MicroRNA-15a-5p suppresses triple-negative breast cancer cell proliferation and invasion by modulating T-cell immunoreceptor associated with immunoglobulin and ITIM domain expression

Data collection and differentially expressed gene (DEG) analysis

The expression data from the GSE22513 RNA microarray dataset, which comprises data from TNBC specimens and surrounding normal tissues, were obtained from the GEO database and normalized. Raw data were preprocessed (background correction using RMA algorithm, quantile normalization, and log2 transformation) using the limma package in R to ensure comparability across samples. DEGs were determined using “limma” in R (version 4.20) based on the following criteria: |log2 fold-change (FC)| ≥ 1.5 and false discovery rate (FDR) ≤0.05.^[29,30] The expression levels were categorized as follows based on the log2 FC values: Upregulation, 0 and downregulation, <0. Volcano plots and heat maps were produced using ggplot2 and pheatmap packages, respectively.

Protein–protein interaction (PPI) networks and enrichment analyses

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed through the clusterProfiler package (Bioconductor 3.14, Sun Yat-sen University, Guangzhou, China) in R to characterize DEGs. GO functions encompass molecular function (MF), biological process (BP), and cellular component (CC). A PPI network was constructed utilizing the STRING database (https://string-db.org/).^[31,32] The TSV-formatted dataset was loaded into Cytoscape platform for network visualization to generate graphical representations of messenger RNA (mRNA) regulatory relationships. Six candidate genes exhibiting the highest node connectivity in the PPI network were designated as central regulators for further investigation.

Hub gene association with survival

The associations between hub genes and survival were examined using Kaplan-Meier plotter. The mRNA profiles of BC specimens from the TCGA database were examined using GEPIA2 and UALCAN.

miRNA–mRNA network construction

TargetScan, miRDB, and miRWalk were used to predict potential miRNAs that can target TIGIT. The intersecting miRNAs were visualized using Venn diagrams generated with FunRich software (v3.13, http://www.funrich.org/). Cytoscape was used to develop regulatory networks for TIGIT and miRNAs.

Cells

Human BC cell lines MCF-7 [HTB-22], AU565 [CRL-2351], SKBR-3 [HTB-30], and MDA-MB-468/231 [HTB-132/26 cells] as well as MCF-10A [CRL-10317] epithelial cells were sourced from the American Type Culture Collection (Manassas, VA, USA). HEK293T (GNHu17) cells were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). Cells were identified using short tandem repeat profiling. Mycoplasma test was performed to ensure the quality of cell culture and the reliability of the experimental results. BC and HEK293T cell lines were routinely maintained in Dulbecco’s Modified Eagle Medium (DMEM) (SH30243.02, HyClone, Utah, USA) containing 10% fetal bovine serum (FBS) at 37℃ and 5% carbon dioxide. MCF-10A cells were propagated using DMEM/F12 (CM-0525, Gibco, Carlsbad, CA, USA) supplemented with 5% horse serum, thymidine, putrescine, zinc, and hypoxanthine.

RNA isolation and quantitative real-time polymerase chain reaction (qRT-PCR)

TRIzol (12183555, Invitrogen, California, USA) was employed for mRNA isolation from cells, while AxyPrep miRNA isolation kit (R401-01, Vazyme, Nanjing, China) was utilized to isolate miRNA. The isolated RNA was reverse transcribed into complementary DNA (cDNA) using HiScript III RT SuperMix for qPCR (+ gDNA wiper) and miRNA 1^st strand cDNA synthesis kit (by stem-loop) (D7168M, Beyotime, Shanghai, China). qRT-PCR analysis was performed using BeyoFast™ SYBR Green qPCR Mix (D7260, Beyotime, Shanghai, China). The reactions were configured as follows: Initial denaturation at 95°C for 3 min, followed by 40 cycles of denaturation at 95°C for 10 s and annealing/extension at 60°C for 30 s. Relative expression levels of mRNA and miRNA were analyzed quantitatively through the 2^−ΔΔCt method, with GAPDH and RNU6-1 as endogenous reference genes for normalization of mRNA and miRNA, respectively. The primer sequences are detailed in Table 1.

Cell transfection

Lentiviruses were procured from Genechem Co., Ltd. (Genechem, Shanghai, China) for the development of stably transfected cell lines. MDA-MB-231/468 cells in logarithmic growth were inoculated in six-well plates and transfected following the lentivirus transfection protocol. The multiplicity of infection (MOI) was set at 10. Stably expressing TIGIT cell populations were established through antibiotic selection with 2 μg/mL puromycin. At 72 h post-transfection, green fluorescent protein expression was analyzed under a laser confocal microscope (LSM 900, ZEISS, Oberkochen, Germany). Cells exhibiting strong fluorescence signals and stable growth were cultured for subsequent experiments. miR-15a-5p mimics and negative controls (NCs) were procured from Shanghai GenePharma Co., Ltd. (Genechem, Shanghai, China). MDA-MB-231/468 cells were transferred to six-well plates 1 day before transfection, cultured until 80% confluency, and transfected using Lipofectamine 3000 (L3000150, Invitrogen, California, USA). Transiently transfected cell lines were then constructed. Subsequent experiments were performed at designated time points.

Dual-luciferase reporter (DLR) assay

The BiBiServ database was employed to preliminarily estimate the potential docking location between miR-15a-5p and TIGIT. GenePharma generated TIGIT wild-type (WT) and mutant (Mut) constructs, along with miR-15a-5p mimics. The TIGIT-WT and TIGIT-Mut sequences were then cloned into the pmirGLO vector. The resulting pmirGLO-TIGIT-WT and pmirGLO-TIGIT-Mut constructs were transfected into HEK293T (GNHu17, Shanghai, China) cell lines together with miR-15a-5p mimics and miR-15a-5p NC using Lipofectamine 3000 (L3000150, Invitrogen, California, USA). Luciferase (luc) activity was measured 48 h post-transfection by utilizing a DLR assay system (Vazyme, Nanjing, China), with Renilla luc as the normalization control for firefly luc activity.

Western blotting (WB) assay

Following 48 h of exposure, cellular proteins were obtained from all cohorts through extraction with radioimmunoprecipitation assay (RIPA) lysis buffer comprising protease inhibitors, and their levels were measured (BCA Protein Assay Kit, P0011, Beyotime, Shanghai, China). The samples were separated by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE). The resolved proteins were electroblotted onto a polyvinylidene fluoride membrane (E801-01, Vazyme, Nanjing, China). The membrane was incubated with 5% skim milk for blocking and probed with anti-TIGIT antibodies (rabbit, 1:2000, #99567, Cell Signaling Technology, USA) overnight at 4°C. Following three washes with Tris-buffered saline-Tween 20, the membrane was incubated with secondary antibodies (1:10000, #58802, Cell Signaling Technology, USA) for 1.5 h under ambient conditions. The membrane was washed three times, and protein bands were developed using Tanon 2500 (Bio-Rad, Shanghai, China) chemiluminescence imaging system. Protein expression levels were measured with ImageJ software (version 6.0; Media Cybernetics, Shanghai, China).

Cell counting kit-8 (CCK-8) assay

Transfected cells were inoculated in 96-well microplates. At designated intervals (0, 12, 24, 36, and 48 h post-transfection), the cellular samples were administered with CCK-8 detection kit (A311-01, Vazyme, Nanjing, China) and maintained for 1 h. The absorbance of the resultant solution was quantified at 450 nm using a multi-mode microplate reader (Bio-Tek, Vermont, USA).

Colony formation

Cellular suspensions (10³ cells/well) were inoculated in six-well plates (CCP01096-B, Vazyme, Nanjing, China) and cultured for 24 h before transient transfection. After a 14-day culture period, the treated cells were immobilized with 4% paraformaldehyde (28908, Invitrogen, California, USA) for 30 min. After removing the fixative, samples were rinsed with phosphate buffer saline (PBS) 3 times and stained with 0.5% crystal violet (R40052, Invitrogen, California, USA) for 20 min. After rinsing, the samples were air-dried before imaging using a microscope (Smartzoom 5, ZEISS, Oberkochen, Germany). The images were analyzed using ImageJ software (version 6.0; Media Cybernetics, Shanghai, China).

Wound healing assay

After lentiviral transfection, 5 × 10⁵ BC cells were seeded into each well of six-well plates. A standardized wound was generated in the cell monolayer using a sterile 200 μL pipette tip (PTR02014-01, Vazyme, Nanjing, China). The monolayer was immediately imaged. The culture medium was removed by aspiration, and adherent cells were flushed two to three times with PBS (G101, Vazyme, Nanjing, China) to remove dislodged cellular debris. The scratch width at identical positions was documented after 24 h of cultivation. The rate of scratch closure was evaluated.

Transwell invasion assay

Matrigel (A1569601, Invitrogen, California, USA) in combination with serum-free medium was uniformly coated onto the base of the transwell chamber (140627, Invitrogen, California, USA). The transfected BC cells were resuspended in serum-free medium, and 150 μL of cell suspension was loaded into the upper compartment. The lower chamber was filled with 700 μL of culture medium supplemented with 10% FBS (A5670701, Invitrogen, California, USA). After 24 h of culture, the transwell chamber was detached. The Matrigel, along with the cells within the chamber, was gently wiped using PBS-soaked cotton. Invaded cells adhering to the underside of the membrane were fixed with 4% paraformaldehyde for 30 min and stained with 0.1% crystal violet solution for 10 min. Images of three to five microscopic fields were captured. The cells were counted from the images using ImageJ software.

Flow cytometry

Cells were incubated for 24 h, transfected, subjected to trypsin digestion, and pooled. The samples were incubated in cooled PBS. The resulting cellular mixture was stabilized in cooled 70% ethanol and maintained at 4°C overnight. The cells were rinsed twice with PBS, treated with 100 μL of RNase A (#4087; Cell Signaling Technology, USA) at 37°C for 30 min, and incubated with 400 μL of propidium iodide (PI) for 30 min at 4°C. Cell cycle phase distribution and apoptotic proportions were evaluated using a flow cytometer (CytoFLEX S, Beckman Coulter Co., Ltd., California, USA).

Statistical analysis

Data were analyzed with R 4.1.0 (v4.1.0, R Foundation for Statistical Computing, Vienna, Austria) and GraphPad Prism 9.0 (v9.0, San Diego, CA, USA). Intergroup comparisons were conducted using unpaired t-tests for two-group analyses and one-way analysis of variance, followed by Tukey’s post hoc testing for multi-group comparisons. Numerical results are expressed as mean ± SD. A threshold of P < 0.05 was statistically significant. Each experiment was repeated three times to verify reproducibility.

Screening of DEGs

Gene expression data from the GSE22513 dataset were normalized and transformed into a log2 scale. The filtering parameters were as follows: FDR <0.05; log2 FC ≥ 1.5; log2 FC ≤ −1.5. This study identified 139 DEGs (37 upregulated and 102 downregulated genes), which were visualized as volcano plots [Figure 1a] and heatmaps [Figure 1b] using the R package. In the volcano plot, red dots signify upregulated genes, while green dots denote those that were downregulated.

Close

Figure 1:

TIGIT levels are elevated in TNBC and modulated by miR-15a-5p based on bioinformatic analysis. (a) Differential gene expression visualization through volcano plotting, with upregulated and downregulated transcripts depicted in red and green, respectively. (b) Expression heatmap of DEGs in the GSE22513 dataset. (c-f) GO and KEGG functional enrichment analyses of DEGs. Scatter plots showing BP (c), CC (d), MF (e), and KEGG pathway (f) annotations. (g) Network analysis of core genes extracted from DEGs and PPI, with principal modules and essential genes identified through Cytoscape MCODE plugin. (h) Kaplan–Meier analysis revealed the correlation of TIGIT dysregulation with unfavorable outcomes in patients with BC. (i and j) Analysis of GEPIA2 and UALCAN datasets demonstrated that TIGIT expression in BC specimens, especially in luminal BC, HER2-positive BC, and TNBC, was upregulated when compared with that in normal tissue controls. (k) Venn diagram illustration of TIGIT-targeting microRNAs (miRNAs) predicted through integrated analysis of miRWalk, miRDB, and TargetScan databases. (l) Computational prediction of TIGIT-miR-15a-5p interaction site using BiBiServ. ns: No significant difference; ^✶P <0.05. PPI: Protein-protein interaction, FDR: False discovery rate, KEGG: Kyoto Encyclopedia of Genes and Genomes, MF: Molecular function, BP: Biological process, TIGIT: T cell immunoreceptor with Ig and ITIM domains, CC: Cellular component, GO: Gene Ontology, TNBC: Triple-negative breast cancer, BC: Breast cancer, DEGs: Differentially expressed genes.

Export to PPT

GO and KEGG analyses

The enrichment of DEGs in different GO terms was as follows: BP term: “negative Negative hydrolase activity modulation,” “negative peptidase activity modulation,” “digestion,” “negative small molecule metabolic pathway modulation,” and “digestive system process” [Figure 1c]; CC term: “Perinuclear region of cytoplasm,” “specific granule,” and “specific granule lumen” [Figure 1d]; and MF term: “Carbohydrate binding,” “dystroglycan binding,” and “aldoketo reductase NADP activity” [Figure 1e]. The DEGs were enriched in the following KEGG pathways: “Arachidonic acid metabolism,” “mineral absorption,” “long-term depression,” “salivary secretion,” and “glucagon signaling pathway” [Figure 1f].

Identification of hub genes in the PPI network

A PPI network was developed by importing 139 DEGs into the STRING database. The exported data were filtered to retain entries with a combined score of >0.4. This dataset was subsequently uploaded to Cytoscape software. The MCODE plugin was used to identify key modules for further investigation. The hub genes identified through this process included CCR6, ICOS, PRF1, TIGIT, GATA3, and IDO1 [Figure 1g].

Survival analysis

Survival analysis of the hub genes indicated that TIGIT levels were markedly linked to recurrence-free survival in individuals with BC (P < 0.05), [Figure 1h]. Analysis of the GEPIA2 dataset revealed that TIGIT expression in BC samples, including luminal BC, human epidermal growth factor receptor 2 (HER2)-positive BC, and basal-like BC (TNBC), was markedly higher than in non-cancerous samples [Figure 1i]. TIGIT exhibited a similar expression pattern in the UALCAN dataset [Figure 1j]. The analysis of GEPIA2 and UALCAN datasets revealed that TIGIT levels in TNBC were markedly upregulated when compared with those in luminal BC and HER2-positive BC.

miR-15a-5p directly targets and downregulates TIGIT

miRNAs targeting TIGIT were identified using miRDB, miRWalk, and TargetScan databases. Venn diagrams revealed that 41 common miRNAs identified from the three databases were potential direct regulators of TIGIT [Figure 1k]. The potential binding site between miR-15a-5p and TIGIT was determined through the BiBiServ online platform [ Figure 1l] to validate their explicit molecular linkage. The Pearson’s correlation analysis of the TCGA datasets revealed that TIGIT was negatively correlated with miR-15a-5p in different cancer types, especially in thymoma and bladder urothelial carcinoma [Figure S1]. qRT-PCR indicated that five BC cell lines exhibited marked upregulation of TIGIT expression and concurrent suppression of miR-15a-5p compared with their normal counterparts [Figure 2a and b]. The WB analysis demonstrated that TIGIT protein levels in BC cell lines were higher than in normal cells [Figure 2c and d]. Subsequent DLR experiments suggested that introducing miR-15a-5p mimics to HEK293T cells substantially decreased the luc signal of the pmirGLO-TIGIT-WT plasmid versus the miR-15a-5p NC cohort. By contrast, the luc signals of the pmirGLO-TIGIT-MUT plasmid were unchanged [Figure 2e and f], providing evidence that miR-15a-5p directly interacts with TIGIT. For further experiments, MDAMB-468 and MDA-MB-231 cells were selected. Fluorescence microscopy analysis demonstrated that lentivirus infection efficiency varied at different MOI values (MOI = 2, 10, or 50). At an MOI of 10, the infection efficiency was higher than 80% [Figure 2g]. These findings indicate that the lentivirus infection efficiency was sufficient for further analyses.

Close

Figure 2:

Validation of TIGIT as the target gene of miR-15a-5p and analysis of its expression. (a and b) Quantitative assessment of miR-15a-5p and TIGIT contents between different BC cells and normal breast epithelial cells by utilizing qRT-PCR methodology. (c and d) TIGIT protein abundance in different BC cells was evaluated through WB analysis. (e and f) Direct interaction between miR-15a-5p and TIGIT was validated using the DLR assay. (g) Green fluorescent protein expression patterns in cells transfected with lentiviral constructs at different MOI levels were examined using laser confocal microscopy (Scale bar = 50 μm) (h-k). Effects of transfection with LV-vector, LV-TIGIT, miR-15a-5p mimics, NC, or their combinations in BC cells on the expression profiles were assessed using qRT-PCR (h-k) and WB analyses (l and m). Values are expressed as mean ± SD from three distinct experimental replicates. ns: No significant difference, ^✶P <0.05, ^✶✶P < 0.01, ^✶✶✶P <0.001, and ^✶✶✶✶P <0.0001 compared with the NC-transfected cells. qRT-PCR: Quantitative real-time polymerase chain reaction, LV: Lentivirus vector, NC: Negative control, MOI: Multiplicity of infection, WB: Western blotting, SD: Standard deviation, BC: Breast cancer.

Export to PPT

qRT-PCR analysis revealed that the miR-15a-5p levels were upregulated, and the TIGIT levels were downregulated in miR-15a-5p mimic-transfected cells compared with those in NC-transfected cells. Moreover, TIGIT expression was upregulated, and miR-15a-5p expression was downregulated in LV-TIGIT-transfected cells compared with those in LV-vector-transfected cells. Meanwhile, TIGIT expression was downregulated, and miR-15a-5p expression was upregulated in cells co-transfected with miR-15a-5p mimics and LVTIGIT compared with those in LV-TIGIT-transfected cells [Figure 2h-k]. The WB analysis confirmed the findings of qRT-PCR analysis [Figure 2l and m]. These observations suggest that elevated miR-15a-5p levels potentially downregulate TIGIT gene expression. Hence, miR-15a-5p might function as a direct regulator of the TIGIT level.

miR-15a-5p mitigates TIGIT-induced upregulation of TNBC cell growth, migration, and invasion by modulating cell cycle and inducing apoptosis

The viability of transfected TNBC lines (MDA-MB-231/468) was evaluated at multiple time points (0, 12, 36, and 48 h) using the CCK-8 assay. Compared with NC-transfected cells, miR-15a-5p mimic-transfected cells exhibited decreased viability. Conversely, LV-TIGIT-transfected cells exhibited enhanced viability when compared with LV-vector-transfected cells. Notably, the miR-15a-5p mimics and LV-TIGIT cotransfection reduced cell proliferation compared with that in the LV-TIGIT cohort [Figure 3a and b]. The proliferative potential of transfected TNBC cells was examined using colony formation assay. Cells containing miR-15a-5p mimics showed markedly diminished colony development compared with NC controls. By contrast, TIGIT overexpression enhanced colony formation relative to the LV-vector cohort. Nevertheless, the miR-15a-5p mimics and LV-TIGIT co-incorporation diminished colony formation ability versus the LV-TIGIT cohort [Figure 3c-f]. The migration of TNBC cells, shown by wound-healing assessments, was diminished after miR-15a-5p mimic-transfection versus the NC-transfected. Conversely, the migration of LV-TIGIT-transfected cells was higher than that of LV-vector-transfected cells. Furthermore, the miR-15a-5p mimics and LV-TIGIT co-incorporation decreased migration ability compared with that in the LVTIGIT cohort [Figure 4a-d]. Transwell assay was performed to examine the invasive potential of transfected cells. The incorporation of the miR-15a-5p mimic diminished the TNBC cell quantity traversing the basement membrane in comparison with that in the NC cohort. The elevated TIGIT expression increased the number of invasive cells compared with that in the LV-vector cohort. Furthermore, the miR-15a-5p mimics and LV-TIGIT co-incorporation reduced invasion when compared with the LV-TIGIT cohort [Figure 4e-h]. The impact of transfection on TNBC cell cycle progression and apoptotic rates was assessed by flow cytometry. Cell cycle profiling indicated that cells transfected with miR-15a-5p mimics displayed G2/M phase arrest relative to the NC-transfected group. By contrast, TIGIT overexpression reduced the duration of the G2/M phase compared with that in the LV-vector group. Notably, co-transfection with miR-15a-5p mimics and LV-TIGIT resulted in elevated G2/M phase accumulation [Figure 5a-d]. The apoptosis rate (Annexin V+/PI+) in miR-15a-5p mimic-transfected cells was markedly higher than that in NC-transfected cells, whereas the apoptosis rate in LV-TIGIT-transfected cells was lower than that in LV-vector-transfected cells. Furthermore, the apoptosis rate in cells co-transfected with miR-15a-5p mimics and LV-TIGIT was significantly higher than that in cells transfected with LV-TIGIT [Figure 5e-h]. These observations suggest that miR-15a-5p overexpression may inhibit or even counteract the effects of TIGIT in promoting TNBC growth, migration, and invasiveness while triggering cell cycle arrest and enhancing apoptosis.

Close

Figure 3:

Effect of miR-15a-5p and TIGIT on TNBC cell proliferation and clone formation ability. (a and b) Effect of transfection with LV-vector, LV-TIGIT, miR-15a-5p mimics, NC, or both on MDA-MB-231/468 cell proliferation was systematically evaluated using CCK-8 proliferation analysis. (c-f) Effect of transfection with LV-vector, LV-TIGIT, miR-15a-5p mimics, NC, or both on the clone formation ability of MDA-MB-231 (c and e) and MDA-MB-468 (d and f) cells was evaluated using the colony formation assay. Values are expressed as mean ± SD from three distinct experiments. ^✶✶✶✶P <0.0001 compared with the NC-transfected group. TNBC: Triple-negative breast cancer, LV: Lentivirus vector, CCK-8: Cell counting kit-8, NC: Negative control, SD: Standard deviation.

Export to PPT

Close

Figure 4:

Impacts of miR-15a-5p and TIGIT on TNBC cell migration and invasion. (a-d) Cell migration, as evidenced by wound healing assessment (Scale bar = 500 μm), following the incorporation of LV-vector, LV-TIGIT, miR-15a-5p mimics, NC, or both into MDA-MB-231 (a and b) and MDA-MB-468 (c and d) cells. (e-h) Effect of transfection with LV-vector, LV-TIGIT, miR-15a-5p mimics, NC, or both on the invasive ability of MDA-MB-231 (e and f) and MDA-MB-468 (g and h) cells was evaluated using the transwell assay (Scale bar = 200 μm). Values are represented as mean ± SD from three distinct experiments. ^✶✶✶✶P <0.0001 compared with the NC-transfected group. SD: Standard deviation, NC: Negative control, LV: Lentivirus vector, TNBC: Triple-negative breast cancer.

Export to PPT

Close

Figure 5:

Impacts of miR-15a-5p and TIGIT on TNBC cell cycle and apoptosis. (a-d) Effect of transfection with LV-vector, LV-TIGIT, miR-15a-5p mimics, NC, or both on the cell cycle of MDA-MB-231 (a and c) and MDA-MB-468 (b and d) cells was examined using flow cytometry. (e-h) Effect of transfection with LV-vector, LV-TIGIT, miR-15a-5p mimics, NC, or both on the apoptosis of MDA-MB-231 (e and g) and MDA-MB-468 (f and h) cells was examined using flow cytometry. Values are represented as mean ± SD from three distinct experiments. ^✶✶✶P <0.001, ^✶✶✶✶P <0.0001 compared with the NC-transfected cells. NC: Negative control, LV: Lentivirus vector, SD: Standard deviation.

Export to PPT