Background
Breast cancer still remains one of the most common malignancies in women with multiple risk factors [
1]. Any solid tumor derived from breast epithelial tissue is supported by tumor stroma – a non-malignant tumor compartment composed from multiple cell types and non-cellular components. The tumor microenvironment creates a complex signaling network which substantially affects tumor biology and therapeutic responsiveness [
2,
3]. Adipose tissue is the most abundant stromal constituent in the breast and also a rich source of mesenchymal stromal cells (MSCs) which contribute to mammary carcinogenesis [
4]. As a fat grafting procedure is frequently used in breast reconstruction, breast contour deformity correction or even in breast augmentation, it also carries potential oncological risk of de novo breast cancer and/or its recurrence [
5,
6].
The MSCs derived from the adipose tissue (AT-MSCs) share a number of key characteristics with the bone marrow-derived MSCs (BM-MSCs) [
7‐
9]. MSCs from both sources were demonstrated to integrate into tumor-associated stroma and exhibit multiple regulatory functions in the tumor microenvironment [
10‐
12]. Experimental data revealed the capability of BM-MSCs to differentiate into tumor-associated fibroblasts [
13‐
15] and even create a cancer stem cell niche [
16] when exposed to tumor-conditioned medium. The interaction of BM-MSCs and breast cancer cells was also shown to promote metastatic spread as a result of bidirectional paracrine signaling [
17]. Although the effect on proliferation of the tumor cells was not stimulatory in general, MSCs were shown to promote tumor cell migration, an epithelial-to-mesenchymal transition (EMT), mediate release from the hormone-dependence, and increase chemoresistance in breast cancer cells [
18‐
23]. MSCs-secreted factors increased mammosphere formation and the exosomes from MSCs were sufficient to support the growth of tumor xenografts [
24‐
26]. Taken together these data suggest that BM-MSCs promote breast cancer growth and/or metastatic spread. However, a suppression of the tumor growth by MSCs was reported for the tumor types other then breast; and the role of MSCs in tumor growth remains a matter of further investigations [
12,
27‐
29]. Better understanding of the underlying mechanisms might lead to the therapeutic intervention with the aim to increase an antitumor response [
30,
31]. MSCs themselves can be specifically engineered for the increased tumor-targeting and efficiency of the anti-tumor treatment [
32]. The introduction of specific transgene(s) into the AT-MSCs sensitized the breast cancer cells MDA-MB-231 to the chemotherapeutic drug 5FU for
in vitro[
33].
We have previously characterized the effect of AT-MSCs on the proliferation of breast cancer cells; and linked it to the cytokine secretion profile of AT-MSCs [
23]. In this study we have focused on the multiple alterations induced in human Her2-positive breast cancer cell line SKBR3 by the AT-MSCs. We have extended our investigation also on the effect of stromal cells on drug responses in the tumor cells. We have observed that the AT-MSCs induced an EMT, decreased proliferation, increased migration and other molecular changes in the SKBR3 cells. We have shown that the AT-MSCs could alter chemosensitivity of the tumor cells.
Methods
Cells
Human tumor cell line SKBR3 (ATCC® Number HTB-30™) was used for the study. Tumor cells were maintained in high-glucose (4.5 g/l) DMEM (PAA Laboratories GmbH) containing 10% FBS (Biochrom AG), 10.000 IU/ml penicillin (Biotica, Part. Lupca, Slovakia), 5 μg/ml streptomycin, 2 mM glutamine and 2.5 μg/ml amphotericin (PAA Laboratories GmbH).
For mammosphere cultures, 4×104 EGFP-SKBR3 cells per well were plated in non-adherent 6-well plates (Ultra-low attachments plates, Corning, Amsterdam, NL) in serum free DMEM/F12 medium (GIBCO-Invitrogen BRL) supplemented with 10.000 IU/ml penicillin (Biotica, Part. Lupca, Slovakia), 5 μg/ml streptomycin, 2 mM glutamine, and 2.5 μg/ml amphotericin (Sigma, St. Louis, MO), 10 ng/ml bFGF (Miltenyi Biotec), 10 ng/ml EGF (Miltenyi Biotec), 4 μg/ml heparin (Sigma, St. Louis, MO), 2 μg/ml insulin (Sigma, St. Louis, MO) and B27 supplement (diluted 1:100, Gibco-Invitrogen BRL) and cultivated at 37°C in humidified atmosphere and 5% CO2 for 5 days. Specific inhibitors 1.63 μM LY294002 (Sigma, St. Louis, MO) or 0.5 μM SB203580 (Sigma, St. Louis, MO) were added to the MSCs-CM mammosphere medium as indicated.
AT-MSCs were isolated and characterized by immunophenotype and differentiation potential as previously described in [
34] (Additional file
1). The AT-MSCs were expanded in low glucose (1.0 g/l) DMEM supplemented with 10% HyClone® AdvanceSTEM™ supplement (Thermo Scientific) and antibiotic/antimycotic mix (10.000 IU/ml penicillin, 5 μg/ml streptomycin, 2 mM glutamine, and 2.5 μg/ml amphotericin). Different isolates were used for the experiments (n = 4), each experiment was run at least twice with each isolate to draw the conclusions. Cells were maintained at 37°C in humidified atmosphere and 5% CO
2.
Cell-free AT-MSCs conditioned medium (MSCs-CM) was collected from 80–90% confluent cultures after 24 hours of cultivation with fresh tumor cell culture medium or mammosphere culture medium, respectively, and filtered through 0.45 μm filters. Fresh MSCs-CM was always used for the experiments.
EGFP expression
Stable transduction of SKBR3 to express enhanced green fluorescent protein (EGFP) was done by retrovirus gene transfer as described elsewhere [
33]. Transgene incorporation and EGFP expression was confirmed by PCR, reverse transcription coupled PCR and flow cytometric analysis performed on BD Canto II cytometer (Becton Dickinson, USA) equipped with FACS Diva program. FCS Express software was used for evaluation. The identity of SKBR3 and EGFP-SKBR3 cells was further confirmed by sustained expression of epithelial cell adhesion molecule (CD326, ≥98% positivity) verified by flow cytometry with specific antibody anti-EpCAM-PE (Miltenyi Biotec GmbH, Germany). Mouse IgG1-PE (Miltenyi Biotec GmbH, Germany) was used as negative isotype control.
Analysis of morphological changes in EGFP-SKBR3
Three ×105 EGFP-SKBR3 cells were mixed with 1.5×105 DiI-stained AT-MSCs and cocultured for 5–9 days. For a comparison, EGFP-SKBR3 cells alone were seeded and cell morphology was analyzed by fluorescent microscopy (Axiovert 200, Zeiss, Germany). Alternatively, quadruplicates of 4×104 tumor cells were seeded in MSC-CM or culture medium in 96-well plates. Phase-contrast images were taken in the IncuCyte ZOOM™ Kinetic Imaging System (Essen BioScience, UK). Cell confluence was evaluated by IncuCyte ZOOM™ 2013A software (Essen BioScience, UK) based on the confluence masks as recommended by manufacturer.
Migration assay
Fifty thousand EGFP-SKBR3 per well were plated in triplicates in ImageLock 96-well plates (Essen BioScience, UK) and let to adhere for 16 hrs. Confluent monolayers were wounded with wound making tool (Essen BioScience, UK), washed twice and supplemented with MSC-CM or culture medium. As indicated, medium was supplemented with receptor-tyrosine kinase inhibitors 150 nM Pazopanib, 250 nM Sorafenib or 200 nM Sunitinib (inhibitors kindly provided by National Cancer Institute, Bratislava). Images were taken every two hours for next 72 hrs in the IncuCyte ZOOM™ Kinetic Imaging System (Essen BioScience, UK). Cell migration was evaluated by IncuCyte ZOOM™ 2013A software (Essen BioScience, UK) based on the relative wound density measurements and expressed as means of three independent experiments run in triplicates ± SD.
Gene expression analysis
EGFP-SKBR3 tumor cells were cultured with or without MSC-CM for 6 days with everyday medium replenishment. Total RNA was isolated from 5×10
6 EGFP-SKBR3 cultured with or without MSC-CM. Cultured cells were collected by trypsinization, RNA isolated by NucleoSpin® RNA II (Macherey-Nagel) and treated with RNase-free DNase (Qiagen, Hilden, Germany). Total RNA was subjected to control PCR to confirm the absence of genomic DNA contamination. RNA was reverse transcribed with RevertAid™ H minus First Strand cDNA Synthesis Kit (Fermentas, Hanover, MD). 200 ng of cDNA was amplified in standard PCR performed in 20 μl 1x PCR master mix (Fermentas, Canada) with 0.5 μl respective specific primers (20 pmol/μl) and DNase free water (Fermentas, Canada) in DNA Engine Dyad™ Peltier Thermal Cycler (MJ Research, UK) with pre-set amplification profile and horizontal electrophoresis was used for detection of amplicons. Each reaction was run with appropriate no template controls and negative control (RNA template without reverse transriptase). Primer sequences were listed in Additional file
2.
Quantitative PCR was performed in 1 × ABsolute™ QPCR SYBR® Green Mix (ABgene, Surrey, UK), 0.16 μM primers and 200 ng of template cDNA on Bio-Rad CFX96™ and analyzed by Bio-Rad CFX Manager software version 1.6. Relative gene expression change was calculated according to ΔΔCt method. GAPDH and HPRT1 gene expression was taken as endogenous reference. Analysis was performed twice in triplicates and data expressed as means ± SD.
Multiplex and SDF-1α secretion analysis
5×104 EGFP-SKBR3, 2.5×104 AT-MSCs alone, and 5×104 SKBR3 cells mixed with 2.5×104 AT-MSCs (ratio 2:1) were plated in the wells of 24-well plates and cultured in 2 ml of complete culture medium for two days. Cell-free supernatants were collected and subjected to human Bio-Plex™ 27-plex Cytokine Assay (Bio-Rad Laboratories Inc, Hercules, CA). Measurements were performed on Luminex 100 System (Luminex Corporation, Austin, TX) in duplicates with two different AT-MSCs isolates. Results were expressed as mean pg/ml of culture medium ± SD.
In order to confirm the SDF-1α secretion SDF1-α Quantikine Immunoassay (R&D Systems Inc.) was used. SDF-1α levels in cell free supernatants were determined on xMark™ Microplate Spectrophotometer (BIO-RAD).
Cell proliferation
The effect on tumor cell proliferation was evaluated as a relative fluorescence determined by green fluorescence readout (Ex. 485, Em. 520) on PolarStar OPTIMA reader (BMG Labtechnologies, Offenberg, Germany) in direct cocultures. Quadruplicates of 1×104 EGFP-SKBR3 cells were seeded in black-walled 96-well plates (Greiner Bio-One Intl. AG) with increasing numbers of AT-MSCs and cultured for 6 days. Green fluorescence was directly proportional to the number of viable tumor cells within the wells and the fluorescence value in the untreated cells was set to 100% by default. Experiments were evaluated as mean of quadruplicates ± SD.
In order to dissect the role of SDF-1α/CXCR4 axis in proliferation of EGFP-SKBR3 cells in cocultures with AT-MSCs, specific inhibitor of this signaling axis AMD 3100 (Sigma, St. Louis, MO) was used. Final concentration of 5 μg/ml AMD 3100 was added to EGFP-SKBR3 cells alone, cultured in MSC-CM or in coculture with AT-MSCs. The effect on proliferation was evaluated as a relative fluorescence as described above.
Relative cell viability was evaluated by CellTiter-Glo™ Luminescent Cell Viability Assay (Promega Corporation, Madison, WI) based on the ATP quantitation representative of metabolically active cells. Quadruplicates of 6×103 SKBR3 cells per well were seeded in 96-well plates overnight. Diluted MSCs-CM was added to the adherent tumor cells on the next day. Relative proliferation was determined on LUMIstar GALAXY reader (BMG Labtechnologies, Offenburg, Germany). Values were expressed as mean relative luminescence ± SD, when luminescence of control cells was taken as reference. Experiments were repeated at least twice with similar results and a representative result is shown.
Chemosensitivity
Following drugs were used: 5-fluorouracil (5FU, Sigma, St. Lois, MO), doxorubicin (DOX, EBEWE Pharma, Austria) and cis-platin (EBEWE Pharma, Austria). For the evaluation of chemosensitivity, either 6×103 EGFP-SKBR3 cells alone or mixed with AT-MSCs (ratio 2:1) were seeded in 96-well plates. On day 0, treatments were started with doxorubicin (6.25 -100 ng/ml), 5FU (6.25-1000 ng/ml) or cis-platin (0.001-10 μg/ml). The chemosensitivity was determined by fluorescence measurements as described above 6 days later. Experiments were evaluated as means of three different experiments run in quadruplicates and the relative fluorescence in untreated cells was taken as 100% by default. Alternatively, 8×103 EGFP-SKBR3 were seeded in 96-well plates overnight and treated with the drugs diluted in MSCs-CM. Relative fluorescence and cell proliferation was determined as above.
Caspase-3/7 assay
Quadruplicates of 2×104 SKBR3 per well were seeded in 96-well white-walled plates (Corning Costar Life Sciences, Amsterdam, NL) overnight. Doxorubicin (100 ng/ml) or 5FU (100 μg/ml and 500 μg/ml) diluted in MSC-CM or culture media was added to the cells for the indicated period of time and a Caspase-3/7 activity was determined by the Caspase-Glo® 3/7 Assay (Promega Corporation, Madison, WI) on LUMIstar GALAXY reader (BMG Labtechnologies, Offenburg, Germany) at indicated timepoints. Values were determined as mean values of RLU ± SD.
Annexin V assay
In order to quantify a proportion of viable, apoptotic and necrotic cells in cocultures, adherent AT-MSCs were labeled with 5 μM carboxy-fluorescein diacetate, succinimidyl ester (CFDA-SE, Molecular Probes, Eugene, OR) in a serum-free DMEM for 15 min at 37°C. Medium was replaced for standard culture medium to incubate overnight. Next day, SKBR3 cells were mixed with CFDA-SE labeled AT-MSCs in a ratio 2:1 and plated onto 6-well plate (5×104 SKBR3, 5×104 AT-MSCs, or 5×104 SKBR3 with 2.5×104 AT-MSCs/well) for direct co-culture. Doxorubicin at final concentration 50 ng/ml was added to the respective wells one day later and cells were treated for 48 hrs. Apoptotic cells were stained with Phycoerythrin-labeled Annexin V (eBioscience, San Diego, CA); dead cells were detected with DAPI viability dye. Cells were analyzed using BD CantoII cytometer (Becton Dickinson, USA) equipped with FACSDiva program. FCS Express software was used for the evaluation.
Statistical analysis
Studies involving comparison between the two groups were analyzed by an unpaired Student's t-test in GraphPad Prism® software (LA Jolla, CA). The value of p < 0.05 was considered statistically significant.
Discussion
MSCs represent multipotent cells valuable for regenerative therapies including augmentation of tissue regeneration in breast reconstruction after cancer-related surgery. Although recent results suggested that AT-MSCs might improve a long-term retention of the grafts, the risks of this cellular treatment still remain unresolved specifically in the context of a patient with cancer history [
5,
6]. Tumors always encompass both malignant part and non-malignant cells of various cell lineages with complex mutual interactions between particular cell types [
2,
40]. MSCs can contribute to the tumor microenvironment and play a role in mammary carcinogenesis [
11]. Our data showed that AT-MSCs did not increase the proliferation of the HER2-overexpressing, estrogen/progesterone receptor negative breast cancer cells SKBR3. However, AT-MSCs induced an EMT in tumor cells with increased tumor cell migration and mammosphere formation, potentially leading to increased aggressiveness and metastatic capability. MSCs derived from bone marrow were already described to affect breast cancer cell proliferation, migration, invasiveness, metastasis, morphology, chemoresistance and hormone responsiveness (reviewed in [
11,
41]). According to our data the MSCs can alter tumor biology regardless of their tissue origin. Similarities in the MSCs secretome dictate the nature of the interaction with the other cell types [
9]. It has been shown that a gene expression profile of the MSCs derived from breast adipose tissue is comparable to the MSCs originating from abdominal adipose tissue resulting in comparable stimulation of proliferation in breast cancer cells MCF7 and MDA-MB-231 [
42]. Moreover, the MSCs from primary breast cancer tissues were also shown to exert stimulatory effect on MCF7 proliferation and tumor growth [
43]. Detailed study of migration properties of the tumor-cell exposed MSCs have unraveled increased migration of the MSCs isolated from breast adipose tissues in comparison to the migration of the MSCs derived from abdominal adipose tissue [
44]. Gene expression profile of these migratory MSCs was close to the profile of MSCs isolated from the tumor-adjacent breast adipose tissues [
44]. Thus the MSCs derived from abdominal adipose tissue with lower responsiveness to tumor-induced motility might be preferred exogenous cell source for fat grafting and breast augmentation to limit the effect on mammary carcinogenesis.
MSCs-secreted cytokines induced an EMT, increased expression of pluripotency genes and mammosphere formation in breast cancer cells (Figure
1C-D and [
24]) which might suggest the capability of MSCs to increase the proportion of tumor initiating cells as a consequence of the EMT [
35,
45]. MSC-CM induced expression of VEGFR2 concomitant with high VEGFA expression in SKBR3 cells could generate autocrine loop directly affecting a tumor cell survival and potentially more invasive phenotype [
46]. Based on these data, we hypothesized that SKBR3 cells in combination with AT-MSCs might have increased tumorigenicity. However, no increase in the tumor-forming capabilities was observed when AT-MSCs were coinjected with EGFP-SKBR3 cells
in vivo. AT-MSCs could not support the xenotransplant growth in immunodeficient mice (data not shown). The EMT and upregulation of pluripotency genes induced by MSC-CM was not sufficient to promote tumor growth in low tumorigenic SKBR3 cells
. Recently Karnoub's group demonstrated that the MSCs-mediated EMT was neither sufficient nor necessary for a generation of cancer stem cell phenotype, although it contributed to the increased metastasis
in vivo[
47]. Future studies will be focused on the attempt to develop tumor xenotransplant model to test the MSCs-mediated alterations in the tumor behavior and its chemosensitivity
in vivo.
Our data further support the dual role of MSCs in tumor cell proliferation. Previously we have reported increased proliferation of breast cancer cells T47D, MCF7 and MDA-MB-361 in response to AT-MSCs [
19,
23] in contrast to antiproliferative action on SKBR3 cells (Figure
3). Our data correspond with the findings by Donnenberg
et al., who did not show the capability of the AT-MSCs to increase the proliferation of dormant tumor cells [
6]. Several studies reported that the MSCs could actually inhibit tumor growth
in vivo[
29,
48‐
50] although in different tumor types (glioblastoma, leukemic, thyroid and colon cancer cells). More importantly, substantially altered composition of the chemokine secretome in tumor-stromal coculture indicated how an inflammatory component of the tumor might arise
in vivo[
51,
52]. IP-10 (chemokine CXCL10) is an important mediator in bidirectional MSCs/breast cancer signaling [
53]. Its increase in the normoxic conditions and different AT-MSCs/SKBR3 coculture model further extends its importance in stromal/breast cancer interactions.
MSCs were also suggested to contribute to altered tumor drug resistance [
21,
22,
54,
55]. Recently the study by Roodhart
et al. demonstrated that cis-platin-preexposed MSCs mediated systemic resistance to cis-platin in tumor models including breast cancer cells MDA-MB-231 [
22]. However our experiments indicated that soluble factors present in the MSC-CM or the AT-MSCs concomitantly exposed to chemotherapeutic drug in direct coculture were not able to mediate chemoresistance (Figures
4 and
5). SKBR3 tumor cells in the presence of AT-MSCs had significantly increased sensitivity to chemotherapeutic drugs doxorubicin and 5FU that are frequently used for the breast cancer treatment. No significant difference in sensitivity to cis-platin (Figure
5C) or paclitaxel (data not shown) was detected when the AT-MSCs and tumor cells were exposed to the drug in cocultures. We believe that a concomitant exposure of stromal and tumor cells to the drug might actually increase the treatment efficiency. Contrastingly the exposure of (circulating) MSCs to the chemotherapy might induce secretion of mediators which subsequently contributed to increased tumor cell resistance [
22,
55]. It remains to be further evaluated, which mechanisms are drug-specific, tumor cell type-specific or context specific. Taken together the mutual tumor/stromal interactions do not only determine the biological behavior of tumor as a complex organ, but also its response to the chemotherapeutic treatment. The effects of MSCs on tumor cells are multiple and depend on the state of the tumor cell (dormant vs. actively-proliferating), the properties of specific MSCs populations, and interactions with other cell types, such as tumor infiltrating immune cells origin [
56]. It is important to focus on the evaluation of interactions of MSCs with primary tumor cells to shed more light into the operating interactions and signaling pathways.
Competing interests
The authors declare that they have no competing interests.
Authors’ contribution
Concept, design and development of methodology: LK, MM, MB; acquisition of data LK, SS, MM, ZK; analysis and interpretation of data LK, MM, SS, ZK; writing of the manuscript and review: LK, SS; technical and material support: MB. All authors read and approved the final manuscript.