Introduction
Globally, lung cancer ranks first as the cause of tumor-related mortality [
1]. The 5-year overall survival (OS) of this disease remains ≤ 20% [
2]. As one of the first-line drugs of epidermal growth factor receptor-tyrosine kinase inhibitors (EGFR-TKIs), gefitinib prolonged the survival of patients with NSCLC to a certain extent [
3]. Nonetheless, several patients acquire resistance to gefitinib after taking the drug for 8–12 months [
4]. Invasive diagnosis and delayed treatment of the acquired resistance often result in non-negligible harm to patients, and there is no simple mechanism to clarify the complex pathway interactions in such cases of acquired resistance. In recent years, an increasing number of studies have demonstrated that non-coding RNAs play an important role in the resistance mechanism of EGFR-TKIs, lncRNA CRNDE promoted the eIF4A3/MUC1/EGFR axis and apoptotic activity, and restored the sensitivity to EGFR-TKIs [
5]. MiR-200c-3p has been reported to play a role in EGFR TKI sensitivity by regulating the EMT process [
6]. Therefore, in light of the literature, new gefitinib-resistance mechanisms and effective therapeutic targets seem worth exploring.
An exosome is an extracellular vesicle of 30–150 nm [
7] in diameter and is composed of a lipid bilayer. The exosome can encapsulate the genetic information of the maternal cells, including circRNA [
8], long non-coding RNA (lncRNA) [
9], and micro-RNA (miRNA) [
10]. The exosomes deliver these RNAs to the target cells [
11]. Scientists have discovered that exosomes are essential for developing drug resistance [
12]. CircRNAs are non-coding RNA molecules that possess a closed circular structure [
13]. In recent years, several studies have indicated that circRNA plays a significant role in the proliferation [
14], metastasis [
15], immune escape [
16], and drug resistance [
17] of malignant tumors. CircRNA is a competitive endogenous RNA (ceRNA) to regulate downstream targets by absorbing miRNA [
18,
19]. MiRNAs are short non-coding RNAs of 21–22 nucleotide length [
20]. CircRNA contains a specific miRNA response element (MRE) that regulates the mRNA containing the same MRE from miRNA [
21]. CircRNA inhibits miRNA activity, regulates tumor-related genes, alters key pathways, and participates in the cancer process via the ceRNA network [
22]. In addition, circRNAs are widely enriched in humoral exosomes [
23]. We had previously reported that the upregulated circRNA PTPRA inhibits the epithelial–mesenchymal transition (EMT) and the metastatic growth of NSCLC by targeting miR-96-5p [
24].
MEF2A, a DNA-binding transcription factor [
25], is involved in several cellular processes, including cellular mitochondrial metabolism [
26], cell growth [
27], and apoptosis [
28]. It has also been reported that MEF2A can directly inhibit the expression of the cell cycle genes [
29] and activate the apoptosis-related pathways [
30]. Cyclin-dependent kinase 4 (CDK4, a kinase that drives the G1/S progression) [
31] and the BCL2-associated X (BAX, an apoptotic activator) [
32] are crucial for the development and treatment of acquired resistance to gefitinib. Furthermore, although the Warburg effect suggests that cancer cells prefer to use glycolysis over the aerobic cycle [
33], the mitochondrial OXPHOS function of EGFR-TKI-resistant cells is significantly increased [
34]; therefore, modulation of cellular OXPHOS may be an important approach to restore gefitinib sensitivity.
In this study, we obtained the differential expression profiles of circRNAs in the exosomes from gefitinib-resistant and -sensitive cells by high-throughput sequencing. A significantly poorly expressed circRNA, circKIF20B, was screened and validated in the serum exosomes and tumor tissues of patients with NSCLC. Accordingly, we hypothesized that exosomal circKIF20B regulated gefitinib resistance by the miR-615-3p/MEF2A axis. Therefore, we explored the biological function and the specific molecular mechanism of circKIF20B in developing gefitinib-acquired resistance.
Materials and methods
Clinical specimens
A total of 85 pairs of NSCLC and matched normal tissues were collected from The First Affiliated Hospital of Anhui Medical University (Anhui, China). The evaluation standard of normal tissues was > 5 cm away from the tumor edge [
35]. All tissues were frozen in liquid nitrogen. A total of 48 patients with NSCLC who were treated with gefitinib at The First Affiliated Hospital of Anhui Medical University were enrolled in this study. A total of 24 serum samples were sourced from patients whose disease had progressed after taking gefitinib; another half of the samples were sourced from patients who had achieved a complete response or partial response after taking the drug. We evaluated the efficacy of patients receiving gefitinib according to the solid tumor evaluation standard RECIST (Version 1.1). Patients with CR (complete remission) and PR (partial remission) are regarded as a sensitive group, and patients with PD (disease progression) are regarded as a resistant group. All serums were stored at -80℃.
All patients were diagnosed with NSCLC by professional pathologists at The First Affiliated Hospital of Anhui Medical University. The clinical pathological information has been recorded in Tables
1 and
2. This study was approved by the Medical Ethics Committee from The First Affiliated Hospital of Anhui Medical University. The experiment was conducted under the World Medical Association Declaration of Helsinki.
Table 1
Correlation between circKIF20B expression and clinicopathological characteristics in NSCLC patients (n = 85)
n | 43 | 42 | |
Gender, n (%)
| | | 0.448 |
Female | 19 (22.4%) | 23 (27.1%) | |
Male | 24 (28.2%) | 19 (22.4%) | |
Histological type, n (%)
| | | 0.789 |
Adenocarcinomas | 36 (42.4%) | 37 (43.5%) | |
Squamous cell carcinoma | 7 (8.2%) | 5 (5.9%) | |
Smoking status, n (%)
| | | 0.845 |
No | 32 (37.6%) | 33 (38.8%) | |
Yes | 11 (12.9%) | 9 (10.6%) | |
Tumor size, n (%)
| | | 0.012* |
> 3 | 19 (22.4%) | 7 (8.2%) | |
≤ 3 | 24 (28.2%) | 35 (41.2%) | |
Lymph node metastasis, n (%)
| | | 0.116 |
No | 31 (36.5%) | 37 (43.5%) | |
Yes | 12 (14.1%) | 5 (5.9%) | |
Metastasis, n (%)
| | | 1.000 |
No | 39 (45.9%) | 38 (44.7%) | |
Yes | 4 (4.7%) | 4 (4.7%) | |
Pathologic stage, n (%)
| | | 0.020* |
I-II | 29 (34.1%) | 38 (44.7%) | |
III-IV | 14 (16.5%) | 4 (4.7%) | |
Differentiation, n (%)
| | | 0.030* |
Poor | 21 (24.7%) | 10 (11.8%) | |
well and moderate | 22 (25.9%) | 32 (37.6%) | |
Age, mean ± SD
| 60.37 ± 9.93 | 60.74 ± 11.55 | 0.876 |
Table 2
Correlation between serum exosomal circKIF20B expression and clinicopathological characteristics of NSCLC patients (n = 24/24)
n
| 24 | 24 | |
Gender, n (%)
| | | 0.317 |
Female | 4 (8.3%) | 8 (16.7%) | |
Male | 20 (41.7%) | 16 (33.3%) | |
Smoking status, n (%)
| | | 1.000 |
No | 19 (39.6%) | 18 (37.5%) | |
Yes | 5 (10.4%) | 6 (12.5%) | |
Tumor size, n (%)
| | | 0.018* |
> 3 | 10 (20.8%) | 19 (39.6%) | |
≤ 3 | 14 (29.2%) | 5 (10.4%) | |
Pathologic stage, n (%)
| | | 0.040* |
IIIB | 18 (37.5%) | 10 (20.8%) | |
IV | 6 (12.5%) | 14 (29.2%) | |
EGFR mutation, n (%)
| | | 0.724 |
19DEL | 18 (37.5%) | 20 (41.7%) | |
L858R | 6 (12.5%) | 4 (8.3%) | |
Differentiation, n (%)
| | | 0.563 |
Poor | 11 (22.9%) | 14 (29.2%) | |
Well and moderate | 13 (27.1%) | 10 (20.8%) | |
Age, mean ± SD
| 63.62 ± 12.31 | 63.54 ± 9.84 | 0.979 |
Cell culture
PC9GR and HCC827GR (gefitinib-resistant NSCLC cell lines) were generated after PC9 and HCC827 upon continuous exposure to gefitinib (M1749, Abmole, China) to step up to 20 µM concentration for at least 6 months [
36]. PC9 cell line and HCC827 cell line were obtained from the Fuheng Biology in 2019 (Shanghai, China), the cell line has been authenticated by STR, and the cell line has been tested for mycoplasma contamination. DMEM or RPMI 1640 (SH30022.01, Hyclone, USA) supplemented with 10% fetal bovine serum (FBS) (10099141, Gibco, USA) and 1% penicillin and streptomycin (C0222, Beyotime, China) were used as the culture media. The plates were incubated under a 5% carbon dioxide atmosphere at 37℃. The Serum-free Media (UR51102, Umibio, China) was used before the cell extraction of the exosomes. Half maximal inhibitory concentration (IC50 value) to gefitinib of PC9, PC9GR, HCC827, and HCC827GR were calculated to determine the resistance.
Isolation and identification of exosomes
As directed by the manufacturer’s instruction, the Quick Exosome Isolation Kit (41201ES25, Yeasen, China) was used to isolate the exosomes from the serums and cells. The exosomes for circRNA sequencing were extracted utilizing an ultracentrifuge (Beckman, USA) [
37]. The morphological observation of exosomes was performed by transmission electron microscopy (TEM) (FEI, USA). Western blotting demonstrated the presence of exosomal markers, including CD63, CD9, and TSG101. Particle size measurements were performed using a nanoparticle tracking analyzer (NTA) (Malvern, UK).
High-throughput RNA sequencing
The MiRNeasy Micro Kit (217084, Qiagen, USA) was used to extract RNA from exosomes. RNA high-throughput sequencing was performed using the Total RNA-seq (H/M/R) Library Prep Kit (NR603, Vazyme, China) and PE150 (Illumina, USA). The results of this sequencing are listed in Table
S1. The selection criteria were ∣FC (Fold Change)∣>1 and P-value<0.05.
The miRNA-seq and RNA-seq were downloaded from the lung adenocarcinoma (LUAD) and lung squamous carcinoma (LUSC) projects in The Cancer Genome Atlas (TCGA). R studio (3.6.3) was used for subsequent data analyses. The websites of relevant databases in this study are listed under Table
S2.
RNA and gDNA extraction, and real-time quantitative PCR (qRT-PCR)
TRIzol (15596,018, Ambion, USA) was used to isolate total RNAs from the cells, tissues, and exosomes. Tiagen (KG203, Beijing, China) provided the DNA extraction kit for cell genomic DNA extraction. The cDNA Synthesis SuperMix (11120ES60, Yeasen) was used to reverse transcript circRNA and mRNA into cDNA. The miRNA cDNA Synthesis Kit (R601, EnzyArtisan Biotech, China) was used to reverse transcript miRNA into cDNA. The QPCR SYBR Green Mix (11201ES08, Yeasen, China) was used for conducting qPCR. Electrophoresis was performed on 2% agarose gel after PCR. For qRT-PCR, the PCR products were quantified on the MX3000P (Agilent, USA). The primers were synthesized by the EnzyArtisan Biotech (China), and the 2
-ΔΔCt method was used for calculation [
38]. The primer sequences are listed in Table S3.
Subcellular RNA fractionation
The cells were seeded in a 6-well plate, and the cytoplasmic and nuclear RNAs were extracted using the PARIS Kit (AM1921, Ambion) as described previously [
39] when the cells reached 80% confluence. The expression of circKIF20B was detected by qRT-PCR.
Western blotting
We collected and quantified the total proteins with the RIPA lysis buffer and the enhanced BCA protein assay kit (P0010, Beyotime, China). Briefly, 20 µg protein from each sample was loaded into the gels. Then, 12.5% SDS-PAGE (PG113, Epizyme, China) was used to separate the total protein and transfer it onto the PVDF membranes (IPVH00010, Millipore, Germany). The membranes were then blocked for 1 h with 5% non-fat milk at room temperature. The primary antibodies, including CD9, CD63, TSG101, β-actin, BAX, CDK4, and MEF2A, were incubated at 4℃ overnight and the membranes were washed by TBST. A secondary antibody linked to HRP was used for 1 h at room temperature, and the membranes were washed thrice with TBST for 10 min each. The Tanon Chemiluminescent Imaging System (China) was used to detect the Western blotting. The brand and dilution ratios of antibodies are given in Table S4.
Fluorescence in situ hybridization (FISH)
The circKIF20B probe was fluorescently labeled with FAM, and the miR-615-3p probe was fluorescently labeled with cy3. These probes were synthesized by Genepharma (Shanghai, China), and their sequences are listed in Table S3. Then, 2 × 104 cells were seeded on a coverslip overnight. Then, PBS was used to wash the cells twice, and 4% paraformaldehyde was used to fix the cells for 15 min at room temperature. Next, the cells were treated with 0.1% Triton-X 100 for 15 min at room temperature and blocked for 30 min at 37℃ incubator. Then, the cells were hybridized with the probe in the dark overnight at 37℃ incubator for 16 h. Then, 0.1% Tween 20 was used to wash the cells for 10 min at 37℃, and 2 × saline sodium citrate was used to wash the cells at 37℃ and 60℃. Then, DAPI was used as a nuclear stain and observed under a confocal microscope (Zeiss, Germany) to image the coverslip.
Transfection of knockdown and overexpression vectors
CircKIF20B siRNAs, knockdown vector LV3-circKIF20B, and matched negative control vector LV3-NC, the mimics and inhibitors of miR-615-3p, MEF2A’s siRNAs, the MEF2A overexpression vector pEX3-MEF2A, and matched negative control vector pEX3-NC were manufactured by the Genepharma (Shanghai, China). The circKIF20B overexpression vector GV689-circKIF20B and matched negative control vector GV689-NC were synthesized by Genechem (Shanghai, China). HEK293T cells were seeded into a 6-well plate (3471, Corning, USA) and grown to 80% confluency. Then, the abovementioned siRNAs and vectors were co-transfected with Lipofectamine 3000 reagent (L3000001, Invitrogen, USA) and Opti-MEM (Gibco) for 8 h. Fresh medium was added to the cells after removing the transfection medium. After transfection for 48 h, qRT-PCR was used to detect the transfection efficacy. The sequences are listed under Table S3.
Construction of stably transfected cell lines
HEK293T cells were seeded in a 100-mm cell culture dish (430167, Corning, USA) and transfected with the circKIF20B knockdown or overexpression vectors and the helper plasmids (Hanbio, China) via Opti-MEM and the Lipofectamine 3000 reagent when cells reached 80% confluence [
40]. The transfection medium was removed after 8 h and a fresh medium was added to the cells. After transfected for 48 h, the supernatant of cell was harvested and centrifuged at 1500 x
g for 5 min and purified via filtration through a 0.45-µm filter (UFC810096, Millipore, USA) and concentrated by centrifugation at 4000 x
g for 15 min. Then, the lentiviral titer was detected for the subsequent experiments. The target cells were seeded into a 6-well plate and transfected with lentivirus and 5 µg/mL polybrene. After transfection for 24 h, fresh 10% FBS/DMEM was added to continue the incubation of the target cells for 48 h. Because the lentivirus also expressed GFP, flow sorting (Beckman, USA) was used to screen the GFP-positive cells [
41]. Harvesting and incubating the GFP-positive cells to obtain the stably transfected cell lines. Then, the expression of target genes was detected by qRT-PCR.
Stability verification
For RNase R treatment, we used 10 units of RNase R (R0301, Geneseed, China) to 2.5 µg of total RNA and incubated it at 37℃ for 30 min. For ACTD treatment, the cells were simultaneously treated with 2 µg/mL ACTD (SBR00013, Sigma, Germany) for 0, 4, 8, 12, and 24 h, and the total RNA was extracted with TRIzol. The expression of circKIF20B and KIF20B was detected by qRT-PCR.
RNA pulldown
The circKIF20B biotin-labeled probe and the control probe were synthesized by Jinkairui (China), and the sequences are listed in Table S3. Following cell lysis, the cell lysates were incubated with the probe. According to the manufacturer’s instructions, the experiment was performed using the Pierce Magnetic RNA-Protein Pull-Down Kit (20164, Thermo, USA). The total RNA was extracted with TRIzol from the complex to generate biotin-labeled complementary RNA targets, followed by the preparation of RNA-conjugated beads and binding and elution of RBP. The expression of circKIF20B and miR-615-3p was detected by qRT-PCR [
42].
RNA immunoprecipitation (RIP)
To analyze the level of interaction between miR-615-3p and MEF2A, the Millipore’s Magna RIP RNA Binding Protein Immunoprecipitation Kit (17–700, USA) was used to perform RIP according to the manufacturer’s protocols. The harvested cells were lysed directly. Then, 100 µL of the cell lysate was pipetted onto anti-Ago2 magnetic beads or anti-IgG magnetic beads for immunoprecipitation [
43]. After washing with RIP buffer, RNA purification and qRT-PCR detected the expression of miR-615-3p and MEF2A, respectively.
Dual-luciferase reporter gene assay
Wild-type (WT) and mutant-type (MUT) vectors of miR-615-3p MRE were synthesized by Bioogenetech (China). Then, WT or MUT vectors were co-transfected with miR-615-3p into HEK293T. The cells were then lysed and added to the reaction solution and processed using a dual-luciferase reporter gene kit (RG042S, Beyotime). The fluorescence expression was calculated using a multifunctional microplate reader (Synergy, USA). The sequences are listed in Table S3.
Cell viability assay and IC50 value determination
For the cell viability detection, 5 × 103 cells were cultured in a 96-well plate for 0, 24, 48, and 72 h with 0.05 µM or 5 µM gefitinib, respectively. For the IC50 values analyses, different concentrations of gefitinib were added to a 96-well plate and incubated for 48 h. Each well was incubated for 1 h at 37℃ with 10 µL of the Cell Counting Kit-8 (CCK-8) (C0037, Beyotime). The microplate readers (Synergy) read the absorbance at 450 nm.
Proliferation assay (5-ethynyl-20-deoxyuridine; EdU incorporation)
The cells (2 × 104) were seeded on a coverslip, and 10 µM of EdU (C0075S, Beyotime) was added, followed by incubation for 2 h at 37℃. The cells were then fixed in 4% paraformaldehyde for 15 min at room temperature, followed by permeabilization with Trixon-X 100 for 10 min at room temperature. The cells were then treated with the click reaction mixture as per the manufacturer’s instructions. Then, DAPI was used to stain the nuclear. The coverslips were photographed under a microscope (Leica, Germany).
Cell cycle assay and apoptosis assay
The cells (1 × 10
5) were seeded in a 12-well plate. For cell-cycle analysis, the supernatant containing dead cells was collected to terminate tryptic digestion and centrifuged at 1000 x
g for 5 min. Then, PBS was used to wash the cells twice, and the cells were pipetted into 1 mL of ice-cold 70% ethanol at 4 ℃ overnight. Then, PBS was used to wash the cells and stained them with propidium iodide (C1052, Beyotime) for 30 min. Then, the cell cycle was analyzed by flow cytometry (Beckman). For apoptosis, the supernatant was collected to terminate tryptic digestion and centrifuged at 1000 x
g for 5 min, followed by washing with PBS. The proportion of apoptotic cells was detected by the Annexin V-APC/7-AAD (40310ES50, Yeasen) staining [
44]. All analyses were performed by flow cytometry (Beckman) as per the manufacturer’s instructions.
Exosome labeling and tracing
The exosomes were stained with 5 µM of 3,3-dioctadecyloxacarbocyanine perchlorate (DiO) (40725ES10, Yeasen) and incubated at 37℃ for 20 min. The exosomes were filtered through a 0.22-µM filter and then added to the cell culture medium for co-culturing with the cells for 48 h [
45]. A microscope (Leica) was then used to detect the green fluorescence.
ATP generation assay
The cells (2 × 10
5) were seeded into a 6-well plate, and 200 µL of the cell lysates were added to each well, followed by centrifugation at 12,000 ×
g for 5 min after cell lysis. After discarding the precipitation, add 100 µL of the ATP detection solution (S0027, Beyotime) was added to each sample [
46]. A multifunctional microplate reader (PE, USA) was used to calculate the chemiluminescence absorbance.
Mitochondrial membrane potential assay
The cells (2 × 10
5) were seeded in a cell culture dish (801001, NEST, China), to which 0.5 mL of the TMRE working solution (C2001S, Beyotime) was added, and the dish was incubated at 37℃ for 40 min [
47]. As a positive control, the cells were treated with 10 µM of carbonyl cyanide m-chlorophenyl hydrazone (CCCP) for 20 min before adding tetramethylrhodamine ethyl ester (TMRE). At the end of the incubation period, the nuclei were stained with DAPI, and the images were captured using a confocal microscope (Zeiss).
Oxygen consumption rate (OCR) assay
According to the manufacturer’s instructions, OCR was measured with the Seahorse XF 96 Analyzers (Seahorse Bio, USA) using the XF Mito Stress Test Kit (103015, Agilent, USA). A 96-well XF special cell culture plate was seeded with 2 × 10
4 cells/well, and the sensor probe plate was hydrated with the seahorse XF calibration solution overnight at 37℃ in a CO
2-free incubator. Then, oligomycin (1.5 µM), 3,3-dioctadecyloxacarbocyanine perchlorate (FCCP) (1.5 µM), and rotenone/antimycin A (0.5 µM) were added successively, and the cells were measured for OCR in the medium (pH: 7.4) containing 2 mM glutamine, 10 mM glucose, and 1 mM pyruvate [
48].
Establishment of a xenograft model
Overexpressed circKIF20B PC9 (1 × 107) was subcutaneously injected into the mice’s armpits. The GV689-NC was a control group. The tumor size and weight were measured every week. After 7 weeks, the mice were sacrificed, and the tumor was collected for subsequent experiments. The Anhui Medical University’s Animal Ethics Committee approved all the procedures.
Immunohistochemistry and hematoxylin-eosin (HE) staining
We next used 4% paraformaldehyde to fix the tissues, followed by paraffin embedding and sectioning. The cells were then treated with the primary antibodies MEF2A and Ki67 and incubated overnight at 4℃, followed by treatment with the secondary antibody at room temperature for 1 h. Subsequently, a DAB reagent was used to visualize the IHC, and hematoxylin was used for counterstaining. For HE staining, after fixing the sections, hematoxylin staining was performed for 15 min and eosin staining for 5 min. The cells were then observed under a microscope (Leica). The brand and dilution ratios of these antibodies used in this study are listed under Table S4.
Statistical analysis
In this study, each experiment was repeated thrice. The GraphPad prism (9.0, USA), Image J (1.53e, USA), and Figdraw (China) were employed for graph drawing and processing. Statistical analysis was performed by SPSS (23.0, USA) and Excel 2016 (Microsoft, USA), and the data were presented as the means and standard deviation. A Wilcoxon rank-sum test was applied to analyze the unpaired differential expression in TCGA. Log-rank and COX regression techniques were employed to examine the survival probability. Pearson’s test was performed for correlational analyses. Student’s t-tests were performed to determine the differences between the groups. The difference was statistically significant at P < 0.05. (*P < 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001).
Discussion
Gefitinib, as the first-line treatment for EGFR-mutated NSCLC significantly improves patients’ survival [
49], albeit acquired resistance strongly limits the long-term clinical efficacy of gefitinib. In the past few years, some scientists have studied the resistance mechanism of gefitinib and found that gene mutation is the leading cause of gefitinib resistance [
50]. Although some targeting genetic mutations EGFR-TKIs, such as osimertinib and afatinib, have demonstrated reliable results in second-line therapy [
51,
52], their vast cost and secondary resistance have caused severe harm to the patients. Therefore, exploring new gefitinib resistance mechanisms is urgent. This study revealed a novel circRNA, circKIF20B, that was significantly downregulated in the gefitinib-resistant cellular exosomes. The expression of circKIF20B was decreased in both the serum exosomes of patients with gefitinib resistance and the tumor tissues of patients with NSCLC. This study sheds light on how exosome-derived circKIF20B regulates the cell cycle, proliferation, apoptosis, and OXPHOS to affect gefitinib resistance. This discovery is novel and differs from conventional resistance mechanisms.
This study demonstrated the advantages of exosomal circKIF20B as an alternative liquid biopsy candidate for diagnosing and treating gefitinib resistance in NSCLC in terms of the following points: first, characterization experiments signified that the circular structure made circKIF20B more resistant to ACTD and RNase R. The lipid membrane encapsulation of the exosomes reduced the risk for degradation [
53]. Moreover, the stable expression of circKIF20B in the human tissues and serum exosomes was verified in this study, and the expression in the tissues was observed to be consistent with that in the serum exosomes. This finding suggests that exosomal circKIF20B is highly specific and stable as a noninvasive biopsy biomarker and a candidate for gene therapy. In addition, the results from gain and loss experiments implied that the overexpression of circKIF20B significantly decreased the IC50 value of gefitinib and the proliferation of the NSCLC cells. Contrarily, the downregulation of circKIF20B enhanced the gefitinib resistance and cell proliferation. This finding provides strong scientific evidence for diagnosing, monitoring, and treating acquired drug resistance. Finally, the recipient cells were intervened with exosomes extracted from NSCLC cell lines with artificially upregulated circKIF20B expression to mimic biological communication between the cancer cells. After exosomes loaded with upregulated circKIF20B entered the target cells, exosomal circKIF20B, as a unique gene therapy modality, suppressed target cells’ gefitinib resistance and proliferation properties. Exosome-derived circKIF20B appears to hold considerable potential in engineered therapy for reversing gefitinib resistance, which agrees with the findings of Huang’s research [
54].
Mechanistically, this study found that circKIF20B was significantly enriched in the cytoplasm, suggesting the potential of circKIF20B as a miRNA sponge to function as a ceRNA [
55]. MiR-615-3p was screened out by bioinformatics analysis, luciferase reporter gene, and RNA pulldown. CircKIF20B has a natural binding ability to the miR-615-3p-response element, which can downregulate the inhibitory effect of miR-615-3p on target mRNA. Notably, miR-615-3p appeared to be a bidirectionally regulated tumor-associated molecule. Liu found that circITCH increases bortezomib resistance in multiple myeloma by upregulating miR-615-3p [
56]. Lei found that miR-615-3p acts as an oncogenic factor that promotes EMT and metastasis in breast cancer by regulating the PICK1/TGFBRI axis [
57]. However, Professor Pan’s research reported that miR-615-3p is one of the critical factors in the fight against erlotinib resistance [
58]. In our study, the expression of miR-615-3p was significantly upregulated in both the serum exosomes of 23 patients with gefitinib resistance and 85 NSCLC tissues, which is consistent with the miRNA-seq data of the TCGA; moreover, its expression was negatively correlated with the expression of circKIF20B. The IC50 value of the gefitinib assay indicated that miR-615-3p could significantly promote the level of gefitinib resistance. In corroboration with previous reports, we found that circKIF20B acted as a tumor suppressor gene in the progression of gefitinib resistance via binding the oncogene miR-615-3p.
Mature miRNAs bind to the 3’ untranslated region (UTR) of target mRNAs and trigger mRNA degradation or translational repression. We found that the 3’UTR of MEF2A can directly bind to the MRE of miR-615-3p through bioinformatics analysis, dual-luciferase reporter gene assay, and RIP. To date, this study is the first to report the endogenous existence of the circKIF20B/miR-615-3p/MEF2A signaling axis in NSCLC cells. MEF2A has been reported to be related to the cell cycle and apoptosis [
59], as also validated by Western blotting in our study. The interaction of the MEF2 family with EGFR inhibits cell proliferation in drug-resistant metastatic colorectal cancer [
60]. In our study, the expression of MEF2A was significantly reduced in the NSCLC tissues and the serum exosomes of patients with gefitinib resistance, which is consistent with the RNA-seq data in TCGA. The qRT-PCR and Western blotting indicated that circKIF20B and miR-615-3p could independently regulate the transcriptional and translational expression of MEF2A. Moreover, we found that circKIF20B regulates the cell cycle, proliferation, and apoptosis in gefitinib resistance via the miR-615-3p/MEF2A axis.
GSEA showed that the expression of MEF2A was significantly negatively correlated with mitochondrial metabolism and OXPHOS. A past study showed that miRNA could enhance the mitochondrial function and activate OXPHOS by downregulating MEF2A [
61], while tumor-related genes enhance OXPHOS through metabolic reprogramming to inhibit sensitivity to TKI drugs [
62]. Moreover, it is interesting that Gong’s research showed that circPUM1 accelerated the development of cancer by increasing the mitochondrial membrane potential of the esophageal squamous cell carcinoma cells and enhancing the function of OXPHOS [
63]. This finding prompted an investigation of whether circKIF20B could regulate the mitochondrial function and OXPHOS in the NSCLC cells via MEF2A. As we all know that cancer cells have specific energy metabolism, aerobic glycolysis. It can support cancer cells surviving the same as OXPHOS. It has been reported that tumor tissues have an increased glucose metabolism and lactic acid production. However, the availability of glutamine is low in both normal and tumor tissues. Meanwhile, tumor tissues need more glucose as the main carbon source in the TCA cycle compared to healthy tissues. All the evidence demonstrated that glucose carbon is essential for tumor formation [
64]. Both aerobic glycolysis and oxidative phosphorylation are involved in cancer progression. It has been reported that upregulating MEF2A inhibited the function of mitochondria [
65]. And MEF2A played roles in muscle cells and drove DIK1-Dio3 cluster to inhibit mitochondrial activity. Meanwhile, MEF2A can regulate glycolytic versus oxidative myofiber development [
66]. In our study, we detected that the level of ATP was increased in circKIF20B-KD group and decreased in circKIF20B-OE group compared with the NC group. It suggested that circKIF20B may perform a critical role in cancer cells’ energy metabolism. To further confirm the role of circKIF20B and investigate the potential mechanism, we regulated the expression of MEF2A and measured the OCR and Mitochondrial membrane potential. Our result showed that circKIF20B significantly inhibited cellular ATP synthesis, mitochondrial membrane potential, and cellular OXPHOS. However, MEF2A could reverse these effects. Furthermore, exosomes loaded with dysregulated circKIF20B could significantly alter the OXPHOS in the target cells.
This result implied that circKIF20B/miR-615-3p/MEF2A signal axis reduced the gefitinib resistance by arresting the cell cycle, promoting cell apoptosis, and altering the cellular energy metabolism. This axis is a novel mechanism that differs from other conventional drug resistance models. The overexpression of circKIF20B could not only alter the energy metabolism and growth ability of parent cells but also be endogenously loaded by exosomes to inhibit the proliferation and gefitinib resistance of the recipient cells.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.