Oncotarget64330
www.impactjournals.com/oncotarget
www.impactjournals.com/oncotarget/
Oncotarget, Vol. 7, No. 39
The iron chelator deferasirox induces apoptosis by targeting
oncogenic Pyk2/β-catenin signaling in human multiple myeloma
Yusuke Kamihara
1,*
, Kohichi Takada
1,*
, Tsutomu Sato
1
, Yutaka Kawano
1
, Kazuyuki
Murase
1
, Yohei Arihara
1
, Shohei Kikuchi
1
, Naotaka Hayasaka
1
, Makoto Usami
1
,
Satoshi Iyama
1
, Koji Miyanishi
1
, Yasushi Sato
1
, Masayoshi Kobune
1
, Junji Kato
1
1
Department of Medical Oncology and Hematology, Sapporo Medical University School of Medicine, Japan
*
These authors have contributed equally to this work
Correspondence to: Junji Kato, email: jkato@sapmed.ac.jp
Keywords: deferasirox, Pyk2, β-catenin, apoptosis, multiple myeloma
Received: October 06, 2015 Accepted: August 21, 2016 Published: September 02, 2016
ABSTRACT
Deregulated iron metabolism underlies the pathogenesis of many human cancers.
Recently, low expression of ferroportin, which is the only identied non-heme iron
exporter, has been associated with signicantly reduced overall survival in multiple
myeloma (MM); however, the altered iron metabolism in MM biology remains unclear.
In this study we demonstrated, by live cell imaging, that MM cells have increased
intracellular iron levels as compared with normal cells. In experiments to test the
effect of iron chelation on the growth of MM cells, we found that deferasirox (DFX),
an oral iron chelator used to treat iron overload in clinical practice, inhibits MM cell
growth both in vivo and in vitro. Mechanistically, DFX was found to induce apoptosis
of MM cells via the inhibition of proline-rich tyrosine kinase 2 (Pyk2), which is known
to promote tumor growth in MM. Inhibition of Pyk2 is caused by the suppression of
reactive oxygen species, and leads to downregulation of the Wnt/β-catenin signaling
pathway. Taken together, our ndings indicate that high levels of intracellular iron,
which might be due to low ferroportin expression, play a role in MM pathophysiology.
Therefore, DFX may provide a therapeutic option for MM that is driven by deregulated
iron homeostasis and/or Pyk2/Wnt signaling.
INTRODUCTION
Iron is essential for many fundamental cellular
functions, including proliferation and DNA synthesis
[1]. Accumulating evidence suggests that abnormal iron
metabolism plays an important role in carcinogenesis and
in the progression of many tumors [2]. In cancer cells,
the demand for iron increases in response to sustained,
accelerated cell proliferation and DNA synthesis [3].
Multiple myeloma (MM) is characterized by clonal
proliferation of long-lived plasma cells within the bone
marrow. Despite recent advances in its treatment, MM
remains an incurable disease, underlining the continuing
need to explore its molecular characteristics [4]. Recently,
two independent groups reported that serum ferritin,
which is used as a marker for iron overload, can be a
negative prognostic indicator in MM [5, 6]. However,
the mechanisms underlying this effect have not been
claried, and the signicance of iron metabolism in MM
cells remains unclear. Ferroportin has also been shown
to have a role in MM progression [7]. This cell-surface
transmembrane protein is the only non-heme iron exporter
identied in mammalian cells, and is a pivotal protein of
iron homeostasis [8]. Reduced expression of ferroportin
can lead to intracellular iron overload.
We previously reported that excess iron is a feature
of myelodysplastic syndrome (MDS) and hepatocellular
carcinoma pathogenesis [9, 10]. Moreover, we found
that iron chelation with deferasirox (DFX) suppresses
leukemic transformation in MDS patients, and long-term
phlebotomy and low iron diet therapy lower the risk of
hepatocarcinogenesis in patients with chronic hepatitis
C. Remarkably, iron chelation therapy has been shown
to suppress cancer cell growth and is considered to be a
promising cancer treatment [11]. Several mechanisms for
the cytotoxic effects of iron chelation therapy have been
reported; however, its action might depend on both the cell
type and the iron chelation agent [12-14].
Research Paper
Oncotarget64331
www.impactjournals.com/oncotarget
Proline-rich tyrosine kinase 2 (Pyk2) is a member
of the focal adhesion kinase (FAK) family. It is a non-
receptor protein kinase that shares a similar central domain
with FAK, showing 60% sequence identity. In addition,
Pyk2 has a conserved arrangement of proline-rich regions
and tyrosine phosphorylation sites [15]. Pyk2 plays a
crucial role in cell proliferation, migration and invasion in
several cancers [16]. Specically, it has been implicated in
the progression of MM and in micro-environment-specic
MM cell survival [17, 18]. Hence, it is an attractive target
for MM therapeutics.
In the present study, we observed that reduced
ferroportin mRNA levels might lead to increased
intracellular iron concentrations in MM cells. These
observations prompted us to investigate the effects of iron
chelation with DFX on MM. Both in vitro and in vivo,
studies demonstrated that DFX induces apoptosis via the
inhibition of Pyk2/β-catenin signaling in MM. Overall, our
ndings reveal that dysregulation of iron metabolism is a
characteristic of MM and provide a strong rationale for using
DFX as a therapeutic Pyk2/β-catenin inhibitor to treat MM.
RESULTS
Expression of ferroportin correlates with clinical
outcomes
We previously demonstrated that an excess of
intracellular iron contributes to the pathogenesis of MDS
and hepatocellular carcinoma [9, 10]. To explore the
hypothesis that iron overload might be also associated
with MM pathogenesis, we initially analyzed the
prognostic relevance of genes related to iron metabolism
(Supplementary Tables S1 and S2) utilizing a data set in
the public domain containing microarray proles and MM
patient outcomes. As Gu et al. [7] have shown, transcript
levels of ferroportin are correlated with event-free survival
(EFS) and overall survival (OS) in this data set, with a
statistically signicant negative correlation between
ferroportin levels and EFS or OS (Supplementary Figure
S1A, S1B). We therefore evaluated ferroportin mRNA
expression in MM cell lines and primary MM cells.
Quantitative reverse transcription (qRT)-PCR studies
showed lower ferroportin mRNA expression in both MM
cell lines and primary MM cells as compared with control
cells (Figure 1A, 1B).
We also measured ferroportin transcript levels in
other hematological malignancies, including leukemia
and lymphoma. ferroportin mRNA expression was lower
in B cell malignant cells than in leukemic cells or normal
peripheral blood mononuclear cells (PBMCs) from healthy
volunteers (Supplementary Figure S2).
Increased intracellular iron in MM cells
We considered that low expression of ferroportin
might result in an increase in intracellular iron. To examine
intracellular iron levels in MM cell lines, we determined
the cellular iron content using FeRhoNox-1 staining [19].
Live cell imaging showed that cellular iron levels were
apparently higher in MM cell lines than in PBMCs from
healthy volunteers (Figure 1C). Together with another
study [7], our data suggested that excess intracellular iron
might contribute to MM pathogenesis.
Inhibition of MM cell proliferation by DFX
Our observations suggested that iron chelation might
be an effective therapeutic intervention against MM. To
examine whether DFX, an oral, long-acting iron chelator
that is approved for iron overload, might be effective in
MM, we tested its inuence on the proliferation of MM
cell lines using WST-1 assays. DFX signicantly reduced
the proliferation of MM cell lines, with IC
50
values ranging
from 3.2 to 47.9 μM (Figure 2A, 2B).
To assess the threshold level of intracellular iron
for predicting DFX-induced cytotoxicity, the correlation
between intracellular iron content and the IC
50
for DFX
was investigated in four MM cell lines. As shown in
Figure 2B, a signicant correlation was observed between
the FeRhoNox-1-positive area and the IC
50
of DFX (r
s
=
-0.74, P = 0.004). Next, because bone marrow stromal
cells (BMSCs) promote MM cell survival and induce drug
resistance in MM cells, we tested the effect of BMSCs on
the sensitivity of MM cells to DFX treatment. As shown
in Figure 2D, DFX overcame the survival advantage
conferred by the bone marrow microenvironment.
DFX induces apoptosis in human MM cells
To elucidate the cytotoxic mechanism of DFX on
MM cells, we examined proteins involved programmed
cell death by immunoblotting and performed ow
cytometry analysis. We observed increased proteolytic
cleavage of Caspase 9, Caspase 3, and PARP, but not
Caspase 8 (data not shown) in all MM cell lines (Figure
3A), indicating cell apoptosis. Treatment with the pan-
Caspase inhibitor Z-VAD signicantly inhibited DFX-
induced apoptosis (Supplementary Figure S3). These
results indicated that the cytotoxicity triggered by DFX is
mediated, at least in part, via caspase-dependent (intrinsic)
apoptosis.
As shown in Figure 3B, Annexin V/7-AAD staining
showed a markedly higher percentage of apoptosis among
MM1S and RPMI8226 cells (Figure 3B). A previous study
has shown that DFX induces autophagy in MM cells [20].
In the present study, however, autophagic cell death was not
indicated by immunoblotting of LC3A/B (data not shown).
To assess the contribution of iron chelation by
DFX to cytotoxicity, we investigated whether iron
supplementation might affect apoptosis. Pretreatment
of MM cells with ferric chloride (FeCl
3
, 100 μM) in
combination with DFX prevented DFX-induced apoptosis,
as evaluated by ow cytometry (Figure 3B). These
Oncotarget64332
www.impactjournals.com/oncotarget
observations strongly indicate that DFX-induced apoptosis
depends on iron chelation by DFX. Furthermore, DFX-
induced apoptosis was detected in MM cell lines and
primary MM cells (#3) under co-culture systems with
BMSCs (Figure 3C).
DFX inhibits phosphorylation of Pyk2
To clarify the mechanisms of DFX-induced
apoptosis in MM, we analyzed components of the
phosphoinositide 3-kinase (PI3K)/Akt signaling pathway
by PCR array (Qiagen, Hilden, Germany), because this
pathway plays a crucial role in MM pathogenesis [21].
Unexpectedly, as shown in Supplementary Figure S4,
FA K transcript expression was substantially suppressed
after DFX treatment in both MM1S and RPMI8226
cells. Expression of FAK protein was therefore tested
by immunoblotting in a panel of MM cells. However,
no expression of FAK protein was observed in three of
the four cell lines (Figure 4A), indicating that FAK is not
associated with DFX-induced apoptosis.
Next, we investigated Pyk2, which is a member
of the FAK family and shares 48% amino acid sequence
identify with FAK. Pyk2 was expressed differentially in
the four different MM cell lines (Figure 4B). Consistent
with a previous report [17], basal Pyk2 expression
was relatively low in U266 cells. Consequently, we
assessed Pyk2 activation after treatment with DFX.
DFX signicantly reduced the level of phosphorylated
Pyk2(Tyr402), indicating in a loss of Pyk2 activity
(Figure 4B). There were no substantial differences in
total Pyk2 levels. In addition, we assessed the janus
Figure 1: Decreased ferroportin expression and intracellular iron accumulation in MM cells. A, B. Expression of
ferroportin mRNA in MM cell lines (A) and in CD138
+
cells from MM patients (B) evaluated by qRT-PCR. Data are the mean of triplicate
measurements. Error bars represent the SD. C. Intracellular iron content was monitored by live cell microscopy. Intracellular Fe
2+
was
stained by a FeRhoNox-1 probe. Scale bars, 10 μM.
Oncotarget64333
www.impactjournals.com/oncotarget
kinases Jak1, Jak2 and Jak3, because these molecules
are also relevant in MM pathogenesis [22]. As shown
Supplementary Figure S5, Jak1, Jak2 and Jak3 transcript
levels were not suppressed by DFX treatment.
Recently, it was reported that reactive oxygen
species (ROS) induce phosphorylation of Pyk2 in acute
monocytic leukemia cells [23]. Furthermore, we have
previously shown that DFX can inhibit ROS production
in MDS [9]. These two observations prompted us to
analyze changes in ROS production in DFX-treated MM
cells. Flow cytometric analysis using CellROX deep red
probes revealed that DFX signicantly suppressed ROS
production in MM cells (Figure 4C). Taken together, these
data suggest that DFX exerts anti-myeloma activities by
inhibiting Pyk2 phosphorylation, following a decrease in
ROS production.
Pyk2 inhibition by DFX suppresses the Wnt/β-
catenin signaling pathway
Next, we analyzed Pyk2 downstream target
molecules in order to clarify the mechanism of DFX-
induced apoptosis in MM. It has been shown that
inhibition of Pyk2 promotes β-catenin degradation via
the activation of GSK-3β [17]. In MM cells, as expected,
β-catenin protein levels were reduced by DFX treatment,
accompanied by decreased levels of phosphorylated-
GSK-3β (inactive form), resulting in the degradation of
β-catenin (Figure 5A).
Wnt signaling activities were also examined by
qRT-PCR and immunoblotting. Treatment with DFX
reduced the mRNA levels of Axin2 and c-Myc, which
are target genes of Wnt; in particular, Axin2 is a robust
Figure 2: Iron chelation with DFX inhibits the proliferation of cultured MM cells. A, C. MM cell lines (A) and primary
MM cells isolated from two MM patients (C) were cultured with DFX (0-50 μM) for 48 hours. Cell proliferation was assessed in triplicate
cultures by WST-1 assays. Data are the mean ± SD of triplicate measurements (n = 3; P < 0.05 for all cell lines). IC
50
values are for growth
inhibition by DFX. (B) Correlation between the FeRhoNox-1-positive area and IC
50
values for DFX in four MM cell lines. The relationship
was tested by two-tailed Spearman correlation (r
s
). D. RPMI8226 cells were cultured for 48 hours in HS-5-coated, primary BMSC-coated,
or uncoated wells with DFX at its IC
50
concentration. Primary BMSCs were derived from two MM patients (#1, #2). Cell proliferation was
assessed by BrdU assays. Data are the mean of triplicate measurements. Error bars represent the SD. * P < 0.01.
Oncotarget64334
www.impactjournals.com/oncotarget
and specic Wnt target gene (Figure 5B). In addition,
immunoblotting showed that c-Myc and cyclin D1, which
function downstream of β-catenin in the Wnt pathway,
were signicantly decreased by DFX treatment (Figure
5C). Collectively, these results supported the hypothesis
that DFX inhibits activation of Wnt/β-catenin signaling.
Lastly, to pinpoint the function of Pyk2 in MM cells,
we performed in vitro Pyk2 loss-of-function analysis.
Similar to DFX treatment, Pyk2 knockdown with siRNA
in MM cells led to a decrease in β-catenin and c-Myc
protein levels (Figure 5D).
Anti-tumor activity of DFX in mouse xenograft
models
To test the therapeutic potential of DFX, we
evaluated its ability to suppress tumor growth in vivo
by using the subcutaneous RPMI8226 murine xenograft
Figure 3: DFX induces apoptosis in MM cells. A. Four MM cell lines were treated with DFX for 0, 24 and 48 hours. Whole-cell
lysates were subjected to immunoblotting using anti-Caspase 9, anti-Caspase 3, anti-PARP, and anti-β-actin antibodies. FL; full length,
CL; cleaved form. B. MM1S and RPMI8226 cells were pre-incubated with or without 100 mM FeCl
3
for 2 hours. Cells were washed three
times with phosphate-buffered saline and then treated with the IC
50
concentration of DFX for 48 hours. Apoptotic cells were analyzed
by ow cytometry using Annexin V/7-AAD staining. Data are the mean of triplicate measurements. Error bars represent the SD. * P <
0.01. C. MM1S, RPMI8226 and primary MM (#3) cells were cultured for 48 hours in BMSC (HS-5)-coated or uncoated wells with DFX.
Apoptotic cells were analyzed by ow cytometry using Annexin V/7-AAD staining. Data are the mean of triplicate measurements. Error
bars represent the SD. * P < 0.01.