Vemurafenib resistance selects for highly

Cancer Letters 361 (2015) 86–96
Contents lists available at ScienceDirect
Cancer Letters
j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / c a n l e t
Original Articles
Vemurafenib resistance selects for highly malignant brain and
lung-metastasizing melanoma cells
Inna Zubrilov a, Orit Sagi-Assif a, Sivan Izraely a, Tsipi Meshel a, Shlomit Ben-Menahem a,
Ravit Ginat a, Metsada Pasmanik-Chor b, Clara Nahmias c, Pierre-Olivier Couraud d,e,f,
Dave S.B. Hoon g, Isaac P. Witz a,*
a
Department of Cell Research and Immunology, The George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel Aviv, Israel
Bioinformatics Unit, The George S. Wise Faculty of Life Science, Tel Aviv University, Tel-Aviv, Israel
c
Inserm U981, Institut Gustave Roussy, 94800 Villejuif, France
d
Inserm, U1016, Institut Cochin, Paris, France
e CNRS, UMR8104, Paris, France
f University Paris Descartes, Paris, France
g Department of Molecular Oncology, John Wayne Cancer Institute, Saint John’s Health Center, Santa Monica, CA, USA
b
A R T I C L E
I N F O
Article history:
Received 2 February 2015
Received in revised form 19 February 2015
Accepted 19 February 2015
Keywords:
Vemurafenib resistance
Cancer stem cells
Metastatic microenvironment
Biomarkers
A B S T R A C T
V600E being the most common mutation in BRAF, leads to constitutive activation of the MAPK signaling pathway. The majority of V600E BRAF positive melanoma patients treated with the BRAF inhibitor
vemurafenib showed initial good clinical responses but relapsed due to acquired resistance to the drug.
The aim of the present study was to identify possible biomarkers associated with the emergence of drug
resistant melanoma cells. To this end we analyzed the differential gene expression of vemurafenibsensitive and vemurafenib resistant brain and lung metastasizing melanoma cells. The major finding of
this study is that the in vitro induction of vemurafenib resistance in melanoma cells is associated with
an increased malignancy phenotype of these cells. Resistant cells expressed higher levels of genes coding
for cancer stem cell markers (JARID1B, CD271 and Fibronectin) as well as genes involved in drug resistance (ABCG2), cell invasion and promotion of metastasis (MMP-1 and MMP-2). We also showed that
drug-resistant melanoma cells adhere better to and transmigrate more efficiently through lung endothelial cells than drug-sensitive cells. The former cells also alter their microenvironment in a different
manner from that of drug-sensitive cells. Biomarkers and molecular mechanisms associated with drug
resistance may serve as targets for therapy of drug-resistant cancer.
© 2015 Elsevier Ireland Ltd. All rights reserved.
Introduction
The MAPK signaling pathway involves activation of BRAF which
phosphorylates and activates MEK which in turn phosphorylates and
activates ERK. These reactions result in activation of transcription
factors that regulate cell survival, proliferation and differentiation [1].
BRAF mutations have been found in different malignancies including melanoma. V600E is the most common mutation in BRAF
leading to constitutive activation of the MAPK signaling pathway
[2]. Several small molecule inhibitors targeting the V600E BRAF mutation such as vemurafenib were developed [3]. Treatment of V600E
BRAF positive metastatic melanoma with vemurafenib showed initial
good clinical responses. However most of the patients relapsed due
to acquired resistance [4].
* Corresponding author. Tel.: +972 3 640 6979; fax: +972 3 640 6974.
E-mail address: isaacw@tauex.tau.ac.il (I.P. Witz).
http://dx.doi.org/10.1016/j.canlet.2015.02.041
0304-3835/© 2015 Elsevier Ireland Ltd. All rights reserved.
Acquired drug resistance is one of the major obstacles in cancer
treatment and management [5,6]. Several approaches have been
adopted to overcome drug resistance, among them attempts to
detect novel markers that can be targeted on resistant cells [7–10].
We have previously generated xenograft human melanoma brain
metastasis models, consisting of local, cutaneous variants as well
as of brain and lung-metastasizing variants yielding either dormant
micrometastasis or overt metastasis. These cell lines comprise
BRAFV600E mutation. All the variants originated from single melanomas thus sharing a common genetic background. Genes that
are differentially expressed by these variants can, thus, be assigned to the differential malignancy phenotype of the different
variants [11]. Using these models we demonstrated that brainmetastasizing melanoma variants expressed a set of genes whose
expression pattern differed from that of cutaneous melanoma variants [11].
In this study we analyzed the differential gene expression of
vemurafenib-sensitive brain and lung metastasizing melanoma cells
and corresponding cells in which resistance to this bio-drug was
I. Zubrilov et al./Cancer Letters 361 (2015) 86–96
induced by repeated cycles of in vitro exposure to the drug. The
vemurafenib sensitive melanoma cells and their resistant counterparts originated from a single melanoma tumor having therefore
a common genetic background [11]. Any difference in gene expression between these metastatic variants can therefore be attributed
to the difference in the metastatic microenvironment they originated from (brain versus lungs) and their drug sensitivity/resistance
status.
87
Table 1
qRT-PCR oligonucleotide primers.
Gene
name
Reaction
specificity
Accession
no.
Sequence
IL1R1
Human
NM_000877.3
JARID1B
Human
NM_006618.3
CYR61
Human
NM_001554.4
MMP-1
Human
NM_002421.3
SPINK1
Human
NM_003122.4
CEACAM1
Human
NM_001712.4
ABCG2
Human
NM_004827.2
Nestin
Human
NM_006617.1
Oct4
Human
NM_002701.5
E-cadherin
Human
NM_004360.3
CCL17
Mouse
NM_011332.3
1.5 × 106 human melanoma cells comprising BRAFV600E mutation were plated in
normal growth medium until adherent. The medium was then removed and replaced with 5% FCS medium containing 5 μM Vemurafenib for 72 hrs. Melanoma
cells grown in 5% FCS medium containing the same amount of DMSO served as control.
Following incubation, cells were rinsed with fresh growth medium and cultured in
a drug-free medium for a week. This process was repeated 3 times, then the concentration of vemurafenib was elevated to 10 μM for two more cycles. At the end
of each cycle total cell death was examined using a MEBCYTO® Apoptosis Kit (MBL,
Woburn, MA) according to the manufacturers’ instructions. Melanoma cell variants were considered resistant when more than 70% of the cells survived the
treatment.
CCL22
Mouse
NM_009137.2
IL-1β
Mouse
NM_008361.3
TNF-α
Mouse
NM_013693.2
Flow cytometry
RS-9
Human
NM_001013.3
β2M
Human
NM_004048.2
β2M
Mouse
NM_009735.3
S – 5′-GGACATTACTATTGCGT
GGTAAG-3′
AS – 5′-TGCTTAAATATGGCTT
GTGCAT-3′
S – 5′-AGCAGACTGACCGAA
GCTCA-3′
AS – 5′-AATTCCATCTCGCTT
CCCTC-3′
S – 5′-CTTAACGAGGACTGCA
GCAA-3′
AS – 5′-GTCTGCCCTCTGACT
GAGCT-3′
S – 5′-GTGCCTGATGTGGCTC
AGTT-3′
AS – 5′-ATGGTCCACATCTGCT
CTTG-3′
S – 5′-CCAAGATATATGACCCT
GTCTGT-3′
AS – 5′-TTCTCAGCAAGGCCC
AGATT-3′
S – 5′-GTCACCTTGAATGTCAC
CTATG-3′
AS – 5′-TGGACGGTAATAGGT
GTCTG-3′
S – 5′-TGGCTTAGACTCAAGCA
CAGC-3′
AS – 5′-TCGTCCCTGCTTAGA
CATCC-3′
S – 5′-AAGATGTCCCTCAGC
CTGGA-3′
AS – 5′-GAGGGAAGTCTTGGA
GCCAC-3′
S – 5′-GAAGGAGAAGCTGGAG
CAAA-3′
AS – 5′-CATCGGCCTGTGTAT
ATCCC-3′
S – 5′-CTCAGAAGACAGAAGAGA
GACTG-3′
AS – 5′-GTCAGAGAGAAGACA
GAAGACTC-3′
S – 5′-ATCAGGAAGTTGGTG
AGCTG-3′
AS – 5′-CAGTCAGAAACACG
ATGGCA-3′
S – 5′-CTCGTCCTTCTTGCT
GTGGC-3′
AS – 5′-TCTTCCACATTGGCA
CCATA-3′
S – 5′-CAGGCAGGCAGTATC
ACTCA-3′
AS – 5′-GAGGATGGGCTCTTC
TTCAA-3′
S – 5′-AGTTCTATGGCCCAG
ACCCT-3′
AS – 5′-CACTTGGTGGTTTGC
TACGA-3′
S – 5′-CGGAGACCCTTCGAGA
AATCT-3′
AS – 5′-GCCCATACTCGCC
GATCA-3′
S – 5′-ATGTAAGCAGCATCAT
GGAG-3′
AS – 5′-AAGCAAGCAGAATTTG
GAAT-3′
S – 5′-CTGGTCTTTCTGGTGC
TTGT-3′
AS – 5′-GGCGTGAGTATACTTG
AATTTGAG-3′
Materials and methods
Cells
All human melanoma cells (YDFR.CB3, YDFR.SB3, YDFR.CB3CSL3) were grown
in RPMI 1640 medium supplemented with 10% heat-inactivated fetal calf serum (FCS),
2 mmol/ml L-glutamine, 100 units/ml penicillin, 0.1 mg/ml streptomycin, 12.5 units/
ml nystatin and 1% Hepes (Biological Industries, Beit-Haemek, Israel). Medium of
melanoma cells resistant to Vemurafenib was supplemented with 1 μM PLX-4032
(Vemurafenib) (Selleck, Houston, TX) dissolved in Dimethyl Sulfoxide (DMSO) (SigmaAldrich, St. Louis, MO). Medium of the non-resistant melanoma cells was
supplemented with the same amount of DMSO. Human embryonic kidney 293T cells
were maintained as described by Izraely et al. [12]. Immortalized human brain microvascular endothelial cells (hCMEC/D3) were maintained as described by Weksler
et al. [13]. Immortalized human pulmonary endothelial cells (hPMEC) were maintained as previously described by Unger et al. [14]. Cells were routinely cultured in
humidified air with 5% CO2 at 37 °C. The cultures were tested and determined to
be free of Mycoplasma.
Animals
Male athymic nude mice (BALB/c background) were purchased from Harlan Laboratories (Jerusalem, Israel). Mice were housed and maintained in laminar flow cabinets
under specific pathogen-free conditions in the animal quarters of Tel Aviv University and in accordance with current regulations and standards of the Israel Ministry
of Health. The mice were used in accordance with institutional guidelines when they
were 7 to 10 weeks old.
Orthotopic inoculation of tumor cells and in-vivo tumorigenicity assays
An orthotopic sub-dermal inoculation of nude mice and measurements of the
tumorigenic properties were performed as described previously by Izraely et al. [12].
Mice were sacrificed 6 weeks after inoculation and brain, lungs and liver were
harvested. The organs were immediately stored at −80 °C, until used for RNA
extraction.
Drug resistance assessment
Cells were detached with trypsin-EDTA (Biological Industries) into single cell
suspension. 5 × 105 cells/sample were incubated for 1 hr at 4 °C with primary antibodies: α-CCR4 (1 μg/sample, R&D systems, Minneapolis, MN), α-CD271 (0.5 μg/
sample, BioLegend, San Diego, CA), α-CD133 (0.5 μg/sample, Miltenyi Biotec, Bergisch
Gladbach, Germany), α-VCAM1 (2 μg/sample, BD Pharmingen™, San Jose, CA) or with
corresponding isotype controls. After washing, the cells were incubated for 45 minutes
at 4 °C with FITC-conjugated secondary antibody (1:50, Jackson Laboratories, Baltimore, MD). Following an additional wash the cells were suspended in 300 μl
phosphate-buffered saline (PBSX1) containing 0.1%NaN3. Antigen expression was determined using Becton Dickinson FACSort and CellQuest software. Baseline staining
was obtained by labeling the cells with appropriate isotype control.
S, Sense; AS, Anti-sense.
Quantitative real-time PCR (qRT-PCR)
Total RNA was extracted using EZ-RNA Total RNA Isolation Kit (Biological Industries) and processed to cDNA with the M-MLV Reverse Transcriptase (Ambion
Inc., Austin, TX). For cDNA amplification, primers were designed based on the GenBank
Nucleotide Database of the NCBI website (Table 1). Amplification reactions were performed with SYBR Green I (Thermo Fisher Scientific, Waltham, MA) in triplicates
in Rotor-gene 6000™ (Corbett life science, Hilden, Germany). PCR amplification was
performed over 40 cycles, 95 °C for 15 s, 59 °C for 20 s, 72 °C for 15 s. Detection of
88
I. Zubrilov et al./Cancer Letters 361 (2015) 86–96
Table 2
Uniformly altered genes in vemurafenib resistant versus sensitive melanoma cells.
RefSec number
Gene name
NM_004827
NM_130847
NM_001657
NM_001130046
NM_004360
NM_006851
NM_003484
NM_000576
NM_000877
NM_001199640
ABCG2
AMOTL1
AREG
CCL20
CDH1 (E-cadherin)
GLIPR1
HMGA2
IL1β
IL1R1
IL33
NM_000214
NM_006618
NM_002309
JAG1
KDM5B (JARID1B)
LIF
NR_029672
NR_029493
NR_029635
MIR128-1
MIR21
MIR221
NR_029636
MIR222
NM_001145938
NM_002422
NM_003122
MMP1
MMP3
SPINK1
Role in cancer
Fold change (FC)
Cancer drug resistance [49]
Angiogenesis via regulating endothelial cell function [20]
Cell proliferation, motility and invasion via EGFR binding [21]
Cancer development and metastasis [22]
Cell–cell adhesion and tumor suppression [23]
In correlation with invasive potential of melanoma [24]
Induction of neoplastic transformation and promotion of metastasis [25,26]
Tumor-mediated angiogenesis and stimulates tumor invasiveness [27,28]
IL-1β receptor, decreases the recognition of melanoma by immune cells [29]
Contributes to invasive behavior via epithelial-to-mesenchymal transdifferentiation in cancer cells [30]
Cancer cell growth, migration and invasion via Notch signaling [31]
Cancer stem cell marker in melanoma [32]
Promotes invasive tumor microenvironment and self-renewal of embryonic
stem cells [33,34]
Tumor suppressor and regulator of stem cell factors in prostate cancer [35]
Downregulates tumor suppressor genes [36]
Downregulates tumor suppressor genes and is involved in resistance
acquisition to numeral chemotherapies [37,38]
Downregulates tumor suppressor genes and is involved in resistance
acquisition to numeral chemotherapies [37,38]
Cancer invasion and metastasis [39,40]
Cancer invasion and metastasis [41]
Cancer aggressiveness and epithelial to mesenchymal transition [42,43]
CB3 vem
vs con
SB3 vem
vs con
CB3CSL3 vem
vs con
1.91
1.57
1.50
1.71
−1.84
1.60
1.31
8.15
2.20
1.78
1.19
1.38
1.24
1.93
−1.71
1.50
1.30
1.36
1.22
1.21
4.27
1.56
2.75
1.92
−4.11
16.79
1.48
7.37
2.78
2.22
1.39
1.23
3.24
1.47
1.29
1.49
1.48
1.43
3.33
−1.27
1.43
1.50
−1.30
1.37
2.42
−1.29
1.93
1.83
1.27
2.63
3.12
2.83
2.57
3.41
1.79
1.23
3.49
43.97
34.62
4.08
The table displays genes uniformly up or down-regulated (FC ≥ 1.25 or FC ≤ −1.25 respectively) in three melanoma cell variants (YDFR.CB3, YDFR.SB3, YDFR.CB3CSL3) comparing vemurafenib-resistant to the corresponding non-resistant cells, as obtained from GeneChip® Human Transcriptome Array 2.0 analysis. FC – Fold Change, vem – vemurafenibresistant cells, con – control vemurafenib-sensitive cells.
human cells (micro-metastases) in mouse tissue by qRT-PCR was performed as previously described by Izraely et al. [11].
Stable expression of mCherry vector
To produce the MLV infectious viruses, the 293T packaging cell line was cotransfected using calcium phosphate method with the retroviral backbone plasmid
mCherry-pQCXIP, packaging plasmid gag/pol and envelope plasmid pVSV-G (Clontech
Laboratories, Mountain View, CA). After 48 hrs of incubation, the virus particles in
the medium were collected and filtrated (0.45 μm, Millipore, Billerica, MA). 2 × 106
melanoma cells were infected in the presence of 8 μg/ml polybrene overnight. The
cells were infected for the second time with new virus particles for 4 hrs and then
the virus-containing medium was replaced with fresh medium. After 72 hrs 1 μg/ml
Puromycin (InvivoGen, Toulouse, France) was added for additional 7 days in order
Fig. 1. Melanoma cells resistant to vemurafenib show increased malignant phenotype in vitro. mRNA expression of the following genes: IL1R1, ABCG2, SPINK1, JARID1B,
MMP-1 and E-cadherin was tested in human melanoma cells resistant and sensitive to vemurafenib using qRT-PCR analysis. The obtained values were normalized to hβ2M.
All data are means of at least three independent experiments ± SD. Significance was evaluated using Student’s t-test for each vemurafenib-resistant variant compared to its
vemurafenib-sensitive counterpart, *p ≤ 0.05, **p ≤ 0.005, ***p ≤ 0.0005. Control – vemurafenib-sensitive cells, vemur resist – vemurafenib-resistant cells.
I. Zubrilov et al./Cancer Letters 361 (2015) 86–96
to select a stably infected cell population. After selection, Puromycin was continuously added to the cultures.
Adhesion to brain and lung endothelial cells
Wells of 96-well plates were coated with 100 μg/ml rat-tail collagen type I (BD
Pharmingen™) for 1 hr at 37 °C. Following one wash with PBSX1, 5 × 104 hCMEC/
D3 or hPMEC cells in 100 μl EBM2 or M-199 growth, medium, respectively were
cultured for 24 hrs to form a confluent monolayer. Then, endothelial cells were
activated with 0.1 μg/ml IFNγ and 0.1 μg/ml TNFα (PeproTech, Rocky Hill, CT) in
suitable starvation medium for additional 24 hrs. At the end of the incubation
period, cells were gently washed twice with PBSX1, and 1 × 105 mCherry-infected
89
melanoma cells suspended in tris-buffered saline (TBSX1), containing 2 mM CaCl2
and 1 mM MgCl2, were added onto the stimulated hCMEC/D3 or hPMEC monolayer. Cells were incubated for 30 minutes at 37 °C to allow adhesion to occur. The
total fluorescence signal of melanoma cells added to each well was measured by a
Synergy HT fluorescent multi-well plate reader (BioTek, Winooski, VT) at wavelength of 590/645 nm before removing the non-adherent cells. Then, the wells
were washed twice with PBSX1 to remove the non-adherent cells and fluorescence of the adherent cells was measured again. The fluorescence rate of the
adherent cells was normalized to the total fluorescence in the same well. Wells
containing only endothelial cells served as a blank control. The activation of the
endothelial cells with IFNγ and TNFα was monitored by measuring the expression
of VCAM-1 using flow cytometry [15].
Fig. 2. Vemurafenib-resistant melanoma cells exhibit a higher level of CCR4 and up-regulate the expression of its ligands CCL17 and CCL22 within the remote mice organs
in vivo. (A) Flow cytometry analysis of CCR4 in vemurafenib-resistant and non-resistant melanoma cells. The FACS diagrams shown are of a representative experiment. The
continuous line in the diagrams represents negative isotype control; the discontinuous line represents anti-CCR4 immunostaining. Significance was evaluated using Student’s t-test for each vemurafenib-resistant variant compared to its vemurafenib-sensitive counterpart. All data are means of at least three independent experiments ± SD.
(B) qRT-PCR analysis of CCL17 and CCL22 mRNA extracted from brain, lungs and liver of mice inoculated with vemurafenib resistant YDFR.SB3 cells, vemurafenib sensitive
YDFR.SB3 cells or healthy mice. The data shown are average values from each group of mice: 5, 5 and 3 respectively. Significance was evaluated using Student’s t-test comparing between each group of mice. *p ≤ 0.05, **p ≤ 0.005. Control – vemurafenib sensitive cells, vemur resist – vemurafenib resistant cells.
90
I. Zubrilov et al./Cancer Letters 361 (2015) 86–96
Cell migration through collagen layer
Trans-well inserts (6.5 mm diameter polycarbonate membrane with 8.0 μm pores,
Corning Costar Corp., New York, NY) were coated with 100 μg/ml collagen type I
for 1 hr at 37 °C. After washing the trans-wells with PBSX1 twice, 1 × 105 melanoma cells were added to the upper chamber and allowed to transmigrate for 24 hrs
after adding serum-free RPMI-1640 medium containing 2 mM CaCl2 and 1 mM MgCl2
to the lower chamber. At the end of the incubation period, cells in the upper chamber
were removed using cotton swabs. The bottom side of the trans-well inserts was
gently washed in PBSX1 and fixed in ice-cold methanol for 5 minutes, then the cells
were stained with Dif-stain Kit (Kaltek, Padova, Italy), according to the manufacturer’s instructions. The number of melanoma cells that migrated through the membrane
was determined by counting cells from five independent fields in duplicates using
OlympusIX53 light microscope (Olympus, Center Valley, PA).
Trans-endothelial migration
hCMEC/D3 or hPMEC (5 × 104 cells/200 μl) were plated on trans-well inserts, precoated with 100 μg/ml collagen type I, for 48 hrs, to form a confluent monolayer.
Medium was added only into the upper compartment to prevent the formation of endothelial bilayer [16,17]. 1 × 105 melanoma cells expressing mCherry
were added to the upper chamber and allowed to transmigrate for 24 hrs after adding
serum-free RPMI-1640 medium containing 2 mM CaCl2 and 1 mM MgCl2 to the lower
chamber. At the end of the incubation period, cells in the upper chamber were
removed using cotton swabs. The bottom side of the transwell inserts was gently
washed in PBSX1 and fixed in 4% paraformaldehyde for 15 minutes. Then, the
transwell inserts were washed again in PBSX1 and stained with DAPI using Dapi–
Fluoromount–GTM solution (SouthernBiotech, Birmingham, AL). The number of
melanoma cells that migrated through the membrane was determined by counting
five independent fields under fluorescence microscopy (Olympus) in duplicates.
Hoechst staining
1 × 104 melanoma cells were plated in 96-well plates for 24 hrs in triplicates.
Adherent cells were incubated with 5 μg/ml of Hoechst 33342 dye (Sigma-Aldrich)
in 100 μl RPMI medium for 60 minutes at 37 °C. After the incubation, cells were
rinsed twice with growth medium and PBSX1 was added for measurement in
Synergy HT fluorescent multi-well plate reader at wavelength of 360/450 nm [18].
The results were normalized to the amount of cells as obtained from XTT assay
(Cell Proliferation Kit, Biological Industries) that was performed simultaneously
according to manufacturer’s instructions.
mRNA microarray
mRNA of Vemurafenib-resistant and sensitive melanoma cells was hybridized
to GeneChip® Human Transcriptome Array 2.0 (Affymetrix, Santa Clara, CA). Partek
Genomics Suite™ software (version 6.5; Partek, St. Louis, MO; http://www.partek.com/
pgs) was used for microarray analysis. Raw data (CEL files) were normalized at
the transcript level using robust multiaverage method [19]. Median summarization of transcript expressions was calculated. Gene-level data were then filtered to
include only those probe sets that are in the “core” meta-probe list, which represent RefSeq genes and full-length GenBank mRNAs. Fold change (FC) cutoff of ≥1.25
or ≤−1.25 was performed to obtain differentially expressed genes for the various
conditions.
Gelatin zymography and western blotting analysis
4.5 × 105 melanoma cells were plated in 24-well plates for 24 hrs, then growth
medium was removed and replaced by 400 μl serum-free RPMI-1640 medium for
additional 24 hrs. For Matrix Metalloproteinase-2 (MMP-2) activity detection, conditioned medium was collected and prepared without boiling or reduction, then
subjected to SDS polyacrylamide gel electrophoresis containing 0.1% gelatin. After
electrophoresis, the gels were washed for 15 minutes 4 times with reaction buffer
(50 mM Tris–HCl, pH 7.5, 200 mM NaCl, 10 mM CaCl2, 5 μM ZnCl2, 0.02% NaN3) containing 2.5% Triton X-100 and incubated in reaction buffer at 37 °C overnight. Finally,
the gels were stained with Coomassie Blue (BioRad, Hercules, CA) and destained in
a 20% methanol, 10% acetic acid solution. The results were validated by Western Blot
analysis using α-MMP-2 antibody (1:200; R&D systems), as described previously [11].
The obtained MMP-2 activity values were normalized to the amount of viable cells
in each well.
Fig. 3. Vemurafenib resistant cells are enriched for cancer stem cell markers. (A) Flow cytometry analysis of CD271. The FACS diagrams shown are of a representative experiment. The continuous line in the diagrams represents negative isotype control; the discontinuous line represents anti-CD271 immunostaining. (B) qRT-PCR analysis of
FN1 mRNA expression in human melanoma cells resistant and sensitive to vemurafenib. The obtained values were normalized to hβ2M. (C) Uptake of Hoechst 33342 dye
by different melanoma variants. The results were normalized to the amount of cells as obtained from XTT assay. All data are means of at least three independent experiments ± SD. Significance was evaluated using Student’s t-test for each vemurafenib-resistant variant compared to its vemurafenib-sensitive counterpart, *p ≤ 0.05, ***p ≤ 0.0005.
Control – vemurafenib-sensitive cells, vemur resist – vemurafenib-resistant cells.
I. Zubrilov et al./Cancer Letters 361 (2015) 86–96
Statistical analysis
Paired or unpaired Student’s t-test was used to compare in vitro and in vivo results.
Results
Vemurafenib resistance of melanoma cells is associated with an
altered gene expression profile
In previous studies we generated variants of human
melanoma cells that metastasize to the brain and lungs of
xenotransplanted nude mice [11]. In this study we utilized variants that metastasize specifically and spontaneously to brain and
lungs of nude mice forming micro-metastasis (YDFR.SB3 and
YDFR.CB3CSL3, respectively) in these organs following an orthotopic sub-dermal inoculation. We also used a variant that generates
brain macro-metastasis following an intra-cardiac inoculation
(YDFR.CB3). All three variants carry a BRAFV600E mutation. Exposing these cells in vitro to several cycles of increasing concentrations
of vemurafenib we generated drug-resistant melanoma cells. Using
expression microarray analysis we then compared the mRNA expression profile of resistant and sensitive cells. Analyzing the results,
we filtered out genes that expressed a change fold of ≥1.25 or ≤−1.25
between the variants, focusing on a cluster of genes that uniformly changed in vemurafenib resistant versus sensitive cells. The
most prominent genes that changed in the vemurafenib-resistant
cells were those involved in promotion of metastasis, drug resistance, cancer stem cell markers and cell invasion (Table 2). The
alteration of the following genes: ABCG2, E-cadherin, IL1R1, JARID1B,
MMP1, SPINK1 was validated using qRT-PCR (Fig. 1).
In a previous study we demonstrated that an up-regulated
expression of the chemokine receptor CCR4 on human melanoma
91
cells is associated with brain metastasis in nude mice xenografted
with human-melanoma cells. It was also shown that the expression of this chemokine receptor is regulated by the brain
microenvironment [44,45]. Results of the present study indicated
that CCR4 was significantly up-regulated on vemurafenib resistant cells (p < 0.05) (Fig. 2A).
Since metastasis is regulated to a significant part by interactions of tumor cells with their microenvironment [46,47] we
hypothesized that the interaction of CCR4 on melanoma cells with
the corresponding ligands (CCL17 and CCL22) in the metastatic
microenvironment could be functionally involved in metastasis formation [44].
Six weeks after inoculation, mice bearing sub-cutaneous tumors
generated by both vemurafenib-sensitive as well as resistant
melanoma cells showed an up-regulated CCL17 expression in the
lungs compared to lungs of healthy mice (p < 0.05) (Fig. 2B). The brain
and liver of mice bearing tumors formed by vemurafenib-resistant
cells expressed significantly higher levels of CCL17 than vemurafenibsensitive cells (p < 0.05). CCL22 expression in the brain of mice
bearing tumors generated by vemurafenib-resistant melanoma cells
was also higher than that in brains of mice bearing tumors generated by vemurafenib-sensitive cells (p < 0.05).
Vemurafenib-resistant cells are enriched for characteristics of
side-population tumor cells
It is widely accepted that cancer stem cells resist various anticancer treatments [6,48]. We therefore analyzed the expression of
cancer stem cell markers in vemurafenib sensitive and resistant cells.
Figures 1 and 3A,B show that the resistant cells were enriched for
cells bearing several stem cell markers such as JARID1B, fibronectin
Fig. 4. MMP-2 activity and cell migration through collagen layer is more efficient in vemurafenib resistant cells. (A) The activity of MMP-2 secreted from melanoma cells
was measured by gelatin zymography protease assay. The obtained data were normalized to the amount of cells per well. The results were validated by Western Blot analysis. A representative blot is presented. (B) Cell migration through collagen layer was tested using trans-well inserts coated with collagen type I. The images shown are of
a representative experiment. All data are means of at least three independent experiments ± SD. Significance was evaluated using Student’s t-test for each vemurafenibresistant variant compared to its vemurafenib-sensitive counterpart, *p ≤ 0.05, **p ≤ 0.005. Control – vemurafenib-sensitive cells, vemur resist/vemur – vemurafenibresistant cells.
92
I. Zubrilov et al./Cancer Letters 361 (2015) 86–96
Fig. 5. The migration of vemurafenib-resistant melanoma cells through brain and lung vascular endothelial layer is enhanced. (A) Melanoma cell adhesion to brain or lung
vascular endothelial cell layer was tested by seeding mCherry expressing melanoma cells on adhered activated endothelial cells for 30 mins. Percentage of the adhered
melanoma cells was determined by dividing the fluorescence rate after washing the wells by total fluorescence rate before washing. (B) Melanoma cell migration through
brain or lung vascular endothelial cell layer was tested using trans-well inserts coated with collagen type I, on which endothelial cells were plated two days prior to adding
mCherry expressing melanoma cells. Representative images are presented. Red staining – mCherry transfected melanoma cells, blue staining – DAPI. All data are means of
at least three independent experiments ± SD. Significance was evaluated using Student’s t-test for each vemurafenib-resistant variant compared to its vemurafenibsensitive counterpart, *p ≤ 0.05, **p ≤ 0.005. Control – vemurafenib-sensitive cells, vemur resist – vemurafenib-resistant cells. (For interpretation of the references to color
in this figure legend, the reader is referred to the web version of this article.)
I. Zubrilov et al./Cancer Letters 361 (2015) 86–96
Table 3
Adhesion of melanoma cells to endothelia and TEM – summary.
Vemurafenib-resistant versus sensitive melanoma cells
CB3
SB3
CB3CSL3
Adhesion to
brain endothelia
NSD
Adhesion to
lung endothelia
NSD
TEM – brain
endothelia
NSD
TEM – lung
endothelia
The table summarizes the results of adhesion of melanoma cells to brain and lung
vascular endothelial cells and trans-migration of melanoma cells through monolayers of these endothelia (Fig. 4). The results shown are of vemurafenib-resistant
cells relative to their non-resistant counterparts. TEM = trans-endothelial migration, NSD = no significant difference.
and CD271. The resistant cell variants also expressed elevated levels
of ATP-binding cassette transporter ABCG2 (Fig. 1). This cellular transporter is directly involved in cancer drug resistance [49]. The
chemotherapy extrusion by ABCG2 transporters correlates with the
93
extrusion of the Hoechst 33342 dye. Cancer cells that efficiently
exclude this dye are referred to as side-population cells. These cells
share characteristics of cancer stem cells [50]. Figure 3C demonstrates that vemurafenib-resistant cells exhibited increased extrusion
of the Hoechst 33342 dye.
The in vitro malignancy phenotype of vemurafenib-resistant
melanoma cells
The 3 vemurafenib-resistant melanoma variants and their original vemurafenib-sensitive variants were tested for several
malignancy-linked in vitro functions including MMP secretion, migration through collagen coated membranes, adhesion to brain and
lung endothelium and trans-migration through monolayers of these
endothelia. Figure 4 indicates that vemurafenib resistant cells express
a significant increased MMP activity and a higher transmigration
through collagen coated membranes than sensitive cells (p < 0.05).
Figure 5A shows that vemurafenib-resistant brain metastasizing melanoma cells (both macro and micro-metastases) adhere significantly
better than sensitive cells to lung endothelium (p < 0.05), whereas
they adhere significantly less efficiently than vemurafenib-sensitive
cells to brain endothelial cells (p < 0.05). Figure 5B demonstrates that
vemurafenib-resistant brain metastasizing melanoma cells
Fig. 6. The tumorigenicity of vemurafenib-resistant melanoma cells is increased. (A) The graph shows average tumor volume of mice inoculated sub-cutaneously with YDFR.SB3
vemurafenib-resistant and non-resistant melanoma cells. Each group consisted of 13 mice. Error bars depict standard error around the mean. Representative pictures of
mice from each group 4, 5 and 6 weeks after inoculation are shown. Significance was evaluated using Student’s t-test for tumor volume of mice inoculated with vemurafenib
resistant cells compared to mice inoculated with their non-resistant equivalents, *p ≤ 0.05. (B) Metastasis formation in the remote organs of mice after orthotopic inoculation with vemurafenib resistant or sensitive melanoma cells. The metastasis presence was determined by qRT-PCR analysis.
94
I. Zubrilov et al./Cancer Letters 361 (2015) 86–96
Fig. 7. The metastatic cascade is enhanced in vemurafenib-resistant melanoma cells. Vemurafenib-resistant melanoma cells express decreased E-cadherin levels and increased levels of CD271, CCR4, JARID1B, MMP-2 and ABCG2 compared to their non-resistant counterparts. The resistant cells also generate larger primary tumors in nude
mice. Moreover, mRNA expression of CCL17 in brain and liver of mice inoculated with vemurafenib-resistant cells was elevated. The in vitro results showed increased transendothelial migration through both brain and lung vascular endothelial monolayer by vemurafenib-resistant cells and increased ability to adhere by these cells to lung, but
not to brain endothelia. Correspondingly, the capacity of the vemurafenib-resistant melanoma cells to metastasize to the lungs, but not to the brain was increased, comparing to vemurafenib-sensitive melanoma cells.
transmigrate through both brain and lung vascular endothelial cells
much more efficiently than their sensitive counterparts (p < 0.05).
Table 3 summarizes these results: The in vitro ability of melanoma
cells to adhere to brain and lung endothelia does not predict ability
to trans-migrate through these endothelial layers.
The in vivo malignancy phenotype of vemurafenib-resistant
melanoma cells
1 × 106 vemurafenib-sensitive or resistant micrometastatic melanoma cells (YDFR.SB3) were inoculated sub-dermally into nude mice.
The formation of local tumors was followed up to 6 weeks following inoculation. Figure 6A shows that the tumors formed by
vemurafenib-resistant cells were significantly larger than tumors
formed by vemurafenib-sensitive cells.
We reported previously that melanoma cells inoculated subdermally into nude mice form spontaneous micro-metastasis [11].
The capacity of vemurafenib-sensitive or resistant melanoma cells
to form micro-metastasis in the brain and in the lungs was determined by qRT-PCR analysis of human cells in these organs [11].
Figure 6B shows that whereas the vemurafenib-sensitive or resistant melanoma cells did not differ in their ability to form brain
micro-metastasis, the capacity of the latter cells to form lung and
liver micro-metastasis was increased.
We expected that an up-regulated expression of CCR4 by
vemurafenib resistant melanoma cells and an increased expression of the 2 CCR4 ligands in the brain of mice (as reported in section
“Vemurafenib resistance of melanoma cells is associated with an
altered gene expression profile”) will lead to an augmentation of
the metastatic load in the brain of mice inoculated with these cells
[47]. However the results summarized in this section indicated that
the load of brain metastasis in mice bearing vemurafenib resistant cells was not higher than that of mice bearing vemurafenibsensitive cells. These results, while not discarding the possible
involvement of the CCR4-CCL17/CCL22 axis in brain metastasis [47]
support the findings that this axis is only one of several determinants of melanoma brain metastasis [11,12].
Discussion
A major finding of this study is that the in vitro induction of
vemurafenib-resistance of melanoma cells is associated with an increased malignancy phenotype of these cells. It is not unlikely that
selecting for drug resistance selects for tumor and metastasis initiating cells [51,52]. This possibility is supported by the findings that
vemurafenib-resistant cells express higher levels of certain stem cell
markers and that these cells also express higher levels of ABCG2
functioning as a key molecule in multidrug-resistance of cancer cells
by mediating efflux of various compounds, from such cells [53,54].
Another important finding is that vemurafenib-resistant melanoma cells express novel characteristics such as gene products and
functions that are not expressed by the drug-sensitive cells. This
raises the possibility of attacking resistant cells by specifically targeting these markers.
Figure 7 shows a model summarizing the phenotypic and functional alterations that may accompany vemurafenib-resistant
melanoma cells. Vemurafenib-resistant melanoma cells show downregulation of E-cadherin and up-regulation of molecules involved
I. Zubrilov et al./Cancer Letters 361 (2015) 86–96
in promotion of metastasis and genes coding for cancer stem cell
markers such as MMP-2, CD271, CCR4, Fibronectin, JARID1B and
ABCG2. The resistant cells also generate larger primary tumors in
nude mice. The in vitro trans-endothelial migration by vemurafenibresistant cells through both brain and lung vascular endothelial
monolayer was significantly intensified. However, only the ability
to adhere to lung, but not to brain endothelia was increased in
vemurafenib-resistant cells. Correspondingly, the capacity of the
vemurafenib-resistant melanoma cells to metastasize to the lungs,
but not to the brain was increased, as compared to vemurafenibsensitive melanoma cells.
Our results show that an efficient trans-endothelial migration
in vitro does not predict metastasis formation in vivo. However, there
is a correlation between the ability of melanoma cells to adhere to
the endothelium of a specific organ and the capacity to form metastasis within the parenchyma of that organ.
In a recent publication, Paulitschke et al. [55] established a
proteome signature of vemurafenib resistant melanoma cells. Some
of the results reported above which were obtained with
vemurafenib-resistant and sensitive cell variants that originated in
the same parent population confirmed their findings. For example
our results indicate that vemurafenib-resistant cells express an increased potential for metastasis, higher migratory and adherence
functions, down-regulated levels of E-cadherin and higher levels of
MMP-2 activity than vemurafenib-sensitive cells.
Numerous attempts are being undertaken to overcome drug resistance [56–59]. Here we suggest that intervening in the interactions
between drug-resistant tumor cells and their specific microenvironment may constitute an additional approach to overcome
drug resistance. It had been repeatedly shown that tumormicroenvironment interactions are a crucial component in the
survival, propagation and progression of cancer cells [46,60,61]. It
is therefore not unlikely that the microenvironment of drugresistant cancer cells could be harnessed to overcome drug resistance.
However such attempts have not been sufficiently explored. The observation that vemurafenib-resistant cells reprogram their
microenvironment in a different manner from that of drug-sensitive
cells suggest that different signaling pathways may operate in interactions between drug-sensitive and drug resistant cancer cells
with their respective microenvironments. This may provide the possibility to selectively manipulate and target the signaling pathways
of the resistant cells leading to their growth arrest or apoptosis.
Organ specificity of metastasis i.e. the differential biological
behavior of cancer cells metastasizing to different organ microenvironments is another point that comes up when analyzing the
results of this study. As demonstrated brain and lung metastasizing melanoma cells originating from the same melanoma may
express differences in the expression of certain molecules and
may interact differentially with microenvironmental components.
These results support the notion that metastasis cannot be characterized in generic terms (such as metastatic melanoma or
metastatic breast cancer) [62–64]. Characterizing the specific and
unique molecular phenotype, biologic behavior, response to
treatment etc. of cancer cells that metastasize to different organ
sites is essential in the era of personalized medicine [60,65–69].
Acknowledgements
This study was supported by The Dr. Miriam and Sheldon G.
Adelson Medical Research Foundation (04-7023433) (Needham, MA).
Conflict of interest
None.
95
References
[1] C. Peyssonnaux, A. Eychene, The Raf/MEK/ERK pathway: new concepts of
activation, Biol. Cell 93 (1–2) (2001) 53–62.
[2] H. Davies, et al., Mutations of the BRAF gene in human cancer, Nature 417 (6892)
(2002) 949–954.
[3] P.B. Chapman, et al., Improved survival with vemurafenib in melanoma with
BRAF V600E mutation, N. Engl. J. Med. 364 (26) (2011) 2507–2516.
[4] R.J. Sullivan, K.T. Flaherty, Resistance to BRAF-targeted therapy in melanoma,
Eur. J. Cancer 49 (6) (2013) 1297–1304.
[5] A.M. Menzies, G.V. Long, Systemic treatment for BRAF-mutant melanoma:
where do we go next? Lancet Oncol. 15 (9) (2014) e371–e381.
[6] E.L. Niero, et al., The multiple facets of drug resistance: one history, different
approaches, J. Exp. Clin. Cancer Res. 33 (2014) 37.
[7] M. Panczyk, Pharmacogenetics research on chemotherapy resistance in
colorectal cancer over the last 20 years, World J. Gastroenterol. 20 (29) (2014)
9775–9827.
[8] D.R. Bell, et al., Detection of P-glycoprotein in ovarian cancer: a molecular marker
associated with multidrug resistance, J. Clin. Oncol. 3 (3) (1985) 311–315.
[9] A. Sorrentino, et al., Role of microRNAs in drug-resistant ovarian cancer cells,
Gynecol. Oncol. 111 (3) (2008) 478–486.
[10] H. Aguilar, et al., VAV3 mediates resistance to breast cancer endocrine therapy,
Breast Cancer Res. 16 (3) (2014) R53.
[11] S. Izraely, et al., The metastatic microenvironment: brain-residing melanoma
metastasis and dormant micrometastasis, Int. J. Cancer 131 (5) (2012) 1071–
1082.
[12] S. Izraely, O. Sagi-Assif, A. Klein, T. Meshel, S. Ben-Menachem, A. Zaritsky, et al.,
The metastatic microenvironment: Claudin-1 suppresses the malignant
phenotype of melanoma brain metastasis, Int. J. Cancer 15 (2015) 1296–1307.
[13] B.B. Weksler, et al., Blood-brain barrier-specific properties of a human adult
brain endothelial cell line, FASEB J. 19 (13) (2005) 1872–1874.
[14] R.E. Unger, et al., In vitro expression of the endothelial phenotype: comparative
study of primary isolated cells and cell lines, including the novel cell line
HPMEC-ST1.6R, Microvasc. Res. 64 (3) (2002) 384–397.
[15] H. Duan, et al., Targeting endothelial CD146 attenuates neuroinflammation by
limiting lymphocyte extravasation to the CNS, Sci. Rep. 3 (2013) 1687.
[16] H. Yamamoto, J.B. Sedgwick, W.W. Busse, Differential regulation of eosinophil
adhesion and transmigration by pulmonary microvascular endothelial cells, J.
Immunol. 161 (2) (1998) 971–977.
[17] R.A. Worthylake, et al., RhoA is required for monocyte tail retraction during
transendothelial migration, J. Cell Biol. 154 (1) (2001) 147–160.
[18] G.M. Seigel, L.M. Campbell, High-throughput microtiter assay for Hoechst 33342
dye uptake, Cytotechnology 45 (3) (2004) 155–160.
[19] R.A. Irizarry, et al., Exploration, normalization, and summaries of high density
oligonucleotide array probe level data, Biostatistics 4 (2) (2003) 249–264.
[20] Y. Nakajima, Y. Nakamura, W. Shigeeda, M. Tomoyasu, H. Deguchi, T. Tanita, et al.,
The role of tumor necrosis factor-alpha and interferon-gamma in regulating
angiomotin-like protein 1 expression in lung microvascular endothelial cells,
Allergol. Int. 62 (3) (2013) 309–322.
[21] N.E. Willmarth, A. Baillo, M.L. Dziubinski, K. Wilson, D.J. Riese 2nd, S.P. Ethier,
Altered EGFR localization and degradation in human breast cancer cells with
an amphiregulin/EGFR autocrine loop, Cell. Signal. 21 (2) (2009) 212–219.
[22] L. Zhao, J. Xia, X. Wang, F. Xu, Transcriptional regulation of CCL20 expression,
Microbes Infect. 16 (2014) 864–870.
[23] X. Liu, K.M. Chu, E-cadherin and gastric cancer: cause, consequence, and
applications, Biomed Res. Int. 2014 (2014) 637308.
[24] A. Awasthi, A.G. Woolley, F.J. Lecomte, N. Hung, B.C. Baguley, S.M. Wilbanks,
et al., Variable expression of GLIPR1 correlates with invasive potential in
melanoma cells, Front. Oncol. 3 (2013) 225.
[25] B. Liu, B. Pang, X. Hou, H. Fan, N. Liang, S. Zheng, et al., Expression of highmobility group AT-hook protein 2 and its prognostic significance in malignant
gliomas, Hum. Pathol. 45 (8) (2014) 1752–1758.
[26] D. Kong, G. Su, L. Zha, H. Zhang, J. Xiang, W. Xu, et al., Coexpression of HMGA2
and Oct4 predicts an unfavorable prognosis in human gastric cancer, Med. Oncol.
31 (8) (2014) 130.
[27] R.N. Apte, Y. Krelin, X. Song, S. Dotan, E. Recih, M. Elkabets, et al., Effects of
micro-environment- and malignant cell-derived interleukin-1 in carcinogenesis,
tumour invasiveness and tumour-host interactions, Eur. J. Cancer 42 (6) (2006)
751–759.
[28] E. Voronov, Y. Carmi, R.N. Apte, The role IL-1 in tumor-mediated angiogenesis,
Front. Physiol. 5 (2014) 114.
[29] O. Kholmanskikh, N. van Baren, F. Brasseur, S. Ottaviani, J. Vanacker, N. Arts,
et al., Interleukins 1alpha and 1beta secreted by some melanoma cell lines
strongly reduce expression of MITF-M and melanocyte differentiation antigens,
Int. J. Cancer 127 (7) (2010) 1625–1636.
[30] S.F. Chen, S. Nieh, S.W. Jao, M.Z. Wu, C.L. Liu, Y.C. Chang, et al., The paracrine
effect of cancer-associated fibroblast-induced interleukin-33 regulates the
invasiveness of head and neck squamous cell carcinoma, J. Pathol. 231 (2) (2013)
180–189.
[31] P. Ranganathan, K.L. Weaver, A.J. Capobianco, Notch signalling in solid tumours:
a little bit of everything but not all the time, Nat. Rev. Cancer 11 (5) (2011)
338–351.
[32] M. Held, M. Bosenberg, A role for the JARID1B stem cell marker for continuous
melanoma growth, Pigment Cell Melanoma Res. 23 (4) (2010) 481–483.
96
I. Zubrilov et al./Cancer Letters 361 (2015) 86–96
[33] G. Chen, X. Xu, L. Zhang, Y. Fu, M. Wang, H. Gu, et al., Blocking autocrine VEGF
signaling by sunitinib, an anti-cancer drug, promotes embryonic stem cell
self-renewal and somatic cell reprogramming, Cell Res. 24 (9) (2014) 1121–1136.
[34] J. Albrengues, I. Bourget, C. Pons, V. Butet, P. Hofman, S. Tartare-Deckert, et al.,
LIF mediates proinvasive activation of stromal fibroblasts in cancer, Cell Rep.
7 (2014) 1664–1678.
[35] M. Jin, T. Zhang, C. Liu, M.A. Badeaux, B. Liu, R. Liu, et al., miRNA-128 suppresses
prostate cancer by inhibiting BMI-1 to inhibit tumor-initiating cells, Cancer Res.
74 (15) (2014) 4183–4195.
[36] M.S. Song, J.J. Rossi, The anti-miR21 antagomir, a therapeutic tool for colorectal
cancer, has a potential synergistic effect by perturbing an angiogenesisassociated miR30, Front. Genet. 4 (2014) 301.
[37] M. Garofalo, C. Quintavalle, G. Romano, C.M. Croce, G. Condorelli, miR221/222
in cancer: their role in tumor progression and response to therapy, Curr. Mol.
Med. 12 (1) (2012) 27–33.
[38] Y. Chen, M.S. Zaman, G. Deng, S. Majid, S. Saini, J. Liu, et al., MicroRNAs 221/222
and genistein-mediated regulation of ARHI tumor suppressor gene in prostate
cancer, Cancer Prev. Res. 4 (1) (2011) 76–86.
[39] Z. Cierna, M. Mego, P. Janega, M. Karaba, G. Minarik, J. Benca, et al., Matrix
metalloproteinase 1 and circulating tumor cells in early breast cancer, BMC
Cancer 14 (2014) 472.
[40] J.S. Blackburn, I. Liu, C.I. Coon, C.E. Brinckerhoff, A matrix metalloproteinase-1/
protease activated receptor-1 signaling axis promotes melanoma invasion and
metastasis, Oncogene 28 (48) (2009) 4237–4248.
[41] N. Sun, Q. Zhang, C. Xu, Q. Zhao, Y. Ma, X. Lu, et al., Molecular regulation of
ovarian cancer cell invasion, Tumour Biol. 35 (2014) 11359–11366.
[42] O. Itkonen, U.H. Stenman, TATI as a biomarker, Clin. Chim. Acta 431 (2014)
260–269.
[43] C. Wang, L. Wang, B. Su, N. Lu, J. Song, X. Yang, et al., Serine protease inhibitor
Kazal type 1 promotes epithelial-mesenchymal transition through EGFR
signaling pathway in prostate cancer, Prostate 74 (7) (2014) 689–701.
[44] S. Izraely, et al., Chemokine-chemokine receptor axes in melanoma brain
metastasis, Immunol. Lett. 130 (1–2) (2010) 107–114.
[45] A. Klein, et al., The metastatic microenvironment: brain-derived soluble factors
alter the malignant phenotype of cutaneous and brain-metastasizing melanoma
cells, Int. J. Cancer 131 (11) (2012) 2509–2518.
[46] I.P. Witz, The tumor microenvironment: the making of a paradigm, Cancer
Microenviron. 2 (Suppl. 1) (2009) 9–17.
[47] A. Zoccoli, et al., Premetastatic niche: ready for new therapeutic interventions?
Expert Opin. Ther. Targets 16 (Suppl. 2) (2012) S119–S129.
[48] C.E. Eyler, J.N. Rich, Survival of the fittest: cancer stem cells in therapeutic
resistance and angiogenesis, J. Clin. Oncol. 26 (17) (2008) 2839–2845.
[49] M. Kim, H. Turnquist, J. Jackson, M. Sgagias, Y. Yan, M. Gong, et al., The multidrug
resistance transporter ABCG2 (breast cancer resistance protein 1) effluxes
Hoechst 33342 and is overexpressed in hematopoietic stem cells, Clin. Cancer
Res. 8 (1) (2002) 22–28.
[50] V. Tirino, et al., Cancer stem cells in solid tumors: an overview and new
approaches for their isolation and characterization, FASEB J. 27 (1) (2013) 13–24.
[51] M. Cojoc, K. Mäbert, M.H. Muders, A. Dubrovska, A role for cancer stem cells
in therapy resistance: cellular and molecular mechanisms, Semin. Cancer Biol.
31C (2015) 16–27.
[52] H. Easwaran, H.C. Tsai, S.B. Baylin, Cancer epigenetics: tumor heterogeneity,
plasticity of stem-like states, and drug resistance, Mol. Cell 54 (5) (2014)
716–727.
[53] M.M. Gottesman, V. Ling, The molecular basis of multidrug resistance in
cancer: the early years of P-glycoprotein research, FEBS Lett. 580 (4) (2006)
998–1009.
[54] K. Noguchi, K. Katayama, Y. Sugimoto, Human ABC transporter ABCG2/BCRP
expression in chemoresistance: basic and clinical perspectives for molecular
cancer therapeutics, Pharmgenomics. Pers. Med. 7 (2014) 53–64.
[55] V. Paulitschke, W. Berger, P. Paulitschke, E. Hofstaetter, B. Knapp, R.
Dingelmaier-Hovorka, et al., Vemurafenib resistance signature by proteome
analysis offers new strategies and rational therapeutic concepts, Mol. Cancer
Ther. (2015) doi:10.1158/1535-7163.MCT-14-0701.
[56] S. Kapse-Mistry, et al., Nanodrug delivery in reversing multidrug resistance in
cancer cells, Front. Pharmacol. 5 (2014) 159.
[57] J.L. Markman, et al., Nanomedicine therapeutic approaches to overcome cancer
drug resistance, Adv. Drug Deliv. Rev. 65 (13–14) (2013) 1866–1879.
[58] Y.D. Livney, Y.G. Assaraf, Rationally designed nanovehicles to overcome cancer
chemoresistance, Adv. Drug Deliv. Rev. 65 (13–14) (2013) 1716–1730.
[59] A.K. Iyer, et al., Role of integrated cancer nanomedicine in overcoming drug
resistance, Adv. Drug Deliv. Rev. 65 (13–14) (2013) 1784–1802.
[60] A. Klein-Goldberg, S. Maman, I.P. Witz, The role played by the microenvironment
in site-specific metastasis, Cancer Lett. 352 (1) (2014) 54–58.
[61] I.P. Witz, Tumor-microenvironment interactions: dangerous liaisons, Adv. Cancer
Res. 100 (2008) 203–229.
[62] B. Homet, A. Ribas, New drug targets in metastatic melanoma, J. Pathol. 232
(2) (2014) 134–141.
[63] C.I. Lee, et al., Comparative effectiveness of imaging modalities to determine
metastatic breast cancer treatment response, Breast 24 (1) (2015) 3–11.
[64] P.F. Peddi, S.A. Hurvitz, Ado-trastuzumab emtansine (T-DM1) in human
epidermal growth factor receptor 2 (HER2)-positive metastatic breast cancer:
latest evidence and clinical potential, Ther. Adv. Med. Oncol. 6 (5) (2014)
202–209.
[65] A. Irmisch, J. Huelsken, Metastasis: new insights into organ-specific
extravasation and metastatic niches, Exp. Cell Res. 319 (11) (2013) 1604–1610.
[66] M.J. Carlini, M.S. De Lorenzo, L. Puricelli, Cross-talk between tumor cells and
the microenvironment at the metastatic niche, Curr. Pharm. Biotechnol. 12 (11)
(2011) 1900–1908.
[67] A. Ben-Baruch, Organ selectivity in metastasis: regulation by chemokines and
their receptors, Clin. Exp. Metastasis 25 (4) (2008) 345–356.
[68] S. Gout, P.L. Tremblay, J. Huot, Selectins and selectin ligands in extravasation
of cancer cells and organ selectivity of metastasis, Clin. Exp. Metastasis 25 (4)
(2008) 335–344.
[69] E. Fokas, et al., Metastasis: the seed and soil theory gains identity, Cancer
Metastasis Rev. 26 (3–4) (2007) 705–715.