ARTICLE
Recei ed 7 Ma 2016 |Accep ed 19 Sep 2016 |Published 2 Dec 2016
L- ype calcium channels egula e filopodia s abili y
and cance cell in asion downs eam o in eg in
signalling
Guillaume Jacqueme 1, Habib Baghi o 1,w, Ma ia Geo giadou1, Ha i Sih o2, Emilia Peuhu1, Pie e Ce ou -Jane 1,
Tao He3,w, Me ja Pe a
¨la
¨3,w, Pauliina K onq is 4, Heikki Joensuu2,5 & Johanna I aska1,6
Moun ing in i o,in i o and clinical e idence sugges an impo an ole o filopodia in
d i ing cance cell in asion. Using a high- h oughpu mic oscopic-based d ug sc een, we
iden i y FDA-app o ed calcium channel blocke s (CCBs) as po en inhibi o s o filopodia
o ma ion in cance cells. Unexpec edly, we disco e ha L- ype calcium channels a e
unc ional and equen ly exp essed in cance cells sugges ing a p e iously unapp ecia ed ole
o hese channels du ing umo igenesis. We u he demons a e ha , a filopodia, L- ype
calcium channels a e ac i a ed by in eg in inside-ou signalling, in eg in ac i a ion and S c.
Mo eo e , L- ype calcium channels p omo e filopodia s abili y and ma u a ion in o alin- ich
adhesions h ough he spa ially es ic ed egula ion o calcium en y and subsequen
ac i a ion o he p o ease calpain-1. Al oge he we unco e a no el and clinically ele an
signalling pa hway ha egula es filopodia o ma ion in cance cells and p opose ha cycles
o filopodia s abiliza ion, ollowed by ma u a ion in o ocal adhesions, di ec s cance cell
mig a ion and in asion.
DOI: 10.1038/ncomms13297 OPEN
1Tu ku Cen e o Bio echnology, Uni e si y o Tu ku, FIN-20520 Tu ku, Finland. 2Labo a o y o Molecula Oncology, T ansla ional Cance Biology p og am,
Uni e si y o Helsinki, FIN-00290 Helsinki, Finland. 3VTT Medical Bio echnology, Technical Resea ch Cen e o Finland, FIN-20520 Tu ku, Finland. 4Depa men
o Pa hology, Uni e si y o Tu ku and Tu ku Uni e si y Hospi al, FIN-20520 Tu ku, Finland. 5Depa men o Oncology, Helsinki Uni e si y Hospi al,
FIN-00290 Helsinki, Finland. 6Depa men o Biochemis y, Uni e si y o Tu ku, FIN-20520 Tu ku, Finland. wP esen add esses: No wegian Uni e si y o Science
and Technology, NO-7491 T ondheim, No way (H.B.); Radiome e Tu ku Oy, 20750 Tu ku, Finland (T.H.); Na u al Resou ces Ins i u e Finland, FI-31600
Jokioinen, Finland (M.P.). Co espondence and eques s o ma e ials should be add essed o G.J. (email: guillaume.jacqueme @u u.fi) o o
J.I. (email: Johanna.i aska@u u.fi).
NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.na u e.com/na u ecommunica ions 1
Cell mo ili y is in ol ed a e e y s age o umo igenesis
and con ibu es o p ima y umou g ow h, cance cell
dissemina ion and me as asis o ma ion1,2. As me as asis
emains he leading cause o cance - ela ed mo bidi y in pa ien s
wi h solid umou s3, he e is an immedia e need o gain a
mo e comp ehensi e unde s anding o he cellula s uc u es and
signalling pa hways ha d i e cance cell mig a ion. To mig a e,
cells in e ac and sense he su ounding ex acellula ma ix
(ECM) ia ansmemb ane adhesion ecep o s such as
in eg ins4–6. In eg in unc ion is con olled by a con o ma ional
swi ch be ween ac i e and inac i e s a es ha de e mine ECM
ligand in e ac ion and subsequen ecep o signalling5. In eg in
ac i a ion om wi hin he cell (in eg in inside-ou signalling) is
p omo ed by se e al mechanisms including he Rap1-RIAM- alin
pa hway and leads o in eg in-ECM engagemen (in eg in
ou side-in signalling) and he ec ui men and ac i a ion o a
la ge numbe o p o eins including he oncogenic kinases ocal
adhesion kinase (FAK) and S c o he in eg in4,7.
Filopodia a e ac in- ich finge -like p o usions ha ex end
om he plasma memb ane and ha e been implica ed in cell
mig a ion and in asion bo h in i o and in i o8. Filopodia a e
assembled a he on o in ading cance cells8–10 and
filopodia-like s uc u es p omo e cance cell su i al a
me as a ic si es11,12. Se e al filopodia-inducing p o eins such as
he molecula mo o myosin-X (MYO10) o he ac in-bundling
p o ein ascin p omo e cance cell in asion bo h in i o and in
mouse models and a e associa ed wi h poo pa ien p ognosis in
mul iple ca cinoma ypes8,13,14. Thus, in e e ing wi h filopodia
o ma ion could be a iable s a egy o inhibi cance me as asis
in i o.MYO10 is a homodime ic molecula mo o which is
up egula ed in b eas cance whe e i s exp ession co ela es
wi h mu an p53, poo p ognosis and inc eased me as a ic
po en ial13,15. Monome ic MYO10 is inac i e and localizes o
he cy osol o o Rab7-posi i e esicles16.MYO10 ac i a ion,
p omo ed by PI(3,4,5)P3, esul s in mo o dime iza ion and
d i es filopodia o ma ion by anspo ing ac in egula o s, cell–
cell adhesion ecep o s and in eg ins o filopodia ips.
He e, we desc ibe a no el d uggable and clinically ele an
pa hway egula ing MYO10-posi i e filopodia o ma ion and
s abili y. Unbiased high- h oughpu mic oscopy sc eens e eal
ha L- ype calcium channels, h ough egula ion o calcium en y
a filopodia ips, d i e filopodia s abiliza ion. Unexpec edly,
L- ype calcium channels a e exp essed and equen ly al e ed in
many human cance s and con ibu e o cance cell in asion by
egula ing filopodia downs eam o b1 in eg in and S c ac i a ion.
Resul s
L- ype calcium channel blocke s inhibi filopodia o ma ion.
To iden i y no el egula o s o filopodia o ma ion, cance cells
exp essing MYO10-GFP ( o induce and isualize filopodia) we e
ea ed wi h a lib a y comp ising o e 500 compounds o 1 h
and imaged using high- h oughpu mic oscopy. The numbe
o MYO10-posi i e spo s was au oma ically quan ified o
de e mine he a e age numbe o filopodia pe cell
(Supplemen a y Fig. 1A–D, see me hods o de ails). F om his
sc een, se e al L- ype calcium channel blocke s (CCBs)
we e iden ified as compounds ha consis en ly inhibi filopodia
o ma ion (Supplemen a y Figs 1D and 2A). In alida ion
expe imen s, ou s uc u ally dis inc CCBs (amlodipine
besyla e, elodipine, manidipine dichlo ide and cilnidipine)
we e demons a ed o significan ly educe he numbe o
MYO10-induced filopodia in b eas cance cells as e ficien ly as a
PI3K inhibi o (posi i e con ol o block MYO10 ac i i y16),
whe eas a ea men wi h zonisamide (inhibi s - ype calcium
channels, ol age-ga ed sodium channels and ca bonic
anhyd ase) o bume anide (inhibi s he Naþ/Kþ/2Cl
co anspo e ) ailed o a ec filopodia numbe (Fig. 1a).
Simila esul s we e ob ained in panc ea ic cance cells
ollowing CCB ea men (Supplemen a y Fig. 2B). In addi ion,
o e all inhibi ion o calcium en y in o cells by EGTA-media ed
chela ion o ex acellula calcium d ama ically educed he
numbe o MYO10-induced filopodia (Supplemen a y Fig. 2C).
Toge he , hese da a indica e ha calcium en y in o cells
ia L- ype calcium channels posi i ely egula es filopodia
o ma ion in cance cells.
L- ype calcium channel exp ession in cance cell lines.
The iden ifica ion o L- ype calcium channels as egula o s o
filopodia o ma ion in cance cells was unexpec ed as hei
exp ession and ac i i y is p incipally hough o be es ic ed
o exci able cells17,18. L- ype calcium channels a e composed
o mul iple subuni s (a1, a2d,band g), o which he a1
subuni o ms he co e channel anspo ing calcium ac oss he
plasma memb ane and he o he subuni s o m egula o y
componen s17,19. Fou di e en genes (CACNA1C,CACNA1D,
CACNA1F and CACNA1S) encode he a1 subuni ha is a ge ed
by CCBs. In e es ingly, all ou genes we e ound o be widely
exp essed a a iable le els ac oss cance cell lines20 (Fig. 1b)
ega dless o hei issue o o igin (Supplemen a y Fig. 3A–C).
Fu he mo e, we ound ha L- ype calcium channels a e
unc ional in cance cell lines MDA-MB-231 and PDAC
p53R172H as ea men wi h a specific L- ype calcium channel
ac i a o (BAY K8644) igge ed a apid and ansien inc ease
in in acellula calcium h oughou he cell body (Fig. 1c;
Supplemen a y Fig. 3D; Supplemen a y Mo ies 1 and 2)
(3–4 old inc ease a 1 min pos s imula ion; de ec ed wi h a
GFP-based calcium p obe21) and a filopodia ips (Fig. 1d).
Impo an ly, calcium en y media ed by he L- ype calcium
channel ac i a o was inhibi ed in he p esence o an L- ype
calcium channel inhibi o (Fig. 1e).
L- ype calcium channels a e clinically ele an in cance .
Analysis o public da ase s using cBioPo al22,23 e ealed ha
L- ype calcium channels a e commonly al e ed in pa ien samples
o di e en cance ypes (Supplemen a y Fig. 4A). In pa icula ,
o e 29% o pa ien samples in he B eas In asi e Ca cinoma24
da a se displayed al e a ions in CACNA1C,CACNA1D,
CACNA1F o CACNA1S and hese al e a ions showed a
significan associa ion wi h un a ou able pa ien su i al
(Supplemen a y Fig. 4B–D). In e es ingly, he wo s su i al
a es we e obse ed when al e a ions in CACNA1C,CACNA1D
and CACNA1S we e analysed oge he (Supplemen a y Table 1).
Fu he mo e, while all ou L- ype calcium channel a1 subuni s
a e exp essed a low le els in bo h heal hy b eas and b eas
ca cinoma samples (IST Online), CACNA1D is he mos
commonly o e exp essed a1 subuni in b eas ca cinoma while
CACNA1F is o en down egula ed (Supplemen a y Fig. 4E).
CACNA1D exp ession was also ound o be up egula ed in b eas
cance Oncomine da a se s25 and was he mos commonly
exp essed L- ype calcium channel a1 subuni in b eas and
panc ea ic cance cell lines (Supplemen a y Fig. 3B,C). Taken
oge he , hese da a indica e ha L- ype calcium channels a e
equen ly al e ed in b eas cance samples and ha al e a ion in
hese genes may co ela e wi h poo p ognosis. The exp ession o
he indi idual L- ype calcium channel a1 subuni in clinical
samples will equi e u he s udies using specific an ibodies.
L- ype calcium channels, cance cell mig a ion and in asion.
Filopodia suppo h ee-dimensional (3D) cell mig a ion
and cance in asion, pa icula ly in cance s ha bou ing p53
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297
2NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.na u e.com/na u ecommunica ions
mu a ions8,13. Impo an ly, simila o PI3K inhibi ion, CCB
ea men o p53 mu an b eas and panc ea ic cance cells
(MDA-MB-231, P53R280K; PDAC, P53R172H; Su.86.86, P53G245S)
significan ly impai ed cance cell in asion (Fig. 2a–c), in a
concen a ion-dependen manne (Supplemen a y Fig. 5A,B). In
con as , ea men wi h zonisamide o bume anide had no e ec
on cance cell in asion (Fig. 2a,b). In addi ion, up egula ion o
MYO10 exp ession lead o filopodia o ma ion and was su ficien
a
Zonisamide
Bume anide
MYO10-GFP
Ac in
MYO10-GFP
Ac in
MYO10-GFP
Ac in
MYO10-GFP
Ac in
DMSO Manidipine
Amlodipine Zonisamide Bume anide LY294002
MYO10-GFP
Ac in
MYO10-GFP
Ac in
MYO10-GFP
Ac in
MYO10-GFP
Ac in
0
20
40
60
80
100
120
DMSO
Manidipine
dichlo ide
Cilnidipine
Felodipine
Amlodipine
LY294002
Numbe o MYO10-posi i e
ilopodia pe cell
T- ype calcium channel blocke
Sodium channel inhibi o
L- ype calcium channel blocke s
DMSO
PI3K inhibi o
*** *** *** *** ***
CACNA1S
CACNA1C
CACNA1F
CACNA1D
MCF10DCIS.com
MCF-7
MDA-MB-453
Caco 2
MDA-MB-468
QGP-1
SNU-719
MDA-MB-361
PANC-1
Capan-2
Capan-1
DMS-53
Hs-683
SUM-1315M02
U-937
A2780
U-698-M
BT-474
HPAC
NCI-H889
NCI-H2171
COR-L279
U2OS
SW-480
A-375
SU.86.86
HeLa
K-562
NCI-H82
MDA-MB-231
A-431
COLO-704
b
ND
Gene exp ession
ead coun s
Cell lines:
CACNA1S
CACNA1C
CACNA1F
CACNA1D
c
–1 min +1 min +10 min
Calcium p obe
BAY K8644
Calcium p obe
Fold inc ease in
calcium p obe in ensi y
0
2
4
6
8
1 min
10 min
de
Fold inc ease in
calcium p obe in ensi y
DMSO
Amlodipine
0
2
4
6***
ROI
MYO10-mChe y– 1 min
–1 min BAY K8644
+1min
Calcium p obe MYO10-mChe y
–1 min BAY K8644
+1min
11
Calcium p obe
BAY K8644
FelodipineCilnidipine
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 ARTICLE
NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.na u e.com/na u ecommunica ions 3
o d i e cance cell in asion in non-in asi e cell lines in a CCB
sensi i e manne (Fig. 2d; Supplemen a y Fig. 5C,D). Func ional
L- ype calcium channels and PI3K ac i i y we e also impo an
o egula ion o di ec ionali y du ing 3D cell mig a ion on
cell-de i ed ma ices (Fig. 2e). Thus, MYO10 exp ession
acili a es filopodia o ma ion and in asion whe eas inhibi ion
o L- ype calcium channels using CCBs o inhibi ion o PI3K
impedes filopodia o ma ion, di ec ional cell mig a ion and
cance cell in asion.
O he L- ype calcium channel a1 subuni s, exp ession o
CACNA1D,CACNA1S, bu no CACNA1C, was de ec ed in
MDA-MB-231 cells a he mRNA le el (Fig. 3a; CACNA1F was
no es ed as i appea ed o be o en down egula ed in b eas cance
clinical samples). Impo an ly, indi idual silencing o CACNA1D
o CACNA1S gene exp ession using mul iple siRNA oligos
(Supplemen a y Fig. 6A,B) dec eased bo h filopodia o ma ion
and cance cell in asion (Fig. 3b,c), indica ing ha CACNA1D and
CACNA1S a e likely o syne gis ically suppo filopodia o ma ion
and cance cell in asion in MDA-MB-231 cells. The cell biological
eagen s o s udy L- ype calcium channels a e a he limi ed and
hus we we e only able o analyse CACNA1D localiza ion
(Supplemen a y Fig. 6B). Impo an ly, endogenous CACNA1D
localized wi h MYO10 a filopodia ips (Fig. 3d). In addi ion,
CACNA1D p o ein exp ession could be de ec ed in all majo b eas
cance sub ypes (cance ype, posi i e exp ession; luminal, 3/6;
iple-nega i e, 7/8; He 2-posi i e, 4/4; heal hy b eas , 0/1) om a
se o b eas ca cinoma pa ien samples (Fig. 3e). The e o e, L- ype
calcium channels, in pa icula CACNA1S and CACNA1D a e
exp essed in b eas cance and con ibu e o filopodia o ma ion
and cance cell in asion.
Calcium concen a ion and filopodia s abili y. Nex we sough
o u he s udy he ela ionship be ween calcium and filopodia
using wo dis inc app oaches: a GFP-based calcium p obe
(GCaMP6s) and a dye-based calcium indica o (Fluo4-AM). Li e-
cell TIRF imaging o cells ansien ly co-exp essing he GFP-based
calcium p obe and MYO10-mChe y e ealed ha s able filopodia
(41min MYO10 spo li e ime) display highe le els o calcium
han uns able filopodia (o1min MYO10 spo li e ime) (Fig. 4a;
Supplemen a y Mo ies 3) while no di e ence in he fluo escence
signal was obse ed be ween s able and uns able filopodia in cells
exp essing GFP as a con ol (Fig. 4b). Co espondingly, Fluo4-AM
demons a ed highe in ensi y in s able filopodia (Supplemen a y
Fig. 6C). These da a sugges ha inc eased calcium le els a
filopodia ips co ela es wi h filopodia s abili y. In line wi h his
no ion, CCB- ea ed cells apidly los mos o hei calcium-
posi i e s able filopodia and displayed a highe p opo ion o
uns able filopodia compa ed wi h DMSO- ea ed cells (Fig. 4c,d;
Supplemen a y Mo ies 4–6). As a second ead-ou o filopodia
s abili y, he a e age eloci y o MYO10 spo s was measu ed and
was ound o be much highe in CCB- ea ed cells indica i e o
as e filopodia u no e ollowing L- ype calcium channel
inhibi ion (Fig. 4e). CCB ea men also dec eased he numbe and
he s abili y o endogenous filopodia in bo h b eas and panc ea ic
cance cell lines (see me hod o de ails; Fig. 4 ,g; Supplemen a y
Fig. 6D). Taken oge he , calcium en y a filopodia ia
L- ype calcium channels con ibu es o filopodia s abili y as well as
filopodia o ma ion.
Ac i e in eg ins, filopodia and L- ype calcium channels.
In eg ins a e one o he es ablished ca gos o MYO10 ha a e
anspo ed o filopodia ips o media e cell–ECM adhesion26.In
addi ion o he filopodia ip localiza ion we obse ed ac i e b1
in eg ins in he sha s o MYO10-posi i e filopodia (Fig. 5a).
Mo eo e , and cong uen wi h a s udy desc ibing Rap1-RIAM
localiza ion o filopodia27, he in eg in ac i a o alin-1 was ound
o localize wi h MYO10 a filopodia ips (Fig. 5b). In addi ion, he
Rap1/ alin axis was impo an o filopodia o ma ion as inhibi ion
o Rap1 (Fig. 5c; Supplemen a y Fig. 7A) o silencing o alin-1
exp ession (Fig. 5d; Supplemen a y Fig. 7B,C) significan ly educed
filopodia numbe . Con e sely, o e exp ession o a cons i u i ely
ac i e mu an o Rap1 (CA-Rap1, Fig. 5e; Supplemen a y Fig. S7D)
o o he FERM domain o alin-1 ( alin head, Fig. 5 ;
Supplemen a y Fig. 7E), known o p omo e in eg in ac i i y28,
significan ly inc eased filopodia o ma ion. Gi en ha Rap1 can be
ac i a ed by inc eases in calcium le els29, we es ed whe he L- ype
calcium channels egula e filopodia o ma ion h ough Rap1.
In e es ingly, he exp ession o CA-Rap1 o alin head did no
es o e filopodia numbe s ollowing CCB ea men (Fig. 5e, ),
sugges ing ha L- ype calcium channels could ac downs eam o
he Rap1-RIAM- alin pa hway. In addi ion, CCB ea men did no
inhibi o e all in eg in ac i i y in cells sugges ing ha L- ype
calcium channels do no p omo e filopodia by inducing in eg in
ac i a ion (Supplemen a y Fig. 7F).
To u he s udy he ela ionship be ween in eg in ac i i y,
filopodia numbe and calcium concen a ion a filopodia ips, cells
exp essing a calcium p obe and MYO10-mChe y we e pla ed on
con o ma ion-specific an i-b1 in eg in an ibodies which lock b1
in eg in in ei he an ac i e o inac i e con o ma ion30. Using his
sys em, b1 in eg in ac i a ion significan ly inc eased filopodia
numbe as well as calcium le els a filopodia ips (Fig. 5g;
Supplemen a y Fig. 8). These in eg in-media ed e ec s we e ully
inhibi ed by CCB ea men (Fig. 5g; Supplemen a y Fig. 8), u he
demons a ing ha in eg in ac i a ion ac s ups eam o L- ype
calcium channels o egula e filopodia o ma ion (Fig. 5h).
In eg ins p omo e filopodia o ma ion and s abili y ia S c.As
he ac i a ion o FAK and S c a e downs eam e en s ollowing
b1 in eg in-ECM engagemen , we nex assessed a po en ial ole
o hese kinases in filopodia o ma ion. The S c inhibi o
Figu e 1 | L- ype calcium channel blocke s (CCBs) inhibi filopodia o ma ion and L- ype calcium channels a e unc ional in cance cells. (a) MDA-MB-
231 cells ansien ly exp essing MYO10-GFP and adhe ing o fib onec in (FN) we e ea ed wi h a ious compounds (10 mM) o 1 h, fixed, s ained o ac in
and imaged on a TIRF mic oscope (scale ba , 20 mm). The numbe o MYO10-posi i e filopodia was coun ed o each cell and displayed as a box plo
( h ee biological epea s, n4100 cells, ***P alueo8.3 1017). (b) Rela i e exp ession o he ou genes encoding he L- ype calcium channel a1 subuni
ac oss 676 commonly used cance cell lines19. Gene exp ession ead coun s a e displayed. The alue 7.99 co esponds o non-de ec ed (ND). Selec ed cell
lines a e anno a ed. (c) MDA-MB-231 cells ansien ly exp essing he calcium p obe (pGP-CMV-GCaMP6s) and adhe ing o FN we e ea ed wi h an
L- ype calcium channel ac i a o (BAY K8644; 1 mM) while being imaged on a TIRF mic oscope (63 objec i e). The ela i e inc ease in he in acellula
in ensi y o he calcium p obe was measu ed a 1 and 10 min pos s imula ion. Cell bounda ies a e indica ed by do ed lines ( h ee biological epea s, n¼74
cells; scale ba , 20 mm). (d) MDA-MB-231 cells ansien ly exp essing he calcium p obe (pGP-CMV-GCaMP6s) and MYO10-mChe y we e seeded on FN
and ea ed wi h an L- ype calcium channel ac i a o (BAY K8644; 1 mM) while being imaged on a TIRF mic oscope (100 objec i e; scale ba , 10 mm). The
inse shows a ep esen a i e MYO10-posi i e filopodia ip delinea ed by a do ed line. ROI: egion o in e es . (e) MDA-MB-231 cells ansien ly exp essing
he calcium p obe (pGP-CMV-GCaMP6s) and adhe ing o FN we e ea ed wi h an L- ype calcium channel ac i a o (BAY K8644; 1 mM) in combina ion
wi h DMSO o amlodipine besyla e (1 mM). The ela i e inc ease in he in acellula in ensi y o he calcium p obe was measu ed a 1 min ( h ee biological
epea s, n421 cells; ***P alueo1.39 10 4). P alues we e calcula ed using S uden ’s - es (unpai ed, wo- ailed, unequal a iance).
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297
4NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.na u e.com/na u ecommunica ions
dasa inib dec eased filopodia numbe in he o iginal mic oscopy-
based d ug sc een (60% ewe filopodia han DMSO a 20 mM).
This esul was alida ed using a small molecule inhibi o
o S c (PP2) which igge ed a significan loss in filopodia
compa ed wi h con ols (PP3 and DMSO). In con as , inhibi ion
o FAK had no e ec on filopodia o ma ion (Fig. 6a;
a
LY294002
DMSO
Felodipine
Amlodipine
Su.86.86
In asion (>45 μm)
0.8
1.2
0
0.4
***
***
***
c
0 μm 150 μm
DMSO
Cilnidipine
Felodipine
Manidipine
Amlodipine
Zonisamide
Bume anide
45 μm
Un ea ed
DMSO
Cilnidipine
Felodipine
Amlodipine
Zonisamide
Ca bamazepine
Bume anide
LY294002
In asion (>45 μm)
0
0.1
0.2
0.3
*** ***
*** ***
***
Manidipine
dichlo ide
MDA-MB-231
NS NS NS
LY294002
DMSO
Cilnidipine
Felodipine
Amlodipine
Zonisamide
Bume anide
Manidipine
dichlo ide
DMSO
Bume anide
PDAC P53–/– PDAC P53R172H
In asion (>45 μm)
0.8
1.2
0
0.4
ND
***
***
***
***
***
***
b d
DMSO
Amlodipine
DMSO
Amlodipine
U2OS
GFP
U2OS
MYO10-GFP
0
0.4
0.2
0.6
0.8
In asion (>45 μm)
***
PDAC P53–/–, DMSO
Cilnidipine
Felodipine
Manidipine
Amlodipine
Zonisamide
Bume anide
0 μm 150 μm45 μm
DMSO
PDAC P53R172H
Veloci y (μm min–1)
Un ea ed
DMSO
Cilnidipine
Felodipine
Amlodipine
Ca bamazepine
Bume anide
LY294002
0
4
8
12
Manidipine
dichlo ide
*********
Di ec ionali y
0
0.2
0.4
0.6
0.8
Un ea ed
DMSO
Cilnidipine
Felodipine
Amlodipine
Ca bamazepine
Bume anide
LY294002
Manidipine
dichlo ide
***
***
***
***
***
e
L- ype calcium channel blocke s
DMSO
T- ype calcium channel blocke
Sodium channel inhibi o s
PI3K inhibi o
ND
***
Figu e 2 | L- ype calcium channels egula e filopodia o ma ion and cance cell in asion. (a) MDA-MB-231 cells we e seeded in o an in e ed in asion
assay in he p esence o a ious compounds (10 mM) whe e indica ed o 48 h. The ela i e in asion o e 45 mm was quan ified (n¼ h ee biological
epea s, ***P alueo1.3 10 5). (b) P53/and P53R172H PDAC cells we e seeded in o an in e ed in asion assay in he p esence o a ious compounds
(10 mM) o 4 days. The ela i e in asion o e 45 mm was quan ified (n¼ h ee biological epea s, ***P alueo4.1 108). (c) Su.86.86 panc ea ic
ca cinoma cells we e seeded in o an in e ed in asion assay and allowed o in ade o 4 days in he p esence o a ious compounds (10 mM). Rela i e
in asion o e 45 mm was quan ified (n¼ h ee biological epea s, ***P alueo9103). (d) U2OS cells s ably exp essing ei he GFP o MYO10-GFP we e
seeded in o an in e ed in asion assay and allowed o in ade o 4 days in he p esence o amlodipine besyla e (10 mM) o DMSO. Rela i e in asion o e
45 mm was quan ified (n¼ h ee biological epea s, ***P alueo4.05 106). (e) MDA-MB-231 cells we e seeded on fib oblas -gene a ed cell
de i ed ma ices ( ep esen a i e image is shown) in he p esence o a ious compounds (10 mM), and cell mig a ion was eco ded o e 24 h. O e 65 cells
we e manually acked o each condi ion and mig a ion speed and di ec ionali y we e measu ed (n¼ wo biological epea s, scale ba , 200 mm;
***P alueo2.6 105). P alues we e calcula ed using S uden ’s - es (unpai ed, wo- ailed, unequal a iance). All e o ba s ep esen s.e.m.
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 ARTICLE
NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.na u e.com/na u ecommunica ions 5

Supplemen a y Fig. 9A). Simila o CCB ea men , PP2-media ed
inhibi ion o S c also p omo ed filopodia ins abili y
(Supplemen a y Fig. 9B). Impo an ly, o e exp ession o a
cons i u i ely-ac i e mu an o S c (CA-S c) inc eased
filopodia numbe while o e exp ession o a dominan -nega i e
mu an o S c (DN-S c) inhibi ed filopodia o ma ion (Fig. 6b;
Supplemen a y Fig. 9C). Immunofluo escence analyses u he
e ealed ac i e S c (pS cY416) localiza ion o filopodia (Fig. 6c)
sugges ing a cen al ole o S c in filopodia induc ion. These S c-
dependen e ec s appea ed o be downs eam o in eg in sig-
nalling as ea men wi h a S c inhibi o (Fig. 6d) o o e -
exp ession o DN-S c (Fig. 6e) significan ly educed filopodia
numbe and calcium le els a filopodia ips in cells pla ed on he
an i-ac i e b1 in eg in an ibody (Fig. 6d,e). Co espondingly,
o e exp ession o CA-S c was su ficien o bypass he equi e-
men o in eg in ac i a ion and p omo ed filopodia o ma ion
and calcium inc ease a filopodia ips in cells pla ed on he
an i-inac i e b1 in eg in an ibody (Fig. 6e). Taken oge he , hese
da a indica e ha S c ac i i y plays a key ole in filopodia
o ma ion and s abili y downs eam o in eg in signalling.
Impo an ly, as CA-S c-induced filopodia emained sensi i e o
CCB ea men (Fig. 6b) and as CCBs did no a ec o e all S c
ac i i y (Supplemen a y Fig. 9D), hese esul s indica e ha
S c ac s ups eam o L- ype calcium channels (Fig. 6 ). Howe e ,
how S c p omo es L- ype calcium channel ac i a ion emains o
be de e mined.
CACNA1D MYO10-mChe y CACNA1D MYO10-mChe y
d
cba
0
20
40
60
Numbe o MYO10-posi i e
ilopodia pe cell
siCTRL
siCACNA1C
siCTRL
siCACNA1D #1
***
siCACNA1S #1
***
siCACNA1D #2
***
siCACNA1S #2
***
In asion (>45 μm)
0.4
0
0.2
0.6
siCTRL
siCACNA1C
siCTRL
siCACNA1D #1
siCACNA1S #1
siCACNA1D #2
siCACNA1S #2
***
***
***
***
CACNA1D
CACNA1S
mRNA exp ession ela i e
o 106 copies o GAPDH
CACNA1C
0
0.05
0.10
0.15
0.20
0.25
ND
MDA-MB-231
CACNA1D
Luminal #1 Luminal #2 He 2-posi i e
eT iple-nega i e
Figu e 3 | CACNA1D and CACNA1S egula e filopodia o ma ion and cance cell in asion. (a) Rela i e exp ession o CACNA1C,CACNA1D and CACNA1S
in MDA-MB-231 cells as de e mined by Q-RT-PCR (n¼ h ee biological epea s). (b) MDA-MB-231 cells p e iously silenced o CACNA1D o CACNA1S
using mul iple RNAi oligos as indica ed and ansien ly exp essing MYO10-GFP we e pla ed on FN o 2 h, fixed and he numbe o MYO10-posi i e
filopodia pe cell was quan ified (n495 cells, h ee biological epea s ***P alueo2.4 1010). A sma pool o oligos a ge ing CACNA1C was used as an
addi ional con ol as no CACNA1C exp ession was de ec ed in hese cells. (c) MDA-MB-231 cells p e iously silenced o CACNA1D o CACNA1S using
mul iple oligos we e seeded in o an in e ed in asion assay and allowed o in ade o 48 h. Rela i e in asion o e 45 mm was quan ified (n¼ h ee biological
epea s, ***P alueo8.9 10 3). (d) MDA-MB-231 cells ansien ly exp essing MYO10-mChe y we e pla ed on FN o 2 h, s ained o endogenous
CACNA1D and imaged on a TIRF mic oscope (scale ba , 20 mm). (e) Rep esen a i e images o a ious ypes o b eas cance issue samples s ained o
CACNA1D (scale ba , 200 mm). P alues we e calcula ed using S uden ’s - es (unpai ed, wo- ailed, unequal a iance). All e o ba s ep esen s.e.m.
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297
6NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.na u e.com/na u ecommunica ions
Inc eased MYO10 and S c p o ein exp ession and/o ac i i y
ha e been epo ed independen ly in b eas cance and a e
linked o me as asis and poo pa ien su i al13,15,31. Gi en ha
S c is c i ical o he o ma ion o MYO10-posi i e filopodia, we
in es iga ed a possible associa ion be ween MYO10 and S c
p o ein le els and/o ac i i y in pa ien samples. Using
cBioPo al, we de e mined ha inc eased MYO10 le els s ongly
co ela es wi h inc eased S c mRNA le els in he B eas In asi e
Ca cinoma da a se (Log Odds a io: 0.826; P alue, o0.001).
Mo eo e , phosphop o eomic analyses indica ed ha pa ien
A e age numbe o
endogenous ilopodia
pe ame pe cell
0
10
20
30
DMSO
Felodipine
A
mlodipine
DMSO
Felodipine
A
mlodipine
A e age li e ime o
endogenous ilopodia (s)
30
20
10
0
40
50
dMYO10 spo
eloci y
DMSO
Felopidine
Amlodipine
A e age MYO10
spo eloci y (μm s
–1
)
*
0
0.04
0.08
0.12
***
e
0 min
1 min
2 min
0 min 1 min 2 min
MYO10-mChe yCalcium p obe
DMSO Felodipine
0
2 min
4 min
0
2 min
4 min
F
Amlodipine
0
2 min
4 min
a
Calcium p obeMYO10-mChe y
Calcium p obe in ensi y a
ilopodia ips
0
0.2
0.4
0.6
0.8 ***
<1 min
>1 min
Myo10 spo
li e ime
c
DMSOFelopidineAmlodipineDMSOFelopidineAmlodipine
<1 min
MYO10 spo s (% pe cell)
MYO10 spo li e ime
0
20
40
60
80
Felodipine
DMSO
Amlodipine
*** ***
*** ***
Calcium p obe
Myo10 spo
li e ime
0 min 1 min 2 min
*** ***
**
GFP in ensi y a ilopodia ips
NS
<1 min
>1 min
0
0.2
0.4
0.6
0.8
GFP
b
g
1–2 min 2–3 min 3–4 min >4 min
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 ARTICLE
NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.na u e.com/na u ecommunica ions 7
samples wi h high MYO10 exp ession (130 pa ien s ou o 1,080;
12%) also exhibi ed highe S c ac i i y (mean o al e a ion in
al e ed MYO10,0.05; mean o al e a ion in unal e ed g oup,
0.23; P alue, 1.408e 3; q alue, 0.0162). Fu he mo e,
e-analysis o a issue mic oa ay comp ising 1,212 samples
om a Finnish na ionwide popula ion-based b eas cance se ies
(FinP og coho ), p e iously s ained o MYO10 ( e . 13) and
pS cY416 ( e s 13,32), e ealed ha MYO10 p o ein le els
co ela e wi h S c ac i i y in pa ien samples (Fig. 6g; Table 1).
The e o e, MYO10 exp ession co ela es wi h bo h S c exp ession
and S c ac i i y in b eas cance pa ien samples.
Calpain 1 and filopodia o ma ion. Nex we sough o iden i y
he calcium- egula ed pa hway(s) ha is/a e in ol ed in
filopodia o ma ion. Cells exp essing MYO10-GFP we e ea ed
wi h inhibi o s a ge ing mul iple calcium- egula ed p o eins and,
among hese inhibi o s, only he compound a ge ing calpain
ac i i y significan ly diminished filopodia o ma ion (Fig. 7a;
Supplemen a y Fig. 10A). Mo e specifically, silencing o
calpain-1, and no calpain-2, exp ession dec eased filopodia
numbe compa ed wi h con ol (Fig. 7b; Supplemen a y
Fig. 10B–E). In addi ion, calpain-1 and calpain-2 appea ed o
ha e coope a i e unc ions du ing he in asi e p ocess as
silencing o bo h calpain-1 and calpain-2 inhibi ed cell in asion
mo e e ficien ly han he indi idual deple ion o ei he iso o m
(Fig. 7c). Immunofluo escence analyses e ealed ha calpain-1
localizes o filopodia (Fig. 7d) and a iome ic FRET analyses o
cells co-exp essing a calpain senso 33 and MYO10-mChe y
demons a ed ha calpains a e ac i e wi hin filopodia (Fig. 7e;
con ols ela ed o he use o he calpain FRET p obe a e shown
in Supplemen a y Fig. 11). To assess whe he calpains egula e
filopodia o ma ion downs eam o in eg in ac i a ion, cells
exp essing he calcium p obe and MYO10-mChe y we e
pla ed on an i-b1 in eg in an ibodies and ea ed wi h a
calpain inhibi o . Blocking calpain ac i i y dec eased filopodia
numbe bu no calcium le els a he emaining filopodia ips in
cells pla ed on he an i-ac i e b1 in eg in an ibody (Fig. 7 ).
Al oge he , hese da a indica e ha calpain-1 egula es filopodia
o ma ion downs eam o in eg in ac i a ion and calcium
accumula ion a filopodia ips.
Filopodia s abiliza ion leads o ocal adhesion o ma ion. The
impo ance o filopodia in he p ocess o umo igenesis has been
desc ibed in se e al cance s. Howe e , he significance o a
mechanism ha could egula e filopodia s abili y du ing cell
mig a ion has ne e been explo ed. CCB ea men only weakly
impac ed ocal adhesion dynamics as hei o e all li e ime
dis ibu ion was una ec ed (Fig. 8a). CCB ea men appea ed o
dec ease bo h he assembly and disassembly a e o ocal
adhesions as well as hei maximal size (Fig. 8a). Using li e-cell
TIRF imaging we ound ha alin-1-GFP localizes o bo h ocal
adhesions and MYO10-posi i e filopodia and in many ins ances
was obse ed o mo e oge he wi h MYO10 spo s along filopodia
sha s (Fig. 8b; Supplemen a y Mo ies 7). In s abilized filopodia,
bo h MYO10 and alin-1 we e ound o accumula e a filopodia
ips (Fig. 8b; Supplemen a y Mo ies 7). Subsequen ad ancemen
o he plasma memb ane coincided wi h MYO10 lea ing he
g owing alin-1-posi i e s uc u e and he o ma ion o new
MYO10-posi i e filopodia ollowed by accumula ion o a
alin-posi i e pa ch esembling a classical ocal adhesion.
These obse a ions a e in ag eemen wi h p e ious s udies
demons a ing a ole o filopodia in di ec ing lamellipodia
o ma ion in fib oblas s34 and ma u a ion o filopodial sha
adhesions in o ocal adhesions upon lamellipodia ad ancemen 35.
Taken oge he , hese da a allow us o p opose a model by which
filopodia o ma ion and s abiliza ion ia an in eg in/S c/L- ype
calcium channel/calpain pa hway, desc ibed he e, con ibu es o
di ec ional cell mo ili y and cance cell in asion (Fig. 8c).
Discussion
He e, we define he ounda ion o a d uggable and clinically
ele an signalling pa hway (Fig. 8c) ha egula es filopodia
o ma ion and s abili y in cance cells. No ably, L- ype calcium
channels a e exp essed in human cance and a ge ing hei
unc ion wi h FDA-app o ed CCBs impai s filopodia o ma ion
and blocks cance cell in asion. Mechanis ically, we es ablish a
link be ween a localized inc ease in calcium concen a ion a
filopodia ips, media ed by L- ype calcium channels and filopodia
s abili y. Mo eo e , in eg in inside-ou ac i a ion and ligand
binding is indispensable o filopodia o ma ion and appea s o be
he fi s c i ical s ep in his p ocess ollowing MYO10-dependen
deli e y o in eg ins o filopodia ips. Subsequen s eps equi e
in eg in ou side-in signalling o igge S c ac i a ion and he
spa ially es ic ed calcium en y a filopodia ips and he
ac i a ion o he calcium- egula ed p o ease calpain-1. Finally,
we demons a e ha filopodia s abiliza ion appea s o p ecede
ocal adhesion ma u a ion and p opose ha cycles o filopodia
s abiliza ion and ocal adhesion ma u a ion di ec cell mig a ion
and in asion.
The cen al posi ion o L- ype calcium channels in cance cell
in asion was e y unexpec ed as hese ol age-ga ed channels
(exp ession and ac i i y) a e conside ed o be es ic ed o
exci able cells (neu onal and muscle cells). Ne e heless, se e al
s udies epo oles o L- ype calcium channels in many o he
cell ypes including fib oblas s, kidney cells and endome ial and
p os a e cance cells36–38. In addi ion, analyses o a ailable gene
exp ession da a se s by us ( his s udy) and o he s25 e ealed a
wide exp ession o L- ype calcium channels in many cance cell
Figu e 4 | Calcium a filopodia ips egula es filopodia s abili y. (a) MDA-MB-231 cells ansien ly exp essing he calcium p obe (pGP-CMV-GCaMP6s)
and MYO10-mChe y we e pla ed on FN and imaged li e using a TIRF mic oscope (1 pic u e e e y 5 s; scale ba , 20 mm). The in ensi y o he calcium p obe
a MYO10-posi i e filopodia ips was measu ed and compa ed be ween ansien (o1 min li e ime) and s able filopodia (41 min li e ime) (n¼276
filopodia, h ee biological epea s, ***P alueo1.8 109). (b) MDA-MB-231 cells ansien ly exp essing GFP and MYO10-mChe y we e pla ed on FN and
imaged li e using a TIRF mic oscope. The in ensi y o GFP a MYO10-posi i e filopodia ips was measu ed and compa ed be ween ansien and s able
filopodia (n¼293 filopodia, h ee biological epea s). (c–e) MDA-MB-231 cells ansien ly exp essing he calcium p obe (pGP-CMV-GCaMP6s) and
MYO10-mChe y we e pla ed on FN, ea ed wi h DMSO, elodipine o amlodipine besyla e (10 mM) and imaged li e using a TIRF mic oscope (1 pic u e
e e y 5 s; scale ba , 20 mm). Rep esen a i e images a e shown (c). Fo each condi ion, MYO10-posi i e pa icles we e au oma ically acked and MYO10
spo li e ime (calcula ed as a pe cen age o he o al numbe o filopodia gene a ed pe cell) (d) and a e age MYO10 spo eloci y (e) we e plo ed
(see me hod o de ails; h ee biological epea s, n45,600 pa icles acked in mo e han 16 cells, *P alue ¼0.015, ***P alueo2.98 105).
( ,g) MDA-MB-231 cells ansien ly exp essing li eac -GFP we e pla ed on FN, ea ed wi h DMSO, elodipine o amlodipine besyla e (10 mM), and imaged
li e on a TIRF mic oscope. Mo ies we e segmen ed and endogenous filopodia au oma ically iden ified using CellGeo51 (See me hod o de ails). A e age
filopodia numbe pe ame and pe cell ( ) and a e age filopodia li e imes a e displayed (g)(n416 cells, wo biological epea s; *P alue ¼0.035,
***P alueo5.6 105). P alues we e calcula ed using S uden ’s - es (unpai ed, wo- ailed, unequal a iance). All e o ba s ep esen sem.
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297
8NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.na u e.com/na u ecommunica ions
lines and clinical samples. In pa icula , CACNA1D, which
localizes o filopodia ips and con ibu es o cance cell in asion
in i o (shown in his s udy), was ound o be exp essed in
pa ien samples in all majo b eas cance sub ypes.
The pa hway delinea ed he e as undamen al o filopodia
o ma ion in cance cells sha es some simila i ies wi h
calcium- egula ed p ocesses iden ified in no mal exci able cells.
In a smoo h muscle cells a5b1 in eg in signalling has been
g
Felodipine
A
mlodipine
An i-ac i e β1 in eg in mAb
An i-inac i e β1 in eg in mAb
Numbe o MYO10-posi i e
i lopodia pe cell
DMSO
DMSO
Felodipine
A
mlodipine
0
100
200
Calcium p obe in ensi y
a ilopodia ips
DMSO
DMSO
0
1
2
ec
Rap1
inhibi o
DMSO
Numbe o MYO10-posi i e
ilopodia pe cell
d
siCTRL
siTalin-1 #1
siTalin-1 #2
Felodipine
Amlodipine
DMSO
Felodipine
Amlodipine
DMSO
GFP
0
20
40
60
10
30
50
Felodipine
Amlodipine
DMSO
Felodipine
Amlodipine
DMSO
0
100
20
40
60
80
GFP CA-Rap1
b
aAc i e β1 in eg in
Talin-1
MYO10-mChe y
Ac i e β1 in eg in
MYO10-mChe y
Talin-1
AmlodipineFelodipineDMSODMSO
DMSO DMSO AmlodipineFelodipine
An i-inac i e
β1 in eg in mAb An i-ac i e β1 in eg in mAb
Calcium p obe
Calcium p obe
MYO10-mChe y
Numbe o MYO10-posi i e
ilopodia pe cell
Numbe o MYO10-posi i e
ilopodia pe cell
Numbe o MYO10-posi i e
ilopodia pe cell
Inac i e
in eg ins
Ac i e
in eg ins
Talin
Rap1
Ca
2+
L- ype Ca2+
channels
?
?
Filopodia s abili y
h
MYO10-mChe y
MYO10-mChe y
0
100
20
40
60
80
***
0
20
40
10
30
50
***
***
*** *** *** ***
***
***
***
*** ***
***
***
***
***
***
***
Talin head
***
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 ARTICLE
NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.na u e.com/na u ecommunica ions 9
Hs00913730_m1; CACNA1S: Hs00163885_m1) and we e analysed wi h he
7900HT as RT-PCR Sys em (Applied Biosys ems). Da a we e s udied using RQ
Manage So wa e (Applied Biosys ems).
Clinical sample analyses.Analyses o publicly a ailable da a se s we e pe o med
using IST online (MediSapiens L d; h p://is .medisapiens.com/) o cBioPo al
(h p://www.cbiopo al.o g/index.do). In e oga ion o he B eas In asi e
Ca cinoma (TCGA, Cell 2015)24 which con ains 1,105 samples was pe o med
using cBioPo al. The genomic p ofiles selec ed we e mu a ions, pu a i e
copy-numbe al e a ions om GISTIC and mRNA exp ession da a (mRNA
Exp ession z-Sco es (RNA Seq V2 RSEM)). Su i al analyses we e pe o med
using he Su i al ab and he co-exp ession analyses we e pe o med using he
Co-exp ession o he En ichmen abs.
Po en ial co ela ion be ween MYO10 and pS cY416 le els in pa ien samples
was assessed in he FinP og se ies, a Finnish na ionwide b eas cance coho .
Pa ien and sample inclusion c i e ia o FinP og a e desc ibed in de ail
elsewhe e56. The FinP og se ies (1,212 samples) was p e iously s ained o pS cY416
and MYO10 exp ession as desc ibed ea lie 13,32. F om hese da a a po en ial
associa ion be ween he exp ession g oups was assessed using he w2- es .
Pe mission o use pa ien samples o esea ch pu poses was p o ided by he
Minis y o Social A ai s and Heal h, Finland (pe mission 123/08/97).
CACNA1D s aining o pa ien s samples.The use o b eas cance issue samples
was pe o med wi h app o al om he e hical commi ee o Tu ku Uni e si y
Hospi al (149/6/2002) and Au ia Biobank (AB15-9859) and acco dingly in o med
consen was ob ained om all human pa icipan s. B eas issue ma e ial
was p epa ed acco ding o s anda d his ology p ac ice, ha is, fixed in bu e ed
o malin (pH 7.0) and embedded in o pa a fin blocks. TMAs we e p epa ed by
collec ing issue co es (4 mm in diame e ) om he ep esen a i e umou a ea o
each pa ien . Tissue sec ions (3 mm hick) we e cu and s ained wi h an i-CAC-
NA1D an ibody (1/600) using Lab Vision Au os aine 480 (The mo-Fishe Sci-
en ific, F emon , CA, USA) and de ec ed wi h Powe Vision þpolyme ki ,
acco ding o s anda d p o ocols (DPVB þ110HRP; Immuno ision Technologies,
Vision Biosys ems, No well, MA, USA), using diaminobenzidine as he
ch omogen.
S a is ical analysis.S a is ical analyses we e pe o med when app op ia e, and
P alues indica ed by an as e isk in he figu e legends. Unless o he wise indica ed,
he S uden ’s - es was used (unpai ed, wo- ailed, unequal a iance). F equency
ables we e analysed by using wo- ailed w2- es .
Da a a ailabili y.The au ho s decla e ha he da a suppo ing he findings o his
s udy a e a ailable wi hin he a icle and om he au ho s on eques .
Re e ences
1. Waclaw, B. e al. A spa ial model p edic s ha dispe sal and cell u no e limi
in a umou he e ogenei y. Na u e 525, 261–264 (2015).
2. Gundem, G. e al. The e olu iona y his o y o le hal me as a ic p os a e cance .
Na u e 520, 353–357 (2015).
3. Gup a, G. P. & Massague
´, J. Cance Me as asis: Building a F amewo k. Cell 127,
679–695 (2006).
4. Lega e, K. R., Wicks o
¨m, S. A. & Fa
¨ssle , R. Gene ic and cell biological analysis
o in eg in ou side-in signaling. Genes De . 23, 397–418 (2009).
5. Aska i, J. A., Buckley, P. A., Mould, A. P. & Humph ies, M. J. Linking in eg in
con o ma ion o unc ion. J. Cell Sci. 122, 165–170 (2009).
6. Jacqueme , G., Humph ies, M. J. & Caswell, P. T. Role o adhesion ecep o
a ficking in 3D cell mig a ion. Cu . Opin. Cell Biol. 25, 627–632 (2013).
7. Hynes, R. O. In eg ins: bidi ec ional, allos e ic signaling machines in hei oles
as majo adhesion ecep o s, in eg ins. Cell 110, 673–687 (2002).
8. Jacqueme , G., Hamidi, H. & I aska, J. Filopodia in cell adhesion, 3D mig a ion
and cance cell in asion. Cu . Opin. Cell Biol. 36, 23–31 (2015).
9. Jacqueme , G. e al. RCP-d i en a5b1 ecycling supp esses Rac and p omo es
RhoA ac i i y ia he RacGAP1-IQGAP1 complex. J. Cell Biol. 202, 917–935
(2013).
10. Paul, N. R. e al. a5b1 in eg in ecycling p omo es A p2/3-independen cance
cell in asion ia he o min FHOD3. J. Cell Biol. 210, 1013–1031 (2015).
11. Shibue, T., B ooks, M. W., Inan, M. F., Reinha d , F. & Weinbe g, R. A. The
ou g ow h o mic ome as ases is enabled by he o ma ion o filopodium-like
p o usions. Cance Disco . 2, 706–721 (2012).
12. Shibue, T., B ooks, M. W. & Weinbe g, R. A. An in eg in-linked machine y o
cy oskele al egula ion ha enables expe imen al umo ini ia ion and
me as a ic coloniza ion. Cance Cell 24, 481–498 (2013).
13. A jonen, A. e al. Mu an p53-associa ed myosin-X up egula ion p omo es
b eas cance in asion and me as asis. J. Clin. In es . 124, 1069–1082 (2014).
14. Li, A. e al. Fascin is egula ed by slug, p omo es p og ession o panc ea ic
cance in mice, and is associa ed wi h pa ien ou comes. Gas oen e ology 146,
1386 (2014).
15. Cao, R. e al. Ele a ed exp ession o myosin X in umou s con ibu es o b eas
cance agg essi eness and me as asis. B . J. Cance 111, 539–550 (2014).
16. Plan a d, L. e al. P dIns(3,4,5)P3 is a egula o o myosin-X localiza ion and
filopodia o ma ion. J. Cell Sci. 123, 3525–3534 (2010).
17. Ca e all, W. A. Vol age-ga ed calcium channels. Cold Sp ing Ha b. Pe spec .
Biol. 3, a003947 (2011).
18. Lipscombe, D., Hel on, T. D. & Xu, W. L- ype calcium channels: he low down.
J. Neu ophysiol. 92, 2633–2641 (2004).
19. Tyson, J. R. & Snu ch, T. P. Molecula na u e o ol age-ga ed calcium
channels: s uc u e and species compa ison. WIREs Memb . T ansp. Signal 2,
181–206 (2013).
20. Klijn, C. e al. A comp ehensi e ansc ip ional po ai o human cance cell
lines. Na . Bio echnol. 33, 306–312 (2015).
21. Chen, T.-W. e al. Ul asensi i e fluo escen p o eins o imaging neu onal
ac i i y. Na u e 499, 295–300 (2013).
22. Ce ami, E. e al. The cBio cance genomics po al: an open pla o m o
explo ing mul idimensional cance genomics da a. Cance Disco . 2, 401–404
(2012).
23. Gao, J. e al. In eg a i e analysis o complex cance genomics and clinical
p ofiles using he cBioPo al. Sci. Signal. 6, pl1 (2013).
24. Ci iello, G. e al. Comp ehensi e molecula po ai s o in asi e lobula b eas
cance . Cell 163, 506–519 (2015).
25. Wang, C.-Y., Lai, M.-D., Phan, N. N., Sun, Z. & Lin, Y.-C. Me a-analysis o
public mic oa ay da ase s e eals ol age-ga ed calcium gene signa u es in
clinical cance pa ien s. PLoS ONE 10, e0125766 (2015).
26. Zhang, H. e al. Myosin-X p o ides a mo o -based link be ween in eg ins and
he cy oskele on. Na . Cell Biol. 6, 523–531 (2004).
27. Laga igue, F. e al. A RIAM/lamellipodin- alin-in eg in complex o ms he ip
o s icky finge s ha guide cell mig a ion. Na . Commun. 6, 8492 (2015).
28. Calde wood, D. A. e al. The Talin head domain binds o in eg in be a
subuni cy oplasmic ails and egula es in eg in ac i a ion. J. Biol. Chem. 274,
28071–28074 (1999).
29. Bos, J. L. Linking Rap o cell adhesion. Cu . Opin. Cell Biol. 17, 123–128
(2005).
30. By on, A. e al. An i-in eg in monoclonal an ibodies. J. Cell Sci. 122, 4009–4011
(2009).
31. Elsbe ge , B. T ansla ional e idence on he ole o S c kinase and ac i a ed S c
kinase in in asi e b eas cance . C i . Re . Oncol. Hema ol. 89, 343–351 (2014).
32. Vassile , B. e al. Ele a ed le els o S AR- ela ed lipid ans e p o ein 3 al e
choles e ol balance and adhesi eness o b eas cance cells: po en ial
mechanisms con ibu ing o p og ession o HER2-posi i e b eas cance s.
Am. J. Pa hol. 185, 987–1000 (2015).
33. S ockholm, D. e al. Imaging calpain p o ease ac i i y by mul ipho on FRET in
li ing mice. J. Mol. Biol. 346, 215–222 (2005).
34. Johnson, H. E. e al. F-ac in bundles di ec he ini ia ion and o ien a ion o
lamellipodia h ough adhesion-based signaling. J. Cell Biol. 208, 443–455
(2015).
35. Hu, W., Weh le-Halle , B. & Vogel, V. Ma u a ion o filopodia sha adhesions
is up egula ed by local cycles o lamellipodia ad ancemen s and e ac ions.
PLoS ONE 9, e107097 (2014).
36. Hao, J. e al. Ca2 þchannel subuni a1D p omo es p oli e a ion and
mig a ion o endome ial cance cells media ed by 17b-es adiol ia he G
p o ein-coupled es ogen ecep o . FASEB J. 29, 2883–2893 (2015).
37. Chen, R. e al. Ca 1.3 channel a1D p o ein is o e exp essed and modula es
and ogen ecep o ansac i a ion in p os a e cance s. U ol. Oncol. 32, 524–536
(2014).
38. Yang, S. & Huang, X.-Y. Ca2 þinflux h ough L- ype Ca2 þchannels con ols
he ailing ail con ac ion in g ow h ac o -induced fib oblas cell mig a ion.
J. Biol. Chem. 280, 27130–27137 (2005).
39. Wu, X. e al. Modula ion o calcium cu en in a e iola smoo h muscle by
a b3 and a5b1 in eg in ligands. J. Cell Biol. 143, 241–252 (1998).
40. Wu, X., Da is, G. E., Meininge , G. A., Wilson, E. & Da is, M. J. Regula ion o
he L- ype calcium channel by alpha 5be a 1 in eg in equi es signaling be ween
ocal adhesion p o eins. J. Biol. Chem. 276, 30285–30292 (2001).
41. Gui, P. e al. In eg in ecep o ac i a ion igge s con e ging egula ion o
Ca 1.2 calcium channels by c-S c and p o ein kinase A pa hways. J. Biol. Chem.
281, 14015–14025 (2006).
42. Robles, E., Hu enloche , A. & Gomez, T. M. Filopodial calcium ansien s
egula e g ow h cone mo ili y and guidance h ough local ac i a ion o calpain.
Neu on 38, 597–609 (2003).
43. Gomez, T. M., Robles, E., Poo, M. & Spi ze , N. C. Filopodial calcium ansien s
p omo e subs a e-dependen g ow h cone u ning. Science 291, 1983–1987
(2001).
44. Lau e milch, N. J. & Spi ze , N. C. Regula ion o calcineu in by g ow h cone
calcium wa es con ols neu i e ex ension. J. Neu osci. 20, 315–325 (2000).
45. Wen, Z., Gui land, C., Ming, G.-L. & Zheng, J. Q. A CaMKII/calcineu in swi ch
con ols he di ec ion o Ca(2 þ)-dependen g ow h cone guidance. Neu on 43,
835–846 (2004).
ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297
16 NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.na u e.com/na u ecommunica ions

46. Yang, M. & B ackenbu y, W. J. Memb ane po en ial and cance p og ession.
F on . Physiol. 4, 185 (2013).
47. F ase , S. P. e al. Vol age-ga ed sodium channel exp ession and po en ia ion o
human b eas cance me as asis. Clin. Cance Res. 11, 5381–5389 (2005).
48. Hein, M. Y. e al. A human in e ac ome in h ee quan i a i e dimensions
o ganized by s oichiome ies and abundances. Cell 163, 712–723 (2015).
49. Paul, N. R., Jacqueme , G. & Caswell, P. T. Endocy ic a ficking o in eg ins in
cell mig a ion. Cu . Biol. 25, R1092–R1105 (2015).
50. By on, A. e al. A p o eomic app oach e eals in eg in ac i a ion s a e-
dependen con ol o mic o ubule co ical a ge ing. Na . Commun. 6, 6135
(2015).
51. Sbalza ini, I. F. & Koumou sakos, P. Fea u e poin acking and ajec o y
analysis o ideo imaging in cell biology. J. S uc . Biol. 151, 182–195 (2005).
52. Tsyganko , D. e al. CellGeo: a compu a ional pla o m o he analysis
o shape changes in cells wi h complex geome ies. J. Cell Biol. 204, 443–460
(2014).
53. Be ginski, M. E., Vi iol, E. A., Hahn, K. M. & Gomez, S. M. High- esolu ion
quan ifica ion o ocal adhesion spa io empo al dynamics in li ing cells. PLoS
ONE 6, e22025 (2011).
54. Jacqueme , G. e al. Rac1 is deac i a ed a in eg in ac i a ion si es
h ough an IQGAP1-filamin-A-RacGAP1 pa hway. J. Cell Sci. 126, 4121–4135
(2013).
55. Hennigan, R. F., Hawke , K. L. & Ozanne, B. W. Fos- ans o ma ion ac i a es
genes associa ed wi h in asion. Oncogene 9, 3591–3600 (1994).
56. Joensuu, H. e al. Amplifica ion o e bB2 and e bB2 exp ession a e supe io
o es ogen ecep o s a us as isk ac o s o dis an ecu ence in pT1N0M0
b eas cance : a na ionwide popula ion-based s udy. Clin. Cance Res. 9,
923–930 (2003).
Acknowledgemen s
We hank J. Sii onen and P. Laasola o echnical assis ance, M. Saa i o help wi h
he mic oscopes and H. Hamidi o scien ific w i ing and edi ing o he manusc ip .
This s udy has been suppo ed by he Academy o Finland, ERC S a ing G an ,
ERC Consolida o G an , he Sig id Juselius Founda ion, and he Finnish Cance
O ganiza ion. G.J. and M.G. a e suppo ed by an EMBO Long-Te m Fellowship. E.P. is
suppo ed by an Academy o Finland pos doc o al ellowship.
Au ho con ibu ions
G.J designed, ca ied ou and analysed he majo i y o he expe imen s wi h help om J.I,
M.G, E.P and P.C. G.J w o e he manusc ip wi h help om J.I. H.B pe o med he
o iginal d ug sc een wi h c ucial help om T.H and M.P. H.S, H.J and P.K p o ided he
clinical samples and he issue sample analyses. J.I supe ised he s udy.
Addi ional in o ma ion
Supplemen a y In o ma ion accompanies his pape a h p://www.na u e.com/
na u ecommunica ions
Compe ing financial in e es s: The au ho s decla e no compe ing financial in e es s.
Rep in s and pe mission in o ma ion is a ailable online a h p://npg.na u e.com/
ep in sandpe missions/
How o ci e his a icle: Jacqueme , G. e al. L- ype calcium channels egula e filopodia
s abili y and cance cell in asion downs eam o in eg in signalling. Na . Commun.
7, 13297 doi: 10.1038/ncomms13297 (2016).
Publishe ’s no e: Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in
published maps and ins i u ional a filia ions.
This wo k is licensed unde a C ea i e Commons A ibu ion 4.0
In e na ional License. The images o o he hi d pa y ma e ial in his
a icle a e included in he a icle’s C ea i e Commons license, unless indica ed o he wise
in he c edi line; i he ma e ial is no included unde he C ea i e Commons license,
use s will need o ob ain pe mission om he license holde o ep oduce he ma e ial.
To iew a copy o his license, isi h p://c ea i ecommons.o g/licenses/by/4.0/
The Au ho (s) 2016
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 ARTICLE
NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.na u e.com/na u ecommunica ions 17