MicroRNA profiling in serum: Potential signatures for breast cancer diagnosis

Abstract

BACKGROUND:

Circulating microRNAs (miRNAs) prove to be potential non-invasive indicators of cancers. The purpose of this study is to profile serum miRNA expression in breast cancer (BC) patients to find potential biomarkers for BC diagnosis.

METHODS:

The miRNA expression patterns of serum samples from 216 BC patients and 214 normal control subjects were compared. A four-phase validation was conducted for biomarker identification. In the screening phase, the Exiqon miRNA qPCR panel was employed to select candidates, which were further analyzed by quantitative reverse transcriptase PCR in the following training, testing, and external validation phases.

RESULTS:

A 12-miRNA (let-7b-5p, miR-106a-5p, miR-19a-3p, miR-19b-3p, miR-20a-5p, miR-223-3p, miR-25-3p, miR-425-5p, miR-451a, miR-92a-3p, miR-93-5p, and miR-16-5p) panel in serum was constructed. The diagnostic performance of the panel was assessed using ROC curve analyses. The area under the curves (AUCs) were 0.952, 0.956, 0.941 and 0.950 for the four separate phases, respectively. Additionally, the expression features of the 12 miRNAs were further explored in 32 pairs of BC tumor and para-tumor tissues, and 32 pairs of serum exosomes samples from patients and healthy subjects. miR-16-5p, miR-106a-5p, miR-25-3p, miR-425-5p, and miR-93-5p were highly overexpressed and let-7b-5p was conversely downregulated in tumor tissues. Excluding miR-20a-5p and miR-223-3p, the 10 other miRNAs were all significantly upregulated in BC serum-derived exosomes.

CONCLUSION:

A signature consisting of 12 serum miRNAs was identified and showed potential for use in non-invasive diagnosis of BC.

Keywords

Serum microRNA breast cancer biomarker qRT-PCR exosomes

1. Introduction

Table 1
Clinical characteristics of 216 BC patients and 214 NCs in the study

Characteristics	Screening phase		Training phase		Testing phase		External validation phase		Tissue samples
	BC	NC	BC	NC	BC	NC	BC	NC	BC
Number	24	24	32	32	124	122	36	36	32
Age at diagnosis
$<$ 50	10	11	12	14	47	41	14	13	13
$\geqslant$ 50	14	13	20	18	77	81	22	23	19
TNM stage
In situ	1		1		8		3		1
I	10		2		41		10		9
II	10		19		61		19		15
III	3		10		14		4		7
Grade
I	2		1		7		7		2
II	7		7		43		9		7
III	15		24		74		26		23
Subtype
Luminal A	8		9		46		14		11
Luminal B	1		2		18		2		1
HER2-enriched	8		11		28		9		10
Triple-negative	7		10		32		11		10

BC: breast cancer; NC: normal control.

Breast cancer (BC) has caused a heavy health burden for females in both developed and developing countries [1, 2]. According to the 2018 Global Cancer Statistics, BC is the most frequently diagnosed cancers and remains one of the leading causes of cancer-related deaths for females worldwide [3]. In China, it is now the most frequently diagnosed cancer among women, and has shown a consistent increase in global morbidity and mortality rates in recent decades [4]. Despite effective treatments such as surgical operation in early stages, many patients remain undiagnosed until the disease develops distal metastases [5]. Thus, in some ways, early diagnosis, rather than delayed treatment, can provide a greater benefit for BC patients. Conventional diagnostic strategies often show high rates of false positives or over-diagnosis; this may cause direct or indirect damage to patients, greatly limiting their range of applications [6, 7, 8, 9]. Commonly used tumor biomarkers like cancer antigen 15-3 also have some limitations, including low sensitivity and specificity for BC diagnosis [10]. Therefore, it is very important to discover novel non-invasive biomarkers.

MicroRNAs (miRNAs) are families of small, single-stranded RNA molecules (containing about 22 nucleotides) that regulate gene expression by binding target mRNA sequences [11]. They participate in many cancer-related processes including apoptosis, differentiation, and cell cycle regulation [12, 13, 14]. The dysregulation of specific miRNAs is reported to be majorly important in tumor development and progression [15]. Importantly, the stable presence of miRNAs in the blood enables them to be potent signatures for cancer detection, including BC [16, 17]. Different studies have reported various circulating miRNAs (e.g. miR-21, miR-145, miR-373, and miR-199a) as potential biomarkers for BC [18, 19, 20, 21]. However, these results were often inconsistent, possibly because they analyzed different subtypes and stages of the disease, or the studies had different designs and subjects. Furthermore, many studies only included a small number of cases and controls [17]. Systematic research with a larger study cohort and a more precise study method needs to be performed for the discovery of more reliable biomarkers.

Here, we conducted a four-phase experiment to profile serum miRNA expression using the miRCURY Ready-to-Use PCR, Human panel I $+$ II V1.M (Exiqon) and qRT-PCR for validation. The identified differentially expressed miRNAs were also detected in exosomes and tissues to further decipher their roles in disease progression. These findings aim to promote the convenient surveillance of BC in clinical practice.

2. Methods

2.1 Study subjects

Totally, 248 histopathologically diagnosed BC patients (all previously untreated) and 214 healthy normal controls (NC) were included in the study. All participants were recruited from the Jiangsu Provincial People Hospital during 2016 and 2017. All experiments were conducted under the approval of the Institutional Ethical Committee. Written informed consent was obtained from each participant.

Serum samples (216 BC vs. 214 NCs) obtained from the participants were divided into four sets. Table 1 shows the clinical features of the study subjects in each set. The distribution of patients’ age, TNM stage, grade, or molecular subtypes was unbiased across the four phases with $P>$ 0.05. Furthermore, between the BC and NC groups, there was no significant difference in age distribution in any stage ( $P>$ 0.05).

2.2 Study design

The experiment was performed in four phases: the screening, training, testing, and external validation phases. The flow chart of experiment design is shown in Fig. 1.

Figure 1.

Flow chart of the experimental design. BC: breast cancer; NC: normal control; qRT-PCR: quantitative reverse transcription PCR.

The initial screening phase was performed in two steps: first, possible candidate miRNAs were identified using the commercial plasma/serum miRNA profiling panel (Exiqon miRCURY-Ready-to-Use PCR-Human-panel-I $+$ II-V1.M, Vedbaek, Denmark) from 4 BC serum pools and 1 NC serum pool (1 pool included 10 serum samples); second, the candidate miRNAs was re-confirmed by qRT-PCR (24 BC vs. 24 NCs). Exiqon miRNA qPCR panels are commonly used platforms that show promising sensitivity and linearity for miRNA quantification, especially in blood samples that usually have limited miRNA abundance [22]. In this step, samples were pooled mainly to reduce the impact of individual sample variations on the final results. The four BC pools were composed of one HER2-positive pool, one HR-positive (ER or PR-positive) pool, one triple-negative pool, and one mixed pool, to avoid possible confounding effect of varied molecular subtypes. In each pool sample, miRNA profiling was conducted based on 174 miRNAs known to be commonly enriched in serum/plasma. Using qRT-PCR, the Ct value was evaluated, and miRNAs with Ct value $>$ 37, or 5 higher than that of the negative control (No Template Control), were omitted from candidate list. A serum miRNA was considered a candidate for BC detection if it was differentially expressed in any one of the four BC pools compared to that in the NC pool (FC $>$ 1.5 or $<$ 0.67). The final candidate miRNAs were chosen based on the combined results of all four lists; thus, if the expression of a miRNA was significantly different in any BC pool samples, it would be re-analyzed using qRT-PCR in an additional 24 pairs of individual serum samples.

The training (32 BC vs. 32 NCs) and testing phases (124 BC vs. 122 NCs) were designed for precise confirmation of promising serum miRNA biomarkers. Based on the candidate list obtained in the screening stage, the expression of each candidate miRNA was validated using qRT-PCR. The external validation phase (36 BC vs. 36 NC) was designed for the final verification of the identified miRNA signatures. Moreover, 32 pairs of BC tumor and para-tumor tissue samples, as well as serum exosomes samples (32 BC vs. 32 NCs) were analyzed to explore the expression of the identified miRNAs.

2.3 Sample processing

Serum samples were separated from whole blood (5 ml) samples collected with SST Advance tubes after clotting for 30 min (BD Co., Franklin Lakes, NJ, USA), using a centrifugation process (3000 $\times$ rpm for 10 min, 4 ${{}^{\circ}}$ C) within 2 h. The supernatant serum samples were separated and kept at $-$ 80 ${}^{\circ}$ C until analysis. Tissue specimens were collected from patients undergoing surgery and stored in liquid nitrogen until use.

Exosomes samples were isolated from serum using the Exo-Quick Exosome Precipitation Solution (System Biosciences, Mountain View, CA, USA). According to the manufacturer’s protocol, 50 $\mu$ l ExoQuick exosome precipitation solution was added to 200 $\mu$ l serum samples. The obtained exosomes pellets were dissolved in 200 $\mu$ l RNase-free water.

2.4 RNA extraction

The mirVana PARIS Kit (Ambion, Austin, TX, USA) was used to extract total RNA from serum and exosome samples following the manufacturer’s protocol. For tissue specimens, Trizol (Invitrogen, Carlsbad, CA, USA) was used for RNA extraction. 5 $\mu$ l synthetic Caenorhabditis elegans miRNA cel-miR-39 (5 nM/L, RiboBio, Guangzhou, China) was added to each sample for normalization. RNA samples were finally eluted into 50–150 $\mu$ l RNase-free water and stored at $-$ 80 ${}^{\circ}$ C. The NanoDrop 2000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) was applied to assess the purity and concentration of the acquired RNA samples.

2.5 Quantitative reverse transcription PCR (qRT-PCR)

The extracted miRNAs were amplified using the Bulge-Loop miRNA qRT-PCR Primer Set (RiboBio, Guangzhou, China) on the LightCycler 480 Real-Time PCR System (Roche Diagnostics, Basel, Switzerland). The procedure of RT reaction was 42 ${}^{\circ}$ C for 60 min, and then 70 ${}^{\circ}$ C for 10 min; the procedure of PCR was 95 ${}^{\circ}$ C for 20 s, followed by 40 cycles of 95 ${}^{\circ}$ C for 10 s and 60 ${}^{\circ}$ C for 20 s, with a final extension at 70 ${}^{\circ}$ C for 10 s. The fluorescence level of SYBR Green (SYBR Premix Ex Taq II, TaKaRa Bio, Shiga, Japan) was detected to measure the concentration of PCR production. According to a previous study, the 2 ${}^{-\Delta\Delta\text{Ct}}$ method ( $\Delta$ Ct $=$ Ct ${}_{\text{miRNA}}$ $-$ 0.5 $\times$ Ct ${}_{\text{endogenous reference}}$ $-$ 0.5 $\times$ Ct ${}_{\text{cel-miR-39}}$ ) was used to determine the miRNA expression levels. For serum and exosomes, miR-1228 was chosen as endogenous reference as previously described [23, 24, 25]. RNU6B (U6) was measured for tissue normalization. If the expression of a miRNA showed a fold change (FC) of $>$ 1.5 or $<$ 0.67 in BC samples compared to that in NC samples, and the Wilcoxon test confirmed a significant $P$ -value, the identified miRNA was defined as differentially expressed serum miRNAs in BC.

2.6 Statistical analysis

Differences in miRNA expression was assessed using the Wilcoxon rank-sum test for serum and exosomes samples. The paired nonparametric test was performed to analyze paired tissue samples. The one-way analysis of variance or chi-square test were conducted to analyze the association between the identified miRNAs and the demographic and clinical features of the BC patients. A logistic regression model was then built to construct a BC-related predictive panel. A receiver-operating characteristic (ROC) curve was constructed, and the area under the ROC curve (AUC) was calculated to assess the predictive performance of the identified miRNA signature for BC diagnosis. SPSS 20.0 (SPSS Inc., Chicago, IL, USA) and GraphPad Prism 7.0 (GraphPad Software, San Diego, CA, USA) were employed for analyzing data and constructing graphs. A two-sided value of $P<$ 0.05 was defined as statistically significant.

2.7 Bioinformatics analysis

The potential targets and involved pathways of the identified miRNAs were predicted using DIANA-miRPath v3.0, a tool for the analysis of miRNA-involved pathways. Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) analyses were conducted on the basis of experimentally verified miRNA-pathway interactions from the DIANA-TarBase7.0.

3. Results

3.1 MiRNA quantification in screening phase

In the screening phase, a total of 28 dysregulated miRNAs were identified using the Exiqon miRNA quantification platform, with 22 upregulated and 6 downregulated. Then the 28 miRNAs were re-confirmed by qRT-PCR using 24 BC and 24 NC serum samples. Finally, a total of 19 miRNAs with consistent aberrant expression in BC serum samples were identified as candidate miRNAs (Supplementary Table S1).

3.2 Identification of miRNA biomarkers in serum by qRT-PCR

In the training phase, 13 (let-7b-5p, miR-106a-5p, miR-19b-3p, miR-19a-3p, miR-20a-5p, miR-223-3p, miR-25-3p, miR-425-5p, miR-451a, miR-92a-3p, miR-93-5p, miR-16-5p, and miR-24-3p) out of the 19 miRNAs were consistently upregulated in BC serum samples compared with NCs ( $P<$ 0.05; FC $>$ 1.5). Of the 13 miRNAs, only miR-24-3p failed to pass subsequent screening among larger cohorts in the testing phase. When the data of two phases was combined, all the 12 identified miRNAs (let-7b-5p, miR-106a-5p, miR-19a-3p, miR-19b-3p, miR-20a-5p, miR-223-3p, miR-25-3p, miR-425-5p, miR-451a, miR-92a-3p, miR-93-5p, and miR-16-5p) were significantly upregulated in the serum of BC patients ( $P<$ 0.05; FC $>$ 1.5) (Fig. 2). The expression of these 12 miRNAs in BC and NC serum samples is described in Table 2; miRNAs that did not pass the two phases are described in Supplementary Table S2.

Table 2
Expression levels of the 12 serum miRNAs in the training and testing phases (presented as mean $\pm$ SD; $\Delta$ CT, relative to combination of cel-miR-39 and miR-1228)

miRNA	Training phase						Testing phase						Combined
	BC		NC		FC	$P$ value	BC		NC		FC	$P$ value	FC	$P$ value
let-7b-5p	1.19	$\pm$ 0.77	2.75	$\pm$ 0.86	2.93	$<$ 0.001	$-$ 0.14	$\pm$ 1.13	1.53	$\pm$ 1.23	3.18	$<$ 0.001	3.14	$<$ 0.001
miR-106a-5p	4.34	$\pm$ 1.06	4.99	$\pm$ 1.03	1.57	0.024	1.44	$\pm$ 1.27	2.54	$\pm$ 1.22	2.15	$<$ 0.001	2.02	$<$ 0.001
miR-16-5p	$-$ 2.94	$\pm$ 1.19	$-$ 0.96	$\pm$ 2.97	3.93	0.007	$-$ 3.61	$\pm$ 1.39	$-$ 1.91	$\pm$ 1.99	3.25	$<$ 0.001	3.39	$<$ 0.001
miR-19a-3p	5.13	$\pm$ 1.14	6.30	$\pm$ 1.21	2.26	$<$ 0.001	3.24	$\pm$ 2.07	4.65	$\pm$ 1.84	2.66	$<$ 0.001	2.59	$<$ 0.001
miR-19b-3p	1.19	$\pm$ 1.88	2.75	$\pm$ 2.08	2.94	0.003	$-$ 0.08	$\pm$ 1.04	1.82	$\pm$ 1.44	3.73	$<$ 0.001	3.56	$<$ 0.001
miR-20a-5p	8.82	$\pm$ 0.83	9.42	$\pm$ 0.81	1.51	0.013	1.78	$\pm$ 0.98	2.60	$\pm$ 0.68	1.76	$<$ 0.001	1.73	$<$ 0.001
miR-223-3p	$-$ 0.47	$\pm$ 1.02	0.95	$\pm$ 1.10	2.67	$<$ 0.001	$-$ 1.73	$\pm$ 1.28	$-$ 0.36	$\pm$ 1.97	2.59	$<$ 0.001	2.61	$<$ 0.001
miR-25-3p	$-$ 0.04	$\pm$ 0.85	1.72	$\pm$ 1.03	3.39	$<$ 0.001	$-$ 0.99	$\pm$ 1.06	0.81	$\pm$ 1.40	3.47	$<$ 0.001	3.46	$<$ 0.001
miR-425-5p	4.38	$\pm$ 0.93	5.76	$\pm$ 1.65	2.60	0.001	2.09	$\pm$ 1.18	3.37	$\pm$ 1.25	2.43	$<$ 0.001	2.47	$<$ 0.001
miR-451a	3.83	$\pm$ 1.62	5.69	$\pm$ 2.55	3.63	0.003	$-$ 0.65	$\pm$ 1.21	1.12	$\pm$ 1.67	3.41	$<$ 0.001	3.49	$<$ 0.001
miR-92a-3p	0.34	$\pm$ 1.89	1.67	$\pm$ 2.20	2.51	0.022	$-$ 1.87	$\pm$ 1.10	$-$ 0.53	$\pm$ 1.44	2.52	$<$ 0.001	2.53	$<$ 0.001
miR-93-5p	1.46	$\pm$ 1.06	3.26	$\pm$ 0.83	3.49	$<$ 0.001	0.21	$\pm$ 1.17	2.05	$\pm$ 1.25	3.59	$<$ 0.001	3.58	$<$ 0.001

BC: breast cancer; NC: normal control; FC: fold change.

Figure 2.

Expression levels of twelve identified serum miRNAs in the combined training and testing phases. N: normal control; T: tumor; Horizontal line: mean with 95% CI.

Figure 3.

Receiver operative characteristic (ROC) curve analysis of the 12-miRNA panel for breast cancer (BC) detection. A: Combined training and testing phase cohorts; B: Training phase; C: Testing phase; D: External validation phase. AUC: area under the ROC curve.

Figure 4.

Expression levels of the twelve miRNAs in paired tissue samples. N: normal control; T: tumor; Horizontal line: mean with SEM.

Figure 5.

Expression levels of the twelve miRNAs in serum-derived exosomes. N: normal control; T: tumor; Horizontal line: mean with SEM.

3.3 Diagnostic value of identified miRNAs

To assess the diagnostic performance of the 12 serum miRNAs, we built ROC curves for each of the miRNAs using the combined data of training and testing phases (Supplementary Fig. S1). We then combined the 12 serum miRNAs and established a 12-miRNA panel for BC detection using logistic regression analysis. The probability of BC prediction using the panel was determined by the following formula: Logit (P) $=$ 1.919 $-$ (0.695 $\times$ S ${}_{\text{let-7b-5p}}$ ) $-$ (0.632 $\times$ S ${}_{\text{miR-106a-5p}}$ ) $-$ (0.324 $\times$ S ${}_{\text{miR-19a-3p}}$ ) $-$ (0.721 $\times$ S ${}_{\text{miR-19b-3p}}$ ) $+$ (0.387 $\times$ S ${}_{\text{miR-20a-5p}}$ ) $-$ (0.434 $\times$ S ${}_{\text{miR-223-3p}}$ ) $-$ (1.22 $\times$ S ${}_{\text{miR-25-3p}}$ ) $+$ (0.861 $\times$ S ${}_{\text{miR-425-5p}}$ ) $+$ (0.392 $\times$ S ${}_{\text{miR-451a}}$ ) $-$ (0.149 $\times$ S ${}_{\text{miR-92a-3p}}$ ) $-$ (0.73 $\times$ S ${}_{\text{miR-93-5p}}$ ) $+$ (0.506 $\times$ S ${}_{\text{miR-16-5p}}$ ). If the miRNA had higher expression levels in BC than in the corresponding controls, the S ${}_{\text{miRNA}}$ was set as 1 (otherwise as 0). The combined data of the two phases was analyzed, and the AUC of the 12-miRNA panel for BC detection remained as high as 0.941 (95% confidence interval (CI): 0.915–0.967 (Fig. 3A). The corresponding optimal cut-off value was found to be 0.56, with a sensitivity being 0.872 and a specificity being 0.893. For the training set, the 12-miRNA signature yielded an AUC of 0.952 (95% CI: 0.906–0.998 (Fig. 3B); for the testing set, the corresponding AUC was 0.956 (95% CI: 0.933–0.979 (Fig. 3C). The sensitivity and specificity were 0.844 and 0.916 for the training set, and 0.871 and 0.907 for the testing set, respectively, when 0.56 was regarded as the cut-off value. The 12-miRNA panel exactly showed higher diagnostic accuracy than any single miRNA. The detection ability of the 12 identified miRNAs was further verified in an external validation set containing 36 BC and 36 NC serum samples. As in the previous phases, the difference in expression of these 12 serum miRNAs remained the same ( $P<$ 0.05; FC $>$ 1.5) (Supplementary Table S3). In the external validation phase, the AUC of the 12-miRNA panel for BC detection remained high, at 0.950 (95% CI: 0.895–1.000) (Fig. 3D).

However, this panel may still be unsuitable for clinical use due to the high testing cost caused by the large number of tested items. Therefore, we tried to reduce the number of miRNAs to be analyzed to construct a smaller panel with a similar diagnostic ability. Through evaluation among multiple combinations, we found the combination of serum miR-25-3p, miR-451a, and miR-93-5p (the three miRNAs with the highest variations between BC and NCs) also exhibited relatively high diagnostic performance while requiring fewer miRNAs to be detected. The equation for the logistic regression model from this 3-miRNA panel was Logit (P) $=$ 0.811 $-$ (1.07 $\times$ S ${}_{\text{miR-25-3p}}$ ) $+$ (0.435 $\times$ S ${}_{\text{miR-451a}}$ ) $-$ (0.859 $\times$ S ${}_{\text{miR-93-5p}}$ ). The corresponding AUC for the 3-miRNA signature was 0.881 (95% CI: 0.844–0.918) when the training and testing sets were combined (Supplementary Fig. S2).

3.4 External validation in public databases

The diagnostic value of the identified 12-miRNA and 3-miRNA signatures was further verified in two public datasets: The Cancer Genome Atlas Breast Invasive Carcinoma (TCGA-BRCA) dataset and GSE73002 dataset (Supplementary Fig. S3). In the TCGA-BRCA dataset, the identified signatures still performed well in discriminating BC tumor tissues ( $n=$ 1091) and normal tissues ( $n=$ 100). The corresponding AUC of the 12-miRNA signature for BC detection was 0.974 (95% CI: 0.960–0.988, specificity: 94.9%, sensitivity: 96.2%). The diagnostic value of the 3-miRNA signature was also stable with the AUC being 0.929 (95% CI: 0.902–0.956, specificity: 90.4%, sensitivity: 88.2%). The GSE73002 dataset was the largest public dataset which examined the expression profiles of serum miRNAs in a total of 1280 BC patients, 93 patients with prostate cancer, and 2686 healthy volunteers [26]. In GSE73002, all the 12 identified miRNAs were consistently upregulated in BC serum samples compared to normal controls with $P<$ 0.001(data not shown). The AUCs of the 12-miRNA and 3-miRNA panels to distinguish BC patients from healthy people were 0.966 (95% CI: 0.960–0.972, specificity: 91.3%, sensitivity: 91.6%) and 0.961 (95% CI: 0.954–0.967, specificity: 90.1%, sensitivity: 90.2%), respectively. Moreover, the identified signatures could also accurately distinguish BC patients from prostate cancer patients. The AUC of the 12-miRNA panel to differentiate BC patients from prostate cancer patients was nearly 1.000 with extremely high sensitivity and specificity. The corresponding AUC of the 3-miRNA panel was 0.988 (95% CI: 0.984–0.992, specificity: 96.5%, sensitivity: 98.9%). Taken together, the diagnostic value of the identified signature was successfully verified in the large cohorts of the two public datasets.

3.5 Comparison with previously identified signatures

The diagnostic capability of the two signatures was compared with seven previously reported signatures that could similarly differentiate BC patients from healthy people using the data from TCGA-BRCA dataset [18, 21, 26, 27, 28, 29, 30]. As shown in Supplementary Table S4, the 12-miRNA signature had the highest AUC value compared with the other markers, and the performance of the 3-miRNA panel was only behind the signature proposed by Fermin et al. The two signatures identified in this study are superior to the other serum miRNA biomarkers in BC diagnosis.

3.6 Association of identified miRNAs with pathological features

The associations of the 12 miRNAs with clinical characteristics such as TNM stage and molecular subtype were further analyzed in the 216 BC patients and 214 NCs. In BC patients at different TNM stages, all the 12 serum miRNAs were consistently significantly upregulated when compared to those in NCs ( $P<$ 0.05 (Supplementary Table S5, Supplementary Fig. S4). ROC curve analysis was performed for patients in each stage to further evaluate the diagnostic performance of the panel. The 12-miRNA panel consistently exhibited good performance in distinguishing BC patients from NCs at all TNM stages, showing high AUCs of 0.959 (95% CI: 0.928–0.990), 0.926 (95% CI: 0.887–0.965), 0.938 (95% CI: 0.910–0.967) and 0.973 (95% CI: 0.951–0.994) for the in situ, stage I, II, and III subgroups, respectively (Supplementary Fig. S5). When the data were combined, the panel showed an AUC of 0.931 (95% CI: 0.896–0.965) for early-stage cases (in situ and stage I) and 0.943 (95% CI: 0.919–0.968) for advanced-stage cases (stage II and III) (Supplementary Fig. S5).

Subgroup analyses based on molecular subtypes were also conducted. The 216 BC cases were divided into four subtypes: Luminal A ( $n=$ 77), Luminal B ( $n=$ 23), HER2 overexpressed ( $n=$ 56) and triple-negative ( $n=$ 60). Nearly all the 12 identified serum miRNAs were consistently upregulated in each subgroup when compared to those in NCs ( $P<$ 0.05) (Supplementary Table S6, Supplementary Fig. S6). According to ROC curve analysis, this signature successfully differentiated BC patients with different molecular subtypes from NCs, with high AUCs of 0.927 (95% CI: 0.891–0.962) for Luminal A subtype, 0.993 (95% CI: 0.981–1.000) for Luminal B subtype, 0.941 (95% CI: 0.910–0.972) for HER2 overexpressed subtype, and 0.942 (95% CI: 0.909–0.974) for triple-negative subtype (Supplementary Fig. S7).

3.7 Tissue miRNA expression

In addition, the expression patterns of the identified miRNAs were explored in 32 pairs of BC tumor and para-tumor normal tissue samples (Fig. 4). Among the 12 miRNAs, miR-16-5p, miR-106a-5p, miR-25-3p, miR-425-5p, and miR-93-5p were significantly overexpressed in both serum and tissue samples of BC patients; however, the expression level of let-7b-5p was significantly lowered in BC tissues compared to that in para-tumor tissues ( $P<$ 0.05). No significant differences were observed for the other six miRNAs.

3.8 Serum exosome miRNA expression

We analyzed the expression of the 12 miRNAs in serum exosomes samples (32 BC vs. 32 NCs) by qRT-PCR for further exploration. All the identified miRNAs, except miR-20a-5p and miR-223-3p, were consistently upregulated in serum exosomes in BC patients compared to NCs ( $P<$ 0.05; Fig. 5), indicating the strong association of these miRNA biomarkers with BC-related biological processes.

3.9 Functional exploration using bioinformatics analysis

We conducted Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) analyses for functional exploration of the 12- and 3-miRNA signatures, and the constructed heatmaps and summary sheets are described in Supplementary Fig. S8 (12-miRNA), Supplementary Fig. S9 (3-miRNA), and Supplementary Table S7. We concluded that many of the 12 identified miRNAs were commonly involved in biological processes relating to various cancers, such as colorectal cancer (10/12 miRNAs), prostate cancer (9/12 miRNAs), and thyroid cancer (9/12 miRNAs). Several typical cancer-related pathways also showed a close association with the 12 identified miRNAs, including the p53 signaling pathway (8/12), Hippo signaling pathway (9/12), EGFR signaling pathway (11/12), cell death (11/12), and others. The common pathways (i.e. DNA damage response, signal transduction by p53 class mediator resulting in cell cycle arrest, G1/S transition of mitotic cell cycle) probably involved by the 3-miRNA signature was also summerized (Supplementary Fig. S9). Functional annotation results indicated that the 12-miRNA and 3-miRNA panels may have a critical role in tumor initiation and development, and indirectly demonstrated the reliability of the miRNA expression differences discovered in this study between BC patients and NCs. These results show potential for further large-scale verification and an investigation of the related mechanisms in future studies.

4. Discussion

BC is the leading cause of cancer deaths in females worldwide. The accurate diagnosis of BC is of critical importance to improve clinical outcomes. However, common screening and diagnostic strategies such as mammograms have several limitations [6]. Tumor-derived miRNAs in blood have been established as potent markers for cancer diagnosis and prognosis prediction [25]. Until now, serum miRNA biomarkers such as miR-21, miR-484, miR-1246, and miR-199a have been reported to be useful for BC diagnosis in various studies [17, 21, 26, 29, 31, 32, 33]. However, these studies often have conflicting results due to the use of diverse miRNA quantification platforms, study subjects, or analytical methods.

In the present study, miRNA expression in serum was profiled using qRT-PCR for 216 BC patients and 214 NCs. A total of 12 serum miRNAs (let-7b-5p, miR-106a-5p, miR-19a-3p, miR-19b-3p, miR-20a-5p, miR-223-3p, miR-25-3p, miR-425-5p, miR-451a, miR-92a-3p, miR-93-5p, and miR-16-5p) were significantly overexpressed in BC patients compared to NCs. A 12-miRNA panel with steady diagnostic performance for BC detection was further established. Moreover, we found that the 12-miRNA panel successfully discriminated BC patients in different stages or with different molecular subtypes from NCs, which further confirmed its diagnostic value. In tissues, among the 12 miRNAs, miR-16-5p, miR-106a-5p, miR-25-3p, miR-425-5p and miR-93-5p were consistently overexpressed, and let-7b-5p was conversely downregulated, in BC tissues compared to those in NCs.

MiR-106a-5p, miR-19b-3p, and miR-92a-3p are derived from the miR-106a-363 cluster. The tumor promoting function of the miR-106a-363 cluster has been demonstrated in several reports [34, 35, 36, 37]. In this experiment, we discovered that the expression level of miR-106a-5p was significantly higher in both the serum and tissue of BC patients. According to previous reports, the aberrant expression of serum miR-106a-5p has been observed in many tumors, such as prostate cancer, astrocytoma, osteosarcoma, and lung squamous cell carcinoma, but was rarely reported in BC [39, 40, 41]. Jonsdottir et al. discovered that increased miR-106a-5p and miR-25 expression in BC tissues had some associations with high proliferation, cytokeratin 5/6 positivity, and ER-negativity status, indicating a poor prognosis for patients [42]. In contrast, it was reported by Jin et al. that miR-19b-3p inhibited the proliferation of BC cells [43]. Circulating miR-19b-3p was found to be a potential biomarker for gastric cancer [44, 45]. However, there is little information about its relationship with BC and requires further investigation. MiR-92a-3p has been defined as a circulating BC marker in several researches, indicating its underlying correlation with the disease state and biological process of BC [18, 46, 47].

MiR-25-3p, along with miR-93-5p, belongs to the miR-106b-25 cluster which locates on chromosome 7. MiR-19a-3p and miR-20a-5p, identified here, belongs to the miR-17-92 cluster on chromosome 13. These two clusters are paralogs of the miR-106a-363 cluster, which play important roles in cancer-related processes [48, 49, 50, 51, 52]. In our study, miR-25-3p and miR-93-5p were upregulated in both serum and tissue samples from BC patients. We proposed that these miRNAs might be released by BC tumor cells into circulation and have a major association with BC development. Similar results, as well as their determined functions in tumor progression via different mechanisms in BC, have also been reported by several studies [53, 54, 55]. For instance, Chen et al. once reported the overexpression of miR-25-3p in triple-negative BC tissue promoted tumor proliferation by negatively regulating the tumor suppressor BTG2 [56]. Li et al. found that miR-93 can regulate cell cycle progression in BC by suppressing PTEN via the PI3K/Akt pathway [57]. Bai et al. recently reported that miR-20a-5p can promote tumor cell growth by targeting RUNX3 in triple-negative BC [58]. Serum miR-25-3p and miR-93-5p have been proposed as markers for cancer diagnosis in prostate cancer, lung cancer, and osteosarcoma, but were rarely reported in BC [38, 39, 40]. Similar results were observed for miR-20a-5p and miR-19a-3p [39, 40, 59, 60].

In this study, miR-425-5p and miR-16-5p were significantly upregulated in both serum and tissue in BC, while miR-451a and miR-223-3p was upregulated in serum but had no expressing differences in tissue. However, according to previous studies, miR-451a was downregulated in BC tumor tissues, which could promote cell growth, proliferation, migration, and invasion by weakening its tumor-suppressive function [61, 62, 63]. The dysregulation of miR-223-3p in blood circulation has been discovered in colorectal adenocarcinoma, esophageal squamous cell carcinoma, and lung cancer [64, 65, 66]. MiR-223-3p may also inhibit tumor progression in a variety of other cancer types [67, 68, 69]. However, there is scarce evidence of its role in BC [45, 60, 70, 71, 72, 73]. More research is required to decipher the exact roles of these identified miRNAs in BC.

In addition to serum and tissue samples, miRNA expression was also evaluated in serum-derived exosomes. Exosomes are membrane-derived vesicles released by various cell types that carry various molecular cargos and function in cell-to-cell communication [74, 75]. Exosomes are a major potential existing form of circulating miRNAs in addition to argonaute2 complexes [76, 77]. The dysregulation of exosomal tumor-related miRNAs may hint to their potential functions. In this study, 10 of the 12 identified miRNAs (excluding miR-20a-5p and miR-223-3p) were upregulated in serum-derived exosomes. These results confirmed our hypothesis that these miRNAs and their exosomal forms are not only potential promising biomarkers for BC detection, but also might play important roles in disease progression.

In summary, we identified a 12-miRNA signature in serum (let-7b-5p, miR-106a-5p, miR-19a-3p, miR-19b-3p, miR-20a-5p, miR-223-3p, miR-25-3p, miR-425-5p, miR-451a, miR-92a-3p, miR-93-5p, and miR-16-5p) which may serve as a promising signature for the accurate diagnosis of BC after four-phase screening. However, more research is required to further verify of our results and explore the functions of these miRNAs in BC-related biological processes. While there is still a long way to go for these miRNA signatures to be widely applied in clinical practice, this may be useful in combination with other standard detection methods.

Footnotes

Acknowledgments

This work was supported by the National Natural Science Foundation of China [Grant numbers: 81672400, 81672788, 81702364].

Conflict of interest

The authors declare that they have no conflict of interest.

Supplementary data

The supplementary files are available to download from https://dx-doi-org.web.bisu.edu.cn/10.3233/CBM-201547.

References

DeSantis

Bryan

and Jemal

, Breast cancer statistics, 2013, CA Cancer J Clin 64 (2014), 52–62.

Hong

and Dong

, The past, present and future of breast cancer research in China, Cancer Lett 351 (2014), 1–5.

Torre

L.A.

Bray

Siegel

R.L.

Ferlay

Lortet-Tieulent

and Jemal

, Global cancer statistics, 2012, CA Cancer J Clin 65 (2015), 87–108.

Fan

Strasser-Weippl

J.J.

St Louis

Finkelstein

D.M.

K.D.

Chen

W.Q.

Shao

Z.M.

and Goss

P.E.

, Breast cancer in China, Lancet Oncol 15 (2014), e279–289.

Jemal

Bray

Center

M.M.

Ferlay

Ward

and Forman

, Global cancer statistics, CA Cancer J Clin 61 (2011), 69–90.

de Gelder

Heijnsdijk

E.A.

van Ravesteyn

N.T.

Fracheboud

Draisma

and de Koning

H.J.

, Interpreting overdiagnosis estimates in population-based mammography screening, Epidemiol Rev 33 (2011), 111–121.

Feng

Huang

Fan

Wang

Cao

Wang

Xie

Wang

and Ouyang

, Efficacy of physical examination, ultrasound, and ultrasound combined with fine-needle aspiration for axilla staging of primary breast cancer, Breast Cancer Res Treat 149 (2015), 761–765.

Hubbard

R.A.

Kerlikowske

Flowers

C.I.

Yankaskas

B.C.

Zhu

and Miglioretti

D.L.

, Cumulative probability of false-positive recall or biopsy recommendation after 10 years of screening mammography: A cohort study, Ann Intern Med 155 (2011), 481–492.

Matsen

C.B.

and Neumayer

L.A.

, Breast cancer: A review for the general surgeon, JAMA Surg 148 (2013), 971–979.

10.

Duffy

M.J.

Evoy

and McDermott

E.W.

, CA 15-3: Uses and limitation as a biomarker for breast cancer, Clin Chim Acta 411 (2010), 1869–1874.

11.

Macfarlane

L.A.

and Murphy

P.R.

, MicroRNA: Biogenesis, function and role in cancer, Curr Genomics 11 (2010), 537–561.

12.

Bhattacharyya

N.P.

Das

Bucha

Das

and Choudhury

, Regulation of cell cycle associated genes by microRNA and transcription factor, Microrna 5 (2016), 180–200.

13.

Harfe

B.D.

, MicroRNAs in vertebrate development, Curr Opin Genet Dev 15 (2005), 410–415.

14.

Hashimi

S.M.

Good

D.A.

Cao

Duan

Plummer

P.N.

Mellick

A.S.

and Wei

M.Q.

, Apoptosis and microRNA aberrations in cancer, Clin Exp Pharmacol Physiol 39 (2012), 739–746.

15.

Di Leva

Garofalo

and Croce

C.M.

, MicroRNAs in cancer, Annu Rev Pathol 9 (2014), 287–314.

16.

Bertoli

Cava

and Castiglioni

, MicroRNAs: New biomarkers for diagnosis, prognosis, therapy prediction and therapeutic tools for breast cancer, Theranostics 5 (2015), 1122–1143.

17.

Pimentel

Bonilla

Ravishankar

Y.G.

Contag

Gopal

LaCour

Lee

and Niemz

, Technology in MicroRNA profiling: Circulating MicroRNAs as noninvasive cancer biomarkers in breast cancer, J Lab Autom 20 (2015), 574–588.

18.

Chan

Liaw

C.S.

S.M.

Tan

H.H.

Wong

C.Y.

Thike

A.A.

Tan

P.H.

G.H.

and Lee

A.S.

, Identification of circulating microRNA signatures for breast cancer detection, Clin Cancer Res 19 (2013), 4477–4487.

19.

Eichelser

Stuckrath

Muller

Milde-Langosch

Wikman

Pantel

and Schwarzenbach

, Increased serum levels of circulating exosomal microRNA-373 in receptor-negative breast cancer patients, Oncotarget 5 (2014), 9650–9663.

20.

Kumar

Keerthana

Pazhanimuthu

and Perumal

, Overexpression of circulating miRNA-21 and miRNA-146a in plasma samples of breast cancer patients, Indian J Biochem Biophys 50 (2013), 210–214.

21.

Zhang

Jin

Wang

Zhao

Chen

Wang

Cao

Xie

and Liang

, A circulating miRNA signature as a diagnostic biomarker for non-invasive early detection of breast cancer, Breast Cancer Res Treat 154 (2015), 423–434.

22.

Jensen

S.G.

Lamy

Rasmussen

M.H.

Ostenfeld

M.S.

Dyrskjot

Orntoft

T.F.

and Andersen

C.L.

, Evaluation of two commercial global miRNA expression profiling platforms for detection of less abundant miRNAs, BMC Genomics 12 (2011), 435.

23.

Zhou

Xia

Zhou

Huang

Zhang

Zhu

Ding

and Wang

, Circulating microRNAs from the miR-106a-363 cluster on chromosome X as novel diagnostic biomarkers for breast cancer, Breast Cancer Res Treat 170 (2018), 257–270.

24.

Livak

K.J.

and Schmittgen

T.D.

, Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method, Methods 25 (2001), 402–408.

25.

Mitchell

P.S.

Parkin

R.K.

Kroh

E.M.

Fritz

B.R.

Wyman

S.K.

Pogosova-Agadjanyan

E.L.

Peterson

Noteboom

O’Briant

K.C.

Allen

Lin

D.W.

Urban

Drescher

C.W.

Knudsen

B.S.

Stirewalt

D.L.

Gentleman

Vessella

R.L.

Nelson

P.S.

Martin

D.B.

and Tewari

, Circulating microRNAs as stable blood-based markers for cancer detection, Proc Natl Acad Sci U S A 105 (2008), 10513–10518.

26.

Shimomura

Shiino

Kawauchi

Takizawa

Sakamoto

Matsuzaki

Ono

Takeshita

Niida

Shimizu

Fujiwara

Kinoshita

Tamura

and Ochiya

, Novel combination of serum microRNA for detecting breast cancer in the early stage, Cancer Sci 107 (2016), 326–334.

27.

Mar-Aguilar

Mendoza-Ramirez

J.A.

Malagon-Santiago

Espino-Silva

P.K.

Santuario-Facio

S.K.

Ruiz-Flores

Rodriguez-Padilla

and Resendez-Perez

, Serum circulating microRNA profiling for identification of potential breast cancer biomarkers, Dis Markers 34 (2013), 163–169.

28.

Guo

Liu

Wang

Liu

Chen

Xiang

and Luo

, Identification of serum miR-1915-3p and miR-455-3p as biomarkers for breast cancer, PLoS One 13 (2018), e0200716.

29.

Mangolini

Ferracin

Zanzi

M.V.

Saccenti

Ebnaof

S.O.

Poma

V.V.

Sanz

J.M.

Passaro

Pedriali

Frassoldati

Querzoli

Sabbioni

Carcoforo

Hollingsworth

and Negrini

, Diagnostic and prognostic microRNAs in the serum of breast cancer patients measured by droplet digital PCR, Biomark Res 3 (2015), 12.

30.

Shaker

Maher

Nassar

Morcos

and Gad

, Role of microRNAs -29b-2, -155, -197 and -205 as diagnostic biomarkers in serum of breast cancer females, Gene 560 (2015), 77–82.

31.

Chen

and Wang

, MicroRNA-21 in breast cancer: Diagnostic and prognostic potential, Clin Transl Oncol 16 (2014), 225–233.

32.

Sun

Wang

Lin

Sun

and Li

, Serum microRNA-155 as a potential biomarker to track disease in breast cancer, PLoS One 7 (2012), e47003.

33.

Zearo

Kim

Zhu

Zhao

J.T.

Sidhu

S.B.

Robinson

B.G.

and Soon

, MicroRNA-484 is more highly expressed in serum of early breast cancer patients compared to healthy volunteers, BMC Cancer 14 (2014), 200.

34.

Kan

Sato

Ito

Matsumura

David

Cheng

Agarwal

Paun

B.C.

Jin

Olaru

A.V.

Selaru

F.M.

Hamilton

J.P.

Yang

Abraham

J.M.

Mori

and Meltzer

S.J.

, The miR-106b-25 polycistron, activated by genomic amplification, functions as an oncogene by suppressing p21 and Bim, Gastroenterology 136 (2009), 1689–1700.

35.

Tan

Neo

T.W.

Aung

M.O.

Wasser

Lim

S.G.

and Tan

T.M.

, Role of the miR-106b-25 microRNA cluster in hepatocellular carcinoma, Cancer Sci 100 (2009), 1234–1242.

36.

Poliseno

Salmena

Riccardi

Fornari

Song

M.S.

Hobbs

R.M.

Sportoletti

Varmeh

Egia

Fedele

Rameh

Loda

and Pandolfi

P.P.

, Identification of the miR-106b∼25 microRNA cluster as a proto-oncogenic PTEN-targeting intron that cooperates with its host gene MCM7 in transformation, Sci Signal 3 (2010), ra29.

37.

Smith

A.L.

Iwanaga

Drasin

D.J.

Micalizzi

D.S.

Vartuli

R.L.

Tan

A.C.

and Ford

H.L.

, The miR-106b-25 cluster targets Smad7, activates TGF-beta signaling, and induces EMT and tumor initiating cell characteristics downstream of Six1 in human breast cancer, Oncogene 31 (2012), 5162–5171.

38.

Cochetti

Poli

Guelfi

Boni

Egidi

M.G.

and Mearini

, Different levels of serum microRNAs in prostate cancer and benign prostatic hyperplasia: Evaluation of potential diagnostic and prognostic role, Onco Targets Ther 9 (2016), 7545–7553.

39.

Zhang

Liu

L.H.

Ouyang

Guo

H.B.

Zhang

and Xiao

, MicroRNA screening identifies circulating microRNAs as potential biomarkers for osteosarcoma, Oncol Lett 10 (2015), 1662–1668.

40.

Zhang

Shan

Wang

Zhu

Huang

Zhang

Zhou

Cheng

Shu

Zhu

and Liu

, A three-microRNA signature for lung squamous cell carcinoma diagnosis in Chinese male patients, Oncotarget 8 (2017), 86897–86907.

41.

Zhi

Shao

Wang

Deng

Xue

Wang

Zhang

Shi

Xia

Wang

Lan

and Yang

, Identification of 9 serum microRNAs as potential noninvasive biomarkers of human astrocytoma, Neuro Oncol 17 (2015), 383–391.

42.

Jonsdottir

Janssen

S.R.

Da Rosa

F.C.

Gudlaugsson

Skaland

Baak

J.P.

and Janssen

E.A.

, Validation of expression patterns for nine miRNAs in 204 lymph-node negative breast cancers, PLoS One 7 (2012), e48692.

43.

Jin

Sun

Yang

Tang

Chen

and Guan

, miR-19b-3p inhibits breast cancer cell proliferation and reverses saracatinib-resistance by regulating PI3K/Akt pathway, Arch Biochem Biophys 645 (2018), 54–60.

44.

Wang

Yang

Gong

Xiao

and Liu

, A serum exosomal microRNA panel as a potential biomarker test for gastric cancer, Biochem Biophys Res Commun 493 (2017), 1322–1328.

45.

Zhang

Song

Zhang

Zhi

Chen

Pan

Wang

Jin

and Cui

, Circulating MiR-16-5p and MiR-19b-3p as two novel potential biomarkers to indicate progression of gastric cancer, Theranostics 5 (2015), 733–745.

46.

Haakensen

V.D.

Nygaard

Greger

Aure

M.R.

Fromm

Bukholm

I.R.

Luders

Chin

S.F.

Git

Caldas

Kristensen

V.N.

Brazma

Borresen-Dale

A.L.

Hovig

and Helland

, Subtype-specific micro-RNA expression signatures in breast cancer progression, Int J Cancer 139 (2016), 1117–1128.

47.

Shin

V.Y.

Siu

J.M.

Cheuk

E.K.

and Kwong

, Circulating cell-free miRNAs as biomarker for triple-negative breast cancer, Br J Cancer 112 (2015), 1751–1759.

48.

Petrocca

Vecchione

and Croce

C.M.

, Emerging role of miR-106b-25/miR-17-92 clusters in the control of transforming growth factor beta signaling, Cancer Res 68 (2008), 8191–8194.

49.

Petrocca

Visone

Onelli

M.R.

Shah

M.H.

Nicoloso

M.S.

de Martino

Iliopoulos

Pilozzi

Liu

C.G.

Negrini

Cavazzini

Volinia

Alder

Ruco

L.P.

Baldassarre

Croce

C.M.

and Vecchione

, E2F1-regulated microRNAs impair TGFbeta-dependent cell-cycle arrest and apoptosis in gastric cancer, Cancer Cell 13 (2008), 272–286.

50.

Tan

Lim

S.G.

and Tan

T.M.

, miR-106b-25/miR-17-92 clusters: Polycistrons with oncogenic roles in hepatocellular carcinoma, World J Gastroenterol 20 (2014), 5962–5972.

51.

Ventura

Young

A.G.

Winslow

M.M.

Lintault

Meissner

Erkeland

S.J.

Newman

Bronson

R.T.

Crowley

Stone

J.R.

Jaenisch

Sharp

P.A.

and Jacks

, Targeted deletion reveals essential and overlapping functions of the miR-17 through 92 family of miRNA clusters, Cell 132 (2008), 875–886.

52.

Xiao

Srinivasan

Calado

D.P.

Patterson

H.C.

Zhang

Wang

Henderson

J.M.

Kutok

J.L.

and Rajewsky

, Lymphoproliferative disease and autoimmunity in mice with increased miR-17-92 expression in lymphocytes, Nat Immunol 9 (2008), 405–414.

53.

Caiazza

and Mallardo

, The roles of miR-25 and its targeted genes in development of human cancer, Microrna 5 (2016), 113–119.

54.

Dong

Wang

L.E.

Liu

Zhao

Tang

Chen

Dai

Wei

Zhang

and Shen

, Serum microRNA profiling and breast cancer risk: The use of miR-484/191 as endogenous controls, Carcinogenesis 33 (2012), 828–834.

55.

M.Z.

Cheng

W.C.

Chen

S.F.

Nieh

O’Connor

Liu

C.L.

Tsai

W.W.

C.J.

Martin

Lin

Y.S.

K.J.

L.F.

and Izpisua Belmonte

J.C.

, miR-25/93 mediates hypoxia-induced immunosuppression by repressing cGAS, Nat Cell Biol 19 (2017), 1286–1296.

56.

Chen

Pan

Qian

Zhou

and Liu

, MiR-25-3p promotes the proliferation of triple negative breast cancer by targeting BTG2, Mol Cancer 17 (2018), 4.

57.

Miao

Shan

Liu

Zhao

and Jia

, MiR-106b and miR-93 regulate cell progression by suppression of PTEN via PI3K/Akt pathway in breast cancer, Cell Death Dis 8 (2017), e2796.

58.

Bai

Han

Liu

Jiang

and He

, MiRNA-20a-5p promotes the growth of triple-negative breast cancer cells through targeting RUNX3, Biomed Pharmacother 103 (2018), 1482–1489.

59.

Cheng

Zhao

Tang

Zhang

Sun

Jiang

Yue

Wang

Zhang

Liu

Sun

Zhou

Qin

Zhang

Yan

Wen

and Peng

, MicroRNA-20a-5p promotes colorectal cancer invasion and metastasis by downregulating Smad4, Oncotarget 7 (2016), 45199–45213.

60.

Fan

Teng

Chen

Wang

and Xia

, Identification of serum miRNAs by nano-quantum dots microarray as diagnostic biomarkers for early detection of non-small cell lung cancer, Tumour Biol 37 (2016), 7777–7784.

61.

Chen

Wang

Y.W.

Zhu

W.J.

Liu

Yin

and Gao

, A 4-microRNA signature predicts lymph node metastasis and prognosis in breast cancer, Hum Pathol 76 (2018), 122–132.

62.

Liu

Miao

Feng

Jiang

Zhou

and Li

, miR-451a inhibited cell proliferation and enhanced tamoxifen sensitive in breast cancer via macrophage migration inhibitory factor, Biomed Res Int 2015 (2015), 207684.

63.

Ouyang

Chang

Wang

and Wang

, MicroRNA profiling implies new markers of chemoresistance of triple-negative breast cancer, PLoS One 9 (2014), e96228.

64.

Huang

Zhang

Zhu

Shan

Zhou

L.W.

Zhu

Cheng

Zhang

Chen

Zhu

Wang

and Liu

, A novel serum microRNA signature to screen esophageal squamous cell carcinoma, Cancer Med 6 (2017), 109–119.

65.

Yang

Zhou

Zhao

Lei

Chen

Shen

Chen

Dai

and Huang

, The clinical use of circulating microRNAs as non-invasive diagnostic biomarkers for lung cancers, Oncotarget 8 (2017), 90197–90214.

66.

Zheng

Yang

Zhang

Wang

Yang

and Wang

, Serum microRNA panel as biomarkers for early diagnosis of colorectal adenocarcinoma, Br J Cancer 111 (2014), 1985–1992.

67.

Ding

Shen

Nie

Pan

Yan

Zhou

and Xu

, MiR-223-3p overexpression inhibits cell proliferation and migration by regulating inflammation-associated cytokines in glioblastomas, Pathol Res Pract 214 (2018), 1330–1339.

68.

Guo

Cao

Xiao

and Chen

, MicroRNA-223-3p inhibits human bladder cancer cell migration and invasion, Tumour Biol 39 (2017), 1010428317691678.

69.

Song

Tai

Goodman

S.B.

Jiao

Maloney

W.J.

and Wang

, miR-223-3p inhibits human osteosarcoma metastasis and progression by directly targeting CDH6, Mol Ther 26 (2018), 1299–1312.

70.

Sun

Jiang

Wang

and Mu

, MicoRNA-425-5p is a potential prognostic biomarker for cervical cancer, Ann Clin Biochem 54 (2017), 127–133.

71.

Liu

Zhang

Guo

Liu

Luo

Chang

Liu

Sun

and Wang

, Let-7b-5p regulates proliferation and apoptosis in multiple myeloma by targeting IGF1R, Acta Biochim Biophys Sin (Shanghai) 46 (2014), 965–972.

72.

Yan

Y.F.

Gong

F.M.

Wang

B.S.

and Zheng

, MiR-425-5p promotes tumor progression via modulation of CYLD in gastric cancer, Eur Rev Med Pharmacol Sci 21 (2017), 2130–2136.

73.

Zhu

Huang

Zhu

Zhou

Shan

L.W.

Cheng

Zhu

Zhang

Chen

Zhu

Wang

and Liu

, A panel of microRNA signature in serum for colorectal cancer diagnosis, Oncotarget 8 (2017), 17081–17091.

74.

Simons

and Raposo

, Exosomes-vesicular carriers for intercellular communication, Curr Opin Cell Biol 21 (2009), 575–581.

75.

Mathivanan

and Simpson

R.J.

, Exosomes: Extracellular organelles important in intercellular communication, J Proteomics 73 (2010), 1907–1920.

76.

Arroyo

J.D.

Chevillet

J.R.

Kroh

E.M.

Ruf

I.K.

Pritchard

C.C.

Gibson

D.F.

Mitchell

P.S.

Bennett

C.F.

Pogosova-Agadjanyan

E.L.

Stirewalt

D.L.

Tait

J.F.

and Tewari

, Argonaute2 complexes carry a population of circulating microRNAs independent of vesicles in human plasma, Proc Natl Acad Sci U S A 108 (2011), 5003–5008.

77.

Zhang

Guo

Yao

and Mi

, Exosome and exosomal microRNA: Trafficking, sorting, and function, Genomics Proteomics Bioinformatics 13 (2015), 17–24.

MicroRNA profiling in serum: Potential signatures for breast cancer diagnosis

Abstract

BACKGROUND:

METHODS:

RESULTS:

CONCLUSION:

Keywords

1. Introduction

Table 1 Clinical characteristics of 216 BC patients and 214 NCs in the study

2.1 Study subjects

2.2 Study design

2.4 RNA extraction

2.5 Quantitative reverse transcription PCR (qRT-PCR)

2.6 Statistical analysis

2.7 Bioinformatics analysis

3. Results

3.1 MiRNA quantification in screening phase

3.2 Identification of miRNA biomarkers in serum by qRT-PCR

Table 2 Expression levels of the 12 serum miRNAs in the training and testing phases (presented as mean ± SD; Δ CT, relative to combination of cel-miR-39 and miR-1228)

3.4 External validation in public databases

3.5 Comparison with previously identified signatures

3.6 Association of identified miRNAs with pathological features

3.7 Tissue miRNA expression

3.8 Serum exosome miRNA expression

3.9 Functional exploration using bioinformatics analysis

4. Discussion

Footnotes

Acknowledgments

Conflict of interest

Supplementary data

References

Table 1
Clinical characteristics of 216 BC patients and 214 NCs in the study

Table 2
Expression levels of the 12 serum miRNAs in the training and testing phases (presented as mean $\pm$ SD; $\Delta$ CT, relative to combination of cel-miR-39 and miR-1228)