Adipose-Derived Stem Cells Primed with Paclitaxel Inhibit Ovarian Cancer Spheroid Growth and Overcome Paclitaxel Resistance

Adipose-derived stem cells (ADSCs) primed with paclitaxel (PTX) are now hypothesized to represent a potential Trojan horse to vehicle and deliver PTX into tumors. We analyzed the anticancer activity of PTX released by ADSCs primed with PTX (PTX-ADSCs) (~20 ng/mL) in a panel of ovarian cancer (OvCa) cells sensitive or resistant to PTX. We used two (2D) and three dimensional (3D) in vitro models (multicellular tumor spheroids, MCTSs, and heterospheroids) to mimic tumor growth in ascites. The coculture of OvCa cells with PTX-ADSCs inhibited cell viability in 2D models and in 3D heterospheroids (SKOV3-MCTSs plus PTX-ADSCs) and counteracted PTX-resistance in Kuramochi cells. The cytotoxic effects of free PTX and of equivalent amounts of PTX secreted in PTX-ADSC-conditioned medium (CM) were compared. PTX-ADSC-CM decreased OvCa cell proliferation, was more active than free PTX and counteracted PTX-resistance in Kuramochi cells (6.0-fold decrease in the IC50 values). Cells cultivated as 3D aggregated MCTSs were more resistant to PTX than 2D cultivation. PTX-ADSC-CM (equivalent-PTX) was more active than PTX in MCTSs and counteracted PTX-resistance in all cell lines. PTX-ADSC-CM also inhibited OvCa-MCTS dissemination on collagen-coated wells. In conclusion, PTX-ADSCs and PTX-MSCs-CM may represent a new option with which to overcome PTX-resistance in OvCa.


Introduction
Improving the delivery of cancer therapies to tumor sites is fundamental to decreasing their negative side effects. In this context, mesenchymal stem cells (MSCs) have been proposed as cellular vehicles for targeted cancer therapies, thanks to their tumor homing properties. MSCs are present in many different mammalian tissues (adipose tissue, bone marrow, skin, umbilical cord blood, placenta, etc.) and are easy to isolate and expand [1].
Based on these characteristics, several laboratories set up models of engineered MSCs as vehicles for anticancer molecules, such as interferons, growth factors, chemokines and drugs [2,3]. Alternatively, the ability of MSCs to incorporate and release drugs in the conditioned medium (CM) or secretome was used against cancer cells [4][5][6].
MSCs from different sources, including adipose tissues (adipose-derived stem cells, ADSCs), without any genetic manipulation, acquire strong anti-tumor activity after priming with the chemotherapeutic drug paclitaxel (PTX) through their capacity to uptake and release drugs in

Drugs
Paclitaxel (PTX) (ACTAVIS) was a surplus drug from the clinical pharmacy of CRO Aviano.

Cell Culture
Ovarian cancer cell lines A2780 and A2780cis were from Sigma-Aldrich; SKOV3 (HTB-77) and COV318 were obtained from the American Type Culture Collection (ATCC); OVCAR5 (NIH) cells were provided by Dr. Baldassarre (CRO, Aviano, Italy); Kuramochi cells (JCRB0098), resistant to paclitaxel [15], were from JCRB Cell Bank; and OVCAR8 cells were from the National Cancer Institute Developmental Therapeutics Program Tumor Repository. All cell lines were routinely tested for mycoplasma, with negative results, and authenticated in our laboratory using PowerPlex 16 HS System (Promega, Madison, WI, USA) and GeneMapper ID version 3.2.1 to identify DNA short tandem repeats. ADSCs (Lonza) were maintained in Mesenchymal-Stem-Cell Growth Medium Bulletkit MSCGM (Lonza, Verviers, Belgium). Ovarian cancer cell lines were maintained in RPMI-1640 medium (Sigma-Aldrich Co., St. Louis, MO, USA) containing 10% heat-inactivated FBS, 100 U/mL penicillin and streptomycin (complete medium), at 37 • C in a 5% CO2 atmosphere. A2780cis cells were maintained in 1 µM cisplatin and cultured without the drug for 72 h before use in cellular assays. To obtain 3D-multicellular tumor spheroids (MCTSs), plates were coated twice with 20 mg/mL of poly(2-hydroxyethyl methacrylate) (poly-HEMA; Sigma) in 95% ethanol and washed once with PBS before cell seeding (2.0 × 10 4 cells in 24-well).

PTX-Uptake and Release by ADSCs in Conditioned Medium (CM)
Priming of ADSCs was performed as previously described [3,16]. Briefly, sub-confluent cultures of ADSCs were exposed to 2 µg/mL PTX for 24 h. Then cells were washed twice with PBS, detached and then seeded in a new flask. After 24 h of culture, CM was collected and PTX-primed ADSCs (PTX-ADSCs) were used to evaluate cell cycle phases and migration. Three separate samples from PTX-ADSCs were pooled together to analyze the concentration of released PTX [17] by LC-MS/MS, as described in Supplementary Materials; then PTX-ADSC-CM was used to perform experiments. ADSC-CM obtained in the same experimental conditions but without PTX loading was used as control.

Cultivation of Ovarian Cancer Cells as a 2D Model in the Presence of ADSC-CM
Cell lines (1.0 × 10 3 cells/well) were seeded in 96-well flat-bottomed microplates in 100 µL complete medium. Cells were allowed to adhere for 24 h, and then cultured with increasing concentrations of the reference drugs PTX (0-50 ng/mL) (free PTX), equivalent amounts of PTX released in PTX-ADSC-CM (0-100% v/v) (100% of PTX-ADSC-CM corresponds to 20 ng/mL secreted PTX) and ADSC-CM. After 7 days, cell viability was assayed using the MTT assay. Each experiment was conducted in triplicate. The IC50 values were calculated using the CalcuSyn software [18]. IC50-fold decrease was calculated as the ratio of the IC50 for free PTX to that of PTX secreted in PTX-ADSC-CM.

Cultivation of Ovarian Cancer Cells as a 3D Model (MCTSs) in the Presence of ADSC-CM
For cultivation, 1 × 10 4 MCTS ovarian cancer cells were plated into poly-HEMA-coated 48-well plates (non-adherent conditions) for 72 h in complete medium and cultured with increasing doses of free PTX (0-50 ng/mL) or of PTX-ADSC-CM (0-75%) for 7 days. Then viable cells were evaluated using PrestoBlue Cell Viability Reagent (Thermo Fisher Scientific, Frederick, MD, USA).
The antitumor activity of PTX released in PTX-ADSC-CM was compared to equivalent amounts of free PTX in preformed single MCTSs. Briefly, 1.0 × 10 3 SKOV3 cells were dispensed into poly-HEMA coated round-bottom 96-well. After 3 days, formed MCTSs (1 MCTS/well) were treated with increasing concentrations of PTX or PTX-ADSC-CM. Spheroid size was measured at day 7 after drug treatment initiation using an inverted microscope (Eclipse TS/100, Nikon, Tokyo, Japan) with photomicrographic systems DS Camera Control Unit DS-L2. Spheroid volumes were calculated using the formula: (width 2 × length × 3.14)/6 [21].

Migration (2D) and Dissemination/Invasion Assay
Cell migration was assessed using the in vitro scratch assay. Briefly, cells were grown to confluence and then treated with free PTX and PTX-ADSC-CM (equivalent to 5 ng/mL free PTX). After 72 h, monolayers were washed twice with PBS, scraped with a pipette tip to create a "wound" in the monolayer and washed again. Culture medium with 2% FBS was added and cells were cultured for 48 h. Wounds were photo-graphed using an inverted microscope (EclipseTS/100, Nikon, Instruments Europe BV Amsterdam, Netherlands) at magnification 4×. Migration was assessed by measuring the covered area (in pixels) with ImageJ-NIH (National Institutes of Health) tool software after 24 and 48 h. Tumor spheroid-based migration assay: this assay was performed in 96-well plates pre-coated with 10 µg/mL collagen I (Sigma Aldrich) and blocked with BSA (1 mg/mL) for 2 h [22]. Pre-formed tumor spheroids from SKOV3 and OVCAR8 cells were transferred (1 spheroid/well) into 96-well plates in the absence or in the presence of free PTX, ADSC-CM and PTX-ADSC-CM (equivalent concentrations). Image analysis software was used to calculate the area covered by migrated cells. The extent of migration was determined using Adobe Photoshop by outlining the entire area of the dispersed cells [23]. The fold-change (increase) in covered area was calculated dividing the pixel area of the spheroid at 48 h by the pixel area at time 0.

Statistical Analyses
Statistical analysis was carried out using GraphPad Prism version 6.0 software (GraphPad, LJ, USA), using the most appropriate test, as specified in each figure. Two sets of data differences were analyzed by Student's t-test. One-way ANOVA, followed by the Bonferroni correction, was used for multiple comparisons. One-way ANOVA, followed by Dunnett's test, was used to compare each of a number of treatments with a single control. p-value < 0.05 was considered significant.

The Effect of PTX Uptake on ADSCs' Activities
We evaluated PTX's effects on ADSCs viability ( Figure 1A, blue bar charts), proliferation ( Figure 1A, red bar charts), cell cycle phase distribution ( Figure 1B) and migration towards OvCa and CM from OvCa cells ( Figure 1C). We assessed the sensitivity of ADSCs to PTX in a 24 h cytotoxicity test and in an anti-proliferation assay after 7 d of treatment ( Figure 1A). ADSCs had low sensitivity to the cytotoxic and antiproliferative effects of PTX ( Figure 1A). A short incubation with PTX did not affect ADSCs viability and only the highest concentration of free-PTX (10 µg/mL) determined a reduction of proliferation of about 40% respect to control (untreated ADSCs) after 7 d of treatment ( Figure 1A). ADSCs primed with PTX (2 µg/mL for 24 h) (PTX-ADSCs) secreted about 20 ng/mL of the drug (Supplementary Figure S1). PTX-ADSCs showed a slight decrease of the S phase and an increase in the G2/M phase of the cell cycle with respect to control ADSCs ( Figure 1B). To evaluate migration of ADSCs after priming with PTX, we used as stimulus CM from SKOV3 cells (SKOV3-CM) or CM from SKOV3-MCTSs (SKOV3-MCTS-CM) and a layer of SKOV3 cells ( Figure 1C). All chemoattractants, SKOV3 cells and SKOV3-CM (from SKOV3 layer and SKOV3-MCTSs), increased the percentage of migrated ADSCs with respect to the control (medium). Priming with PTX (PTX-ADSCs) reduced the capability of ADSCs to migrate towards SKOV3 cells, SKOV3-CM and SKOV3-MCTS-CM ( Figure 1C).

PTX-ADSCs Inhibit OvCa Growth
To evaluate the anticancer activity of PTX-ADSCs, we cultured a fixed number of OvCa cells with increasing numbers of ADSCs and PTX-ADSCs (2D-direct contact) (Figure 2A). PTX-ADSCs decreased in a number-dependent manner tumor cell viability (Figure 2A). A2780 and its cisplatin-resistant clone A2780cis were the most sensitive to PTX secreted by PTX-ADSCs, as evaluated by the number of PTX-ADSCs capable of decreasing 50% of the cell viability ( Figure 2B). On the contrary, ADSCs alone did not affect cancer cell viability of OvCa cells (Figure 2A). ADSCs added to SKOV3-MCTS under non-adherent conditions, aggregate to form heterospheroids (3D model) (Supplementary Figure S2). The cytotoxic effects of PTX-ADSCs were confirmed in heterospheroids formed by SKOV3 cells and PTX-ADSCs. While the cultivation of SKOV3-MCTS cells with ADSCs did not affect cell viability, PTX-ADSCs (SKOV3/PTX-ADSCs heterospheroids) decreased in a number-dependent manner the percentage of viable cells in heterospheroids ( Figure 2C).

PTX-ADSCs-CM is More Active than Free PTX and Overcomes PTX-Resistance in 2D Cultures
Then we compared free PTX cytotoxicity to PTX released in PTX-ADSC-CM (20 ng/mL) (Supplementary Figure S1) by testing different amounts of CM in parallel to the respective free PTX concentrations. Under these experimental conditions, ADSC-CM did not affect OvCa cell proliferation ( Figure 3A), while both PTX-ADSC-CM and PTX decreased in a dose-dependent manner OvCa cell growth ( Figure 3A). PTX-ADSC-CM (equivalent free PTX concentration) was more active than free PTX, with the lower IC50 (ranging from 1.1 to 3.9 ng/mL) than that obtained with free PTX (ranging from 3.0 to 23.3 ng mL) ( Figure 3B). PTX secreted by PTX-ADSCs counteracted PTX resistance in Kuramochi cells [15]. This cell line demonstrated an IC50 of 23.3 ng/mL for PTX that decreased to 3.9 ng/mL with PTX-ADSC-CM (equivalent to free PTX concentrations). Overall, PTX-ADSC-CM decreased the IC50 for PTX in all cell lines tested ( Figure 3C) with fold decrease ranging from 1.6 (OVCAR5) to 6.3 (COV318) ( Figure 3C).

PTX-ADSC-CM Inhibits OvCa Cell Migration
The effect of PTX-ADSC-CM on OvCa cell migration was evaluated using the in vitro scratch assay (2D model). Treatment with free PTX, and especially with PTX-ADSC-CM, slowed the ability of SKOV3 and OVCAR8 cells to refill an empty area ("scratch") of the monolayer compared to untreated cells: 48 h after SKOV3 monolayer was scratched, the remaining uncovered area was about 52% in PTX-ADSC-CM, 32% in free PTX-pretreated cells, and almost totally covered in control cells ( Figure 4A,B); in OVCAR8 the uncovered areas were 58%, 31.5%, and 9% for PTX-ADSC-CM, free PTX and medium, respectively ( Figure 4C,D).

Discussion
Peritoneal carcinomatosis with formation of malignant ascites often characterizes the late stage of OvCa [12]. In ascitic fluid, exfoliated OvCa cells aggregate to form MCTSs and heterospheroids, composed by tumor cells and normal cells, including ADSCs [24,25]. Both MCTSs and heterospheroids in ascitic fluids are enriched in OvCa stem cells and contribute to drug resistance and spreading to secondary sites [24]. Thus, since MCTSs and heterospheroids mimic tumor growth in ascitic fluid and are considered effective first-line approaches to study in vitro drug activity, we used both cellular models to evaluate the anticancer effects of ADSCs primed with PTX (PTX-ADSCs).
Here, we found that the direct contact with PTX-ADSCs inhibited OvCa viability and overcame PTX resistance. Equivalent concentrations of PTX released by PTX-ADSCs were more cytotoxic than free PTX in 2D and 3D models (MCTSs) of OvCa; overcame intrinsic PTX resistance (Kuramochi cells); and overcame the resistance induced by culturing OvCa cells as MCTSs. Moreover, PTX released in PTX-ADSC-CM inhibited OvCa cell migration and dissemination of preformed MCTSs.
Primary debulking surgery followed by platinum and taxane chemotherapy is the standard treatment for advanced OvCa [12], but surgery can favor proliferation and invasion of residual cancer cells. We found that ADSCs primed with PTX released the drug and their cocultivation with OvCa cells was cytotoxic against both PTX and cisplatin resistant tumor cells. Thus, PTX-ADSCs used after surgery could counteract the proliferation/expansion of the residual drug-resistant cancer cells [24,26] and prevent the protective effects of ADSCs also in OvCa [25,26].
PTX-ADSCs maintained the capability to migrate towards molecules secreted by OvCa cells. PTX-ADSCs aggregated together with OvCa cells to form heterospheroids and inhibited both PTX and cisplatin-resistant MCTSs viability. These results suggest that PTX-ADSCs could be alternatively used during intraperitoneal chemotherapy (IP), where anticancer drugs are infused directly into the peritoneal cavity of patients, characterized by poor prognosis (stage III OvCa) and drug-resistance [12,27]. PTX efficacy is decreased by its low aqueous solubility and by multiple intrinsic or acquired drug resistance that can be mediated by different mechanisms, including the overexpression of the drug efflux transporter P-glycoprotein (Pgp) [28]. PTX released in CM from PTX-primed MSCs is secreted as free PTX and as PTX stored in EVs [5,6]. Our results suggest that PTX-ADSCs might provide PTX stored in EVs that, similarly to liposomal drugs, may cross plasma membranes more efficiently, and by increasing PTX accumulation, counteract Pgp-mediated efflux [29,30].
Most ovarian cancer patients' present disseminated disease at the time of their diagnosis, which is one of the main reasons for their poor prognosis. OvCa cells shed from the primary tumor into the peritoneal cavity [31]; they have to survive to anoikis, a programmed cell death due to detachment from extracellular matrix [32] and immune surveillance [33]. Upon successful detachment from primary tumor, OvCa cells can survive by forming MCTSs or heterospheroids and then metastasize predominantly to the omentum and peritoneum via a direct mechanism [31,34]. Conditioned medium from ADSCs primed with PTX decreased, more efficiently than free PTX, OvCa cell 2D migration and also OvCa-MCTSs dissemination onto collagen gels, suggesting that PTX secreted by PTX-ADSCs may be more active in reducing metastasis formation as well.
In conclusion, both PTX-ADSCs and PTX-ADSC-CM decreased OvCa growth as MCTSs, and as heterospheroids, overcame the intrinsic PTX resistance of Kuramochi cells, the extrinsic resistance induced by aggregation of tumor cells as MCTSs and inhibited OvCa cell migration/dissemination. Results of our study suggest that both PTX-ADSCs and PTX-ADSC-CM may represent new natural systems, absent the necessity to engineer MSCs, with which to overcome intrinsic and extrinsic PTX resistance in OvCa, and potential new cancer therapies based on cell drug delivery.
Supplementary Materials: The following are available online at http://www.mdpi.com/1999-4923/12/5/401/s1, Figure S1. LC-MS/MS quantification of taxol (PTX) released in PTX-ADSC-CM. Representative multi reaction monitoring (MRM) chromatogram for PTX quantification in control culture medium (A,A') and released in ADSC-CM (B,B'). Red and blue lines: are referred to MRM-chromatogram for docetaxel used as internal standard (IS) and PTX respectively. The normalized PTX signal after. Figure S2. Heterospheroid formation. Time-Lapse Imaging of ADSC internalization in SKOV3 spheroids by Phase-Contrast and Fluorescence Microscopy. Representative photomicrographs of four-day-old SKOV3 spheroids cocultured with fluorescent DiI stained ADSC cells (ratio 1:1) (red). ADSC internalization was followed by time-lapse imaging using the Leica DMI6000 B Microscope for 96 h. Figure S3