MCF10A, MCF7, ZR751 and MDA-MB-231 cells were cultivated in Leibovitz's L-15 Medium containing 10% fetal bovine serum plus 2 mM L-glutamine. The cells were split before confluence and incubated at 37°C in a humidified incubator with 5% CO₂. miR-136 mimics and antisense oligonucleotides (ASO) were purchased from Shanghai GenePharma Co., Ltd. (Shanghai, China). RASAL2 full length CDS were cloned to pCMV2-myc. Transfection of miRNA mimics and ASO was carried out using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions.

Forty TNBC samples or adjacent normal mucosa tissues were obtained from patients with TNBC. Detailed pathological and clinical data were collected for all samples including WHO tumor grade, invasion and metastasis. The diagnoses of these samples were verified by pathologists. The collection of human tissue samples was approved and supervised by the Ethics Committee of Harbin Medical University.

Female immune-deficient nude mice (strain BALB/c nu/nu; 4–5 weeks old) used in the present study were bred at the Department of Pathology, Harbin Medical University. For orthotopic injection, 1×10⁷ MDA-MB-231 cells were injected into the mammary gland fat pad of each CB-17 SCID mouse in a volume of 0.1 ml. miR-136 mimics and its negative control miRNA were injected into the tumor every 2 days from day 7 of the graft. Twenty-one days later, the tumors were collected and sectioned for immunohistochemical analysis.

Tumor tissues of xenograft mice, human colon cancer samples or adjacent normal mucosa tissues were fixed in 10% neutral buffered formalin for 24 h and then embedded in paraffin. Sections (4 µm) were cut and stained for histological examination. E-cadherin (BD Biosciences; cat. BD 610182), RASAL2 (Santa Cruz Biotechnology; cat. sc-67935) antibody was diluted in PBS with 1% (wt/vol) BSA. Images were obtained with a Nikon DP70 camera mounted on a Nikon Bx60 microscope with Cell-F imaging software (Soft Imaging System).

RNA was extracted from cells or tissue samples using the mirVana miRNA isolation kit (Ambion, Foster City, CA, USA) according to the manufacturer's instructions. Small RNA fraction (<200 nt) was separated and purified according to the procedure. cDNA was obtained using M-MLV (Promega, Madison, WI, USA) and 1 µg RNA. The relative level of miR-136 was detected by stem-loop RT-PCR with the following conditions: denaturing the DNA at 94°C for 4 min, followed by 40 cycles of amplification: 94°C for 60 sec, 58°C for 60 sec, 72°C for 60 sec for data collection. U6 snRNA was used as an endogenous control. Quantitative PCR was performed on an ABI 7500 thermocycler (Applied Biosystems) using SYBR^® Premix Ex Taq™ (perfect real-time) kits (Takara Bio, Inc., Shiga, Japan) according to the manufacturer's instructions.

Equal number of MDA-MB-231 cells transfected with miR-136 mimics and control miRNA were seeded on BioCoat™ collagen I coated 6-well tissue culture dishes (BD Biosciences; cat. 354400) respectively, and allowed to grow to confluent for 48 h. Then scratches were made using p200 pipette tips, however, floating cells were carefully washed away with fresh growth medium. The wound-healing of the scratch regions were monitored and imaged at designated time-points.

A total of 5×10⁴ MDA-MB-231 cells (in 0.2 ml RPMI-1640 with 5% FBS) were seeded into the upper part of a Transwell chamber (Corning Life Sciences, Corning, NY, USA). For migration invasion assay, the chamber was pre-coated with 1 mg/ml Matrigel (Growth Factor Reduced BD Matrigel™ Matrix) for 2 h. In the lower part of the chamber, 0.6 ml RPMI-1640 with 20% FBS was added. After incubating for 30 h, chambers were disassembled and the membranes were stained with 2% crystal violet for 10 min and placed on a glass slide. Then, cells on the bottom of the membranes were counted in 5 random visual fields under a light microscope. All assays were performed in triplicate and independently performed three times.

The human RASAL2 3′UTR harboring three miR-136 potential target-binding sequences was synthesized by Shanghai GenePharma. Luciferase constructs were made by ligating the synthesized 3′UTR as well as the seed-sequence mutated version after the lucORF in the pMIR-Report luciferase vector (Ambion). For the fluorescent reporter assay, cells were seeded in a 48-well plate the day before transfection. The cells were co-transfected with miRNA mimics or ASO as well as the controls and RASAL2-3UTR or RASAL2-3UTRmut. The cells were lyzed 48 h later and the intensity of luciferase was detected.

Western blotting was performed to determine protein expression and the GAPDH was used as the internal control. Total protein from cells were lysed by RIPA buffer and then measured by Micro BCA protein assay kit (Pierce Biotechnology). Protein (50 µg/lane) was resolved on sodium dodecyl sulfate-polyacrylamide gel followed by transferred to a nitrocellulose membrane (Life Technologies, Carlsbad, CA, USA). The nitrocellulose membrane was incubated with polyclonal rabbit anti-human RASAL2 (1:3,000) and alkaline phosphatase-conjugated goat anti-rabbit antibody, respectively. After incubation the nitrocellulose membrane was evaluated by ECL.

MDA-MB-231 cells were seeded at 4×10⁵ cells/well in 6-well culture plates. Twenty-four hours later, the attached cells were transfected with 30 µM microRNA mimics and allowed to grow further for 72 h. Then the post-treatment cells were trypsinized and re-seeded at a density of 1.5×10⁵ cells/well on 8-mm coverslips in 12-well plates. After additional 48 h, coverslips with cells were fixed in methanol, and probed with primary E-cadherin (BD Biosciences; cat. BD 610182), or vimentin (Santa Cruz Biotechnology; cat. SC-6260) antibodies in 1:100 to 1:1,000, dilution then subsequently with florescent labeled secondary antibodies. Cell nuclei were stained with DAPI. After cell staining the coverslips were mounted with FluorSave reagent (Calbiochem, Darmstadt, Germany). Cells were imaged using Zeiss Meta upright microscope under 63X oil objective.

Student's t-test was performed to analyze the significance of differences between the sample means obtained from three independent experiments. One-way ANOVA was used for multiple group comparisons. Differences were considered statistically significant at P<0.05.

qPCR analysis revealed that miR-136 transcripts were reduced in tumor tissues compared to normal tissues (Fig. 1A). Further analysis showed that the miR-136 expression were negatively correlated with the WHO grades (Fig. 1A). We also determined miR-136 levels in the immortalized mammary epithelial cells (MCF10A) and three breast cancer cell lines (MCF7, ZR751 and MDA-MB-231). The level of miR-136 is much lower in the TNBC cell lines (MDA-MB-231, ZR751 and MCF7) than control cells (Fig. 1B). However, the miR-136 expression is negatively correlated with the invasive ability of these breast cancer cell lines (Fig. 1B). These results indicated a role for miR-136 in the development and/or metastasis of TNBC.

We have shown that miR-136 expression was negatively correlated with WHO grade in TNBC, we wonder whether it was involved in the invasion of this cancer. Wound healing assay result suggested that overexpression of miR-136 significantly hampered the migration of these cells (Fig. 2A). Transwell assay also revealed that miR-136 overexpression decreased the migration ability of MDA-MB-231 cells (Fig. 2B). We used a Matrigel coated Transwell assay system in vivo to further test the function of miR-136 in cancer invasion. MDA-MB-231 cells transfected with miR-136 showed a decreased ability of invasion through the Matrigel (Fig. 2C). These results suggested that miR-136 may act as a suppressor of tumor invasion.

Epithelial-to-mesenchymal transition (EMT) is a critical step for metastatic dissemination (26) thus, we hypothesized that miR-136 may affect EMT. As shown, the epithelial marker E-cadherin fluorescence was visibly enhanced in cells transfected with miR-136 compared to those transfected with scramble miRNA (Fig. 3A). Consistently, the result of confocal imaging analysis indicated that mesenchymal marker vimentin expression was weakened by overexpression of miR-136 (Fig. 3A). Further assessment of a panel of EMT-related genes by western blot analysis showed that, miR-136 treatment induced the expression of E-cadherin and decreased the expression of vimentin, Snail and SLUG, confirmed the results of immunofluorescence assay (Fig. 3B). To further confirmed these results, we established a mouse xenograft model by injecting MDA-MB-231 cells into the mammary gland fat pads. miR-136 mimics and its negative control miRNA were injected into the tumor every 2 days after day 7 of the graft. In miR-136 overexpressed group, E-cadherin expression were significant higher than the control group (Fig. 3C), and the levels of E-cadherin were correlated with the injection dose of miR-136 mimics (Fig. 3D). These results elucidated that miR-136 is a suppressor of EMT in the TNBC cell line.

Previously we showed that miR-136 may act as tumor suppressor of TNBC metastasis, then we attempted to uncover the underlying mechanisms. RASAL2 is a newly identified cancer-promoting gene in TNBC and it drives mesenchymal invasion and metastasis (14). We investigated the regulatory sequence of miR-136 in the 3′UTR of RASAL2 mRNA. Indeed, RASAL2 harbors a binding site of miR-136 (Fig. 4A). The result of luciferase activity detection revealed miR-136 overexpression repressed, whereas the ASO elevated the luciferase activities (Fig 4B, left panel). Then we mutated the binding site of miR-136 in the 3′UTR of RASAL2 mRNA and the regulatory effect of the mimics or ASO could not be observed (Fig. 4B, right panel). To further confirm these results, the miR-136 mimics, ASO and their negative control were co-transfected into MCF10A and MDA-MB-231 cells. miR-136 mimics significantly decreased the mRNA level of RASAL2 and conversely, the ASO increased it (Fig. 4C and D). Western blot analysis confirmed these results (Fig. 4E). Conclusively, those results demonstrated that RASAL2 is a direct target of miR-136.

Our results showed that RASAL2 was negatively regulated by miR-136 which functioned as a tumor suppressor in TNBC. Thus, we inferred that RASAL2 may be a cancer-promoting gene. We determined the expression of RASAL2 in 20 samples of TNBC tissues and their normal adjacent tissues (NAT) by qPCR. The results showed that RASAL2 was unregulated in TNBC tissues and its levels were correlated with WHO grade 90 (Fig. 5A). Western blot analysis and IHC results were used as further confirmation (Fig. 5B and C). These results supported that RASAL2 may function as an oncogene in TNBC.

We identified miR-136 as a tumor suppressor of TNBC, and RASAL2 may served as its target. Whether miR-136 regulate cell invasion through RASAL2 remains unclear. To validate this hypothesis, we co-transfected miR-136 mimics and RASAL2 into MCF10A and MDA-MB-231 cells. Compared with control, miR-136 suppressed RASAL2 expression both at mRNA and protein levels (Fig. 6A and B). Moreover, RASAL2 restoration significantly elevated the migration and invasion of MDA-MB-231 cells which were inhibited by miR-136 (Fig. 6C and D), indicating that RASAL2 could rescue miR-136 mediated suppression of migration and invasion of TNBC cells. These results suggested that RASAL2 may be a functional target of miR-136 in TNBC cell invasion.