Exosomes loaded with the anti-cancer molecule mir-1-3p inhibit intrapulmonary colonization and growth of human esophageal squamous carcinoma cells

Abstract

Background

The overall prognosis of patients with esophageal cancer (EC) is extremely poor. There is an urgent need to develop innovative therapeutic strategies. This study will investigate the anti-cancer effects of exosomes loaded with specific anti-cancer microRNAs in vivo and in vitro.

Methods

Specific miRNAs that were significantly down-regulated in EC tissues were screened using the miRNAs profiling data of human EC tissue samples in TCGA, and the role of their exogenous expression in the proliferation and migration of human EC cell lines, KYSE150 and Eca109, were detected using CCK-8 and Transwell assays. Exosomes were loaded with miRNAs using electroporation.

Results

The expression of miR-1-3p was significantly down-regulated in human EC tissues with potential anti-cancer effects. Exosomes loaded with miR-1-3p significantly inhibited the proliferation, migration and invasion of KYSE150 and Eca109 cells in vitro, as well as the intrapulmonary colonization and growth of KYSE150 cells in vivo. In addition, miR-1-3p could directly bind to the 3’UTR of the transcription factor E2F5 mRNA, down-regulate the protein expression of E2F5, and inhibit the activation of the MAPK/ERK signaling pathway.

Conclusion

Exosomes loaded with miR-1-3p may be applicable to the treatment of EC.

Background

According to the latest data from Global Cancer Observatory (https://gco.iarc.fr/), esophageal cancer (EC) is the seventh leading cause of cancer-related deaths, with 445,391 deaths reported globally in 2022. The two main subgroups of EC are squamous cell carcinoma (ESCC) and adenocarcinoma (EAC), which account for 95% of all cases. ESCC accounts for 90% of all cases, and is predominantly found in the Central Asian esophageal cancer belt (extending from Iran to China), Africa, and South America. EAC occurs predominantly in Western countries with high Human Development Index (HDI) scores, but its incidence in other parts of the world has been steadily increasing over the last few years [1]. ESCC and EAC differ in tissue sites of occurrence, risk factors, and molecular mechanisms of progression [2].

Multidisciplinary treatments have been developed for EC. Endoscopic resection and ablation are usually indicated for stage 0 and T1a patients, surgical resection alone may be indicated for stage T1b patients, and surgical resection combined with neoadjuvant chemotherapy or neoadjuvant radiochemotherapy is usually indicated for patients with stage T2 or higher. However, early EC is usually asymptomatic, and is generally discovered incidentally during endoscopic examination for other diseases or indications [3]. Patients are usually in locally advanced or metastatic disease by the time they become aware of any symptoms, including dysphagia, weight loss, chest discomfort, and worsening heartburn. Currently, immunotherapy in combination with chemotherapy is recommended as the first-line treatment for advanced EC. However, the overall prognosis of EC patients remains poor, with a 5-year relative survival rate of approximately 26% for locally advanced disease and 5% for distant metastases [4]. In addition, patients with stage IV disease usually receive only treated palliative care. Currently developed molecularly targeted drugs have also failed to improve the survival of EC patients [5]. Therefore, there is an urgent need to develop innovative therapeutic approaches to improve the prognosis of EC patients.

Gene therapy is the delivery of genetic material (specific DNA, mRNA, microRNA, small interfering RNA, cyclic RNA, etc.) into target cells to repair, add, or delete disease-causing genes. Gene therapy makes it possible to control cancer, although it is not yet a substitute for conventional therapies [6]. miR-1-3p is encoded by the miR-1-2 gene located on chromosome 18q11.2. It is commonly expressed silently in cancer tissues, and its exogenous expression has shown significant tumor suppression in a variety of tumor cells and animal models, which is considered a promising gene therapy [7]. However, nucleic acids are negatively charged and hydrophilic, cannot directly penetrate cell membranes, and are highly unstable and susceptible to enzymatic degradation. Therefore, the most critical aspect in gene therapy is the delivery system of nucleic acids.

Exosomes are extracellular vesicles with a particle size of 30–200 nm surrounded by a bilayer of lipids that is secreted by almost any cell. It can travel in the blood circulation and various body fluids, and mediate intercellular communication within an organism. Exosomes have several properties and advantages over conventional nanoparticles (viruses and synthetic nanocarriers), making them a highly desirable natural nanogene carrier [8]. First, exosomes are produced by autologous cells, have extremely low cytotoxicity and immunogenicity, and are not phagocytosed by monocytes and macrophages, unaffected by opsonin and coagulation factors, and are highly stable in the bloodstream, which facilitates the administration of drugs via intravenous injection. Secondly, exosomes can easily cross blood vessels and extracellular matrix in tumor tissues to target tumor cells. In addition to lipid-soluble drugs, their hydrophilic core allows them to carry water-soluble nucleic acids. In addition, the unique vesicular structure of exosomes ensures that the contained nucleic acids are highly stable in a variety of environments. This stability ensures the production, transportation and use of exosomes as drug delivery vehicles. Finally, its bilayer lipid membrane structure combined with specific targeting modifications allows exosomes to easily dock with the surface of target cells and be taken up by the cells.

In this study, we prepared miR-1-3p-containing exosomes, and reported their key therapeutic role in hindering the development of EC and the downstream signaling pathways regulated by their action.

Materials and methods

Download and analysis of miRNAs profiling data

Loop miRNAs profiling data of 197 samples (including 184 human EC tissue samples and 13 human normal esophageal tissue samples) were downloaded from UCSC Xena (https://xenabrowser.net/datapages/), which derived from TCGA [9]. Loop miRNAs that were significantly differentially expressed in EC tissues compared to normal samples were screened using |FC| > 1.5 & adj.P.Val < 0.01 as the screening threshold, and the results were displayed as volcano plots and bidirectional clustering heatmaps. The differentially expressed loop miRNAs were relabeled using the R language miRBaseConverter package to obtain the names of mature miRNAs [10].

Cell culture and transfection

Human ESCC cell lines (Eca109 and KYSE150) and HEK293 cells were obtained from the National Biomedical Experimental Cell Repository (Shanghai, China), and cultured in the Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS. The culture environment was a sterile, humidified cell culture incubator at 37 °C with 95% air and 5% CO₂.

The E2F5 cDNA sequence was cloned into the pcDNA3.1 plasmid to construct the expression plasmid. miRNA mimics and inhibitor sequences are listed in Table 1. RNA sequences (50 nM) and plasmid (2 µg) were transfected into the cells using siTran 2.0 siRNA transfection reagent (TT320002, ORIGENE). Control cells (CON) were not subjected to any transfection and treatment, only cultured normally. Negative control cells (NC) were transfected with nonsense miRNA sequences (and pcDNA3.1 empty vector).

Table 1 RNA sequence information

Extraction and enrichment, characterization, mir-1-3p electrotransformation and PKH67 labeling of exosomes

Exosomes were collected from cell cultures using ultracentrifugation. Briefly, when HEK293 cells grew to 70% confluence, the medium was replaced with DMEM supplemented with 10% exosome-free FBS (JKF1001-100, JEKAIYER, Shandong, China). After 48 h of incubation, the medium was collected and centrifuged at 2,000 g for 15 min. After removing the precipitate, the supernatant sample was continued to be centrifuged at 10,000 g for 30 min. After removal the precipitate, the supernatant sample then was centrifuged at 120,000 g for 70 min (using an Optima XPN-100 Ultracentrifuge, BECKMAN), at which the main components of the precipitate were exosomes. The precipitate was resuspended in PBS and centrifuged again at 120,000 g for 70 min to remove adsorbed impurities. The precipitate obtained was exosomes, which were solubilized in PBS and stored at -80 °C. Exosomes were identified by examining morphology, size, and expression of positive markers using transmission electron microscopy (TEM) (HT7800, HITACHI), a nanoparticle size detector (ZetaView PMX110, Particle Metrix), and western blot assay.

The miR-1-3p mimics were loaded into exosomes using an electrotransfer method. Briefly, exosomes and miR-1-3p mimics were ice-bathed for 10 min in advance. The miR-1-3p mimics were added to 0.5 mg/ml of exosome solution, and the final concentration of miR-1-3p mimics after addition was 10 nM. After mixing, the mimics were transferred to a pre-cooled 2 mm cuvette for 5 pulses (400 V, 125 uF, 5 ms) (the electrotransfer apparatus used was the Gene Pulser Xcell, Bio-Rad). The electroporation was followed by an ice bath for 10 min. The mixture was diluted with PBS and then centrifuged at 120,000 g for 70 min to remove free miR-1-3p mimics, and the precipitate was the exosome after electroporation (Exo-miR-1-3p). Exosomes electrotransformed with nonsense miRNAs were used as negative controls (Exo-NC).

Exosomes were labeled with PKH67 green fluorescent dye (MIDI67-1KT, Sigma) according to the procedure in the manufacturer’s instructions. DAPI (KGA215-10, KeyGEN BioTECH) was used for nuclear staining.

Western blot

Cells or exosomes were lysed using a Strong RIPA Lysis Buffer (CW2333, CWBIO) supplemented with protease inhibitors and phosphatase inhibitors on ice and at 4 °C ambient temperature for 20 min. The lysate was centrifuged for 10 min at 4 °C and 14,000 g, and the supernatant obtained was the total protein solution. After detection of protein concentration by BCA method (KGP902, KeyGEN BioTECH), the protein solution was added with up-sampling buffer and boiled in boiling water for 5 min to denaturation. 20 µg proteins of each group were electrophoresed by SDS-PAGE (10% separating gel concentration) at a constant current of 80 mA for 30 min for concentrated gel and 120 mA for 1 h for separating gel. Proteins separated by molecular weight size in the gels were electrotransferred (200 mA, 50 min) onto PVDF membranes using the wet-transferring method. 5% skimmed milk was used for closure, the membrane was incubated with primary antibodies overnight at 4 °C, and then incubated with secondary antibody at room temperature for 1 h. Luminescence reactions were performed using ultrasensitive ECL reagent (PK10002, Proteintech).

Antibodies information used in this study is provided below: anti-TSG101 (1:2000, 14497-1-AP, Proteintech), anti-CD81 (66866-1-Ig), anti-CD9 (60232-1-Ig), anti-E2F5 (1:2000, bs-1734R, Bioss), anti-GAPDH (1:5000, 60004-1-Ig, Proteintech), anti-p-p38 (1:1000, 4511, CST), anti-p38 (1:5000, 66234-1-lg, Proteintech), anti-p-ERK (1:1000, AP0974, ABclonal), anti-ERK (1:700, A4782, ABclonal), anti-p-Akt (1:2000, 28731-1-AP, Proteintech), anti-p-mTOR (1:500, AF3308, Affinity), HRP Goat anti-Rabbit IgG (1: 5000, AS014, ABclonal), HRP Goat anti-Mouse IgG (1: 5000, AS003, ABclonal).

RT-qPCR

Total RNA was extracted using TRIzon Reagent (CW0580, CWBIO) and reverse transcribed using HiFiScript cDNA Synthesis Kit (CW2569, CWBIO). RT-qPCR reactions were performed using UltraSYBR Mixture (CW0957, CWBIO). The primer sequences used were: U6 Forward 5’-GTGCTCGCTTCGGCAGCACATAT-3’, Reverse 5’-AGTGCAGGGTCCGAGGTATT-3’; miR-1-3p Forward 5’- GCGCGTGGAATGTAAAGAAGT-3’, Reverse 5’-AGTGCAGGGTCCGAGGTATT-3’; E2F5 Forward 5’-CTCTGCAGCAGACATCAGCT-3’, Reverse 5’-TGTAGTCATCTGCCGGGGTA-3’; β-actin Forward 5’-CCACCATGTACCCTGGCATT’, Reverse 5’-CGGACTCGTCATACTCCTGC-3’. The relative levels of target genes were calculated by the 2^−∆∆Ct method using U6 or β-actin as an internal reference.

CCK-8 assay

2000 cells were seeded into each well of a 96-well plate. After incubation for the indicated times (24, 48, 72, 96 h), 10 µl of 5 mg/mL CCK-8 reagent (HY-K0301, MCE) was added to the medium, and the cells were further incubated at 37 °C for 2 h. Subsequently, the absorbance (OD) of each well at 450 nm was measured using a microplate reader (tecan F50). Finally, the growth curve of the cells was plotted based on the OD values.

Transwell assay

Cells were incubated with 20 µg/ml, 50 µg/ml or 100 µg/ml of exosomes loaded with miRNA for 48 h. Subsequently, the cells were digested and resuspended, and the invasion and migration abilities were examined using Transwell chambers. For invasion assays, the membrane of the Transwell chamber needed to be coated with Matrigel gel (serum-free DMEM 1:6 dilution) (incubated at 37 °C for 4 h), whereas for migration assays it was not required, and the rest of the experimental steps were consistent. 1 × 10⁴ cells suspended in 100 µl serum-free medium were added to the upper chamber, and the lower chamber was supplemented with 600 µl of DMEM containing 10% FBS. After 24 h of incubation, the number of cells that had invaded or migrated to the lower chamber was detected.

Mouse model and histopathological analysis

Fifteen BALB/c nude mice (SPF, female, 5 weeks old) were randomly divided into 3 groups of 5 mice each after 1 week of acclimation feeding. 1 × 10⁶ KYSE150 cells/0.1 ml cell suspension was injected into mice via tail vein, and then 0.1 ml of exosomes (2 µg/g) electrotransformed with miRNA was injected via tail vein 2 days later. The Exo-NC group was injected with exosomes electrotransformed with meaningless miRNA, the Exo-miR-1-3p group was injected with exosomes electrotransformed with miR-1-3p, and the model group was injected with no exosomes and an equal amount of saline. After 4 weeks of feeding, tail vein injection of KYSE150 cells and exosomes was repeated once. The purpose of injecting the cells twice in the animal model was to increase the modeling success rate. After another 4 weeks of continued feeding, mice were anesthetized and executed by 1% pentobarbital sodium overdose. Lung, liver, kidney, and skin tissues were extracted for HE staining and pathological analysis. Mice were weighed weekly and growth curves were plotted.

Tissue section information was scanned and imaged using a PANNORAMIC panoramic section scanner (3DHISTECH, Hungary). Images were viewed and captured using CaseViewer 2.4 software (3DHISTECH). The number of metastatic foci in each section was calculated and the corresponding tissue area was measured using Image-Pro Plus 6.0 software (Media Cybemetics, USA). Metastatic foci density = number of metastatic foci/tissue area. The liver, kidney and skin tissues of mice were pathologically scored according to the four-level grading system of the International Harmonization of Nomenclature and Diagnostic Criteria for Lesions in Rats and Mice proposal (INHAND). Pathological scoring items for liver tissue included hepatocyte edema, hepatocyte vacuolar degeneration, gore, and inflammation, for kidney tissue included renal tubular epithelial cell edema, intra-tubular eosinophils, gore, and inflammation, and for skin tissue included stratum spinosum cell degeneration, necrotic cell debris, and inflammation. The sum of the scores for each item was the final score for histopathology.

Immunohistochemistry

Mouse lung tissues were formalin-fixed for 48 h at 4 °C, and subsequently paraffin-embedded. The paraffin tissue blocks were sectioned into 0.4 μm sections. Sections were incubated with anti-E2F5 antibody solution (1:200) overnight at 4 °C, and then with secondary antibody solution for 30 min at room temperature. Color reaction was performed with DAB solution. H-score scoring was performed by two independent pathologists.

High-throughput transcriptome sequencing

Eca109 cells cultured with Exo-NC or Exo-miR-1-3p (100 µg/ml) for 48 h were subjected to high-throughput transcriptome sequencing on the Illumina sequencing platform. Each subgroup consisted of 3 independent cell samples. Clean reads were aligned to reference sequences using Bowtie2, and the gene and transcript expression levels were calculated using RSEM. Differential gene detection was performed using the DEseq2 algorithm with screening parameters of Fold Change ≥ 2.00 and Adjusted P value ≤ 0.05. KEGG enrichment analysis was performed using the phyper function in R software.

Dual luciferase reporter assay

The E2F5 3’UTR WT sequence and the E2F5 3’UTR Mut sequence were synthesized by TsingkeBiotechnologyCo.,Ltd. (Beijing, China). After cleaving out the sticky ends by restriction endonucleases NheI (FD0974, Thermo Scientific) and XbaI (FD0684, Thermo Scientific), the insert sequence was ligated into the pmirGLO vector. The recombinant plasmid (50 ng) was co-transfected with miRNA mimics (20 µM) into HEK293 cells. After 48 h of culture, the fluorescence values of the samples were read using the Dual-Luciferase Reporter Assay System (E1910, Promega) and GloMax^®-Multi Multifunctional Enzyme Labeler (Promega). Relative luciferase activity = (firefly luciferin value - background value)/(endosperm sea kidney luciferin value - background value).

Statistical analysis

Data were analyzed using the GraphPad Prism 8.0, and values were expressed as mean ± standard deviation (SD). Student’s t-test was used for two-group comparisons, and one-way ANOVA was used for multiple comparisons. The difference of P < 0.05 was considered statistically significant.

Results

miR-1-3p may have potential anticancer effects, and its expression is significantly down-regulated in human EC tissues

As shown in the volcano plot (Fig. 1A) and bidirectional clustering heatmap (Fig. 1B), we screened 112 loop miRNAs, which were significantly differentially expressed in human EC tissues (N = 184) compared to normal esophageal tissues (N = 13), of which 85 were significantly up-regulated and 27 were significantly down-regulated.

Fig. 1

miR-1-3p level is significantly down-expressed in human EC tissues. The differentially expressed genes in human EC tissues (N = 184) compared to human normal esophageal tissues (N = 13), shown in the volcano plot (A) and bidirectional clustering heatmap (B). The levels of the 27 significantly down-regulated loop miRNAs in human normal and EC tissues (C). Data were downloaded from UCSC Xena, derived from TCGA. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Because of the desire to find key molecules that could play a role in suppressing EC, we focused on 27 significantly down-regulated loop miRNAs, and re-annotated them as mature miRNAs (Table 2). The expression levels of these 27 loop miRNAs in human normal esophageal tissues and EC tissues are listed in Fig. 1C.

Table 2 Re-annotation of 27 loop miRNAs

Based on the significance of the expression differences and the progress of the study, we preliminarily selected miR-378c, miR-1-3p, miR-139-5p, miR-133a-3p, and miR-145-5p as potential cancer suppressors for subsequent studies. As shown in Fig. 2, after transfection of the corresponding mimics into human ESCC cell lines KYSE150 and Eca109, respectively, among the five mimics, only miR-1-3p could significantly inhibit the proliferation and migration of KYSE150 and Eca109 cells in vitro.

Fig. 2

miR-1-3p significantly inhibits the proliferation and migration of KYSE150 and Eca109 cells. The proliferation (A and B) and migration (C-F) of KYSE150 and Eca109 cells transfected with miR-378c, miR-1-3p, miR-139-5p, miR-133a-3p, or miR-145-5p, were detected by CCK-8 and Transwell assays. *P < 0.05

Exosomes load mir-1-3p and mediate mir-1-3p entry into target cells

Extracted and enriched exosomes were senn under TEM with a typical teato-like exosome morphology (Fig. 3A). Nanoparticle tracking analysis yielded particle sizes consistent with the size range of exosomes (30–200 nm) (Fig. 3B). Western blot also determined the expression of exosome-positive markers TSG101, CD9, and CD63 in the enriched particles (Fig. 3C). All these results confirmed the successful extraction and enrichment of exosomes.

Fig. 3

Exosomes load miR-1-3p and mediate miR-1-3p entry into target cells. Observed images under a transmission electron microscope (A), the particle sizes analyzed by NTA (B), and the expression of positive markers detected by western blot (C) of collected exosomes. The miR-1-3p levels in the exosomes after electrotransformation measured with RT-qPCR (D). PKH67 green fluorescent dye labeling and tracking of exosomes (E)

RNA was extracted from the electrotransformed exosomes and subjected to RT-qPCR, which showed that the exosomes were successfully loaded with miR-1-3p mimics (Fig. 3D). In addition, using PKH67 green fluorescent dye tracking, it was observed that the incubated cells gradually ingested and retained more exosomes as the incubation time increased (Fig. 3E). In summary, this study successfully loaded exosomes with miR-1-3p and mediated their entry into target cells.

Exosomes loaded with miR-1-3p inhibit migration and invasion of ESCC cells in vitro

20 µg/ml, 50 µg/ml or 100 µg/ml of exosomes loaded with miRNA were incubated with KYSE150 and Eca109 cells, and the results of CCK-8 assays showed that none of the different concentrations of Exo-NC had a significant effect on the cell proliferation, whereas Exo-miR-1-3p significantly inhibited the proliferation of KYSE150 and Eca109 cells in vitro in a dose-dependent manner (Fig. 4A). Subsequently, their effects on the invasive and migratory abilities of these cells in vitro were examined using Transwell chambers. The results showed that incubation (48 h) with different concentrations of Exo-NC had no significant effect on the migration and invasion of both KYSE150 and Eca109 cells (Fig. 4B-E). Whereas, incubation with Exo-miR-1-3p for 48 h resulted in a significant decrease in the number of cells completing migration and invasion, which indicated a significant reduction in their migratory and invasive abilities, especially those incubated with 100 µg/ml Exo-miR-1-3p (Fig. 4B-E).

Fig. 4

Exosomes loaded with miR-1-3p inhibit migration and invasion of ESCC cells in vitro. The proliferation (A), migratory (B and C) and invasive (D and E) abilities of KYSE150 and Eca109 cells incubated with 20 µg/ml, 50 µg/ml or 100 µg/ml of exosomes loaded with miRNA. *P < 0.05

Exosomes loaded with miR-1-3p inhibit lung colonization and growth of ESCC cells in vivo

Tail vein injections of KYSE150 cells were performed on BALB/c nude mice (model group) to construct a mouse model for in vivo experiments. During the 8-weeks feeding period, starting from week 5, the body weights of mice in the model group and Exo-NC group (tail vein injection of exosomes loaded with meaningless miRNAs was performed 2 days after the injection of KYSE150 cells) were abruptly and consistently reduced until the end of the experiment (Fig. 5A). In contrast, the weight loss of mice in the Exo-miR-1-3p group (tail vein injection of exosomes loaded with miR-1-3p performed 2 days after the injection of KYSE150 cells) appeared at week 6, and the magnitude was smaller (Fig. 5A). At the end of the experiment at week 8, the body weights of mice in the Exo-miR-1-3p group were significantly different from those of mice in the model group and Exo-NC group, whereas there was no significant difference in the body weights of mice in the model group and Exo-NC group (Fig. 5A). Furthermore, neither Exo-NC nor Exo-miR-1-3p showed significant systemic or local specific toxicity, and pathological tests of liver, kidney, and skin tissues did not show any significant abnormalities (Fig. 5B and C). Total RNA from different tissues of mice was extracted and subjected to RT-qPCR, which showed that different tissues of mice in the Exo-miR-1-3p group efficiently absorbed the exosomes loaded with miR-1-3p, and the levels of miR-1-3p were significantly elevated in the lungs, livers, and kidney tissues (Fig. 5D).

Fig. 5

Exosomes loaded with miR-1-3p inhibit lung colonization and growth of ESCC cells in vivo. (A) Body weight changes of mice. The representative section imaging (B) and pathological scores (C) of liver, kidney, and skin tissues. (D) miR-1-3p levels in lung, liver, and kidney tissues. (E) Representative section imaging, number and density of metastases of lung tissues. Black arrows indicate metastases. *P < 0.05, ***P < 0.001

Furthermore, obvious metastasis foci of cancer cell (indicated by black arrows) and abnormal angiogenesis within the foci were seen in the lung tissues of mice in the model group and Exo-NC group (Fig. 5E). In contrast, the lung tissues of mice in the Exo-miR-1-3p group were morphologically normal, and did not show obvious carcinomas or lesions (Fig. 5E). These data suggest that exosomes loaded with miR-1-3p have significant anticancer or cancer therapeutic effects.

miR-1-3p can directly bind to the 3’UTR of E2F5 mRNA, and down-regulates E2F5 expression in ESCC cells

To explore the potential molecular mechanisms by which exosomes loaded with miR-1-3p antagonize ESCC progression, Eca109 cells incubated with Exo-NC or Exo-miR-1-3p (100 µg/ml) for 48 h were subjected to high-throughput transcriptome sequencing. Compared with the Exo-NC group, 194 genes were significantly differentially expressed in the Exo-miR-1-3p group, of which 126 genes were significantly down-regulated and 68 genes were significantly up-regulated (Fig. 6A), and all differentially expressed genes were analyzed by KEGG pathway (Fig. 6B).

Fig. 6

miR-1-3p directly targets and down-regulates E2F5 expression. The differentially expressed genes in the Exo-miR-1-3p group compared with the Exo-NC group, shown in the volcano plot (A), and the KEGG pathway analysis of differentially expressed genes (B). (C) Predicted target genes of miR-1-3p with TargetScan and miRWalk were cross-analyzed with significantly expression down-regulated genes. (D) E2F5 mRNA level after Exo-miR-1-3p incubation. (E) Putative binding sequence of miR-1-3p to the E2F5 mRNA 3’UTR and results of dual luciferase reporter assays. E2F5 protein expression of KYSE150 and Eca109 cells transfected with miR-1-3p mimics (F) or miR-1-3p inhibitor (G). Immunohistochemical staining (H) and H-scores (I) of E2F5 in mouse lung tissues. *P < 0.05, ***P < 0.001

TargetScan 8.0 and miRWalk 3.0 predicted 698 and 88 potential target mRNAs of miR-1-3p, respectively, which were cross-analyzed with the 126 significantly expression down-regulated genes obtained by sequencing, yielding E2F5 and PTPRG as common intersections (Fig. 6C) (Supplementary Table 1). E2F5 was enriched in “Cell growth and death” and “Signal transduction” pathways (Fig. 6B). RT-qPCR verified the sequencing results, and the mRNA level of E2F5 in KYSE150 and Eca109 cells was significantly reduced after incubation with Exo-miR-1-3p (100 µg/ml, 48 h) (Fig. 6D). Moreover, dual luciferase assay verified the direct binding of miR-1-3p to the putative sequence of E2F5 mRNA 3’UTR (Fig. 6E). In addition, miR-1-3p mimics transfection significantly down-regulated E2F5 protein expression (Fig. 6F), whereas miR-1-3p inhibitor transfection significantly increased E2F5 protein levels in KYSE150 and Eca109 cells (Fig. 6G).

Moreover, the expression level of E2F5 protein was significantly higher in the lung tissues of mice in the model group and Exo-NC group, especially in the metastatic foci fraction, which was mainly concentrated in the nucleus and also accumulated in the cytoplasm due to excessive expression (Fig. 6H and I). The expression level of E2F5 protein was significantly lower in the lung tissues of mice in the Exo-miR-1-3p group (Fig. 6H and I). These results suggest that miR-1-3p can directly bind to the 3’UTR of E2F5 mRNA, and down-regulate E2F5 protein expression in ESCC cells.

miR-1-3p inhibits the proliferation, migration and invasion of ESCC cells in vitro by down-regulating the expression of E2F5, which may involve the MAPK/ERK signaling pathway

miR-1-3p mimics and E2F5 expression plasmids were co-transfected into KYSE150 and Eca109 cells in vitro, and their effects on cell proliferation, migration, and invasion were examined. The intracellular expression levels of miR-1-3p and E2F5 after co-transfection were verified in Fig. 7A-C. The results of CCK-8 and Transwell assays indicated that overexpression of E2F5 neutralized the inhibitory effects of miR-1-3p on the proliferation, migration and invasion of KYSE150 and Eca109 cells (Figs. 7D-F and 8A-B).

Fig. 7

miR-1-3p inhibits proliferation and migration of ESCC cells in vitro by down-regulating E2F5. The miR-1-3p levels (A), E2F5 mRNA levels (B), and E2F5 protein levels in KYSE150 and Eca109 cells in vitro co-transfected with miR-1-3p mimics and E2F5 expression plasmids. The proliferation (D) and migration (E and F) of transfected cells. ns, no significance; *P < 0.05, ***P < 0.001

Fig. 8

miR-1-3p inhibits invasion of ESCC cells in vitro by down-regulating E2F5, which may involve the MAPK/ERK signaling pathway. (A and B) The invasion of KYSE150 and Eca109 cells in vitro co-transfected with miR-1-3p mimics and E2F5 expression plasmids. (C) The protein levels of p-p38, p38, p-ERK, ERK in transfected cells, detected with western blot. ns, no significance; ***P < 0.001

In addition, further exploration of signaling pathway activation revealed that transfection of miR-1-3p significantly inhibited the phosphorylation of p38 and ERK in KYSE150 and Eca109 cells, and which was further counteracted by E2F5 overexpression (Fig. 8C). This result suggests that miR-1-3p may impede ESCC progression by down-regulating the expression of E2F5 and inhibiting the activation of the MAPK/ERK signaling pathway.

Discussion

Based on the analysis of miRNAs profiling data of human EC tissue samples in TCGA, we preliminarily confirmed that the level of miR-1-3p was significantly down-regulated in human EC tissues, and this trend was consistent with the trend of miR-1-3p expression dysregulation in other types of solid tumors. Although, we failed to collect more clinical EC samples to re-validate the expression level of miR-1-3p in EC tissues by wet experiments. In addition, EAC is a major subclass of EC. The cell lines used in this study were limited to ESCC, so further expansion is necessary to confirm the role of miR-1-3p in EAC.

MiR-1-3p is a key anti-cancer molecule. However, the mechanisms underlying its dysregulated expression in tumors remain unclear. The main mechanisms reported so far include binding and silencing by long-stranded non-coding RNAs [11] or cyclic RNAs [12], hypermethylation of gene promoters [13], and expression dysregulation of RNA-binding protein [14]. Further exploring the regulatory mechanism of miR-1-3p expression silencing in cancer and targeting this mechanism may be beneficial for the development of targeted molecular therapy and/or gene therapy for cancer.

Exogenous expression of miR-1-3p significantly inhibited the proliferation, migration and invasion of KYSE150 and Eca109 cells in vitro. More critically, miR-1-3p was loaded into exosomes by electroporation in this study. Exosomes loaded with miR-1-3p not only inhibited the proliferation, migration and invasion of KYSE150 and Eca109 cells in vitro, but also significantly hindered the intrapulmonary colonization and growth of KYSE150 cells through blood circulation in mice. At the endpoint of the animal experiments (day 54 of the first exosome inoculation and day 26 of the second exosome inoculation), the detection results of miR-1-3p content in various tissues of mice indicated that exosomes could effectively mediate the stable presence of their inclusions (miRNAs) in various tissues in vivo. This result suggests that loading of exosomes enables miRNAs to remain quite stable in the animal’s internal environment, including blood circulation and tissues, and to be taken up by cells of various tissues. In addition, we did not find any systemic or local-specific toxic side effects of miRNA-loaded exosomes, which are relatively safe. The excellent anticancer effect, stability and safety of exosomes loaded with specific anticancer miRNAs suggest its potential as cancer therapeutic agents.

The loading strategy used in this study was post-production loading, where miRNA molecules were loaded directly into isolated and purified exosomes. In addition to electroporation, commonly used methods include sonication. In addition, another exosome loading strategy is preloading, in which nucleic acids or drugs are pre-loaded into parental cells first, and then the exosomes they produce may contain these nucleic acids or drugs. For example, Zhang et al. achieved miR-29a-3p overexpression in human mesenchymal stem cells and then further collected miR-29a-3p-loaded exosomes [15]. They reported that these exosomes attenuated glioma cell migration and angiogenesis [15]. The source cells, enrichment methods, and engineering modifications of the exosomes, as well as the loading methods of miR-1-3p and co-administration with other drugs, need to be further determined before possible clinical applications.

In addition, this study reported a novel downstream target and pathway of miR-1-3p, including the transcription factor E2F5 and the MAPK/ERK signaling pathway. In the KEGG pathway enrichment results of differentially expressed genes obtained by high-throughput transcriptome sequencing, ITGA3, ITGB2, PIP4K2C and TMSB4X were enriched in the “Cell motility” pathway (Fig. 6B). They are proteins that can directly alter cell motility-related phenotypes, including cell surface adhesion and migration. On the other hand, their altered expression is a reflection or marker of altered cell motility-related phenotypes. However, after our validation, none of them were direct targets of miR-1-3p (results not shown). Many details of the regulatory network, from miR-1-3p to its direct target (E2F5, and possibly others), to the downstream signaling pathway it affects (MAPK/ERK signaling pathway, and possibly others), to the expression dysregulation of proteins that can directly alter cellular phenotypes (e.g., cell motility, cell proliferation), need to be added.

E2F5, a transcription factor that regulates cell proliferation and cell cycle, has been reported to be significantly overexpressed in poorly differentiated EC tissues and independently predicted patients’ prognosis [16]. EC patients with E2F5 overexpression have a poorer prognosis [16]. However, how E2F5 expression further affects ESCA cell colonization and growth in the lungs needs to be explored in greater depth. Currently, there are also relatively few findings on E2F5 expression and tumor metastasis. In addition, most of the current studies reported the inhibitory effect of miR-1-3p on the activation of Akt/mTOR signaling pathway [17, 18], while the effect on MAPK/ERK signaling pathway has not been reported. However, the results of the present study indicated that miR-1-3p did not affect the activation of Akt/mTOR signaling pathway in KYSE150 and Eca109 cells, which may be related to cell specificity (Supplementary Fig. 1).

Conclusion

Exosomes loaded with miR-1-3p significantly inhibited the proliferation, migration and invasion of ESCC cells in vitro, as well as the intrapulmonary colonization and growth of ESCC cells in vivo, which may be applicable in the treatment of EC.