Abstract
Background: MicroRNAs (miRNAs) have been linked to a number of cancer types including breast cancer. The rate of brain metastases is 10-30% in patients with advanced breast cancer which is associated with poor prognosis. The potential application of miRNAs in the diagnostics and therapeutics of breast cancer with brain metastasis is an area of intense interest. In an initial effort to systematically address the differential expression of miRNAs and mRNAs in primary breast cancer which may provide clues for early detection of brain metastasis, we analyzed the consequent changes in global patterns of gene expression in Gene Expression Omnibus (GEO) data set obtained by microarray from patients with in situ carcinoma and patients with brain metastasis. Materials and Methods: The miRNA-pathway regulatory network and miRNA–mRNA regulatory network were investigated in breast cancer specimens from patients with brain metastasis to screen for significantly dysregulated miRNAs followed by prediction of their target genes and pathways by Gene Ontology (GO) analysis. Results: Functional coordination of the changes of gene expression can be modulated by individual miRNAs. Two miRNAs, hsa-miR-17-5p and hsa-miR-16-5p, were identified as having the highest associations with targeted mRNAs [such as B-cell lymphoma 2 (BCL2), small body size/mothers against decapentaplegic 3 (SMAD3) and suppressor of cytokine signaling 1 (SOCS1)] and pathways associated with epithelial–mesenchymal transitions and other processes linked with cancer metastasis (including cell cycle, adherence junctions and extracellular matrix–receptor interaction). mRNAs for two genes [HECT, UBA and WWE domain containing 1 (HUWE1) and BCL2] were found to have the highest associations with miRNAs, which were down-regulated in brain metastasis specimens of breast cancer. The change of 11 selected miRNAs was verified in The Cancer Genome Atlas (TCGA) breast cancer dataset. Up-regulation of hsa-miR-17-5p was detected in triple-negative breast cancer tissues in TCGA. Furthermore, a negative correlation of hsa-miR-17-5p with overall survival and phosphatase and tensin homolog (PTEN) and BCL2 target genes was found in TCGA breast cancer specimens. Conclusion: Our findings provide a functionally coordinated expression pattern of different families of miRNAs that may have potential to provide clinicians with a strategy to treat breast cancer with brain metastasis from a systems-rather than a single-gene perspective.
Brain metastases occur in 10-30% of patients with breast cancer and these patients had a median overall survival from 4 to 5.5 months prior to surgery and radiation (1, 2). Most patients with brain metastasis of breast cancer are diagnosed at an advanced stage due to lacking symptoms at an early stage, which limits the possible therapeutic interventions and usually leads to a poor prognosis. The death rate of patients with brain metastasis of breast cancer is about two-fold higher compared to those with bone metastasis (3). Such patients tend to have poor prognosis, with short overall survival (4). Therefore, a comprehensive overview of the molecular mechanisms involved in brain metastasis of breast cancer, which could be identified in matching primary tumors, may help find potential predictive biomarkers and drug targets for prevention and treatment of brain metastases (5).
Studies in molecular biology have increased our understanding of the pathophysiology of breast cancer with brain metastasis. Genome-wide microRNAs (miRNAs) and mRNA expression profiling by microarray-based approaches has provided important insights into the phenotypic characteristics of breast cancer with distant metastasis. A large number of genes are associated with the development and progression of breast cancer, such as the mutation of phosphatase and tensin homolog (PTEN) (6), tumor protein p53 (TP53) (5), matrix metalloproteinase 1 (MMP1) (7) and vascular endothelial growth factor (VEGF) (7). miRNAs are small (~22 nucleotides) non-coding RNAs with gene-regulatory functions which act by binding to the 3’-untranslated region of their target mRNA, resulting in either repression of translation or mRNA degradation (8). miRNAs are remarkably tissue specific and have been used to identify the tissue in which cancer of unknown primary origin arose (9). miRNAs critically regulate global mRNA expression both in physiological and pathological processes, including cancer development and metastasis (1, 10). miRNAs are differentially expressed in breast cancer and function as oncogenes, tumor-suppressive genes, and modulators of distant metastasis via regulation of their target genes (11, 12). Many miRNAs have been reported to alter proliferation and migration of breast cancer cells in vivo (13), such as miR-200, miR-9 and miR-210, have been found to regulate breast cancer cell proliferation and metastasis (14, 15). Studies have suggested that some miRNAs are only expressed in primary or metastatic sites, indicating an organ and tumor type-dependent manner of miRNA expression. A complicated feedback network between miRNAs and their target mRNAs has been suggested to underlie breast cancer initiation and metastasis (13). Therefore, it is of great importance to elucidate the roles of the miRNA–mRNA network in breast cancer progression, especially with brain metastasis, to further reveal miRNA-driven pathways.
Extensive research has focused on establishing a set of biomarkers that could be used to predict the brain metastasis-free survival in patients with breast cancer. To date, there is no report regarding the comprehensive regulatory network of miRNA–mRNA in breast cancer cases with brain metastases. Since it is hard to collect brain tumors, few studies are available which studied brain metastasis sites from breast cancer directly. The aim of the present study was to screen the miRNAs associated with pathways underlying brain metastasis in breast cancer; the study design is presented in Figure 1. In this study, we analyzed the profiles of mRNAs and miRNAs from primary breast cancer specimens from patients with brain metastasis and compared them with the primary breast cancer in situ to reveal the gene-regulatory network between miRNAs and mRNAs. Pathways regulated by these miRNAs or miRNAs, respectively, and common pathways regulated by both were found as the regulatory networks implicated in breast cancer brain metastasis. Our data may provide important information helping to elucidate the molecular mechanisms of brain metastasis from breast cancer and may provide a strategy to treat patients with brain metastasis from breast cancer from a systematic perspective.
Workflow of the study design. BC: Breast cancer; BCBM: breast cancer with brain metastasis. TNBC: triple-negative breast cancer.
Materials and Methods
Gene-expression profiles. The Gene Expression Omnibus database (GEO, http://www.ncbi.nlm.nih.gov/geo) was used for mRNA and miRNA expression profiling studies in primary breast cancer specimens from patients with brain metastasis. GEO served as a public platform providing high-throughput gene expression data. Above all, the mRNAs in primary breast cancer specimens differentially expressed between patients with brain metastasis (n=11) and in those with in situ breast cancer (n=41) were identified by GEO2R from the GSE46928 microarray dataset. The differentially expressed miRNAs between primary breast cancer with brain metastasis (n=14) and cancer in situ (n=17) were identified by GEO2R from the GSE37407 microarray dataset.
Characteristics of mRNA and miRNA expression profiling of breast cancer specimens with brain metastasis.
The following information was extracted from each identified study: GEO accession number, platform, number of cases and controls, country, time and author.
Differential expression of miRNAs and mRNAs. After GEO2R analysis, the results of mRNA and miRNA were further analyzed, and significant differences between miRNAs and mRNA was found by means of differential expression between breast cancer brain metastasis and in situ groups. The differentially expressed mRNAs were selected with a criterion of p<0.05, while differentially expressed miRNA with a criterion of p<0.05 and log fold change (FC) >0.58.
Pathway enrichment analysis and target prediction. Target pathways of the differentially expressed miRNAs were predicted by the DIANA miRPath v.3 web-based computational tool (http://snf-515788.vm.okeanos.grnet.gr/) to investigate signaling pathways related to miRNAs differentially expressed between breast cancer with brain metastasis and breast cancer in situ. The software performs enrichment analysis for miRNA targets which are correlated to each set of miRNA target genes to Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. Target pathways for differentially expressed mRNAs were predicted by the WEB GASTALT (http://www.webgestalt.org/webgestalt_2013/). The intersection by these two methods revealed the common pathways which were regulated by both miRNAs and mRNAs. All pathways showing p<0.05 were considered significantly enriched between the compared groups.
Functional annotation. Gene Ontology (GO) analysis is used to analyze the gene function (16). Functional annotation of the differentially expressed 872 genes was performed by GO classification, including three aspects of biological pathway, molecular function, and cell component. A p-Value of less than 0.05 was considered as a significant change and the pie chart was drawn for the top 10 changes with the most statistical significance.
Constructing regulatory network between miRNAs and their targets. The miRNA–target gene pairs are negatively correlated because of degradation. A regulatory network of miRNA and target genes and pathways which were differentially expressed in the primary breast cancer of patients with brain metastasis was established by Cytoscape mapping (17).
TCGA breast cancer data set analysis. The miRNA data containing 1,538 patients with breast cancer and 163 normal tissues from TCGA database were downloaded as an independent dataset for verification of the results from GEO dataset. The levels of 11 differentially expressed miRNAs obtained from GEO dataset were compared between primary breast cancer with brain metastasis and cancer in situ. In addition, miRNAs differentially expressed between triple-negative and non-triple-negative breast cancer was also analyzed by PRISM GRAPH software (GraphPad Software, Inc. La Jolla CA, USA).
Gene set enrichment analysis (GSEA). GSEA was carried-out to analyze the PTEN and BCL2 gene sets to verify their correlation with hsa-mir-17-5p. PTEN target gene set was obtained from PTEN_DN.V2_UP from Broad Institute website (http://software.broadinstitute.org/gsea/msigdb/cards/PTEN_DN.V2_UP.html). BCL2 gene set was obtained from HANN_RESISTANCE_TO _BCL2_INHIBITOR_DN from Broad Institute (http://software.broadinstitute.org/gsea/msigdb/cards/HANN_RESISTANCE_TO_BCL2_INHIBITOR_DN.html). Tests were performed by using the default settings, and the number of permutations was set at 1,000. A false discovery rate (FDR) of <0.2 was considered statistically significant.
Results
Differential expression of miRNAs and mRNAs in breast cancer brain metastasis. In this study, we compared the miRNA and mRNA expression profiles of primary breast cancer and tumors with distant metastasis in two data sets. GEO miRNA data set (GSE37407) has a total of 17 patients with primary breast carcinoma in situ and 14 with brain metastases of breast cancer. GSE46928 mRNA data set contains 41 cases of breast carcinoma in situ and 11 cases of breast cancer with brain metastases. The full list of the genes contained in these datasets is shown in Table I. Global miRNA expression profiling was performed to screen the most significantly changed miRNAs when comparing breast cancer with brain metastasis to primary breast cancer in situ. In total, 35 miRNAs, with a threshold of p<0.05 and logFC>0.58, were considered as significantly differentially expressed, 25 were up-regulated and 10 down-regulated (Figure 2). The most significantly down-regulated miRNA was hsa-miR-199a-3p (Table II). Significant up-regulation of hsa-miR-1246 was detected in breast cancer with brain metastasis (Table II).
The differences between brain metastasis and primary breast cancer in situ are not only related to their morphology and microenvironment, but also are largely associated with the difference in their transcriptional responses. Therefore, it is necessary to examine the differential expression of members of the miRNA–mRNA network in metastatic brain tumor in order to understand the miRNA–mRNA network involved in brain metastasis in a more extensive way. A total of 872 genes were identified as being differentially expressed in breast cancer with brain metastasis under the threshold of p<0.05, including 304 up-regulated and 568 down-regulated genes (Table III).
Differentially expressed miRNAs in breast tumor specimens from patients with breast cancer with brain metastasis. A: Microarray data from GSE37407 were analyzed by GEO2R to screen the most significantly changed miRNAs in breast tumor specimens from patients with brain metastasis of breast cancer as compared with those with breast cancer in situ. A total of 35 miRNAs (red dots), for which p<0.05 and the logarithm of the fold change (logFC)>0.58 were selected and considered as the most significantly differentially expressed miRNAs.
GO classification of miRNA target genes. To gain insights into the biological roles of differentially expressed miRNA target genes in breast cancer brain metastatic tumors, we performed GO classification enrichment analysis for the 872 genes which were differentially expressed in breast cancer with brain metastasis. Genes that had a nominal significance level of p<0.01 were selected and were tested against the background set of all genes with GO annotations. Firstly, it was found that these miRNA-targeted genes were significantly enriched for biological processes of cell growth and metabolism (Figure 3A). Secondly in cell component analysis, these genes were associated with cell junctions and cell parts (Figure 3B), which have been found to be involved in the migration of cancer cells, and are likely to be one of the major causes of breast cancer metastasis to the brain. Thirdly, in the analysis of molecular function, these genes mainly interact with protein binding, which indicates that protein–protein interaction, in which protein levels may also change, are also associated with breast cancer brain metastatic process (Figure 3C). These findings suggest that regulation of proteins responsible for cell junctions and cell parts during cell growth and metabolism process are associated with breast cancer brain metastasis.
List of differentially regulated miRNAs in GSE37407.
The significantly enriched functional annotation of differentially expressed miRNA target genes. Gene Ontology (GO) enrichment analysis was performed to find the differentially expressed mRNAs in order to determine the function of the mRNAs for breast cancer with brain metastasis. Biological processes (A), cellular components (B) and molecular functions (C) of the miRNA target genes. D: Histogram of the top 10 signaling pathways for miRNA/mRNAs significantly different between breast cancer specimens with brain metastasis and breast cancer in situ.
The most significantly changed top-10 target genes were screened in functional annotations. It was found that the genes with the lowest p-value were involved in metabolic function. It has been hypothesized that predisposition or adaptation of cancer cells to energy metabolism such as glucose oxidation is a key factor in breast cancer brain metastasis (18). Our findings provide evidence for this hypothesis and indicate important roles of miRNAs in energy metabolism in breast cancer brain metastatic cells.
The regulatory network of miRNA–mRNA pathway in breast cancer brain metastasis. In order to obtain genuine miRNA targets, the combination of miRNA and mRNA expression data were analyzed by KEGG analysis for Diana Tools database to predict miRNA target pathways in a total of 35 differentially expressed miRNA-controlled signaling pathways. There were six miRNAs which were not recognized by Diana Tools or their 3p/5p were not identified and these were consequently removed from the analysis. Therefore, in total, 29 differentially expressed miRNA were used and results under the threshold of top-10 minimal p-value were considered as target pathways of these miRNAs. We also performed KEGG pathway enrichment analysis for the Web Gestalt database to predict the signaling pathways mediated by 872 differentially expressed mRNAs. Hypergenometric test with p<0.05 was used as the criterion for pathway detection. In total, 120 signaling pathways related to these mRNAs were detected. In order to better demonstrate which pathways are more important in brain metastasis from breast cancer, we took the intersection of the two sets of pathways (Figure 4A and 4B). Ultimately, seven pathways obtained were considered critical in the brain metastasis of breast cancer that were jointly controlled by miRNA and mRNA (Figures 5 and 6).
List of differentially regulated mRNAs in GSE46928.
The target pathways of significantly expressed miRNAs in breast cancer with brain metastasis. A: Venn diagram showing the overlap of seven pathways between Web Gestalt-predicted target pathways of differentially expressed mRNAs (total of 120 pathways) and Diana tool-predicted target pathways of differentially expressed miRNAs (total of 10 pathways). B: Coordinates of the top seven target pathways which correspond to the differentially expressed miRNAs selected by the value of −log p. The coordinate system shows the relative significance of these target pathways with miRNAs and mRNAs, respectively.
Based on the results from Web Gestalt, the mRNAs corresponding to the seven pathways were obtained. In addition, the targets of miRNA from each of the seven pathways were predicted by the Diana Tools. We then screened the common mRNAs and their matched miRNAs from these two sets of mRNAs obtained via the above methods. Finally a miRNA–mRNA–pathway relationship was shown (Table IV). The miRNA-target gene regulatory network was constructed with the miRNA-target gene pairs by the Cytoscape software. In miRNA–mRNA regulatory network, hsa-miR-16b-5p and hsa-miR-17-5p regulate nine and eight mRNAs and had the highest connectivity (Figure 6), suggesting that these genes may play an important role in the metastasis of breast cancer. In miRNA-pathways control network, hsa-miR-16b-5p and hsa-miR-17-5p had high connectivity and regulated seven and six pathways, respectively. Among them, pathways in cancer, adherens junction, and ubiquitin-mediated proteolysis were connected to 17, 13, and 13 miRNAs respectively, with high connectivity (Figure 6).
Verification of miRNA expression levels in TCGA dataset. It is reasonable to confirm the above findings of GEO data in other independent datasets. The levels of selected 29 miRNAs differentially expressed between primary breast cancer with brain metastasis and cancer in situ were analyzed in the TCGA dataset. Results for hsa-miR-938 and hsa-miR-193a-3p were not consistent with GEO data. In addition, data for hsa-miR-199a-3p, hsa-miR-199b-3p/5p, hsa-miR-202-3p, hsa-miR-30d-5p, hsa-miR-1246, hsa-miR-29c-3p and hsa-miR-489-3p were not statistically significant (p>0.05). However, data from a total of 11 miRNAs were in line with the GEO results and showed prominent statistical significance (p<0.0001) (Table V). For example, hsa-miR-143-3p was significantly down-regulated in breast cancer tissues compared with adjacent tissues. Up-regulation of 10 selected miRNAs, such as hsa-miR-9-5p and hsa-miR-17-5p, were detected in breast cancer tissues.
Nearly half the patients with advanced triple-negative breast cancer develop brain metastases (19). The expression levels of the above selected 11 miRNAs were compared between triple-negative and non-triple negative breast cancer. The levels of hsa-miR-17-5p (p<0.0001) (Figure 7A) and hsa-miR-15b-3p (p<0.03) (data not shown) were significantly up-regulated, which indicates that these two miRNAs may be involved in the brain metastatic process of triple-negative breast cancer. In order to analyze whether the level of hsa-miR-17-5p is correlated with breast cancer prognosis, the overall survival of patients with breast cancer was evaluated according the average level of hsa-miR-17-5p. Results showed that patients with lower levels of hsa-miR-17-5p in the primary breast cancer had better survival compared with the patients with higher levels of hsa-miR-17-5p (Figure 7B). PTEN mRNA was significantly down-regulated in breast cancer brain metastases (6).
In order to confirm the target genes of hsa-miR-17-5p in the TCGA breast cancer dataset, GSEA was used to test the enrichment value of PTEN and BCL2 target genes and their relation with hsa-miR-17-5p levels. The hsa-miR-17-5p levels were negatively correlated with gene signatures of PTEN and BCL2 in TCGA breast cancer specimens (Figure 7C and D). These data are consistent with data shown in Figure 6. These findings suggest that a global changed of specific miRNAs in the primary site of breast cancer may contribute to the brain metastatic process by regulation of specific signaling pathways via targeting tumor-suppressive genes such as PTEN and BCL2.
The regulatory network between miRNAs and target pathways in breast cancer specimens with brain metastasis. The diamond nodes and square nodes represent miRNAs and pathways, respectively. The red and green colors represent relatively high and low expression, respectively. There is a total of 77 groups of relationship between each miRNA and each pathway when there is a line connecting them. Among them, hsa-miR-16-5p and hsa-miR-17-5p have the highest connectivity, being connected with seven and six pathways, respectively. The greater the number of lines emanating from a pathway, the more miRNAs interacted with it.
The regulatory network between miRNAs and target genes in breast cancer specimens with brain metastasis. The diamond nodes and square nodes represent miRNAs and target genes, respectively. The red and green colors represent relatively high and low expression, respectively. A total of 104 relationship groups were found between each miRNA and each mRNA. The larger the symbol, the more miRNAs or genes interacted with it. hsa-miR-16-5p and hsa-miR-17-5p have the highest connectivity, and are connected with nine and eight mRNAs respectively.
The levels of hsa-miR-17-5p are negatively correlated with activation of phosphatase and tensin homolog (PTEN) B-cell lymphoma 2 (BCL2) target genes in breast cancer. A: The expression levels of hsa-miR-17-5p in triple-negative breast cancer (TNBC) (n=80) and non-TNBC (n=702) in TCGA dataset. B: The influence of hsa-miR-17-5p expression on the survival rates of patients with breast cancer (n=782). Enrichment plots of gene expression signatures for PTEN (C) and BCL2 (D) according to hsa-miR-17-5p expression level by Gene set enrichment analysis (GSEA) of The Cancer Genome Atlas (TCGA) breast cancer database. The breast cancer samples were divided into high and low hsa-miR-17-5p level groups based on the average expression of hsa-miR-17-5p. The results showed that hsa-miR-17-5p level had significant negative correlation with PTEN. FDR: False discovery rate for target genes in breast cancer samples. NES: Normalized enrichment score.
Discussion
The dissemination of breast cancer cells from the primary tumor to form metastasis in the brain is a multistep process that has not yet been fully elucidated. It is difficult to obtain specimens of brain metastasis from breast cancer. Therefore, analysis for the signature of brain metastatic genes from primary tumor samples and cell lines may provide predictive value for brain metastases in primary tumor cohorts. In the present study, global profiling of miRNA expression from primary breast tumors in situ or with brain metastases of a panel of patients from GEO dataset found 35 pairs of important miRNA/mRNA clustered in the primary breast cancer which may be critical in driving metastasis of cancer cells to the brain. The differentially expressed miRNAs were screened and their function was predicted by negative correlation and interaction between mRNA and miRNA and their co-regulative pathways. In total, 22 miRNAs, 24 mRNAs and seven pathways came together into the pairing of 134 miRNA/mRNA pathways (Figures 2, 3, 4, 5 and 6). Our results show that these participated in a series of biological processes and signaling pathways including metabolic pathway, cell cycle regulation, extracellular matrix–receptor interaction and adherens junction (Figure 4).
The miRNA-pathway regulatory network was drawn according to 77 pairs of miRNAs and their target pathways, in which, hsa-miR-16b-5p and hsa-miR-17-5p regulated seven and six pathways, respectively; and pathways in cancer, ubiquitin-mediated proteolysis and adherens junction, were connected with 18, 14 and 13 miRNAs, showing higher connectivity (Figure 5).The miRNA–mRNA regulatory network was drawn based on 103 negative miRNA–mRNA correlation pairs, in which, BCL2 and HUWE1 were all associated with 11 miRNAs, as was the highest connective mRNA; and hsa-miR-16b-5p and hsa-miR-17-5p regulated nine and eight mRNAs respectively (Figure 6). For instance, in down-regulated miRNAs, hsa-miR-199a-3p showed the most significant change and this miRNA had a larger FC. hsa-miR-199a-3p was found to play a tumor-suppressive role in various types of cancer such as hepatocellular carcinoma (20), ovarian cancer (21), endometrioid adenocarcinoma (22), and breast cancer (23). Reports show that hsa-miR-199a-3p inhibited invasion and migration of breast cancer cells by suppression of mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinases (ERK) signaling (23), which further confirms the anticancer role of miR-199a-3p. A study showed that expression of hsa-miR-141-3p was elevated in the serum of patients with breast cancer and associated with shorter brain metastasis-free survival. In addition, knockdown of hsa-miR-141-3p inhibited brain metastases from breast cancer in a mouse model (24). Furthermore, hsa-miR-16b-5p and hsa-miR-17-5p were identified as the miRNAs with highest connectivity and regulated eight mRNAs. hsa-miR-17-5p was reported to promote breast cancer invasion and metastasis via targeting WNT/β-catenin (25), and its high expression was shown in triple-negative breast cancer cells, where PTEN was revealed as its target (26), which was consistent with findings in the present study.
The miRNA target genes correlated with pathways.
For all differentially expressed mRNAs, a number of mRNAs were reported to be associated with breast cancer brain metastasis; for instance, BCL2 was reported to be an important anti-apoptotic protein. It has been confirmed that the knockout of BCL2 can induce breast cancer to metastasize to lung, liver and bone (27), and BCL2 was regulated by siRNA to result in metastases from epithelial cell cancer, including gastric cancer. Furthermore, results from the same study also suggested that down-regulation of BCL2 was correlated with breast cancer metastasis (28). In addition, most of the selected miRNAs with differential expression were confirmed in the TCGA breast cancer dataset. Results from TCGA breast cancer specimens showed that hsa-miR-17-5p expression levels were significantly up-regulated in triple-negative breast cancer (Figure 7A). PTEN and BCL2 have been found as miRNA target genes involved in breast cancer tumorigenesis and metastasis (29). Overall survival analysis and GSEA indicated that patients with breast cancer with lower hsa-miR-17-5p levels had better survival compared to those with higher hsa-miR-17-5p levels (Figure 7B), which may be attributed to the inhibition of PTEN and activation of BCL2-targeted genes by hsa-miR-17-5p (Figure 7C and D). The inhibition of hsa-miR-17-5p may prevent brain metastases of breast cancer. The above findings further verified the reliability of this regulatory network.
Change of selected miRNAs in breast cancer dataset.
It has been indicated that change in multiple factors control the proliferation and metastasis of cancer cells. Understanding the detailed network involved in the molecular processes of breast cancer brain metastasis is, therefore, very critical. miRNAs present in the corresponding primary tumors may be initiators in promoting brain metastasis of breast cancer. Metastasis of cancer cells from original sites to the brain requires crossing through the blood–brain barrier. Breast cancer cells have been shown to secrete exosomes, which are able to breach the blood–brain barrier to enter the brain (30). Recently, a report indicated that hsa-miR-1246, which was up-regulated in our study, was significantly elevated in the plasma of patients with breast cancer (31) and enriched in breast cancer exosomes (32). It is possible that hsa-miR-1246 may facilitate the breach of the blood–brain barrier to induce brain metastasis of breast cancer cells.
Since original sites lack the expression of certain genes expressed only in brain, the signature of genes for brain metastasis from primary tumor samples may not be sufficient to predict brain metastases in primary tumor cohorts due to the brain microenvironment, our findings suggest important roles of these differentially expressed miRNAs in breast cancer brain metastasis which may serve as potential biomarkers or therapeutic targets.
Acknowledgements
This work was supported by Beijing Municipal Natural Science Foundation (7152014) and the National Natural Science Foundation of the People's Republic of China (no. 81372739, and 81672521)
Footnotes
↵* These Authors contributed equally to this study.
Conflicts of Interest
No conflicts of interests were disclosed.
- Received June 1, 2017.
- Revision received June 19, 2017.
- Accepted June 21, 2017.
- Copyright© 2017, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
