Citrus limon L. purchased from a local market, carefully washed and squeezed. The juice was sequentially centrifuged at 3,000 × g for 30 min to remove dead cells, cell debris, and large particles. The supernatant was centrifuged at 10,000 × g for 60 min. Subsequently, the supernatant was filtered at 0.22 μm pore filter and centrifuged at 100,000 × g for 90 min in a fixed angle rotor Type 50.2 Ti (Beckman Coulter Inc., Brea, CA, USA). To further purify small EVs, the pellet was suspended in 5 mL cold PBS (1X) and transferred to a 30% sucrose/D₂O cushion for ultracentrifugation. After centrifugation at 100,000 × g, 4°C for 120 min, the nanovesicle-containing fraction was resuspended in PBS (1X) and was centrifuged twice at 100,000 × g, 4°C for 90 min. The pellet was collected and resuspended in PBS (1X) for the subsequent experiments.

The protein concentration of nanovesicles was determined using the BCA assay kit (Thermo Fisher Scientific, Waltham, MA), following the manufacturer’s instructions. Samples mixed with working reagent were incubated at 37°C for 30 min and then at room temperature for 20 min before being read. Protein concentration standard curve was obtained by using a series of bovine serum albumin (BSA) ranging from 25 μg/mL to 1,000 μg/mL. The absorbance was measured at 562 nm in a Cytation3 Cell Imaging Multi-Mode Reader (BioTek, Winooski, VT, USA).

CLENs (10 μL) were added to Formvar/Carbon 200 mesh copper grids for 2 min at room temperature. The grids were dried using filter paper to remove excess moisture. After rinse with 10 μL of ddH₂O, 10 μL of 5% uranyl acetate were dropped onto the copper grids. After 2 min, the excess staining solution was absorbed using filter paper, and the samples were air-dried for 30 min under the biological hood. The samples were observed by a Zeiss Libra 120 transmission electron microscope (TEM) at 80 kV.

Size distribution of CLENs was determined by dynamic light scattering (DLS) performed with a Zetasizer Nano ZS90 (Malvern Panalytical, Malvern, United Kingdom). Samples were diluted 1:100 with 0.22 μm filtered PBS prior to measurement. Average size distributions determined from triplicate measurements are reported. Data analysis was performed by Malvern Zetasizer software (Palmieri et al., 2014; Guglielmi et al., 2020).

The size distribution and concentration of CLENs were characterized by NanoSight NS300 (Malvern Technologies, Malvern, United Kingdom). Accurate nanoparticle tracking was verified using a 30 nm gold nanoparticle standard (Sigma Aldrich, product #753629, St. Louis, MO, USA) prior to examination of the samples. CLENs were diluted 100-fold with sterile filtered (0.22 μm pore size) PBS for NTA measurement. Five videos of each sample were captured with a high-resolution camera to obtain the mean and standard error, and outcomes were analyzed with NTA 3.4 software.

The human embryonic kidney-293 (HEK-293) and murine mammary tumor cell line 4T1 were maintained in culture in Dulbecco’s Modified Essential Medium (DMEM, Sigma-Aldrich, St. Louis, MO, USA) enriched with 10% fetal bovine serum (FBS, Gibco, Life Technologies) and 2% penicillin–streptomycin (Sigma-Aldrich, St. Louis, MO, USA). Human basal-like TNBC cell line HCC-1806 was maintained in culture in Roswell Park Memorial Institute (RPMI) 1,640 medium (Thermo Fisher Scientific, Waltham, MA, USA) llemented with 10% fetal bovine serum (FBS) and 2% penicillin–streptomycin. The cells were purchased from American Type Culture Collection (ATCC, Rockville, MD, USA) and maintained at 37°C with 5% CO₂ under a humidified atmosphere.

The CLENs (100 μL) isolated as described above were stained with 50 μM Calcein-AM for 30 min at 37°C. After incubation, the labeled CLENs were resuspended in 5 mL of 0.22 μm filtered PBS and subjected to ultracentrifugation at 100,000 x g for 60 min. The pellet was carefully resuspended in 100 μL PBS, ready for subsequent incubation with cells. 4T1 and HCC-1806 cells were seeded at a density of 1 × 10⁴ cells/cm² per well in an 8-well chambered coverglass (Cellvis, Cat. #: C8-1.5-H-N). The day after, cells were incubated with 20 μg/mL of labeled-nanovesicles for 4 h at 37°C or 4°C. After labelling, the cells were fixed with 4% paraformaldehyde for 20 min at room temperature, and then washed two times with PBS. Fixed cells were permeabilized with 0.1% Triton^™ X-100 (Sigma-Aldrich) for 15 min at RT, followed by two washes with PBS. Subsequently, cells were stained with one unit of rhodamine phalloidin (Life Technologies, Molecular. Probes) for 1 h at RT, and the nucleus was stained with DAPI for 15 min in the dark condition at room temperature. Confocal microscopy was used to visualize the cellular uptake of CLENs after 4 h of incubation. Fluorescence images were acquired using Nikon’s A1 MP + multiphoton confocal microscope, equipped with a ×60 oil immersion objective.

Cell viability of HEK-293, 4T1 and HCC-1806 cells were evaluated by CellTiter-Glo^® Luminescent Cell Viability Assay (Promega, Madison, WI, USA). The 1 × 10⁵ cells were seeded in each well of the 96-well plate in a complete DMEM medium and incubated overnight at 37°C with 5% CO₂. After 24 h, cells were treated with CLENs at serial concentrations of 5, 10, 20, 40 and 80 μg/mL in DMEM supplemented with 2% FBS, as low serum conditions can reduce the availability of growth factors and nutrients necessary for cell survival. After incubating for 24, 48, or 72 h at 37°C, the cells were thoroughly washed with PBS to remove excess CLENs, and then equal volume of CellTiter-Glo^® reagent were added to the cell culture medium present in each well and incubated for 2 min with an orbital shaker to induce cell lysis. The plate was incubated for an additional 10 min at room temperature in the dark to stabilize the luminescence signal. After incubation, luminescence was measured by Cytation 3 Cell Imaging Multi-Mode Reader (BioTek, Winooski, VT, USA).

Live/dead staining was performed using Calcein AM and propidium iodide (PI). Calcein-AM itself is a non-fluorescent analog of calcein, but the highly lipophilic acetoxymethyl ester of calcein permeates cell membranes, and intracellular esterase generate calcein from Calcein-AM, which emits intense green fluorescence. Calcein AM, a cell-permeable dye that stains only living cells, whereas PI is a cell-impermeable DNA-binding dye that penetrates the cell membrane of dead or dying cells. The 1 × 10⁵ of HEK-293, 4T1 and HCC-1806 cells were seeded in the 96-well plate in a complete DMEM medium and incubated overnight at 37°C with 5% CO₂. The cells were treated with CLENs at concentrations of 5, 10, 20, 40, and 80 μg/mL for 24, 48, and 72 h. Subsequently, the cells were incubated with green-fluorescent Calcein-AM for 20 min, followed by a 5 min co-culture with PI. The fluorescence images were observed by Cytation 3 Cell Imaging Multi-Mode Reader (BioTek, Winooski, VT, USA), with excitation at 488 nm and emission recorded at 525 nm.

A density of 5 × 10⁵ cells was seeded in a 12-well plate and incubated overnight at 37°C in a 5% CO2 incubator. The day after, when the cells reached density of 90% confluence, the cell monolayers were scratched using a sterile P-200 pipette tip and washed twice with PBS to remove the floating cells and debris. The PBS was pre-warmed at 37°C to avoid detaching cells during the wash. Subsequently, the cells were treated with low serum (2% FBS) medium containing CLENs at concentrations of 40 and 80 μg/mL or without treatment. Wound healing images were continuously captured at indicated time points using Cytation 3 Cell Imaging Multi-Mode Reader (BioTek, Winooski, VT, USA) with ×4 magnification. The remaining wound area was measured using ImageJ software.

The cells were trypsinized and resuspended in cold serum free DMEM mixed with cold Matrigel (Corning; catalog no.: 354,234) at a final concentration of 8 mg/mL. Four drops of Matrigel containing 7,500 cells per drop were plated in six well plates by BioX (Cellink) 3D-bioprinter. All materials were pre-chilled and experiments were performed on ice to avoid gelation. The drops were allowed to solidify by upside-down the plate and placing it in a humidified incubator at 37°C, under 5% CO₂ for 30 min. Afterwards, the plate was inverted again and left for an additional 30 min, and then 3 mL of complete medium is added to each well. Images were taken at day 0, day 7, and day 10 post-incubation, both in the absence and presence of CLENs. Quantification of cell aggregates and their respective areas were analyzed using ImageJ software (Moriconi et al., 2017; Perini et al., 2022).

After 72 h of incubation, cellular protein extraction was performed by homogenizing cells in cell lysis buffer (#FNN0011, Invitrogen) supplemented with an EDTA-free Protease Inhibitor (Thermo Scientific™) and PMSF Protease Inhibitor (Thermo Scientific™). Subsequently, the resulting homogenates were subjected to Western blot analysis using established protocols, as previously described (Cui et al., 2022). Briefly, a total of 20 µg of proteins extracted from cell lysates were loaded onto a 4%–20% Mini-PROTEAN^® TGX^™ Precast Protein Gel (Bio-Rad, Hercules, CA, USA), and transferred to a polyvinylidene difluoride (PVDF) transfer membrane (Thermo Scientific^™). The membrane was blocked by EveryBlot Blocking Buffer (Bio-Rad, Hercules, CA, USA) for 10 min to prevent non-specific binding. Subsequently, the membrane was incubated overnight at 4 °C with primary antibodies (PI3K #4257, p-PI3K #17366, Akt #9272, p-Akt #9271, Erk #4695, p-Erk #4370 and β-actin #93473, from Cell Signaling Technology, with a dilution ratio of 1:1,000). The membranes were washed three times with TBS-T and then incubated with secondary antibody (Anti-rabbit IgG, HRP-linked Antibody #7074, from Cell Signaling Technology, 1:2000) at room temperature for 1 h. After three washes with TBS-T, the protein bands were incubated with Pierce^™ ECL Western blotting Substrate (Thermo Scientific^™) and detected using the Alliance Q9 Advanced System (UVITEC, Cambridge, United Kingdom). Densitometric analysis was performed by using ImageJ software.

Quantitative data were obtained from three independent experiments and are presented as means ± SEM. Statistical analysis was performed using one-way ANOVA and two-way ANOVA, followed by Tukey’s multiple comparison post hoc test, to determine the significance of differences. Statistical analysis was executed with GraphPad Prism9 software (San Diego, CA, USA), using p < 0.05 as the critical level of significance (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). Three independent experiments done in triplicate.

One-step sucrose cushion ultracentrifugation approach (Figure 1A) was employed to extract extracellular nanovesicles from C. limon L. juice. This methodology is widely regarded as the “gold standard” for EV isolation due to its established effectiveness and reliability (Gupta et al., 2018). A representative TEM image was reported in Figure 1B shows that CLENs are spherically shaped with a diameter of approximately 50–100 nm, which is consistent with the results obtained from dynamic light scattering (DLS) analysis (Figure 1C) and Nanoparticle Tracking Analysis (NTA) measurements (Figure 1D), ensured unambiguous determination of the nanovesicle size, thus strengthening the reliability of the findings. Furthermore, the NTA revealed a concentration of the nanovesicles to be 1.03 e+11 ± 1.71 e+10 particles/mL.

The cellular uptake of nanovesicles offers valuable insights into their potential therapeutic response, as a high uptake specifically by cancer cells implies the potential for enhanced treatment outcomes. The human HCC-1806 cells and the mouse 4T1 cells were used as representative models of TNBC. The uptake behavior of CLENs in 4T1 and HCC-1806 TNBC cells at 37°C and 4°C was obtained by confocal microscopy, allowing a thorough investigation of its internalization pattern. As shown in Figures 2A, B, following 4 h of incubation at 37°C, the extracellular nanovesicles was co-localized with intracellular vesicular structures, highlighting the successful uptake by both cell lines. A large number of labeled nanovesicles can be observed within the cytoplasm (Figures 2C, E). In general, the nanovesicles uptake occurs via energy-dependent (ED) mechanisms, energy-independent (EI) mechanisms, or a combination of both. Endocytosis is a well-studied ED pathway for the transportation of nanoparticles across the cellular membrane, facilitating their internalization into specific cells and subsequent accumulation, which includes phagocytosis, macropinocytosis, clathrin- or caveolin-mediated endocytosis. Conversely, EI internalization mechanisms involve fusion, embedment, and direct translocation (Quagliarini et al., 2022). Traditional small molecular drugs primarily enter cells through passive or active process, whereas nanomedicines enter cells via endocytosis (Longfa et al., 2013). To evaluate the uptake process, we exploited temperature inhibition. Results in Figures 2D, F show that the uptake was inhibited at lower temperatures (4°C) in both cell lines. The fluorescence profiles of 4T1 and HCC-1806 cells are shown in Figure 2G. Notably, the majority of labeled-nanovesicles are intercepted in the cell membrane, providing evidence to suggest that the uptake likely occurred through endocytosis, facilitates efficient interactions of CLENs with recipient cells, resulting in profound biological effects.

We used the human cell line HEK-293 as a model to determine the specificity of nanovesicles towards TNBC cells compared to non-aggressive cell lines. The cells were incubated with indicated concentrations of CLENs for 24h, 48h and 72 h prior to bioluminescent assay (Figures 3A–C). CLENs exhibited a significant inhibition of cell viability in both TNBC cell lines, and the magnitude of the effect was time- and dose-dependent. In particular, the administration of 80 μg/mL of CLENs to 4T1 cells resulted in a cell viability of 85.1% and 81.2% after 24 and 48 h of incubation, respectively. Notably, upon extending the incubation period to 72 h, a substantial reduction of 73.4% in cell viability was observed. A similar trend was observed in the HCC-1806 cell line, wherein a 24-h incubation with the highest concentration of CLENs resulted in a cell viability of 81.9%, which slightly decreased to 81.8% after 48 h. However, the viability dropped further to 70.3% after 72 h of incubation. It is worth noting that no cytotoxicity was observed in HEK-293 cells after CLENs administration, highlighting the specific selectivity of CLEN for TNBC cells. Furthermore, to validate the results obtained from the bioluminescent assay, we conducted an additional Calcein-PI live-dead assay (Figures 3D–F). Notably, the administration of CLENs at the highest concentration (80 μg/mL) to 4T1 cells led to successive decreases in cell viability of 8.1%, 18.3%, and 20.8% after 24, 48, and 72 h of incubation, respectively. In the case of HCC-1806 cells, there was a decrease of 14.9%, 24.2%, and 25.9% following the respective incubation periods. Figures 3G, H illustrate the most representative images obtained from the Cal-PI live-dead assay after incubating with 40 and 80 μg/mL CLENs for 24, 48, and 72 h. The Supplementary Figure S3 provides images of live and dead cells showing the effect of administering different concentrations of CLENs. The results obtained from the CellTiter-Glo assay in line with the findings from the Cal-PI live-dead assay, suggest that CLENs have potential anticancer effects by inhibiting tumor cell proliferation. Notably, no cytotoxic effects were observed in normal Human cells HEK-293 cell line, implying CLENs themself possess anti-cancer properties specifically towards TNBC cells.

To elucidate the inhibitory effect of CLENs on the migratory behavior of 4T1 and HCC-1806 cells, we conducted a wound-healing assay. The results revealed a significant finding regarding the impact of CLENs on cell migration in TNBC cell lines. Following a 24-h incubation period in the absence or presence of 40 and 80 μg/mL of CLENs, a significant and dose-dependent inhibitory effect of CLENs on the migratory behavior of both 4T1 and HCC-1806 cell lines was observed (Figures 4A, B). Quantitative analysis was performed to assess the migratory cell population that infiltrated the wound area. Notably, upon exposure to CLENs at concentrations of 40 and 80 μg/mL, the migratory potential of the 4T1 cell line exhibited a substantial reduction of −87.1% and −93.8% respectively (Figure 4C). Equally remarkable, the HCC-1806 cell line demonstrated a pronounced inhibition of migration by −61.2% and −87.5% with the corresponding CLENs concentrations (Figure 4D). It is noteworthy that in human TNBC HCC-1806 cell line, the inhibitory effects reached 40.8% and 75.8% respectively, and this inhibitory effect persisted even after a 48-h. Furthermore, the inhibitory effects demonstrated remarkable persistence, as evidenced by the observation of 37.2% and 72.9% inhibition after a 72-h, indicating a prolonged and sustained impact on the migratory behavior of the cells (Supplementary Figure S3).

Next, we conducted a Matrigel drop evasion assay to assess the evasion and migration capability of TNBC cells. 3D structured Matrigel droplets prepared with 4T1 and HCC-1806 cells were incubated for 10 days in culture medium supplemented with or without CLENs. Notably, exposure to the presence of CLENs resulted in a significant inhibition of cell migration in both cell lines. In contrast, invadopodia formation was observed in the control group of cells, representing specialized protrusions that exert a pivotal role in facilitating the dissemination of cancer cells. In detail, the aggressive growth of 4T1 cells was distinctly observed after 7 and 10 days in the absence of CLENs, whereas the presence of 40 and 80 μg/mL CLENs effectively regulated cell evasion (Figure 5A). Migrated cell counts showed a significant increase, with levels in the control group soaring 5-fold and 7-fold relative to CLENs treatment on day 7 and 10, as shown in Figure 5C. Moreover, the control group showed cell aggregation areas that were 1.5 and 2 times larger than those observed in the groups treated up to day 7 and day 10, respectively (Figure 5D). Interestingly, we noticed that HCC-1806 cells have a unique proliferative behavior within 3D Matrigel droplets that differs from patterns observed in 4T1 cells under same conditions. The control group displayed a notable tendency of cell aggregation and strong intercellular interactions in the absence of CLENs. Conversely, the presence of CLENs effectively suppressed the formation of cell aggregates in HCC-1806 cells (Figure 5B). By days 7 and 10, the number of aggregated cells in the control group increased 5- and 6-fold compared to those treated with CLENs (Figure 5E). In terms of aggregate formation area, on days 7 and 10, the inhibitory effects of 40 μg/mL and 80 μg/mL CLENs on HCC-1806 cells were 1 and 2 times that of the control group, respectively (Figure 5F). A schematic diagram is shown in Figure 5G. Our findings provided compelling evidence supporting the inhibitory effects of CLENs on the metastatic potential of TNBC cells.

The phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT) and mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (ERK) signaling pathways are the key signaling pathways involved in the regulation of multiple cellular physiological processes by activating downstream corresponding effector molecules, which serve an important role in the cell cycle, growth and proliferation that mediate cancer cell survival and proliferation (Pascual and Turner, 2019). To delve into the molecular mechanisms underlying the anticancer effects of CLENs, we performed Western blot analysis on 4T1 and HCC-1806 cells, both untreated and treated with varying concentrations (40 μg/mL and 80 μg/mL) of CLENs over a 72-h period. Specifically, we examined the levels of total and phosphorylated forms of PI3K, AKT, and ERK. The representative images were shown in Figures 6A, C, with a dose-dependent downregulation of phosphorylated PI3K (p-PI3K), phosphorylated AKT (p-AKT), and ERK (p-ERK) in both 4T1 and HCC-1806 cells following exposure to 40 and 80 μg/mL CLENs (Figures 6B, D). Collectively, our findings suggest that CLENs exert suppressive effects on the PI3K/AKT and MAPK/ERK signaling pathways in both 4T1 and HCC-1806 TNBC cells.