Sublethal doxorubicin promotes migration and invasion of breast cancer cells: role of Src Family non-receptor tyrosine kinases

MCF7, MDA-MB-231, and SKBR3 breast carcinoma cells, HeLa cervical adenocarcinoma cells, U-2OS osteosarcoma cells are from American Type Culture Collection (ATCC, Manassas, VA). ZR75-1 breast carcinoma cells are a gift from Dr. Natalia Marchenko (Stony Brook University). RPMI, DMEM, F12/DMEM, and McCoy’s 5A culture medium, fetal bovine serum (FBS), and superscript III reverse transcriptase (RT) are from Life Technologies (Carlsbad, CA). Bio-Rad protein assay was from Bio-Rad (Hercules, CA). Antibodies for total PARP, total Src, phospho-Src (Y416), Yes, p53, ATM, ATR, E-cadherin, n-cadherin, and Vimentin were from Cell Signaling Technology (Danvers, MA). Anti-actin antibody, doxorubicin-HCl (Dox), and Dasatinib were from Sigma (St. Louis, MO). HRP-labeled secondary antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). Chemiluminescence kit was from Thermo Scientific (Suwanee, GA). Fluorescence-labeled antibodies for CD44 and CD24 were from BD Biosciences (San Jose, CA). Rhodamine phalloidin was from Invitrogen (Carlsbad, CA). Fluoroshield mounting media with DAPI was from Abcam (Cambridge, UK).

MCF-7 and ZR75-1 cells were maintained in RPMI media containing 10% or 15% FBS respectively; MDA-MB-231 and HeLa cells were maintained in DMEM containing 10% FBS. U-2 OS cells were maintained in McCoy’s 5A containing 10% FBS; SKBR3 cells were maintained in F12/DMEM containing 10% FBS. Cell lines were maintained at 37 °C, 5% CO₂ in a humidified atmosphere and tested for mycoplasma contamination bi-monthly. For experiments, cells were sub-cultured in 60 mm (200 K cells) and 100 mm (500 K cells) dishes with media being changed 1–2 h prior to the start of experiments. For siRNA, cells were transfected using both forward and reverse transfection methods. For forward transfections, cells were seeded in 60 mm (150 K) and transfected the next day with 20 nM negative control (AllStars, Qiagen) or siRNA using oligofectamine or lipofectamine RNAimax (Life Technologies) according to the manufacturer’s protocol (Top2α, Top2β, ATR, ATM, p53, Fyn, Yes siRNA from Life Technologies). For reverse transfection, cells were seeded into media containing siRNA complexes using lipofectamine RNAimax according to the manufacturer’s protocol. In both cases, cells were incubated for 48 h before a media change prior to stimulation.

The pcDNA-p53-WT plasmid was a generous gift from Dr. Ute Moll (Stony Brook University). MCF7 cells were seeded in 60 mm (1 × 10⁶) and transfected the next day with 500 ng of empty vector (pcDNA) or p53 plasmid using Xtreme gene transfection reagent (Roche, Basel, Switzerland) according to the manufacturer’s protocol for 24 h. Cells were harvested for protein and RNA extractions as described below.

To extract cellular protein, cells were scraped in RIPA buffer, lysed by sonication on ice (1 time, 10s) and protein concentration estimated by the Bradford assay. Lysate aliquots were mixed with one-third volume of 4X Laemmli buffer (Bio-Rad) containing 2-mercaptoethanol (Sigma), vortexed for 2–3 s and boiled for 5–10 min. Protein was separated by SDS-PAGE using the Criterion system (Bio-Rad) and immunoblotted as described previously [47].

Total mRNA was extracted using the PureLink RNA Mini Kit (Invitrogen). 0.5–1 μg of RNA was converted to cDNA with the Superscript III Kit for first-strand synthesis (Invitrogen) and samples were diluted to 100 μl with molecular biology-grade dH₂O. Real-time RT-PCR with Taqman assays were performed on the ABI 7500 real-time system using iTaq mastermix (Bio-Rad). Reactions were run in triplicates in 96-well plates with each reaction containing 10 μl of 2 × iTAQ mastermix, 5 μl of cDNA, 1 μl of Taqman primer probe, and 4 μl of water. The qPCR protocol consisted of 2 min enzyme activation at 95 °C followed by 40 cycles consisting of a 10-s melt at 98 °C, and a 60-s anneal and extension at 60 °C. Ct values were converted to mean normalized expression using the ΔΔCt method using actin as a reference gene. Taqman assays were purchased from Life Technologies. For real-time RT-PCR reactions with sybr green, primers were designed with the Thermo Fisher oligoperfect program (https://​www.​thermofisher.​com/​us/​en/​home/​life-science/​oligonucleotides​-primers-probes-genes/​custom-dna-oligos/​oligo-design-tools/​oligoperfect.​html) and validated in silico using the University of California, Santa Cruz (UCSC) in silico-PCR program (http://​mgc.​ucsc.​edu/​cgi-bin/​hgPcr). Two micrograms total RNA were used for cDNA synthesis using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Thermo Fisher).

This was carried out as previously described with minor modifications [48]. MCF7 cells were grown on poly-D-lysine-coated 35-mm confocal dishes (MatTek Corporation) and treated the following day. Cells were fixed using 4% paraformaldehyde, washed with 1× PBS, and permeabilized with 0.1% Triton X-100. Cells were incubated in rhodamine phalloidin at 1:200 in PBS for 20 min at room temperature and protected from light. Following two washes with PBS, cells were stained with DAPI in mounting media for 10 min according to the manufacturer’s protocol and imaged using a Leica TCS SP8 laser scanning confocal microscope.

Cells were seeded at 90% confluence in 24-well plates in triplicate for each condition. After 24 h, a scratch was introduced across the center of each well using a p10 pipette tip. Cells were washed once with 1× PBS and replaced with fresh media. After 1 h, the wound was imaged using the EVOS XL Core Cell Imaging System and treated immediately after. Images were taken at 24 and 48 h and area of the scratch was measured at 0, 24, and 48 h using the ImageJ wound healing assay.

Migration assay was performed as previously described with minor modifications [28]. Cells were seeded and treated in 60-mm plates for 24 h or 100-mm plates with siRNA for 48 h followed by stimulation with Dox. Cells were trypsinized and resuspended at 200,000 cells/ml in serum-free media. In total, 500 μl was added to transwell inserts with 8-μm pores (Corning, NY, USA) with 600 μl of complete media added to the lower chamber. Following 24 h in normal culture conditions, excess cells were removed from the upper side of the membrane with cotton swabs and the lower side of the membrane was fixed with 70% ethanol for 10 min. After drying (10–15 min), membranes were stained in 0.2% crystal violet for 5–10 min. Wells were washed with ddH₂O and air-dried overnight. Migrated cells were imaged using EVOS XL Core Cell Imaging System and quantified using ImageJ multi-point and threshold tool.

Cell invasion assay was carried out using the Corning Tumor Invasion System (Corning, NY, USA) following the manufacturer’s protocol with minor modifications as previously described [49]. Briefly, cells were starved for 4 h, resuspended at 200,000 cells/ml, and 500 μl was added to the apical chamber of the preactivated cell invasion plate. Serum-free medium or medium with chemoattractant was added to the lower chamber at a volume of 750 μl and cells were allowed to invade under normal cell culture conditions. Calcein AM dye (Life Technologies) was used to stain invading cells. Fluorescence was quantified using the SpectraMax M5 plate reader (Molecular Devices, Sunnyvale, CA, USA).

Cell-matrix adhesion assay was adapted from Piccolo protocol [50] and reported previously [51]. MCF-7 cells were plated in 100-mm dishes at 1 million cells per each treatment condition in RPMI-high glucose, 10% FBS. Next day, culture media were changed to RMPI-high glucose without FBS, for 16 h, and then treatments were added for the required time. Cells were washed in FBS-free media and detached with 10 mM EDTA in PBS for 20 min and collected with the help of a cell lifter. Cells were collected and centrifuged at 600×g for 3 min and resuspended at 10⁵ cells/ml in FBS-free media. Cell culture 12-well plates or 35-mm confocal dishes were previously coated with fibronectin for 16 h at 1μg/ml, and cells were added at 100 K/cells/ml and incubated at 37 °C for 60 min. Culture dishes were washed 3 times with PBS, and cell number was estimated using the MTT assay. Alternatively, cells were fixed with warmed 4% paraformaldehyde in PBS (w/v), permeabilized in 0.1% Triton X-100 (v/v) in PBS, and stained using rhodamine-phalloidin and DAPI. Attached cells were visualized under a laser scanning confocal microscope Leica TCS SP8 at × 20 magnification. Images were processed for automatic cell counting using Python-Bioformats-OpenCV.

The LDH assay was carried out using a commercially available kit according to the manufacturer’s protocol. Briefly, at the end of treatment period, 50 μL of medium were placed in duplicates in a 96-well plate. Equal volume of LDH reaction mix was added to each well and covered with aluminum foil to protect from light. The plate was incubated at 37 °C for 30 min and endpoint absorbance was detected at 490 nm and 680 nm wavelength using the SpectraMax plate reader.

Following treatment, cells were washed once with warm PBS. Medium containing 5 mg/ml MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium Bromide) was added for 30 min at normal cell culture conditions. The insoluble formazan product of MTT was dissolved with DMSO and quantified by measuring its absorbance at 570 nm using the SpectraMax plate reader.

MCF-7 cells were sub-cultured in 100-mm dishes (400 K) and treated with vehicle (Veh) or Dox as described. Total mRNA was extracted using the PureLink RNA Mini Kit (Invitrogen) and > 2 μg of DNase-treated RNA samples were submitted to New York Genome Center for deep sequencing analysis. The transcriptome was analyzed by 2 × 50 paired end sequencing to a depth of 30 million reads using a HiSeq2500. Aligned genes were normalized using Salmon [52] and the resulting lists for each condition were filtered to remove gene counts having a value < 1. Gene counts were compared as a ratio of treated and untreated conditions. List of genes having a ratio value of > 1.95 were input into the Database for Annotation, Visualization and Integrated Discovery (DAVID) for gene ontology analysis (https://​david.​ncifcrf.​gov/​tools.​jsp).

The assay was carried out as previously described [53] with minor changes. Briefly, cells were sub-cultured in 150-mm dishes (10⁶) and, following treatment, were washed with warm 1× PBS, trypsinized, and centrifuged at 700×g for 5 min. Cell pellets were resuspended in PBS at 500,000 cells/100 μl and incubated with FITC-CD24 and PE-CD44 antibodies (1:20 dilution, 4 °C, 30 min). FITC-IgG and PE-IgG were used as negative controls. Following incubation, cells were washed twice with PBS, resuspended in 500 μl PBS, and analyzed using BD Accuri C6 Plus flow cytometer. Data were analyzed using FlowJo software.

To establish appropriate sublethal doses of Dox and to assess the cellular effect of sublethal Dox treatment, MCF7 cells were treated with Dox ranging from 0.2 to 1.0 μM for 24 h. Immunoblot analysis showed PARP cleavage did not occur at Dox doses of 0.2 μM, 0.4 μM, and 0.6 μM but was observed at 0.8 μM and 1.0 μM Dox, indicating initiation of apoptosis at these higher concentrations (Fig. 1A). Consistent with this, lactate dehydrogenase (LDH) release was not observed at doses up to 0.8 μM Dox, but significant release of LDH occurred at the apoptotic dose of 1.0 μM, indicating loss of membrane integrity at this dose (Fig. 1B). Based on these results, doses of 0.6 μM or less were denoted as sublethal. Notably, cells treated with sublethal doses of Dox displayed changes in cell shape (Fig. 2A), with protrusions extending from the cell body following treatment that gave the cells a stellate shape. This filopodia-like morphology was most prominent in at 0.2 μM and 0.4 μM Dox, and while protrusions were present at 0.6 μM, 0.8 μM, and 1.0 μM Dox, the cells appeared more elongated. Phalloidin staining highlighted that the change in cell shape indeed affected the actin cytoskeleton. Untreated cells were polygonal in shape while Dox-treated cells appeared more multipolar (Fig. 2B). As changes in cell shape at sublethal doses of Dox suggested a more invasive phenotype, we sought to evaluate this experimentally. Accordingly, a wound healing assay was performed. Following introduction of a scratch, cells were treated with vehicle or 0.4 μM Dox, and wound closure was monitored. As can be seen, Dox-treated cells closed the wound significantly more compared to vehicle-treated cells (Fig. 2C, D). To rule out effects on cell growth as a reason for wound closure, viable cell numbers of vehicle- and Dox-treated cells were assessed (Fig. 2E). Strikingly, vehicle-treated cells retained normal cell growth while a growth arrest was observed in Dox-treated cells. Thus, the wound closure observed is not a result of enhanced cell growth (Fig. 2E). Consolidating these data, cells treated with sublethal doses of Dox showed significantly increased motility in a transwell migration assay compared to vehicle-treated cells (Fig. 2F). Consistent with induction of a more invasive phenotype, there was a significant increase in expression of MMPs, most evident with MMP-1 and MMP-2 increasing by 2.17-fold and 1.92-fold respectively (Fig. 2G). This was further reflected in the increased secreted activity of MMP-2 and MMP-9 following sublethal Dox treatment with little to no activity in untreated cells (Fig. 2H). Finally, cells treated with sublethal Dox showed a significant reduction in the ability of cells to attach on fibronectin (Fig. 2I). Of note, while we also observed a reduction of cell growth following Dox treatment, removal of Dox allowed cells to resume growth (Supplemental Fig. 1). Taken together, these results show that sublethal Dox treatment induces a pro-migratory and pro-invasive program in non-invasive MCF7 cells.

The epithelial-mesenchymal transition (EMT) is a well-established promoter of cell invasion and can lead to the generation of cells with cancer stem cell (CSC) characteristics. In the EMT, cells lose epithelial characteristics and transition to display more mesenchymal features resulting in increased motility [54]. To determine if EMT is relevant in the effects of Dox here, EMT markers were assessed by immunoblot (Fig. 3A). Interestingly, cells treated with sublethal Dox doses retained expression of the epithelial marker E-cadherin and did not express the mesenchymal markers Vimentin or N-cadherin. Consistent with this, sublethal Dox treatment did not significantly increase the expression of Twist (Fig. 3B), a key transcription factor for EMT. Consistent with these findings, there was no appreciable change in the CSC population (CD44⁺/CD24^low/−) between vehicle-treated and Dox-treated cells (Fig. 3C, D), indicating sublethal Dox treatment did not result in an enrichment of CSCs. Taken together, these results showed that sublethal Dox treatment induces migration and invasion independently of EMT or increasing CSC population.

To understand how sublethal Dox may affect the biology of the cell, we performed a discovery RNAseq analysis on MCF7 cells in order to define changes in gene expression following sublethal Dox treatment. Consistent with the observed phenotypes, data analysis showed that genes regulating cell adhesion and cellular migration were enriched in cells treated with sublethal doses of Dox (Fig. 4). Interestingly, multiple Src Family Kinases (SFK) were induced by Dox in MCF7 cells (Table 1). To confirm SFK upregulation was indeed at the mRNA level, Src expression was assessed by real-time PCR. Results in MCF7 cells showed significant increases in Src mRNA at 0.4 μM (1.90-fold) and 0.6 μM Dox (2.10-fold) yet at 1 μM Dox—an apoptotic Dox dose—Src was not induced (Fig. 5A). Beyond Src, there are currently nine members of the SFK including Fyn, Yes, Blk, Yrk, Fgr, Hck, Lck, and Lyn. Of these, Src, Fyn, and Yes are expressed in a wide range of tissues [33] and have been shown to promote cell migration/invasion [35, 40, 55]. Extending our analysis revealed marked induction of Fyn by Dox, peaking at 0.4 μM (10.7-fold) and markedly lower at apoptotic doses of 1 μM (2.6-fold) (Fig. 5D). Similarly, Yes expression was induced by Dox with a peak at 0.4 μM (2.66-fold) with no induction at higher doses (Fig. 5G). Finally, Fgr was induced at 86.5-fold at 0.6 μM (Fig. 5J). To consolidate and extend these results, we utilized ZR75-1, MDA-MB-231, and SKBR3 breast carcinoma and U-2OS osteosarcoma cells (Fig. 5, Supp. Fig. 2). Real-time analysis showed that sublethal Dox treatment of ZR75-1 cells leads to similar induction of Src (3.54-fold at 0.4 μM)(Fig. 5B), Fyn (3.62-fold)(Fig. 5E), Yes (2.82-fold)(Fig. 5H), and Fgr (15.4-fold) (Fig. 5K). Similarly, Dox treatment of U-2OS cells showed increased Src (1.81-fold at 0.4 μM)(Fig. 5C), Fyn (1.75-fold), (Fig. 5F) Yes (1.87-fold)(Fig. 5I), and Fgr (39-fold)(Fig. 5L). To evaluate if the increase in mRNA was reflected at the protein level, Dox effects on Src and Yes in MCF7, ZR75-1, and U-2 OS cells were assessed by immunoblotting (Supp. Fig. 4A-C). Results showed that Dox treatment showed a marked increase of Src at 0.4 μM to 1.0 μM in MCF7 cells (Supp. Fig. 4A). In ZR75-1 cell, 0.4 μM Dox showed an increase in both Src and Yes levels (Supp. Fig. 4B). Similar results were seen in U-2 OS cells with increased levels of Src at 0.6 μM and Yes at 0.4 μM (Supp. Fig. 4C). Importantly, in all cell lines, increased levels of Src were accompanied by increased levels of phospho-Src (Tyr-416), indicating that Dox is inducing the active form of the protein (Supp. Fig. 4).

Dox binds to DNA and Topoisomerase II (Top2) and generates DNA strand breaks which leads to a cascade of events involving p53 ref. [56]. Indeed, in MCF7 cells, there was a marked dose-dependent increase in the levels of p53 in cells treated with Dox (Fig. 6A). To understand the mechanism of action, we sought to explore whether p53 played a role in the induction of Fyn by Dox using an siRNA approach. Results showed a partial dependence of Fyn induction on p53 in MCF7 cells (Fig. 6B, C) but this was not seen for Yes or Fgr (Fig. 6D, E). Similarly, Fyn induction was partially dependent on p53 in ZR75-1 cells (Supp. Fig. 5A-C). In contrast, in U-2OS cells, Yes induction was dependent on p53 expression (Fig. 6F–I). Mutations of p53 that lead to a disruption of its transcriptional activity is a common event in many cancers. Accordingly, to determine if Dox had similar effects in the context of p53 mutations, we extended our studies into MDA-MB-231 (basal) and SKBR3 (HER2+) cells (Supp. Fig. 2). Both cell lines harbor p53 mutation and, importantly, represent distinct subtypes of breast cancer than the luminal, ER-positive MCF7 and ZR75-1 cells. Results in MDA-MB-231 cells showed that Dox did induce SFKs, increasing Src (1.98-fold at 0.4 μM Dox), Fyn (2.17-fold at 0.6 μM Dox), and Yes (1.75-fold at 0.6 μM Dox) although this was considerably less than seen in MCF7 and ZR75-1 cells. Interestingly, Dox treatment of SKBR3 cells had no effect on Src, Fyn, and Yes within the range of 0.4–1.0 μM. Finally, to determine if p53 is sufficient to drive SFK levels, the effects of transient p53 overexpression on the mRNA levels of SFKs were assessed. As can be seen, we achieved a modest overexpression of p53 at the protein level. Importantly, this was associated with a modest but statistically significant increase in Fyn expression but not in Src and Yes (Supp. Fig. 5D-G). Taken together, these results suggest that Dox effects on SFKs are at least partially p53 dependent.

The involvement of p53 in SFK induction led us to speculate that this could be linked to the DNA damage response (DDR). To begin to explore this, the role of the upstream DDR kinases ATM and ATR in SFK induction was assessed in MCF7 cells (Fig. 7A–C). Here, knockdown of ATR but not ATM showed a modest but statistically significant effect on both Fyn and Yes induction (Fig. 7B, C). The interaction of Dox with Top2α and Top2β is an important part of how it induces DNA double-strand breaks. However, siRNA knockdown of either Top2α or Top2β had no significant effects on Fyn induction (Fig. 7D–F). The connection of SFK induction to the ATR-p53 axis prompted us to assess if other DNA-damaging agents had the same effects. Interestingly, MCF7 cells treated with Camptothecin and Etoposide did not show increases in either Fyn (Fig. 8A) or Yes (Fig. 8B). Similarly, treatment with Taxol, a cytoskeletal targeting drug, also did not change expression levels of Fyn and Yes mRNA (Fig. 8A, B). Consistent with these findings, treatment with Camptothecin or Taxol did not induce migration of MCF7 cells in the wound healing assay (Fig. 8C). Collectively, while these agents are all known to induce the DDR and p53, these results suggest a Dox-specific mechanism is required for induction of SFKs in MCF7 cells.

The increase in SFK expression following sublethal dose of Dox prompted us to assess its role in the migratory phenotype. For this, and as multiple SFKs were induced on Dox treatment, a pharmacological approach with the SFK inhibitor Dasatinib was used. Co-treatment with 0.1 μM Dasatinib and 0.4 μM Dox showed inhibition of SFK phosphorylation on Tyr-416 (Fig. 9A). Results from wound healing assays showed Dasatinib co-treatment significantly attenuated the migration of Dox-treated cells (Fig. 9B, C). As before, this was not a consequence of alterations in cell growth (Fig. 9D). Furthermore, Dasatinib co-treatment also significantly mitigated the invasion of Dox-treated cells through Matrigel (Fig. 9E). To further assess the effect of specific SFKs on migration of Dox-treated cells, a genetic approach was employed (Supp. Fig. 6A-C, Fig. 9F,G). As can be seen, knockdown of Fyn showed significant reduction of cell migration through the transwell (Fig. 9F). Migration was also reduced by knockdown of Yes, albeit not significantly (Fig. 9G). Taken together, these results suggest a role of SFKs in the Dox-induced invasion/migration in MCF7 cells and identify Fyn as a major contributor for this effect.