UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)
UvA-DARE (Digital Academic Repository)
TRPM7: Ca2+ signaling, actomyosin remodeling and metastasis
Visser, J.P.D.
Publication date
2014
Link to publication
Citation for published version (APA):
Visser, J. P. D. (2014). TRPM7: Ca 2+ signaling, actomyosin remodeling and metastasis.
Boxpress.
General rights
It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s)
and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open
content license (like Creative Commons).
Disclaimer/Complaints regulations
If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please
let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material
inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter
to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You
will be contacted as soon as possible.
81
-Chapter 3
Daan Visser
*
, Jeroen Middelbeek*
, Linda Henneman*
, Edwin Lasonder, Adrian Tsang, Yi-Jing Song, Jan Koster, Rogier Versteeg, Kristopher Clark‡, Kees Jalink‡ and Frank van Leeuwen‡The TRPM7-interactome defines a cytoskeletal
complex involved in neuroblastoma metastasis
*
and ‡C
ontributed equallyManuscript in preparation
Abstract
Metastases form when tumor cells escape the primary tumor and migrate across tissue
boundaries to colonize distant sites. Tumor cell dissemination requires extensive
reorganization of the actomyosin cytoskeleton to drive the coordinated turnover of cell
adhesions and protrusions. We previously demonstrated that overexpression of the
channel-kinase TRPM7 in the mouse neuroblastoma cell line N1E-115 induces cell spreading and
strongly increases cell adhesion, accompanied by formation of invadosome-type adhesions.
Here, we extend these observations by showing that TRPM7 expression also strongly affects
neuroblastoma cell migration in vitro, as well as invasion to liver and bone marrow in vivo.
We further used a proteomics approach to show that TRPM7 is part of a large complex of
proteins, many of which function in actomyosin remodeling and invadosome formation.
Expression of a subset of these TRPM7 interacting proteins strongly correlates with disease
outcome in a cohort of 88 neuroblastoma patients as well as in an independent validation
cohort. Thus, TRPM7 is part of a large cytoskeletal complex that controls the metastatic
potential of tumor cells by regulating the actomyosin cytoskeleton and cell adhesion
formation.
Keywords
TRPM7; actomyosin remodeling; adhesion; migration; metastasis;
82
-Introduction
Metastatic spread of cancer cells is the hallmark of malignant cancer and the main cause of
cancer-related mortality (Chaffer and Weinberg, 2011). Despite a conceptual understanding
of the steps that comprise the metastatic cascade (i.e. dissociation, migration towards
and transport through the vasculature, extravasation and colonization), the biological and
molecular mechanisms that underlie these different steps remain elusive. Metastasizing
tumor cells have to spatially and temporally modulate cell-cell and cell-matrix interactions,
adjust cell shape and control cellular contractility (Wirtz et al., 2011). All these processes
depend on the ability of cells to respond to mechanical forces and require dynamic
remodeling of the actomyosin cytoskeleton and the extracellular matrix (ECM).
The ECM and actomyosin cytoskeleton are typically connected via specialized integrin
containing cell adhesions, such as focal adhesions (FAs) and invadosome-type adhesions.
These structures are sensitive to mechanical forces from both inside and outside the cell
and transmit these reciprocally across the plasma membrane, which affects ECM properties
as well as cytoskeletal dynamics (Geiger et al., 2009). A proper balance between external
mechanical forces and internal cytoskeletal tension is critical to cell behavior and disruptions
to this tensional homeostasis contribute to tumor progression and metastasis (DuFort et al.,
2011).
Besides integrins, cell adhesions also contain stretch-activated ion channels, by which
cells can mechanically probe their environment (Munevar et al., 2004). Amongst these are
cation channels of the Transient Receptor Potential (TRP) superfamily, which have various
functions in sensory transduction (Clapham, 2003; Venkatachalam and Montell, 2007). In
addition, several family members, including TRPM7, are well known to initiate cytoskeletal
remodeling (reviewed in (Clark et al., 2008c; Kuipers et al., 2012)). TRPM7 consists of a
channel with a C-terminal a-kinase domain (Middelbeek et al., 2010; Nadler et al., 2001;
Runnels et al., 2001; Ryazanova et al., 2001) and it has since its original discovery been
ascribed a variety of functions, including the regulation of cellular Mg
2+homeostasis (Schmitz
et al., 2003), cell proliferation (Abed and Moreau, 2007; Guilbert et al., 2009; Hanano et al.,
2004; Yee et al., 2011) and the control of anoxic neuronal cell death through Ca
2+overload
(Aarts et al., 2003).
More recently, a role for TRPM7 in controlling the actomyosin cytoskeleton, cell
adhesion and migration has become subject of intense investigation (Abed and Moreau,
2009; Callera et al., 2009; Chen et al., 2010; Clark et al., 2006; Gao et al., 2011; Liu et al.,
2011; Su et al., 2006; Su et al., 2011; Wei et al., 2009). Su and colleagues demonstrated
that TRPM7 modifies adhesion and migration in HEK-293 cells via the calcium-dependent
protease m-calpain (Su et al., 2006). Simultaneously, using N1E-115 mouse neuroblastoma
83
-3
and regulates cytoskeletal organization, most likely through myosin II heavy chain
phosphorylation. Furthermore, TRPM7 activation induced the formation of
invadosome-type adhesions and it localized to these structures together with its substrate myosin
IIA (Clark et al., 2006; Clark et al., 2008b). This lead us to postulate that TRPM7 may set
up a local Ca
2+microenvironment to promote activation of its α-kinase domain which
facilitates actomyosin relaxation and ultimately leads to cell flattening and modification
of cell adhesion architecture (Clark et al., 2008b). In agreement with this model, Wei
and colleagues presented evidence that TRPM7-dependent Ca
2+microdomains (‘calcium
flickers’) at the leading lamella of lung fibroblasts steer directional migration (Wei et al.,
2009). Taken together, these reports firmly establish TRPM7 as a pivotal player in the
regulation of actomyosin remodeling and adhesion dynamics.
Cell migration depends highly on actomyosin remodeling and indeed TRPM7 has
been found to modulate tumor cell migration (Abed and Moreau, 2009; Chen et al., 2010;
Gao et al., 2011; Middelbeek et al., 2012), as well as metastasis (Middelbeek et al., 2012).
However, how TRPM7 contributes to migration and the metastatic spread of tumor cells in
vivo remains elusive and studies are significantly hampered by the lack of insight into its
interactions with the actomyosin cytoskeleton.
In this study, we set out to further investigate the role of TRPM7 as regulator of
actomyosin remodeling in cell adhesion and migration. Furthermore we explore its
potential involvement in tumor cell dissemination. Using previously characterized N1E-15/
TRPM7 cells, we first explore how TRPM7-mediated actomyosin remodeling contributes to
cell adhesion and the acquisition of metastatic properties, in vitro and in vivo. Next, using
mass spectrometry and gene expression analyses in two neuroblastoma patient cohorts,
we provide evidence that indicate that TRPM7 may function as part of a large, primarily
cytoskeletal, protein complex to regulate cancer progression and metastasis.
Results
TRPM7 regulates neuroblastoma cell-matrix interactions and enhances cell motility
In line with our previous observations (Clark et al., 2006), stable expression of TRPM7-HA
(
~2 fold over endogenous levels) conveys an adhesive phenotype onto otherwise contractile
and poorly adhesive mouse N1E-115 neuroblastoma cells [Fig. 1; Fig. S1A], without affecting
cell viability and proliferation [Fig. S1B]. The striking increase in cell adhesive properties may
be appreciated from both the fraction of cells that spread in time-lapse microscopy series
[Fig. 1A] as well as from the increased fraction of the basal membrane area of each cell
that adheres directly to the underlying substrate, as quantified by Total Internal Reflection
Fluorescence (TIRF) images [Fig. 1B].
84
-We next set out to further characterize adhesive structures in these cells.
Invadosome-type adhesions in the cell center (underneath the nucleus) represented the most commonly
found cell adhesions in both empty-vector control cells and N1E-115/TRPM7 cells. These
stained positive for invadosome markers such as cortactin and Tks4 [Fig. 2A; Fig. S2]. In
addition, as reported previously (Clark et al., 2006), TRPM7 overexpression promoted the
formation of large peripheral invadosome-type adhesions in a subfraction of N1E-115/
TRPM7 cells, an effect that could be augmented by activation of TRPM7 by stimuli such
as bradykinin [Fig. 2B]. Furthermore, TRPM7-mediated peripheral invadosomes formed in
response to topological cues [Fig. 2C].
Invadosomes are thought to increase the metastatic properties of cells by favoring
adhesion, guiding ECM remodeling and assisting in directional migration and transmigration
B
EV TRPM7
Merge
DIC TIRF DIC TIRF
EV TRPM7 0 20 40 60 80 Ad he si ve contact (% ) * * A 0 20 40 60 80 0 10 20 30 40 50 60 EV TRPM7 Time (min) Sp re ad c el ls (% ) t = 1 t = 5 t = 10 t = 20 EV TRPM7
Figure 1 | TRPM7 overexpression modifies neuroblastoma cell-matrix interactions. (A) Top: Quantification (mean ± SEM) of cell spreading in time shows that TRPM7 overexpression enhances cell spreading of N1E-115 cells (EV = 108 cells; TRPM7 = 100 cells). Bottom: Representative images of cell spreading at indicated time points of control and TRPM7 overexpressing cells. (B) Representative DIC and TIRF images of adherent N1E-115/EV and N1E-115/TRPM7 cells loaded with CellTrackerGreen (left) illustrate that N1E-115/TRPM7 cells make more contact with the underlying substrate. The adhesive surface of cells was quantified by normalizing TIRF area (i.e. cell surface area in contact with substrate) to the DIC area (i.e. total cell surface area). Data are mean ± SEM of 2 independent experiments performed in duplo (EV = 114 cells; TRPM7 = 96 cells, ** is p<0.01, one-tailed unpaired t-test).
85
-3
(Linder et al., 2011). Indeed, invadosomes in N1E-115/EV and N1E-115/TRPM7 cells are
proteolytically active and readily degrade fluorescently-labeled gelatin, as assessed by
confocal microscopy [Fig. 2D]. Note, however, that since central invadosomes generally
greatly outnumber their peripheral counterparts, the former dominate matrix degradation.
Consequently, we did not observe a noteworthy increase in degradative capacity upon
TRPM7 overexpression [data not shown].
Vinculin F-actin Merge Vinculin Merge F-actin
B
F-actin Vinculin Reflection
F-actin/Vinculin Merge
F-actin Vinculin Reflection
F-actin/Vinculin Merge
C
D
F-actin Gelatin-OG Merge EV F-actin Gelatin-OG Merge TRPM7A
Lifeact-GFP RFP-cortactin Merge RFP-cortactin Lifeact-GFP Merge EV Lifeact-GFP RFP-cortactin Merge Lifeact-GFP RFP-cortactin Merge TRPM7Figure 2 | TRPM7 overexpression alters cell adhesion formation, but does not affect matrix degradation. (A) Small-sized dot-like cell adhesions were revealed by RFP-cortactin (red) and Lifeact-GFP (green) and formed centrally at the ventral membrane of both N1E-115/EV and N1E-115/TRPM7 cells, resembling invadosomes. (B) Typical example of larger peripheral invadosomes that formed in N1E-115/TRPM7 cells stimulated with bradykinin (20 nm, 15 minutes), but not in N1E-115/EV cells.#1 (C) N1E-115/TRPM7 cells seeded on patterned
substrates form peripheral invadosomes that concentrate and align in response to topological cues (merge). Patterned substrates were generated by sanding of coverslips, which introduces superficial scratches (reflection).#1 Typical result of > 20 experiments. (D) Matrix degrading
activity of invadosomes in N1E-115/EV (left) and N1E-115/TRPM7 (right) cells plated on cross-linked Oregon Green-labeled gelatin (Gelatin-OG). Focal matrix degradation by invadosomes is apparent from dark spots in Gelatin-OG. Invadosomes are visualized by staining the F-actin-dense core with phalloidin-Alexa568 (red in B-D) and the surrounding ring by vinculin immuno-labeling (green in B + C). Dashed boxes indicate zoomed regions depicted in the second line in all subpanels. Scalebar = 10 µm. #1 Please note that the outcome of these
responses are variable between experiments (% of cells with peripheral invadosomes is 2-5% in unstimulated N1E-115/TRPM7 cells and 10-60% upon BK-stimulation or plating on a patterned substrate).
86
-As cytoskeletal organization and adhesion dynamics guide malignant transformation
and subsequent migration and metastasis of tumor cells (Paszek et al., 2005; Samuel and
Olson, 2011), we compared migratory properties of N1E-115/TRPM7 and control cells by
assaying migration towards a serum gradient in Boyden chambers. Whereas control cells
remained immobile during 48 hours, TRPM7 overexpression dramatically increased the
ability of neuroblastoma cells to cross the transwell membrane (EV = 4.1 ± 2.2 cells; TRPM7
= 336 ± 106.7 cells, p = 0.027, n = 6) [Fig. 3A]. A similar effect was found using gelatin-coated
filters [Fig. 3B]. Taken together, our results indicate that TRPM7 may facilitate cell migration
through its control over cell-matrix interactions.
TRPM7 promotes tumor metastasis in mice
We next assessed a role for TRPM7 in malignant transformation in vivo using an experimental
metastasis model. Following intravenous injection of luciferase expressing N1E-115/EV
and N1E-115/TRPM7 cells in Rag2
-/-Il2rg
-/-immunodeficient mice, we used non-invasive
bioluminescence imaging to monitor tumor cell dissemination and growth. Bioluminescence
signals were observed from the start-point at day 7 post injection and progressively
increased [Fig. 4A & B], showing that injected cells successfully survived, proliferated and
formed metastases. In good agreement with earlier reports on metastasis of neuroblastoma
cells in mice (Bogenmann, 1996), bioluminescence originated predominantly from the
abdominal region. Strikingly, the abdominal signal in N1E-115/TRPM7 injected mice was
much higher at all time points (day 7: EV = 5.41 x 10
4± 9.92 x 10
3photons/s, n = 9; TRPM7 =
9.75 x 10
5± 1.63 x 10
5photons/s, n = 9) [Fig. 4A & B], which indicates that increased TRPM7
expression indeed enhances metastatic spread of N1E-115 cells in vivo. Note further that the
progressive increase in bioluminescence was comparable in both groups, showing that in
vivo the proliferation rates of the two cell lines are identical as well [Fig. 4B]. Thus, increased
TRPM7 expression levels markedly enhance (N1E-115) neuroblastoma metastasis formation
A EV TRPM7 EV TRPM7 0 50 100 R el at iv e m ig ra tio n * B Gelatin-coated 0 25 50 R el at iv e m ig ra tio n * EV TRPM7
Figure 3 | TRPM7 promotes N1E-115 cell migration. Transwell migration assays reveal strongly improved migration of N1E-115/TRPM7 cells as compared to N1E-115/EV cells on both uncoated (A) and gelatin-coated (B) filters (n = 6 and n = 3 independent experiments performed in triplicate, respectively). Equal numbers of cells were seeded on transwell filters and allowed to migrate towards a serum gradient for 48 hours. Data are mean ± SEM. * is p<0.05, two-tailed unpaired t-test with Welch correction).
87
-3
in vivo, consistent with our previous observations in breast cancer cells (Middelbeek et al.,
2012).
TRPM7 promotes metastasis to liver and bone marrow: Histopathological characterization.
As the proliferation rate of N1E-115/TRPM7 cells did not differ significantly from control
cells both in vitro and in vivo, the increased abdominal signals of mice injected with these
cells suggest that dissemination is more widespread. Histopathological analysis [Fig. 4C;
Fig. S3A] showed that the dissemination pattern is similar in both groups, with metastases
predominantly present in the liver [Fig. S3A]. However, image analysis of liver paraffin
sections stained with H&E or the neuronal marker P75 [Fig. S3B] demonstrated that much
more tumors formed in mice injected with N1E-115/TRPM7 cells (EV = 9.5 ± 3.3; TRPM7 =
79.7 ± 13.4; p<0.01, n = 9) [Fig. 4C & D, left panel]. Importantly, TRPM7 overexpression had
no effect on tumor size, consistent with our in vitro observations (EV = 0.35 ± 0.13 mm
2,
n = 118 tumors; TRPM7 = 0.26 ± 0.03 mm
2, n = 786 tumors; p=0.23) [Fig. 2D, right panel;
Fig. S3C]. Therefore, we conclude that TRPM7 enhances the metastatic potential, but not
proliferation rate, of N1E-115 neuroblastoma cells in vivo.
In addition to the liver, metastases in bone and bone marrow are commonly observed
in neuroblastoma patients (DuBois et al., 1999). We therefore isolated bone marrow from
each limb and analyzed luciferase activity [Fig. 4E]. Limbs were considered positive when
bioluminescence was five times above background. Strikingly, only one limb of a single
N1E-115/EV injected mouse met this criterion, whereas all mice injected with N1E-115/TRPM7
cells scored positive with an average of 3 infested legs per mouse (EV = 1 limb in 1 out of
9 mice; TRPM7 = 9 out of 9 mice affected with 27 out of 36 limbs scoring positive, p<0.01)
[Fig. 4E]. When N1E-115/EV and N1E-115/TRPM7 cells from bone marrow isolations were
put to culture, they had retained viability as well as their relative TRPM7 expression levels
and typical morphologies [Fig. S3D]. These data show that our xenograft mouse model
closely resembles human neuroblastoma pathogenesis (DuBois et al., 1999). We conclude
that elevated TRPM7 expression renders N1E-115 neuroblastoma cells much more effective
in establishing metastases, without noticeably affecting the dissemination pattern.
TRPM7 associates with the dynamic actomyosin cytoskeleton
TRP channels often function in large multimeric protein complexes (Kuipers et al., 2012). To
gain insight into how TRPM7 may modulate the actomyosin cytoskeleton and cell adhesion
to facilitate migration and neuroblastoma metastasis, we next set out to identify TRPM7
interacting proteins using a proteomic approach [Fig. 5A]. The TRPM7 complex was purified
by immunoprecipitation from mouse N1E-115/TRPM7 neuroblastoma cells, and associated
proteins were resolved by SDS-PAGE. Silver staining of the gels revealed several proteins that
were specifically present in the TRPM7 fraction [Fig. 5B]. All proteins present in the control
88
-*
D E V TR P M 7 0 0.1 0.2 0.3 0.4 0.5 Tu m or s iz e (m m 2) 0 25 50 75 100 N um be r o f l iv er tu m or s E V TR P M 7**
E P os iti ve m ice (%) E V TR P M 7 0 25 50 75 100 EV TR PM 7 25 50 75 P os iti ve li m bs (% ) 0**
A N1E-115/EV N1E-115/TRPM7 5.0 4.0 3.0 2.0 1.0 x 10 7 p/s Day 7 Day 20*
*
3 6 9 12 15 18 21 104 105 106 107 108 109 TRPM7EVDays after injection
Fl ux (p /s ) B C N1E-115/EV N1E-115/TRPM7 1000 µm 1000 µm 100 µm 100 µm 1000 µm 1000 µm 100 µm 100 µm
Figure 4 | TRPM7 increases the metastatic potential of N1E-115 cells and promotes metastasis to liver and bone marrow. (A) Metastasis formation was followed over time by non-invasive bioluminescence imaging in an experimental metastasis assay. Representative bioluminescent images of mice, 7 (top) and 20 (bottom) days after intravenous injections with empty-vector control cells or TRPM7 overexpressing cells illustrate clearly higher signals in the mice injected with N1E-115/TRPM7 cells (n = 9 mice in each group). Photon fluxes are to the same scale (photons/s). Asterisk indicates an example of bioluminescence observed in a limb. (B) Quantification of bioluminescence measured in the abdominal region between day 7 and day 21 post-injection shows higher signals at all time points for mice injected with N1E-115/TRPM7 cells as compared to mice injected with N1E-115/EV cells. The similar trend of increasing signal over time of both graphs indicates that the proliferation rate of N1E-115/EV and N1E-115/TRPM7 cells in vivo are comparable. (C) Representative Haemotoxylin and Eosin-stained liver sections, collected 21 days after injection with 115/EV (left) and N1E-115/TRPM7 (right) cells. Prominent tumors in liver sections illustrate that both empty-vector control cells and TRPM7 overexpressing cells metastasize to the liver. (D) Left: N1E-115/TRPM7 cells form 7-8 times more liver tumors as compared to N1E-115/EV cells. Right: Mean size of liver tumors in mice (continued on next page)
89
-3
and TRPM7 immunoprecipitations were identified by nano liquid chromatography tandem
mass spectrometry (LC-MS/MS) and proteins were considered enriched when the
iBAQ-ratio between the TRPM7 fraction and control fraction was greater than 10. This analysis
led to the identification of 64 proteins that were specifically enriched in TRPM7 precipitates
[Table 1]. Subsequently, proteins were classified according to the Gene Ontology (GO)
annotation for molecular function, cell component and biological process. A distribution
of GO terms between the control and TRPM7 fractions showed that proteins involved in
the organization and biogenesis of the actomyosin cytoskeleton were specifically enriched
in TRPM7 immunoprecipitates [Fig. 5D]. This set of proteins includes conventional myosin
II and non-conventional myosins (myosins I, V and VI) as well as proteins regulating actin
dynamics such as the Arp2/3 complex, F-actin capping proteins, drebrin, tropomyosins,
tropomodulins, gelsolin and cofilin [Table 1]. Moreover, a group of structural proteins
involved in regulating cell adhesion and migration, including
α-actinin4 and plectin, was
also present [Table 1]. We confirmed by immunoprecipitation and Western blotting that
these interactions were specific, as abundant actin-binding proteins such as
α-actinin1,
talin, cortactin and vinculin, were not detected in TRPM7 immunoprecipitations [Fig. 5C].
Strikingly, this ‘TRPM7 interactome’ resembles cytoskeletal complexes that mediate
the formation of like filopodia, neuronal growth cones, dendritic spines and podocytes
(Kuipers et al., 2012; Mattila and Lappalainen, 2008), suggesting that TRPM7 is part of a
cytoskeletal machinery responsible for the dynamic formation of such cellular protrusions.
We previously showed that TRPM7 localized to peripheral invadosomes, where it may
contribute to the assembly of such structures (Clark et al., 2006). Consistent with these
observations, a significant number of proteins we identified are known constituents of
either the invadosome core, such as
β-actin, F-actin capping proteins, Arp2/3, gelsolin,
cofilin, or the ring nonmuscle myosin IIA and tropomyosin (Linder et al., 2011) [Table 1:
column “Invadosome”]. Intriguingly, many other components of the TRPM7 complex that
are implicated in the formation of cellular protrusions, appear to localize to invadosomes as
well [Fig. 5E; in Table 1 marked as ‘new’]. These include drebrin, myosin IIB and C, myosin
V, α-actinin4 and SIPA1-L1/SPAR1. Taken together, these results suggest that TRPM7 may
control actomyosin remodeling and the formation of cell adhesions like invadosomes as part
of a cytoskeletal complex.
injected with N1E-115/EV is not different from mice injected with N1E-115/TRPM7 cells, indicating similar proliferation rate in vivo (** p<0.01, two-tailed unpaired t-test with Welch correction). (E) Quantification of the percentage of mice (left) and limbs (right) with bone marrow metastases in N1E-115/EV and N1E-115/TRPM7-injected mice (** p<0.01, two-tailed unpaired t-test with Welch correction). Data shown in B, D and E are mean ± SEM of n = 9 mice per group. Data shown are representative examples of two independent experiments.
A
IP from N1E-115/EV and N1E-115/TRPM7 cells using anti-HA antibodies (12CA5)
1D SDS-PAGE (6-12% PAA) Silver stain gel Cut gel into 24 pieces In-gel digest with trypsin
nanoLC-MS/MS Reverse database search
to set false positve rate Identify enriched proteins:
iBAQ-ratio > 10 Bioinformatic analysis → → → → → → → → D G O : 0 03 00 36 G O : 0 030029 G O : 0 04 44 30 G O : 0 01 56 29 60 50 40 30 20 10 0 Molecular Function Cell Component Biological Process G O : 0 00 80 92 G O : 0 00 3779
TRPM7 interactome proteins (%) Actin filament-based process Actin cytoskeleton organization
C cortactin talin vinculin α-actinin1 α-actinin4 myosin IIA myosin IIB myosin IIC drebrin p116Rip TL EV TRPM7 IP EV TRPM7 E myosin V actin merge myosin IIC actin merge myosin IIB actin merge myosin IIA actin merge SIPA1-L1 actin merge drebrin actin merge p116Rip actin merge α-actinin4 actin merge B 250 150 100 75 50 37 25 20 15 kD EV TRPM7
Cytoskeletal protein binding Actin binding Cytoskeletal part Actin cytoskeleton
immunoprecipitation using anti-HA monoclonal antibodies (12CA5) from empty-vector control cells and N1E-115/ TRPM7 cells. Proteins were separated by SDS-PAGE on 6% (top) and 12% (bottom) gels and subjected to silver staining. Proteins that co-immunoprecipitated with TRPM7 are indicated. (C) The interactions between TRPM7 and the cytoskeleton are specific. Protein complexes were immunoprecipitated from empty-vector control cells and N1E-115/TRPM7 cells using anti-HA antibodies. Proteins present in immunoprecipitations (IP, left) and total lysates (TL, right) were detected by Western blot. TRPM7 interactors that were identified as novel invadosome components (in E) are underlined. (D) A distribution of GO terms for the TRPM7 fraction shows that proteins involved in the organization and biogenesis of the actomyosin cytoskeleton were specifically enriched in TRPM7 immunoprecipitations. GO analysis was performed on the set of proteins enriched in TRPM7 IP (iBAQ > 10). Next, GO enrichment relative to the mouse whole genome was determined. GO terms in the categories ‘biological process’, ‘molecular function’ and ‘cell component’ that were most significantly enriched in the TRPM7 IP are indicated. The fraction of proteins within the TRPM7 IP, annotated to these terms, are presented in a bar chart. (E) Representative confocal images show the localization of selected TRPM7 interactome components to peripheral invadosomes in N1E-115/TRPM7 cells. Protein localization was evaluated by immunolabeling, if available (myosin IIA, B and C, drebrin, p116Rip, and α-actinin), or by expression of GFP- or myc-tagged proteins (myosin V and
SIPA1-L1). F-actin was visualized by Alexa568-phalloidin. Scale bar = 5 µm. Figure 5 | TRPM7 interacts with a cytoskeletal
complex that is involved in actomyosin remodeling and adhesion formation. (A) Flow chart for proteomic analysis of the TRPM7 complex. (B) Detection of TRPM7 associated proteins by silver staining of SDS-PAGE gels. HA-tagged TRPM7 was isolated by
Protein name Gene symbol M ol ec ul ar M ot or s Myosin-11 MYH11 17880 4629
Myosin light chain 3 MYL3 17897 4634
Myosin light chain, regulatory B-like MYL9 67268 10398
Myosin light polypeptide 6 MYL6 17904 4637
Myosin-1e MYO1E 71602 4643 Myosin-Ib MYO1B 17912 4430 Myosin-Ic MYO1C 17913 4641 Myosin VI MYO6 17920 4646 Sc af fo ld / St ru ct ur al p ro te in Alpha-actinin-4 Desmoplakin
LIM and calponin homology domains-containing protein 1 LIM domain and actin-binding protein 1 Myosin phosphatase Rho-interacting protein
Neurabin-2 Plectin ACTN4 DSP LIMCH1 LIMA1 MPRIP PPP1R9B PLEC 60595 109620 77569 65970 26936 217124 18810 81 1832 22998 51474 23164 84687 5339 O th er Annexin A2 Calmodulin Fructose-bisphosphate aldolase A Junction plakoglobin Kinase D-interacting substrate of 220 kDa
Pericentriolar material 1 protein Tax1-binding protein 1 homolog
ANXA2 CALM1 ALDOA JUP KIDINS220 PCM1 TAX1BP1 12306 12313 11674 16480 77480 18536 52440 302 801 226 3728 57498 5108 8887 En zy m at ic a ct iv
ity Protein phosphatase Slingshot homolog 2 Signal-induced proliferation-associated 1-like protein 1 Transient receptor potential cation channel subfamily M member 7
SSH2 SIPA1L1 TRPM7 237860 217692 58800 85464 26037 54822 N on -c yt os ke le to n-re la te d
116 kDa U5 small nuclear ribonucleoprotein component 14-3-3 protein sigma
AT-rich interactive domain-containing protein 1A ELAV-like protein 3
Histone H3.2 Importin-7 Importin-9 Prelamin-A/C
Probable ATP-dependent RNA helicase DDX5 Protein transport protein Sec16B Sarcoplasmic/endoplasmic reticulum calcium ATPase 2
Transcription intermediary factor 1-beta Ubiquitin carboxyl-terminal hydrolase 15 Voltage-dependent anion-selective channel protein 1
EFTUD2 SFN ARID1A ELAVL3 HIST1H3E IPO7 IPO9 LMNA DDX5 SEC16B ATP2a2 TRIM28 USP15 VDAC1 20624 55948 93760 15571 319151 233726 226432 16905 13207 89867 11938 21849 14479 22333 9343 2810 8289 1995 8353 10527 55705 4000 1655 89866 488 10155 9958 7416 Ac tin D yn am ic s Actin, cytoplasmic 1 Actin, cytoplasmic 2 Actin-related protein 2 Actin-related protein 2/3 complex subunit 1A Actin-related protein 2/3 complex subunit 1B Actin-related protein 2/3 complex subunit 2 Actin-related protein 2/3 complex subunit 3 Actin-related protein 2/3 complex subunit 4 Actin-related protein 2/3 complex subunit 5 Actin-related protein 2/3 complex subunit 5-like protein
Cofilin-1 Drebrin
F-actin-capping protein subunit alpha-1 F-actin-capping protein subunit alpha-2 F-actin-capping protein subunit beta
Gelsolin Tropomodulin-2 Tropomodulin-3 Tropomyosin alpha-1 chain Tropomyosin alpha-3 chain, isoform 2
Tropomyosin alpha-4 chain
ACTB ACTG1 ACTR2 ARPC1A ARPC1B ARPC2 ARPC3 ARPC4 ARPC5 ARPC5L CFL1 DBN1 CAPZA1 CAPZA2 CAPZB GSN TMOD2 TMOD3 TPM1 TPM3 TPM4 11461 11465 66713 56443 11867 76709 56378 68089 67771 74192 12631 56320 12340 12343 12345 227753 50876 50875 22003 59069 326618 60 71 10097 10552 10095 10109 10094 10093 10092 81873 1072 1627 829 830 832 2934 29767 29766 7168 7170 7171 Mouse Human
Category Gene ID Invadosome
Myosin-9 (NMHC IIA) MYH9 17886 4627
new
new
Myosin-14 (NMHC IIC) MYH14 71960 79784 new
Myosin-10 (NMHC IIB) MYH10 77579 4628 new
Myosin Va MYO5A 17918 4644 new
new CA + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + 1 1 1 + + + + + + + + + + + + + + + + + +
Table 1 | The TRPM7 interactome defines a largely cytoskeletal complex that contains proteins involved in actomyosin remodeling. All 64 components with an iBAQ-ratio score >10 identified in the mass spec analysis are listed within categories of biological function. Components are represented by protein name, gene symbol and gene ID (for mouse and human). Known (+) and here-identified (new) invadosome components are indicated, as well as established associations with cancer progression and metastasis (CA) as determined by literature search; 1 = associated with neuroblastoma progression.
92
-Components of the ‘TRPM7 interactome’ correlate with human neuroblastoma metastasis
An extensive literature survey [see Material and Methods] illustrated that expression of
many of the TRPM7 interactome components (~54%) have been previously associated with
outcome in various cancers, although only few in neuroblastoma (~5%) [Table 1: column
“CA”]. We therefore tested to what extent components of the TRPM7 interactome may
predict outcome in human neuroblastoma using a large neuroblastoma gene expression
dataset (Molenaar et al., 2012). This set (NB-88)
contains expression profiles of 88 primary
tumor biopsies
, along with clinical information on overall survival (OS), relapse-free survival
(RFS) and metastasis to bone and bone marrow. Probes for every validated component of
the TRPM7 interactome, with the exception of MYH14, were present on this microarray. For
TRPM7 our analysis revealed no significant correlation (not shown), but expression levels
of a substantial number of interactome components correlated significantly with overall
survival
(12 genes), relapse-free survival (11 genes) or both (9 genes)
after Bonferroni
correction for multiple testing [
Fig. 6A & B, columns OS and RFS].
We validated these positive hits
in a second independent cohort of 251 neuroblastoma
patients (Oberthuer et al., 2006). Probes for seven of the 14 genes were present in the
Oberthuer set and six of these (i.e. MYO5A, TMOD2, TRIM28, IPO7, TAX1BP1 and ALDOA)
correlated significantly with OS, RFS or both in this data set [Fig. 6B; Table S1]. Moreover,
four of the six validated genes (i.e. MYO5A, TMOD2, TRIM28 and TAX1BP1) correlate with
metastasis to bone or bone marrow in the NB-88 cohort [Table S2], suggesting that these
genes and their products may be of general significance for neuroblastoma progression.
Tax1-binding protein 1 (TAX1BP1) inhibits tumor necrosis factor (TNF) alpha-induced
apoptosis and nuclear factor kappa B (NFκB) activation as part of the A20 ubiquitin-editing
complex (De Valck et al., 1999; Shembade et al., 2007). With transcriptional repression as
its primary function (Moosmann et al., 1996), transcription intermediary factor 1-beta
(TRIM28) directly interacts with the E3 ubiquitin ligase Mdm2 to promote ubiquitylation
and degradation of the tumor suppressor p53 (Wang et al., 2005a). As in our cohorts,
TRIM28 upregulation is associated with poor survival and peritoneal dissemination in
gastric cancer patients (Yokoe et al., 2010). Neuron-specific tropomodulin-2 (TMOD2) is an
actin-regulatory protein with capping activity by which it regulates actin polymerization and
neurite formation (Fath et al., 2011). It was also part of a 55-gene expression signature that
identified subgroups within clinically indistinguishable high-risk metastatic neuroblastomas
(Asgharzadeh et al., 2006). Finally, the actin-based motor protein myosin Va (MYO5A) is
involved in vesicle and organelle transport (Desnos et al., 2007), filopodia and neurite
extension (Eppinga et al., 2008; Wang et al., 1996) and has been recently linked to metastasis
in colorectal cancer (Lan et al., 2010). If and how the interactions between these proteins
and TRPM7 affect adhesion, migration and metastasis remains to be elucidated.
93
-3
A B Cohort: Oberthuer-251 OS RFS 1.66e-11 high 3.47e-09 low 1.28e-06 low 1.29e-02 high 5.36e-20 8.73e-14 KM-arm 3.80e-08 1.14e-04 1.13e-04 1.22e-02 high 7.19e-03 n.s. n.s. n.s. low -Gene Cohort: NB-88 RFS MYO5A OS KM-arm 1.68e-10 TMOD2 7.20e-10 EFTUD2 8.91e-07 TRIM28 1.74e-05 KIDINS220 2.36e-04 ACTR2 2.36e-04 IPO7 3.48e-04 TAX1BP1 1.98e-03 low low high high low low high low 4.30e-06 2.56e-05 4.23e-04 7.27e-04 9.06e-04 2.92e-03 1.37e-02 2.58e-02 TPM3 3.64e-03 n.s. highALDOA 1.60e-02 n.s. high
CALM1 3.01e-02 2.72e-02 low
CFL1 4.11e-02 n.s. high
ANXA2 2.45e-02 low
TPM4 n.s.n.s. 2.72e-02 high high (n = 24) low (n = 64) p = 1.1e-06 0.10 0.30 0.50 0.70 0.90 Follow up in months TRIM28 0 24 48 72 96 120 144 168 192 216 high (n = 78) low (n = 10) p = 2.5e-04 0.10 0.30 0.50 0.70 0.90 Follow up in months TAX1BP1 0 24 48 72 96 120 144 168 192 216 high (n = 53) low (n = 35) p = 2.3e-11 0.10 0.30 0.50 0.70 0.90
overall survival pro
bability Follow up in months TMOD2 0 24 48 72 96 120 144 168 192 216 high (n = 57) low (n = 31) p = 2.7e-12 0.10 0.30 0.50 0.70 0.90
overall survival probability
Follow up in months MYO5A 0 24 48 72 96 120 144 168 192 216 high (n = 66) low (n = 22) p = 2.2e-05 0.10 0.30 0.50 0.70 0.90 Follow up in months ACTR2 0 24 48 72 96 120 144 168 192 216 high (n = 18) low (n = 70) p = 4.2e-08 0.10 0.30 0.50 0.70 0.90 Follow up in months EFTUD2 0 24 48 72 96 120 144 168 192 216
Figure 6 | Expression levels of TRPM7 interactome components correlate with disease outcome in two independent neuroblastoma patient cohorts. (A) Kaplan-Meier curves for overall survival of selected TRPM7 interactome genes in the NB-88 cohort. Low expression of ACTR2, MYO5A, TAX1BP1 and TMOD2, and high expression of EFTUD2 and TRIM28 strongly correlate with poor disease outcome, respectively. The number of patients evaluated in each arm is indicated between brackets. P-values are based on log-rank test and were corrected for multiple testing (Bonferroni). (B) Overview of TRPM7 interactome components that correlate with overall survival (OS) and relapse-free survival (RFS) in the NB-88 cohort (left) and their validation in the Oberthuer-251 cohort (right). Please note that the Bonferroni-corrected p-values determined in Kaplan-Meier analysis (in A) were subjected to additional False Discovery Rate (FDR) correction for multiple testing. N.s. = not significant; - = gene not present on array; KM-arm ‘low’ and ‘high’ indicate that low or high gene expression correlates with poor outcome in the Kaplan-Meier plot.
94
-We conclude that expression of a subset of genes in the TRPM7 interactome correlates
strongly with patient survival and metastasis in two independent neuroblastoma cohorts. As
most of these genes have not been previously recognized in neuroblastoma pathogenesis,
it will be interesting to further investigate their involvement in neuroblastoma progression
and metastasis formation.
Discussion
We here show that TRPM7 increases the migratory and metastatic potential of neuroblastoma
cells in vitro as well as in vivo. Our data further support the link between TRPM7 and
actomyosin remodeling, cell adhesion, migration and metastasis that has recently emerged
in literature (Abed and Moreau, 2007; Abed and Moreau, 2009; Chen et al., 2010; Clark et
al., 2006; Gao et al., 2011; Jiang et al., 2007; Kim et al., 2008; Middelbeek et al., 2012; Su et
al., 2006; Su et al., 2010; Su et al., 2011). We further provide hitherto unknown molecular
insight into how TRPM7 may bring about these changes that provides a start for future
investigations to resolve the underlying molecular mechanisms.
Given the limited number of TRPM7 interaction partners described to date, we argued
that a detailed knowledge of the molecular composition of protein complexes surrounding
TRPM7 is pivotal to our understanding of the complex biology of this bifunctional protein.
A detailed proteomic analysis of TRPM7 immune precipitations revealed that TRPM7 is part
of a large macromolecular complex involved in the dynamic regulation of the actomyosin
cytoskeleton and cell adhesion. Many of these proteins are established components of
dynamic cell protrusions such as filopodia and invadosomes, and in the current study we
extended the list of invadosome constituents with myosin IIB, myosin IIC and myosin V as
well as SIPA1-L1, drebrin, p116
Ripand α-actinin4. The interactions between TRPM7 and the
actomyosin cytoskeleton appear to be highly specific. First of all, none of the interactors
were found in significant quantities in the control sample. Moreover, actin-binding proteins
such as a-actinin1, cortactin, talin and vinculin, which are abundantly expressed in these
cells and thus likely contaminants in immunoprecipitations, were not detected either by
mass spectrometry or Western blotting. Finally, the detection of myosin IIA, IIB and IIC
heavy chains, known substrates of the TRPM7
α-kinase domain, (Clark et al., 2006; Clark
et al., 2008a) served as a positive control for our proteomics approach. The TRPC5/6
immunoprecipitate (Goel et al., 2005) also largely consists of cytoskeletal proteins, and
our findings are consistent with the notion that TRP channels commonly function in
macromolecular complexes linked to the cytoskeleton.
If and how the interactome components affect the activity and functioning of
TRPM7, and vice versa, remains to be determined. Similar to TRP channels in Drosophila
photoreceptors (Tsunoda et al., 2001), TRPM7 might function as an anchor to localize and
95
-3
maintain the integrity of the complex. On the other hand, ion influx through the TRPM7
channel could (locally) modulate the activity of calcium- and magnesium-sensitive proteins
irrespective of direct physical protein interactions. The TRPM7 interactome component
calmodulin, for example, is a calcium sensor known to activate the serine/threonine
phosphatase calcineurin (Li, 1984). Calcineurin, in turn, regulates cytoskeletal organization
and neurite extension by activation of the phosphatase slingshot and its substrate cofillin,
(Descazeaud et al., 2012; Wang et al., 2005b), both of which are present in the TRPM7
interactome. Interestingly, differential calcineurin expression in small cell lung cancer cells
was shown to affect cell adhesion, migration, invasion and bone metastasis (Ma et al., 2011).
TRPM7 expression did not significantly correlate with clinically relevant parameters in
the NB-88 cohort while the custom array of the Oberthuer-251 cohort did not contain probes
for TRPM7. In contrast, TRPM7 was recently shown to correlate with disease progression
in pancreatic cancer (Rybarczyk et al., 2012) and breast cancer (Dhennin-Duthille et al.,
2011; Middelbeek et al., 2012). Discrepancies between expression profile analyses, even
within cancer subtypes, are frequently observed and remain a matter of concern (Ahmed
and Brenton, 2005; Ein-Dor et al., 2005; Subramanian and Simon, 2010). Further study is,
therefore, required to establish the significance of TRPM7 as a prognostic factor in different
cancer subtypes. The cell biological findings in the current study are, however, in line with
and complement a report in which short-hairpin mediated TRPM7 knockdown was used to
show that TRPM7 guides breast cancer metastasis formation (Middelbeek et al., 2012).
Via the TRPM7 interactome we identified several genes with potentially general
significance in neuroblastoma pathogenesis. In particular, we present the transcriptional
regulator TRIM28 and the cytoskeleton-associated genes MYO5A, TAX1BP1 and TMOD2
to strongly correlate with patient survival in two independent neuroblastoma cohorts, as
well as with bone marrow metastasis formation. It is noteworthy that both MYO5A and
TMOD2, as well as other TRPM7 interactome components, have been shown to function
in neuritogenesis (Fath et al., 2011; Lise et al., 2009; Wang et al., 1996). A whole-genome
sequence analysis on tumors of the NB-88 cohort recently revealed that defects in genes
encoding cytoskeletal regulators of neuritogenesis or growth cone guidance strongly
correlated with aggressive, high-stage tumors, implicating neuritogenesis-regulating genes
in neuroblastoma pathogenesis (Molenaar et al., 2012). Therefore, we propose that TRPM7
functions in a cytoskeletal complex involved in cell adhesion and protrusion formation to
mediate neuroblastoma migration and metastasis.
Material and methods
Constructs and cell lines96
-2006). The recombinant proteins contain an HA-tag at the C-terminus. Luciferase cDNA was isolated from pMX-luciferase-YFP-neo and subcloned into the retroviral vector pLZRS-IRES-zeocin. Both constructs were verified by DNA sequencing. Mouse N1E-115 neuroblastoma cells were cultured in DMEM supplemented with 10% FCS and 1% penicillin-streptomycin. Stable N1E-115 cells mildly overexpressing TRPM7-HA (~3-fold endogenous), empty vector control and luciferase were generated by retroviral transduction. Cells were selected by the addition of 0.8 mg/ml G418 or 0.4 mg/ml zeocin, respectively and the selection was complete within 7 days.
Quantitative RT-PCR
Total mRNA was isolated (RNeasy Mini kit, Qiagen) and cDNA synthesized (SuperScript IIaa reverse
transcription enzyme, Invitrogen). PCR reactions were performed with Power SYBR-Green reagent (Applied Biosystems, Carlsbad, CA) and primers for mouse TRPM7 (forward: tagcctttagccactggacc, reverse: gcatcttctcctagattggcag) and the GAPDH-housekeeping gene (forward: gccaaggtcatccatgacaac, reverse: gaggggccatccacagtctt) to control for loading.
Immunoprecipitation
N1E-115 control and TRPM7-transduced cells were washed twice in ice-cold PBS and subsequently, lysed on ice for 30 min in MyoII lysis buffer (50 mM Tris pH 7.5, 300 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM DTT, 1% Triton X-100 and protease inhibitors). The lysate was cleared by centrifugation for 10 min at 16 000 g. HA-tagged proteins were immunoprecipitated by incubating the supernatant with proteinG-sepharose beads that were blocked with 0.5 %w/v BSA and precoupled with 12CA5 antibodies. The samples were incubated on an end-over-end rotor for 3 h at 4 °C. Subsequently, the immunocomplexes were washed in MyoII lysis buffer and solubilized in Laemmli buffer.
Western blotting
Proteins were separated by SDS-PAGE and transferred to nitrocellulose. Subsequently, proteins were detected by immunoblotting using anti-α-actinin1 (1:1000; Sigma), anti-cortactin (1:500 ), anti-talin (1:500; Sigma), anti-vinculin (1:2000; Sigma), anti-myosin IIA (1:1000; Sigma), anti-myosin IIB (1:1000; Sigma), anti-myosin IIC (1:1000), anti-drebrin (1:1000, Progene), anti-p116Rip (1:2000) and
anti-α-actinin4 (1:500, K. Cho) followed by HRP-conjugated secondary antibodies (1:5000; Dako). Antibody-reactive bands were visualized by treating the blots with ECL (Amersham) followed by autoradiography.
Mass Spectrometry
1. Nano LC-MS/MS measurements
Proteins were separated by SDS-PAGE on 6% and 12% polyacrylamide gels and subsequently, detected by silver staining. Gels were sliced into pieces and digested with trypsin overnight at 37 °C. Peptide mass spectrometric experiments were performed using a nano-HPLC Agilent 1100 system connected to a 7-Tesla linear quadruple ion trap-Ion Cyclotron Resonance Fourier transform (LTQFT) mass spectrometer (Thermo Fisher). Peptides were separated on 15 cm 100 µm ID PicoTip (New Objective) columns packed with 3 µm Reprosil C18 beads (Dr. Maisch GmbH) using a 45 min gradient from 10% buffer B to 35% buffer B (80% acetonitrile in 0.5% acetic acid). Peptides eluting from the column tip were electrosprayed directly into the mass spectrometer with a spray voltage of 2.1 kV. Peptide selection and fragmentation was set by the Xcalibur 1.4 data acquisition software (Thermo Electron). The mass spectrometer was operated in the data-dependent mode to sequence the four most intense ions per duty cycle. Briefly, full-scan MS spectra of intact peptides (m/z 350–1500) with an automated gain control accumulation target value of 106 ions were acquired in the Fourier transform ion cyclotron
97
-3
resonance (FT ICR) cell with a resolution of 50 000. The four most abundant ions were sequentiallyisolated and fragmented in the linear ion trap by applying collisionally induced dissociation using an accumulation target value of 20 000 (capillary temperature, 150°C; normalized collision energy, 30%). A dynamic exclusion of ions previously sequenced within 180 s was applied. All unassigned charge states were excluded from sequencing. A minimum of 500 counts was required for MS2 selection.
RAW spectrum files were converted into a Mascot generic peaklist using DTA Supercharger (http:// msquant.sourceforge.net).
2. Peptide identification by MASCOT database searches
Proteins were identified by searching peak lists containing fragmentation spectra with Mascot version 2.1 (Matrix Science) against an in-house mouse International Protein Index (IPI, version 3.16) database supplemented with the sequences of known external contaminants such as human keratins (68291 entries). A reversed protein sequence database of the mouse IPI database was also installed locally and searched to determine the false-positive rate of protein identifications. Mascot search parameters for protein identification specified an initial mass tolerance of 20 ppm for the parental peptide and 0.8 Da for fragmentation spectra and a trypsin enzyme specificity allowing up to 3 miscleaved sites. Carbamidomethylation of cysteines was specified as a fixed modification, oxidation of methionines, deamidation of glutamine or asparagine and protein N-acetylation were set as variable modifications. Internal mass calibration of measured ions was performed simualtenously with parsing Mascot search result html files using MSQuant open-source software (www.msquant.sourceforge.net) into text files. A final absolute mass tolerance for the parental peptide was determined at 6 ppm. Only multiple charged peptides with precursor masses larger than 350 were considered in the subsequent validation analysis. A minimal Mascot peptide score was set at 24. Reverse database searches revealed no false-positive identifications for proteins identified by 2 or more non-redundant peptides. More stringent criteria were required to identify proteins sequenced by one peptide with high confidence. A false-positive rate of 1.0 % was obtained for proteins identified by 1 unmodified peptide with a peptide cut-off score of 40, and a Mascot peptide delta score of 10. A total of 6267 validated peptides with an average absolute mass accuracy of 1.71 ppm were sequenced in this study that led to the identification of 298 non-redundant mouse proteins.
3. Functional annotation of identified proteins
Validated peptides of all samples combined were remapped to mouse IPI database version 3.16 to remove protein redundancy between different samples using an in-house Perl script. Priority of redundant IPI entries was given to Swiss prot, TREMBL, and REFSEQ entries respectively for maximising Gene Ontology (GO) annotation of identified proteins. External contaminating proteins (keratins, trypsin) were excluded for further analysis. Entrez gene identifiers were obtained using the Gene Conversion ID tool of DAVID (http://david.abcc.ncifcrf.gov/). Proteins were divided into two categories: 1) proteins enriched in TRPM7 immunoprecipitations and 2) contaminants, which were found in both TRPM7 and control samples. Enrichment of each protein was determined by the intensity-based absolute quantification (iBAQ) approach according to the formula: iBAQ-score of TRPM7 sample/ iBAQ-score of control sample > 10. All proteins with a ratio less than 10 were considered contaminants. The distribution of GO terms between the two groups of proteins was tested using the webbased tool FatiGO (http://babelomics.bioinfo.cipf.es). Statistical significance was determined using the Fisher exact test. All the p-values are adjusted by False Discovery Rate (FDR) approach. Differences were considered significant if p<0.05.
98
-Immunohistochemistry
N1E-115/TRPM7 cells were seeded on glass coverslips and serum starved (0.1% FCS) overnight prior to stimulation with 20nM bradykinin for 15 min. Cells were fixed with 4% paraformaldehyde in PBS for 10 minutes at room temperature and subsequently permeabilized with 0.1% triton X-100 in PBS for 3 minutes. Antibodies were used to reveal the presence of myosin IIA (1:100; BTI), anti-myosin IIB (1:100), anti-anti-myosin IIC (1:100), anti-drebrin (1:100), anti-p116Rip (1:200), anti-a-actinin4
(1:200) and alexa488-conjugated secondary antibodies (1:1000; Molecular Probes). As no antibodies were available against cortactin, Tks4, myosin V and SIPA1-L1, GFP- and Myc-tagged proteins, respectively, were introduced into N1E-115/TRPM7 cells. Transfections were carried out using Fugene HD Transfection Reagent (Roche Applied Science) according to manufacturer’s protocol. F-actin was detected using texas-red phalloidin (1:250; Molecular Probes). Cells were viewed using a Leica TCS SP5 confocal microscope. To determine cell adhesive surface in contact with underlying substrate, cells were loaded with CellTracker-Green CMFDA (Invitrogen) prior to fixation.
Cell migration
Migration assays were performed in 24-well transwell plates (8 µm pore polycarbonate membrane insert, Product #3422, Corning, NY, USA). Cells (5 x 104) were suspended in serum-free DMEM and
seeded onto the inserts. Subsequently, inserts were placed in the bottom chamber containing DMEM supplemented with 10% FBS. Transwell plates were placed in a humidified CO2 incubator at 37°C and cell migration was quantified after 48 hours. In short, non-migrated cells were first removed from the top membrane surface with cotton swabs. Next, migrated cells at the bottom of the membrane were fixed (75% MeOH, 25% HAc), stained (0.25% Coomassie Blue (w/v), 45% MeOH, 10% HAc and 45% H2O) and counted in 4 random fields under a light microscope. Migrated cells were expressed as the average number of cells per field.
Cell proliferation measurements
Cells were seeded in a 96-well plate and cell numbers were measured at various time points using the colorimetric MTS assay according to manufacturer’s instructions (Promega, Madison, WI). The experiments were performed in triplicate.
Experimental metastasis assays in mice
All animal experiments were performed in accordance with institutional guidelines and national ethical regulations. Rag2-/-Il2rg-/- immunodeficient mice, backcrossed on a Balb/c background, were used for metastasis experiments at 5-8 weeks old. N1E-115 mouse neuroblastoma cells were trypsinized and washed 3 times with PBS. Subsequently, 0.2 ml PBS containing 105 cells was injected into a tail vein.
Beetle luciferin (Promega, Fitchburg, WI, USA) was dissolved at 15 mg/ml in sterile PBS and stored at -20°C. Animals were anaesthetized with 2-3% isoflurane. Luciferin solution was injected i.p. (0.01 ml per g body weight). Light emission was measured 15 min later, using a cooled CCD camera (IVIS; Xenogen), coupled to Living Image acquisition and analysis software over an integration time of 1 min. Signal intensity was quantified as the Flux (photons/s) measured over the region of interest.
Histology
Organs and tissues were collected at day 20 after injection, fixed in EAF (ethanol-acetic acid-formol saline fixative, 40:5:10:45 v/v) and processed routinely for histology preparations. The paraffin sections were stained with Haematoxylin and Eosin (H&E). For quantitative analysis of the neoplastic lesions, liver sections (9 sections from each liver) were stained with both H&E and immunohistochemistry
99
-3
of P75 (neurotrophin receptor). The primary antibody P75 was purchased from Chemicon (AB1554,dilution 1:8000) and the secondary antibody was goat anti-rabbit from DAKO (E0432, dilution 1:1000). The lesions in each section were counted under a microscope and the data were further processed by Image J.
Bone marrow isolation
Under sterile conditions, both femur and tibiae from each mouse were excised. Muscle and the entire connective tissue were detached. The ends of the bones were removed and the shaft of the bones was flushed with DMEM supplemented with 10% FBS and 100 IU/ml penicillin-100 mg/ml streptomycin using a 25-gauge needle. Bone marrow suspension was centrifuged at 1400 rpm for 5 min and the cell pellet was resuspended in 1 ml of red cell lysisbuffer and incubated 10 min at rT. Subsequently, 10 ml of medium was added and cells were centrifuged at 1400 rpm for 5 min. Cell pellet was resuspended in 300 μl of medium and plated in a 96-well plate (100 μl per well). After 3 hours at 37°C, 10 μl beetle luciferin (Promega, Fitchburg, WI, USA; 15 mg/ml) was added per well and luminescence recorded as described for in vivo imaging.
Affymetrix DNA microarray hybridization and analysis
The Affymetrix NB tumor dataset NB-88 contains the expression profiles of 88 NB tumors with documented genetic and clinical features (GEO accession=GSE16476). This set is called ‘‘NB-88’’. Total RNA was extracted from frozen NBs containing >95% tumor cells, and Affymetrix HG U133 Plus 2.0 micro-array analysis was performed as described in Geerts et al., 2010. All analyses were performed using R2, a microarray analysis and visualization platform developed in the Department of Oncogenomics at the Academic Medical Center in the University of Amsterdam (http://r2.amc. nl; J. Koster, unpublished data).. All gene transcript levels were determined from data image files using GeneChip operating software (MAS5.0 and GCOS1.0, from Affymetrix). The probe sets selected for a gene showed the highest expression in samples containing a present call for that gene. The TranscriptView tool, embedded within the R2 platform was used to check if the probe set selected was valid.
Kaplan-Meier scanner and other statistical analyses
Kaplan-Meier scanning was performed within the R2 platform. In short, to determine the optimal value to set as cut-off in the Kaplan-Meier curve for expression of gene ‘‘X’’ in the NB-88 (see earlier), we sorted the NB tumors on the expression gene X and subsequently divided them in two groups based on the gene X expression value of every tumor. For every group separation (higher or lower than the current expression of gene X), the logrank significance was calculated using the log-rank test. The best p value obtained was used to represent the final gene expression cut-off value for gene X. To correct for multiple testing, the resulting p value was multiplied by the number of tests performed (Bonferoni correction).
Literature survey
Association of TRPM7 interactome components with cancer progression was evaluated by an online Pubmed search, using the search terms “Gene ID (or common alias) and cancer or tumor or metastasis” followed by hand screening of titles, abstracts and full text articles. Genes were considered cancer-associated when shown to correlate with patient outcome in microarray-based gene expression profiles or when differential expression in tumor cells affects tumorigenesis in mouse xenograft experiments.
100
-Acknowledgements
We thank D. Clapham and O. van Tellingen for TRPM7 in pTracer-CMV and pMX-luciferase-YFP-neo, respectively; J. Klarenbeek for technical assistance, and members of the Division of Cell biology I and group members for support, discussions and critical reading of the manuscript.
Grant support
This work was supported by KWF grants (KUN 2007-3733 and NKI 2010-4626) and KiKa (104) to KJ and FvL.
References
Aarts, M, Iihara, K, Wei, WL, Xiong, ZG, Arundine, M, Cerwinski, W, MacDonald, JF, and Tymianski, M. (2003) A key role for TRPM7 channels in anoxic neuronal death. Cell; 115 (7); 863-877.
Abed, E and Moreau, R. (2007) Importance of melastatin-like transient receptor potential 7 and cations (magnesium, calcium) in human osteoblast-like cell proliferation. Cell Prolif.; 40 (6); 849-865. Abed, E and Moreau, R. (2009) Importance of melastatin-like transient receptor potential 7 and magnesium in the stimulation of osteoblast proliferation and migration by platelet-derived growth factor. Am.J.Physiol.Cell Physiol.; 297 (2); C360-C368.
Ahmed, AA and Brenton, JD. (2005) Microarrays and breast cancer clinical studies: forgetting what we have not yet learnt. Breast Cancer Res.; 7 (3); 96-99.
Asgharzadeh, S, Pique-Regi, R, Sposto, R, Wang, H, Yang, Y, Shimada, H, Matthay, K, Buckley, J, Ortega, A, and Seeger, RC. (2006) Prognostic significance of gene expression profiles of metastatic neuroblastomas lacking MYCN gene amplification. J.Natl.Cancer Inst.; 98 (17); 1193-1203.
Bogenmann, E. (1996) A metastatic neuroblastoma model in SCID mice. Int J.Cancer; 67 (3); 379-385. Callera, GE, He, Y, Yogi, A, Montezano, AC, Paravicini, T, Yao, G, and Touyz, RM. (2009) Regulation of the novel Mg2+ transporter transient receptor potential melastatin 7 (TRPM7) cation channel by bradykinin in vascular smooth muscle cells. J.Hypertens.; 27 (1); 155-166.
Chaffer, CL and Weinberg, RA. (2011) A perspective on cancer cell metastasis. Science; 331 (6024); 1559-1564.
Chen, JP, Luan, Y, You, CX, Chen, XH, Luo, RC, and Li, R. (2010) TRPM7 regulates the migration of human nasopharyngeal carcinoma cell by mediating Ca(2+) influx. Cell Calcium; 47 (5); 425-432. Clapham, DE. (2003) TRP channels as cellular sensors. Nature; 426 (6966); 517-524.
Clark, K, Langeslag, M, van Leeuwen, B, Ran, L, Ryazanov, AG, Figdor, CG, Moolenaar, WH, Jalink, K, and van Leeuwen, FN. (2006) TRPM7, a novel regulator of actomyosin contractility and cell adhesion. EMBO J.; 25 (2); 290-301.
Clark, K, Middelbeek, J, Dorovkov, MV, Figdor, CG, Ryazanov, AG, Lasonder, E, and van Leeuwen, FN. (2008a) The alpha-kinases TRPM6 and TRPM7, but not eEF-2 kinase, phosphorylate the assembly domain of myosin IIA, IIB and IIC. FEBS Lett.; 582 (20); 2993-2997.
Clark, K, Middelbeek, J, Lasonder, E, Dulyaninova, NG, Morrice, NA, Ryazanov, AG, Bresnick, AR, Figdor, CG, and van Leeuwen, FN. (2008b) TRPM7 regulates myosin IIA filament stability and protein
101
-3
localization by heavy chain phosphorylation. J.Mol.Biol.; 378 (4); 790-803.Clark, K, Middelbeek, J, and van Leeuwen, FN. (2008c) Interplay between TRP channels and the cytoskeleton in health and disease. Eur.J.Cell Biol.; 87 (8-9); 631-640.
De Valck, D, Jin, DY, Heyninck, K, Van de Craen, M, Contreras, R, Fiers, W, Jeang, KT, and Beyaert, R. (1999) The zinc finger protein A20 interacts with a novel anti-apoptotic protein which is cleaved by specific caspases. Oncogene; 18 (29); 4182-4190.
Descazeaud, V, Mestre, E, Marquet, P, and Essig, M. (2012) Calcineurin regulation of cytoskeleton organization: a new paradigm to analyse the effects of calcineurin inhibitors on the kidney. J.Cell Mol. Med.; 16 (2); 218-227.
Desnos, C, Huet, S, and Darchen, F. (2007) ‘Should I stay or should I go?’: myosin V function in organelle trafficking. Biol.Cell; 99 (8); 411-423.
Dhennin-Duthille, I, Gautier, M, Faouzi, M, Guilbert, A, Brevet, M, Vaudry, D, Ahidouch, A, Sevestre, H, and Ouadid-Ahidouch, H. (2011) High expression of transient receptor potential channels in human breast cancer epithelial cells and tissues: correlation with pathological parameters. Cell Physiol. Biochem.; 28 (5); 813-822.
DuBois, SG, Kalika, Y, Lukens, JN, Brodeur, GM, Seeger, RC, Atkinson, JB, Haase, GM, Black, CT, Perez, C, Shimada, H, Gerbing, R, Stram, DO, and Matthay, KK. (1999) Metastatic sites in stage IV and IVS neuroblastoma correlate with age, tumor biology, and survival. J.Pediatr.Hematol.Oncol.; 21 (3); 181-189.
DuFort, CC, Paszek, MJ, and Weaver, VM. (2011) Balancing forces: architectural control of mechanotransduction. Nat.Rev.Mol.Cell Biol.; 12 (5); 308-319.
Ein-Dor, L, Kela, I, Getz, G, Givol, D, and Domany, E. (2005) Outcome signature genes in breast cancer: is there a unique set? Bioinformatics; 21 (2); 171-178.
Eppinga, RD, Peng, IF, Lin, JL, Wu, CF, and Lin, JJ. (2008) Opposite effects of overexpressed myosin Va or heavy meromyosin Va on vesicle distribution, cytoskeleton organization, and cell motility in nonmuscle cells. Cell Motil.Cytoskeleton; 65 (3); 197-215.
Fath, T, Fischer, RS, Dehmelt, L, Halpain, S, and Fowler, VM. (2011) Tropomodulins are negative regulators of neurite outgrowth. Eur.J.Cell Biol.; 90 (4); 291-300.
Gao, H, Chen, X, Du, X, Guan, B, Liu, Y, and Zhang, H. (2011) EGF enhances the migration of cancer cells by up-regulation of TRPM7. Cell Calcium; 50 (6); 559-568.
Geerts, D, Koster, J, Albert, D, Koomoa, DL, Feith, DJ, Pegg, AE, Volckmann, R, Caron, H, Versteeg, R, and Bachmann, AS. (2010) The polyamine metabolism genes ornithine decarboxylase and antizyme 2 predict aggressive behavior in neuroblastomas with and without MYCN amplification. Int J.Cancer; 126 (9); 2012-2024.
Geiger, B, Spatz, JP, and Bershadsky, AD. (2009) Environmental sensing through focal adhesions. Nat. Rev.Mol.Cell Biol.; 10 (1); 21-33.
Goel, M, Sinkins, W, Keightley, A, Kinter, M, and Schilling, W. (2005) Proteomic analysis of TRPC5- and TRPC6-binding partners reveals interaction with the plasmalemmal Na+/K+-ATPase. Pflugers Archiv
European Journal of Physiology; 451 (1); 87-98.
