Enhancing Chemosensitivity of Breast Cancer Stem Cells by Downregulating SOX2 and ABCG2 Using Wedelolactone-encapsulated Nanoparticles

Prevalence of triple-negative breast cancers (TNBC), the most aggressive form of breast cancer, has increased substantially (1). Although different treatment strategies have been implemented, it remains a challenge in breast cancer management today because of the high occurrence of disease relapse (2). Tumor recurrence has been implicated to acquired chemoresistance and enhanced metastasis due to the existence of the cancer stem cells (CSC; ref. 2). CSCs retain the capacity for survival, self-renewal, and differentiation into tumorigenic cancer cells and are substantially insensitive to most conventional anticancer therapies (3), thereby suggesting that removal of CSCs is crucial for effective and more comprehensive cancer therapy. In addition, epithelial-to-mesenchymal transition (EMT) can induce enrichment and differentiation of cancer cells into CSC phenotypes (4), and plays an important role in invasion and metastasis (5) by modulating markers such as E-cadherin, cytokeratins, N-cadherin, and vimentin (6). Multiple signaling pathways govern these transitions by regulating the expression of crucial EMT-related transcription factors, such as Snail, Slug, and Twist (7). Thus, if cancers have to be eradicated, new selective treatments specifically targeting CSCs should be developed.

According to the World Health Organization, 80% people across the globe use medicinal plants for the treatment of cancer, because they are easily accessible, cost effective, and have less toxic side effects (8). Wedelolactone (7-Methoxy-5, 11, 12-trihydroxycoumestan, mol wt. 314.2), the principle active polyphenolic compound found in extracts of Wedelia calandulaceae and Eclipta prostrata (9) is known for treatment of liver diseases, viral infections, human bronchial epithelial cell injury, and snake bites (10). Wedelolactone also regulates osteoclastogenesis in breast cancer (11–13), inhibits androgen-independent prostate cancer (14), and endometrial and ovarian cancer cell growth (15). Wedelolactone holds great promise for development as an effective anticancer drug because it is widely used for prevention of inflammation and is mostly nontoxic to humans (16), although its effects on obliterating the CSC population has not been investigated till date. However, shortcomings such as poor solubility and bioavailability prevent clinical application of wedelolactone. Therefore, synthesized polymeric nanoparticles with efficient biodegradation and biocompatibility properties, and low antigenicity, are currently being tested to overcome these drawbacks and facilitate efficient delivery and effective functioning of the compound, specifically in the CSCs (17, 18). Poly-lactide-co-glycolide (PLGA) is one of the most effectively used polymers in nanomedicine, known to effectively deliver drugs into cells because of its biodegradable and biocompatible properties, and drug products containing PLGA have been approved for parenteral use by regulatory authorities (19). On the basis of this, we hypothesized that encapsulation of wedelolactone in PLGA nanoparticles will improve its efficacy by enhancing the anticancer effects, reducing toxic side effects, and eventually result in clinically favorable outcome (20). The process of nano-encapsulation not only protects poorly soluble and unstable payloads from the biological milieu but also is minute enough for capillary penetration, internalization, and endosomal escape (21). In addition, these particles have controlled release properties owing to pH and temperature sensitivity (22). Overall, PLGA-coated nanoparticles have significant advances over metal and other nonbiodegradable nanoparticles because they have been proven to be safe in clinical studies, are surface-tunable, and control the rate of polymer degradation and drug release (22, 23). This study describes the synthesis of wedelolactone-loaded nanoparticles (nWdl) for obliterating breast cancer stem cells (BCSC), with reduced side effects, improved pharmacokinetics, modified biodistribution, and enhanced functional efficacy, both in vitro and in vivo.

Human breast cancer cell line MDA-MB-231 was obtained from National Centre for Cell Sciences, Pune, India. They were initially validated by immunohistochemistry. Cells were maintained in complete DMEM supplemented with 10% FBS and 1% penicillin–streptomycin–neomycin, and incubated at 37°C in 5% CO₂. Cells were passaged every 72 hours and cells in log phase were used for subsequent experiments.

The solvent displacement method was applied to prepare bioactive PLGA-encapsulated nano wedelolactone (nWdl). Briefly, 50 mg of PLGA and 10 mg of wedelolactone was dissolved in 3 mL of acetone. The mixture was injected drop-wise into 20 mL of aqueous solution containing Pluronic nonionic surfactant F68 and stirred continuously at room temperature until the organic solvent evaporated. After removing the redundant F68 from the nanoparticles, the pellet was resuspended in Milli-Q water. Nanoparticle-loaded suspensions were stored at 4°C. For blank nanoparticles, similar method was deployed without the addition of wedelolactone.

For synthesis of FITC-tagged PLGA-loaded wedelolactone nanoparticles (FITC-nWdl), 50 mg PLGA and 2 mg of FITC (in DMSO) were dissolved in 5 mL of acetone. FITC-PLGA blank nanoparticles were centrifuged and the precipitate was dried under vacuum for 24 hours before determining the weight. A volume of 300 mL acetone was then added to completely dissolve the precipitate. Fluorescence intensity of FITC was measured and the standard calibration curve of FITC was used to quantify and measure the concentration of FITC in 1 mL of FITC-PLGA blank nanoparticle and FITC-nWdl solution (24).

DLS was carried out using a Zetasizer, Nano-ZS Instrument (Malvern Instruments). The intensity of scattered light was detected at 90° to an incident beam. The data was analyzed in the automatic mode (25). Measured size was presented as the average value of 20 runs, with triplicate measurements within each run. The size and shape of nanoparticles were analyzed by transmission electron microscopy (JEM-2100 HR-TEM, JEOL). The chemical properties of nWdl were characterized by Fourier Transform Infrared (FTIR) Spectroscopy (Perkin Elmer) with samples as KBr pellets.

Wedelolactone (2 mg) was dissolved in 2 mL acetone for complete extraction of wedelolactone into acetone for the loading and encapsulation estimations. The wedelolactone concentrations were determined at 351 nm (26). A standard plot of wedelolactone (0–100 μmol/L) was prepared under identical conditions. The loading content and encapsulation efficiency (EE) of wedelolactone were calculated according to the following formula: EE (%) = [(drug fed − drug loss)/(drug fed)] × 100.

The release kinetics of nWdl was determined by dispersing 50 mg of nWdl in 10 mL PBS following which nWdl release profile was determined (24). For in vitro cellular uptake studies, nWdl solution was added to MDA-MB-231 cells and incubated for 30 minutes to 8 hours, after which the nanoparticles present in the medium were removed (27). Cells were co-stained with propidium iodide and harvested. Samples were analyzed by flow cytometry (BD Accuri C6) or subjected to fluorescence imaging using FV 12000 (Olympus) at 330 nm, viewed with the Fluoview software.

To determine the release rate of nWdl in normal physiologic conditions, and to resolve whether proteins would impede drug release from the nanoparticle, the release kinetics was performed in 50% serum diluted in PBS. Release profile was compared at different time points (30 minutes to 72 hours) and the release percentage was calculated on the basis of O.D_351nm (24).

Cell viability was determined MTT assay. In brief, 5,000 cells/well were seeded and treated either with wedelolactone or nWdl for 24, 48, and 72 hours followed by addition of MTT. After incubation, 100 μL of detergent reagent was added to each well and absorbance intensity was recorded at 570 nm (SpectraMax 190 device microplate reader, Molecular Devices).

MDA-MB-231 cells were grown in 6-well plates. A linear wound was gently created in the monolayer. Cells were incubated for 24 hours in fresh media containing wedelolactone and nWdl. The wound closure was documented using an AxioVision Microscope (Carl Zeiss) at 20× magnification and captured from five randomly selected fields in each sample, at 0 and 24 hours. The migration rate of control cells was taken as 100% and healing rates of treated cells were compared with respect to control cells. Images were captured at 0 and 24 hours and the wound areas were calculated by NIH ImageJ software (28). The distance between the opposing edges of the wound was measured in micrometers (29).

Cells (2 × 10⁵) were seeded onto the porous membranes of BioCoat Matrigel invasion chambers (8-mm pore size). Treatment commenced after 24 hours with indicated concentrations of wedelolactone and nWdl. After 24 hours, the nonmigrated cells in the top compartment were wiped off gently. The cells in the insert were washed, fixed with 100% methanol, and the number of migrated cells in four quadrants was observed. Quantification was done in triplicates.

MDA-MB-231 cells were seeded in ultra-low attachment plates (Corning Inc) in serum-free DMEM/F12 supplemented with 5 μg/mL bovine insulin, 1× B27, 20 ng/mL EGF, 10 μg/mL heparin, 1% antibiotic–antimycotic solution, and 100 μg/mL gentamicin (30). They were treated without or with wedelolactone and nWdl till the appearance of primary spheres (P1) for 6 days. Mammospheres with diameter ≥50 μm were counted manually. To assess sphere numbers during secondary (P2) passages, mammospheres (P1) were collected on day 6, dissociated with 0.05% trypsin, filtered using a 40 μmol/L sieve, and replated in ultra-low attachment plates, with no additional treatments (31). Treatments carried out in quadruplicates were determined from at least three independent experiments.

For each condition, 10⁶ cells were suspended in DMEM supplemented with 5% FBS. After addition of rhodamine123 (Rh123) with or without the inhibitor verapamil, cells were incubated in the dark for 30 minutes at 37°C. IC₅₀ concentrations of nWdl were used for the assay. Cells were allowed to efflux in substrate-free media for 1 hour, centrifuged, resuspended in ice-cold PBS, and kept on ice until further analysis. Cells were gated for forward versus side scatter and the geometric mean of fluorescence intensity (cellular uptake of fluorescent substrate) was recorded for a total of 1 × 10⁴ cells using a flow cytometer (BD Biosciences) under the green excitation emission wavelengths. All data were analyzed using FlowJo software (Tree Star, Inc.).

Total RNA was prepared from adherent MDA-MB-231 cells and mammospheres collected on day 6 (P1), and converted into cDNA. After initial denaturation at 95°C for 1 minute, PCR was performed for 40 cycles (15 seconds at 95°C and 45 seconds at 60°C) using KAPA SYBR FAST Universal qPCR Kit (Kapa Biosystems; ref. 31). The primers used for gene expression analyses are presented in Supplementary Table S1. 18S mRNA was used for normalizing RNA and fold change was calculated on the basis of vehicle-treated normalized values for each transcript.

Monolayer cultures of MDA-MB-231 cells, 6-day mammospheres, and tissue samples were lyzed using 100 μL of RIPA buffer. After resolving by PAGE and transfer, the membranes were blocked with 5% BSA. Next, they were incubated overnight with primary antibodies at 1:1,000 dilution at 4°C followed by secondary antibodies for 1 hour at room temperature. Protein band intensities were visualized by an Enhanced Chemiluminescence Detection System (Optiblot ECL Detect Kit, Abcam) in Bio-Rad ChemiDoc and relative quantification was done using the ImageJ software (31).

To enrich the CSCs, 1 × 10⁷ cells/mL were resuspended in Hank's Balanced Salt Solution containing 2% FBS and 100 mmol/L HEPES and stained with primary antibodies (CD24-FITC and CD44-PE; 1:100 dilution) or isotype controls for 15 minutes at room temperature. Cells with CD44⁺/CD24^−/low phenotype were plated for mammosphere formation at a seeding density of 1 × 10⁴ cells per well. P1 mammospheres were collected for gene expression analyses by qRT-PCR (31).

Cells (1 × 10⁶)/mL cells were suspended in Aldefluor assay buffer containing ALDH substrate (bodipy-aminoacetaldehyde) and incubated for 45 minutes at 37°C. As a reference control, the cells were incubated in the presence of diethylaminobenzaldehyde (DEAB), a specific ALDH1A1 enzyme inhibitor. The brightly fluorescent ALDH1A1-expressing cells (ALDH1A1^high) were detected in the green fluorescence channel (520–540 nm) of FACS Aria III (BD Biosciences) and ALDH1A1^high populations were sorted out (32).

The untreated and treated MDA-MB-231 cells (10⁶ cells/mL) were resuspended in wash buffer and incubated in the presence of antibodies against CD44-PE, CD24-FITC, and their corresponding isotype controls at 4°C in the dark for 40 minutes. Subsequently, the cells were washed, resuspended in FACS buffer, and processed using BD FACS Accuri C6 (31). The results were analyzed using BD Accuri C6 software (BD Biosciences).

All animal experiments were conducted as per the approval and guidelines of the Institutional Animal Ethical Committee, Government of India (registration no. 885/ac/05/CPCSEA). Female Swiss Albino mice, body weight 20–22 g, from Central Research Institute, (Kolkata, India) were housed in polypropylene cages under standard laboratory conditions of 50% ± 10% relative humidity, 22 ± 2°C temperature, and 12/12 light–dark cycle for 10 days prior to commencement of experiments. Mice were fasted overnight and water was given ad libitum before experimentation. Food was restored after injections but withdrawn 6 hours prior to sacrifice and analysis.

Solid tumors were developed in animals (n = 5) except in the normal group (Group I) or normal mice treated with nWdl (5 mg/kg body weight; Group II) by intraperitoneal injection of 2 × 10⁶ viable cancer cells (0.2 mL) in the mammary fat pad of mice. Palpable tumors were allowed to develop for 14 days. Same set of experiment was repeated in the left flank of mice and tumor was developed for 14 days. Mice were subsequently divided into three groups, each containing five animals, viz., (i) untreated tumor control (no drug treatment; Group III), (ii) wedelolactone-treated tumor-bearing mice (5 mg/kg body weight; Group IV), and (iii) nWdl-treated tumor-bearing mice (5 mg/kg body weight; Group V). Groups II, IV, and V animals were subjected to a treatment schedule for 7 days. Nanoparticle (nWdl) toxicity was assessed in normal animals (Group II) and results of other groups were normalized to the values of those observed in Group I. After 7 days, the tumors were excised out, weighed in a pan balance, measured with slide calipers, and processed for Western blot analysis.

The immediate toxic impact of nanoparticles was elucidated from interactions with different blood cells in vivo (33), such as red blood cells (RBC), white blood cells (WBC, total and differential), and hemoglobin content, of normal mice and mice with tumors. Plasma drug levels were estimated by UV spectral scan (300–400 nm). For in vivo retention and biodistribution of nanoparticles, animals were injected with 5 mg/kg FITC-tagged nWdl on day 0 and sacrificed after days 3, 5, and 7. Tumors and livers were collected, perfused in PBS, and subjected to flow cytometric analysis for presence of FITC-tagged nanoparticles. Furthermore, tumors were processed for Aldefluor assay and immunophenotyping to determine any changes in the CSC population due to drug retention.

Values are shown as SEM. Data were analyzed; when appropriate, significance of the differences between mean values was determined by using Student t test and one-way ANOVA by post hoc testing. The software used for the analysis was SPSS 14.0. Results were considered significant at a P value of not more than 0.05. For Student t test, *, P < 0.05; **, P < 0.01 and ***, P < 0.001 were considered significant. For sample size less than 10, nonparametric Mann–Whitney U test was performed.

Wdl-encapsulated PLGA nanoparticles (nWdl) were formulated by varying the ratio of poly-glycolic acid and poly-lactic acid, and best composition was found to be 50:50 ratio of each. Results of DLS studies showed that nWdl had a mean hydrodynamic diameter of 95 ± 0.34 nm (Fig. 1A) with polydispersity index (PDI) 0.77 ± 0.065 and zeta potential value −8.5 ± 2.35 mV (Fig. 1B). We observed a uniformly narrow distribution of PDI value and negative zeta potential, which prevents particle aggregation. Transmission electron microscopy revealed spherical structures coated with clearly distinguishable PLGA biopolymer. The average size of the nanoparticles was 60–80 nm (Fig. 1C). nWdl was stable at pH 6–7 (Fig. 1D). The nanoparticles showed 83.3% ± 2.15% encapsulation efficiency which indicated that majority of wedelolactone was entrapped into the PLGA-coated nanoparticles during synthesis. FTIR spectra, shown in Fig. 1E, indicated strong peaks in the range of 1,000–1,500 cm⁻¹ and showed the stretching mode of wedelolactone which had prominently disappeared in the spectrum of polymer (nWdl), thus confirming polymerization.

Cells were seen to take up the nanoparticles as early as 30 minutes (Fig. 2A) and significant time-dependent increase in cellular uptake was observed up to 2 hours (Fig. 2B). In addition, cytosolic accumulation of nWdl was seen within 30 minutes whereas nuclear translocation was observed to be significant by 60 minutes. Uptake within both compartments was maximum by 8 hours (Fig. 2C), after which no further uptake was observed.

Sustained release of wedelolactone from PLGA was observed over a period of 3 days at pH 7.4 and 5.0, where 72.75% and 89.74% of wedelolactone was released, respectively (Fig. 2D). The pH sensitivity during drug release by the nanoparticles implicated efficiency in reducing their side effects while in circulation and enhancing specific release in tumors which usually have an acidic pH. In vitro drug release was also estimated in PBS containing 50% serum. The kinetics in PBS-containing serum indicated a similar profile of drug release as in PBS only, further confirming that presence of serum proteins did not impede or alter drug release from nanoparticles (Supplementary Fig. S1).

Cell viability assays indicated that wedelolactone exhibited an IC₅₀ value of 80 ± 3.59 μg/mL, whereas the nano-encapsulated form showed an IC₅₀ value of 20 ± 1.94 μg/mL after 24 hours of incubation (Fig. 2E, a), which did not change significantly after 48 and 72 hours of incubation (Fig. 2E, b and c). Effects of nanoparticles assessed in two other breast cancer cell lines, MDA-MB-468 and MDA-MB-435, indicated insignificant differences in the IC₅₀ doses between different cell lines (Supplementary Fig. S2).

To reduce interference in anti-invasive evaluation, all the assays were performed with half IC₅₀ concentrations (40 μg/mL for wedelolactone and 10 μg/mL for nWdl) for 24 hours. Untreated MDA-MB-231 cells exhibited prominent wound closure activity, whereas wedelolactone interfered with invasion of MDA-MB-231 cells (P < 0.001; Fig. 3A). However, compared with the free drug, nWdl was more effective (distance = 130 ± 0.125 nm; P < 0.001) in retarding invasion of the cells at a much lower concentration than wedelolactone (distance = 197 ± 0.82 nm; P < 0.001). To further confirm our results, transwell migration assays revealed that significantly fewer number of cells migrated to the lower surface of the transwell inserts when treated with nWdl (61.7% ± 2.94%; P = 0.002) compared with wedelolactone treatment (81.4% ± 0.73%; P = 0.012; Fig. 3B).

Analysis of cell death revealed that nWdl more effectively upregulated expression of proapoptotic markers and downregulated expression of antiapoptotic markers (Fig. 3C). Furthermore, in contrast to wedelolactone, nWdl significantly downregulated protein levels of mesenchymal markers, like N-cadherin, vimentin, Twist, Snail, and Slug and upregulated the epithelial markers E-cadherin and cytokeratin-19, both at the transcriptional (Fig. 3D) and translational (Fig. 3E) levels.

To determine whether reduction in cell migration is related to obliteration of the CSC population, we analyzed the percentage of ALDH-positive (ALDH⁺) cells when MDA-MB-231 cells were treated with 40 μg/mL wedelolactone or 10 μg/mL nWdl. The effects were compared with paclitaxel, a drug conventionally used for treatment of patients with breast cancer, because earlier studies in our laboratory have confirmed that chemo-treatment enriched the CSC population, increasing the risk of disease relapse (31). Paclitaxel (2 nmol/L) treatment led to an increase in ALDH⁺ BCSCs by 5.05-fold (32% ± 1.75; P = 0.019) relative to untreated cells. However, expression of ALDH⁺ cells significantly decreased by 1.3-fold when treated with paclitaxel + wedelolactone (24.2% ± 1.48%; P = 0.005) and 2.6-fold when treated with paclitaxel + nWdl (12.1% ± 1.02; P = 0.021), relative to paclitaxel alone (Fig. 4A). In support of the above, in vitro uptake studies specifically in mammospheres showed that the nanoparticles were effectively taken up by the spheres within 2 hours of treatment (Fig. 4B).

The above findings were subsequently validated in both primary and secondary mammospheres formed from MDA-MB-231 cells (Fig. 4C). In the primary spheres, 20 μg/mL nWdl led to a significant reduction of ALDH⁺ cells by 3.2-fold (P = 0.018) and 2.5-fold (P = 0.014) relative to untreated control and wedelolactone treatment, respectively (Fig. 4D), and preferentially reduced the CD44⁺/CD24^−/low cells by 2.1-fold (P = 0.014) relative to control and 1.5-fold (P = 0.006) relative to the free drug (Fig. 4E). Furthermore, nWdl significantly inhibited formation of spheroids (Fig. 4F), reduced the size of the mammospheres by 2.6-fold (P = 0.011) and 1.4-fold (P = 0.005; Fig. 4G), and decreased the number of mammospheres by 2.9-fold (P = 0.016) and 1.8-fold (P = 0.006; Fig. 4H) compared with control and wedelolactone treatment, respectively.

Concomitantly, in secondary mammospheres, nWdl reduced the percentage of ALDH⁺ cells by 4-fold (P = 0.025) and 2.7-fold (P = 0.014) compared with untreated and wedelolactone-treated cells, respectively (Fig. 4I). Furthermore, cells that initially formed primary spheroids in the presence of nWdl did not form secondary spheres compared with the control (Fig. 4J). In addition, nWdl reduced the size of secondary mammospheres by 2.6-fold (P = 0.014) and 1.5-fold (P = 0.006; Fig. 4K), and the number by 3.2-fold (P = 0.017) and 2.1-fold (P = 0.009) relative to control and wedelolactone, respectively (Fig. 4L).

Effects of wedelolactone and nWdl on the viability of spheroids indicated that both wedelolactone and nWdl inhibited cell viability of primary (Fig. 5A) and secondary mammospheres (Fig. 5B). Concomitantly, the sensitivity of the spheres was considerably enhanced for nWdl (IC₅₀ 20 μg/mL for primary spheres, Fig. 5A; and IC₅₀ 10 μg/mL for secondary spheres, Fig. 5B) compared with wedelolactone (IC₅₀ 80 μg/mL for primary spheres, Fig. 5A; and IC₅₀ 40 μg/mL for secondary spheres; Fig. 5B). This finding was further ascertained by reduced drug efflux, as indicated by retention of rhodamine in mammospheres (Fig. 5C), and attenuated expressions of ABCG2 and ALDH1A1 at mRNA transcript (Fig. 5D) and protein (Fig. 5E) levels.

Expression of the transcription factors Nanog, Sox2, and Oct4 in mammospheres indicated that nWdl inhibited their expression measured by qRT-PCR (Fig. 5F) and Western blot analysis (Fig. 5G). Interestingly, nWdl treatment significantly reduced Sox2 level (P = 0.004), whereas wedelolactone did not show any effect on Sox2 mRNA and protein levels. Protein expressions of Sox2, ABCG2, and ALDH1A1 were also examined in MDA-MB-468 and MDA-MB-435 cells, where reduced expression of the proteins was observed on nWdl treatment, similar to MDA-MB-231 cells (Supplementary Fig. S3).

Treatment of BCSCs with nWdl resulted in significant downregulation of mesenchymal markers, concomitant with significant upregulation of the epithelial marker, E-cadherin, by 5.2-fold (P = 0.0001) and 1.5-fold (P = 0.003) compared with control and wedelolactone treatment, respectively, at transcript levels (Fig. 5H) and protein levels (Fig. 5I). Protein expression of Slug was also examined in MDA-MB-468 and MDA-MB-435 cells, where reduced expression of the protein was observed on nWdl treatment, similar to MDA-MB-231 cells (Supplementary Fig. S3).

The anti-invasive effects of wedelolactone are known to be mediated through NFκB signaling (34). Subsequently, analysis of 20 μg/mL nWdl in BCSCs downregulated the expression of nuclear NFκB protein in BCSCs compared with wedelolactone (Fig. 5J), indicating that anti-invasive effects of nWdl on BCSCs are mediated by the inhibition of NFκB.

As observed for MDA-MB-231 adherent cells, treatment of mammospheres with 2 nmol/L paclitaxel in combination with 20 μg/mL nWdl significantly reduced their size by 3.1-fold (P = 0.021) and 1.5-fold (P = 0.003) relative to the control spheres and the paclitaxel + wedelolactone-treated spheres, respectively. Similarly, where paclitaxel treatment alone enhanced the number of mammospheres by 1.4-fold (P = 0.006), nWdl reduced the number of mammospheres by 5.2-fold (P = 0.026) relative to the control spheres (Fig. 6A). These mammospheres barely produced any visible secondary mammospheres after drug withdrawal, compared with the control. In addition, combination of nWdl and paclitaxel reduced the expression of all pluripotency and chemoresistance markers relative to either the paclitaxel-treated group, or the wedelolactone and paclitaxel-treated group (Fig. 6B). Furthermore, compared with control spheres, significantly higher ALDH⁺ cells were observed when mammospheres were treated with paclitaxel alone whereas treatment with paclitaxel + nWdl markedly reduced the percentage of ALDH⁺ population by 12.5-fold (P = 0.051) relative to paclitaxel-treated and 7.3-fold (P = 0.032) relative to paclitaxel + wedelolactone-treated spheres, respectively (Fig. 6C).

On the basis of the pilot studies for dose determination, 5 mg/kg body weight of wedelolactone and nWdl were used for treatment of tumor-bearing Swiss albino mice. It was observed that treatment with wedelolactone and nWdl significantly reduced tumor size by 1.5-fold and 2.3-fold, respectively, compared with that of the untreated mice (Fig. 6D). Interestingly, tumor size and weight reduction was more prominent after treatment with nWdl (P < 0.001) rather than wedelolactone (P < 0.05; Fig. 6D) relative to untreated control. Assessment of apoptosis in the tumors revealed that nWdl significantly upregulated expression of proapoptotic markers (Fig. 6E). This clearly shows that the drug in its nano-encapsulated form was able to suppress metastasis through enhanced cell death. Concomitantly, compared with wedelolactone, nWdl treatment caused significant decline in the CD44⁺/CD24^−/low population harbored in tumors, indicating that nWdl effectively obliterated the BCSC population within a tumor (Fig. 6F). Effects of nWdl on solid tumors developed in left flank also showed marked reduction in tumor volume, size, and weight, and decline in the CD44⁺/CD24^−/low population (Supplementary Fig. S4).

The effect of nWdl on hematologic parameters indicated that nWdl did not affect the RBC and WBC counts or the hemoglobin (Hb) level in normal mice even after 7 days of treatment (Supplementary Table S2). The hematologic parameters however significantly differed in mice bearing solid tumors, as compared with the control group. Significant decrease in the Hb content (by 2.1-fold; P < 0.05), RBC count (by 1.3-fold; P < 0.05), lymphocyte (by 5.8-fold; P < 0.05), and monocyte count (by 0.3-fold), along with a sharp increase in total WBC (by 4.2-fold; P < 0.05) and neutrophil counts (by 3.6-fold; P < 0.05) were observed in the mice bearing solid tumors. Treatment with wedelolactone could partially restore the values of Hb, RBCs, and WBCs, (Supplementary Table S2), whereas treatment with nWdl was more efficient in normalizing the levels of these parameters as compared with wedelolactone alone (P < 0.05).

When mice bearing tumors were treated with nWdl and drug retention was assessed for up to 7 days, it was observed that 3.71% ± 1.49% of nWdl was present in the tumor within 2 hours (Supplementary Fig. S5A) and almost 8% ± 2.71% of the nano-encapsulated drug was retained within the tumor till day 7 (Fig. 6G; P < 0.05). Comparative drug levels in plasma assessed by absorbance spectrum with peak absorbance at 351 nm, corresponding to wedelolactone, indicated baseline concentrations of the released drug in the plasma at days 3, 5, and 7, as assessed by lack of absorbance peak in the samples (Supplementary Fig. S5B). Retention was also assessed in liver and the results indicated insignificant changes on days 3, 5, and 7 (Supplementary Fig. S5C).

A time- and day-wise comparison of obliteration of the BCSC population by nWdl in mice tumors indicated significant and progressive reduction of ALDH⁺ population by 1.74-fold within 12 hours, 2.02-fold with 24 hours (Fig. 6H). Reduction by 7.7-fold was observed by day 3 (Fig. 6I). There was no significant difference in ALDH⁺ cell population between days 3 and 5 but a further 4.6-fold reduction was observed on day 7 (Fig. 6I).

Drug efflux, a major caveat in the management of breast tumors often leading to treatment failures, has been implicated to CSCs which overexpress efflux pumps (31). Persistence of CSCs after chemotherapy mostly leads to disease recurrence, rendering these cells as major therapeutic targets. Designing novel strategies, such as drug nanoformulation, would increase retention and sustained release of the concerned drug within the CSCs and eventually better patient prognosis. Size compatibility of PLGA nanoparticles facilitates intracellular internalization and eludes lysosomal compartments, avoiding degradation of the drug and reduction in drug efficiency (35). There are three mechanisms of drug-loaded nanoparticle incorporation into CSCs (i) caveolin-mediated endocytosis, (ii) clathrin-mediated endocytosis, and (iii) passive transport (36). Anticancer drugs encapsulated in nanoparticles can eventually actively or passively target the CSCs, because sustained drug release in the cytoplasm improves therapeutic effect at the target site (35). Such modifications additionally reduce systemic toxicity of chemotherapy drugs and bypass certain forms of multidrug resistance (36). In contrast to other formulations like microparticles, hydrogels, and implants, PLGA nanoparticles can be functionally modified to provide multiple efficacies including controlled drug release, cell-specific targeting, and increased cellular uptake (37). This work emphasizes that spatiotemporal chemistry utilizing such functionalities can be used to treat breast cancer, so that sustained release and continuous exposure to the encapsulated drugs can effectively obliterate CSCs, and eventually lead to better patient prognosis in future.

PLGA-based nanoparticles demonstrated several advantages over other biopolymers for drug delivery because of their small size, which helps them penetrate specific tissues via the fenestrations present in the tumor endothelial cells. This allows not only increased time in circulation but specific delivery and enhanced cellular uptake of the encapsulated drugs by their target tissues, thereby reducing side effects (38). This technology will generate more effective therapies capable of overcoming several biological barriers and side-effects that the body encounters during treatment with conventional anticancer drugs. Nanoformulation of wedelolactone not only enhanced its biodistribution and bioavailability, along with sustained release of the drug in the CSCs, it further sensitized them to the therapeutic effects of paclitaxel, possibly by downregulation of the ABCG2 drug efflux pumps. Consequently, increased stability of nWdl, due to their entry into cells through endocytosis and delayed hydrolysis, helped them escape the ABCG2 pumps, thereby overcoming drug resistance. The main advantage of these nanoparticles is that they can facilitate gradual and more effective drug release, thereby prolonging the life of drugs in circulation and increasing their half-life in plasma (39).

Metastasis is characterized by a transition from epithelial-to-mesenchymal phenotype. Circulating tumor cells in TNBC metastasis are reported to exhibit mesenchymal characteristics (40) and the gene signatures of these mesenchymal-type cells induced by EMT are highly correlated with tumor aggressiveness and chemoresistance (41). We demonstrated that nWdl could inhibit EMT by triggering molecular reprogramming in MDA-MB-231 cells, with enhanced efficiency compared with wedelolactone. In addition, nWdl modulated “cadherin switching,” by increasing the expression of E-cadherin and reducing the expression of N-cadherin (42, 43). Snail and Slug, which are known to regulate expression of E-cadherin in TNBC cells (44), were also regulated by nWdl. Overall, the processes of tumor metastasis and invasion, which appear to be inextricably linked to EMT, was effectively reduced by nWdl in human BCSCs in vitro. In addition, a clear indication of apoptosis induced by the drug and enhanced by nanoformulation of the drug supports reduction of metastatic potential after treatment, both in vitro and in vivo.

A link between the acquisition of molecular and functional traits of CSCs and the induction of EMT has several implications in tumor progression (45). This study demonstrated for the first time that nWdl can target both the bulk tumor population, as well as the BCSCs, by inhibiting their self-renewal capacity and their ability to successfully connect to different metastatic niches. Interestingly, although paclitaxel enriches the BCSC population (31), nWdl reduced the BCSC population when treated alone and more effectively in combination with paclitaxel. In addition, nWdl reduced key regulators of BCSC self-renewal and pluripotency (46). We have already reported that SOX2 is significantly over expressed in most of the TNBC cases in humans (31). We have also knocked down and overexpressed SOX2 to further confirm that SOX2 played a major role in metastasis and CSC self-renewal. Silencing SOX2 could induce chemosensitivity in the BCSCs, which were otherwise chemoresistant (31). Here, a remarkable finding was that, unlike the other stemness markers, wedelolactone was unable to alter SOX2 expression; however, the nano-encapsulated form of the drug could significantly reduce the expression of SOX2 in mammospheres. Therefore, nWdl treatment could be considered a novel approach to sensitize the BCSCs, which were otherwise resistant to conventional chemotherapy drugs, like paclitaxel, and eventually prevent initiation of secondary tumors after a chemotherapeutic regime. In addition, NFκB which is postulated to play a major role in EMT and invasiveness of TNBCs (47, 48) was significantly reduced by nWdl, conforming that reduced NFκB expression possibly helps suppress invasiveness and migratory properties of CSCs present in MDA-MB-231 cells.

The effect of nWdl on solid tumors (both in the mammary fat pad and in the flank) in Swiss albino mice confirmed that drug-loaded nanoparticles did not affect the hematologic parameters of normal mice, indicating minimum toxicity of these particles in vivo. However, nWdl significantly and specifically reduced tumor weight and volume, and improved hematologic parameters of tumor-bearing mice to levels observed in normal mice. Anemia and myelosuppression are evident side effects of cancer chemotherapy (49). Results indicated that tumor bearing mice developed anemia due to reduced Hb and concomitant hemolytic conditions. Administration of both wedelolactone and nWdl reinstated the Hb content, along with restoration of hematologic parameters towards normalcy, although nWdl was significantly more effective. Dose equivalence of nWdl in humans was calculated using the method of Reagan-Shaw and colleagues (50), which indicated that only 0.4 gm of nWdl will be more effective in simulating anticarcinogenic effects compared with an equivalent dose of wedelolactone. That the BCSC population in solid tumors can be effectively reduced by nWdl has furthered credence to obliteration of metastasis and disease recurrence. Therefore, formulation of PLGA-based coumestan nano-therapeutics for targeting BCSCs will eventually provide a more effective and complete therapy for patients with TNBC. Lower toxicity and sustained release profile of these nanoparticles decipher significant reduction in size, number, and metastatic potential of mammospheres, thereby rendering nWdl suitable for development as an effective chemotherapeutic drug in future.

No potential conflicts of interest were disclosed.

Conception and design: S. Das, U. Chatterji

Development of methodology: S. Das, P. Mukherjee, R. Chatterjee, U. Chatterji

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S. Das, P. Mukherjee, R. Chatterjee, Z. Jamal

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Das, Z. Jamal, U. Chatterji

Writing, review, and/or revision of the manuscript: S. Das, P. Mukherjee, U. Chatterji

This work was funded by the DS Kothari Fellowship to S. Das from University Grants Commission, Government of India [F.4-2/2006(BSR)/BL/14-15/0072] and partially supported by a grant to U. Chatterji from the Department of Biotechnology, Government of India [BT/PR5731/MED/31/165/2012]. We are grateful to the Centre for Research in Nanoscience and Nanotechnology, University of Calcutta, India, for instrument support. The generous help of Sudeshna Mukherjee, Department of Physiology, University of Calcutta, with the hematologic experiments is also humbly acknowledged. The authors also acknowledge support from DST-FIST, UGC-SAP, and DST-PURSE programs in the department for instrument and infrastructure support.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.