Assessment of a panel of miRNAs in serum and pleural fluid for the differential diagnosis of malignant and benign pleural effusion

Abstract

BACKGROUND:

Differential diagnosis between malignant pleural effusion (MPE) and benign pleural effusion (BPE) remains a clinical challenge.

OBJECTIVE:

The aim of the study is to assess the efficacy of the serum and pleural fluid (PF) miRNA panels in distinguishing MPE from BPE.

METHODS:

Fourteen candidate miRNAs which were shown aberrant expression in lung cancer based on previous studies were tested by quantitative real-time PCR (qRT-PCR) in 20 MPE patients and 20 BPE patients. Significantly aberrantly expressed miRNAs were further assessed by qRT-PCR in all patients enrolled in this study. A receiver operating characteristic (ROC) curve was constructed, and the area under the ROC curve (AUC) was calculated to evaluated the diagnostic performance of the miRNAs.

RESULTS:

miR-21, miR-29c and miR-182 were found to be significantly aberrantly expressed in the serum and PF of MPE patients. The AUCs for the combination of miR-21, miR-29c and miR-182 in serum and PF were 0.832 and 0.89 respectively in distinguishing MPE from infection-associated PE including tuberculous pleurisy and parapneumonia PE, and 0.866 and 0.919 respectively for differentiating MPE from heart failure-associated PE, which were superior to AUC of each individual miRNAs.

CONCLUSIONS:

miR-21, miR-29c and miR-182 in serum and PF could be useful biomarkers for diagnosis of MPE.

Keywords

Pleural effusion differential diagnosis biomarkers miR-21 miR-29c miR-182

Table 1
Patient demographics and etiology of pleural effusion (PE)

	Malignant PE	Benign PE
Age, yr (median)	57	48
Range	42–76	19–73
Sex (%)
Male	67 (54.5%)	63 (56.3%)
Female	56 (45.5%)	49 (43.7%)
Etiology of PE	Lung adenocarcinoma (64)	Tuberculous pleurisy (54)
	Lung squamous cell cancer (26)	Pneumonia (33)
	Small cell lung cancer (10)	Heart failure (25)
	Breast cance r(12)
	Colorectal cancer (4)
	Ovarian cancer (3)
	Gastric cancer (2)
	Renal cancer (1)
	Pancreatic cancer (1)

1. Introduction

Pleural effusion (PE) are common complications in patients with advanced tumor, infection or inflammation [1]. Cancer is one of the main causes of PE, and $>$ 90% of malignant pleural effusion (MPE) are due to pleural metastases of malignant tumors [2]. The most common primary tumor of causing pleural metastasis is lung cancer, followed by breast cancer, malignant lymphoma, gynecological and digestive tract cancers [3, 4]. Diagnosis of PE is of particular importance for cancer patients and a co-existing MPE implies an advanced-stage of tumor and intricate clinical management. Although cytological examination is considered a standard method for the diagnosis of MPE, the sensitivity is still limited to 50%–60% even with repeated thoracenteses [5]. Blind needle pleural biopsy confer little additive diagnostic accuracy for false-negative cytology [6]. Thoracscopy can achieve a diagnosis in $>$ 90% of patients with MPE, but this procedure requires access to a skilled operator, is invasive, unsuitable for frail patients and has established risk of complications, such as infection and hemorrhage [7]. A series of tumor markers have been well studied for their ability to improve the diagnosis of MPE, including carcinoembryonic antigen (CEA), CYFRA21-1, carbohydrate antigen (CA) 125, CA15-3, and CA19-9, but researchers have been unable to identify a reliable tumor marker with satisfactory sensitivity and specificity [8, 9].

microRNAs (miRs) represent a class of naturally occurring non-coding RNA molecules with regularly 19–25 nucleotides. The role of miRNA deregulation in carcinogenesis, cancer progression, and metastasis has been well established. miRNA signature may be used to identify specific type of cancer and thus are considered a promising tool for patient diagnosis, prognosis and clinical management [10]. miRNAs may be extracted from all biological samples, including tissue and liquid specimens, which include blood, saliva, sputum, urine, and bronchial lavage and serous fluids [11, 12]. In recent years, miRNA analyses conducted in surgical specimens, plasma, serum, sputum, and urine from patients with cancer have demonstrated its potential role in diagnosis and prognosis. However, the expression of miRNA in PE specimens has been less investigated [13, 14]. In the present study, we assessed the utility of miRNAs in serum and PF samples as biomarkers of differential diagnosis of MPE and BPE. Because most of MPE originate from lung cancer, in this study, 14 miRNAs proven to has aberrant expression in the tissue and blood specimens of lung cancer patients in previous studies were chosen as candidate diagnostic biomarkers (Supplementary Table 1 and Supplementary References). Then these miRNAs were examined by quantitative real-time PCR (qRT-PCR) in serum and pleural fluid (PF) specimens from 20 patients with MPE caused by lung cancer and 20 patients with benign pleural effusion (BPE). Significantly aberrantly expressed miRNAs were further analyzed for their diagnostic value in distinguishing MPE from BPE.

2. Material and methods

2.1 Study subjects, specimens, and study design

We prospectively enrolled 235 PE patients who were administrated to the Affiliated Hospital of Jiangsu University in China from March 2015 to December 2018. All patients with PE had a definite etiology documented by examination of medical imaging, PF biochemistry, microbiology, cytology, pleural biopsy, or endoscopic biopsy. The demographics and etiology of PE patients are summarized in Table 1. Etiology of MPE included lung adenocarcinoma (LAD), lung squamous cell carcinoma (LSCC), breast cancer, colorectal and ovarian cancer, etc. BPE consisted of tuberculous pleurisy (TBP), parapeumonia PE and heart failure-associated PE (HFAPE). In order to evaluate precisely diagnostic value of serum and PE miRNAs for MPE and BPE, the PE diagnosis protocol follow the recommendations set out in the guidelines published by different medical societies [15, 16, 17]. The diagnosis of MPE was based on the presence of cancer cells or tissue in either PF or biopsy pleural specimen, or if patients had disseminated malignancy and there was no alternative explanation for the PE [15, 16]. TBP was defined by the presence of pleural caseating granulomas, or the detection of Mycobacterium tuberculosis in patients sputum, PF or pleural biopsy specimens, or a high levels of PF adenosine deaminase (ADA) was determined ( $>$ 40 U/L) with an improvement of the pleurisy after antituberculous therapy [16]. Parapneumonia PE was diagnosed as any PE associated with bacterial pneumonia, lung abscess or bronchiectasis [16]. The diagnosis of HFAPE was based on the guideline of the 2012 European Society of Cardiology for heart failure [17], and patients met following criteria: typical symptoms and signs of HF; a left ventricular eject fraction (LVEF) $<$ 50% on echocardiography; N-terminal pro B-type natriuretic peptide (NT-proBNP) level elevated to three or more times of the upper limit of normal; with transudative PE classified by Light’s criteria [18].

PF samples were collected by thoracocentesis or during pleural biopsy before start any treatment. Peripheral blood sample were collected from PE patients and healthy controls. The time interval between collection of PF and blood samples in PE patients was 0–2 days in order to prevent possible miRNA data inconsistency between PF and matched serum samples caused by longer sampling interval. All samples were transported to the laboratory within 30 min of collection. Then the PF were centrifuged at 3000 g for 20 min and supernatants were transferred to new microfuge tubes (500 ul in each tube). Blood samples were centrifuged at 5000 g for 30 min to completely remove any cell debris and serum was transferred to new microfuge tubes. PF supernatants and serum were frozen instantly and stored at $-$ 80 ${{}^{\circ}}$ C until use. The experiments were undertaken with the understanding and written consent of each subject, and that the study conforms with The Code of Ethics of the World Medical Association (Declaration of Helsinki), printed in the British Medical Journal (18 July 1964). All patients provided written informed consent to participate in the study. The study was approved by the Ethics Review Board of Affiliated Hospital of Jiangsu University, China (No. 20150016).

In the initial phase of this study, we selected 14 candidate miRNAs which have been reported to be aberrantly expressed in the tissue or blood samples of lung cancer patients (Supplementary Table 1 and Supplementary References). The PF and serum levels of each of these miRNAs were tested by quantitative real-time PCR (qRT-PCR) in 20 patients with MPE caused by lung cancer (15 LAD and 5 LSCC) and 20 patients with BPE (10 TBP, 5 parapneumonia-PE and 5 HFAPE). When comparing MPE patients with BPE patients, only those miRNAs with a mean fold change $\geqslant$ 2 and $P$ -value $<$ 0.05 in both PF and serum samples were chosen for further analysis. According to the criteria, 3 of the 14 miRNA, including miR-21, miR-29c and miR-182 were found to be markedly aberrantly expressed in both PF and serum samples from MPE patients (Supplementary Table 1). Among them, miR-21 and miR-182 were significantly up-regulated, whereas miR-92c was markedly down-regulated. Then we further tested the levels of the 3 miRNAs in PF and/or serum samples using qRT-PCR in subsequent all patients and healthy controls enrolled in this study, and evaluated the clinical significance of the 3 miRNAs as biomarkers of distinguishing between MPE and BPE.

2.2 RNA extract and quantitative RT-PCR

Total RNA was extracted from each pleural fluid supernatant and serum sample using miRVana PARISKit (Ambion, TX) according to the manufacturer’s protocol. RNA concentration was quantified using Nano Drop 1000 Spectrophotometer (NanoDrop Technologies, MA), total RNA concentration $<$ 10 ng/ml was not included in the analysis. qRT-PCR was performed using a Taqman miRNA PCR kit (Applied Biosystems, Foster, CA) following the manufacturer’s instructions. Briefly, 5 $\mu$ l (5–10 ng/ $\mu$ l) of total RNA was reverse-transcribed to cDNA using AMV reverse transcriptase (TakaRa, Dalian, China) and the stem-loop RT primers (Applied Biosystems, Foster, CA). qRT-PCR was performed on an Applied Biosystem, 7900 HT thermocycler at 90 ${{}^{\circ}}$ C for 10 min, followed by 40 cycles of 90 ${{}^{\circ}}$ C for 15 sec and 60 ${{}^{\circ}}$ C for 1 min. All reaction were run in triplicate. Since there is no consensus on the optimal housekeeping gene in plasma, the relative miRNA expression level in serum was calculated using the equation 2 ${}^{-\rm\Delta CT}$ , in which $\Delta\rm CT=$ cycle threshold (CT) of MPE, BPE or healthy control individuals – mean of CT of all healthy controls. The relative miRNA expression in pleural fluid was calculated using the equation 2 ${}^{-\rm\Delta CT}$ , $\Delta\rm CT=$ CT of MPE or BPE individuals – mean of CT of all HFAPE. The calculation method was reported by Hu et al. [19]. The fold change of miRNA value was $\log$ 2 transformed.

2.3 Statistical analysis

Results were expressed as mean $\pm$ standard deviation (SD). A Mann-Whitney $U$ -test was used to analyze differences in continuous variable between the two groups. A receiver operating characteristic (ROC) curve was plotted, and the area under the ROC curve (AUC) was calculated to evaluate the diagnostic performance of the identified miRNAs. According to the optimal cut-off values provided by ROC curve analyses, values for sensitivity, specificity, and accuracy were calculated. In each analysis, $P$ -value of $<$ 0.05 were assessed as being statistically significant. Risk scores were assigned to all patients according to a linear combination of the expression levels of the miRNAs in serum and pleural fluid, weighted according to the regression coefficient. A logistic regression model was conducted.

Figure 1.

Scatter plots showing the expression levels (log2 values) of miR-21, miR-29c and miR-182 in serum (A–C) and pleural samples (D–F), as well as the values of their pleural fluid/serum ratio (G–H) in different types of cancer, different groups of benign pleural effusion and healthy control. ADC, lung adenocarcinoma; SCC, lung squamous cell carcinoma; SCLC, small cell lung cancer; BC, breast cancer; TB, tuberculous pleurisy; HF, heart failure; HC, healthy control. Other cancer included colorectal, ovarian, gastric, renal, and pancreatic cancer.

Figure 2.

Serum levels of miR-21, miR-29c and miR-182, and their ROC curves. (A–C) The diagram at the top of each panel show serum levels of the 3 miRNAs in MPE and different groups of BPE. The diagram at the bottom of each panel show the ROC curves of miR-21, miR-29c and miR-182 in distinguishing MPE from IAPE, HF and healthy controls, respectively. MPE, malignant pleural effusion; IAPE, infection associated pleural effusion; HF, heart failure; HC, healthy control; ROC, receiver-operating characteristic.

3. Results

3.1 Expression levels of candidate miRNAs in serum and PF samples of MPE patients and BPE patients

Results showed that the levels of miR-21 and miR-182 expression in serum and PF samples were significantly up-regulated in all subgroup MPE patients as compared with all BPE patients (Fig. 1A, C, D and F). No difference was found in the serum or PF miR-21 and miR-182 expressions between lung cancer patients and other cancer patients. Serum and PF miR-29c levels were found to be markedly down-regulated in the MPE patients in comparison with the BPE patients (Fig. 1B and E). Likewise, there was no significant difference in the serum and PF miR-29c levels between MPE caused by lung cancer and MPE caused by other cancer.

Interestingly, the serum and PF miR-21 levels were elevated in patients with pneumonia compared with patients with TB and heart failure, but they were still significantly lower in patients with pneumonia than in patients with MPE caused by various cancer (Fig. 1A and D). In addition, miR-29c levels in serum and PF were markedly higher in patient with TBP than in patients with pneumonia and heart failure. Nevertheless, serum and PF miR-29c levels in all subgroup BPE patients were markedly higher than those in all MPE patients (Fig. 1B and E). In addition, we noted that the levels of miR-21 and miR-182 in the patients with MPE were higher in PF than in serum. Hence PF/serum (P/S) ratios of the 2 miRNAs were evaluated. Because the levels of serum and PF miR-29c have declined below zero in varying degrees, their P/S ratio can not be evaluated. We found that the P/S ratios of miR-21 and miR-182 were all markedly elevated in the MPE patients in compared to the BPE patients (Fig. 1G–H).

3.2 Diagnostic performance of miR-21, miR-29c and miR-182 for distinguishing MPE from BPE

To assess the diagnostic performance of the 3 miRNAs for discriminating MPE from BPE, the AUC of each miRNA in serum and PF was calculated by the ROC analysis. TBP and parapneumonia-PE were combined into one group which defined as infection-associated PE (IAPE). The individual AUC values of the serum miR-21, miR-29c and miR-182 ranges between 0.767–0.808 in distinguishing MPE from IAPE. These AUC values went up to 0.819–0.845 for discriminating MPE from HFAPE, and 0.825–0.868 for discriminating MPE from healthy controls (Fig. 2A–C, Tables 2 and 3, and Supplementary Table 2). When the optimal cut-off value of each miRNA was determined based on ROC curve analyses, the 3 serum miRNAs yielded sensitivity of 66.7–80.5%, specificity of $\sim$ 80.5%, and accuracy of 72.4–79.0%, respectively (Tables 2 and 3 and Supplementary Table 2).

Table 2
Diagnostic performance of three miRNAs in serum and pleural fluid and their combined use in distinguishing between MPE and IAPE

	Parameter	AUC (95% CI)	Sensitivity	Specificity	Accuracy
Serum	miR-21	0.767 (0.703–0.831)	66.7%	80.5%	72.4%
	miR-29c	0.776 (0.714–0.839)	68.3%	80.5%	73.3%
	miR-182	0.808 (0.750–0.866)	69.1%	80.5%	73.8%
	Three miRNAs combined	0.832 (0.779–0.885)	76.4%	82.8%	79.0%
PF	miR-21	0.805 (0.747–0.863)	73.2%	79.3%	75.7%
	miR-29c	0.795 (0.736–0.854)	72.4%	78.2%	74.8%
	miR-182	0.844 (0.790–0.897)	75.6%	80.5%	77.6%
	Three miRNAs combined	0.890 (0.846–0.933)	82.1%	82.8%	82.4%
P/S ratio	miR-21	0.856 (0.802–0.909)	84.6%	80.5%	82.9%
	miR-182	0.868 (0.821–0.916)	86.2%	80.5%	83.3%
	Two miRNAs combined	0.933 (0.902–0.965)	87.8%	85.1%	86.7%

MPE, malignant pleural effusion; IAPE, infection-associated pleural effusion; PF, pleural fluid; P/S, pleural fluid/serum.

Table 3

Diagnostic performance of three miRNAs in serum and pleural fluid and their combined use in distinguishing between MPE and HFAPE

	Parameter	AUC (95% CI)	Sensitivity	Specificity	Accuracy
Serum	miR-21	0.819 (0.764–0.874)	69.1%	80.5%	73.8%
	miR-29c	0.815 (0.759–0.871)	69.9%	80.5%	74.3%
	miR-182	0.845 (0.793–0.897)	74.0%	80.5%	76.7%
	Three miRNAs combined	0.866 (0.818–0.914)	84.6%	83.9%	84.3%
PF	miR-21	0.824 (0.768–0.879)	74.8%	79.3%	76.7%
	miR-29c	0.838 (0.786–0.890)	76.4%	79.3%	77.6%
	miR-182	0.878 (0.832–0.925)	83.7%	80.5%	82.4%
	Three miRNAs combined	0.919 (0.822–0.956)	90.2%	86.2%	88.6%
P/S ratio	miR-21	0.895 (0.851–0.938)	84.6%	80.5%	82.9%
	miR-182	0.916 (0.879–0.953)	87.8%	80.5%	84.3%
	Two miRNAs combined	0.945 (0.915–0.975)	94.3%	88.5%	91.9%

MPE, malignant pleural effusion; HFAPE, heart failure-associated pleural effusion; PF, pleural fluid; P/S, pleural fluid/serum.

Then, we evaluated the diagnostic performance of the 3 PF miRNAs in distinguishing MPE from BPE. The ROC analyses revealed that the AUC values of the 3 PF miRNAs were 0.805–0.844 (MPE vs IAPE), and 0.824–0.878 (MPE vs HFAPE). The miR-182 showed better diagnostic performance than PF miR-29c and miR-21 (Fig. 3A–C and Tables 2 and 3), corresponding sensitivity, specificity and accuracy are shown in Tables 2 and 3. We observed that the AUC values of the 3 PF miRNAs were higher than those of the 3 serum miRNAs in discriminating MPE from IAPE or HFAPE.

Figure 3.

Pleural effusion levels of miR-21, miR-29c and miR-182, and their POC curves. (A–C) The diagram at the top of each panel show pleural fluid levels of the 3 miRNAs in MPE and different groups of BPE. The diagram at the bottom of each panel show the ROC curves of miR-21, miR-29c and miR-182 in distinguishing MPE from IAPE and HF, respectively. PE, pleural effusion; MPE, malignant pleural effusion; IAPE, infection associated pleural effusion; HF, heart failure; ROC, receiver-operating characteristic.

Finally, the capacity of P/S ratios of miR-21 and miR-182 to discriminate MPE from BPE was examined with ROC analyses. As shown in Fig. 4A and B and Tables 2 and 3, the AUCs of the P/S ratios of the 2 miRNAs were 0.856 and 0.868 in differentiating MPE from IAPE, and 0.895 and 0.916 in discriminating MPE from HFAPE. The sensitivity, specificity and accuracy of the P/S ratios of the 2 miRNA in diagnosing MPE were listed in Tables 2 and 3, which were better than those of the 2 serum and PF miRNAs. These results suggest that serum and PF miR-21, miR-29c and miR-182, as well as the P/S ratios of miR-21 and miR-182 may be useful tools for diagnosing MPE.

Figure 4.

Pleural fluid/serum ratios of miR-21 and miR-182, and their POC curves. (A–C) The diagram at the top of each panel show pleural fluid/serum ratios of the 2 miRNAs in MPE and different group of BPE. The diagram at the bottom of each panel show the ROC curves of pleural fluid/serum ratio of miR-21 and miR-182 in differentiating MPE from IAPE and HF, respectively. MPE, malignant pleural effusion; IAPE, infection associated pleural effusion; HF, heart failure; P/S, pleural fluid/serum; ROC, receiver-operating characteristic.

3.3 Combination of miR-21, miR-29c and miR182 for diagnosis of MPE

To assess if an optimal miRNA panel, instead of the single individual miRNA, could be applied to the diagnosis of MPE with higher sensitivity and accuracy, we used a 3-miRNA signature by combining the 3 miRNAs using the logistic regression analysis to construct the diagnostic model of the 3 miRNA. Results showed that the AUCs for the combination of the 3 miRNA in serum and PF were 0.832 and 0.89 respectively, and the AUC for the combination of P/S ratios of miR-21 and miR-182 was 0.933, for distinguishing MPE from IAPE, which were considerably higher than AUC value of each individual miRNAs (Table 2 and Fig. 5A).

Figure 5.

ROC curve analyses of the 3-miRNA panel (combination of miR-21, miR-29c and miR-182) in serum, pleural fluid and their pleural fluid/serum ratio. (A) Distinguishing MPE patients from IAPE patients. (B) Distinguishing MPE patients from HF patients. (C) Distinguishing MPE patients from healthy control. MPE, malignant pleural effusion; IAPE, infection associated pleural effusion; HF, heart failure; P/S, pleural fluid/serum; ROC, receiver-operating characteristic.

By using the logistic regression model again, we found that the 3-miRNA signature set diagnostic performance equal well in differentiating MPE from HFAPE, the corresponding AUCs were 0.866 and 0.919 respectively in serum and PF samples. In addition, combining miR-21 and miR-182 P/S ratios achieved an AUC of 0.945 (Table 3 and Fig. 5B), which were also substantially higher than the AUC value of each individual miRNA. When we applied this logistic regression model to analyze the 3 miRNAs combination discriminating between MPE and healthy control only in serum, we obtained an AUC of 0.898 (Fig. 5C and Supplementary Table 2). The sensitivity, specificity and accuracy in discriminating MPE from IAPE, HFAPE, or healthy control by combining the 3 miRNAs of serum and PF, and combining miR-21 and miR-182 P/S ratios were shown in Tables 2 and 3 and Supplementary Table 2, all of them were superior to those of each individual miRNA.

Considering that PF cytology is one of the gold standard for diagnosis of MPE, it is worthwhile to assess the potential clinical applications of the 3 PF miRNAs when combined with PF cytological examination. The sensitivity of initial cytological examination for MPE diagnosis in the present study was 41.5% indicating that only of 51/123 MPE patients were initially examined as positive cytology in PF samples. When cytology in combination with PF miR-21, miR-29c, or miR-182, the sensitivity increased from 41.5% for cytology alone to 80.4%, 79.7% and 82.9%, respectively (Supplementary Table 3). Combination of cytology and the 3-miRNA signature further raised the sensitivity to 91.1% (112/123). This results indicate that 61 of 72 MPE samples missed by initially cytological examination were rescued by the 3-miRNA signature in PF.

4. Discussion

Although the diagnosis of MPE is critical, making correct diagnosis is still challenging. Cytological examination of pleural fluid obtained through thoracetesis is a standard and non-invasive method for the diagnosis of MPE. However, false-negative results of cytological examination in PF are serious problem. Being able to establish a diagnosis of MPE in a non-invasive way through tumor markers is of great investigational interest. Consequently, there have been a number of studies on the utility tumor marker measurement in PE [8, 9, 20, 21, 22].

In the present study, we found that a 3-miRNA panel consisting of miR-21, miR-29c and miR-182 can efficiently differentiating MPE patients from BPE patients. We demonstrate that the combination of the 3 miRNAs in either serum or PF exhibited better diagnostic performance in compared to each individual miRNA, as evidenced by increased AUC values, sensitivity and accuracy (Tables 2 and 3). Furthermore, compared to serum and PF samples, the combination of the P/S ratios of miR-21 and miR-182 showed the best diagnostic performance in differentiating MPE from BPE. Meantime, our results indicate that the 3 PF miRNA contributed the additional information to PF cytology alone. Frequently, the 3 miRNAs were positive in the PF of patients exhibiting malignant cytology. However, a striking finding was the demonstration of the 3 miRNAs positivity in 61 of 72 initially cytological negative MPE.

In the present study, we noted that the levels of serum and PF miR-21 in patients with parapneumonia PE were markedly elevated than those in patients with BPE caused by TBP or heart failure. Nevertheless, the levels of serum and PF miR-21 in MPE patients were dramatically higher than those in all BPE patients including parapneumonia PE cases. Abd-El-Fattah et al reported that the levels of serum miR-21, miR-155 and miR-197 in pneumonia patients were markedly higher than healthy controls, and could be used to distinguish pneumonia patients from healthy controls but the serum miR-21 and miR-155 levels in pneumonia patients were still lower than those in lung cancer patients [23], which are similar to our results.

Also we found that the levels of serum and PF miR-29c in patients with TBP were not only significantly elevated than those in patients with MPE, but also much higher than those of patients with parapneumonia PE and heart failure. A study showed that six serum miRNAs (miR-378, miR-483-5p, miR-22, miR-29c, miR-101, and miR-3206) could serve as distinct biosignature of pulmonary tuberculosis compared to healthy controls, and groups of pneumonia, lung cancer, and chronic obstructive pulmonary disease (COPD) patients [24]. Among six miRNAs, serum miR-29c was up-regulated in the serum of the pulmonary tuberculosis patients than those of the healthy control, pneumonia, lung cancer and COPD patients [24]. These results resemble our finding.

miR-21 functions as an oncogene in several cancer. In non-small cell lung cancer (NSCLC), it was stated that NF- $\kappa$ B influenced the expression of miR-21 through a mechanism that involves the binding of NF- $\kappa$ B to the promoter region of miR-21, leading to its up-regulation [25]. PTEN, PDCD4, HIF1 $\alpha$ and hMSH2 have been identified as the target genes of miR-21 [25, 26]. A number of studies indicated that independent of the sample type, up-regulation of miR-21 was correlated with worse prognosis of patients in tissue, plasma or serum samples of NSCLC [27]. In breast cancer patients, the expression of miR-21 appeared to be up-regulated in tumor compared with normal breast tissue [28]. Moreover, the expression of miR-21/miR-155/miR-365 in serum samples could help oncologists to discriminate between breast cancer patients and healthy controls [29]. In gastric and colon cancer patients, miR-21 was found as up-regulation in tissue samples and the expression was associated with recurrence and poor survival in these individual [30, 31].

miR-29c plays the role as tumor suppressor in several kinds of tumor. miR-29c was shown to inhibit cell proliferation, migration and invasion by targeting VEGFA in LAD. Its expression down-regulation in tumor tissue was significantly associated with unfavorable prognosis in LAD [32]. In LSCC, the miRNA-29 family including miR-29c was significantly reduced in tumor tissue compared to normal tissue, loss of miR-29c enhanced cancer cell migration and invasion through direct regulation of oncogenic LOXL2 [33]. It was reported that the expression of miR-29c was decreased with the progression of breast cancers and was closely associated with an overall survival of patients. Overexpression of miR-29c inhibited the migration, colony formation and growth in breast cancer cells by regulating the TIMP3/STAT1/FOXD1 pathway [34]. miR-29c was found to be down-regulated in gastric cancer tissue and cell lines, and miR-29c overexpression suppressed cell proliferation, promoted cell apoptosis and arrested cell cycle at G1/G0 phage by targeting NASP [35]. Although many studies showed that miR-29c functioned as a tumor suppressor in various type of cancer and its expression was down-regulated in tumor tissue or serum of cancer patients [33, 34, 35, 36], some researchers reported that the levels of miR-29c expression were up-regulated in NSCLC tissue and/or serum samples compared to normal lung tissue or healthy control serum, and it was a member of panel of miRNAs proposed as biomarker for NSCLC early detection. For example, a panel of miRNAs (miR-29c, miR-93, and miR-429) detected in serum samples that includes also miR-29c, could serve as a potential tool for early diagnosis, as their elevated levels could discriminate between NSCLC patients and healthy controls [37]. In another study, along with serum miR-1466, miR-205, and miR-306, serum miR-29c was up-regulated and present as a member of the 4 miRNAs signature with high potential for early diagnosis of NSCLC [38]. The reason for the discrepancies between these studies are not clear, which may be explained by variance in the methods applied for miRNA analysis, the criteria for analyzed miRNA of interest, the ethnicity and/or clinicpathological features of recruited patients. It is necessary to further study miR-29c expression trend in various specimens (tissue, serum, plasma and PF) of different types of cancer patients with reasonable stratification of other relevant features.

Some studies signified that overexpression of miR-182 showed increased tumor cell growth and proliferation, highly invasive features, and tumor progression through repression of FBXW 7 and FBXW 11 in NSCLC [39], PECK in breast cancer [40], and the ST6GALNACa and PI3K/AKT pathway in colorectal cancer [41]. Some researchers demonstrated that circulating miR-182 can stand as a biomarker for tumor diagnosis, including NSCLC and breast cancer [42, 43]. A meta-analysis has revealed that the expressions of miR-183 family including miR-182 were up-regulated in many type of cancer, in which NSCLC was ranked in third place [44]. Thus it is generally believed that miR-182 acts as a oncogene in the biological process of cancer. However, other studies showed different conclusions. Zhang et al reported that overexpression of miR-182 inhibits the epithelial to mesenchymal transition and metastasis via inactivation of Met/AKT/Snail in NSCLC cells [45]. Sun et al reported that the miR-182 induced cervical cancer cell apoptosis by suppressing DNMT32 expression [46]. The reason of these conflicting results is unknown, but it may be related to the limited sample size and study population, or different design and experimental methods. This is also a subject worthy of in-depth study.

5. Conclusion

Our results show that the 3-miRNA panel is more reliable diagnostic signature than each individual miRNA in differentiating MPE from BPE. Overall, the combination of the 3 PF miRNAs are superior to the combination of the 3 serum miRNAs in diagnosing MPE. Although the positive of the 3 miRNA in serum or PF would not be used to confirm the diagnosis of MPE, the molecular analyses can serve as an important adjunct to cytological examination in MPE diagnosis.

Author contributions

Conception: Li-Rong Zhu, Rong-Xia Yuan and Jian Li.

Interpretation or analysis of data: Li-Rong Zhu, Rong-Xia Yuan, Xian-Bin Xia, Yi Wang, Yu-Min Zhu and Ling Fi.

Preparation of the manuscript: Li-Rong Zhu, Rong-Xia Yuan and Jian Li.

Revision for important intellectual content: Jian Li.

Supervision: Yi Wang.

Supplementary data

The supplementary files are available to download from https://dx-doi-org.web.bisu.edu.cn/10.3233/CBM-210090.

sj-docx-1-cbm-10.3233_CBM-210090.docx - Supplemental material

Supplemental material, sj-docx-1-cbm-10.3233_CBM-210090.docx

Footnotes

Acknowledgments

This work was supported by The Social Development Foundation of Zhenjiang (SH2013076, SH2014076).

Conflict of interest

The authors declare that they have no conflict of interest.

References

Feller-Kopman

and Light

R.W.

, Plerual disease, N Engl J Med 378 (2018), 740–751.

Bennett

and Maskell

, Management of malignant pleural effusion, Curr Opin Pulm Med 11 (2005), 296–300.

Lombardi

Zustovich

Nicoletto

M.O.

Donach

Artioli

and Pastorelli

, Diagnosis and treatment of malignant pleural effusion: A systematic literature review and new approaches, Am J Clin Oncol 33 (2013), 420–423.

Penz

Watt

K.N.

Hergott

C.A.

Rahman

N.M.

and Psallidas

, Management of malignant pleural effusion: Challenges and solutions, Cancer Manag Res 9 (2017), 229–241.

Motherby

and Bocking

, Diagnostic accuracy of effusion cytology, Diagn Cytopathol 20 (1999), 350–357.

Specter

and Polak

J.S.

, Management of malignant pleural effusions, Semin Respir Crit Care Med 29 (2008), 405–413.

Casal

R.F.

Eapen

G.A.

Morice

R.C.

and Jimenez

C.A.

, Medical thoracoscopy, Curr Opin Pulmon Med 15 (2009), 313–320.

Liang

Q.-L.

Shi

H.-Z.

Qin

X.-J.

Liang

X.-D.

Jiang

X.-D.

and Yang

H.B.

, Diagnostic accuracy of tumor markers for malignant pleural effusion: A meta-analysis, Thorax 63 (2008), 35–41.

Yang

Liu

Y.-L.

and Shi

H.Z.

, Diagnostic accuracy of combinations of tumor markers for malignant pleural effusion: A updated meta-analysis, Respiration 94 (2017), 62–69.

10.

Berindan-Neagoe

Monroig

P.D.C.

Pasculli

and Calin

G.A.

, MicroRNAome genome: A treasure for cancer diagnosis and therapy, CA Cancer J Clin 64 (2014), 311–336.

11.

Weber

J.A.

Baxter

Zhang

Huang

D.Y.

Huang

K.H.

Lee

M.J.

Galas

D.J.

and Wang

, The microRNA spectrum in 12 body fluids, Clin Chem 56 (2010), 1733–1741.

12.

Xiao

Y.-F.

Yong

Fang

Y.-H.

M.-H.

Yang

S.-H.

and Hu

C.-J.

, microRNA detection in faces, sputum, pleural effusion and urine: Novel tool for cancer screening (review), Oncol Rep 30 (2013), 535–544.

13.

Han

H.-S.

Yun

Lim

S.-N.

Han

J.-H.

Lee

K.-H.

Kim

S.T.

Kang

M.-H.

Son

S.-M.

Lee

Y.-M.

Choi

S.-Y.

Yun

S.J.

Kim

W.-J.

and Lee

O.-J.

, Down-regulation of cell-free miR-198 as a diagnostic biomarker for lung adenocarcinoma-associated malignant pleural effusion, Int J Cancer 133 (2013), 645–653.

14.

Lin

Wan

Zhou

Y.-Q.

Chen

Huang

Liu

Y.-M.

Zhang

Yang

W.-M.

Min

Q.-H.

Sun

S.-Q.

Gao

Q.-F.

and Wang

X.-Z.

, Differential miRNA expression in peural effusion derived from extracellular vesicles of patients with lung cancer, pulmonary tuberculosis, or pneumonia, Tumor Biol 37 (2016), 15865–15845.

15.

Roberts

M.E.

Neville

Berrisford

R.G.

Antunes

Ali

N.J.

and BTS Pleural Disease Guideline Group, Management of a malignant pleural effusion: British Thoracic Society Pleural Disease Guideline 2010, Thorax 65(supp/12) (2010), ii32–ii40.

16.

Villena Garrido

Cases Viedma

Fernandez Villar

de Pablo Gafas

Perez Rodriguez

Porcel Perez

J.M.

Panadero

F.R.

Martínez

C.R.

Velázquez

A.S.

and Cuadrado

L.V.

, Recommendation of diagnosis and treatment of pleural effusion. Update, Arch Bronconeumol 50 (2014), 235–249.

17.

McMurray

J.J.

Adamopoulos

Anker

S.D.

Auricchio

Böhm

Dickstein

Falk

Filippatos

Fonseca

Gomez-Sanchez

M.A.

Jaarsma

Køber

Lip

G.Y.

Maggioni

A.P.

Parkhomenko

Pieske

B.M.

Popescu

B.A.

Rønnevik

P.K.

Rutten

F.H.

Schwitter

Seferovic

Stepinska

Trindade

P.T.

Voors

A.A.

Zannad

and Zeiher

, ESC Guideline for the diagnosis and treatment of acute and chronic heart failure 2012: The task force for the diagnosis and treatment of acute and chronic heart failure 2012 of the european society of cardiology. Developed in collaboration with the Heart Failure Association (HFA) of the ESC, Eur Heart J 33 (2012), 1787–1847.

18.

Light

R.W.

Macgregor

M.I.

Luchsinger

P.C.

and Ball

W.C.

, Jr, Pleural effusion: The diagnostic separation of transudates and exudates, Ann Intern Med 77 (1972), 507–513.

19.

Chen

Zhao

Tian

Jin

Shu

Chen

Zen

Zhang

and Shen

, Serum microRNA signatures identified in a genome-wide serum microRNA expression profiling predict survival of non-small-cell lung cancer, J Clin Oncol 28 (2010), 1721–1726.

20.

Son

S.-M.

Han

H.-S.

J.Y.

Choe

K.H.

Lee

K.M.

Lee

K.H.

Kim

S.-S.

Lee

Y.-M.

Lee

H.-C.

Song

H.G.

and Lee

O.-J.

, Diagnostic performance of DC66c in lung adenocarcinoma-associated malignant pleural effusion: Comparison with CEA, CA 19-9 and CYFRA21-1, Pathology 47 (2015), 123–129.

21.

Zhai

Wang

Liu

J.-Y.

Zhou

and Shi

H.-Z.

, Diagnostic accuracy of tumor marker for malignant pleural effusion: 9 derivation and validation study, J Thorac Dis 9 (2017), 5220–5229.

22.

Wang

W.-W.

Zhou

X.-L.

Song

Y.-J.

C.-H.

Zhu

W.-G.

and Tong

Y.-S.

, Combination of long noncoding RNA MALAT1 and carcinoembryonic antigen for the diagnosis of malignant pleural effusion caused by lung cancer, OncoTargets Ther 11 (2018), 2333–2344.

23.

Abd-EI-Fattah

A.A.

Sadik

N.A.H.

Shaker

O.G.

and Aboulfilouh

M.L.

, Differential microRNAs expression in serum of patients with lung cancer, pulmonary tuberculosis, and pneumonia, Cell Biochem Biophys 67 (2013), 875–884.

24.

Zhang

Guo

Fan

Wei

Yang

Jiang

Chen

Wang

Liu

Ping

Wang

Qiu

and Li

J.-C.

, Screening and identification of six serum microRNAs as novel potential combination biomarkers for pulmonary tuberculosis diagnosis, PLOS One 8 (2013), e81076.

25.

Yan

Fang

Ying

Tan

and Gu

, Modulation of NF-kappa B/miR-21/PTEN pathway sensitize non-small cell lung cancer to cisplatin, PLOs One 10 (2015), e0121547.

26.

Jiang

L.P.

C.Y.

and Zhu

Z.T.

, Role of micreRNA-21 in radiosensitivity in non-small cell lung cancer cells by targeting PDCD4 gene, Oncotarget 8 (2017) 23675–23689.

27.

Bica-Pop

Cojocneanu-Pefric

Magdo

Raduly

Gulei

and Berindan-Neagoe

, Overview upon miR-21 in lung cancer: Focus on NSCLC, Cell Mol Life Sci 75 (2018), 3539–3551.

28.

Huang

G.L.

Zhang

X.H.

Guo

G.L.

Huang

K.T.

Yang

K.Y.

Shen

You

and Hu

X.-Q.

, Clinical significance of miR-21 expression in breast cancer: SYBR-GreenI-based real-time RT-PCR study of invasive ductal carcinoma, Oncol Rep 21 (2009), 673–679.

29.

Han

J.G.

Jiang

Y.D.

Zhang

C.H.

Yang

Y.M.

Pang

Song

Y.N.

and Zhang

G.-Q.

, A novel panel of serum miR-21/miR-155/miR-365 as a potential diagnostic biomarker for breast cancer, Ann Surg Treat Res 92 (2017), 55–66.

30.

Park

S.K.

Park

Y.S.

Ahh

J.Y.

E.J.

Kim

J.E.

Jung

Byeon

J.-S.

B.D.

Yang

D.H.

Park

S.H.

Hwang

S.W.

Jung

H.-Y.

and Myung

S.-J.

, MiR-21-5p as a predictor of recurrence in young gastric cancer patients, J Gastroenterol Hepatol 31 (2016), 1429–1435.

31.

Oue

Anami

Schetter

A.J.

Moehler

Okayama

Khan

M.A.

Bowman

E.D.

Mueller

Schad

Shimomura

Hinoi

Aoyagi

Sasaki

Okajima

Ohdan

Galle

P.R.

Yasui

and Harris

C.C.

, High miR-21 expression from FFPE tissue is associated with poor survival and response to adjuvant chemotherapy in colon cancer, Int J Cancer 134 (2014), 1926–1934.

32.

Liu

Ding

Zhou

Zhang

Shi

Song

and Wang

, MicroRNA-29c functions as a tumor suppressor by targeting VEGFA in lung adenocarcinoma, Mol Cancer 16 (2017), 50.

33.

Mizuno

Seki

Mataki

Matsushita

Kamikauaji

Kumamoto

Takagi

Goto

Nishikawa

Kato

Enokida

Nakagawa

and Inoue

, Tumor-suppressive microRNA-29 family inhibits cancer cell migration and invasion directly targeting LOXL2 in lung squamous cell carcinoma, Int J Oncol 48 (2016), 450–460.

34.

Zheng

Liu

Ren

Hoon

D.S.B.

Wang

and Du

, MiR-29c plays a suppressive role in breast cancer by targeting the TIMP3/STAT1/FOXO1 pathway, Clin Epigenetics 10 (2018), 64.

35.

Chen

Zhu

and Liu

, MicroRNA-29c inhibits cell proliferation by targeting NASP in human gastric cancer, BMC cancer 17 (2017), 109.

36.

Kuo

T.-Y.

Hsi

Yang

I.-P.

Tsai

P.-C.

Wang

J.-Y.

and Hank Juo

S.-H.

, Computational analysis of mRNA expression profiles identifies microRNA-29a/c as predictor of colorectal cancer early recurrence, PLOs One 7 (2012), e31587.

37.

Zhu

Chen

Zhang

Liu

Zhang

and Le

, Expression of miR-29c, miR93, miR429 as potential biomarkers for detection for of early stage non-cell lung cancer, PLOs One 9 (2014), e87780.

38.

Yang

Zhang

Wang

Zhang

Beer

D.G.

Yang

and Chen

, Serum microRNA signature is capable of early diagnosis for non-small cell lung cancer, Int J Biol Sci 15 (2019), 1712–1722.

39.

Chang

Liu

Y.-H.

Wamh

L.-L.

Wang

Zhao

Z.-H.

J.-F.

and Wang

S.-F.

, MiR-183 promotes cell proliferation by suppressing FBXW7 and FBXB11 in non-small cell lung cancer, Am J Transl Res 10 (2018), 1131–1142.

40.

Chiang

C.-H.

Hou

M.-F.

and Hung

W-C.

, Up-regulation of miR-182 by β-catenin in breast cancer increases tumorigenicity and invasiveness by targeting the matrixmetalloproleinase inhibitor RECK, Biochim Biophys Acta 1830 (2013), 3068–3076.

41.

Jia

Luo

Ren

Liu

Zhao

Shen

and Zhou

, MiR-182 and miR-1356 mediate the tumorigensis and invasiveness of colorectal cancer cells via targeting ST6 GALNAC2 and PI3K/AKT pathway, Dig Dis Sci 62 (2017), 3447–3459.

42.

Zhu

Zhou

Zha

Chen

J.Y.

Liu

and Zhang

, Diagnostic value of serum miR-182, miR-183, miR-210, and miR-126 levels in patients with early-stage non-small cell lung cancer, PLOs One 11 (2016), e0153046.

43.

Wang

P.-Y.

Gong

H.-T.

B.-F.

C.-L.

Wang

H.-T.

Zhou

H.-H.

X.-X.

Xie

S.-Y.

and Jiang

B.-F.

, Higher expression of circulating miR-182 as a novel biomarker for breast cancer, Oncol Lett 6 (2013), 1681–1686.

44.

Zhang

Q.H.

Sun

H.M.

Zheng

R.Z.

Y.C.

Zhang

Cheng

Tang

Z.H.

and Huang

, Mata-analysis of microRNAs-183 family expression in human cancer studies comparing cancer tissue with noncancerous tissues, Gene 527 (2013), 26–32.

45.

Zhang

Zhao

Fan

Liu

and Chen

, MiR-182 inhibits the epithelial to lung cancer cells by targeting the met gene, Mol Carncinog 57 (2018), 125–136.

46.

Sun

Huo

Song

and Zhang

, miR-182 induces cervical cancer cell apoptosis through inhibiting the expression of DNMT3α, Int J Clin Exp Pathol 8 (2015), 4755–4763.

Assessment of a panel of miRNAs in serum and pleural fluid for the differential diagnosis of malignant and benign pleural effusion

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

Table 1 Patient demographics and etiology of pleural effusion (PE)

2. Material and methods

2.1 Study subjects, specimens, and study design

2.2 RNA extract and quantitative RT-PCR

2.3 Statistical analysis

3.1 Expression levels of candidate miRNAs in serum and PF samples of MPE patients and BPE patients

3.2 Diagnostic performance of miR-21, miR-29c and miR-182 for distinguishing MPE from BPE

Table 2 Diagnostic performance of three miRNAs in serum and pleural fluid and their combined use in distinguishing between MPE and IAPE

5. Conclusion

Author contributions

Supplementary data

sj-docx-1-cbm-10.3233_CBM-210090.docx - Supplemental material

Footnotes

Acknowledgments

Conflict of interest

References

Table 1
Patient demographics and etiology of pleural effusion (PE)

Table 2
Diagnostic performance of three miRNAs in serum and pleural fluid and their combined use in distinguishing between MPE and IAPE