Identification of novel exosomal miRNAs and their role in diagnosis and prognosis of triple negative breast cancer.

Introduction
Triple-negative breast cancer (TNBC) constitutes a highly aggressive and therapeutically challenging subtype of breast cancer, accounting for approximately 15–20% of all cases [1–3]. Lacking expression of estrogen receptor (ER), progesterone receptor (PR), and HER2, TNBC is not amenable to endocrine or HER2-targeted therapies. As a result, treatment relies heavily on chemotherapy, which, despite initial effectiveness, is often followed by early recurrence, rapid disease progression, and poor overall survival [2, 3]. The absence of molecularly defined targets has driven an urgent need to discover novel biomarkers that can enable early detection, stratify risk, and identify new avenues for intervention.
Within the TNBC landscape, cancer stem-like cells (CSCs) have been implicated as key contributors to treatment resistance, disease persistence, and metastatic relapse. These cells exhibit self-renewal, phenotypic plasticity, and the ability to evade cytotoxic therapies [4, 5]. Their survival following treatment may underlie minimal residual disease and tumor recurrence, making CSCs a critical focus for next-generation diagnostics and effective therapeutic strategies [6].
Exosome-derived microRNAs (miRNAs) have emerged as promising candidates for non-invasive biomarker development [7]. Encased in lipid bilayer vesicles, these small non-coding RNAs remain stable in circulation and can reflect the transcriptional landscape of the tumor from which they originate. Exosomal miRNAs have been shown to mediate cell–cell communication, influence metastatic behaviour, and participate in therapy resistance. While several exosomal miRNAs have been explored in breast cancer broadly, the landscape of secretory miRNAs specifically enriched in TNBC and its CSC population remains poorly characterized [8–12]. Moreover, the mechanistic significance of these miRNAs in TNBC biology has not been systematically addressed.
Our study was intentionally designed as a pilot-scale, discovery-phase investigation focusing on exosomal miRNA profiling to identify potential diagnostic and prognostic biomarkers, with mechanistic and in-vivo validation planned for future work. We present an integrative approach combining meta-analysis, experimental validation, and clinical correlation to identify a novel panel of five exosomal miRNAs enriched in TNBC and TNBC stem cells (TNBCSCs). We hypothesized that these secretory miRNAs not only reflect key molecular features of TNBC aggressiveness but also contribute to its invasive and therapy-resistant phenotype. Our findings aim to define exosomal miRNAs as functional biomarkers with translational potential for early detection, prognostication, and future therapeutic targeting in TNBC.

Materials and methods

Identification and analysis of miRNA profiling datasets specific to TNBC
A search was conducted on PubMed and Web of Science electronic databases to find all the relevant literature studies on miRNA expression in TNBC [13]. Referencing a method described previously by Chen et.al(1), The search algorithms applied included “((microRNA OR micro RNA OR micro ribonucleic acid OR miRNA) AND ((breast carcinoma OR breast carcinomas OR breast cancer OR breast cancers OR triple negative breast cancer OR Triple Negative Breast cancers OR breast tumors) OR (TNBC carcinoma OR TNBC carcinomas OR TNBC OR adenocarcinoma of breast OR TNBC cancers OR TNBC)) AND (Humans [Mesh] AND English[lang]))”. Additionally, a total of 120 miRNA datasets were searched within the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/), with each dataset's title, abstract, and full text were thoroughly reviewed. The selection criteria included: (a) Original experimental studies that compare miRNA expression among different groups (TNBC vs. normal, TNBC vs. non-TNBC, and breast cancer vs. normal) using human tumor tissue samples, and (b) studies that report both upregulated and downregulated miRNAs, including specific cutoff parameters such as fold change and p-values. These specific inclusion criteria enabled the identification of all qualifying miRNA expression datasets. Accordingly, we excluded datasets according to the criteria: (1) Any duplicated publications; (2) in-vitro or pre-clinical studies [14–17]; (3) reviews, reports, editorials [18–33], and (4) studies not characterised by miRNA expression analysis [34, 35]. A total of 53 studies on breast cancer were mined from the public domain and literature survey was conducted. The expression value of miRNA was calculated from each study. After selecting the datasets, the data matrices were downloaded, and differential analyses were conducted using tools available on GEO. Various miRNA microarray platforms were employed, and uniquely expressed miRNAs from each dataset were annotated using miRbase. The fold change (FC) in miRNA expression was normalized by expressing it as log2FC to standardize the obtained miRNA expression values. Five miRNAs showing high differential expression in breast cancer were selected for downstream experiments. Among the five oncogenic miRNAs identified through our meta-analysis, hsa-miR-1180 and hsa-miR-4728 were prioritized for downstream functional investigation based on multiple complementary criteria:(i) Both displayed the highest and most consistent differential expression (≥ eightfold upregulation, adjusted p < 0.05) across all TNBC cell lines, stem-like cell subpopulations, and patient tissue datasets, whereas the remaining three miRNAs showed moderate or variable expression patterns.(ii) They exhibited selective enrichment within exosomal fractions derived from TNBC stem-like cells, with negligible expression in exosomes from non-stem or non-tumorigenic controls, indicating biological specificity for aggressive TNBC phenotypes. (iii) Bioinformatic network analyses (TargetScan, miRDB, miRTarBase, miRmap) revealed that both miRNAs converge on key oncogenic and stemness-associated signalling pathways, including Wnt/β-catenin, Notch, EGFR, and TP53 regulatory axes, supporting their mechanistic relevance.(iv) Kaplan–Meier survival analysis of public datasets demonstrated a strong association between high expression of miR-1180/miR-4728 and poor overall survival in breast cancer cohorts, underscoring their potential prognostic value.(v) Finally, both candidates have limited prior characterization in TNBC, offering novelty and translational potential as previously unreported exosomal biomarkers.

miRNA target prediction
Potential miRNA–target interactions were predicted using established software tools, including TargetScan, miRTarBase, miRDB, and miRmap. The selection criteria included a prediction score of less than 0 for TargetScan, a cumulative score exceeding 95 for miRmap, and a prediction score range of 75 to 100 for miRDB, while all targets from miRTarbase were included for consideration. These miRNA sequences were used as input in conjunction with reference cDNA sequences in the miRanda tool.

Gene ontology and enrichment analysis
Gene Ontology (GO) analysis is commonly used to assess the enrichment of differentially expressed genes (DEGs) in relation to biological processes, cellular components, and molecular functions. The candidate miRNA target genes were subjected to Gene Ontology (GO) and pathway enrichment analyses to elucidate their roles in critical biological pathways. Functional enrichment analysis of the five selected miRNAs’ predicted target genes was carried out using KEGG, Biocarta, Panther, and Reactome databases.

Real-time PCR for miRNA expression analysis
Total RNA was isolated using TRIzol reagent (Invitrogen), with concentration measured at A260 and purity assessed via the A260/A280 ratio.· RNA quality was evaluated using 2% agarose gels. A total of 1 µg of RNA was reverse-transcribed into cDNA using random primers (Thermo), resulting in a final volume of 20 µL of cDNA. Subsequent amplification was performed via PCR for the miRNAs miR-940, miR-1915, miR-6803,miR-1180 and miR-4728, with primer sets listed in Supplementary Table S1. RT-qPCR was conducted using SYBR Green chemistry, normalizing to U6 expression levels, and the ∆∆Ct method was used to calculate normalized fold expression for the target miRNAs [36]. For qRT-PCR data, ΔCt values were obtained from triplicate technical replicates. Error propagation for ΔΔCt calculations was performed using standard deviation estimates from both target and reference Ct values, and fold-change error was computed as (ln 2 × 2^-ΔΔCt × SD_ΔΔCt). Data are expressed as mean ± SD.

Cell culture
In this study, the human mammary epithelial cell line MCF-10A and TNBC cell lines MDA-MB-231 and MDA-MB-468, sourced from the American Type Culture Collection (ATCC, USA) were used. Flow cytometric analysis confirmed enrichment of TNBC stem cells within the TNBC cell lines. The cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% exosome-depleted fetal bovine serum (FBS), 100 IU/mL penicillin, and 100 µg/mL streptomycin. Cultures were routinely maintained at 37 °C in a 5% CO₂ environment.

Transfection
TNBC cells were transfected with 50 nmol/L miRNA antagomiRs (miRNA inhibitors) of hsa-miR-1180 and hsa-miR-4728 using Lipofectamine 3000 reagent (Invitrogen). The cells were treated with the miRNA antagomiR and corresponding scrambled controls (IDT Technologies) in Opti-MEM medium for 4 h, then switched to standard growth medium as per the manufacturer’s instructions. Analysis was performed 48 h after transfection.

Migration and Invasion assays
Following the transfection with anti-miR-1180 and anti-miR-4728, TNBC cells were subjected to wound healing and transwell migration assays. Matrigel (Corning, USA) was coated on top of the transwell chamber and the serum starved transfected cells were added in the upper chamber, seeded at 1 × 104/well. Subsequently, 500 μl of cell growth medium was added to the bottom chamber. Growth media served as a chemoattractant. For 24 to 36 h, the cells were incubated at 37 °C. After the cells migrated to the lower chamber, they were fixed with 70% ethanol and stained using crystal violet. The quantity of migrated (stained) cells was measured by counting the number of stained cells, and the mean cell count per field was computed for each well. Three replicate wells were utilized for every experiment, with representative images captured from randomly chosen fields in each well.

Wound healing assay
TNBC cells (5 × 105 cells/well) were seeded into 6-well plates and grown as a monolayer for 24–48 h. Once cells reached 80% confluence, a horizontal scratch is created in each well using a 200uL pipette tip. Detached cells were removed by rinsing with 500 μL PBS, followed by the addition of 500 μL of fresh medium. The plates were incubated for 12, 18, and 24 h, and images were captured at each interval using an inverted microscope to monitor wound closure progress. Quantitative analysis of wound closure and cell invasion was performed using ImageJ software (NIH, USA). For each experimental condition, four non-overlapping fields per well were randomly selected and analysed. In the wound healing assay, the wound area at each time point was measured using the freehand selection tool, and percent wound closure was calculated as:
[Wound Closure (%) = [(Initial wound area − Final wound area)/Initial wound area] × 100]. Data from replicate fields were averaged to represent each biological replicate (n = 4), and results were plotted as mean ± SD.

Exosome isolation and analysis
Exosomes were isolated using a commercial isolation kit (ExoCan Life Sciences, India) in accordance with ISEV 2018 guidelines. The protocol included differential centrifugation and 0.2 µm filtration to remove cellular debris and larger vesicles (> 200 nm) while preserving canonical exosomes (30–150 nm). Purity and integrity were verified by SEM, TEM, and NTA analyses, enrichment of CD63/CD81, absence of Calnexin, and BCA-based protein quantification, confirming that the isolated vesicles were high-purity exosomes suitable for downstream miRNA profiling. As per manufacturer’s protocol, cells were grown in exosome-depleted, serum-free media. Exosomes were isolated from cell-free conditioned media collected at 48 h by use of Exosome isolation kit (ExoCan Life Sciences, India). The exosome-containing supernatant was filtered through 0.2 μm membrane filters to remove particles exceeding 200 nm in size. Following filtration, the supernatant underwent centrifugation at 20,000 × g for 40 min at 4 °C to isolate the exosomes. The resulting pellet was then resuspended in 1 × PBS for further processing to identify miRNA.

Exosomal characterization (Physical properties)
EV characterization was conducted according to the International Society of Extracellular Vesicles (ISEV) guidelines. The physical properties of exosomes from cancer cell lines and stem cells were analysed using scanning electron microscopy (SEM) and transmission electron microscopy (TEM). A 100 µL aliquot of exosomes was placed on a formvar carbon-coated nickel grid for 1 h. The grid was then rinsed over drops of 0.1 M sodium cacodylate (pH 7.6) and treated with a solution of 2% paraformaldehyde and 2.5% glutaraldehyde in the same buffer for 10 min. The grids were rinsed with 0.1 M sodium cacodylate (pH 7.6), stained with 2% uranyl acetate for 15 min to enhance contrast, washed, treated with 0.4% uranyl acetate for 10 min, air-dried for 5 min, and then examined at 100 kV using a transmission electron microscope.

Exosome characterization using Nanoparticle Tracking Analysis (NTA)
Exosomal size distribution and concentration were determined by Nanoparticle Tracking Analysis (NTA) using a Malvern Nanosight NS300. Freshly isolated exosome pellets were resuspended in phosphate-buffered saline (PBS) and diluted to ~ 1 × 108 particles/mL to achieve 20–100 particles per frame. For each sample, three 60-s videos were recorded at 25 °C with a camera level of 13 and detection threshold of 5, using a continuous syringe pump set at 50 µL/min. Recorded videos were analysed with NTA software v3.4 to generate mean particle diameter, size distribution histograms, and particle concentration. All measurements were performed in triplicate and data are reported as mean ± SD. Reference video is added in supplementary results.

Exosomal characterization (Western blotting)
The BCA Protein Assay Kit (Thermo Fisher Scientific) was employed to quantify exosomal proteins. Following electrophoresis on 4–15% gradient SDS-PAGE gels, 30 µg of protein were transferred to PVDF membranes, which were blocked using 5% bovine serum albumin (BSA) and subsequently incubated with specific primary antibodies, including ACTB, CD63, CD81, and Calnexin (Biolegend, USA) at a dilution of 1:1000 for 24 h at 4 °C. Protein levels were assessed by probing with secondary antibodies against rabbit and mouse conjugated with horseradish peroxidase (HRP) at a dilution of 1:10,000, following three to five washing steps (10 min each). The bound complexes were detected using chemiluminescence methods (ECL; Bio-Rad, USA), and images were acquired using the Amersham ChemiDoc Imaging System.

Exosomal miRNA isolation
Total RNA, including miRNA, was isolated from breast cancer cell lines in vitro using the TRIzol precipitation method. RNA quantity and quality were assessed using the NanoDrop 1000 spectrophotometer.

Immunocytochemical staining and immunofluorescence analysis
Cells were cultured on 0.13 mm thick coverslips (Corning® Cover glass, Corning Life Sciences) for 2 days, followed by washing with PBS and fixation in 1 mL of 100% methanol for 10 min. Permeabilization was then performed using 100% acetone for 30 s, followed by blocking with 1% BSA for 2 h at RT. Samples were incubated with CD81 primary antibodies (1–2 µg) overnight at a dilution of 1:1000, after which the samples were incubated with appropriate secondary antibodies. Finally, cells were mounted using VECTASHIELD mounting medium containing 0.5 µg/mL DAPI (VECTASHIELD; VectorLabs, CA).

Western blotting
Western Blot was performed as described previously by Tyagi Et al [37]. For all western blot analyses of cell lines, proteins were extracted from cells that had reached approximately 80% confluence. The protein concentration in the cell lysates was determined using the BCA Protein Assay Kit (Thermo Fisher Scientific), and 30 µg of protein was loaded into each lane. Protein samples were subjected to SDS–PAGE (12%) and transferred onto PVDF membranes. The immunodetection protocol included protein transfer (15 V for 15 min per membrane) and blocking the membranes with a western blot blocking solution overnight. Following two washes with 1X TBST, membranes were incubated overnight at 4 °C with primary antibodies against SOX2 (1:500) and GAPDH (1:5000) for 1 h at room temperature. After five additional TBST washes, HRP-conjugated secondary antibodies (1:10,000) were applied for 1 h at room temperature, and detection was performed using ECL chemiluminescence. Band intensities were subsequently analysed and compared using ImageJ software and densitometric quantification was carried out.

Culture of tumorspheres
We assessed the capability of cell lines to generate spheres in an anchorage-independent suspension culture. Human breast adenocarcinoma cell lines MDA-MB-231 and MDA-MB-468 were cultured in DMEM (Gibco, ThermoFisher) containing 10% FBS (Gibco, ThermoFisher) and incubated at 37 °C with 5% CO2 for 48 h. Following collection and washing, cells were resuspended in serum-free DMEM/F12 (Gibco, ThermoFisher) supplemented with 10 ng/ml fibroblast growth factor (FGF), 1% B27 and 20 ng/ml epidermal growth factor (EGF). At a density of up to 5000 cells/ml, the cells were seeded in ultra-low attachment 6-well plates (Corning, USA) and incubated in a humidified environment with 5% CO2, set at 37 °C for 4–6 days. Afterwards, the plates were examined for the growth of tumorspheres and measured with an inverted microscope. The tumorspheres were collected through mild centrifugation, followed by dissociation using Accutase (Sigma, USA) to produce individual cells, which were subsequently suspended in a serum-free medium to reform tumorspheres. These tumorspheres were passaged every 5 days. Once the primary tumorspheres grew to a diameter of approximately 100 µm, the samples were collected for downstream analysis.

Flow cytometric analysis and CSC characterization
The expression of the molecular markers CD133⁺, CD44⁺, and CD24⁻ was evaluated by flow cytometry. Tumorspheres were carefully dispersed into a single-cell solution using brief treatment of Trypsin, after they were collected. Cells were labelled with anti-CD44-FITC, anti-CD133-PE, and anti-CD24 AlexaFluor antibodies (BD Biosciences, USA), incubated for 30 min in the dark at 4 °C.The cells were analysed using a flow cytometer (BD FACS ARIA III). The acquisition was set for 10,000 events per sample. Data analysis was performed using the BD FACSDiva™ software (FACSDiva™, BD, USA). Cells were sorted based on surface antigen expression.

Side population analysis
To isolate and identify SP and non-SP fractions, TNBC cells were treated with trypsin, then resuspended in pre-warmed DMEM containing 1% FBS. Dye Cycle Violet reagent (DCV) at a concentration of 5 µg/mL was added in the presence and subsequent absence of verapamil (Sigma), incubated at 37 °C for 90 min with periodic shaking. Following incubation, the cells were washed with PBS containing 1% FBS, centrifuged at 4 °C, and resuspended in ice-cold sheath fluid (BD).Cells were preincubated with the ABCG2 inhibitor fumitremorgin-C(FTC) at a concentration of 10 μg/ml at 37 °C for 30 min before DCV addition. Propidium iodide was added to the cells at a concentration of 1 µg/mL to differentiate viable cells. The Hoechst 33,342 dye was excited at 357 nm, and its fluorescence was subsequently analysed using FACS AriaIII (BD Biosciences, San Diego, CA). The gating for forward and side scatter was stringent, ensuring that debris and non-viable cells were excluded from the analysis. Software used for analysis was BD FACSDiva™ (FACSDiva™, BD, USA).

Immunofluorescence analysis
For tumorsphere immunostaining, cells were plated on glass coverslips (0.11 mm, Corning, USA) in DMEM with 10% FBS for 4 h. Cells were then fixed with 4% paraformaldehyde and incubated with primary antibodies against SOX2 (mouse monoclonal IgG, Santa Cruz; 1:1000), OCT4 (mouse monoclonal IgG, Biolegend; 1:500), and epCAM (mouse monoclonal IgG, Biolegend; 1:500). Corresponding goat anti-mouse secondary antibodies conjugated with FITC, PE, and Cy3 were applied. Tumorspheres were incubated at 37 °C for 60 min, followed by DAPI staining (Sigma) to visualize nuclei. Images were captured with a Zeiss fluorescence microscope and processed using ZEN Blue, ZEN Black microscopy software (Zeiss).

Statistical analysis
Power analysis for miRNA expression comparisons was conducted using observed ΔCt variability from qRT-PCR data. Calculations based on mean effect size (≥ eightfold change) and standard deviation estimates confirmed that the available sample size (n = 15) yields > 80% power at α = 0.05 for detecting significant expression differences between TNBC and control tissues. All quantitative data are presented as mean ± standard deviation (SD) unless otherwise stated. Significance was assessed using two-tailed Student’s t-test or one-way ANOVA with FDR (Benjamini-Hochberg) post-hoc correction, as appropriate with significance levels determined as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Confidence intervals (95%) were included for Kaplan–Meier survival curves. Individual data points are displayed for all quantitative plots to reflect biological replicate variability. All experimental data were quantified from three independent sets of experiments and expressed as mean ± standard deviation (SD). Statistical analyses were conducted using GraphPad Prism software. Additional study workflows are illustrated in supplementary files (Figs. 1, 2).

Results

Identification of differentially expressed miRNAs by meta-analysis in breast cancer
For the meta-analysis, a total of 120 miRNA datasets related to breast cancer were obtained from the Gene Expression Omnibus (GEO) databases. Of these, a total of 53 datasets on breast cancer were shortlisted and mined from the public domain based on various exclusion and inclusion criteria. Following the dataset selection, we downloaded data matrices and differential analyses were conducted using the GEO2R tool provided on the GEO platform. Across these datasets that reported multivariate analyses, expression values of miRNAs were calculated from each study. A set of five miRNAs showing high differential expression in invasive breast cancers was selected for downstream experiments. The predictive significance of the chosen miRNAs in breast cancer patients was validated by performing Kaplan–Meier analysis which demonstrated that all five miRNAs: hsa-miR-6803, hsa-miR-1180, hsa-miR-4728, hsa-miR-1915 and hsa-miR-940 were associated with poor overall survival (OS). Survival analysis indicates increased expression of these miRNAs is correlated with poor prognosis for patients with breast cancer as shown in Fig. 1(a-e). Kaplan–Meier analysis confirmed poorer overall survival for patients with high expression of these miRNAs (HR > 2, 95% CI 1.2–3.8). After FDR (Benjamini–Hochberg) correction, miR-4728, miR-6803, miR-940, and miR-1915 retained significant associations with poorer overall survival (p < 0.05), while miR-1180 showed a non-significant trend. Among these, miR-1915 exhibited the strongest prognostic effect. While these miRNAs have shown poor prognosis across breast cancers, there is limited data to support the same in case of TNBCs as these miRNAs have not been previously reported in TNBC or TNBCSCs. Although these associations are derived from breast-cancer cohorts overall, their consistent overexpression supports further exploration in triple-negative breast-cancer (TNBC) subsets, where survival data remain limited.

Target prediction and Functional annotation for target genes
To further investigate the potential downstream effects of the five selected miRNAs (hsa-miR-4728-3p, hsa-miR-940, hsa-miR-1180, hsa-miR-6803-5p, and hsa-miR-1915-3p), we conducted in silico target prediction using four well-established databases: TargetScan, miRTarBase, miRDB, and miRmap. The union of predicted targets across these platforms was compiled for each miRNA, and a Venn diagram was generated to illustrate both unique and shared targets (Fig. 2a). This analysis revealed that miR-1915-3p had the highest number of predicted targets (7071), followed by hsa-miR-940 (6171), hsa-miR-6803-5p (4709), hsa-miR-4728-3p (3749), and hsa-miR-1180 (2948) (Fig. 2b). These predictions provide a foundational map of genes potentially regulated by the selected miRNAs in the context of breast cancer.
To functionally interpret these targets, we performed enrichment analysis using multiple databases including KEGG, Reactome, Biocarta, and Panther, alongside Gene Ontology (GO) classification. The GO enrichment results (Fig. 2c–d) indicated that the predicted targets were significantly enriched in biological processes related to cell migration, angiogenesis, and regulation of epithelial cell motility, with prominent signalling pathways involving Wnt, Notch, EGFR, and JNK cascades. Additional enrichment was noted for pathways involved in cell division, DNA replication, and stem cell maintenance [38–46]. These findings collectively suggest that the five differentially over expressed exosomal miRNAs may converge on key tumorigenic pathways, contributing to the aggressive phenotypes observed in Triple-Negative Breast Cancer (TNBC). The presence of overlapping target genes also points to potential coordinated regulation of oncogenic processes, warranting further experimental validation.

Exo-miRNAs consistently expressed in CSCs and breast cancer tissues
RT-qPCR was employed to assess the levels of oncogenic miRNAs in TNBC cells and stem cells in vitro (Fig. 3 a, b, d, e). We found a significant upregulation of our target miRNAs hsa-miR-4728, hsa-miR-6803, hsa-miR-940, hsa-miR-1915 and hsa-miR-1180 in TNBC cell lines MDA-MB 231, MDA-MB 468 in comparison to their non-stem counterparts. A similar expression pattern was observed in TNBC and TNBCSCs derived exosomes (Fig. 3c and f). Our panel of five miRNAs: hsa-miR-6803, hsa-miR-1180, hsa-miR-4728, hsa-miR-1915 and hsa-miR-940 were found to be highly expressed in TNBC cells, TNBCSCs and enriched in circulating exosomes. Given their extracellular and circulating nature and consistent expression across TNBC cells and stem cells, these results indicate miRNAs hsa-miR-4728, hsa-miR-6803, hsa-miR-940, hsa-miR −1915 and hsa-miR-1180 may have a significant role to play in TNBC extracellular communications contributing to disease progression, with therapeutic potential if targeted.
Among the five oncomiRs, hsa-miR-1180 and hsa-miR-4728 were selected for downstream characterization due to their significant enrichment and biological relevance in TNBC. These two miRNAs exhibited the highest exosomal fold change in TNBCSCs compared to non-stem counterparts suggesting a potential role in maintaining aggressive stem-like phenotypes. Additionally, clinical validation in patient tissues showed higher expression across all TNBC tissue cohorts, further strengthening their relevance. As shown in Fig. 3(g), both hsa-miR-1180 and hsa-miR-4728 displayed significant upregulation in primary TNBC tumor tissues compared to matched adjacent normal tissues.

Clinical correlation
To clinically substantiate our findings, we assessed the expression of the identified miRNAs in primary breast cancer tissues. Matched tumor and adjacent normal tissue biopsies were obtained from 15 TNBC patients. RT-qPCR analysis of resected tumor specimens revealed consistent overexpression of these miRNAs in TNBC tissues (Fig. 3g). Notably, among the five candidates, hsa-miR-1180 and hsa-miR-4728 exhibited a marked upregulation (exceeding eightfold) in tumor samples relative to matched normal tissues. These results establish a strong association between elevated miRNA expression and TNBC progression. Furthermore, the concurrent enrichment of hsa-miR-1180 and hsa-miR-4728 in circulating tumor-derived exosomes and their detection in clinical tissue specimens underscores their promise as robust, clinically relevant biomarkers for TNBC.

miR-1180 and miR-4728 regulate TNBC cell migration and invasion
To functionally validate the role of these two miRNAs, miR-1180 and miR-4728 in TNBC progression, we performed loss-of-function assays in MDA-MB-231 cells using chemically synthesized antisense oligonucleotides (anti-miRs) targeting each miRNA. Transient knockdown was achieved via lipofection and phenotypic assays were conducted 24–48 h post-transfection.

Wound healing assay
As shown in Fig. 4A, miR-1180 and miR-4728 knockdown (KD) significantly impaired the migratory capacity of TNBC cells. Wound closure was visibly reduced at 18 h in knockdown cells compared to wild-type (WT) controls. This finding was quantitatively validated using ImageJ software, which revealed a statistically significant decrease in wound closure percentage (Fig. 4C; p < 0.05, n = 4), confirming suppression of migratory potential upon miRNA inhibition.

Transwell invasion assay
Consistent with the migration data, Matrigel-coated transwell invasion assays demonstrated a marked decrease in invasive capacity in both miR-1180 and miR-4728 KD cells (Fig. 4B). Crystal violet staining revealed a substantial reduction in the number of invaded cells after 48 h, indicating that silencing either miRNA strongly impairs TNBC invasiveness. These results indicated the role of miR-1180 and miR-4728 in the cell proliferation, migration and invasion in TNBC cells. Quantitative analysis of wound closure (Fig. 4C) confirmed a statistically significant reduction in migratory potential following knockdown. These results implicate miR-1180 and miR-4728 as potential regulators of TNBC cell proliferation, motility, and invasiveness, highlighting their potential as functional oncomiRs in TNBC progression.

Isolation and characterization of exosomes derived from TNBC cells

Morphological assessment
To evaluate the morphological and biophysical properties of extracellular vesicles (EVs) secreted by TNBC cells, exosomes were isolated from conditioned media of MDA-MB-231 and MDA-MB-468 cell lines following 48 h of culture. Comprehensive characterization was performed using scanning electron microscopy (SEM), high-resolution transmission electron microscopy (HRTEM), and nanoparticle tracking analysis (NTA). Further, confocal microscopy was used to qualitatively analyse localization of exosomes within TNBC cells. Phase-contrast imaging (Fig. 5a-b, e-f) illustrated the typical morphology of TNBC cell lines prior to EV harvest while TNBC-derived exosomes were visualized as discrete, spherical particles with sizes ranging from ~ 45 nm to 75 nm using SEM (Fig. 5c, g). HRTEM further confirmed their characteristic lipid bilayer structure and rounded morphology (Fig. 5d, h), consistent with canonical exosomal features. Size distribution observed via TEM ranged between 50–200 nm.

Quantitative profiling
Quantification of vesicle density from SEM micrographs demonstrated a significantly higher number of exosomes per field in TNBC-derived preparations compared to control-derived exosomes (Fig. 5i) suggesting elevated vesicle secretion by TNBC cells. Complementary nanoparticle tracking analysis revealed a unimodal distribution of exosome populations. TNBC-derived exosomes from MDA-MB-468 exhibited a mean diameter of 208 nm, whereas control-derived exosomes from the non-tumorigenic MCF10A line had a smaller mean diameter of 154 nm (Fig. 5j, k). Exosome concentrations were also markedly increased in TNBC-derived samples, indicating enhanced vesicle biogenesis under oncogenic conditions.

Immunofluorescence localization
Immunostaining for CD81, (exosome tetraspanin protein) revealed punctate vesicle-like signals distributed in the perinuclear and cytoplasmic compartments of TNBC cells (Fig. 6a–c), consistent with exosome localization. Magnified images (Fig. 6d) showed intracellular vesicles positive for CD81, highlighted by co-localization with DAPI-stained nuclei.

Protein marker validation
Western blot analysis (Fig. 6e) confirmed the presence of exosomal markers CD81 and CD63 in TNBC-derived exosomes and whole-cell lysates. Calnexin was absent in exosome fractions, validating minimal contamination from intracellular compartments. β-Actin served as a loading control.

Presence of cancer stem-like cells in TNBC
To investigate the presence of cancer stem-like cells (CSCs) within TNBC populations, we employed phenotypic and functional assays including flow cytometry, tumorsphere formation, qRT-PCR, and immunoblotting. TNBC cell lines MDA-MB-231 and MDA-MB-468 demonstrated the capacity to form 3D spheroids under non-adherent conditions, indicative of self-renewal potential and stem-like behaviour (Fig. 7b, d). Side population (SP) analysis using the Hoechst dye efflux assay revealed a discrete population of dye-excluding cells, ranging from 0.9% to 1.1% in both TNBC lines (Fig. 7e, g), consistent with the presence of stem-like cells. Notably, this SP population was abolished upon treatment with verapamil, an ABC transporter inhibitor (Fig. 7f, h), confirming their drug-efflux capability which is a hallmark of CSCs. Marker-based flow cytometric analysis further identified a subpopulation (1–2%) of cells expressing CD133⁺/CD44⁺/CD24⁻, a well-characterized immunophenotype associated with breast CSCs (Fig. 7i–l). These sorted CSC-enriched populations were expanded in vitro as tumorspheres for downstream analysis. Immunofluorescence staining of tumorspheres demonstrated robust expression of key stemness markers, including SOX2 (green, FITC), OCT4 (red, AlexaFluor 647), and EpCAM (yellow, PE) (Fig. 8A–F), confirming the undifferentiated, pluripotent-like nature of these cells. qRT-PCR analysis revealed a significant upregulation of SOX2, ALDH1, and ABCG2 in CSC-enriched spheres compared to parental TNBC populations (Fig. 8G). Consistently, immunoblotting confirmed elevated SOX2 protein levels in tumorspheres relative to monolayer cultures (Fig. 8H), with densitometric quantification (Fig. 8I) supporting a statistically significant increase. Collectively, these findings confirmed the presence of a functionally and molecularly distinct subpopulation of CSC-like cells within TNBC, reinforcing their potential contribution to tumor heterogeneity, drug resistance, and recurrence.

Discussion
This study identifies a novel panel of five exosomal microRNAs: hsa-miR-6803, hsa-miR-1180, hsa-miR-4728, hsa-miR-1915, and hsa-miR-940, that are consistently overexpressed in triple-negative breast cancer (TNBC) cell lines, cancer stem-like cells (CSCs), and clinical tumor samples. Integrative meta-analysis, in vitro functional assays, and clinical correlation collectively demonstrate that these miRNAs are associated with poor prognosis and may serve as key mediators of TNBC aggressiveness. Notably, hsa-miR-1180 and hsa-miR-4728 were functionally validated to promote invasion and migration in TNBC cells, further supporting their role as drivers of tumor progression. Currently, there are hardly any targeted therapies for combat this aggressive and debilitating disease. Through intercellular communication with the tumor microenvironment and potentiated by the CSC population, TNBC cells often acquire treatment resistance leading to metastasis and chemoresistance [25, 45–48]. Additionally, the molecular heterogeneity of TNBC as revealed by the presence of many molecular markers poses challenges to effective treatment [49] Several studies have outlined promising predictive and prognostic markers in breast cancer [50–52]. However, very few studies have identified TNBC specific therapeutic biomarkers [53, 54]. At present, there is no universal biomarker available for diagnosing and targeting TNBC, unlike other breast cancer subtypes, which have biomarkers such as HER2 and specific hormone receptors [50, 51] There is a critical need to identify reliable, non-invasive markers specific to TNBC for rapid screening, early diagnosis, risk assessment, and ultimately for effective treatment and management of breast cancer. This study provides a discovery-phase foundation for future liquid-biopsy investigations and translational validation of specific exosomal miRNAs as non-invasive diagnostic and prognostic markers, as well as potential therapeutic targets.

Biological implications of identified miRNAs
Prior reports have identified exosomal miRNAs such as hsa-miR-21, hsa-miR-1246, and hs-miR-10b as mediators of invasion, chemoresistance, and stemness in TNBC and related breast cancer contexts. Our panel of hsa-miR-6803, hsa-miR-1180, hsa-miR-4728, hsa-miR-1915, and hsa-miR-940 extends this literature by emphasizing less-characterized candidates that show consistent enrichment in TNBC cells, TNBC stem-like subpopulations, and patient tumor tissues. Unlike broadly reported hsa-miR-21/hsa-miR-1246, the candidates prioritized here appear more tightly linked to stemness- and therapy-resistance–associated programmes, thereby complementing established signatures with a distinct biological focus.
Our pathway analysis of the five differentially expressed exosomal miRNAs suggests a strong convergence on key biological processes involved in tumor aggressiveness. Target prediction using four databases (TargetScan, miRTarBase, miRDB, and miRmap) revealed that these miRNAs regulate a wide range of genes implicated in cell motility, proliferation, and survival. Enrichment analysis indicated that these targets are functionally clustered in pathways such as Wnt, Notch, EGFR, and JNK, as well as biological processes like angiogenesis, cell migration, and stem cell maintenance, whose signalling pathways are known to maintain stemness, promote epithelial-to-mesenchymal transition (EMT), and enhance survival in CSC. These findings are consistent with the known roles of exosomal miRNAs in promoting tumor invasiveness and metastatic potential. The persistent expression of these miRNAs in both TNBC cells and their exosomal cargo suggests a mechanism for reinforcing autocrine and paracrine signalling within the tumor microenvironment, thereby contributing to cellular plasticity and therapeutic resistance. Although experimental validation of individual targets remains ongoing, this integrative bioinformatic analysis provides a systems-level overview that supports the functional relevance of our selected miRNAs in driving TNBC pathobiology. Future studies will explore whether combinatorial targeting of these miRNA-regulated pathways can offer therapeutic benefit.
After comprehensive literature review and subsequent meta-analysis of 120 publicly available breast cancer datasets, 53 relevant datasets were identified based on various exclusion and inclusion criteria. Upon analysis of these datasets, a set of five, highly oncogenic miRNAs were found to be highly expressed in TNBCs: hsa-miR 6803, hsa-miR 1180, hsa-miR 4728, hsa-miR 1915 and hsa-miR 940. Though these miRNAs were often found in breast cancers, they have never been shown to be exclusively reported in TNBC, making them ideal candidates for therapeutic interventions. Additionally, there is currently no evidence linking these oncomiRs with disease relapse, we were eager to explore their gene expression profile in TNBCSCs.
Among the five oncomiRs identified in our meta-analysis, hsa-miR-1180 and hsa-miR-4728 were prioritized for further study due to their markedly elevated expression in TNBC stem-like cells and their plausible functional relevance in cancer progression. Specifically, both miRNAs exhibited a ≥ eightfold enrichment in exosomes derived from TNBC stem-like cells compared to non-stem counterparts, suggesting a potential role in maintaining aggressive stem-like phenotypes. In contrast, the remaining miRNAs showed only moderate differential expression. In addition to in vitro models, clinical validation in patient tissues further strengthened their relevance. As shown in Fig. 3g, both hsa-miR-1180 and hsa-miR-4728 displayed significant upregulation in primary TNBC tumor tissues compared to matched adjacent normal tissues suggesting their potential role as biomarkers of tumor aggressiveness. Literature further supports their oncogenic potential: hsa-miR-4728 has been implicated in HER2-positive breast cancer and is known to modulate estrogen receptor signalling and resistance mechanisms [55] while hsa-miR-1180 has been linked to Wnt/β-catenin-driven cell migration and metastasis in colorectal cancer models [56] Bioinformatic target prediction using TargetScan indicated that both miRNAs may regulate key tumor suppressors such as E-cadherin and PTEN, supporting their role in promoting cell motility and invasion. Taken together, their high exosomal enrichment, literature-supported oncogenicity, and strong predicted links to migration-associated pathways made hsa-miR-1180 and hsa-miR-4728 ideal candidates for functional validation in the context of TNBC stemness and aggressiveness.
To uncover potential shared regulatory mechanisms between hsa-miR-1180 and hsa-miR-4728, we performed integrative target prediction using four independent databases: TargetScan, miRTarBase, miRDB, and miRmap. This analysis identified 14 common target genes between the two miRNAs (Fig. 9A), suggesting potential functional convergence despite their distinct genomic origins. To explore the biological relevance of these shared targets, we utilized the STRING database (https://string-db.org) for protein–protein interaction (PPI) and RNA–protein association analysis. Network mapping revealed that these targets are not randomly distributed but instead form coherent interaction modules (Fig. 9B–D). Specifically, we observed three major functional clusters:
(i) Epigenetic regulators, including histone acetyltransferases (HATs; red nodes), implicating these miRNAs in transcriptional control and chromatin remodelling; (ii) TP53-regulated metabolic effectors (blue nodes), highlighting their potential roles in stress responses and metabolic adaptation in cancer cells; and (iii) PPARα-associated transcriptional programs (green nodes), previously linked to lipid metabolism and breast cancer progression. The list of shared target genes for hsa-miR-1180 and hsa-miR-4728 identified by STRING analysis is provided in Supplementary Table S4. The convergence of hsa-miR-1180 and hsa-miR-4728 on these core regulatory axes suggests a cooperative role in modulating pathways central to tumor plasticity, epigenomic reprogramming, and metabolic rewiring—hallmarks of aggressive TNBC phenotypes. This mechanistic overlap reinforces their selection as prioritized candidates for functional validation and therapeutic targeting in the context of TNBC stemness and relapse.

Functional impact on TNBC progression
Our results demonstrate that hsa-miR-1180 and hsa-miR-4728, in particular, significantly enhance TNBC cell motility and invasiveness, as shown by wound healing and invasion assays. These miRNAs also share a subset of predicted targets involved in epigenetic regulation, TP53 signalling, and metabolic adaptation. Several studies have reported exosomal miRNAs such as hsa-miR-21, hsa-miR-1246, and hsa-miR-10b as mediators of invasion [57], metastasis, and stemness in TNBC [58] and other breast cancer subtypes. While these miRNAs are well established, our findings expand the current repertoire by identifying lesser-known candidates (hsa-miR-1180, hsa-miR-4728, hsa-miR-940, hsa-miR-1915, and hsa-miR-6803) that demonstrate selective enrichment in TNBC and TNBC stem-like exosomes.
The ability of these miRNAs to modulate both phenotypic traits and intracellular signalling underlines their functional role in sustaining the aggressive behaviour of TNBC. Of the five miRNAs, two were found to show consistently high expression in TNBC clinical tissue. RT-qPCR analysis revealed a notable increase in the expression of hsa-miR-1180 and hsa-miR-4728 in TNBC tissues when compared to non-tumor breast tissues. Our study has confirmed that hsa-miR-1180 and hsa-miR-4728 act as tumor promoters, suggesting that their upregulation in tumor tissues may contribute to the progression and metastasis of TNBC. Decreased expression of these miRNAs attenuated tumor progression, as demonstrated by wound healing and invasion assays (Fig. 4A, B). Together, these observations suggest that hsa-miR-1180 and hsa-miR-4728 may act as exosomal effectors of stemness-linked invasion, prioritizing them for orthogonal mechanistic validation and prospective cohort testing.

Comparison with existing literature
Previous studies have linked hsa-miR-6803, hsa-miR-1180, hsa-miR-4728, hsa-miR-1915, and hsa-miR-940 to various cancers, including colorectal, gastric, and cervical carcinomas (as indicated in Table 1). However, their roles in TNBC exosomal secretion and CSC biology remain under-characterized. Our findings associate this miRNA panel with TNBC exosomes and stem-like subpopulations and validate consistent expression across in-vitro models and patient tissues in a discovery-phase framework. Furthermore, our integrative approach reveals mechanistic convergence between exosomal miRNA signalling and known oncogenic pathways, a perspective largely absent in prior reports.
Using gene enrichment analysis and target prediction models, these dysregulated miRNAs were also found to be involved in a wide range of crucial biological processes and pathways including regulation of RNA pol II, Wnt signaling, Notch signaling, MAPK signaling pathway and the regulation of JNK signaling cascade. Previous studies have highlighted the correlation of Wnt signaling pathways with maintenance of stem cell niche and expression in cancer stem cells [34, 35, 38–42]. Notch receptor and ligand overexpression is linked to TNBC progression [39, 43]. Notch receptors are also involved in modulating the behaviour of tumor-initiating cells and in the initiation of TNBC [39, 43, 44]. These pathways are pertinent to the therapeutic targeting of CSCs or other cells responsible for diverse prolapse as they play key roles in genesis and regulation of cell survival and their fate [45]. Given that many of these pathways are known to support tumor cell proliferation, cancer stem cell (CSC) survival, epithelial-to-mesenchymal transition (EMT), and invasion. It is likely that these miRNAs contribute to the progression and expansion of TNBC and TNBC stem cells.
Exosome-encapsulated miRNAs have a great potential as prognostic biomarkers. Several studies have demonstrated that exosomes are robust and may be stored for extended periods without compromising the integrity of encased miRNAs. These features greatly increase their potential applicability in a diagnostic or clinical setting. Secretory miRNAs are reflective of their parent cell status and thus may reveal a more specific tumor profile than conventional miRNA profile derived from whole blood or serum. From our identified set of five oncogenic miRNAs, hsa-miR-6803 has been previously identified as a diagnostic and prognostic marker in colorectal cancer, suggesting it may be a circulating marker for aggressive tumours. Yan et al. found hsa-miR-6803-5p promotes tumor cell proliferation and invasion via the NF-κB pathway [61, 92], hyper-activation of which is often seen in TNBC. ​Hsa-miR 1180 appears to be a potent oncomiR across multiple cancers and can activate Wnt pathway. Guo et al. showed that plasma exosomal hsa-miR-1180-3p is a novel diagnostic marker for cutaneous melanoma [64]. However, Tan et al. demonstrated that hsa-miR-1180 confers apoptotic resistance in hepatocellular carcinoma via NF-κB activation [66]. This dual ability to activate Wnt and NF-κB signals suggests hsa-miR-1180 can promote survival, stemness, and therapy resistance, which might explain the aggressive behaviour of TNBC cells and stem-like cells expressing it. Studies have shown that hsa-miR-1180 from mesenchymal stem cells also promoted ovarian cancer cell glycolysis and proliferation via suppressing the Wnt inhibitor SFRP1​. Similarly, hsa-miR-6803 has been shown to activate NF-κB signalling – in colorectal cancer, exosomal hsa-miR-6803-5p drove cancer cell proliferation and invasion via the PTPRO/NF-κB axis and could contribute to invasion or chemoresistance in TNBC.​ [62–67] hsa-miR-1915 has been reported in gastric cancer [68, 73], colorectal carcinoma (CRC) [74–76] and most recently in Breast cancer as a circulating biomarker for breast cancer (along with hsa-miR-455-3p) [77]. This demonstrates that miR-1915 can be detected in blood and has diagnostic value for breast malignancy. Interestingly, hsa-miR-4728 was found to have dual functions. It acts as a tumor suppressor in CRC [78] while being tumorigenic in breast cancer [69–71]. Hsa-miR-4728 has also been identified as a marker of HER2 status in BrCa patients. It is encoded within the HER2 gene, and although TNBC lacks HER2 amplification, recent studies in breast cancer highlight hsa-miR-4728’s oncogenic potential. Floros et al. reported co-amplification of hsa-miR-4728 in HER2-positive breast cancers helps tumors evade anti-HER2 therapy essentially, hsa-miR-4728 upregulation protected cancer cells from targeted drugs, underscoring its role in therapy resistance [71]. While TNBC cells do not overexpress HER2, miR-4728 might still exert pro-tumor effects via HER2-independent pathways or by affecting other growth factor pathways; its presence in TNBC exosomes is therefore intriguing. Hsa-miR 940 has been widely reported as an oncogenic marker for CRC [72], gastric cancer [79], cervical cancer [80] and breast cancer [81, 82]. Zhang et al. demonstrated that miR-940 promotes breast cancer progression by downregulating the tumor suppressor FOXO3, indicating it can drive proliferation and invasion by silencing a key cell-cycle/apoptosis regulator [82]. This may explain the aggressive phenotype observed in TNBC cases with high hsa-miR-940. Rashed et al. found that exosomal hsa-miR-940 from cancer cells can maintain oncogenic SRC signalling in recipient cells. It is involved in four critical pathways: the Wnt/β-catenin pathway, the MAPK pathway, the PD-1 pathway, and the phosphatidylinositol 3-kinase (PI3K)-Akt pathway, all of which are significantly implicated in breast carcinogenesis [83, 84]. While that study was not in TNBC, it reinforces our hypothesis that tumor-derived exosomal hsa-miR-940 can alter the tumor microenvironment or distant sites to favour cancer progression.

Diagnostic and therapeutic potential
The secretory and stable nature of these miRNAs and their detectability in exosomes make them attractive candidates for liquid biopsy-based monitoring [93]. Their stability in circulation and association with tumor aggressiveness suggest they could serve as prognostic markers for early relapse or metastatic risk.
Moreover, hsa-miR-1180 and hsa-miR-4728 may serve dual roles as biomarkers and therapeutic targets, given their contribution to invasive phenotypes. Targeted inhibition of these miRNAs could potentially disrupt key survival pathways and improve treatment responses. While many of these selected miRNAs have been identified in various types of cancer, their secretory roles in TNBC and TNBC stem cells (TNBCSCs) were previously unrecognized. Our study reveals the presence of secretory miRNAs (miRNAs hsa-miR 6803, hsa-miR-1180, hsa-miR-4728, hsa-miR-1915 and hsa-miR-940) in TNBC and TNBC stem cells highlighting their diagnostic value and clinical utility for better management of TNBC. Furthermore, these secretory oncomiRs were found consistently overexpressed in TNBC and TNBCSCs. We subsequently correlated our findings with TNBC tumor tissue samples (n = 15) and found a consistent high expression of all five oncomiRs across tumor biopsies. From this panel of five miRNAs, two oncomiRs: hsa-miR-1180 and hsa-miR-4728 were found to be significantly upregulated across all tumor tissue samples in TNBC patients. The presence of these two secretory miRNAs in TNBC and TNBCSC along with their overexpression in clinical tissue samples indicate their possible role in TNBC progression and metastasis and may serve as reliable prognostic as well as therapeutic marker(s).
The identification of novel secretory miRNAs in TNBC and TNBCSCs is promising with potential application in the diagnosis and treatment of hormone refractory, metastatic breast cancers. Our study identifies novel oncogenic exomiRs in TNBC and TNBCSCs and highlights their utility as therapeutic targets for TNBC. The mechanisms through which these exomiRs influence the development and progression of TNBC are not yet fully elucidated. Nonetheless, it is believed that they may modulate the expression of genes associated with cell signalling pathways, including the Wnt, Notch, and MAPK pathways, which are known to play a role in the progression and relapse of TNBC. Previous studies have similarly demonstrated prognostic potential of several other miRNAs in breast cancer such as hsa-miR-9 [85], hsa-miR-21 [9, 11, 12], hsa-miR-29b [9, 33] and hsa-miR-331 [94]. Our study is the first to assess prognostic behaviour of these novel TNBC and TNBCSC derived exosomal miRNAs. Elucidating the role of secretory miRNAs in the molecular mechanisms driving TNBC initiation and progression is crucial for advancing miRNA-based therapeutics, which hold promise as effective treatment options for TNBC patients.
Our study establishes a novel exosomal miRNA signature associated with TNBC and its stem-like subpopulations. These findings offer new avenues for biomarker discovery and therapeutic innovation, paving the way for more personalized and effective management of this challenging breast cancer subtype.
Our findings provide a strong preliminary framework, we recognize that confounding factors like tumor heterogeneity, inter-patient variability, and the limited cohort size may influence exosomal miRNA expression patterns. Selection bias arising from the availability of matched tissue specimens and variability in sample handling could also impact quantitative accuracy. Moreover, exosomal miRNA enrichment may, in part, reflect stress-induced secretion dynamics rather than strictly tumor-driven packaging. Considering these factors, the clinical translation of exosomal miRNAs as diagnostic or prognostic biomarkers will require larger, multi-institutional validation cohorts, standardized isolation and normalization protocols, and longitudinal follow-up. Nonetheless, the secretory stability and reproducible enrichment of these miRNAs highlight their promise as candidate markers for future liquid-biopsy applications. This study therefore serves as a proof-of-concept, with definitive clinical validation to be pursued through expanded, multi-subtype analyses incorporating ROC-based performance metrics and benchmarking against established TNBC biomarkers.

Study limitations and future directions
The present work represents an exploratory, discovery-phase study aimed at defining exosomal miRNA signatures with diagnostic and prognostic relevance in TNBC, rather than pursuing detailed mechanistic discovery. While our study presents a robust foundation, it is limited by the modest number of clinical samples analysed. Since TNBC is a highly metastatic disease, it is very challenging to obtain fresh surgical tissue samples. We were able to include a small number of tumor tissue samples which may limit the clinical relevance and translational impact of our study. We recognize that analysis of larger datasets and clinical correlation using higher numbers of TNBC tissues could provide greater insights on prognostic significance of these TNBC derived exosomal miRNAs. The number of patient samples (n = 15) was determined by the availability of well-annotated TNBC tissues with matched adjacent normal tissue. A post-hoc power estimation, based on the observed ΔCt variance and ≥ eightfold expression difference, indicated adequate statistical power (> 0.8, α = 0.05) for detecting large-effect signals typical of preliminary biomarker discovery. While this sample size is modest, it provides a reliable foundation for hypothesis generation, and can be expanded in multicentric validation studies to confirm diagnostic accuracy and prognostic strength.
Future work will also incorporate additional control groups, including non-TNBC breast cancer subtypes, benign breast lesions, and healthy donor plasma—to better define subtype-specific and circulating biomarker specificity. Functional studies in vivo are also planned to confirm the role of these miRNAs in tumor progression and metastasis. In the next phase, we aim to establish TNBC xenograft and orthotopic mouse models using hsa-miR-1180 and hsa-miR-4728 knockdown and overexpression systems to assess their effects on tumor initiation, growth kinetics, and metastatic potential. These studies will incorporate Ki-67 immunohistochemistry and bioluminescent imaging for in-vivo tumor tracking. Such murine validation will help delineate the causal role of these exosomal miRNAs in TNBC tumorigenesis and provide translational support for their therapeutic targeting. Future research will also explore combinatorial strategies involving specific miRNA inhibition and standard chemotherapy or immunotherapy to assess synergistic effects. Additionally, the integration of miRNA profiling with multi-omic datasets could provide a deeper understanding of TNBC heterogeneity and refine patient stratification strategies. We also recognize that while our exosome isolation protocol complied with ISEV 2018 recommendations, comparative evaluation using ultracentrifugation and additional purity markers (e.g., CD9, Alix) would further strengthen methodological rigor, and such optimization is planned for subsequent validation studies.
During the systematic literature review, the initial screening was conducted across PubMed, Web of Science, and Cochrane Library, which may have excluded some non-indexed or preprint studies. Additional searches across preprint servers and Google Scholar are planned in future updates to minimize bias. Additionally, the limited availability of publicly listed TNBC clinical data sets in GEO and TCGA may have narrowed the scope of our estimates. We have analyzed potential circulatory prognostic markers for TNBC, validated our findings in TNBC tumor tissues and further validated them in vitro using TNBC cell lines. These combined analyses provide preliminary evidence for the translational potential of exosomal miRNAs, though they cannot yet capture tumor microenvironment complexity or inter-patient variability, which future longitudinal studies will address.
Regarding functional validation, although five exosomal miRNAs were identified as candidates, we prioritized hsa-miR-1180 and hsa-miR-4728 based on their strongest differential expression and clinical relevance. Furthermore, dose–response and time-course optimization for miRNA inhibition was not performed in this exploratory phase, as the focus was on establishing proof-of-concept for functional impact. These experiments are part of our planned validation studies to determine optimal inhibitory concentrations, confirm target specificity, and assess reproducibility across biological replicates. Another limitation of our study is the absence of experimental validation of direct mRNA targets for the selected exosomal miRNAs. While we have computationally predicted several plausible targets based on established databases, we recognize that functional confirmation is essential to substantiate these regulatory interactions. Planned follow-up experiments will include luciferase reporter and RNA immunoprecipitation (RIP) assays to validate predicted miRNA–target pairs (e.g., PTEN, E-cadherin) and elucidate downstream effects on signalling pathways identified through STRING and KEGG analyses. Future studies will also aim to validate common downstream targets and investigate how these genes contribute to TNBC aggressiveness and stemness phenotypes and include expanded statistical modelling with multiple-comparison corrections and receiver-operating-characteristic (ROC) validation in larger datasets to confirm diagnostic performance and reproducibility. Collectively, addressing these limitations through larger patient cohorts, refined exosome isolation, expanded controls, and mechanistic assays will advance this work from an exploratory foundation toward clinically actionable biomarker validation in TNBC.

Conclusion
Collectively, our findings underscore the potential of two exosomal miRNAs, hsa-miR-1180, hsa-miR-4728, and their secretory counterparts as clinically relevant biomarkers for aggressive forms of breast cancer. Future studies involving larger patient cohorts and longitudinal follow-up will be crucial to establish their utility in routine diagnostics or therapeutic targets. This work opens a promising avenue for non-invasive miRNA-based diagnosis, monitoring and risk stratification in TNBC.

Supplementary Information