MicroRNA Expression Profiling on Paired Primary and Lymph Node Metastatic Breast Cancer Revealed Distinct microRNA Profile Associated With LNM

Breast cancer (BC) is the foremost cause of cancer-related deaths in women. BC patients are oftentimes presented with lymph node metastasis (LNM), which increases their risk of recurrence. Compelling data have recently implicated microRNAs in promoting BC metastasis. Therefore, the identification of microRNA (miRNA)-based molecular signature associated with LNM could provide an opportunity for a more personalized treatment for BC patients with high risk of LNM. In current study, we performed comprehensive miRNA profiling in matched primary breast and LNM and identified 40 miRNAs, which were differentially expressed in LNM compared to primary tumors. The expression of 14 miRNAs (Up: hsa-miR-155-5p, hsa-miR-150-5p, hsa-miR-146a-5p, hsa-miR-142-5p and down: hsa-miR-200a-3p, hsa-miR-200b-3p, hsa-miR-200c-3p, hsa-miR-205-5p, hsa-miR-210-3p, hsa-miR-214-3p, hsa-miR-141-3p, hsa-miR-127-3p, hsa-miR-125a-5p, and hsa-let-7c-5p) was subsequently validated in a second cohort of 32 breast and 32 matched LNM tumor tissues. Mechanistically, forced expression of hsa-miR-205-5p, or hsa-miR-214-3p epigenetically inhibited MDA-MB-231 cell proliferation, colony formation, and cell migration. Global gene expression profiling on MDA-MB-231 cells overexpressing hsa-miR-205-5p, or hsa-miR-214-3p in combination with in silico target prediction and ingenuity pathway analyses identified multiple bona fide targets for hsa-miR-205-5p, hsa-miR-214-3p affecting cellular proliferation and migration. Interestingly, interrogation of the expression levels of hsa-miR-205 and hsa-miR-214 in the METABRIC breast cancer dataset revealed significantly poor overall survival in patients with downregulated expression of miR-205 [HR = 0.75 (0.61–0.91)], p = 0.003 and hsa-miR-214 [HR = 0.74 (0.59–0.93) p = 0.008]. Our data unraveled the miRNA-transcriptional landscape associated with LNM and provide novel insight on the role of several miRNAs in promoting BC LNM, and suggest their potential utilization in the clinical management of BC patients.


INTRODUCTION
Despite the advances in cancer management, cancer remains the second foremost cause of death globally. Breast cancer is the most prevalent malignant disease among women, with projected 2.1 million newly diagnosed breast cancer cases and close to 626,679 associated deaths in 2018 worldwide (1). Breast cancer incidence is likely to become a significant healthcare burden with increasing population and aging worldwide, as well as in the gulf region. The age-standardized incidence rates for BC are increasing in several countries, including the Arab region where the observed BC incidence were 9·5-50 cases for every 100,000 women per year. While the reported incidence for BC in this Arab region is less than incidence observed in other geographic regions, the incidence of BC in this region is increasing (1,2). BC patients in the Arab world are usually presented with larger and more aggressive tumors and nearly a decade younger than BC patients from other regions (3). A comparative study of BC patients from Lebanon, Tunisia, and Morocco (south Mediterranean) and France (North Mediterranean) revealed tumors with a more aggressive clinical and pathological features in the south Mediterranean group, which were mostly luminal B, compared to the north BC which were predominantly luminal A molecular subtype (4).
Better understanding of breast cancer lymph node metastasis (LNM) and the identification of novel molecular signature are essential for better disease stratification and management choices. Tumor metastasis is a complex process which is initiated by the spreading, seeding, and subsequent engraftment of malignant breast cancer cells in remote sites (5-7) is the leading cause of breast cancer-related death (8). It is well-known that BC patients with LNM are more likely to exhibit greater invasive clinical presentation at the time of diagnosis than patients without lymph node metastases. BC LNM could be influenced by several factors, including transcriptional factors, chemokines and cellular receptors, and subsequent changes in cytoskeleton organization, adhesion, and directional migration (9). Studies have shown that increased levels of chemokine receptors in lymph nodes, lung, liver, breast, and bone marrow-organs are related to tumor metastasis (10).
One limitation of published reports on LN-associated miRNAs in BC is that these studies are oftentimes conducted on different patient cohorts (29). In current study, we performed comprehensive miRNA profiling on 12 primary breast and their matched LN metastatic samples, which were subsequently validated in a second cohort of 32 breast and 32 matched LN tumor tissues. Functional and bioinformatics analyses highlighted novel insight into the role of miR-205-5p and miR-214-3p in suppressing BC LN metastasis.

Patient and Tissue Collection
Formalin-Fixed Paraffin-Embedded Tissue (FFPE) from 44 lymph node (LN) metastatic and matched primary breast samples was obtained from breast cancer patients undergoing standard treatment at the King Khalid University Hospital (Riyadh, Saudi Arabia); oncology center. Primary breast tumor and their paired Lymph node metastases were selected by an experienced pathologist from achieved FFPE sample blocks. Written informed consent was not required since the study was conducted on archived FFPE tissue blocks. Patient and tumor characteristics are listed in Table 1.

Total RNA Isolation From Formalin-Fixed Paraffin-Embedded Tissue (FFPE)
Total RNA was extracted from FFPE sections using the recover all total nucleic acid isolation kit (Ambion Inc., Life Technologies, USA) according to the manufacturer's protocol with slight modifications. Briefly, 20-µm paraffin sections were cut, deparaffinized by incubation in xylene for 3 min at 50 • C, centrifugation, and then washing twice in 100% ethanol. Subsequently, samples were washed and centrifuge at room temperature, and the pellets were vacuum-dried. Further, proteins were degraded using protease enzyme followed by 2-15 min incubations at 50 and 80 • C and finally nucleic acid was purified using column before DNAse treatment. Finally, RNA was eluted in nuclease free water. The concentration and purity of total RNA was quantified using NanoDrop 2000 (Thermo Scientific, DE, USA).

miRNA and Gene Expression Profiling by Microarray
Expression profiling of mRNA and miRNA was conducted as we previously described (30,31). Initially, total RNA (including miRNA) was labeled and subsequently was hybridized to the Agilent SurePrint G3 Human GE 8 × 60 k or Human 8 × 60 k v21 miRNA microarray chip (Agilent Technologies, Palo Alto, CA, USA). Microarray experiments were performed at the Stem Cell Unit, Department of Anatomy, King Saud University College of Medicine. Raw data was then normalized and differential expression and pathway enrichment analyses were conducted using GeneSpring 13.0 software (Agilent Technologies). P-value < 0.05 and 2-fold cut-off were used. In silico target prediction was carried out using TargetScan release 7.2 database. Ingenuity Pathways Analysis (IPA; Ingenuity Systems; www.ingenuity. com/) was used for functional annotations and network analyses as we previously described (32). miRNA Transfection in MDA-MB-231-Cells To investigate the functional role of selected miRNAs in regulating breast cancer biology, MDA-MB-231 cells (0.168 × 10 6 cells/ml) were transfected with the selected miRNA precursors (pre-miR-negative control, hsa-miR-205-5p and hsa-miR-214-3p) purchased from Ambion. Cell transfection was conducted employing a reverse transfection protocol as we previously described (30,33). Briefly

Measurement of Cell Viability
The alamarBlue (Thermo Scientific, Rockford, IL, USA) assay was employed to assess cell viability as previously described (34).
In brief, MDA-MB-231 cells (4 × 10 3 ) transfected with pre-miR-205, pre-miR-214, and pre-miR-negative control cells were seeded in flat-bottom 96-well plates. Five days later, alamarBlue was added to each well at a final concentration of 10% and plates were subsequently incubated at 37 • C for additional 2 h. The BioTek Synergy II (BioTek Inc., Winooski, VT, USA) plate reader was used to measure fluorescent intensity using 540 nM and 590 nM excitation and emission wavelengths, respectively. The cell viability of transfected breast cancer cell lines (miR-205 and miR-214) compared to pre-miR-negative control cells was presented as percent viability. Each experiment was done in a minimum of triplicates and was conducted at least twice.

Colony Formation Assay
The colony forming ability of transfected MDA-MB-231 cells with miR-205-5p or hsa-miR-214-3p or pre-miR-negative control was assessed employing the clonogenic assay as we described before (31). In brief, transfected cells (4 × 10 4 ) were seeded in 2 ml of DMEM in 12-well tissue culture plate. Subsequently, cells were serially diluted (1:1-1:32). The media was replaced with fresh media twice a week. Ten days later when colonies were formed, plates were washed and subsequently were stained with Diff-Quik Staining Kit (Siemens Healthcare Diagnostics, USA).
The plates were then scanned and the number of formed colonies was counted.

Transwell Migration Assay
MDA-MB-231 cells transfected with miR-205-5p, hsa-miR-214-3p, or pre-miR-negative control were serum-starved for 24 h in 1% serum media. For migration experiments, an 8.0 µm pore polyethylene terephthalate (PET, BD Falcon, MA, USA) inserts were employed as we described before (35). In brief, inserts were placed in 12-well notch plate, and subsequently cells transfected with the indicated miRNAs (5 × 10 4 ) were resuspended in DMEM containing 1.0 % serum and were then added to the insert, while DMEM containing 10% serum was used as attractant in the lower chamber. Twelve and twenty four hours later, cotton swab was used to remove non-migrating cells, while migrated adherent cells on the lower surface of the membrane were stained using the Diff-Quik Stain. For quantitative analysis, the number of migrated cells from six random field was counted using microscope (Axio Observer-A1, Carl Zeiss, Germany).

Statistical Analysis
Statistical analyses were conducted using GraphPad Prism 8.0 software (GraphPad, San Diego, CA, USA). Two-tailed t-test was used to calculate significance. Unless stated otherwise, data are presented as mean ± standard error of the mean (S.E.M), from at least two independent experiments conducted in triplicate.

Identification of Differentially Expressed microRNAs in Paired LN Metastatic and Primary Breast Cancer Tissue
Initial profiling was conducted on 24 samples of BC primary and matched LN metastatic derived from 12 BC patients.

Forced Expression of hsa-miR-205-5p and hsa-miR-214-3p Reduces MDA-MB-231 Cell Growth, Migration and Colony Formation
Several of the identified miRNAs were previous reported to regulate BC cell biology (38). Therefore, we subsequently investigated the biological role for the less studied, hsa-miR-205-5p and hsa-miR-214-3p, miRNAs in regulating BC function. These experiments were conducted on the highly metastatic MDA-MB-231 model, which expresses those miRNAs at low levels (Supplementary Figure 1). As presented in Figure 3, forced expression of hsa-miR-205-5p and hsa-miR-214-3p exhibited remarkable inhibition of cell proliferation, colony formation and cell migration of MDA-MB-231 breast cancer cells (Figures 3A-D), hence corroborating a role for those miRNAs in driving BC metastasis.

Identification of hsa-miR-205-5p and hsa-miR-214-3p Bone Fide Gene Targets in BC Cells
In order to gain more insight into the mechanism by which those two miRNAs regulate BC cell function, global gene expression profiling was conducted on MDA-MB-231 cells over expressing hsa-miR-205-5p and hsa-miR-214-3p compared to negative control cells. Hierarchical clustering revealed different pattern of gene expression in MDA-MB-231 cells under various treatment conditions (Figure 4A). Among the experimentally downregulated genes in MDA-MB-231 cells transfected with hsa-miR-205-5p, 384 genes were predicted as hsa-miR-205-5p targets based on in silicon prediction ( Figure 4B, upper panel). Similarly, among the experimentally downregulated genes in MDA-MB-231 cells transfected with hsa-miR-214-3p, 314 genes were predicted as hsa-miR-214-3p targets based on in silicon prediction ( Figure 4B, lower panel). The expression of selected number of hsa-miR-205-5p and hsa-miR-214-3p gene targets was validated using qRT-PCR (Figures 4C,D), corroborating the microarray data.

DISCUSSION
Although regional lymph node metastasis found at the time of primary diagnosis is curable using surgery and radiotherapy, such regional metastasis confers poor prognostic factor, and is oftentimes associated with a higher risk of subsequent cancer recurrence. While the bulk of genomic and transcriptome analysis have been devoted to primary breast cancers, data available for metastatic breast cancer cells revealed the existence of similar genomic alterations in metastatic cells similar to those seen in primary tumors, in addition to the presence of de novo acquired alterations (40). Currently there are limited resources on the microRNA expression profile of primary and matched metastasis BC tissue especially from this region. In current study, we are the first to conduct global miRNA expression profiling on primary breast cancer and matched lymph node metastatic tissue from this geographic region. Through transcriptome analysis, functional and pathway analyses, our data implicated hsa-miR-205-5p and hsa-miR-214-3p in breast cancer LN metastasis, through regulation of a number of oncogenic networks. Among the 660 expressed miRNAs, only 40 miRNAs were differentially expressed in primary compared to LN metastatic BC cells, suggesting the acquisition of metastatic phenotype could be driven by a handful of miRNAs. The expression of several of the identified miRNAs was subsequently validated in a second cohort of primary BC and matched LN metastasis. Interestingly, several of the downregulated miRNAs in LN metastasis were previously linked to BC stem cells. In particular, downregulated expression of several members of the miR-200 family (hsa-miR-200a, hsa-miR-200b, and hsa-miR-200c) was validated in LN metastatic tumors. Interestingly, a number of previous studies have highlighted a role for MiR-200 family in BC biology. miR-200 members were found to regulate BMI1 expression in breast cancer tumor initiating cells and also to suppress epithelial mesenchymal transition (EMT) through inhibition of zinc-finger E-box binding homeobox (ZEB)1 and ZEB2 and subsequent modulation of BC cell migration (41)(42)(43). Additionally, miR-200c was shown to regulate transforming growth factor β-induced stress fiber formation through targeting formin homology 2 domain containing (FHOD)1, an actinregulatory proteins, and protein phosphatase, Mg 2+ /Mn 2+dependent (PPM)1F (43). Taken together, it is plausible that BC LN metastasis is driven in part through loss of miR-200 family and subsequent upregulation of BMI1 and enrichment in BC TI-Cs, upregulation of ZEB1, ZEB2, and PPM1F and subsequent acquisition of an EMT phenotype.  To provide mechanistic insight into miRNA and LN metastasis, the highly metastatic MDA-MB-231 model was used for miR-205-5p and miR-214-3p functional and mechanistic investigations. Forced expression of miR-205-5p and miR-214-3p suppressed BC cell proliferation, migration, and colony formation suggesting a plausible role for those miRNAs in regulating BC metastatic. Network and functional annotation analysis on the identified miR-205-5p and miR-214-3p gene targets revealed potential regulation of several oncogenic networks in BC cells, hence suppressing cytoskeleton organization, cell proliferation, and migration, and chemotaxis. In particular, miR-205-5p was found to regulate MYC, FOXO1, and AREG networks in BC. Concordant with mechanistic data, lower expression of hsa-miR-205-5p and hsa-miR-214-3p were associated with worse overall survival in 1,262 BC patients from the METABRIC cohort. Our data are in agreement with other reports implicating miR-205 in regulating EMT (44) and as favorable prognostic marker in colorectal (44) and head and neck cancers (45). In summary, our data have identified several miRNAs associated with LN metastasis in BC patients and provided mechanistic insight into regulation of cell proliferation, and migration by hsa-miR-205-5p and hsa-miR-214-3p and their potential involvement in BC LN metastasis for potential utilization as prognostic biomarkers and targets for therapeutic interventions.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are publicly available. This data can be found here: the NCBI Gene Expression Omnibus (GSE148847, GSE148848).

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Institutional Research Ethics Board at the King Saud University College of Medicine. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

AUTHOR CONTRIBUTIONS
RE performed experiment and participated in manuscript writing. RV and MM performed experiments. KA involved in conception and design, in patient selection and clinical interpretation, and obtained funding. AMA, NAE-A, and AAlt involved in patient selection and clinical data interpretation. AA-R involved in pathological examination. MA and AAld involved in conception and design. NMA obtained funding, conceive the study, analyzed data, and finalized manuscript.