HUVECs were cotransfected with reporter plasmids containing the 3’UTR of KLF12 and the Renilla luciferase vector, together with miR-141 mimics or the NC, using the Lipofectamine 3000 transfection reagent. Cell lysates were harvested 48 h later and added to 96-well plates, and luciferase activity was measured with a microplate reader.

A total of 500 μl of growth factor-reduced Matrigel (Corning, cat. No. 356231) mixed with exosomes was injected into the right axillae of C57BL/6 mice (6 weeks old) subcutaneously. The plugs that had formed were removed ten days later, fixed with formalin and sliced for immunohistochemistry (IHC). The presence of the blood vessel-specific marker CD31 was assessed with CD31 antibody (Abcam, ab28364) to measure the density of vessels generated in the plug.

To evaluate whether miR-141 could promote SCLC tumor growth by inducing angiogenesis, 5 × 10⁶ H1048 cells or 1 × 10⁷ H446 cells stably overexpressing miR-141 were injected into BALB/c nude mice (female, 6 weeks of age) subcutaneously. Tumor growth was monitored two times per week. The tumor MVD was evaluated by IHC of CD31. Mouse plasma exosomes were isolated using the Total Exosome Isolation (from plasma) Kit, and exosomal miR-141 was measured by qRT-PCR.

GraphPad Prism software was utilized to perform all analysis of experimental data, with the results presented as the mean ± SD. The statistical differences between two groups were detected by Student’s t-test, and one-way ANOVA was applied for three or more groups. Statistical results with *P < 0.05, **P < 0.01, ***P < 0.001, or ****P < 0.0001, were considered to be statistically significant.

To better understand the clinical importance of miR-141 in patients with SCLC, we measured the expression of miR-141 in both plasma and serum samples. A total of 77 plasma samples from 53 SCLC patients and 24 healthy volunteers were used for exosome isolation followed by RNA extraction and qRT-PCR analysis. We found that miR-141 levels in plasma exosomes were significantly higher in SCLC patients than in healthy controls (Fig. 1A). More importantly, the level of plasm exosomal miR-141 in 12 SCLC patients with stage IV disease was much higher than SCLC patients with stage I-III disease (Fig. 1B), SCLC patients with extensive disease also showed much higher levels of exosomal miR-141 than whom with limited disease (Fig. 1C). To further explore the level of circulating miR-141 in SCLC patients, another set of 101 serum samples from 69 SCLC patients and 32 healthy volunteers was used for total RNA isolation and detection. Consistent with the miR-141 levels in plasma exosomes, those of miR-141 in the serum of SCLC patients were aberrantly increased (Fig. 1D), especially in stage IV patients (Fig. 1E), and SCLC patients with extensive disease also showed much higher levels of circulating miR-141 in serum than whom with limited disease (Fig. 1F). Additionally, clinical correlation analysis demonstrated that higher miR-141 expression was related to larger tumor size, more lymph node metastasis, more distant metastasis and extensive stage (Table 1). As we know, tumor-associated vasculatures transport oxygen and nutrient to tumor cells to promote tumor cell survival and growth and provide opportunities for tumor cell metastasis, which indicated that angiogenesis was closely related to tumor development and progression [5]. Thus, our data implies the potential importance of miR-141 in SCLC-associated angiogenesis.

To determine the function of exosomal miR-141 in SCLC-associated angiogenesis, we treated HUVECs with exosomes derived from SCLC cells to study miR-141 effects on the capacities of cell growth, invasion, migration and tube formation. First, mimics of miR-141 (miR-141) and its NC were transfected into H446 or H1048 cells, which have relatively low expressions of endogenous miR-141 (Supplementary Fig. 1A) and low levels of cell-secreted exosomal miR-141 (Supplementary Fig. 1B). The upregulation of miR-141 in H446 and H1048 cells led to increased levels of miR-141 in exosomes (Fig. 2A-B) with no significant changes in the quantity of released exosomes (Fig. 2C). Vesicles isolated from the CM of SCLC cells were processed for visualization of their appearance by TEM (Fig. 2D), size determination and quantification by nanoparticle tracking technology (Fig. 2C) and protein analysis by western blot (Fig. 2E) to verify the existence of exosomes. Then, we treated HUVECs with the obtained exosomes in which miR-141 levels were upregulated (miR-141-EXO) or the negative control exosomes (NC-EXO) and evaluated their effects on HUVECs. We found that HUVECs incubated with miR-141-EXO grew faster (Supplementary Fig. 2A-B), showed increased migration and invasion (Fig. 2F, Supplementary Fig. 2C-E) and increased tube formation (Fig. 2G). Pretransfection of HUVECs with inhibitors of miR-141 (anti-miR-141) antagonized the promotional effect of miR-141 on cell proliferation, invasion, migration and tube formation (Fig. 2F-G, Supplementary Fig. 2A-E). Moreover, aortic rings that were treated with miR-141-EXO sprouted more and longer microvessels than those treated with NC-EXO, and the increased sprouting was reversed by inhibition of miR-141 (Fig. 2H). Consistent with this observation, CM derived from H446 and H1048 cells transfected with miR-141 mimics also promoted HUVEC proliferation (Supplementary Fig. 3A-B), invasion, migration (Supplementary Fig. 3C) and tube formation (Supplementary Fig. 3D) and increased the number of microvessels sprouting from the aortic rings (Supplementary Fig. 3E). Besides, we treated HUVECs with the plasma of SCLC patients or healthy volunteers to evaluate their effect on the angiogenesis of HUVECs. As shown in Supplementary Fig. 3F-H, compared with the plasma of healthy volunteers, the plasma derived from SCLC patients significantly stimulated the proliferation, invasion, migration and tube formation of HUVECs in vitro.

To further determine whether exosomal miR-141 has direct effects on vascular endothelial cells, we transfected mimics of miR-141 directly into HUVECs and repeated the functional assays. As expected, miR-141 upregulation (Supplementary Fig. 4A) strengthened HUVEC proliferation (Supplementary Fig. 4B), invasion, migration (Supplementary Fig. 4C) and tube formation (Supplementary Fig. 4D). Mouse aortic rings transfected with miR-141 mimics also showed more sprouting microvessels than those transfected with the NC (Supplementary Fig. 4E). To make the result more convincing, another vascular endothelial cell line, EAhy.926, was used to study the impact of miR-141 on angiogenesis. Like the HUVECs, EAhy.926 cells overexpressing miR-141 (Supplementary Fig. 5A) showed enhanced cell proliferation (Supplementary Fig. 5B), invasion, migration (Supplementary Fig. 5C) and tube formation (Supplementary Fig. 5D). Taken together, our results reveal that miR-141 in SCLC cell-secreted exosomes plays critical roles in promoting vascular endothelial cell proliferation, migration, tube formation and angiogenesis in vitro.

Based on the significant effects of exosomal miR-141 in angiogenesis in vitro, we further explored its function in neovascularization in vivo. Exosomes with miR-141 upregulation were isolated from miR-141-mimic-transfected H446 or H1048 cells. Then, the exosomes were added to Matrigel and the mixtures were injected into the axillae of mice subcutaneously. Formed Matrigel plugs were collected and analyzed 10 days later. The results showed that the Matrigel plugs mixed with miR-141-EXO had more new blood vessels than those mixed with NC-EXO (Fig. 3A). We also conducted IHC staining of CD31 to accurately quantify neovascularization, and the results were consistent with those above, namely, there were more microvessels in the plugs containing miR-141-EXO than in those containing NC-EXO (Fig. 3B-C). This in vivo experiment provides additional evidence of the strong proangiogenic effect of SCLC cell-derived exosomal miR-141.

We speculated that SCLC-derived exosomal miR-141 was delivered to and internalized by HUVECs to initiate SCLC-related neovascularization. Thus, we conducted a series of experiments to validate this hypothesis. First, we measured miR-141 expression in HUVECs after incubating them with SCLC cell-derived exosomes and found that miR-141 expression in HUVECs was upregulated after incubation with miR-141-EXO (Fig. 4A-B). Pretreatment of exosomes with Annexin V, an inhibitor of exosome absorption, antagonized the increase in miR-141 in HUVECs (Fig. 4A-B), suggesting that the internalization of exosomes in HUVECs leads to the upregulation of miR-141. Additionally, pri-miR-141 expression in HUVECs remained unchanged after incubation with exosomes (Fig. 4C) and pretreatment of HUVECs with an inhibitor of RNA transcription, 5,6-Dichloro-1-β-D-ribofuranosylbenzimidazole (DRB), could not reverse the upregulation of miR-141 in HUVECs after incubation with miR-141-EXO (Fig. 4A-B), indicating that the upregulated miR-141 did not originated from the transcription of pri-miR-141. Furthermore, we prepared PKH67-labeled exosomes that were isolated from Cy3-miR-141 mimic-transfected H446 cells, treated HUVECs with these exosomes, and took photographs of these cells under a fluorescence microscope. As shown in Fig. 4D, miR-141 with red fluorescence was packaged into green exosomes and absorbed by HUVECs via exosomes. Together, these findings demonstrate that miR-141 can be secreted by SCLC cells and delivered to HUVECs via exosomes to mediate its functions in angiogenesis.

We next explored the molecular mechanisms by which secreted miR-141 promoted HUVEC migration and angiogenesis. We searched three databases, including PicTar, miRDB and TargetScan, to predict target genes of miRNAs, and a total of 262 genes common to the three databases were noted as predicted targets of miR-141 (Fig. 5A). The target genes are listed in Supplementary Table 2, with KLF12 being one of the top candidates. Considering that KLF family members have been reported to have various effects on vascular endothelial cells and tumor-associated angiogenesis [29, 33, 34], we speculated that KLF12 might play a critical role in SCLC angiogenesis and that miR-141 promotes neovascularization via KLF12. To confirm our hypothesis, we first detected the level of KLF12 in HUVECs after transfection with the mimics of miR-141. The results showed that mimics of miR-141 reduced the mRNA and protein levels of KLF12 in HUVECs (Fig. 5B), which suggested that KLF12 might be regulated by miR-141. Therefore, we constructed luciferase plasmids containing the wild-type or mutant target sequences of the KLF12 3’UTR (Fig. 5C) for dual-fluorescence assays. As shown in Fig. 5D, the luciferase activities of three reporter plasmids carrying wild-type target sequences were significantly inhibited by miR-141, whereas this inhibition was abrogated when the predicted sequences were mutated (Fig. 5D). These results indicate that KLF12 is directly regulated by miR-141, but whether miR-141 promoted angiogenesis via KLF12 was still unclear.

Then, we synthesized siRNAs against KLF12 to investigate its function in HUVECs. Interestingly, siRNA-mediated knockdown of KLF12 (Supplementary Fig. 6A) significantly promoted HUVEC proliferation (Supplementary Fig. 6B) and increased the number of HUVECs that had invaded or migrated through transwell inserts (Supplementary Fig. 6C), the number of tubes that formed on the Matrigel and the number of microvessels sprouting from aortic rings were also increased after KLF12 knockdown (Supplementary Fig. 6D and 6E). Next, we cotransfected HUVECs with a KLF12 overexpression plasmid that did not contain the KLF12 3’UTR and the miR-141 mimics to determine whether the proangiogenic effects of miR-141 could be reversed by KLF12. The results showed that the enhancements in HUVEC proliferation, invasion, migration and tube formation as well as the increase in the number of sprouting microvessels stimulated by miR-141 were abrogated by KLF12 overexpression (Supplementary Fig. 6F, Fig. 5E-G). Taken together, these data demonstrate that KLF12 is a direct target gene of miR-141 and that miR-141 promotes HUVEC proliferation, migration and angiogenesis by regulating the level of KLF12.

To further investigate whether miR-141 could promote tumor growth in vivo by inducing neovascularization, we subcutaneously injected miR-141-overexpressing H1048 and H446 cells to the right dorsal flanks of BALB/c nude mice. We found that the tumors that had developed from miR-141-overexpressing SCLC cells were larger, heavier and had more vessels on their surfaces than those that had developed from the NC cells (Fig. 6A-F). To clarify whether this result was attribute to the direct influence of miR-141 on SCLC cells, we conducted CCK8 and transwell assays between H1048/OE-miR-141 and H1048/OE-NC cells or H446/OE-miR-141 and H446/OE-NC cells. As shown in Supplementary Fig. 7A-7D, the proliferation and migration capabilities of both H1048 and H446 cells were not affected by miR-141, which indicated that the increased tumor growth in vivo was not owing to the influence of miR-141 on SCLC cells. Next, we detected the level of miR-141 in mouse plasma exosomes and found that the level of miR-141 in the plasma exosomes derived from the mice with miR-141-overexpressing tumors was significantly increased compared with that of plasma exosomes derived from mice with NC tumors (Fig. 6G and H). We next evaluated the microvessels in the tumor nodules by staining for CD31 and found that both the percentage of CD31-positive cells and the MVD were higher in miR-141-overexpressing tumors than in control tumors (Fig. 6I-K). These results imply that exosomal miR-141 might facilitate SCLC tumor growth in vivo by inducing tumor angiogenesis.