key: cord-1041227-rf89sry8
authors: Yang, Chenghao; Jia, Xiaoyu; Zhou, Jinbao; Sun, Qiangling; Ma, Zhongliang
title: The MiR-17-92 Gene Cluster is a Blood-based Marker for Cancer Detection in Non-small-cell Lung Cancer
date: 2020-05-11
journal: Am J Med Sci
DOI: 10.1016/j.amjms.2020.05.004
sha: 16a363e36b995c059b697ed430ac0a107b0f3fae
doc_id: 1041227
cord_uid: rf89sry8

BACKGROUND: Lung cancer is one of the most malignant cancers threatening human health. The miR-17-92 gene cluster is a highly conserved oncogene cluster encoding six miRNAs: miR-17, miR-18a, miR-19a, miR-19b-1, miR-20a and miR-92a. This study explored whether these miRNAs can be used as diagnostic markers for non-small-cell lung cancer (NSCLC). METHODS: Serum samples were collected from healthy subjects (n = 23) and NSCLC patients at various stages (n = 74). Serum RNA was extracted by the TRIzol-glycogen method, and cDNA libraries were constructed by reverse transcription. Quantitative real-time polymerase chain reaction (qRT-PCR) was utilized to detect the expression levels of the 6 miRNAs. RESULTS: The expression levels of the six miRNAs varied in different stages of NSCLC. Thus, two receiver operating characteristic (ROC) curves, i.e., normal subjects and stage I-III patients and normal subjects and stage IV patients, of each miRNA were established to determine the interval of normal ΔCt values. The two areas under the curve (AUCs) of each miRNA were investigated (miR-17: 0.8097 and 1.000; miR-18a: 0.7388 and 0.9907; miR-19a/19b: 0.8451 and 0.5104; miR-20a: 0.8975 and 1.000; miR-92a: 0.8097 and 0.8342). In addition, a high positive correlation was discovered between miR-17 and miR-20a expression. Combining these two miRNAs can improve the screening effect of NSCLC. CONCLUSION: The miR-17-92 gene cluster can likely serve as a diagnostic marker in NSCLC.

Among all cancers, lung cancer is one of most malignant cancers, with its incidence ranking second and its mortality rate ranking first 1 . Lung cancer has two main types, small-cell lung cancer (SCLC) and non-small-cell lung cancer (NSCLC) 2 , the latter accounting for 85% of lung cancer 3 . Both of them have a low 5-year survival rate. In the United States, the 5-year survival rate of lung cancer is only 17.4% 4 , and this value is even lower in developing countries. Therefore, lung cancer is one of the most serious diseases threatening human health 1 .

MiRNAs are small RNA molecules containing 19-22 nucleotides found in plants, animals and some viruses 5 . They are paramount in maintaining the function of normal cells. They also have special traits such as tissue specificity, high conservation and changes with time 6 .

The atypical expression of miRNAs may induce many diseases, including tumorigenesis, such as lung carcinoma, liver cancer, and breast cancer 7, 8 . Therefore, the regulation of miRNAs in malignant tumors provides a new direction for the prevention and treatment of tumors 9, 10 .

Studies have revealed that miRNAs play a preponderant role in tumor development, metastasis, invasion and treatment 11 . The aberrant expression of some miRNAs may be harnessed as early diagnostic markers in patients who have cancer 12 .

The miR-17-92 gene cluster 13, 14 is a highly conserved gene cluster containing six members: miR-17, miR-18a, miR-19a, miR-19b-1, miR-20a and miR-92a-1 15 . This gene cluster is the first discovered cluster of miRNA oncogenes 16, 17 . In this study, sequences with higher expression levels in the -3p and -5p sequences of each miRNA in the miR-17-92 gene cluster were selected (Table 1) . Among them, miR-19a-3p has exactly the same sequence as miR-19b-1-3p 18 . The miRNAs below, if not indicated by -3p/-5p, are default to the miRNAs shown in Table 1 .

Studies have revealed that the miR-17-92 gene cluster plays an important role in the occurrence and development of a variety of tumors 19, 20 . For this reason, it has received extensive attention from researchers around the world. Liu 23 .

In a previous study, we discovered that miR-18a-5p in the miR-17-92 gene cluster directly targets interferon regulatory factor 2 (IRF2), prominently lowering the expression of its protein 24 . Since IRF2 has the ability to inhibit lung cancer cell growth, miR-18a-5p promotes the proliferation and migration of NSCLC by targeting IRF2. Our lab also investigated whether the miR-17-92 cluster promoted NSCLC by targeting the suppressor gene sprouty homolog 4 (Spry4).

By querying the National Center for Biotechnology Information (NCBI) database, the roles of the miR-17-92 gene cluster in lung cancer were elucidated to a certain extent. Based on our previous research, this study explored the expression of the miR-17-92 cluster in the sera of NSCLC patients to screen for possible miRNAs that can be used as diagnostic markers for NSCLC in the clinic 25 . The diagnostic effects of the existing clinical markers of NSCLC were then compared with those of the miRNAs 26 . Our findings may provide new ideas for improving the clinical screening and diagnosis of NSCLC and thus promoting the precise treatment of lung cancer 27 .

Our procedures were approved by the Ethics Review Board of Shanghai Chest Hospital Affiliated to Shanghai Jiao Tong University, and written informed consent was obtained.

Sera were collected from 97 subjects, including 74 NSCLC patients (from stage I to IV) and 23 normal subjects from the Shanghai Chest Hospital.

In this study, various procedures were performed (Fig. 1) . Serum RNA was extracted by the TRIzol-glycogen method, and then cDNA libraries were constructed by reverse transcription (SYBR Prime Script miRNA RT-PCR Kit, TaKaRa Bio, Shibuya, Japan). Then, quantitative real-time polymerase chain reaction (qRT-PCR) (ChamQ SYBR qPCR Master Mix, Vazyme, Nanjing, China) was utilized to detect miRNA expression.

The receiver operating characteristic (ROC) curves of the six miRNAs were plotted. By observing the area under the curve (AUC), the diagnostic effect of the miRNAs could be judged. The value corresponding to the maximum of the Youden index, i.e., sensitivity + specificity -1, was taken as the best cut-off point.

The total RNA of the serum samples was extracted using TRIzol (TranS, Beijing, China).

Then, 30-50 μL of RNase-free ddH 2 O was added to fully dissolve the RNA. The RNA solutions were stored at -70 °C for subsequent experiments. The concentration of total RNA was measured by a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA). Serum samples whose concentration of total RNA was less than 20 ng/μL were excluded from this study.

The reverse transcription solution was prepared according to the manufacturer's protocol (SYBR Prime Script miRNA RT-PCR Kit, TaKaRa Bio, Shibuya, Japan). The mixture was incubated at 37 °C for 60 min and then at 85 °C for 5 s to inactivate the enzyme. RNase-free ddH 2 O was added to the obtained reverse transcription reaction solution to reach 100 μL 28 .

The diluent was added to the qRT-PCR system for quantitative detection 29, 30 .

qRT-PCR solutions were prepared according to the manufacturer's protocol (ChamQ SYBR qPCR Master Mix, Vazyme, Nanjing, China). The qRT-PCR protocol was conducted on a CFX96 Touch Real-Time PCR Detection System (BIO-RAD, Hercules, CA) in 96-well plates at 95 °C for 3 min, followed by 42 cycles of 95 °C for 10 s and 55 °C for 30 s. A melt-curve was plotted to evaluate the specificity of the PCR product 31 . Cycle threshold (Ct) values less than 40 were used in this study.

The data are analyzed and compared by the 2 -ΔΔCt quantitative method relative to the reference RNU6B (U6), ΔCt = Ct miRNA -Ct U6 32, 33 .

The ROC curve refers to the connection of points drawn under a specific cut-off point, with the false positive rate (1 -specificity) obtained by the subjects under different cut-off points as the abscissa and the true positive rate (sensitivity) as the ordinate.

In this study, we selected different ΔCt values of each miRNA as cut-off points, according to the qRT-PCR results. In stages I-III, several miRNAs are downregulated. In that case, ΔCt values greater than the cut-off point are regarded as positive, while ΔCt values less than the cut-off point are regarded as negative. In stage IV, all miRNAs are upregulated.

Then ΔCt values greater than the cut-off point are regarded as negative, while ΔCt values less than the cut-off point are regarded as positive.

In the control group, the true negative rate (specificity) can be calculated as the "# of controls whose test result is negative / # of controls in total", while the false positive rate can be calculated as 1 -specificity. In the patient group, the true positive rate (sensitivity) can be calculated as the "# of patients whose test result is positive / # of patients in total", while the false negative rate can be calculated as 1 -sensitivity.

By connecting the points drawn under different cut-off points, the AUC can be determined. AUC is the area enclosed by the coordinate axis under the ROC curve. The closer the AUC is to 1, the better the discrimination effect of the detection method.

Typically, an AUC greater than 0.8 indicates that the detection method has relatively good effects.

The best cut-off point is identified by the Youden index, which is sensitivity + specificity -1. The Youden index represents the total ability of the screening method to find true patients and non-patients. In this study, the cut-off point that has the maximum value of the Youden index was chosen as the best cut-off point.

Unpaired T test was used to compare two groups of continuous variables. The diagnostic effect of the miRNAs was evaluated by ROC curves. Then, the AUCs were calculated to compare the diagnostic value of each miRNA. The Pearson correlation coefficient was utilized to measure the degree of correlation between two continuous variables. The maximum Youden index, i.e., sensitivity + specificity -1, was taken as the best cut-off point. Binary logistic regression was used to create the prediction model for combining the diagnostic effects of multiple miRNAs together. The P value 34 was calculated for each statistical analysis to indicate whether the null hypothesis can be refuted 33 .

Serum samples from 23 normal subjects and 74 NSCLC patients were used in this study.

The demographics and clinical characteristics of the patients with NSCLC are listed in Table   S1 . There was no significant difference in the sex or age distributions between the normal subjects and patients with NSCLC in different stages.

Compared with that of normal subjects, the level of miR-17 in NSCLC patients showed no significant change in stage I. In stage II and III patients, it showed significant downregulation, and in stage IV patients, it was markedly upregulated ( Fig. 2A) . From this phenomenon, the ROC curve of normal subjects and stage I-III patients and the ROC curve of normal subjects and stage IV patients could be plotted. Then, the interval of normal human ΔCt values could be determined.

The ROC curve for normal subjects and stage I-III patients determines the upper limit of the ΔCt value, and the ROC curve for normal subjects and stage IV patients determines the lower limit of the ΔCt value.

The results showed that in the ROC curve of normal subjects and stage I-III patients, the AUC was 0.8097, and the best cut-off point was 3.085. The sensitivity and specificity of this value were 0.8250 and 0.6818, respectively (Fig. 2B ). In the ROC curve of normal subjects and stage IV patients, the AUC was 1.0000, and the best cut-off point was -0.1317. The sensitivity and specificity of this value were 1.0000 (Fig. 2C) .

Therefore, the ΔCt value interval of a normal person can be set to [-0.1317, 3.085]. If the test result is less than -0.1317, it is suggested to be stage IV NSCLC; if the result is greater than 3.085, it is suggested to be stage I-III NSCLC.

The expression level of miR-18a in NSCLC patients showed no significant change in stages I and III. In stage II patients, it showed significant downregulation, and in stage IV patients, it showed marked upregulation (Fig. 3A) . Although there was no significant change in stage I and III, patients with low expression still accounted for a considerable proportion.

Therefore, the ROC curve of normal subjects and stage I-III patients and the ROC curve of normal subjects and stage IV patients could be plotted. Then, the interval of normal human ΔCt values could be determined.

The results showed that in the ROC curve of normal subjects and stage I-III patients, the AUC was 0.7388, and the best cut-off point was 4.788. The sensitivity and specificity of this value were 0.8000 and 0.4737, respectively (Fig. 3B ). In the ROC curve of normal subjects and stage IV patients, the AUC was 0.9907, and the best cut-off point was 2.620. The sensitivity and specificity of this value were 1.0000 and 0.8947, respectively (Fig. 3C) .

In that case, the ΔCt value interval of a normal person can be set to [2.620, 4.788 ]. If the test result is less than 2.620, it is suggested to be stage IV NSCLC; if the result is greater than 4.788, it is suggested to be stage I-III NSCLC.

The expression level of miR-19a/19b in NSCLC patients showed no significant change in stages I and IV. In stage II and III patients, it was significantly downregulated (Fig. 4A) . The ROC curve of normal subjects and stage I-III patients and the ROC curve of normal subjects and stage IV patients were plotted.

The results showed that in the ROC curve of normal subjects and stage I-III patients, the AUC was 0.8451, and the best cut-off point was 4.270. The sensitivity and specificity of this value were 0.7885 and 0.5556, respectively (Fig. 4B ). In the ROC curve of normal subjects and stage IV patients, the AUC was only 0.5104, showing that there was almost no effect in the diagnosis of stage IV NSCLC patients using miR-19a/19b (Fig. 4C) .

Nevertheless, miR-19a/19b can be used as a diagnostic marker for stage I-III NSCLC. If the test result is greater than 4.27, it is suggested to be stage I-III NSCLC. Test takers should perform further examinations to confirm or rule out such possibilities.

The level of miR-20a in NSCLC patients was significantly downregulated in stage I, II and III patients. In stage IV patients, it was markedly upregulated (Fig. 5A) . Similar to the previous method, the ROC curve of normal subjects and stage I-III patients and the ROC curve of normal subjects and stage IV patients could be plotted. Then, the interval of normal

human ΔCt values could be determined.

The results showed that in the ROC curve of normal subjects and stage I-III patients, the AUC was 0.8975, and the best cut-off point was 3.425. The sensitivity and specificity of this value were 0.9020 and 0.7273, respectively (Fig. 5B ). In the ROC curve of normal subjects and stage IV patients, the AUC was 1.0000, and the best cut-off point was 0.01167. The sensitivity and specificity of this value were 1.0000 (Fig. 5C) .

Therefore, the ΔCt value interval of a normal person can be set to [0.01167, 3 .425], and if the test result is less than 0.01167, it is suggested to be stage IV NSCLC; if the result is greater than 3.425, it is suggested to be stage I-III NSCLC.

The expression level of miR-92a-1 in NSCLC patients showed no significant change in stage I. In stage II and III patients, it showed significant downregulation, and in stage IV patients, it showed marked upregulation (Fig. 6A) . The ROC curve of normal subjects and stage I-III patients and the ROC curve of normal subjects and stage IV patients could be plotted. Then, the interval of normal human ΔCt values could be determined.

The results showed that in the ROC curve of normal subjects and stage I-III patients, the AUC was 0.8097, and the best cut-off point was 6.852. The sensitivity and specificity of this value were 0.8302 and 0.5909, respectively (Fig. 6B ). In the ROC curve of normal subjects and stage IV patients, the AUC was 0.8342, and the best cut-off point was 5.342. The sensitivity and specificity of this value were 0.7647 and 0.6364, respectively (Fig. 6C) .

Therefore, the ΔCt value interval of a normal person can be set to [5.342, 6 .852], and if the test result is less than 5.342, it is suggested to be stage IV NSCLC; if the result is greater than 6.852, it is suggested to be stage I-III NSCLC.

The following diagnostic markers are used in the clinic: carbohydrate antigen 125 (Ca125), carcinoembryonic antigen (CEA), cytokeratin-19 fragment (CYFRA21-1), neuron-specific enolase (NSE), and squamous cell cancer antigen (SCCAG) 35 .

This study collected data on these commonly used diagnostic markers from 74 patients with NSCLC ( Table 2 ). The expression level of the miR-17-92 cluster in the sera of these patients was previously examined. Table 3 was calculated from Table 2 , and the overall diagnostic effects of the 6 miRNAs are shown in Table 4 . It is apparent that both the sensitivity and specificity of the miRNAs were better than those of the current biomarkers used in the clinic.

Based on the results above, we further analyzed whether there is a synchronization between miRNAs in the miR-17-92 cluster. The qRT-PCR data of patients in each stage were assessed, and the expression of miR-17 and miR-20a showed the same increasing and decreasing trends. Therefore, a correlation analysis between miR-17 and miR-20a was performed.

The correlation analysis showed that in patients with stage I and II disease, the Pearson correlation coefficient (r) was 0.9080 (Fig. 7A ). In patients with stage III and stage IV disease, this value was 0.9924 (Fig. 7B) . These results show that miR-17 is highly positively correlated with miR-20a expression. Therefore, it will be clinically more accurate to use miR-17 in combination with miR-20a for diagnosis.

Since there is a high positive correlation between miR-17 and miR-20a, we then investigated whether these two miRNAs can be combined to improve the screening effect of stage I-III NSCLC. A prediction model was established by using binary logistic regression. This model combined miR-17 and miR-20a as a 2-miRNA panel, with the algorithm, ln(p/1-p) = -1.616 -1.359 × miR-17 + 1.294 × miR-20a. According to the ROC curve plotted with this model, the AUC value reached 0.9479, which was higher than that of either of the two individual miRNAs (Fig. 8) . The best cut-off point was 0.6044. The sensitivity and specificity at this cut-off point were 0.9756 and 0.9090, respectively. These results indicated that this 2-miRNA panel is a stable marker for the diagnosis of NSCLC patients.

For further research, qRT-PCR was utilized to detect the expression level of the miRNAs in normal cell lines and NSCLC cell lines 25 .

In this study, human bronchial epithelial cells (BEAS-2B) were used as a control. The NSCLC cell lines used in this study were A549, H1299, HCC827, PC-9 and 95-D. The results indicated that the miRNAs of the miR-17-92 gene cluster in the NSCLC cell lines showed significantly higher expression compared to that in BEAS-2B (Fig. 9 ).

As mentioned in the introduction, NSCLC, the dominant type of lung cancer, seriously threatens human health 33 . The 5-year survival rate of early-stage NSCLC is more than 90%, while this value is only 15% in late-stage NSCLC 36 . Nevertheless, the detection rate of early-stage NSCLC is low, partly because early-stage NSCLC has almost no or slight symptoms. Patients tend to ignore these symptoms and thus miss the best opportunity to receive treatment.

Another reason is the lack of means to conduct an early diagnosis of lung cancer.

Low-dose spiral computed tomography has been reported to effectively screen early-stage NSCLC 37 . However, this method has disadvantages, such as potentially harmful radiation and high costs 38 . Therefore, a relatively low cost and harmless method should be discovered to provide screening for early-stage NSCLC in a wide range of populations.

Circulating miRNAs in serum may be a promising marker for early diagnostics in NSCLC.

It has several advantages. First, serum can be obtained in an easier way. Second, its costs are more affordable. Last, circulating miRNAs show characteristics such as tissue specificity, high conservation and changes with time.

Various studies have revealed that some circulating miRNAs have the potential to be used as diagnostic markers. Huang et al. suggested that six serum-based miRNAs have the potential to be diagnostic markers for gastric cancer 39 . Wang et al. found that five miRNAs, including miR-205a-5p, miR-145-5p, miR-10a-5p, miR-346 and miR-328-3p, were more highly expressed in ovarian cancer and thus can be used as diagnostic markers 40 . Yang et al.

reported that a four-miRNA panel is effective in diagnosing NSCLC 18 . These studies all confirmed that circulating miRNAs could be diagnostic markers.

However, miRNAs in serum have to be processed via multiple procedures to perform a qRT-PCR test, which will take a relatively long time if conducted manually. This problem also Other studies related to diagnostic markers usually set one cut-off point for each miRNA, while our study set the interval of normal humans, which had two cut-off points.

This is because we discovered that the miR-17-92 gene cluster was often downregulated in stage I-III NSCLC and markedly upregulated in stage IV NSCLC. The expression pattern of this gene cluster enabled us to distinguish patients from different stages of NSCLC, which was a unique superiority compared to other miRNAs.

Studies concerning diagnostic markers usually include 2 phases, the training phase and validation phase. Because of the lack of serum samples, the validation phase has yet to be carried out. More serum samples need to be collected and further investigated.

MiR-17 and miR-20a in the miR-17-92 cluster may be promising markers for screening NSCLC. Moreover, the combination of these 2 miRNAs has a better effect than using them

individually. Further studies with more clinical samples and more sensitive methods are needed to expand the quantity of samples to further confirm these results. Afterward, the results will be able to be used as diagnostic markers in the future.

Z.M. proposed the concept; Q.S. and J.Z. collected patient sera and determined their pathological characteristics; C.Y. and X.J. developed the methodology, conducted experiments, and collected and analyzed data; C.Y. wrote the manuscript; Z.M. administered the whole project and revised the manuscript. 

Molecular histology of lung cancer: From targets to treatments

Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries CA: a

Lung Cancer Screening, Version 3

genomics, biogenesis, mechanism, and function cell

Liquid Biopsy in Tumor Genetic Diagnosis

Expression, regulation and mechanism of action of the miR-17-92 cluster in tumor cells

MicroRNA expression profiles classify human cancers

MicroRNA biogenesis pathways in cancer

Two Plasma MicroRNA Panels for Diagnosis and Subtype Discrimination of

Circulating microRNAs as stable blood-based markers for cancer detection

Manel E Non-coding RNAs in human disease

The miR-17-92 cluster as a potential biomarker for the early diagnosis of gastric cancer: evidence and literature review

Gallicano GI miR-17 family miRNAs are expressed during early mammalian development and regulate stem cell differentiation

Biology of MiR-17-92 cluster and its Progress in lung

Mendell JT miRiad roles for the miR-17-92 cluster in development and disease

A polycistronic microRNA cluster, miR-17-92, is overexpressed in human lung cancers and enhances cell proliferation

Serum microRNA Signature Is Capable of Early Diagnosis for Non-Small Cell Lung

Targeted deletion reveals essential and overlapping functions of the miR-17∼ 92 family of miRNA clusters

Detection of circulating exosomal miR-17-5p serves as a novel non-invasive diagnostic marker for non-small cell lung cancer patients Pathology-Research and Practice

miR-17-92 functions as an oncogene and modulates NF-κB signaling by targeting TRAF3 in MGC-803 human gastric cancer cells

miR-17-92 promotes leukemogenesis in chronic myeloid leukemia via targeting A20 and activation of NF-κB signaling

miR-17-92 plays an oncogenic role and conveys chemo-resistance to cisplatin in human prostate cancer cells

MicroRNA-18a-5p functions as an oncogene by directly targeting IRF2 in lung cancer

High-throughput qRT-PCR validation of blood microRNAs in non-small cell lung cancer

A serum circulating miRNA diagnostic test to identify asymptomatic high-risk individuals with early stage lung cancer

Clinical evaluation and therapeutic monitoring value of serum tumor markers in lung cancer

Jensen SG Evaluation of two commercial global miRNA expression profiling platforms for detection of less abundant miRNAs

Serum microRNA signatures identified in a genome-wide serum microRNA expression profiling predict survival of non-small-cell lung cancer

A novel and universal method for microRNA RT-qPCR data normalization

Diagnostic value of a plasma microRNA signature in gastric cancer: a microRNA expression

Analysis of Relative Gene Expression Data Using Real-Time Quantitative PCR and the 2 −ΔΔ

Lianidou ES Quantification of Circulating miRNAs in Plasma : Effect of Preanalytical and Analytical Parameters on Their Isolation and Stability

A Plasma Biomarker Panel of Four MicroRNAs for the Diagnosis of

Assessment of Seven Clinical Tumor Markers in Diagnosis of Non-Small-Cell Lung Cancer Disease markers

Detterbeck FC The eighth edition TNM stage classification for lung cancer: What does it mean on main street?

Benefits and harms of computed tomography lung cancer screening strategies: a comparative modeling study for the U

Serum and blood based biomarkers for lung cancer screening: a systematic review

Six Serum-Based miRNAs as Potential Diagnostic Biomarkers for Gastric Cancer Cancer epidemiology

The Value of Plasma-Based MicroRNAs as Diagnostic Biomarkers for Ovarian Cancer The American journal of the medical sciences

MiR-17-5p enhances pancreatic cancer proliferation by altering cell cycle profiles via disruption of RBL2/E2F4-repressing complexes

Aberrant microRNAs expression in CD133(+)/CD326(+) human lung adenocarcinoma initiating cells from A549

The authors would give great thanks to Dr. Yanli Li and Yang Shao for their constructive discussions and Nanxiang Mao for his assistance in the experiments.