Serum exosomal microRNAs as predictive markers for EGFR mutations in non–small‐cell lung cancer

Abstract Background Current therapeutic drugs show positive effects on non–small‐cell lung cancer (NSCLC) patients with mutant epidermal growth factor receptor (EGFR) expression, whereas a lesser beneficial effect is generally noted on NSCLC patients with wild‐type EGFR. Therefore, identification of new detection methods for the accurate clinical diagnosis of NSCLC is essential. Methods In this study, tumor‐derived exosomes from the plasma of EGFR mutation and wild‐type NSCLC patients were isolated. Extensive exosomal miRNA profiling of EGFR mutation and wild‐type NSCLC patients, in comparison with healthy individuals, was performed using miRNA‐sequencing analysis. Results The variation of exosomal miRNA expression between control group (NR) and NCSLC samples (AM and AW) was identified. 96 significantly different expressed miRNAs were identified. Of these, 39 miRNAs were upregulated and 57 were downregulated. 11 miRNAs were downregulated, and 31 miRNAs were upregulated in the miRNA expression between NR and AM. Compared with healthy donors, 54 upregulated miRNAs and 36 downregulated miRNAs were observed in samples from AW patients. 40 different expressed miRNAs were identified in AM samples, compared with AW. Ten of upregulated miRNAs are miR‐260, miR‐1169, miR‐117, miR‐15b‐5p, miRNA‐731, miR‐342‐5p, miR‐ 898, miR‐1384, miR‐56, and miR‐1214. Ten of downregulated miRNAs are miR‐99b‐5p, miR‐1116, miR‐689, miR‐818, miR‐604, miR‐72, miR‐955, miR‐403, miR‐1228, and miR‐836. Conclusion The exosomal miR‐1169 and miR‐260 as potential candidates, which contain specific characteristics that can distinguish between wild‐type EGFR and mutant EGFR NSCLC patients in early‐stage cancers.

than 80% in Del19 and L858R. The remainder are rare mutations, including G719X, E709X, Del18, and Ins19 in exon 18, Ins20 and S768I in exon 20, and L861Q in exon 21. 4 In view of the EGFR mutations in NSCLC, the development of EGFR tyrosine kinase inhibitors (EGFR-TKIs), such as gefitinib, afatinib, and oxetinib, has demonstrated enhanced therapeutic effects in NSCLC patients with EGFR mutations. In contrast, EGFR-TKIs have shown poor therapeutic effects in NSCLC patients with wild-type EGFR. 5,6 Therefore, while early diagnosis is important, NSCLC differentiation is essential in the early management of NSCLC patients to identify the appropriate treatments strategies. Tissue biopsy represents the gold standard for the diagnosis of lung cancer and for genotyping mutant EGFR. 7,8 However, due to the invasive nature of the cancer and metastatic state, some tissue samples are difficult to obtain, as well as non-surgical resection tissues not being representative of the whole tumor. Therefore, tissue biopsies may not fulfill the requirements of early screening, dynamic follow-up, accurate, and comprehensive responses to the biological properties of the lung cancer.
MicroRNAs (miRNAs) are reported to play important roles in gene regulation and are detected in many organisms. 9 Accumulating studies demonstrate the involvement of miRNAs in tumorigenesis and development processes, such as cancer cell metastasis, hematopoiesis, proliferation, and cell death. In addition, miRNAs have been extensively studied in the diagnosis and prognosis of lung cancer.
Expression of miR-137, miR-372, let-7, miR-221, and miR-182 in lung cancer may be used as markers for predicting the survival rate of lung cancer patients. 10,11 Microarray data showed that 32 miRNAs are highly expressed in adenocarcinoma tissues. Moreover, miRNAs let-7e, miR-25, miR-191, miR-34a, and miR-34c are associated with prognosis. 12 The role of circulating miRNAs in non-invasive diagnosis has been extensively studied 13,14 ; however, few studies have identified the specific miRNA profiles of mutant EGFR and wild-type EGFR in early NSCLC diagnosis using peripheral blood.
With the advancement of more accurate detection methods in medicine, liquid biopsy has shown great promise due to the availability of samples for analysis. The three main targets of liquid biopsy include, circulating tumor DNA (ctDNA), circulating tumor cells (CTC), and exosomes. Among these, exosomes were discovered in 1983 and more recently have been the focus of extensive research within the liquid biopsy industry due to its wide distribution, high content, and stable structure. Exosomes are composed of 30-100 nm extracellular vesicles. Nucleotides and proteins secreted by specific cell types found in various body fluids, participated in communication between cells. 15 Unlike circulating miRNAs, exosomes are enriched within the circulatory system and prevent RNase degradation. 16 Identification of exogenous miRNA in body fluids confirms their potential as biomarkers for the clinical diagnosis or prognosis of NSCLC. 17,18 Twelve exosomal miRNAs are reported to be specifically expressed in adenocarcinoma patients with lung cancer compared with the control group, indicating a close correlation between circulating miRNA expression of tumor-derived exosomes and tumor miRNA. 19 Moreover, the expression of exosomal let-7f, miR-20b, miR-301, miR-379, and miR-200b in plasma or bronchoalveolar lavage is reported to differentiate NSCLC patients from healthy individuals. 20 Based on these published observations, we sought to further determine whether exosomal miRNAs could differentiate between NSCLC patients with wild-type and mutant EGFR. Therefore, the objective of this study was to identify tumor-derived exosomal biomarkers which are able to discriminate between EGFR mutation and wild-type NSCLC patients during the early stages of diagnosis, in order to improve the specificity and sensitivity of NSCLC diagnosis using this noninvasive method.
In this study, tumor-derived exosomes from the plasma of EGFR mutation and wild-type NSCLC patients were isolated. An extensive exosomal miRNA profiling of EGFR mutation and wild-type NSCLC patients paired with healthy individuals was performed using miRNA-sequencing. spectively. Twenty healthy individuals matched for sex and age in the testing and validation sets were included in the study. A symptomatic set of patients were included, who were suspected of suffering with NSCLC according to preliminary diagnosis, in order to verify the diagnostic accuracy of the selected miRNA panels. Detailed clinical data are summarized in Table 2. Plasma samples of patients were collected in vacuum blood tubes with anticoagulant prior to surgery or pharmacotherapy and handled within 1 h after collection. All individuals gave written consent for the use of their plasma samples and pathologic information to be used in this research. Patients in this study had no other malignant disorders, and the serum samples were nonhemolytic, obtained by centrifuging 2 ml peripheral blood at 280 g for 10 min. Supernatants were stored at -80°C until further analysis.

| Transmission electron microscopy
Exosomes samples were diluted to 0.5 mg/ml with phosphatebuffered saline (PBS), and spotted onto a glow-discharged copper grids covered with filter paper, and left to dry for 10 min under an infrared lamp. Finally, exosomes were stained with a drop of 1% phosphotungstic acid (Sigma, USA) for 1 min and dried using the infrared lamp. Exosomes were observed under transmission electron microscopy (TEM; JEM-1010 microscope) at 80 kV.

| Isolation of exosomes and extraction of RNA
Ultracentrifuge methods were utilized. The supernatant was centrifuged further at 110,000 g for 70 min. All steps were performed at 4°C. The pellets were resuspended in PBS for TEM. For patient serum sample, exosomes were extracted from 500 μl serum using a Total Exosome Isolation Kit (Invitrogen) following the manufacturers' protocol. Isolated exosomes were fully suspended in 100 μl PBS and then mixed with 1 ml of RNAiso reagent (Takara). Finally, the RNAs were dissolved in 20 μl RNase-free water. The quality of RNA was determined using a NanoDrop 2000 (Thermo) by OD260/280 and stored at −80°C until further analysis.

| Quantitative RT-PCR
Complementary DNA (cDNA) of miRNA was synthesized using the PrimeScript RT Reagent Kit (TaKaRa). Stem-loop quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) assays using TaqMan miRNA probes (Thermo) were subsequently conducted to quantitate the levels of mature miRNAs. For serum exosomes, mi-croRNA-16 was used as an internal control in the qRT-PCR analysis.
qRT-PCR reactions were performed on the real-time PCR system in 96-well plates at 95°C for 10 min, followed by 40 cycles at 95°C for 15 s, 60°C for 1 min and finally added a melting curve analysis at 95°C for 15 s, 60°C for 1 min, and 95°C for 15 s. Each sample was analyzed in duplicate.

| Western blot
The proteins of exosomes were extracted with RIPA buffer (Beyotime, China) and centrifuged at 11,809 g for 15 min at 4°C. A small amount of supernatant was quantified using the bicinchoninic acid assay (BCA) kit (Beyotime), and the remaining proteins were separated by SDS-PAGE and transferred onto a PVDF membrane (Millipore) according to standard methods. After blocking with 5% non-fat milk for 2 h, the membrane was incubated overnight at 4°C with primary antibodies for CD63 (proteintech). After washing the non-bound primary antibodies, the membrane was incubated with secondary HRP-conjugated anti-mouse or anti-rabbit antibodies (Beyotime). The antigen-antibody reaction was detected by a chemiluminescence HRP substrate kit (Millipore).

| Small RNA sequencing libraries preparation
Small RNA sequencing was performed using Illumina HiSeq (Illumina).
Total RNA of each sample was extracted using TRIzol Reagent (Invitrogen) and miRNeasy Mini Kit (Qiagen). Total RNA of each sample was quantified and qualified by Agilent 2100 Bioanalyzer (Agilent Technologies), NanoDrop (Thermo) and 1% agrose gel. Total RNA, 2 μg, with a RIN value above 7.5 was used for subsequent library preparations. Next-generation sequencing library preparations were constructed according to the manufacturer's protocols (NEBNext® Multiplex Small RNA library Prep Set for Illumina®). The 3 ´SR Adaptor for Illumina was ligated to the small RNA using 3´ Ligation Enzyme.
To prevent adaptor-dimer formation, the excess of 3 ´SR Adaptor was hybrid with SR RT Primer for Illumina. 5 ´SR Adaptor for Illumina was ligated to the small RNA using 5´ Ligation Enzyme and the first strand of cDNA was synthesized using ProtoScript II Reverse Transcriptase.
Each sample was then amplified by PCR for 12 cycles using P5 and

| Data analysis
Quality control of samples was performed to remove technical sequences, including adapters, PCR primers, or fragments thereof, quality of bases lower than 20 was passed through a FASTQ filter and the data were processed using the Cut adapt software (version 1.9.1)to ensure clean high quality data.
The miRDeep2 software was used to identify novel miRNAs from deep sequencing, and obtaining all the identified microRNA expression data.

For differential expression analysis, the DESeq/DESeq2
Bioconductor package was used, based on the negative binomial distribution. After adjustment using Benjamini and Hochberg's approach for controlling the false discovery rate, the p value <0.05 was set to detect differential expressed miRNAs.
Gene Ontology (GO)-TermFinder was used to identify GO terms, annotating a list of enriched genes with p value <0.05 defined as significant.
KEGG (Kyoto Encyclopedia of Genes and Genomes) is a collection of databases dealing with genomes, biological pathways, diseases, drugs, and chemical substances (http://en.wikip edia.org/ wiki/KEGG). In-house scripts were used to enrich significant differential expression gene in KEGG pathways.
For microRNA target mRNA analysis, miRanda detection software was used to predict the microRNA-mRNA in animals.

| Statistical analysis
Results are expressed as the mean ± standard error of the mean (SEM). Student's t test, the Mann-Whitney U test and chi-square test were used to compare the differences among different groups.
Receiver operator characteristic (ROC) curve and the area under curve (AUC) were performed to estimate the diagnostic accuracy. A p < 0.05 value was considered statistically significant. All statistical analyses were carried out using SPSS 22.0 and GraphPad Prism 8.

| Identification of exosomes and determination of miRNA in serum
To determine the efficacy of the methods used to isolate the serum exosomes, the exosomal marker was analyzed by Western blot and exosomal morphology of microvesicles was characterized by TEM.
Serum samples from NSCLC patients with wild-type or mutant EGFR were obtained before treatment. Exosomes were obtained using the ExoQuick exosome purification kit and observed under TEM.
TEM analysis revealed that the microvesicles separated from the serum were a round or elliptical membranes ( Figure 1A). Vesicle size ranged between 30 and 100 nm and displayed a uniform appearance ( Figure 1B). CD63, a tetraspanin family member which localizes internally in exosomal vesicles, were expressed in isolated serum exosomal pellets as specific bands, but not in the exosomes with depleted serum supernatant ( Figure 1C). These results demonstrated that the separated serum particles exhibit exosome characteristics, confirming the successful exosome separation techniques employed here.
A variety of RNAs, such as messenger RNAs (mRNAs), ribosomal RNAs (rRNAs), transfer RNAs (tRNAs), and miRNAs, were identified in exosomes. Therefore, it is essential to establish strategies for the accurate detection of miRNAs in exosomes. Here, miRNAs were purified from isolated serum exosomes using miRNA isolation kits.
From this, RNAs were subsequently isolated based on size, revealing twenty-five nucleotide-long small RNA representing the miRNAs.
Contrarily, 18S and 28S ribosomal RNA were barely detectable ( Figure 1D). These results confirmed the specificity of our experimental methods for obtaining miRNAs and further suggest that miR-NAs isolated from serum exosomes may be suitable and valuable for clinical sample analysis. The quality of purified exosomes using the ExoQuick system was determined using an Agilent Bioanalyzer (Agilent Technologies) than that of the control groups (57.5 ± 15.3 years); however, no significant differences in gender were noted among both groups.

| miRNA-sequencing analysis of exosomes
The primary content of total exosomes from NSCLC patients was obtained from the tumors within the plasma. Subsequent qRT-PCR validation on the total extracted exosomes from NSCLC patients with wild-type or mutant EGFR, as well as healthy donors was performed ( Figure S1A and B). Exosomal miRNA sequence analysis performed with next-generation sequencing techniques revealed 2284 miRNAs. The thermogram, histogram, venn, and volcano plots showed the variation of exosomal miRNA expression between control group (NR) and NCSLC samples (AM and AW) (Figure 2A-G). In total, 96 significantly different (p < 0.05) expressed miRNAs with greater than twofold changes were identified. Of these, 39 miRNAs were upregulated and 57 were downregulated (Figure 2A,     (Table 1). Therefore, the serum exosomal level of miR-1169 and miR-260 can be used as biomarkers to differentiate between AW and AM in NSCLC patients.

| Validation of miRNA-sequencing data by quantitative RT-PCR analysis
Circulating miRNAs are studied extensively as cancer diagnostic biomarkers in lung cancer and other types of carcinomas. However, target expressions are inconsistent in serum, plasma, or other bodily fluids. [21][22][23] Previous studies have demonstrated the enrichment of miRNAs in exosomes. [24][25][26][27] In this study, the expression of selected miR-NAs, miR-1169, and miR-260 was analyzed in tumor-derived exosomal and circulating plasma, due to their important roles in tumorigenesis.
To exclude the influence of surgery on the levels of exosomal  Table 2, and there are no statistical differences in age, gender, and others listed in Table 2. Exosomal miRNAs were obtained using a combination of the ExoQuick kit and the mirVana microRNA isolation system. qRT-PCR analysis

| Exosomal miRNA panels as diagnostic biomarkers in symptomatic patients
To evaluate the early diagnostic value of the miRNA-sequence profiling data for NSCLC, we selected miR-1169 and miR-260 and ( Figure 4B). These findings suggest that these two miRNA panels could serve as primary diagnostic biomarkers for NSCLC diagnosis.
By involving NSCLC patients with wild-type or mutant-specific miR-NAs in the differential diagnosis, further pathologic diagnosis could be improved.

| Bioinformatics analysis on selected miRNAs and their targets
To explore the potential biological roles of these exosomal miRNAs Furthermore, compared with healthy donors, 30 signaling pathways downstream of the miRNA target genes were enriched (p < 0.05) in NSCLC patients ( Figure 6A). The top six pathways identified were ECM-receptor interaction, cancer miRNAs, focal adhesion, osteoclast differentiation, salivary secretion, and glycerophospholipid metabolism ( Figure 6A). For the EGFR wild-type samples, the top six pathways selected for analysis include ECM-receptor interaction, cancer miRNA in cancer, focal adhesion, PI3 K-Akt signaling pathway, MAPK signaling pathway, and Ras signaling pathway ( Figure 6B). The top eight pathways selected for the EGFR mutation samples include salivary secretion, RAN degradation, ECM-receptor interaction, cancer miRNAs, measles, glycerophospholipid metabolism, MAPK pathway, and focal adhesion ( Figure 6C). Finally, significant differences were observed in three pathways between the EGFR mutation and wild-type samples, namely salivary secretion, RAN degradation, ECM-receptor interaction ( Figure 6D).

| DISCUSS ION
Recently studies have shown that exosomes contain small RNAs, in particular miRNAs. 28,29 Exosomal miRNAs with various biological functions are secreted by primordial cells to silence target mRNAs in other target cells. Exosome miRNAs are involved in the regulation of F I G U R E 6 KEGG pathway analysis and validation of downstream target genes. The top thirty KEGG pathways were enriched in the downstream miRNAs target genes intracellular signal transduction and gene expression through intercellular communication. 30 Several detection and purification methods of exosomes or circulating miRNAs have been reported. [31][32][33] In this study, we extracted the exosomes from plasma of NSCLC patients using exosomal separation systems and observed the characteristic exosome size and shape by TEM and DLS ( Figure 1A and B).
Expressions of the exosome-specific CD63 protein was confirmed by western blot analysis ( Figure 1C). The high quality of the purified exosomal miRNAs was confirmed using a Bioanalyzer ( Figure 1D).
The molecular components of exosomes can affect cancer progression by modulating carcinogenesis, growth, invasion, and metastasis.
Therefore, the circulating exosomal miRNAs may serve as potential biomarkers for differential disease diagnosis, prognosis, and treatment management.
Exosomal RNA profiling analysis is only feasible after the extraction of high quality RNA. Compared with cellular RNAs, exosomal RNAs are more stable, and are reportedly resistant to physical degradation such as prolonged storage and repeated freeze/thaw cycles. In this study, a comprehensive analysis of NSCLC patients derived exosomal miRNA profiles was performed using nextgeneration sequencing. There were 11 and 6 miRNAs expressed at remarkably higher levels, 13 and 8 miRNAs expressed at lower levels in EGFR mutation and wild-type NSCLC patients, respectively, compared with healthy volunteers (Figure 2, Table 2). Furthermore, 6 EGFR mutation-dependent and 2 EGFR wild type-specific tumorderived exosomal miRNAs that were expressed at higher levels, and 7 EGFR mutation-dependent and 2 EGFR wild type-specific miRNAs that were expressed at significantly lower levels were validated.
The reliability of our miRNA-sequencing data was verified with several proven diagnostic miRNAs for NSCLC and other types of carcinomas, such as let-7, miR-21, miR-24, and miR-486. 26,34,35 Novel EGFR mutation-and EGFR wild type-specific biomarkers, such as miR-361b-5p and miR-10b-5p, were also identified suggesting that they can potentially act as effective indicators in EGFR mutation and EGFR wild-type diagnosis.
Thus, the serum exosomal level of miRNAs may act as biomarkers of NSCLC, but may not distinguish between AW and AM. The expression of miR-106b-3p, miR-142-5p, miR-940, miR-548ad-5p, miR-422, and miR-1273 has been previously reported as a diagnostic biomarker in other forms of cancer. 36 However, NSCLC patients with wild-type EGFR-specific miR-1169, and NSCLC patients with mutant EGFR-specific miR-260 have not been reported in NSCLC samples. 10,22,[37][38][39] The findings of this study indicated that miR-1169 and miR-260 exosomal miRNAs may be specific characteristics to distinguish NSCLC patients with wild-type EGFR and mutant EGFR in the early cancer stages. Total miRNAs from circulating plasma may be composed of endogenous cellular miRNAs derived from the debris of all kinds of cells. In addition, circulating plasma miRNAs, which are not associated with exosomes, were reported to have differential stability when treated by RNaseA.
Those findings suggest that exosomal miRNAs are more suitable for developing diagnostic biomarkers due to their stability and miRNA enrichment. The NSCLC patients with wild-type EGFR-or mutant EGFR-specific miRNAs validated in this study demonstrate that exosomes are actively excreted from cancer cells and that their specific miRNA components depend on the cells from which they originated.

ACK N OWLED EG M ENTS
This work was supported by Medical Health Science and Technology Project of Zhejiang Province (Grant No. 2019RC069).
The authors alone are responsible for the content and writing of the paper.

CO N FLI C T O F I NTE R E S T
The authors declare that no conflicts of interests exist.

DATA AVA I L A B I L I T Y S TAT E M E N T
The all of data in this manuscript are availability.