SEARCH

SEARCH BY CITATION

Keywords:

  • HER2;
  • cancer stem cells;
  • salinomycin

Abstract

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

A major obstacle in the successful treatment of cancer is the occurrence of chemoresistance. Cancer cells surviving chemotherapy and giving rise to a recurrence of the tumor are termed cancer stem cells and can be identified by elevated levels of certain stem cell markers. Eradication of this cell population is a priority objective in cancer therapy. Here, we report elevated levels of stem cell markers in MCF-7 mammospheres. Likewise, an upregulation of HER2 and its differential expression within individual cells of mammospheres was observed. Sorting for HER2high and HER2low cells revealed an upregulation of stem cell markers NANOG, OCT4 and SOX2 in the HER2low cell fraction. Accordingly, HER2low cells also showed reduced proliferation, ductal-like outgrowths and an increased number of colonies in matrigel. Xenografts from subcutaneously injected HER2low sorted cells exihibited earlier onset but slower growth of tumors and an increase in stem cell markers compared to tumors developed from the HER2high fraction. Treatment of mammospheres with salinomycin reduced the expression of SOX2 indicating a selective targeting of cancer stem cells. Trastuzumab however, did not reduce the expression of SOX2 in mammospheres. Furthermore, a combinatorial treatment of mammospheres with trastuzumab and salinomycin was superior to single treatment with each drug. Thus, targeting HER2 expressing tumors with anti-HER2 therapies will not necessarily eliminate cancer stem cells and may lead to a more aggressive cancer cell phenotype. Our study demonstrates efficient killing of both HER2 positive cells and cancer stem cells, hence opening a possibility for a new combinatorial treatment strategy.

Although past research has made major progress in the diagnostics and the treatment of breast cancer, it remains a frequent cause of death in women. Classical chemotherapeutic drugs are still primarily used in the clinic and reach response rates from 60 to 100% when administered in a primary or adjuvant setting.1, 2 In spite of this fact, in about 40% of breast cancer cases a relapse is seen accompanied by a decline in chemotherapy response rates to only 20%.3 One of the major reasons for the failure of cancer therapies is the acquired chemoresistance of cancer cells. Extensive research over the years has elucidated many mechanisms such as the constitutive activity and overexpression of growth factor receptors,4–7 the overexpression of members of ATP binding cassette (ABC) membrane transporter family (reviewed in Ref.8) and certain proteins and enzymes such as β tubulin III, thioredoxin, peroxiredoxin (reviewed in Ref.9).

Residual cancer cells after administration of chemotherapy are termed cancer stem cells or tumor-initiating cells because of their ability to give rise to new tumors and are therefore responsible for relapse.10, 11 Recent reports identified markers that are termed as cancer stem cell markers and hence define this subpopulation of tumor cells.11–14 For example, ALDH1 overexpressing leukemia and colonic cancer cells show resistance to cycloheximide.15 Also, CD44+/CD24 cells isolated from MCF-7 and MDA-MB-231 breast cancer cell lines have been linked to resistance to radiotherapy.16 Other cancer stem cell markers (e.g., NANOG, OCT4 and SOX2) have been shown to induce pluripotency in cells and therefore directly influence the stem cell potential. Besides, the addition of transcription factors OCT4 and SOX2 to the medium reprogrammed somatic stem cells to induce pluripotent stem cells.17 Furthermore, Ben-Porath et al. showed association of poorly differentiated breast cancers that are suspected to be enriched for cancer stem cells with NANOG, OCT4 and SOX2 (NOS markers) expression. Hence, indicating their potential as cancer stem cell markers.18–20

Tumors consist of heterogeneous cell populations making it difficult to identify cancer stem cells in a more homogeneous 2D monolayer. Spheroids, termed mammospheres in case of mammary cells, are the models close to tumors as they consist of heterogeneous cells and show differences in cellular properties compared to 2D culture conditions. This is possibly due to the different architectural phenotype of the culture system.21 In an in vitro 3D culture system an increased expression of stem cell markers was observed.13, 22, 23 Hence in toto, mammospheres are excellent systems to study cancer stem cells and their involvement in chemoresistance.

Interestingly, Pickl and Ries showed enhanced signaling of the receptor tyrosine kinase HER2 in mammospheres as compared to 2D monolayer culture.24 Moreover, an increase in cancer stem cells was shown in normal mammary epithelial cells after ectopic overexpression of HER2 making it a relevant target in studying chemoresistance.22 Amplification of the HER2 gene and overexpression of HER2 protein occurs in 20–25% of breast cancers and is often associated with poor prognosis.25–27 This receptor tyrosine kinase from the human epidermal growth factor receptor (EGFR) family regulates cell survival, proliferation, differentiation and migration making it a key molecule in studying breast cancer.28–32 Recent studies have correlated the expression of HER2 with an increase in stem cell markers and side population.12, 22, 29 Consequently, these cells expressing HER2 as well as stem cell markers also show sensitivity to anti-HER2 monoclonal antibody trastuzumab.22 In contrast, Li et al. showed that tumors treated with the HER2 inhibitor lapatinib did not significantly decrease the fraction of breast cancer stem cells by measuring the percentage of CD44+/CD24 fraction in vivo and the ability to form mammospheres in vitro.32 Likewise, Reim et al. showed the presence of a CD44high/CD24low/HER2low breast cancer stem cell population, which successfully circumvents therapies targeting HER2 overexpression.33

In our study, we demonstrate elevation of HER2 levels and its differential expression in individual cells of mammospheres. Sorting of HER2high and HER2low populations from 3D culture system revealed overexpression of stem cell markers such as NANOG, OCT4 and SOX2 and stem-cell-like properties in the HER2low cell fraction. In addition, xenografts of HER2low sorted cells from MCF-7 mammospheres showed elevated levels of stem-cell markers NANOG, OCT4 and SOX2 as compared to xenografts of HER2high confirming stem-cell-like properties of HER2low cells. Thus, the population surviving trastuzumab therapy might be enriched in cancer stem cells and combinatorial treatment with anti-HER2 therapy and chemotherapy will not eradicate this population leading to chemoresistance. Hence, we put forward a new combinatorial treatment strategy using trastuzumab targeting HER2high cells and salinomycin targeting cancer stem cell population, i.e., HER2low cells. Our findings pave the way for novel treatment options, which will efficiently target cancer stem cells and HER2 positive cells from HER2 expressing breast tumors.

Material and Methods

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Reagents and antibodies

c-neu/Ab-2 (mouse monoclonal, OP14 Oncogene Research), Anti-erbB-2/Her2/neu (clone EP1045Y) (rabbit monoclonal, RM-2112-S0, Thermo scientific), Sox-2 (H-65) (rabbit polyclonal, sc 20088, Santa Cruz Biotechnology), Oct3/4 (C-10) (mouse monoclonal, sc 5279, Santa Cruz Biotechnology), CD44 (clone MEM 85) (mouse monoclonal Alexa 488 conjugated, A4-221-T100, Exbio antibodies), donkey anti mouse Cy3 (PA1-29773, Thermo Scientific), goat anti mouse Alexa 488 (A11001, Invitrogen), mouse anti rabbit Cy5 (211-172-171, Jackson Immunoresearch laboratories), donkey anti mouse Cy5 (715-175-151, Jackson Immunoresearch laboratories) 4′-6-diamidino-2-phenylindole (Sigma), Matrigel (BD Biosciences, 354248).

Cell culture and mammospheres generation

The cell lines MCF-7, BT-474 and MDA-MB-231 were obtained from cell line services (Eppelheim, Germany) and cultivated according to supplier's instructions. Briefly, MCF-7 cells were grown in DMEM medium supplemented with 20% fetal calf serum (FCS) and 2 mM glutamine (Gibco/Invitrogen, Karlsruhe, Germany). BT-474 and MDA-MB-231 cells were cultivated in RPMI 1640 medium and L-15 Leibovitz medium (Biocrom AG) respectively with 10% fetal calf serum (FCS) and 2 mM glutamine (Gibco/Invitrogen, Karlsruhe, Germany) respectively. Mammospheres were generated as described previously.23 Dishes were coated with 50 μg/ml Polyhema. Medium from the 2D cells grown to a confluence of 80–85% was taken as conditioned medium for growing mammospheres.

Cell viability assay

Five thousand cells per well of a 96 well plate were seeded for MCF-7 cell line. Mammospheres were divided in two parts. One half was trypsinized and cells were counted to a final concentration of 5000 cells per well. Accordingly, whole mammospheres were seeded in a 50 μg/ml Polyhema coated 96 well plate. Treatment was performed after 24 hrs of seeding for a period of 72 hrs (classical chemotherapeutics) and 7 days (trastuzumab and salinomycin). Combinatorial treatment was performed for 7 days. First cells were treated with trastuzumab followed by salinomycin treatment after 96 hrs. CellTiter-Glo aassay was performed for cell viability according to the manufacturer's instructions (CellTiter-Glo assay kit, Promega, G7571). The luminescence was recorded with a Luminometer (Lumat LB9507 instrument, Berthold, Bad Wildbad, Germany).

Cell lysis and immunoblotting

Cells were washed with PBS and incubated at 4°C with lysis buffer (50 mM HEPES/NaOH, pH 7.5, 150 mM NaCl, 1 mM EDTA, 10% glycerol and 1% Triton X-100) supplemented with phosphatase and protease inhibitors (10 mM sodium pyrophosphate, 1 mM PMSF, 2 mM sodium orthovanadate, 10 μg/ml aprotinin). Cellular debris was removed by centrifugation. Protein concentration measurements were performed using the Micro-BCA-Protein-Assay Kit from Pierce (Bonn, Germany). The samples were suspended in 3 × SDS sample buffer, boiled for 3 min and subjected to SDS-PAGE. For Western blot analysis, proteins were transferred to nitrocellulose membranes and incubated with the appropriate antibodies. Signals were developed via an enhanced chemiluminescence detection system (ECL, Perkin–Elmer, Wellesley, MA). Before reprobing, membranes were stripped with 65 mM Tris/HCl pH 6.8 buffer containing 2% SDS at 50°C for 1 hr.

Flow cytometry and fluorescence-activated cell sorting

Single cell suspension of 2D cell lines and 3D Mammospheres was prepared using non enzymatic cell dissociation solution (C5789, Sigma). For indirect immunofluorescence cells were fixed with 4% paraformaldehyde, permeabilized with 0.2% Triton X 100 and blocked with 10% FBS, 1% gelatin, 0.05% Triton X 100 in 1x PBS. Cells were incubated with indicated primary antibodies and labeled with appropriate fluorescence coupled secondary antibodies. Fluorescence was acquired using FACSCanto™ II flow cytometer (Becton Dickinson).

For cell sorting, single cell suspension was prepared under aseptic conditions. Cells were blocked using FBS for 30–40 min on ice. Incubations with primary (1:100) and secondary antibodies (1:200) were performed for 1 hr each on ice. Just before sorting of cells 20 U/ml DNAase I enzyme (10169, PEQLAB Biotechnologie GMBH) was added to prevent clumping of cells due to free genomic DNA of dead cells and 1 μg/ml propidium iodide (P4170, Sigma) was added to exclude the dead cells. Cell sorting was performed using indicated antibodies with a FACSAria™ II cell sorter (Becton Dickinson). Analysis and overlays of the FACS results are done using Flowjo 7.6.3. Doublets and dead cells were excluded when sorting.

Immunocytology and microscopy

Whole mammospheres were collected and fixed with 4% paraformaldehyde, blocked with 10% FBS, 1% gelatin and 0.05% Triton X 100 and incubated with indicated primary antibodies. Immunofluorescence was visualized using goat anti mouse Cy3 and counterstained with DAPI. Images were captured using 63 x 1.4 Oil DIC objective of Carl Zeiss Laser Scanning Microscope LSM 510 Meta and analyzed using Zeiss image browser version 3.2.0.115.

Quantitative real-time PCR

Total RNA was isolated using miRCURY RNA Isolation Kit (Exiqon) and transcribed with the Transcriptor High Fidelity cDNA Synthesis Kit (Roche) according to manufacturer's protocols. Quantitative real-time PCR was performed using UPL Probes (Roche) and Probes Master (Roche) on a LightCycler 480 system (Roche) with GAPDH as control. Primers used include GAPDH (UPL Probe 60) left: ctctgctcctcc tgttcgac, right: gcccaatacgaccaaatcc; HER2 (UPL Probe 4) left: gggaaacctggaactcaccta, right: ccctgcacctcctggata; NANOG (UPL Probe 69) left: atgcctcacacggagactgt, right: agggctgtcctgaataagca; OCT4 (UPL Probe 35) left: agcaaaacccggaggagt, right: ccacat cggcctgtgtatatc; SOX2 (UPL Probe 35) left: ttgctgcctctttaagact agga, right: ctggggctcaaacttctctc. Experiments were done in triplicates and the obtained average CT values of target genes were normalized to control as ΔCT. Changes in expression levels were shown either as fold increase or as ratio (target gene/ control).

Proliferation assay

Five thousand sorted cells of HER2high and HER2low fractions of MCF-7 2D and MCF-7 mammospheres (7 days old) were seeded and counted using Fuchs-Rosenthal counting chamber at day 2, 4, 6 and 8.

3D matrigel assay

Three-dimensional Matrigel assay was performed as described previously.34 Briefly, eight well chamber slides were coated with 75 μl of pure Matrigel (BD Biosciences, 354248) and allowed to polymerize at 37°C. Five thousand cells were added in a small volume of medium and allowed to adhere to the coating for 2–5 min. Then medium with 20% Matrigel was added on top of it. 50 μl of medium was supplemented once a week and the number of colonies (>20 μm) formed were counted. Phase contrast microscopic images with 10× objective were taken using Infinity capture (Infinity2-3C color 172765) software and were analysed using Axiovision Rel 3.1 software.

In vivo xenograft experiments

1 × 105 immunosorted MCF-7 HER2low and HER2high cells were injected subcutaneously without estrogen pellets into the right flank of 5–7 weeks old female Crl:SHO-Prkdcscid HRhr double homozygous SCID hairless out bred mice supplied by Charles River Laboratories international (seven mice per group). Before inoculation cells were resuspended in PBS and matrigel (1:1) (BD Biosciences). Tumor volume was measured once or twice a week using a calliper beginning at day 9 of injection. The volumes were calculated by the formula 0.5 × d2 × D (where d represents the smaller and D represents the larger diameter). For growth curve data points were expressed as average tumor volume [mm3] +/− SD per group (HER2low and HER2high). After 145 days, mice were sacrificed and tumors were resected for cryosections and qPCR. For xenografts of MCF-7 2D cell line and MCF-7 mammospheres, 1 × 106 single cells were injected subcutaneously without estrogen pellets into the right flank of female Crl:SHO-Prkdcscid HRhr double homozygous SCID hairless out bred mice supplied by Charles River Laboratories international (five mice per group). Mice were sacrificed after 170 days and cryosections were prepared from the resected tumors in tissue tek medium. All animal procedures were approved and controlled by the local ethics committee and carried out according to the guidelines of the German law of protection of animal life.

Immunohistochemistry of tissue sections

Immunohistochemical analysis was performed as described.35 Slides (5–7 μM sections) were incubated for 15 mins with 0.3% H2O2 in methanol (for Oct 4) and 1.5% H2O2 in PBS (for HER 2) to quench endogenous peroxidase. Subsequently sections were boiled in 10 mM sodium citrate buffer (pH: 6.0) for 15 mins (for HER 2) and 10 mins (for OCT 4) for antigen retrieval. Slides were blocked and incubated with relevant primary antibodies overnight at 4°C. Biotinylated secondary antibodies (Dako Cytomation) were applied to sections at 1:200 and incubated for 2 hrs at room temperature. ABC reagent (Vectastain Elite ABC Kit, Vector Labs) was applied to sections and developed in diaminobenzidine (DAB). Slides were then counterstained with haematoxylin. Primary antibodies against the following proteins were used: Rabbit anti HER 2, (1:100). Mouse Oct 4, (1:100). Control sections were processed without primary antibody. Images were captured via the Version 4.7 of the AxioVision image analysis software from Carl Zeiss Microscopy system.

Immunocytochemistry of tissue sections

Tissue sections (5–7 μM) were prepared from xenografts of MCF-7 2D cells and MCF-7 mammospheres injected subcutaneously in female immunodeficient mice after 170 days. The tissue tek layer was removed and the sections were fixed with 4% paraformaldehyde for 10–15 min. After blocking with 5% FBS in PBS twice for 10 min, HER2 antibody (1:200) was added for 2 hrs. Secondary antibody (1:400) and DAPI (1:1000) were added for an hr and sections were treated with fluor-safe. Results were captured using 63 × 1.4 Oil DIC objective of Carl Zeiss Laser Scanning Microscope LSM 510 Meta and analysed using Zeiss image browser version 3.2.0.115.

Results

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Mammospheres are resistant to classical chemotherapeutic drugs, enriched for stem cell markers and show elevated levels of HER2

It is well established that mammospheres resemble tumors because of cellular heterogeneity and enrichment for cancer stem cells that are associated with chemoresistance. Therefore, we exploited these mammospheres to study chemoresistance in comparison with 2D MCF-7 epithelial breast cancer cell line. The results of cell viability assays demonstrate increased resistance of mammospheres to classical chemotherapeutic drugs doxorubicin and paclitaxel as compared to cells cultured in 2D (Fig. 1a).

thumbnail image

Figure 1. Mammospheres are more resistant to classical chemotherapeutic drugs, are enriched for stem cell markers and show elevated levels of HER2 (a) MCF-7 2D cells and 7 days old mammospheres were treated with different concentrations of doxorubicin and paclitaxel for 72 hrs. CellTiter-Glo assay was performed to analyze the cell viability. The experiment was performed in triplicates. Student's t-test was performed to determine the significance (* p-values <0.05, ** p-values <0.01, *** p-values <0.001). (b) qPCR and FACS analysis of 2D cells and 7 days old mammospheres of MCF-7 to check expression of NANOG, OCT4 and SOX2 genes. Levels of these genes in mammospheres were normalized to the expression in 2D cells. FACS staining of mammospheres was performed for SOX2 using Alexa-488 tagged secondary antibody. Total SOX2 levels are shown in the overlay. (c) HER2 overexpression in 7 days old mammospheres as compared to 2D was shown with qPCR, SDS PAGE followed by western blot using total cell lysates and FACS staining. For western blot α-Tubulin was used as loading control. Total HER2 levels as determined by FACS analysis using Alexa-488 tagged secondary antibody are shown in the overlay. Student's t-test was performed to determine the significance (*** p-values <0.001).

Download figure to PowerPoint

Moreover, we demonstrate an overexpression of the stem cell markers NANOG, OCT4, SOX2 and CD44 in mammospheres (Fig. 1b and Supporting Information S1). These results indicate a correlation of stem cell markers and chemoresistance in mammospheres.

Pickl and Ries showed a difference in molecular signaling of HER2 (human epidermal growth factor receptor family protein) in mammospheres and 2D cell lines.24 Because HER2 is an extensively studied growth and survival marker in breast cancer and is associated with chemoresistance, it was intriguing for us to analyze the HER2 levels in 2D cells and mammospheres and to clarify whether HER2 contributes to the increased chemoresistance phenotype of mammospheres. Here, we show an upregulation of HER2 in mammospheres as compared to 2D cells of MCF-7 both at mRNA and protein levels (Fig. 1c). Additionally, tissue sections from xenografts obtained by injecting MCF-7 mammospheres in mice showed expression of HER2 which was absent in tissue sections of injected MCF-7 2D cells (Supporting Information Fig. S2).

Cell sorting of HER2high and HER2low cells from mammospheres

Laser scanning microscopic images of MCF-7 mammospheres stained for HER2 showed differential membrane expression in individual cells of mammospheres (Fig. 2a). Some cells exihibited strong membrane localized HER2 expression (white arrows) whereas some showed no or little HER2 expression (red arrows). To characterize these populations of mammospheres with respect to their stem-cell-like properties, we decided to sort these two cell fractions. Forward and sideward scatter plot of MCF-7 mammospheres showed presence of two distinct cell populations confirming the heterogeneity of mammospheres as compared to 2D cells (Supporting Information Fig. S4). We sorted 10% of the cell population containing highest HER2 levels as HER2high and 10% of cells with lowest HER2 expression as HER2low. Analysis of the location of these two fractions in forward scatter and sideward scatter plot of the whole population showed HER2low as a small cell fraction (Fig. 2b).

thumbnail image

Figure 2. Cell sorting of HER2high and HER2low cells from mammospheres (a) Laser scanning microscopic images of MCF-7 mammospheres stained for HER2 (red). Nucleus is stained with DAPI (blue). White arrows show membrane staining of HER2 and red arrows indicate HER2low cells. (b) Overview of the sorting strategy. From the whole cell population, 10% of the cells expressing highest HER2 levels (HER2high) and 10% of cells expressing lowest HER2 (HER2low) were sorted from mammospheres of MCF-7 grown for 7 days. Similarly these two fractions were also sorted from MCF-7 2D cells and MCF-7 mammospheres grown for 24 hrs. The upper graph displays cells stained with HER2 (FITC) and the autofluorescence (PE-Texas red). The lower right dot plot is the forward and the sideward scatter plot selectively demonstrating the HER2high and HER2low populations of the sorted cells. The lower left dot plot shows the forward and the sideward scatter plot of the whole cell population. HER2low is distinctly present in the small cell fraction and HER2high is primarily present in the big cell fraction as analyzed in forward scatter and sideward scatter plot. (c) qPCR for NANOG, OCT4 and SOX2 of HER2low population sorted from MCF-7 2D cells, mammospheres of MCF-7 grown for 24 hrs and 7 days. The graph shows expression levels of markers in the HER2low fraction normalized to expression levels in the HER2high fraction. The experiment was performed in triplicates with two biological replicates. Student's t-test was performed to determine the significance (* p-values <0.05, ** p-values <0.01).

Download figure to PowerPoint

The quantification of stem cell markers in HER2low versus HER2high cells demonstrated almost equal expression levels of NANOG, OCT4 and SOX2 in 2D as well as in the mammospheres grown for 24 hrs. In contrast, the mammospheres grown for 7 days expressed 4.5-fold, 5-fold and 2.5-fold more NANOG, OCT4 and SOX2 in HER2low cells, respectively. Hence, a time-dependent upregulation of NOS markers in HER2low cells of the mammospheres was observed suggesting an evolution of these markers over time during the formation of mammospheres (Fig. 2c).

Characterization of HER2high and HER2low cells from mammospheres confirm stem-cell-like properties in the HER2low cell population

Because there was an elevation of stem cell markers in HER2low population of the mammospheres, confirmation of their stem-cell-like phenotype was done by physiological assays. Proliferation analysis of these two populations shown in Figure 3a exhibits slow proliferation of HER2low cells in mammospheres of MCF-7. In matrigel, HER2low population displayed presence of ductal-like outgrowths within 7 days (Fig. 3b). Also, the number of colonies obtained in case of this population in matrigel, was higher than that obtained from HER2high (Fig. 3c). This data suggests that the HER2low population manifests properties similar to stem cells such as slow proliferation, ductal-like outgrowth in matrigel and higher colony formation ability.

thumbnail image

Figure 3. Characterization of HER2high and HER2low cells from mammospheres confirms stem-cell like properties in the HER2low cell population (a) Growth curve of HER2high and HER2low sorted populations of MCF-7 mammospheres. Five thousand cells were seeded and counted using Fuchs-Rosenthal counting chamber at day 0, 2, 4, 6 and 8. Student's t-test was performed to determine the significance (* p-values <0.05) (b) Matrigel outgrowths of HER2high and HER2low sorted populations of MCF-7 mammospheres. Five thousand cells were seeded on matrigel and the pictures were taken on day 7. (c) HER2high and HER2low sorted populations from MCF-7 mammospheres were grown on matrigel and the colonies (>20 μm) were counted manually after 7 days. The experiment was performed in triplicates. Student's t-test was performed to determine the significance (*** p-values <0.001). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

Characterization of xenografts of HER2low and HER2high sorted cells from MCF-7 mammospheres

As our in vitro data demonstrated stem-cell-like properties of the HER2low cell fraction, it was appealing to further investigate this finding in vivo. Hence, we injected HER2low and HER2high immunosorted cells from MCF-7 mammospheres subcutaneously in immunodeficient mice. The take rate was similar for both groups. However, when monitoring the growth of these tumors over 145 days, HER2low cell fraction formed visible tumors 6 weeks before tumors of HER2high cells (Fig. 4a). Additionally, tumors from HER2low cells demonstrated slow proliferation rate as compared to tumors from HER2high cells that is in accordance with Figure 3a. The confirmation of HER2 expression in xenografts was done by performing immunohistochemistry for the HER2 protein. HER2high tissue sections displayed some feebly expressed extra- cellular HER2 whereas HER2low showed no expression of HER2 depicting the authenticity of the sorted cells. Moreover, the HER2low cryosections showed abundant expression of OCT4 whereas HER2high showed no expression (Fig. 4b and Supporting Information S5). Besides these results, an upregulation in NANOG, OCT4 and SOX2 was seen at mRNA level too in xenografts of HER2low cells as compared to those of HER2high cells (Supporting Information Fig. S6). Thus, this data shows a slow growth but early onset of tumors from HER2low cells as compared to tumors from HER2high cells along with high expression of NOS markers in HER2low cryosections further indicating stem-cell-like characteristics of HER2low cells.

thumbnail image

Figure 4. Characterization of xenografts of HER2low and HER2high sorted cells from MCF-7 mammospheres (a) 1 × 105 immunosorted MCF-7 HER2low and HER2high cells were injected subcutaneously into the right flank of 5–7 weeks old female immunodeficient mice (seven mice per group). Before inoculation cells were resuspended in PBS and Matrigel (1:1). For HER2low xenografts, two mice developed tumors and for HER2high xenografts three mice developed tumors. The growth curve calculated over 145 days after injection is shown here. (b) Immunohistochemical analysis of cryosections of tumors from injected HER2low and HER2high cells stained with HER2 and OCT4 antibodies.

Download figure to PowerPoint

Targeted treatment of the distinct cell populations of mammospheres

As MCF-7 mammospheres consist of the cell fractions HER2high and HER2low, i.e., cancer stem cells, we treated them using two target specific drugs. On the one hand trastuzumab was used to treat the HER2high subset as it is commonly used in clinics for treating HER2-positive breast cancer patients. In contrast, Gupta et al. have recently found that salinomycin specifically kills the cancer stem cell population of tumors.36 Therefore, we analyzed whether it is effective against the HER2low subset. As expected, HER2high cells exhibited a dose dependent response to trastuzumab. On the contrary, HER2low cells were unaffected by trastuzumab treatment. Salinomycin showed exactly the opposite effect as it killed HER2low cells in a dose dependent manner (Figs. 5a and 5b). Accordingly, the salinomycin sensitive HER2low fraction was more resistant to the classical chemotherapeutic drug doxorubicin than the HER2high fraction (Supporting Information Fig. S7). Thus, we further confirmed our hypothesis that HER2low fraction constitutes cancer stem cells that are effectively killed by salinomycin. FACS analysis of untreated and treated cells of mammospheres stained for SOX2 showed a decrease in expression of SOX2 in case of salinomycin treated cells as compared to untreated as well as trastuzumab treated mammospheres. These findings further support a specific effect of salinomycin against cancer stem cells. As expected, trastuzumab treated cells show an equal expression of these markers (Fig. 5c).

thumbnail image

Figure 5. Targeted treatment of the distinct cell populations (a) HER2high and HER2low sorted cell populations of MCF-7 mammospheres were treated with trastuzumab (0.5, 5 and 50 μg/ml) for 7 days. A cytotoxicity assay was performed subsequently using CellTiter-Glo reagent. Student's t-test was performed to determine the significance (*** p-values <0.001, ** p-values <0.01). (b) HER2high and HER2low sorted cell populations of MCF-7 mammospheres were treated with salinomycin (0.5, 0.25 and 0.1 μM) for 7 days. A cytotoxicity assay was performed subsequently using CellTiter-Glo. Student's t-test was performed to determine the significance (*** p-values <0.001, * p-values <0.05). (c) FACS analysis of mock treated, trastuzumab treated (50 μg/ml) and salinomycin treated (0.1 μM and 0.25 μM) MCF-7 mammospheres is shown. The treatment was performed for 7 days. Staining was performed using secondary antibody tagged with Cy5 for SOX2 and Alexa-488 for HER2. Quantification of SOX2 positive fraction (Q1) is shown in the bar graphs. Student's t-test was performed to determine the significance (n.s.- not significant, * p-values <0.05).

Download figure to PowerPoint

Combinatorial treatment with trastuzumab and salinomycin leads to enhanced cell death in mammospheres

As trastuzumab eliminates HER2high cancer cells and salinomycin targets cancer stem cells, we further investigated the effect of combinatorial treatment of these two drugs on mammospheres. As a control we treated mammospheres of a triple negative cell line MDA-MB-231 with different concentrations of trastuzumab and salinomycin individually as well as in combination (Fig. 6a). We compared the most effective single treatment, in this case 0.5 μM salinomycin to the combinatorial treatment. As expected, there was no synergistic effect of the combination. However, treatment of MCF-7 mammospheres with a combination of trastuzumab and salinomycin demonstrated more cell death in comparison with single treatments (Fig. 6b). The beneficial effect of combination is evident for both combinations with a more significant effect observed when a higher dose of trastuzumab was used for the combination. Furthermore, we treated mammospheres of the cell line BT-474 in similar manner (Fig. 6c). Because this cell line expresses higher levels of HER2 protein, low dose of trastuzumab did not exhibit any significant effect individually or in combination. In contrast, a higher concentration of trastuzumab (100 μg/ml) eliminated almost 40% of the cells from BT-474 mammospheres. Thus, significance of the combinatorial treatment was compared to this trastuzumab concentration. Like MCF-7 mammospheres, BT-474 mammospheres manifested a beneficial effect of combinatorial treatment over single treatments. This data supports a novel treatment approach for HER2-positive tumors, which may contain cancer stem cells in the HER2low fraction which are showing resistance to trastuzumab treatment.

thumbnail image

Figure 6. Combinatorial treatment with trastuzumab and salinomycin leads to enhanced cell death in mammospheres (a) Cell titer glo assay of MDA-MB-231 mammospheres was performed after treatment with 5 μg/ml, 50 μg/ml trastuzumab and 0.5 μM salinomycin individually and in combination. Student's t-test was performed to determine the significance (n.s. not significant). (b) MCF-7 mammospheres were treated with 5 μg/ml, 50 μg/ml trastuzumab and 0.5 μM salinomycin individually and in combination. The cell viability was determined by cellTiter-Glo assay. Student's t-test was performed to determine the significance (*p-values <0.05, *** p-values <0.001). (c) BT-474 mammospheres were treated with 5 μg/ml, 100 μg/ml trastuzumab and 0.25 μM salinomycin individually and in combination. The cell viability was determined by cell titer glo assay. Student's t-test was performed to determine the significance (n.s. not significant, ** p-values <0.01). Tra represents trastuzumab treatment and Sal represents salinomycin treatment. For Student's t-tests in general, the most effective single treatment was compared to the combinatorial treatments.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

With the emerging concept of cells harbouring stem-cell-like properties being involved in the relapse of breast cancer, it has become mandatory to discover new approaches of targeted therapies that will specifically eradicate this cancer cell population.37 Functional heterogeneity of the cells within a tumor or selection barriers such as treatment with chemotherapeutic drugs gives rise to diverse cancer cell populations. Amongst others, cells with stem-cell-like characteristics evolve, what could be a dedifferentiation process. These cells have therapeutic resistance properties, are different from the bulk tumors and must be effectively targeted to achieve definitive curative benefits. Dontu et al. showed an in vitro enrichment of this tumorigenic subpopulation through cultivation of mammospheres.38 Because these mammospheres constitute heterogeneous cell populations and are associated with overexpression of certain stem cell markers such as CD44+/CD24,23 we selected them as tools for our study. In our study, mammospheres reveal an elevation in NANOG, OCT4 and SOX2 levels. These genes are responsible for self-renewal of undifferentiated embryonic stem cells and are also associated with poorly differentiated breast tumors. Thus, they are regarded as cancer stem cell markers.17, 18 Along with this, we also show resistance of mammospheres to classical chemotherapeutics such as doxorubicin and paclitaxel suggesting an involvement of cancer stem cells in chemoresistance.

The EGFR family is one of the most studied receptor tyrosine kinase families involved in chemoresistance (reviewed in Ref.37). Likewise, signalling of HER2/ERBB2 and its activation state have been studied in different tumors with respect to nodal status and receptor positivity.40, 41 Additionally, the difference between and molecular signaling of HER2 protein in 2D cells as well as in mammospheres shows the flexibility of oncogene switching, which may play an important role in development of drug resistance.26 This association of HER2 or activated HER2 with chemoresistance as well as certain stem cell markers as ALDH1 and Notch1 makes it intriguing to investigate the function of this protein in mammospheres.26, 29 Here, we demonstrate elevated HER2 protein levels in mammospheres and xenografts from mammospheres injected in female immunodeficient mice compared to 2D cell lines and xenografts from 2D cells. Furthermore, there was a differential expression of HER2 within individual cells of mammospheres, which made it possible to sort for HER2high and HER2low populations. The differential expression might be due to the association of HER2 with anoikis resistance hence proving its importance as survival factor for mammospheres.42 Moreover, the primary adjuvant therapy for HER2 positive cancers is treatment with trastuzumab that targets HER2. Although it is one of the most effective treatments in oncology, a significant number of patients show resistance to this treatment. Prevention of trastuzumab binding to HER2, upregulation of HER2 downstream signaling pathways, signaling through alternate pathways and failure to trigger an immune mediated destruction of tumor cells are few mechanisms responsible for resistance of tumor cells to trastuzumab treatment (reviewed in Refs.43 and44). The presence of cancer stem cells in the HER2low population of a tumor can be one additional mechanism to acquire trastuzumab resistance. Here, we show for the first time an upregulation of the stem cell markers NANOG, OCT4 and SOX2 in the HER2low fraction along with physiological properties of stem cells such as slow proliferation, matrigel outgrowth and increased no. of matrigel colony formation. Also the tissue sections of xenografts from female immunodeficient mice developed from injection of HER2low cells exhibited augmented expression of NANOG, OCT4 and SOX2 as compared to xenografts of HER2high cells. Hence, abolishment of HER2 positive cells by trastuzumab could give rise to cancer stem cells expressing low or no HER2. Our findings therefore might explain the failure of response of some HER2 positive tumors to trastuzumab therapies and the poor prognosis of relapsed patients. Moreover, a recent study depicted that sorted CD44high/CD24low cells that are regarded as the stem cell fraction of the immunoselected MCF-7 cell line, show enrichment in HER2low cells. This also supports the fact that MCF-7 contains a cancer stem cell fraction that displays reduced HER2 expression.31

One step forward in the treatment of cancer stem cells was the discovery of salinomycin that significantly reduces cell viability of human breast cancer stem-cell-like cells resistant to chemotherapeutic drugs such as paclitaxel, doxorubicin, actinomycin D and campthotecin.37 Salinomycin is able to overcome ABC transporter mediated multidrug resistance in human leukaemia stem-cell-like cells that exhibit resistance to a broad spectrum of chemotherapeutic drugs.45 It is a transmembrane potassium ionophore that is rapidly embedded in biological membranes such as cytoplasmic and mitochondrial membranes.46, 47 As ABC transporters also constitute transmembrane macromolecules and it is unlikely that salinomycin becomes a substrate of ABC transporters. It also increases apoptosis via increased DNA damage and reduced p21 protein levels through proteasome activity.48 Furthermore, a recent publication shows potential of salinomycin in killing cancer stem cells by eradication of CD133+ subpopulations in colorectal cancer cell lines.49 We authenticate the same by demonstrating effectiveness of salinomycin against HER2low population that otherwise is resistant to trastuzumab treatment. Besides, if administered in combination with trastuzumab, it shows an additive effect of increased cell death in mammospheres of different cell lines. Although we are aware of the limitations of a mammosphere model with respect to penetration of the trastuzumab antibody, our results in Figure 5a demonstrates the inefficiency of trastuzumab killing HER2low and an elimination of HER2high cells grown in 2D. Additionally, in Figure 6c we demonstrate an efficient targeting of mammospheres of BT-474, a high HER2 expressing cell line, with trastuzumab. Therefore, this data justifies that HER2high cells will respond to trastuzumab when accessible and, as HER2low cells are unaffected by this treatment, a combinatorial treatment strategy is required. However, our in vitro data forms a basis to investigate and establish a similar effect in breast cancer patients.

In past two decades more than 30 new anticancer drugs have been introduced, but survival rates have improved only marginally for many types of cancers including breast cancer hence making it necessary to formulate new and efficient strategies of treatment.50 Here, we depict a model of heterogeneous tumors consisting of some cells showing HER2 elevation which can be treated using trastuzumab and some cells with low or no HER2 that behaves like cancer stem cells and can be targeted using salinomycin. Our study characterizes trastuzumab resistant cancer stem cell population from mammospheres that could be a cause of therapy resistance in clinics. Furthermore, we show for the first time a scheme targeting HER2high population as well as HER2low population from heterogeneous tumors. Hence, it opens up a new horizon for combinatorial therapy targeting most of cells and can be considered as a baseline to develop a novel treatment strategy for heterogeneous breast cancers.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The authors thank Melinda Kiss for excellent technical support.

References

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
IJC_27595_sm_SuppFig1.tif254KSupporting Information Figure 1
IJC_27595_sm_SuppFig2.tif7012KSupporting Information Figure 2
IJC_27595_sm_SuppFig4.tif5165KSupporting Information Figure 4
IJC_27595_sm_SuppFig5.tif12684KSupporting Information Figure 5
IJC_27595_sm_SuppFig6A.tif4833KSupporting Information Figure 6A
IJC_27595_sm_SuppFig6B.tif5009KSupporting Information Figure 6B
IJC_27595_sm_SuppFig6C.tif4833KSupporting Information Figure 6C
IJC_27595_sm_SuppFig7.tif37KSupporting Information Figure 7

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.