Doxorubicin resistance in breast cancer cells is mediated by extracellular matrix proteins

The MCF-7 (HTB-22™) and MDA-MB-231 (HTB-26™) cell lines utilised in this study were obtained from the American Type Culture Collection. Cells were incubated at 37 °C in a humidified incubator inclusive of 5% carbon dioxide in phenol red-free DMEM/F12 with 10% heat inactivated fetal bovine serum (Life Technologies).

The cell culture assays were performed as previously published [11, 14]. Briefly, for 3D cell cultures 1000 MDA-MB-231 cells or 5000 MCF-7 cells were seeded in 384-well microplates (CellCarrier; PerkinElmer) on top of 7.6 mg/ml Growth Factor Reduced (GFR) Matrigel™ or 7.5 mg/ml PuraMatrix™ (Becton Dickinson Biosciences). Once an average spheroid size of approximately 50-100 μm in diameter was reached (up to a 6-day incubation) drug was applied for a period of 6 days. For 2D monolayer assays, 600 cells per well were seeded into 384-well CellCarrier microplates. Following a 24-h incubation, cells were exposed to doxorubicin and incubated for 6 days. A dose-response curve for doxorubicin was performed with final concentrations per well between 0.0002 μM and 200 μM (3D: 12 point dose response curve, 2D: 20 point dose response curve). Doxorubicin (Tocris Bioscience) was stored at a concentration of 50 mM in dimethyl sulfoxide (DMSO) at −20 °C.

At the conclusion of both assays, a final concentration of 600 μM resazurin (Sigma-Aldrich) was added to each well and incubated for 4–6 h at 37 °C. The fluorescence intensity was detected using an EnVision™ multilabel plate reader (excitation 530 nm, emission 595 nm; PerkinElmer). In addition, 3D cell cultures were imaged using an Operetta™ High Content Imaging System (PerkinElmer) at the assay conclusion with and without the addition of 2 μM calcein AM dye (Life Technologies).

The cellular proliferation rates of breast cancer cell lines cultured in 2D and 3D were assessed via a resazurin reduction assay at 37 °C in a humidified incubator (5% carbon dioxide). A final concentration of 600 μM resazurin was added to each assay well at each specified time period throughout the assay duration, and fluorescence detected on an EnVision multilabel plate reader. Fluorescent intensity values were used to calculate cellular proliferation rates (based on a linear relationship between cell number and total fluorescent units).

To study the diffusion of doxorubicin into 3D cell culture spheroids, a 0.5-2 μM range of concentrations of the inherently fluorescent doxorubicin was applied to cells in 3D cultures and imaged. The total exposure time for breast cancer 3D cell cultures to doxorubicin prior to live cell imaging was between 6 and 72 h. Nuclei were stained with Hoechst 33,342 and incubated for 2 h at 37 °C in a humidified incubator inclusive of 5% carbon dioxide. Using the 10× objective on an Opera™ High-Content Screening System (PerkinElmer), Z-slices at 10 μm intervals through the spheroids were captured. Analysis of a central Z-slice of a spheroid was undertaken using the radial profile plugin in ImageJ (http://​rsbweb.​nih.​gov/​ij/​plugins/​radial-profile.​html). Values generated from ImageJ were graphed using Graphpad Prism.

To assess pro-survival protein expression in cells cultured in 2D and 3D conditions, assays were prepared as described above. Doxorubicin (2D: 90 nM; 3D: 720 nM) or the vehicle control were applied to cells for 24, 48 and 72 h time points. For immunostaining, 2D and 3D cell cultures were fixed for 10–20 min using 4% paraformaldehyde (PFA; Sigma-Aldrich) followed by washing with phosphate buffered saline (PBS; Life Technologies). Cell cultures were blocked with 2% Bovine Serum Albumin (BSA; Sigma-Aldrich) and 0.5% Triton X-100 (Sigma-Aldrich) in PBS. Anti-Bcl-xL (Cell Signaling) and anti-Bcl-2 (Life Technologies) primary antibodies in blocking buffer were added to wells and incubated at room temperature for 16–20 h. Wells were rinsed with PBS, followed by the addition of Hoechst 33,342 (Life Technologies), Alexa Fluor® secondary antibody (Life Technologies) and Alexa Fluor® phalloidin (300 units; Life Technologies) or Cell Mask (Life Technologies) in blocking solution for 2 h at room temperature. Cells were then washed with PBS before imaging.

Confocal fluorescent images were acquired from each assay well using an Opera High Content Screening System. Image analysis was performed with the Columbus™ Image Analysis and Data Management Software (PerkinElmer). In 2D cell cultures, expression levels of the protein of interest were measured by calculating the average intensity (in a region of interest) of a single image plane through a cell monolayer at 20× lens magnification. In 3D cell cultures, expression levels of the protein of interest were measured by calculating the average intensity (in a region of interest) of either a single central slice or a maximum projection image constructed from 10 μm confocal image stack at 10× lens magnification. Representative high magnification images of 3D cultures were acquired at 20× lens magnification to visualise protein localisation. All quantification from image analysis protocols were performed on unmodified files.

The characterisation of β1-integrin expression and morphological properties of the spheroid following inhibition were determined. β1-integrin function blocking antibody (P5D2; R&D Systems) and the equivalent isotype antibody for use in control wells (R&D Systems) were mixed into GFR Matrigel at a concentration of 1.5 μg/ml and layered into a 384-well microplate. The P5D2 antibody has previously been demonstrated to inhibit the function of β1-integrin at a concentration of 1.5 μg/ml [15]. Breast cancer cells were seeded into wells as described above. Cells were exposed to doxorubicin (720 nM; final concentration) for 72 h once average spheroids size reached 50-100 μm in diameter. Immunostaining, confocal imaging and analysis were performed as described above utilising the anti-β1-integrin antibody (R&D Systems) for the β1-integrin expression experiments. Live cell differential interference contrast (DIC) imaging was performed on a CellR microscope (Olympus Life Sciences) with image analysis completed via ImageJ Version 1.46 h (area; images were flat field corrected and then analysed [14] and the AxioVision™ software (longest diameter; ‘length tool’).

Fifteen microlitres of 7.6 mg/ml GFR Matrigel with 1.5 μg/ml of β1-integrin function blocking antibody or the equivalent isotype antibody for use in control wells (R&D Systems) were added to wells of a 384-well microplate. One-thousand MDA-MB-231 cells were added to each well and incubated at 37 °C in a humidified incubator inclusive of 5% carbon dioxide for 6 days (media changes after 3 days incubation). Doxorubicin was applied (0.07-5 μM) for a period of 72 h. At the assay conclusion, a final concentration of 600 μM resazurin was added to each well and incubated for 6–14 h at 37 °C in a humidified incubator inclusive of 5% carbon dioxide. The fluorescence intensity was measured by an EnVision multilabel plate reader. Representative DIC images were acquired on a CellR microscope.

A study was undertaken to evaluate doxorubicin resistance mechanisms exhibited by cells in a 3D ECM-based breast cancer model. Initially, experimentation was undertaken to ascertain if, and to what extent, culturing cells in 3D conditions impacted on doxorubicin activity. The potency (half maximal inhibitory concentration; IC₅₀ value), together with combined efficacy and potency (area under the curve; AUC) were measured.

Doxorubicin was significantly (p ≤ 0.001) more potent against the breast cancer cells grown in 2D cultures in comparison to those cultured in a 3D ECM-based model (Table 1). Furthermore, both MCF-7 and MDA-MB-231 cells exhibited significantly reduced (p ≤ 0.0001) efficacy upon doxorubicin application in 3D conditions in comparison to 2D culture (Table 1). Not only were there significant differences in the potency and efficacy of doxorubicin evaluated against breast cancer cell lines in 2D and 3D culture conditions, the shape of the MCF-7 dose-response curve demonstrated variances in the cellular response to drug in 3D cell culture compared to 2D cell culture (Fig. 1a). The morphological response to doxorubicin observed for the breast cancer cells in the 3D culture system indicated a substantial deterioration of the 3D cellular architecture at 10 μM (Fig. 1b). The data indicates that selected breast cancer cell lines cultured in 3D conditions are more resistant to doxorubicin in comparison to those cells cultured as 2D monolayers.

Investigation into the doxorubicin resistance observed in MCF-7 and MDA-MB-231 cell lines cultured in 3D was undertaken, with initial research conducted on the rates of cellular proliferation between cells cultured in traditional 2D monolayer and 3D cell cultures. Utilising a metabolic indicator dye, previously demonstrated to reflect cell number [14, 16], the number of cells per well under both culture conditions were measured at specific intervals (24 to 72 h) over 6 day (2D) and 9 day (3D) time frames. Outcomes demonstrated that cellular propagation occurred in both the 2D and 3D cell culture systems for both MCF-7 and MDA-MB-231 cell lines (Fig. 2a, b). The total well fluorescence intensity indicated a reduction in the doubling time for MDA-MB-231 (2D: 47.6 ± 10.2, 3D: 69.5 ± 7.2) and MCF-7 (2D: 55.2 ± 3.3, 3D: 190.9 ± 33.9; p ≤ 0.05) cells grown in 3D cell culture compared to those cultured on plastic substrata. Overall, there was a temporal increase in cell number for both breast tumour cell lines in both 2D and 3D culture conditions, and cellular proliferation was decreased in 3D cell cultures for both breast cancer cell lines tested.

The diffusion of doxorubicin within 3D cell cultures was examined to determine if limited cellular exposure was a contributing factor to the doxorubicin resistance observed in breast cancer cells. The diffusion of doxorubicin within 3D cultures of both the MCF-7 and MDA-MB-231 cell lines was investigated at 6, 24 and 72 h time points. Results demonstrate that doxorubicin was detected within both MCF-7 and MDA-MB-231 spheroids at the 6 h time point and the levels of doxorubicin were observed to increase within the breast cancer spheroids over time, particularly upon exposure to 2 μM doxorubicin (Fig. 3).

To investigate the impact cell attachment to the ECM had on cells grown in 3D culture with GFR Matrigel, a synthetic hydrogel, PuraMatrix, was employed to promote spheroid formation in the absence of specific ECM proteins (e.g. laminin, collagen IV). Doxorubicin was applied to spheroids cultured on GFR Matrigel and PuraMatrix and comparisons of the cellular response conducted. The potency of doxorubicin was significantly increased (p ≤ 0.05) against the MDA-MB-231 cell line cultured in 3D on PuraMatrix when compared to GFR Matrigel (Fig. 4a, c). However, there were no significant differences (p > 0.05) in drug sensitivity detected when the MCF-7 cells were cultured on the alternative matrix (Fig. 4b, d). The morphology of cells grown on PuraMatrix (Fig. 4e) was similar to those cultured on GFR Matrigel (Fig. 1b). Thus, the attachment of cells to selected ECM proteins may play a role in mediating drug resistance in MDA-MB-231 cells.

To investigate a potential survival advantage of cells cultured in 3D conditions in an ECM-rich microenvironment, two proteins integral in mediating cell survival, Bcl-2 and Bcl-xL, were examined. Results show that Bcl-2 and Bcl-xL were expressed in untreated and treated MDA-MB-231 cells at 24, 48 and 72 h time points. Following exposure to doxorubicin, there was a significant reduction (p ≤ 0.0001) in the levels of both Bcl-xL and Bcl-2 (48 and 72 h following application) with time in MDA-MB-231 cells cultured in 2D conditions (Fig. 5a, b). In addition, it was observed that Bcl-2 expression was localised to the nucleus of doxorubicin-treated cells following 48 and 72 h of exposure. The presence of Bcl-2 in the nucleus has been observed previously, and associated with a pro-apoptotic function [17, 18]. Conversely, the expression levels of the pro-survival proteins in doxorubicin-treated cells in 3D cultures were equivalent or greater than untreated cells, particularly at the 72 h time point (Bcl-2, p ≤ 0.0001; Fig. 5c, d). Overall, the cellular levels of pro-survival proteins observed in the 2D model were the opposite trend to the levels measured in the 3D ECM-based model.

The involvement β1-integrin signalling has on mediating the resistance of MDA-MB-231 cells grown in 3D cell culture in the presence of ECM proteins was investigated. As a result, the potential role of integrin signalling in doxorubicin resistance on MDA-MB-231 cells was evaluated. MDA-MB-231 cells express β1-integrin, the binding partner for various α-chain integrin heterodimers, in approximately equal quantities with and without the presence of doxorubicin (Fig. 6a, b). When β1-integrin to ECM protein binding was inhibited by inclusion of a function blocking antibody into the GRF Matrigel, morphological modifications were observed in 3D MDA-MB-231 cell cultures, including lack of spheroid integrity (Fig. 6c). The size of these treated 3D cell culture aggregates were measured and results show that inhibition with the β1-integrin function blocking antibody and/or doxorubicin resulted in significant (p ≤ 0.001) reductions in size (diameter and area) compared to the untreated control (Fig. 6d’, d”). To determine the therapeutic potential of blocking β1-integrin signalling in the presence of doxorubicin, the function of β1-integrin on MDA-MB-231 cells was blocked in combination with doxorubicin. Results show there was a dose-dependent enhanced (p ≤ 0.01) sensitivity of MDA-MB-231 cells in 3D culture to doxorubicin (Fig. 6e). These results suggest that blocking the function of β1-integrin prior to the addition of doxorubicin enhanced the efficacy of doxorubicin in a dose-dependent manner.