cancers

Article
Factors Secreted by Cancer-Associated Fibroblasts
that Sustain Cancer Stem Properties in Head and
Neck Squamous Carcinoma Cells as Potential
Therapeutic Targets
Saúl Álvarez-Teijeiro 1,2,†, * , Cristina García-Inclán 1,† , M. Ángeles Villaronga 1,2 ,
Pedro Casado 3 , Francisco Hermida-Prado 1 , Rocío Granda-Díaz 1 , Juan P. Rodrigo 1,2 ,
Fernando Calvo 4 , Nagore del-Río-Ibisate 1 , Alberto Gandarillas 5 , Francisco Morís 6 ,
Mario Hermsen 1,2 , Pedro Cutillas 3 and Juana M. García-Pedrero 1,2, *
1 Department of Otolaryngology, Hospital Universitario Central de Asturias and Instituto de Investigación
Sanitaria del Principado de Asturias; Instituto Universitario de Oncología del Principado de Asturias,
University of Oviedo, 33011 Oviedo, Spain; cristinagarciainclan@gmail.com (C.G.-I.);
angelesvillaronga@gmail.com (M.Á.V.); franjhermida@gmail.com (F.H.-P.); rocigd281@gmail.com (R.G.-D.);
juanpablo.rodrigo@sespa.es (J.P.R.); nagoredelrio@gmail.com (N.d.-R.-I.); mariohermsen@gmail.com (M.H.)
2 CIBERONC, 28029 Madrid, Spain
3 Cell Signalling & Proteomics Group, Barts Cancer Institute, Queen Mary University of London,
London EC1M 6BQ, UK; p.m.casado-izquierdo@qmul.ac.uk (P.Ca.); p.cutillas@qmul.ac.uk (P.Cu.)
4 Tumour Microenvironment Team, Division of Cancer Biology, Institute of Cancer Research,
237 Fulham Road, London SW3 6JB, UK; fernando.calvo@icr.ac.uk
5 Cell Cycle, Stem Cell Fate and Cancer Lab Instituto de Investigación Marqués de Valdecilla (IDIVAL),
39011 Santander, Spain; agandarillas@idival.org
6 EntreChem SL, Vivero Ciencias de la Salud, 33011 Oviedo, Spain; fmv@entrechem.com
* Correspondence: s.alvarezteijeiro@qmul.ac.uk (S.Á.-T.); juanagp.finba@gmail.com (J.M.G.-P.)
† These authors contributed equally to this work.




Received: 22 August 2018; Accepted: 12 September 2018; Published: 17 September 2018


Abstract: This study investigates for the first time the crosstalk between stromal fibroblasts and
cancer stem cell (CSC) biology in head and neck squamous cell carcinomas (HNSCC), with the
ultimate goal of identifying effective therapeutic targets. The effects of conditioned media from
cancer-associated fibroblasts (CAFs) and normal fibroblasts (NFs) on the CSC phenotype were
assessed by combining functional and expression analyses in HNSCC-derived cell lines. Further
characterization of CAFs and NFs secretomes by mass spectrometry was followed by pharmacologic
target inhibition. We demonstrate that factors secreted by CAFs but not NFs, in the absence of
serum/supplements, robustly increased anchorage-independent growth, tumorsphere formation, and
CSC-marker expression. Modulators of epidermal growth factor receptor (EGFR), insulin-like growth
factor receptor (IGFR), and platelet-derived growth factor receptor (PDGFR) activity were identified
as paracrine cytokines/factors differentially secreted between CAFs and NFs, in a mass spectrometry
analysis. Furthermore, pharmacologic inhibition of EGFR, IGFR, and PDGFR significantly reduced
CAF-induced tumorsphere formation and anchorage-independent growth suggesting a role of these
receptor tyrosine kinases in sustaining the CSC phenotype. These findings provide novel insights
into tumor stroma–CSC communication, and potential therapeutic targets to effectively block the
CAF-enhanced CSC niche signaling circuit.

Keywords: head and neck squamous cell carcinoma; cancer-associated fibroblasts; cancer stem cells;
tumor microenvironment; secretome; therapeutic target

Cancers 2018, 10, 334; doi:10.3390/cancers10090334 www.mdpi.com/journal/cancers

Cancers 2018, 10, 334 2 of 17

1. Introduction
Mounting evidence indicates that tumors are highly complex heterogeneous structures in
which growth is supported not only by the cancer cells themselves, but also the surrounding
microenvironment. Since cancer must be considered as a systemic disease, an in-depth understanding
of tumor progression requires knowledge of the role of both tumor cells and infiltrating stroma, as well
as how these distinct cell types interact to drive tumor biology. The tumor stroma is constituted by
various types of stromal cells and the extracellular matrix (ECM), collectively denominated as tumor
microenvironment (TME) [1].
Fibroblasts are a major component within the TME, and in particular, cancer-associated fibroblasts
(CAFs) play a key role in tumorigenesis, as they significantly contribute to important hallmarks
necessary for cancer progression, such as sustained growth, invasion, inflammation, angiogenesis,
metastasis, and therapeutic resistance [2,3]. It is well established that tumor cells can stimulate stromal
cells to release paracrine factors that facilitate cancer growth and dissemination. Specifically, CAFs
emerge as critical players in this process, stimulating cancer progression toward aggressive phenotypes
through cell–cell communication with cancer cells or other stromal cells, remodeling the ECM and
releasing a plethora of growth factors, chemokines, cytokines, and matrix metalloproteinases (MMPs)
in the TME [4]. Accordingly, the presence of CAFs has been widely associated with poor prognosis in
numerous tumor types, including among others, gastric, colorectal, breast, and prostate cancers [5,6].
Tumors are hierarchically organized with different cancer cell subpopulations, where cancer
stem cells (CSCs) are essential for tumor initiation, treatment resistance, relapse, and metastasis [7,8].
CSCs are regulated by, and in turn regulate, cells within the TME. Recent findings have shown
the plasticity and phenotype switching of the different cancer cell subpopulations [9–11]. Thus,
malignant epithelial cells may dedifferentiate, and thereby enter back into the stem cell pool. Therefore,
therapies aimed at targeting CSCs within the tumor will not be curative if the CSCs pool can be
continuously regenerated from plastic non-CSCs capable of dedifferentiating and reentering the CSC
state. There are strong indications that CAFs may regulate CSCs in various ways: (i) acting directly
on the CSCs subpopulation to promote their self-renewal; (ii) re-inducing a stem cell phenotype in
more differentiated tumor cells (reprogramming); or (iii) activating autocrine signaling loops in tumor
cells that maintain them into a stem cell-like state [12]. Consequently, the identification of molecules
responsible for the conversion of non-CSCs into CSCs is indispensable to select the most appropriate
drugs, or combinations of them, to efficiently eliminate CSCs populations, and subsequently reduce
the risk of metastasis outgrowth and tumor relapse.
Previous functional studies have demonstrated that secreted proteins, acting as paracrine factors,
provide an important bidirectional communication system between cancer cells and the surrounding
fibroblasts. Hence, the secretome from both tumor and stromal cells may constitute a rich reservoir of
potential biomarkers and/or new therapeutic targets [13,14]. Likewise, as secreted proteins, they have
the potential of being released into blood circulation or saliva, thus increasing the possibility of their
detection in patient-derived body fluids.
Therefore, a deeper understanding of the molecules involved in the interaction between CSCs
and CAFs is fundamental to find novel targets to block effectively the communication between
them, and ultimately prevent their cooperative roles in promoting tumor progression. This study
investigated for the first time the crosstalk between stromal fibroblasts and CSC in the context of head
and neck squamous cell carcinomas (HNSCC). Using mass spectrometry (MS) we identified various
CAF-secreted molecules potentially responsible for sustaining CSC properties in HNSCC-derived cell
lines. In addition, pharmacological targeting of signaling pathways related to the identified factors
effectively blocked CAF-induced CSC phenotype, thus suggesting their potential as novel therapeutic
targets to overcome CSC-mediated disease progression and resistance to therapy.
Cancers 2018, 10, 334 3 of 17

2. Results

2.1. Fibroblast-Secreted Factors

Cancers 2018, 10, x FOR Sustain Cancer Stem Properties of HNSCC Cells
PEER REVIEW 3 of 17

Cancer stem cells (CSCs) play critical roles in tumor initiation, progression, recurrence, and
2. Results
treatment resistance. However, the crosstalk between CAFs and CSCs in the context of HNSCC has not
2.1. Fibroblast-Secreted Factors Sustain Cancer Stem Properties of HNSCC Cells
yet been explored. This prompted us to investigate the effect of CM from CAFs or NFs on cancer stem
Cancer stem cells (CSCs) play critical roles in tumor initiation, progression, recurrence, and
properties in HNSCC cells, such as tumorosphere formation, anchorage-independent growth, and
treatment resistance. However, the crosstalk between CAFs and CSCs in the context of HNSCC has
expressionnotofyet
CSCbeenmarkers. We first
explored. This assayed
prompted the abilitytheofeffect
us to investigate fibroblast-CM
of CM from CAFsto promote
or NFs onthe formation of
cancer
clonal spheres (tumorspheres)
stem properties in HNSCC in cells,
non-adherent and serum-free
such as tumorosphere formation,culture conditions (Figure
anchorage-independent 1). We found
growth,
and expression
that both FaDu and SCC38 of CSC
cellsmarkers.
were able We to
first assayed
form the ability
orospheres whenof fibroblast-CM
grown in NF-CM,to promoteandthemuch more
formation of clonal spheres (tumorspheres) in non-adherent and serum-free culture conditions
efficiently in CAF-CM, showing bigger size spheres similar to those formed in the presence of stem
(Figure 1). We found that both FaDu and SCC38 cells were able to form orospheres when grown in
supplemented
NF-CM, medium,
and muchusedmoreas a positive
efficiently controlshowing
in CAF-CM, by its high
biggerefficiency
size spherestosimilar
induce orosphere
to those formed formation.
As expected,
in theHNSCC
presence cells
of stemwere unable to
supplemented form used
medium, orospheres in non-supplemented
as a positive medium,
control by its high efficiency to which
served as induce orosphere
a negative formation. As expected, HNSCC cells were unable to form orospheres in non-
control.
supplemented medium, which served as a negative control.

Figure 1.
Figure 1. Effect ofEffect of fibroblast-CM
fibroblast-CM (conditionedmedia)
(conditioned media) on
onthethe
tumorsphere formation
tumorsphere capacity capacity
formation of head of head
and neck squamous cell carcinomas (HNSCC) cells. Representative images of orospheres formed by
and neck squamous cell carcinomas (HNSCC) cells. Representative images of orospheres formed
by (A) FaDu and (B) SCC38 cells in non-supplemented medium, supplemented medium, and CM
from normal fibroblasts (NFs) or cancer-associated fibroblasts (CAFs). Bar chart showing the average
diameter of spheroids formed by (C) FaDu and (D) SCC38 cells in the previous conditions. All data
were expressed as the mean ± SD of at least three independent experiments performed. Scale bar:
500 µm. *** p < 0.001 and ** p < 0.01 by Holm-Sidak’s multiple comparisons test.
 (A) FaDu and (B) SCC38 cells in non-supplemented medium, supplemented medium, and CM from
normal fibroblasts (NFs) or cancer-associated fibroblasts (CAFs). Bar chart showing the average
diameter of spheroids formed by (C) FaDu and (D) SCC38 cells in the previous conditions. All data were
expressed
Cancers 2018, 10, as
334the mean ± SD of at least three independent experiments performed. Scale bar: 500 μm. *** 4 of 17
p < 0.001 and ** p < 0.01 by Holm-Sidak’s multiple comparisons test.

Consistently,we
Consistently, wealso
alsofound
foundthat
thatCAF-CM,
CAF-CM,and andto
toaaless
lessextent
extentNF-CM,
NF-CM,significantly
significantlyincreased
increased
anchorage-independent growth in both FaDu and SCC38 cells, compared to
anchorage-independent growth in both FaDu and SCC38 cells, compared to non-supplementednon-supplemented
medium(Figure
medium (Figure 2).
2). These
These results
results indicate that components
components ofofthe
theCAF-CM
CAF-CMpromote
promoteCSCCSCproperties
propertiesin
inHNSCC
HNSCCcells.
cells.

Effectofoffibroblast-CM
Figure2.2.Effect
Figure fibroblast-CMon onanchorage-independent
anchorage-independentgrowth
growthofofHNSCC
HNSCCcells.cells.(A)
(A)FaDu
FaDuand
and
(B) SCC38 cells were seeded in plates coated with Poly(2-hydroxyethyl methacrylate)
(B) SCC38 cells were seeded in plates coated with Poly(2-hydroxyethyl methacrylate) (polyHEMA) (polyHEMA)
andgrown
and grownininnon-supplemented
non-supplementedmedium,medium,CM CMfrom
fromNFs,
NFs,CM
CMfrom
fromCAFs,
CAFs,ororsupplemented
supplementedmedium.
medium.
Cellproliferation
Cell proliferationwas
wasestimated
estimatedbybytetrazolium-based
tetrazolium-basedMTSMTSassay
assayafter
after44days.
days.Data
Datawere
werenormalized
normalized
totothe
theabsorbance
absorbanceat at day
day 00 and
and relative
relative to
to control
control (non-supplemented)
(non-supplemented)cells.
cells.All
Alldata
datawere
wereexpressed
expressedas
the mean ±
as the mean ± SD of at least three independent experiments performed in quadruplicate. *** pp<<0.001,
SD of at least three independent experiments performed in quadruplicate. *** 0.001,
** p < 0.01 and * p < 0.05 by Holm-Sidak’s multiple comparisons
** p < 0.01 and * p < 0.05 by Holm-Sidak’s multiple comparisons test. test.

2.2. Fibroblast-Secreted Factors Induced the Expression of Stemness-Related Genes in HNSCC Cells
2.2. Fibroblast-Secreted Factors Induced the Expression of Stemness-Related Genes in HNSCC Cells
The results above suggest that CAF-CM may be inducing stem properties in HNSCC cells.
The results above suggest that CAF-CM may be inducing stem properties in HNSCC cells. To
To verify this observation and to gain further insight into the mechanisms by which stromal fibroblasts
verify this observation and to gain further insight into the mechanisms by which stromal fibroblasts
may promote CSC features in HNSCC cells, we analyzed the expression of several CSC markers and
may promote CSC features in HNSCC cells, we analyzed the expression of several CSC markers and other
other genes with well-known functions in pluripotency, self-renewal, and signal transduction in CSCs.
genes with well-known functions in pluripotency, self-renewal, and signal transduction in CSCs.
Thus, we performed RT-qPCR in orospheres formed by FaDu and SCC38 cells treated with
Thus, we performed RT-qPCR in orospheres formed by FaDu and SCC38 cells treated with CAF-
CAF-CM or supplemented medium and gene expression was compared to that of adherent
CM or supplemented medium and gene expression was compared to that of adherent control cultures.
control cultures. Overall, we observed an increase in the mRNA levels of several CSC-related
Overall, we observed an increase in the mRNA levels of several CSC-related genes in orospheres
genes in orospheres formed in CAF-CM and supplemented medium, compared to adherent cells
formed in CAF-CM and supplemented medium, compared to adherent cells (Figure 3). Thus, CAF-
(Figure 3). Thus, CAF-CM potently and consistently increased the mRNA levels of ALDH1, NANOG,
CM potently and consistently increased the mRNA levels of ALDH1, NANOG, SOX2, and OCT4.
SOX2, and OCT4. There were some noticeable differences between supplemented medium and
There were some noticeable differences between supplemented medium and fibroblast-CM-formed
fibroblast-CM-formed orospheres, depending on the gene and also the HNSCC-derived cell line.
orospheres, depending on the gene and also the HNSCC-derived cell line. ALDH1 expression levels
ALDH1 expression levels were highly induced in orospheres formed in supplemented medium in
were highly induced in orospheres formed in supplemented medium in FaDu cells, whereas mRNA
FaDu cells, whereas mRNA levels of NANOG, ABCG2, CD44, CD133, and Nestin were more robustly
levels of NANOG, ABCG2, CD44, CD133, and Nestin were more robustly induced in CAF-CM-
induced in CAF-CM-formed orospheres in HNSCC cells. These results suggest that CAF-CM and
formed orospheres in HNSCC cells. These results suggest that CAF-CM and supplemented medium
supplemented medium regulate stemness by activating different signal transduction programs in
regulate stemness by activating different signal transduction programs in HNSCC cells to sustain the
HNSCC cells to sustain the CSC-phenotype.
CSC-phenotype.
Cancers 2018, 10, 334 5 of 17
Cancers 2018, 10, x FOR PEER REVIEW 5 of 17

Figure
Figure 3.3. Effect
Effect of
of fibroblast-CM
fibroblast-CM on on the
the expression
expression of of stem-related
stem-related genes
genes in
in HNSCC
HNSCC cells.
cells. Bar
Bar chart
chart
showing
showing the the expression
expression analysis
analysis of CSC-related
CSC-related genes by qRT-PCRqRT-PCR analysis
analysis in
in FaDu
FaDu (A)
(A) and
and SCC38
SCC38
(B)
(B) orospheres
orospheresformed
formedinin CAF-CM
CAF-CM andand
supplemented
supplemented medium. Adherent
medium. monolayer
Adherent cultures
monolayer of FaDu
cultures of
or SCC38 cells were used as control. Data were normalized to RPL19 levels and
FaDu or SCC38 cells were used as control. Data were normalized to RPL19 levels and relative to relative to control cells.
All datacells.
control were Allexpressed
data were as the mean ±asSDthe
expressed of mean
at least
± three
SD ofindependent
at least threeexperiments
independentperformed
experiments in
triplicate.
performed* in p <triplicate.
0.05, ** p *<p0.01 p <p 0.01
and*****
< 0.05, < 0.001
andby p < 0.001 t-test.
***Student’s by Student’s t-test.

2.3. Identification of Fibroblast-Secreted Proteins by Mass Spectrometry

2.3. Identification of Fibroblast-Secreted Proteins by Mass Spectrometry
We next performed a Mass Spec analysis in NFs and CAFs to identify the repertoire of
We next performed a Mass Spec analysis in NFs and CAFs to identify the repertoire of CAF-
CAF-secreted proteins that may be responsible for sustaining cancer stem properties in HNSCC
secreted proteins that may be responsible for sustaining cancer stem properties in HNSCC cells.
cells. Comparison of protein expression levels in the secretomes was conducted using three different
Comparison of protein expression levels in the secretomes was conducted using three different
biological replicates for each cell line.
biological replicates for each cell line.
Secretome analysis provided valuable information on the proteins differentially secreted by
Secretome analysis provided valuable information on the proteins differentially secreted by
CAFs compared to NFs (Figure 4). We identified 41 extracellular proteins differentially secreted
CAFs compared to NFs (Figure 4). We identified 41 extracellular proteins differentially secreted by
by CAFs versus NFs (fold change > 2 or <−2) (Table 1). Among the most highly up-regulated
CAFs versus NFs (fold change > 2 or <−2) (Table 1). Among the most highly up-regulated proteins,
proteins, we selected for further study Carboxypeptidase E (CBPE), which has been implicated in
we selected for further study Carboxypeptidase E (CBPE), which has been implicated in cell
cell proliferation and survival in other cancer types [15,16], as well as platelet-derived growth factor
proliferation and survival in other cancer types [15,16], as well as platelet-derived growth factor D
D (PDGFD), epidermal growth factor (EGF)-containing fibulin-like extracellular matrix protein-1
(PDGFD), epidermal growth factor (EGF)-containing fibulin-like extracellular matrix protein-1
(FBLN3), insulin-like growth factor binding protein-5 (IBP5) and insulin-like growth factor binding
(FBLN3), insulin-like growth factor binding protein-5 (IBP5) and insulin-like growth factor binding
protein-7 (IBP7), respectively associated to growth factor-signaling pathways through PDGF, EGF, and
protein-7 (IBP7), respectively associated to growth factor-signaling pathways through PDGF, EGF,
insulin-like growth (IGF) receptors [17–19].
and insulin-like growth (IGF) receptors [17–19].
Cancers 2018, 10, 334 6 of 17
Cancers 2018, 10, x FOR PEER REVIEW 6 of 17

Figure
Figure Mass
4. 4. Mass Spec
Specanalysis
analysisofofextracellular
extracellular proteins differentially
differentiallysecreted
secretedby byCAFs
CAFsversus
versus NFs.
NFs. (A)(A)
Volcano
Volcanoplot showing
plot showing the
theglobal
globalsecretome
secretomechanges,
changes, illustrating foldchange
illustrating fold change(log(logbase
base2)2)
andand p-value
p-value
(−log
(−logbase
base10),
10),between
betweenCAFs
CAFsand and NFs.
NFs. Horizontal
Horizontal bars barsrepresent
representthe thesignificance
significancep p= =
0.05, p =p 0.01
0.05, = 0.01
andand
p =p 0.001
= 0.001 (proteinsunder
(proteins underhorizontal
horizontal barbar of
of pp == 0.05 did
did notnot reach
reach significance).
significance).Vertical
Verticalbars
bars
represent
represent thethe proteins
proteins witha afold
with foldchange
changehigher
higherthanthan 22 or
or −
−2;2;(B)
(B)Heatmap
Heatmaprepresents
representsthe
thechanges
changes in in
thethe growth
growth factors
factors related-proteins
related-proteins foundfound in secretome.
in the the secretome.ThreeThree independent
independent experiments
experiments are
are shown;
redshown;
indicatesredfold
indicates fold>0
changes changes
and blue>0 and blue indicates
indicates fold changes
fold changes <0. <0.

Proteins differentially
Table1.1.Proteins
Table differentially secreted
secretedininCAFs
CAFsversus
versusNFs.
NFs
FoldFold
Change
Change
UNIPROT_ID
UNIPROT_ID Gene
Gene Name
Name p-Value
p-Value
CAFsCAFs
HNRPL_HUMAN heterogeneous
HNRPL_HUMAN heterogeneous nuclear
nuclear ribonucleoprotein
ribonucleoproteinL(HNRNPL)
L(HNRNPL) 6.92 6.92 0.0200.020
CBPE_HUMAN
CBPE_HUMAN carboxypeptidase E(CPE)
carboxypeptidase E(CPE) 6.76 6.76 0.0200.020
CO7_HUMAN
CO7_HUMAN complementC7(C7)
complement C7(C7) 6.08 6.08 0.0460.046
PDGFD_HUMAN
PDGFD_HUMAN platelet
platelet derived growthfactor
derived growth factorD(PDGFD)
D(PDGFD) 5.87 5.87 >0.001>0.001
EGF containing fibulin fibulin
EGF containing like extracellular matrix protein
like extracellular
FBLN3_HUMAN
FBLN3_HUMAN 4.50 4.50 0.0310.031
1(EFEMP1)
matrix protein 1(EFEMP1)
IBP5_HUMAN
IBP5_HUMAN insulin
insulin like
like growth factor binding
growth factor bindingprotein
protein5(IGFBP5)
5(IGFBP5) 3.89 3.89 0.0150.015
DDAH1_HUMAN dimethylarginine dimethylaminohydrolase
DDAH1_HUMAN dimethylarginine dimethylaminohydrolase 1(DDAH1) 1(DDAH1) 3.38 3.38 0.0100.010
PGM1_HUMAN
PGM1_HUMAN phosphoglucomutase
phosphoglucomutase 1(PGM1) 1(PGM1) 3.24 3.24 0.0010.001
GREM1_HUMAN
GREM1_HUMAN gremlin
gremlin 1,
1, DAN family BMP
DAN family BMPantagonist(GREM1)
antagonist(GREM1) 3.09 3.09 0.0100.010
IF4A1_HUMAN
IF4A1_HUMAN
eukaryotic translation initiation factor 4A1(EIF4A1)
eukaryotic translation initiation factor 4A1(EIF4A1) 2.65
2.65 0.047
0.047
RS18_HUMAN ribosomal protein S18(RPS18) 2.40 0.036
RS18_HUMAN ribosomal protein S18(RPS18) 2.40 0.036
TCPQ_HUMAN chaperonin containing TCP1 subunit 8(CCT8) 2.25 0.039
TCPQ_HUMAN chaperonin containing TCP1 subunit 8(CCT8) 2.25 0.039
IBP7_HUMAN insulin like growth factor binding protein 7(IGFBP7) 2.06 0.037
IBP7_HUMAN insulin like growth factor binding protein 7(IGFBP7) 2.06 0.037
TCPD_HUMAN chaperonin containing TCP1 subunit 4(CCT4) −2.07 >0.001
TCPD_HUMAN chaperonin containing
myristoylated alanineTCP1 subunit kinase
rich protein 4(CCT4) C −2.07 >0.001
MARCS_HUMAN myristoylated alanine rich protein −2.22 0.042
MARCS_HUMAN substrate(MARCKS) −2.22 0.042
ANXA2_HUMAN kinase C substrate(MARCKS)
annexin A2(ANXA2) −2.24 0.018
ANXA2_HUMAN
PEDF_HUMAN annexin
serpin family A2(ANXA2)
F member 1(SERPINF1) −2.24
−2.27 0.0180.003
PEDF_HUMAN
MMP3_HUMAN serpin
matrix family F member 1(SERPINF1)
metallopeptidase 3(MMP3) −2.27
−2.29 0.0030.002
MMP3_HUMAN
PSA5_HUMAN matrix metallopeptidase
proteasome subunit alpha3(MMP3)
5(PSMA5) −2.29
−2.31 0.0020.037
PSA5_HUMAN
CALR_HUMAN proteasome subunit alpha 5(PSMA5)
calreticulin(CALR) −2.32
−2.31 0.0370.011
CALR_HUMAN brain abundant calreticulin(CALR)
membrane attached signal protein −2.32 0.011
BASP1_HUMAN −2.39 0.020
1(BASP1)
brain abundant membrane
BASP1_HUMAN −2.39 0.020
VASN_HUMAN attached vasorin(VASN)
signal protein 1(BASP1) −2.45 0.034
LUM_HUMAN
VASN_HUMAN lumican(LUM)
vasorin(VASN) −2.56
−2.45 0.0340.001
CFAD_HUMAN
LUM_HUMAN complement factor D(CFD)
lumican(LUM) −2.62
−2.56 0.0010.002
Cancers 2018, 10, 334 7 of 17

Table 1. Cont.

Fold Change
UNIPROT_ID Gene Name p-Value
CAFs
LA_HUMAN Sjogren syndrome antigen B(SSB) −2.66 0.007
UB2V1_HUMAN TMEM189-UBE2V1 readthrough(TMEM189-UBE2V1) −2.76 0.048
pregnancy specific beta-1-glycoprotein 7
PSG7_HUMAN −3.28 0.003
(gene/pseudogene)(PSG7)
PTGDS_HUMAN prostaglandin D2 synthase(PTGDS) −3.38 0.019
FBLN2_HUMAN fibulin 2(FBLN2) −3.40 0.012
acidic nuclear phosphoprotein 32 family member
AN32B_HUMAN −3.40 0.044
B(ANP32B)
ectonucleotide pyrophosphatase/phosphodiesterase
ENPP2_HUMAN −3.42 0.003
2(ENPP2)
MASP1_HUMAN mannan binding lectin serine peptidase 1(MASP1) −3.58 0.007
EMIL2_HUMAN elastin microfibril interfacer 2(EMILIN2) −3.68 >0.001
CSPG4_HUMAN chondroitin sulfate proteoglycan 4(CSPG4) −4.29 0.010
APOE_HUMAN apolipoprotein E(APOE) −4.45 0.003
TENA_HUMAN tenascin C(TNC) −4.68 0.001
PDIA6_HUMAN protein disulfide isomerase family A member 6(PDIA6) −4.93 0.028
A2GL_HUMAN leucine rich alpha-2-glycoprotein 1(LRG1) −5.24 0.006
RAB2A_HUMAN RAB2A, member RAS oncogene family(RAB2A) −6.09 0.002
RL15_HUMAN ribosomal protein L15(RPL15) −6.99 0.037
PSG4_HUMAN pregnancy specific beta-1-glycoprotein 4(PSG4) −8.01 0.015
Up-regulated proteins are shown in red and down-regulated proteins in blue.

2.4. Targeting EGFR, IGFR, and PDGFR Signaling Effectively Inhibited CAF-Promoted Stemness in
HNSCC Cells
We hypothesized that the CAF-secreted proteins CBPE, PDFGD, FBLN3, IBP5, and IBP7 may
enhance CSC properties in HNSCC cells through the activation of signaling pathways involving
the activity of CBPE or receptors for EGF, IGF, and PDGF. In order to support this hypothesis,
we assessed the ability of CAF-CM to induce anchorage-independent growth in the presence of
specific pharmacologic inhibitors, as these are conveniently available as research tools. We used
drugs whose modes of action involve both receptor blockage (Gefitinib for EGFR, OSI-906 for IGFR,
and CP-673451 for PDFGR) and protein blockage (GEMSA for CBPE) [16–19]. In addition, we also
include the mithramycin analog EC-8042, since it has been recently described as a potent inhibitor of
stemness-related genes and CSCs viability in other cancers [20].
Results showed that GEMSA, Gefitinib, OSI-906, and CP-673451 reduced anchorage-independent
growth in a dose-dependent manner in both FaDu and SCC38 cells grown in CAF-CM (Figure 5).
However, cells grown in supplemented-medium were clearly less sensitive to these drugs, with only
high doses having cytotoxic effects. EC-8042 was an effective blocker of anchorage-independent
growth in both FaDu and SCC38 cells grown in either CAF-CM or supplemented-medium.
We next evaluated the effect of these compounds on orosphere formation. Our previous results
(Figure 5) provided us with information on the most appropriate concentration of compounds to be
used for these experiments. Thus, we found that Gefitinib (1 µM), OSI-906 (10 µM), CP-673451 (5
µM), and EC-8042 (0.01 µM) effectively prevented tumorsphere-forming capability of FaDu cells in
either CAF-CM or supplemented medium (Figure 6). In contrast, 10 µM GEMSA did not significantly
inhibit orosphere formation in CAF-CM, whilst increased orosphere formation in supplemented
medium. Together, these results suggest that modulators of EGFR, PDGF and IGFR activity and also
the mithramycin analog EC-8042 have the potential to inhibit stemness-related properties in HNSCC
cells, consequently emerging as potential therapeutic targets to effectively block the CAF-enhanced
CSC niche signaling circuit.
Cancers 2018, 10, 334 8 of 17
Cancers 2018, 10, x FOR PEER REVIEW 8 of 17

Figure 5.5. Effect

Effect of
of 2-guanidinoethylmercaptosuccinic
2-guanidinoethylmercaptosuccinic acid (GEMSA), Gefitinib, OSI-906, CP-673451,
CAF-CM-mediated anchorage-independent
and EC-8042 on CAF-CM-mediated anchorage-independent growth. (A) (A) FaDu and (B) SCC38 cells
were seeded
were seededin inpolyHEMA-coated
polyHEMA-coatedplates platesand
andgrown
grownininCAF-CM
CAF-CMoror supplemented
supplemented medium.
medium. After
After 2424h,
h, cells
cells were
were treated
treated with
with increasing
increasing concentrations
concentrations of the
of the indicated
indicated drugsdrugs (GEMSA,
(GEMSA, Gefitinib,
Gefitinib, OSI-
OSI-906,
CP-673451,
906, CP-673451,and and
EC-8042). CellCell
EC-8042). proliferation
proliferation waswasestimated
estimatedbybytetrazolium-based
tetrazolium-basedMTS MTSassay
assay after
4 days. Data were normalized to the absorbance at day day 00 and
and relative
relative to
to control
control (vehicle-treated)
(vehicle-treated) cells.
cells.
All data were expressed as the mean
data were expressed as the mean ± ± SD of at least three independent experiments
least three independent experiments performed in
quadruplicate. * p < 0.05, ** p << 0.01
quadruplicate. 0.01 and *** pp << 0.001
and *** 0.001 by
by Student’s t-test.
Student’s t-test.
Cancers 2018, 10, 334 9 of 17
Cancers 2018, 10, x FOR PEER REVIEW 9 of 17

GEMSA, Gefitinib,
Figure 6. Effect of GEMSA, Gefitinib, OSI-906, CP-673451, and EC-8042
EC-8042 onon FaDu
FaDu orosphere
orosphere formation.
formation.
(A) Bar chart showing orospheres formation
(A) Bar chart showing orospheres formation ability of FaDu grown in supplemented
supplemented medium (left)
or CAF-CM (right) and treated with GEMSA (10 µM), μM), Gefitinib (1 µM),
μM), OSI-906 (10 µM),μM), CP-673451
(5 μM),and
(5 µM), andEC-8042
EC-8042 (0.01 µM);
(0.01 (B)(B)
μM); Representative
Representativeimages of FaDu
images orospheres
of FaDu for each
orospheres forcondition shown
each condition
in the bar
shown chart.
in the barSphere formation
chart. Sphere was estimated
formation by tetrazolium-based
was estimated MTS assay
by tetrazolium-based after 10–12
MTS assay days.
after 10–12
All data were expressed as the mean ± SD of at least three independent experiments
days. All data were expressed as the mean ± SD of at least three independent experiments performed performed in
quadruplicate. *** p < 0.001 by Student’s t-test. Scale bar: 100
in quadruplicate. *** p < 0.001 by Student’s t-test. Scale bar: 100 μm.µm.

3. Discussion
3. Discussion
Recent increasing evidence has suggested that the TME is an integral and inseparable part of
Recent increasing evidence has suggested that the TME is an integral and inseparable part of
malignant transformation [21,22], as it plays a significant role during tumor progression, enabling
malignant transformation [21,22], as it plays a significant role during tumor progression, enabling
primary growth, invasion, and metastatic spreading [23–25]. Hence, the study of the role of the
primary growth, invasion, and metastatic spreading [23–25]. Hence, the study of the role of the
different TME components and strategies aimed at interfering with the crosstalk between cancer
different TME components and strategies aimed at interfering with the crosstalk between cancer cells
cells and their cellular partners in the TME is of great interest, since it may provide novel promising
and their cellular partners in the TME is of great interest, since it may provide novel promising anti-
anti-cancer therapies with minimal chance to develop drug resistance [26].
cancer therapies with minimal chance to develop drug resistance [26].
This study provides original evidence demonstrating that CAF-secreted factors sustain and
This study provides original evidence demonstrating that CAF-secreted factors sustain and
robustly enhance stemness in HNSCC-derived cell lines, thereby increasing anchorage-independent
robustly enhance stemness in HNSCC-derived cell lines, thereby increasing anchorage-independent
growth, tumorsphere formation, and expression of various CSC markers, such as NANOG, SOX2,
growth, tumorsphere formation, and expression of various CSC markers, such as NANOG, SOX2,
OCT4, ALDH1, CD133, CD44, and NOTCH1, in the absence of serum or any other supplements.
OCT4, ALDH1, CD133, CD44, and NOTCH1, in the absence of serum or any other supplements.
There are strong indications that CAFs regulate CSCs [12]. In this regard, Donnarumma and colleagues
There are strong indications that CAFs regulate CSCs [12]. In this regard, Donnarumma and
observed that CAFs promoted cancer progression by enhancing stemness, Epithelial-mesenchymal
colleagues observed that CAFs promoted cancer progression by enhancing stemness, Epithelial-
transition (EMT) phenotype, and anchorage-independent growth in breast cancer [27]. Also, CAFs
mesenchymal transition (EMT) phenotype, and anchorage-independent growth in breast cancer [27].
were found to secrete ADAM10-rich exosomes to promote cell motility and activate RhoA and Notch
Also, CAFs were found to secrete ADAM10-rich exosomes to promote cell motility and activate RhoA
signaling in several cancer cell lines [28]. Vermeulen et al. showed that primary colon CAFs released
and Notch signaling in several cancer cell lines [28]. Vermeulen et al. showed that primary colon
HGF to induce nuclear translocation of ß-catenin in tumor cells and a stem cell-like transcription
CAFs released HGF to induce nuclear translocation of ß-catenin in tumor cells and a stem cell-like
profile [29]. In prostate cancer, tumor cells released IL-6 leading to fibroblasts activation, and in
transcription profile [29]. In prostate cancer, tumor cells released IL-6 leading to fibroblasts activation,
turn, fibroblasts, through MMPs secretion, elicited an EMT phenotype in cancer cells, as well as
and in turn, fibroblasts, through MMPs secretion, elicited an EMT phenotype in cancer cells, as well
as enhancement of tumor growth and development of spontaneous metastases. CAF-induced EMT
Cancers 2018, 10, 334 10 of 17

enhancement of tumor growth and development of spontaneous metastases. CAF-induced EMT in

prostate carcinoma cells was accompanied by increased expression of CSC markers, and enhanced
ability to form tumorspheres and self-renewal [30].
Nevertheless, prior to this study, the crosstalk between CAFs and CSC in the context of head and
neck squamous cell carcinomas (HNSCC) had not been explored. To identify the molecules responsible
for mediating the conversion of non-CSCs into CSCs is indispensable to select the most appropriate
drugs, or combinations of them, to efficiently eliminate CSCs populations, and subsequently reduce
the risk of metastasis outgrowth and tumor relapse.
These reasons prompted us to perform an unbiased proteomic analysis using MS to identify the
proteins secreted to extracellular media by stromal fibroblasts that mediate paracrine communication.
Our secretome analysis showed that there were several differentially secreted proteins (over- or
under-expressed) in CAFs compared with NFs. Among the most promising and highly induced factors
that could be responsible for sustaining the CSC phenotype are FBLN3, IBP5, IBP7, and PDGFD.
FBLN3 is an extracellular protein that can bind to EGFR, inducing EGFR autophosphorylation
and the activation of downstream signaling pathways. Several groups have observed that EGFR is
over-expressed in a wide spectrum of tumors, including HNSCC [31,32]. EGFR overexpression results
in aggressive tumor behavior, radiation resistance, and poor prognosis [33]. EGFR activates several
downstream pathways, including Ras/Raf/MAPK/ERK, PI3K/Akt, STAT, and the PLC-γ signaling
pathways to potentiate growth and survival of tumor cells and CSCs [19].
IBP5 and IBP7, as IGF binding proteins, bind IGFs to regulate their activity by prolonging
their half-life and circulation turnover, and by controlling their binding to IGFR, either positively or
negatively, affecting the IGF signaling pathway [34]. IBPs may inhibit mitogenesis, differentiation,
survival, and other IGF-stimulated events by sequestering IGFs away from the IGFR [35,36]. Also, they
can function independently of IGF signaling pathway via interacting with proteins other than IGFs
binding their own membrane receptors [37]. IBP5 has recently been deemed a molecular biomarker
for predicting response to therapy and clinical outcome in patients with different cancers [17]. Also,
it has been implicated as a cancer promoter or cancer repressor in various tumor types to regulate
migration, differentiation, cell attachment, and cell morphology [38,39]. Although there is less evidence
for the role of IBP7 in cancer, some groups have reported that IBP7 can promote cancer progression.
In non-small cell lung carcinoma, IBP7 was associated with metastatic disease [40]. In T-cell acute
lymphoblastic leukemia, IBP7 inhibited proliferation by causing G0/G1 arrest and induced drug
resistance [41]. In gastric cancer, IBP7 overexpression was associated with tumor progression and poor
survival [42].
PDGFD has recently gained tremendous amount of attention due to its involvement in
carcinogenesis. In agreement with the oncogenic function of PDGFD in human malignancies, PDGFD
overexpression has been detected in a variety of cancers including prostate, lung, renal, ovarian, brain,
and pancreatic cancer [18,43]. PDGFD regulates multitude cellular pathways including PI3K/Akt,
NF-κB, Notch, ERK, mTOR, MAPK, VEGF, MMPs, Cyclin D1 and BCL2 [44–49]. Moreover, PDGFD
has been found to regulate the EMT process that is important for tumor metastasis [44,45,50,51].
Prostate cancer cells with an EMT phenotype induced by PDGFD displayed CSC features, including
increased expression of SOX2, NANOG, OCT4, and Notch-1, and enhanced sphere-forming ability
and rapid tumor growth in vivo [44]. In addition, it has also been reported that tissue-resident stem
cells induce EMT through interaction with the TME via PDGFD, thereby leading to increased number
of CSCs and tumor growth [50].
Since FBLN3, IBP5, IBP7, and PDGFD play oncogenic roles through the regulation of tumor cell
growth, invasion, metastasis, EMT, and CSCs, targeting these proteins, their receptors, or downstream
signaling pathways could be valuable strategies to interfere with stromal-cancer cell heterotypic
communication. Interestingly, we found that pharmacologic inhibition of EGFR, IGFR, and PDGFR
signaling pathways efficiently blocked CAF-induced orosphere formation and anchorage-independent
growth in HNSCC cells. Nevertheless, it is plausible that the potent anti-stemness effects observed
Cancers 2018, 10, 334 11 of 17

by targeting EGFR, IGFR, and PDGFR pathways could also be due to a direct effect on the tumor
cells, since these signaling pathways are frequently altered in different cancers, including HNSCC.
A limitation of this study is that only one population of primary CAFs and NFs were used.

4. Materials and Methods

4.1. Drugs
EC-8042 (EntreChem, Oviedo, Spain), OSI-906 (Selleckchem, Suffolk, UK), CP-673451 (Selleckchem,
Suffolk, UK), 2-Guanidinoethylmercaptosuccinic acid (GEMSA) (Abcam, Cambridge, UK), and
Gefitinib (TOCRIS Bioscience, Bristol, UK) were prepared as 1 mM solutions in sterile DMSO or
water, according to manufacturer’s indications, for in vitro experiments, maintained at −20 ◦ C and
brought to the final concentration just before use.

4.2. Cell Culture

FaDu cells were purchased to the American Type Culture Collection, and the HNSCC cell line
SCC38 derived from a primary tumor (T2N0M0) was kindly provided by Dr. R. Grenman (Department
of Otolaryngology, University Central Hospital, Turku, Finland) [52]. Primary cancer-associated
fibroblasts (CAFs) were obtained from minced tumor tissue of surgically resected HNSCC at the
Hospital Universitario Central de Asturias. Normal dermal fibroblasts (NF) were obtained from the
dermis of human neonatal foreskin, by enzymatic cell disaggregation as described [53]. Cell line
authentication was performed by DNA (STR) profiling at the SCT Core Facilities (University of Oviedo,
Asturias, Spain). All cell lines were tested periodically for mycoplasma contamination by PCR to
specifically amplify a conserved region of the mycoplasma 16S ribosomal RNA gene (Biotools Detection
kit, Madrid, Spain).
HNSCC cells and fibroblasts were grown in DMEM (Biowest, Nuaillé, France) supplemented
with 10% fetal bovine serum (FBS) (Gibco, Waltham, MA, USA), 100 U/mL penicillin (Biowest,
Nuaillé, France), 200 mg/mL streptomycin (Biowest, Nuaillé, France), 2 mM L-glutamine, 20 mM
HEPES (pH 7.3) (Biowest, Nuaillé, France), and 100 mM MEM non-essential amino acids (Biowest,
Nuaillé, France).

4.3. Conditioned Media Production

Primary CAFs and NFs were grown in T175 flask (Corning, Corning, NY, USA) with complete
medium until reaching 80–90% confluence (approx 4–5×106 cells). Then, medium was replaced
and cells were grown for 72 h in 20 mL of DMEM-F12 (GE Healthcare, Pittsburg, PA, USA) with
100 U/mL penicillin (Biowest, Nuaillé, France), 200 mg/mL streptomycin (Biowest, Nuaillé, France)
in the absence of supplements or FBS and also without phenol red (for Mass Spec analysis). Next,
conditioned media (CM) were collected and filtered through a 0.45 µm pore filter (Sigma-Aldrich,
St. Louis, MO, USA) and frozen at −80 ◦ C until use.

4.4. Anchorage-Independent Cell Growth

For anchorage-independent cell growth, normal 96-well tissue culture plates (Corning) were
coated with 10 g/L of the anti-adhesive polymer poly-2-hydroxyethyl methacrylate (polyHEMA,
Sigma-Aldrich) in 95% ethanol and dried at 56 ◦ C for 16 h to prevent cell attachment.
PolyHEMA-coated plates were sterilized with UV-light for 30 min before use.
FaDu and SCC38 cells were plated into 96-well tissue culture plates at a density of 10,000 cells
per well. Cell proliferation was measured after 4 days. Quantification of cell number was determined
in quadruplicates using a tetrazolium-based MTS test (CellTiter 96 Aqueous One Solution Cell
Proliferation Assay from Promega, Madison, WI, USA), reading the absorbance at 490 nm with
the use of a Synergy HT plate reader (BioTek, Winooski, VT, USA).
Cancers 2018, 10, 334 12 of 17

4.5. Orosphere Formation Assay

HNSCC-derived cells lines were plated at a density of 500 cells/mL in 75-cm2 flask (5000 cells) or
6-well tissue culture plates (1000 cells/well), treated with a sterile solution of polyHEMA (10 g/L in 95%
ethanol) (Sigma-Aldrich) to prevent cell attachment, and grown in either DMEM-F12 (GE Healthcare,
Pittsburgh, PA, USA) without any supplements, conditioned media from CAFs or NFs, or the
standard cancer stem cell medium [54,55], which is composed of DMEM-F12 supplemented with 1%
Glutamax (Life Technologies, Waltham, MA, USA), 2% B27 Supplement (Life Technologies, Waltham,
MA, USA), 20 ng/mL human EGF (PeproTech, London, UK), 10 ng/mL human bFGF (PeproTech,
London, UK), 4 µg/mL insulin (Sigma-Aldrich), 100 U/mL penicillin, and 200 mg/mL streptomycin.
The above medium will be referred to in this manuscript as supplemented medium as opposed to
non-supplemented DMEM-F12 medium.
After 10–12 days, well-formed spheres were photographed in Leica Microsystems microscope
DMIL T, coupled with a Leica DC500 High-resolution Digital Camera (Leica Microsystems, Barcelona,
Spain). Then the spheres were centrifuged at 500 rpm for 5 min, washed with PBS and collected for
RNA extraction or MTS assay.

4.6. RNA Extraction and Real-Time RT-PCR

Total RNA was extracted from HNSCC cells using Trizol reagent (Invitrogen Life Technologies,
Waltham, MA, USA), and cDNA synthesized with Superscript II RT-PCR System (Invitrogen Life
Technologies), according to manufacturer’s protocols. Gene expression was analyzed by Real-time
PCR using the StepOnePlus Real-Time PCR System (Applied Biosystems, Waltham, MA, USA)
following Applied Biosystems’ SYBR Green Master Mix protocol. Reactions were carried out using the
primers detailed in Table S1. The constitutively expressed RPL19 ribosomal coding gene was used as
endogenous control. The relative mRNA expression was calculated using the 2−∆∆CT method.

4.7. Secretome Analysis by Mass Spectrometry

Conditioned media from NFs and CAFs were obtained as above described, and protein
concentration was determined by Pierce BCA Protein Assay Kit (Thermo Fisher, Waltham, MA,
USA). Three independent batches of CM were produced from each experimental condition and used
as biological replicates.
50 µg of protein (for each CM) were digested overnight with immobilized TLCK-trypsin (20 TAME
units/mg) (Thermo Scientific, Waltham, MA, USA) for 16 h at 37 ◦ C. The resultant peptide solutions
were desalted by solid phase extraction (SPE) using Oasis HLB extraction cartridges (Waters UK Ltd.,
Manchester, UK) according to manufacturer’s instructions. Briefly, cartridges coupled to a vacuum
manifold set at 5 mm Hg, were activated with 1 mL of 100% acetonitrile (ACN) and equilibrated with
1.5 mL of wash solution (1% ACN, 0.1% TFA in water). After the cartridges were loaded with peptide
solution, they were washed with 1 mL of wash solution. Peptides were eluted with 0.5 mL of 50%
ACN containing 0.1% TFA and dried in a speed vacuum centrifuge.
Peptide pellets were solubilized in 0.1% TFA and run in two mass spectrometry platforms
consisting in an LTQ-Orbitrap XL (Thermo Scientific, Waltham, MA, USA) coupled to a nanoACQUITY
ultra performance LC (Waters Corp., Milford, MA, USA) and a QExactive plus (Thermo Scientific)
online connected to a Ultimate 3000 RSLC chromatographer (Thermo Scientific). For the LTQ-Orbitrap
XL system, peptides were loaded into a nanoACQUITY trap column and separated on BEH C18 nano
ACQUITY column. Separation was performed using a 180-min gradient with solvent B 5–25% at a
flow rate of 300 nL/min (mobile phase B; 100% ACN and 0.1% FA; mobile phase A; 100% water and
0.1% FA). The mass spectrometer was operated in data-dependent acquisition mode for top 5 CID
acquisitions. For the QExactive plus system, peptides were loaded into an Acclaim PepMap 100 trap
column and separated on an Acclaim PepMap RSLC analytical column. Separation was performed
using a 120-min gradient with solvent B 5–25% at a flow rate of 300 nL/min (mobile phase B; 100%
Cancers 2018, 10, 334 13 of 17

ACN and 0.1% FA; mobile phase A; 100% water and 0.1% FA). The mass spectrometer was operated in
data-dependent acquisition mode for top 15 CID acquisitions.

4.8. Mass Spectrometry Data Analysis

Raw files were converted to peak lists (in the Mascot Generic Format) using Mascot Distiller
(version 2.3.0) and searched using Mascot Server (version 2.3.01) against the SwissProt Uniprot
database (2012-03-10) restricted to the relevant taxonomy. Mass windows for tolerance for MS scans
were 10 ppm and 600 mmu for MS/MS. Fixed modification of carbamidomethylation of cysteine and
variable modifications of oxidation of methionine and glutamine to pyroglutamate conversion were
permitted. Mascot result files were parsed using a Perl script that uses the Mascot Parser files provided
by Matrix Science. The threshold for accepting peptides as being positively identified was set at an
expectancy score of 0.05.
For individual peptide quantification, we used Pescal, a software that automates the construction
of extracted ion chromatograms for all peptides identified across all samples being compared [56,57].
Independent Pescal analyses were run for the quantification of peptides identified in the LTQ-Orbitrap
XL or QExactive systems. Protein intensity calculation was automated with an in house developed
script that sums the signals of peptides obtained from both platforms and comprised in the same
protein. Data were normalized by dividing each protein intensity by the sum of all protein intensities
within a sample. Fold changes were calculated by averaging the normalized intensities of peptides
within a sample group and dividing these by the intensities of the control group. Fold changes were
then log transformed before calculation of significance using an unpaired t-test.

4.9. Statistical Analysis

The data are presented as the mean ± standard deviation (SD), unless otherwise stated, and
compared using unpaired Student’s t-test or one-way ANOVA test and Holm-Sidak’s multiple
comparisons test. The normality of the data was analyzed using the Kolmogorov-Smirnov test.
Statistical analysis was performed using GraphPad Prism version 6.0 (GraphPad Software Inc., La Jolla,
CA, USA). p values less than 0.05 were considered statistically significant (* p < 0.05; ** p < 0.01;
*** p < 0.005).

5. Conclusions
Together our findings uncover novel insights into the tumor stroma–CSC communication, and
provide also a novel therapeutic rationale to effectively block the CAF-enhanced CSC niche signaling
circuit, to ultimately overcome CSC-mediated disease progression and resistance to therapy.

Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6694/10/9/334/s1,

Supplementary Table S1: Primers used for real-time RT-PCR.
Author Contributions: Conceptualization, S.Á.-T. and J.M.G.-P.; Funding acquisition, J.P.R. and J.M.G.-P.;
Investigation, S.Á.-T., C.G.-I., M.Á.V., F.H.-P. and R.G.-D.; Methodology, S.Á.-T., C.G.-I., M.Á.V., P.C. (Pedro
Casado) and P.C. (Pedro Cutillas); Project administration, S.Á.-T. and J.M.G.-P.; Resources, J.P.R., F.C., A.G., F.M.,
M.H. and P.C. (Pedro Cutillas); Software, P.C. (Pedro Casado) and P.C. (Pedro Cutillas); Supervision, J.M.G.-P.;
Validation, N.d.-R.-I.; Visualization, S.Á.-T., P.C. (Pedro Casado) and P.C. (Pedro Cutillas); Writing-original draft,
S.Á.-T. and J.M.G.-P.; Writing-review & editing, J.P.R., F.C., A.G., F.M. and P.C. (Pedro Cutillas).
Funding: This study was supported by grants from the Plan Nacional de I+D+I 2013-2016 ISCIII (PI13/00259),
PI16/00280, and CIBERONC (CB16/12/00390) Spain, the Principado de Asturias (GRUPIN14-003), Fundación
Merck Salud (17-CC-008), the Instituto de Investigación Sanitaria del Principado de Asturias (ISPA), and the
FEDER Funding Program from the European Union. SAT and MAV were recipients of fellowships from ISCIII
(FI12/00415 and CD13/00157, respectively).
Acknowledgments: We thank René Rodríguez for critical reading of the manuscript, and also Teresa Ortega
Montoliu and OIB staff for the administrative support.
Conflicts of Interest: F.M. reports ownership of stock in EntreChem SL. All other authors declare they have no
competing interests.
Cancers 2018, 10, 334 14 of 17

Abbreviations
CAFs Cancer-associated fibroblasts
CSC Cancer stem cells
HNSCC Head and neck squamous cell carcinomas
EMT Epithelial-mesenchymal transition
NFs Normal fibroblasts
EGF Epidermal growth factor
EGFR Epidermal growth factor receptor
IGF Insulin-like growth factor
IGFR Insulin-like growth factor receptor
PDGF Platelet-derived growth factor
PDGFR Platelet-derived growth factor receptor
ECM Extracellular matrix
TME Tumor microenvironment
MMPs Matrix metalloproteinases
MS Mass spectrometry
GEMSA 2-Guanidinoethylmercaptosuccinic acid
FBS Fetal bovine serum
HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
CM Conditioned media
polyHEMA poly-2-hydroxyethyl methacrylate
FGF Fibroblast growth factor
SPE Solid phase extraction
CAN Acetonitrile
TFA Trifluoroacetic acid
FA Formic acid
CBPE Carboxypeptidase E
PDGFD Platelet-derived growth factor D
EGF-containing fibulin-like extracellular matrix
FBLN3
protein-1
IBPs Insulin-like growth factor binding protein
IBP5 Insulin-like growth factor binding protein-5
IBP7 Insulin-like growth factor binding protein-7
HGF Hepatocyte growth factor
IL-6 Interleukin-6

References
1. Koontongkaew, S. The tumor microenvironment contribution to development, growth, invasion and
metastasis of head and neck squamous cell carcinomas. J. Cancer 2013, 4, 66–83. [CrossRef] [PubMed]
2. Calvo, F.; Ege, N.; Grande-Garcia, A.; Hooper, S.; Jenkins, R.P.; Chaudhry, S.I.; Harrington, K.; Williamson, P.;
Moeendarbary, E.; Charras, G.; et al. Mechanotransduction and YAP-dependent matrix remodelling is
required for the generation and maintenance of cancer-associated fibroblasts. Nat. Cell Biol. 2013, 15, 637–646.
[CrossRef] [PubMed]
3. Koczorowska, M.M.; Tholen, S.; Bucher, F.; Lutz, L.; Kizhakkedathu, J.N.; De Wever, O.; Wellner, U.F.;
Biniossek, M.L.; Stahl, A.; Lassmann, S.; et al. Fibroblast activation protein-alpha, a stromal cell surface
protease, shapes key features of cancer associated fibroblasts through proteome and degradome alterations.
Mol. Oncol. 2016, 10, 40–58. [CrossRef] [PubMed]
4. De Wever, O.; Mareel, M. Role of tissue stroma in cancer cell invasion. J. Pathol. 2003, 200, 429–447. [CrossRef]
[PubMed]
Cancers 2018, 10, 334 15 of 17

5. Bello, I.O.; Vered, M.; Dayan, D.; Dobriyan, A.; Yahalom, R.; Alanen, K.; Nieminen, P.; Kantola, S.;
Laara, E.; Salo, T. Cancer-associated fibroblasts, a parameter of the tumor microenvironment, overcomes
carcinoma-associated parameters in the prognosis of patients with mobile tongue cancer. Oral Oncol. 2011,
47, 33–38. [CrossRef] [PubMed]
6. Vered, M.; Dobriyan, A.; Dayan, D.; Yahalom, R.; Talmi, Y.P.; Bedrin, L.; Barshack, I.; Taicher, S. Tumor-host
histopathologic variables, stromal myofibroblasts and risk score, are significantly associated with recurrent
disease in tongue cancer. Cancer Sci. 2010, 101, 274–280. [CrossRef] [PubMed]
7. Liu, H.X.; Li, X.L.; Dong, C.F. Epigenetic and metabolic regulation of breast cancer stem cells. J. Zhejiang
Univ. Sci. B 2015, 16, 10–17. [CrossRef] [PubMed]
8. Liu, H.; Patel, M.R.; Prescher, J.A.; Patsialou, A.; Qian, D.; Lin, J.; Wen, S.; Chang, Y.F.; Bachmann, M.H.;
Shimono, Y.; et al. Cancer stem cells from human breast tumors are involved in spontaneous metastases in
orthotopic mouse models. Proc. Natl. Acad. Sci. USA 2010, 107, 18115–18120. [CrossRef] [PubMed]
9. He, K.; Xu, T.; Goldkorn, A. Cancer cells cyclically lose and regain drug-resistant highly tumorigenic features
characteristic of a cancer stem-like phenotype. Mol. Cancer Ther. 2011, 10, 938–948. [CrossRef] [PubMed]
10. Chaffer, C.L.; Brueckmann, I.; Scheel, C.; Kaestli, A.J.; Wiggins, P.A.; Rodrigues, L.O.; Brooks, M.; Reinhardt, F.;
Su, Y.; Polyak, K.; et al. Normal and neoplastic nonstem cells can spontaneously convert to a stem-like state.
Proc. Natl. Acad. Sci. USA 2011, 108, 7950–7955. [CrossRef] [PubMed]
11. Chaffer, C.L.; Marjanovic, N.D.; Lee, T.; Bell, G.; Kleer, C.G.; Reinhardt, F.; D’Alessio, A.C.; Young, R.A.;
Weinberg, R.A. Poised chromatin at the ZEB1 promoter enables breast cancer cell plasticity and enhances
tumorigenicity. Cell 2013, 154, 61–74. [CrossRef] [PubMed]
12. Weiland, A.; Roswall, P.; Hatzihristidis, T.C.; Pietras, K.; Ostman, A.; Strell, C. Fibroblast-dependent
regulation of the stem cell properties of cancer cells. Neoplasma 2012, 59, 719–727. [CrossRef] [PubMed]
13. Marimuthu, A.; Chavan, S.; Sathe, G.; Sahasrabuddhe, N.A.; Srikanth, S.M.; Renuse, S.; Ahmad, S.;
Radhakrishnan, A.; Barbhuiya, M.A.; Kumar, R.V.; et al. Identification of head and neck squamous cell
carcinoma biomarker candidates through proteomic analysis of cancer cell secretome. Biochim. Biophys. Acta
2013, 1834, 2308–2316. [CrossRef] [PubMed]
14. Wu, C.C.; Hsu, C.W.; Chen, C.D.; Yu, C.J.; Chang, K.P.; Tai, D.I.; Liu, H.P.; Su, W.H.; Chang, Y.S.; Yu, J.S.
Candidate serological biomarkers for cancer identified from the secretomes of 23 cancer cell lines and the
human protein atlas. Mol. Cell. Proteom. 2010, 9, 1100–1117. [CrossRef] [PubMed]
15. Fan, S.; Li, X.; Li, L.; Wang, L.; Du, Z.; Yang, Y.; Zhao, J.; Li, Y. Silencing of carboxypeptidase E inhibits cell
proliferation, tumorigenicity, and metastasis of osteosarcoma cells. Onco. Targets Ther. 2016, 9, 2795–2803.
[PubMed]
16. Liu, A.; Shao, C.; Jin, G.; Liu, R.; Hao, J.; Shao, Z.; Liu, Q.; Hu, X. Downregulation of CPE regulates cell
proliferation and chemosensitivity in pancreatic cancer. Tumour Biol. 2014, 35, 12459–12465. [CrossRef]
[PubMed]
17. Beattie, J.; Allan, G.J.; Lochrie, J.D.; Flint, D.J. Insulin-like growth factor-binding protein-5 (IGFBP-5):
A critical member of the IGF axis. Biochem. J. 2006, 395, 1–19. [CrossRef] [PubMed]
18. Wang, Z.; Kong, D.; Li, Y.; Sarkar, F.H. PDGF-D signaling: A novel target in cancer therapy. Curr. Drug
Targets 2009, 10, 38–41. [CrossRef] [PubMed]
19. Ma, L.; Zhang, G.; Miao, X.B.; Deng, X.B.; Wu, Y.; Liu, Y.; Jin, Z.R.; Li, X.Q.; Liu, Q.Z.; Sun, D.X.; et al. Cancer
stem-like cell properties are regulated by EGFR/AKT/beta-catenin signaling and preferentially inhibited by
gefitinib in nasopharyngeal carcinoma. FEBS J. 2013, 280, 2027–2041. [CrossRef] [PubMed]
20. Tornin, J.; Martinez-Cruzado, L.; Santos, L.; Rodriguez, A.; Nunez, L.E.; Oro, P.; Hermosilla, M.A.; Allonca, E.;
Fernandez-Garcia, M.T.; Astudillo, A.; et al. Inhibition of SP1 by the mithramycin analog EC-8042 efficiently
targets tumor initiating cells in sarcoma. Oncotarget 2016, 7, 30935–30950. [CrossRef] [PubMed]
21. Hanahan, D.; Coussens, L.M. Accessories to the crime: Functions of cells recruited to the tumor
microenvironment. Cancer Cell 2012, 21, 309–322. [CrossRef] [PubMed]
22. Hanahan, D.; Weinberg, R.A. Hallmarks of cancer: The next generation. Cell 2011, 144, 646–674. [CrossRef]
[PubMed]
23. Joyce, J.A.; Pollard, J.W. Microenvironmental regulation of metastasis. Nat. Rev. Cancer 2009, 9, 239–252.
[CrossRef] [PubMed]
24. Albini, A.; Sporn, M.B. The tumour microenvironment as a target for chemoprevention. Nat. Rev. Cancer
2007, 7, 139–147. [CrossRef] [PubMed]
Cancers 2018, 10, 334 16 of 17

25. Curry, J.M.; Sprandio, J.; Cognetti, D.; Luginbuhl, A.; Bar-ad, V.; Pribitkin, E.; Tuluc, M. Tumor
microenvironment in head and neck squamous cell carcinoma. Semin. Oncol. 2014, 41, 217–234. [CrossRef]
[PubMed]
26. Mueller, M.M.; Fusenig, N.E. Friends or foes—Bipolar effects of the tumour stroma in cancer. Nat. Rev.
Cancer 2004, 4, 839–849. [CrossRef] [PubMed]
27. Donnarumma, E.; Fiore, D.; Nappa, M.; Roscigno, G.; Adamo, A.; Iaboni, M.; Russo, V.; Affinito, A.; Puoti, I.;
Quintavalle, C.; et al. Cancer-associated fibroblasts release exosomal microRNAs that dictate an aggressive
phenotype in breast cancer. Oncotarget 2017, 8, 19592–19608. [CrossRef] [PubMed]
28. Shimoda, M.; Principe, S.; Jackson, H.W.; Luga, V.; Fang, H.; Molyneux, S.D.; Shao, Y.W.; Aiken, A.;
Waterhouse, P.D.; Karamboulas, C.; et al. Loss of the Timp gene family is sufficient for the acquisition of the
CAF-like cell state. Nat. Cell Biol. 2014, 16, 889–901. [CrossRef] [PubMed]
29. Vermeulen, L.; De Sousa, E.M.F.; van der Heijden, M.; Cameron, K.; de Jong, J.H.; Borovski, T.; Tuynman, J.B.;
Todaro, M.; Merz, C.; Rodermond, H.; et al. Wnt activity defines colon cancer stem cells and is regulated by
the microenvironment. Nat. Cell Biol. 2010, 12, 468–476. [CrossRef] [PubMed]
30. Giannoni, E.; Bianchini, F.; Masieri, L.; Serni, S.; Torre, E.; Calorini, L.; Chiarugi, P. Reciprocal activation
of prostate cancer cells and cancer-associated fibroblasts stimulates epithelial-mesenchymal transition and
cancer stemness. Cancer Res. 2010, 70, 6945–6956. [CrossRef] [PubMed]
31. Ishitoya, J.; Toriyama, M.; Oguchi, N.; Kitamura, K.; Ohshima, M.; Asano, K.; Yamamoto, T. Gene
amplification and overexpression of EGF receptor in squamous cell carcinomas of the head and neck.
Br. J. Cancer 1989, 59, 559–562. [CrossRef] [PubMed]
32. Grandis, J.R.; Tweardy, D.J. Elevated levels of transforming growth factor alpha and epidermal growth factor
receptor messenger RNA are early markers of carcinogenesis in head and neck cancer. Cancer Res. 1993, 53,
3579–3584. [PubMed]
33. Ang, K.K.; Berkey, B.A.; Tu, X.; Zhang, H.Z.; Katz, R.; Hammond, E.H.; Fu, K.K.; Milas, L. Impact of
epidermal growth factor receptor expression on survival and pattern of relapse in patients with advanced
head and neck carcinoma. Cancer Res. 2002, 62, 7350–7356. [PubMed]
34. Firth, S.M.; Baxter, R.C. Cellular actions of the insulin-like growth factor binding proteins. Endocr. Rev. 2002,
23, 824–854. [CrossRef] [PubMed]
35. Jones, J.I.; Clemmons, D.R. Insulin-like growth factors and their binding proteins: Biological actions. Endocr.
Rev. 1995, 16, 3–34. [PubMed]
36. Unger, C.; Kramer, N.; Unterleuthner, D.; Scherzer, M.; Burian, A.; Rudisch, A.; Stadler, M.; Schlederer, M.;
Lenhardt, D.; Riedl, A.; et al. Stromal-derived IGF2 promotes colon cancer progression via paracrine and
autocrine mechanisms. Oncogene 2017, 36, 5341–5355. [CrossRef] [PubMed]
37. Mohan, S.; Baylink, D.J. IGF-binding proteins are multifunctional and act via IGF-dependent and
-independent mechanisms. J. Endocrinol. 2002, 175, 19–31. [CrossRef] [PubMed]
38. Johnson, S.K.; Haun, R.S. Insulin-like growth factor binding protein-5 influences pancreatic cancer cell
growth. World J. Gastroenterol. 2009, 15, 3355–3366. [CrossRef] [PubMed]
39. De Bont, J.M.; van Doorn, J.; Reddingius, R.E.; Graat, G.H.; Passier, M.M.; den Boer, M.L.; Pieters, R. Various
components of the insulin-like growth factor system in tumor tissue, cerebrospinal fluid and peripheral
blood of pediatric medulloblastoma and ependymoma patients. Int. J. Cancer 2008, 123, 594–600. [CrossRef]
[PubMed]
40. Zhao, W.; Wang, J.; Zhu, B.; Duan, Y.; Chen, F.; Nian, W.; Sun, J.; Zhang, B.; Tong, Z.; Chen, Z. IGFBP7
functions as a potential lymphangiogenesis inducer in non-small cell lung carcinoma. Oncol. Rep. 2016, 35,
1483–1492. [CrossRef] [PubMed]
41. Bartram, I.; Erben, U.; Ortiz-Tanchez, J.; Blunert, K.; Schlee, C.; Neumann, M.; Heesch, S.; Baldus, C.D.
Inhibition of IGF1-R overcomes IGFBP7-induced chemotherapy resistance in T-ALL. BMC Cancer 2015,
15, e663. [CrossRef] [PubMed]
42. Sato, Y.; Inokuchi, M.; Takagi, Y.; Otsuki, S.; Fujimori, Y.; Yanaka, Y.; Kobayashi, K.; Higuchi, K.; Kojima, K.;
Kawano, T. Relationship between expression of IGFBP7 and clinicopathological variables in gastric cancer.
J. Clin. Pathol. 2015, 68, 795–801. [CrossRef] [PubMed]
43. Wang, Z.; Ahmad, A.; Li, Y.; Kong, D.; Azmi, A.S.; Banerjee, S.; Sarkar, F.H. Emerging roles of PDGF-D
signaling pathway in tumor development and progression. Biochim. Biophys. Acta 2010, 1806, 122–130.
[CrossRef] [PubMed]
Cancers 2018, 10, 334 17 of 17

44. Kong, D.; Banerjee, S.; Ahmad, A.; Li, Y.; Wang, Z.; Sethi, S.; Sarkar, F.H. Epithelial to mesenchymal transition
is mechanistically linked with stem cell signatures in prostate cancer cells. PLoS ONE 2010, 5, e12445.
[CrossRef] [PubMed]
45. Kong, D.; Li, Y.; Wang, Z.; Banerjee, S.; Ahmad, A.; Kim, H.R.; Sarkar, F.H. miR-200 regulates
PDGF-D-mediated epithelial-mesenchymal transition, adhesion, and invasion of prostate cancer cells.
Stem Cells 2009, 27, 1712–1721. [CrossRef] [PubMed]
46. Kong, D.; Wang, Z.; Sarkar, S.H.; Li, Y.; Banerjee, S.; Saliganan, A.; Kim, H.R.; Cher, M.L.; Sarkar, F.H.
Platelet-derived growth factor-D overexpression contributes to epithelial-mesenchymal transition of PC3
prostate cancer cells. Stem Cells 2008, 26, 1425–1435. [CrossRef] [PubMed]
47. Ahmad, A.; Wang, Z.; Kong, D.; Ali, R.; Ali, S.; Banerjee, S.; Sarkar, F.H. Platelet-derived growth factor-D
contributes to aggressiveness of breast cancer cells by up-regulating Notch and NF-kappaB signaling
pathways. Breast Cancer Res. Treat. 2011, 126, 15–25. [CrossRef] [PubMed]
48. Wang, Z.; Kong, D.; Banerjee, S.; Li, Y.; Adsay, N.V.; Abbruzzese, J.; Sarkar, F.H. Down-regulation of
platelet-derived growth factor-D inhibits cell growth and angiogenesis through inactivation of Notch-1 and
nuclear factor-kappaB signaling. Cancer Res. 2007, 67, 11377–11385. [CrossRef] [PubMed]
49. Okada, A.; Yaguchi, T.; Kanno, T.; Gotoh, A.; Nakano, T.; Nishizaki, T. PDGF-D/PDGF-betabeta
receptor-regulated chemotaxis of malignant mesothelioma cells. Cell Physiol. Biochem. 2012, 29, 241–250.
[CrossRef] [PubMed]
50. Devarajan, E.; Song, Y.H.; Krishnappa, S.; Alt, E. Epithelial-mesenchymal transition in breast cancer lines
is mediated through PDGF-D released by tissue-resident stem cells. Int. J. Cancer 2012, 131, 1023–1031.
[CrossRef] [PubMed]
51. Sethi, S.; Sarkar, F.H.; Ahmed, Q.; Bandyopadhyay, S.; Nahleh, Z.A.; Semaan, A.; Sakr, W.; Munkarah, A.;
Ali-Fehmi, R. Molecular markers of epithelial-to-mesenchymal transition are associated with tumor
aggressiveness in breast carcinoma. Transl. Oncol. 2011, 4, 222–226. [CrossRef] [PubMed]
52. Lansford, C.D.; Grenman, R., Bier; Somers, K.D.; Kim, S.-Y.; Whiteside, T.L.; Clayman, G.L.;
Welkoborsky, H.-J.; Carey, T.E. Head and Neck Cancer. In Human Cell Culture, Vol. II; Kluwer Academic
Publisher: Dordrecht, The Netherlands, 1999.
53. Leigh, I.M.; Watt, F.M. Keratinocyte Methods; Cambridge University Press: Cambridge, UK, 1994.
54. Martinez-Cruzado, L.; Tornin, J.; Santos, L.; Rodriguez, A.; García-Castro, J.; Morís, F.; Rodriguez, R. Aldh1
Expression and Activity Increase During Tumor Evolution in Sarcoma Cancer Stem Cell Populations. Sci.
Rep. 2016, 6, e27878. [CrossRef] [PubMed]
55. Weiswald, L.B.; Bellet, D.; Dangles-Marie, V. Spherical cancer models in tumor biology. Neoplasia 2015, 17,
1–15. [CrossRef] [PubMed]
56. Cutillas, P.R.; Vanhaesebroeck, B. Quantitative profile of five murine core proteomes using label-free
functional proteomics. Mol. Cell. Proteom. 2007, 6, 1560–1573. [CrossRef] [PubMed]
57. Alcolea, M.P.; Casado, P.; Rodriguez-Prados, J.C.; Vanhaesebroeck, B.; Cutillas, P.R. Phosphoproteomic
analysis of leukemia cells under basal and drug-treated conditions identifies markers of kinase pathway
activation and mechanisms of resistance. Mol. Cell. Proteom. 2012, 11, 453–466. [CrossRef] [PubMed]

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).