RESEARCH Open Access
SRC inhibi ion p e en s P-cadhe in
media ed signaling and unc ion in basal-
like b eas cance cells
Ana So ia Ribei o
1,2*
, Ana Ri a Nob e
1,2,3
, Nuno Mendes
1,2
,JoãoAlmeida
1,2,3
, And é Filipe Viei a
1,2
, Bá ba a Sousa
1,2
,
Filomena A. Ca alho
4
,JoanaMon ei o
2
,An ónioPolónia
1,2,5
, Ma ina Fonseca
6
, João Miguel Sanches
6
,
Nuno C. San os
4
,RaquelSe uca
1,2,5
and Joana Pa edes
1,2,5
Abs ac
Backg ound: Basal-like b eas cance (BLBC) is a poo p ognosis subg oup o iple-nega i e ca cinomas ha s ill
lack speci ic a ge he apies and accu a e bioma ke s o ea men selec ion. P-cadhe in is equen ly o e exp essed
in hese umo s, p omo ing cell in asion, s em cell ac i i y and umo igenesis by he ac i a ion o S c-Family kinase
(SRC) signaling. The e o e, ou aim was o e alua e i he ea men o BLBC cells wi h dasa inib, he FDA app o ed SRC
inhibi o , would impac on P-cadhe in induced umo agg essi e beha io .
Me hods: P-cadhe in and SRC exp ession was e alua ed in a se ies o in asi e B eas Cance and con ingency ables
and chi-squa e es s we e pe o med. Cell-cell adhesion measu emen s we e pe o med by A omic Fo ce Mic oscopy,
whe e equency his og ams and Gaussian cu es we e applied. 2D and 3D cell mig a ion and in asion, p o eases
sec e ion and sel - enew po en ial we e e alua ed in i o. S uden ’s - es s we e used o de e mine s a is ically
signi ican di e ences. The cadhe in/ca enin complex in e ac ions we e e alua ed by in si u p oximi y-liga ion
assay, and s a is ically signi ican esul s we e de e mined by using Mann-Whi ney es wi h a Bon e oni
co ec ion. In i o xenog a mouse models we e used o e alua e he impac o dasa inib on umo g ow h
and su i al. ANOVA es was used o e alua e he di e ences in umo size, conside ing a con idence in e al
o 95%. Su i al cu es we e es ima ed by he Kaplan-Meie ’s me hod, using he log- ank es o assess
signi ican di e ences o mice o e all su i al.
Resul s: Ou da a demons a ed ha P-cadhe in o e exp ession is signi ican ly associa ed wi h SRC ac i a ion
in b eas cance cells, which was also alida ed in a la ge se ies o p ima y umo samples. SRC ac i i y
supp ession wi h dasa inib signi ican ly p e en ed he in i o unc ional e ec s o P-cadhe in o e exp essing
cells, as well as hei in i o umo igenic and me as a ic abili y, by inc easing mice o e all su i al.
Mechanis ically, SRC inhibi ion a ec s P-cadhe in downs eam signaling, escues he E-cadhe in/p120-ca enin
complex o he cell memb ane, eco e ing cell-cell adhesion unc ion.
Conclusions: In conclusion ou indings show ha a ge ing P-cadhe in/SRC signaling and unc ional ac i i y
may open no el he apeu ic oppo uni ies o highly agg essi e and poo p ognos ic basal-like b eas cance .
Keywo ds: P-cadhe in, Dasa inib, Basal-like b eas cance , S c amily kinase
* Co espondence: [email p o ec ed]
1
Epi helial In e ac ions in Cance (EPIC), i3S - Ins i u o de In es igação e
Ino ação em Saúde, Uni e sidade do Po o, Rua Al edo Allen 208, 4200-135
Po o, Po ugal
2
Ipa imup, Ins i u e o Molecula Pa hology and Immunology o he
Uni e si y o Po o, Po o, Po ugal
Full lis o au ho in o ma ion is a ailable a he end o he a icle
© The Au ho (s). 2018 Open Access This a icle is dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0
In e na ional License (h p://c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion, and
ep oduc ion in any medium, p o ided you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o
he C ea i e Commons license, and indica e i changes we e made. The C ea i e Commons Public Domain Dedica ion wai e
(h p://c ea i ecommons.o g/publicdomain/ze o/1.0/) applies o he da a made a ailable in his a icle, unless o he wise s a ed.
Ribei o e al. Cell Communica ion and Signaling (2018) 16:75
h ps://doi.o g/10.1186/s12964-018-0286-2
Backg ound
Basal-like b eas cance is an in insic molecula sub ype
(BLBC) [1,2] associa ed wi h an agg essi e biological
beha iou and o a wo se pa ien p ognosis, which ypic-
ally does no exp ess ho monal ecep o s and HER2,
cons i u ing he as majo i y o iple-nega i e b eas
ca cinomas [3,4]. Un il now, a ge ed he apies a e un-
a ailable o hese subg oup o b eas cance pa ien s,
mainly due o he absence o well-de ined molecula a -
ge s, being chemo and adio he apy he only he apeu ic
op ions o ea hese pa ien s. Consequen ly, i is u gen
o ind speci ic bioma ke s/pa hways ha could impac
he e ec i e ea men o hese poo p ognosis b eas
ca cinomas o accu a e p edic i e molecula indica o s
ha could be used o selec pa ien s o a ge he apies
al eady a ailable.
In he las yea s, ou esea ch eam has been mainly
ocused in iden i ying key molecules ha could be used
as pu a i e molecula a ge s in BLBC [5,6]. We and
o he s ha e ound ha he majo i y o hese agg essi e
b eas ca cinomas exp ess he basal ma ke P-cadhe in
(a cell-cell adhesion molecule), which o e exp ession is
signi ican ly associa ed o a wo se disease- ee and
o e all pa ien su i al [5,7,8], as well as wi h a
p o-in asi e and s em-like cell beha iou [9–14]. Mo e-
o e , we demons a ed ha P-cadhe in inhibi ion
dec eases in i o cell in asion and sensi izes b eas
cance cells o adia ion, as well as dec eases in i o
umo igenic abili y, being indica ed as a pu a i e a ge
o BLBC ea men [14]. We ha e also shown ha
P-cadhe in-induced unc ional e ec s a e due o he
sec e ion o ac i e p o eoly ic enzymes, such as MMPs,
and o p o-in asi e soluble clea ed- o ms o P-cadhe in
[9], as well as o he ac i a ion o α6β4 in eg in [13] and
o he inhibi ion o he E-cadhe in supp essi e in asi e
unc ion, by he dis up ion o he E-cadhe in/p120c n
complex a he cell memb ane [11]. I is well known ha
bo h cy oplasmic p120c n and α6β4 in eg in ac i a ion
can induce he phospho yla ion o S c Family Kinases
(SRC) and ocal adhesion kinase (FAK), as well as he ac-
i a ion o Rac1 small GTPase [15–17]. The e a e se e al
lines o e idence showing ha y osine phospho yla ion
o hese enzymes may play a ole in he dis up ion o
cell-cell adhesion, due o loss o cadhe in/ca enin associ-
a ion, wi h a huge impac in cance p og ession. Impo -
an ly, we ha e desc ibed ha P-cadhe in exp ession
induces SRC ac i a ion in b eas cance cells, sugges -
ing ha SRC signaling is being ups eam egula ed by
his cell-cell adhesion molecule [13,18]. Thus, in
o de o e alua e he ole o P-cadhe in/SRC signaling
ac i a ion in BLBC cells, we ha e used dasa inib (an
FDA app o ed SRC inhibi o ) o block his pa hway
and p e en P-cadhe in induced agg essi e beha io
o BLBC cells.
Me hods
Cell cul u e and siRNA ans ec ions
Human b eas cance cell (BCC) lines we e ob ained as
ollows: MCF-7/AZ was p o ided by P o . Ma c Ma eel
(Ghen Uni e si y, Belgium); SUM-149 was kindly p o-
ided by D . S ephen E hie (Uni e si y o Michigan,
USA); BT20 and MDA-MB-468 we e ob ained om
Ame ican Type Cul u e Collec ion (Manassas, VA,
USA). The MCF-7/AZ cell line is a a ian o he human
BCC line MCF-7 ha exp esses basal le els o
P-cadhe in. This cell line was e o i ally ansduced o
encode only EGFP (LZRS-IRES-EGFP plasmid, MCF-7/
AZ.Mock cell line) o bo h P-cadhe in and EGFP
(LZRS-P-cad-IRES-EGFP plasmid, MCF-7/AZ.P-cad cell
line), as p e iously desc ibed [19]. MCF-7/AZ.Mock cell
line was used as a con ol.
Cells we e ou inely main ained a 37 °C, 5% CO
2
,in
he ollowing media (In i ogen L d., Paisley, UK): 50%
DMEM + 50% HamF12 (MCF-7/AZ, SUM149) and
DMEM (BT20, MDA-MB-468), supplemen ed wi h 10%
hea -inac i a ed e al bo ine se um (FBS), 100 IU/ml
penicillin and 100 mg/ml s ep omycin (In i ogen, UK).
SUM149 medium was supplemen ed wi h 5 μg/ml o
insulin and 1 μg/ml o hyd oco isone (Sigma-Ald ich,
S . Louis, MO, USA).
T ans ec ions o alida ed small in e e ing RNA
(siRNA) speci ic o P-cadhe in (Hs_CDH3_6, GW Vali-
da ed siRNA, Qiagen, Camb idge, MA, USA) and SRC
(Hs_SRC_1 FlexiTube and Hs_SRC_11 FlexiTube
siRNA, Qiagen) was ca ied ou using Lipo ec amine
2000 (In i ogen), in a inal concen a ion o 50 nM
(CDH3 and SRC1) and 100 nM (SRC11), acco ding o
he manu ac u e ’s ecommended p ocedu es. A e
incuba ion o 5 min, he siRNA and Lipo ec amine
2000 solu ions we e mixed, incuba ed o 20 min, and
added o cell cul u e medium. A siRNA con ol, wi h no
homology o any gene, was also used (Qiagen), o e i y
he speci ici y o he esul s ob ained.
D ugs and an ibodies
Dasa inib was ob ained om Sequoia Resea ch P oduc s
(Pangbou ne, UK), dissol ed in DMSO, and cells we e
ea ed a a inal concen a ion o 100 nM each 24 h o
a o al pe iod o 48 h. PP2 (P0042, Sigma-Ald ich) was
dissol ed in DMSO and added o he cells a a inal
concen a ion o 1 and 10 μg/ml.
Monoclonal an ibodies we e ob ained as ollows:
E-cadhe in (WB) (Taka a Bio Inc., Shiga, Japan); P-cadhe in
(WB) and p120c n (BD T ansduc ion Labo a o ies, Palo
Al o, CA, USA); E-cadhe in (PLA, IF, IHQ), P-cadhe in
(IF), o al S c (L4A1) and pSRC(Ty 416) (Cell Signaling
Technology); β-ac in (San a C uz Bio echnologies, San a
C uz, CA, USA). The de ailed condi ions o each an ibody
a e desc ibed on Addi ional ile 1:TableS1.
Ribei o e al. Cell Communica ion and Signaling (2018) 16:75 Page 2 o 16
Cell-cell adhesion measu emen s by a omic o ce
mic oscopy
Cells we e cul u ed in a issue cul u e dish a a low cell
densi y concen a ion o ha e only dispe sed cells, wi h-
ou eaching cell con luence. On he day o he expe i-
men , cells we e washed wice wi h PBS and 1 mL o
se um- ee DMEM was added. Fo he cell-cell adhesion
expe imen s, ipless a ow TL1 can ile e s (Nanowo ld,
Neucha el, Swi ze land) we e used, wi h a nominal
sp ing cons an o 0.03 N/m. Can ile e s we e cleaned
o 15 min wi h UV ligh and coa ed wi h poly-D-lysine
(50 μg/ml) o a leas 30 min. Can ile e s we e le in
he poly-D-lysine solu ion un il used.
Isola ed cell suspensions we e ob ained a e 1 h o
incuba ion in PBS (wi hou ypsin ea men ) o con lu-
en cells in a sepa a e dish and subsequen pipe e
dispe sion. The cells and he unc ionalized can ile e s
we e moun ed on he CellHesion module (JPK Ins u-
men s, Be lin, Ge many), wi h a 100-μm z- ange piezo-
elec ic scanne , connec ed o he NanoWiza d II a omic
o ce misc oscope moun ed on he op o an Axio e
200 in e ed mic oscope. 100 μL o cells suspension
we e injec ed in o he cells subs a e dish. Cells we e
allowed o se le o 30 s be o e cap u ing by a unc ion-
alized can ile e o one cell using he AFM con ac
mode. A e 30 s o can ile e p essing on o he cell, he
can ile e was aised 100 μm on he z- ange and he
a ached cell allowed o es o 1 min be o e ini ia ing
he con ac wi h an adhe en cell on he subs a e
unde nea h. Cell-cell con ac was es ablished wi h an
applied o ce o 300 pN, in cons an heigh and
closed-loop mode. The AFM ip esonan equency was
main ained a 2 Hz and a cell-cell con ac ime o 5 s
was main ained be o e can ile e e ac ion, wi h a
z- ange displacemen o 50 μm). Fi e o ce-dis ance
cu es we e pe o med on each cell on he subs a e,
wi h a 5 s pause be ween hem. A maximum o 8 di e -
en adhe en cells on he subs a e we e es ed wi h he
same cell a ached o he AFM can ile e .
Wes e n blo
P o ein lysa es we e p epa ed om cul u ed cells in ca-
enin lysis bu e [1% ( / ) T i on X-100 and 1% ( / )
NP-40 (Sigma) in desionized PBS] wi h 1:7 p o eases
inhibi o s cock ail (Roche Diagnos ics GmbH). Cells
we e washed wice in PBS and we e allowed o lyse in
500 μl o ca enin lysis bu e o 10 min, a 4 °C. Lysa es
we e hen submi ed o o ex 3 imes, cen i uged
10 min (14,000 pm) and supe na an s we e collec ed.
P o ein concen a ion was de e mined using a B ad o d
assay (BioRad p o ein quan i ica ion sys em, BioRad,
He cules, CA, USA). P o eins we e dissol ed in Sample
Bu e [Laemmli wi h 5% ( / ) 2-β-me cap oe hanol and
5% ( / ) b omophenol blue] and boiled o 5 min a 95 °
C. Samples we e hen sepa a ed by 8% SDS-PAGE. P o-
eins we e ans e ed in o ni ocellulose memb anes
(Ame sham Hybond ECL, Ame sham biosciences, Buck-
inghamshi e, UK) a 130 V o 1 h. Fo immunos aining,
memb anes we e blocked wi h 5% (w/ ) non- a d y milk
in PBS con aining 0.5% ( / ) Tween-20. Memb anes we e
subsequen ly incuba ed wi h p ima y an ibodies o 1 h
a RT (E-cadhe in –dilu ion 1:1000; P-cadhe in –
dilu ion 1:500; p120c n –dilu ion 1:1000; S c –dilu ion
1:500; pSRC(Ty 416) –dilu ion 1:500 and β-ac in –dilu-
ion 1:1000), ollowed by ou 5 min washes in PBS/
Tween-20 and incuba ion wi h ho se adish pe oxidase-
conjuga ed seconda y an ibodies (donkey an i-goa , goa
an i-mouse, goa an i- abbi –dilu ion 1:2000) o 1 h.
Memb anes we e hen washed o 30 min in PBS/
Tween-20. P o eins we e de ec ed using ECL eagen
(Ame sham biosciences) as a subs a e. Blo s we e ex-
posed o an au o adiog aphic ilm. Band quan i ica ion o
wes e n blo s was pe o med using Quan i y One
(BioRad). Each immunoblo was pe o med a leas h ee
imes, and he ones selec ed o igu es a e ep esen a i e
expe imen s.
3D b eas cance cells sphe oids cul u e
Agg egoids we e ob ained by pe o ming he slow agg e-
ga ion assay, as p e iously desc ibed [19]. B ie ly, 1 × 10
4
cells in 200 μL medium we e seeded on solidi ied aga in
a 96-well pla e o 12 h. Medium con aining he o med
cellula agg ega es (10–100 cells) we e cen i uged and
u he embedded in acid solubilized a ail collagen I
gels (1.7 mg/ml collagen I in PBS; 22.5 mM NaHCO
3
;
8 mM NaOH; 9% o non-supplemen ed cul u e
medium). Cellula agg egoids we e pla ed in a 8-well
co e slip bo om chambe s (IBIDI, Ge many) ollowed
by 30 min incuba ion a 37 °C o allow polyme iza ion.
A e ha ime, medium wi h o wi hou dasa inib was
added o he cellula 3D agg egoid cul u es. The in i o
beha iou was ollowed by ime-lapse mic oscopy (Leica
mic oscope - DMI6000B wi h Adap a i e Focus Con ol)
du ing 18 h. Fu he , he images aken we e con e ed
in o a ime-lapse mo ie, and quan i a i e analysis o he
numbe and leng h o p o usi e s uc u es was pe -
o med using Leica AF6000 so wa e.
Immuno luo escence analysis
Cells we e cul u ed on glass co e slips, ixed wi h 4%
pa a o maldehyde (20 min), ea ed wi h NH
4
Cl (50 mM)
o 10 min, washed wi h PBS, and pe meablilized wi h
0.1% T i on X-100 in PBS o 5 min, a oom empe a u e.
Non-speci ic binding was blocked by ea men wi h 5%
BSA in PBS, o 30 min a oom empe a u e. Cells we e
hen s ained O ange-phalloidin (Sigma, S . Louis, MO, USA)
in o de o isualize F-ac in, o 30 min, o wi h he mono-
clonal mouse an ibody o p120c n (dilu ion 1:100) and
Ribei o e al. Cell Communica ion and Signaling (2018) 16:75 Page 3 o 16
monoclonal abbi an ibody o E-cadhe in (dilu ion
1:100) o 1 h, ollowed by conjuga ed goa an i-mouse
and an i- abbi seconda y IgG (Dako Cy oma ion, Ca pin-
e ia, CA, USA), o 1 h a oom empe a u e. A e an-
o he wash wi h PBS, each sample was moun ed wi h
Vec ashield (Vec o Labo a o ies, Inc., Bu lingame, CA,
USA) con aining 4,6-diamidine-2-phenylindolendihyd o-
ch olide (DAPI) and isualized wi h Leica DM 2000
mic oscope (Leica).
Ma igel in asion assays
Ma igel in asion assay was pe o med using 8 μm po e
size BD BioCoa ™Ma igel In asion Chambe s (BD Bio-
sciences), acco ding o manu ac u e ’s ecommenda-
ions. B ie ly, 24 h a e siRNA ans ec ion, 5 × 10
4
cells
we e added o he uppe compa men o he chambe ,
wi h o wi hou dasa inib ea men . In he lowe com-
pa men , only esh medium supplemen ed wi h 10%
FBS was added. A e 24 h (BT20) o 48 h (MCF-7/AZ)
o incuba ion a 37 °C, he uppe su ace o he il e was
clea ed om non-in asi e cells wi h a co on swab and
washed wi h PBS. The in asi e cells we e ixed wi h cold
me hanol and s ained wi h DAPI. In asi e cells we e
sco ed by coun ing he cells in he il e wi h a luo es-
cence mic oscope (Zeiss Image Z1), a 100× o magni i-
ca ion. In asion index exp esses he a io be ween
in asi e cells om he di e en condi ions ela i e o
he con ol cells.
Wound-healing assay
A e ans ec ion, o 24 h, wi h he espec i e siRNAs,
cells we e ipsinized and seeded o ano he 24 h, o
allow he cells o o m a comple e monolaye . A e hey
became con luen , wounds we e ca e ully in oduced
ac oss he cell monolaye , so ha he su ounding cells
we e no dis u bed. The medium was eplaced wi h esh
comple e medium, wi h and wi hou dasa inib ea men .
Pic u es we e aken a 100× o magni ica ion in an
in e ed mic oscope (Zeiss), and measu emen s o he
dis ance be ween he wound edges we e made a speci ic
ime poin s.
Zymog aphy
The condi ioned medium collec ed om he se e al cell
cul u es, which we e g own 6-well pla es coa ed wi h
collagen ype I, was analysed o p o einases ac i i y
using β-casein zymog aphy. Samples we e mixed wi h
sample bu e [0.03% b omophenol blue, 0.25 M
T is-HCl pH 6.8, 10% SDS (w/ ) and 4% suc ose (w/ )]
and elec opho esed, unde non- educing condi ions, on
10% polyac ilamide gels con aining 0.1% (w/ ) β-casein
om bo ine milk (Sigma). A e elec opho esis, gels
we e washed wice, o 30 min, in 2% ( / ) T i on X-100
(Sigma) a oom empe a u e, in o de o emo e SDS.
β-casein gels we e hen incuba ed in a subs a e eac ion
bu e [0.2 M NaCl, 5 mM CaCl
2
, 1% ( / ) T i on X-100
in 50 mM T is-HCl, pH 7.4], du ing 72 h, and inally
s ained wi h Coomassie Blue S aining Solu ion [0.1% (w/
) Coomassie Blue R250 in 10% ( / ) ace ic acid and
40% ( / ) me hanol], o 25 min. The gels des aining
was pe o med in a solu ion wi h 20% me hanol and
10% ace ic acid, un il bands s a o become isible.
Enzyma ic ac i i y was isualized as a clea band agains
he blue backg ound o s ained casein gels, and MMPs
we e iden i ied by hei molecula weigh . Quan i ica ion
o band densi y was ca ied ou using he Quan i y One
so wa e ( e sion 4.0, BioRad, He cules, CA, USA).
In si u p oximi y liga ion assay (PLA)
Cells we e deposi ed on glass slides and ixed wi h
me hanol o 10 min. PLA was pe o med using he
Duolink ki (Olink Bioscience, Sweden), acco ding o
manu ac u e ’s ecommenda ions. The ollowing com-
bina ion o p ima y an i-human an ibodies we e used
agains : P-cadhe in (dilu ion 1:50, abbi polyclonal IgG,
Cell Signaling), E-cadhe in (dilu ion 1:100, mouse mono-
clonal IgG1, clone HECD-1, Taka a Bio Inc., Shiga,
Japan; o dilu ion 1:50, abbi monoclonal IgG, Cell
Signaling), p120c n (dilu ion 1:100, mouse monoclonal
IgG1, clone 98, BD Biosciences). The nuclei we e coun-
e s ained wi h DAPI. Slides we e analyzed wi h luo es-
cence mic oscopy (Zeiss Image Z1 mic oscope) o
isualiza ion o b igh luo escen ed signals consis en
wi h p o ein-p o ein in e ac ion e en s. Ten s acks pe
image we e aken and a minimum o i e ields pe
expe imen we e e alua ed. The Blob inde V3.2 ee
so wa e (Cen e o Image Analysis, Uppsala, Sweden)
was used o quan i y he numbe o blobs (o do s)
p esen in each condi ion. Fo cell cul u e expe imen s,
an a e age o blobs/cells o a leas h ee independen
expe imen s was pe o med and was no malized o he
con ol condi ion. Fo pa a in embedded umo s, he
speci ic numbe o blobs/cell was analysed. The espec -
i e nega i e con ols we e used in each expe imen .
In e nuclea p o ile analysis
A so wa e applica ion was de eloped in o de o gene -
a e and analyze he in e nuclea p o ile o p o eins. In
each immuno luo escen image, pai s o cells we e
selec ed o analysis in a semi-au oma ed manne . The
mapping and quan i ica ion o he p o ein exp ession
le el was pe o med by compu ing, espec i ely, 1D IN
and RD in ensi y p o iles o wo con iguous cells and
wi hin one single cell. To cope wi h cell size and shape
a iabili y, a geome ic compensa ion algo i hm was
de eloped in a Bayesian amewo k. The me hod was
designed as an i e a i e algo i hm composed by he ol-
lowing s eps: (i) p o ile ex ac ion om selec ed single
Ribei o e al. Cell Communica ion and Signaling (2018) 16:75 Page 4 o 16
cells (in case o RD) o pai s o cells (in case o IN); (ii)
image map building by s acking luo escence p o iles
oge he in columns a e leng h no maliza ion; (iii)
denoising o map image (in which mul iplica i e noise
desc ibed by a Poisson dis ibu ion is assumed); (i ) geo-
me ic compensa ion o each 1D column p o ile minim-
izing he o e all a iabili y o he map along he lines
(ho izon al di ec ion); and ( ) compu a ion o he
a e age and s anda d de ia ion p o iles using he com-
pensa ed map. A e he ex ac ion o he da a, he max-
imum mean a io (MMR) pa ame e was calcula ed
di iding he maximum luo escence alue (nume a o )
by he luo escence mean (denomina o ).
Animal s udies
N:NIH(S)II-nu/nu mice, s ain p oduced by Ipa imup,
we e housed, b eeded and main ained a he CIM-FMUP
Animal House, in a pa hogen- ee en i onmen unde
con olled condi ions o ligh and humidi y.The expe i-
men s consis ed on he o ho opic injec ion in he mam-
ma y a pad o emale mice, wi h 6–8 weeks o age,
wi h 2 × 10
6
cells om each P-cadhe in-o e exp essing
b eas cance models (SUM149PT, n= 17; BT20, n= 15;
MDA-MB-468, n= 13) using a 25G needle. When he
induced p ima y umou s eached a median olume o
100 mm
3
, mice we e andomized in o wo g oups, and
s a ed he o al daily ea men wi h dasa inib, whe eas
he o he s we e ea ed jus wi h he d ug ehicle, un il
he end o he expe imen . Doses o dasa inib o be ad-
minis e ed we e p epa ed weekly by dilu ion in ci a e
bu e 80 mM (pH = 3.1). The ci a e bu e was p e-
pa ed as ollows: sodium ci a e powde was dilu ed in
ci ic acid solu ion 1 M, ob ained by dissol ing ci ic
acid in deionized wa e , o a inal concen a ion o
80 nM and wi h pH = 3.1. The ci a e bu e solu ion
was made e e y 3 weeks, il e ed in low chambe and
kep a 4 °C. E e y week, daily doses o dasa inib we e
p epa ed o a inal concen a ion o 10 mg/kg, in o de
o ea mice o bo h expe imen al models. Mice (6–8
pe g oup) we e weigh ed, and umo wid h and leng h
we e measu ed wi h calipe s, a leas , once a week.
Tumo olume was es ima ed by using he equa ion, V
=0.5×a×b
2
, whe e V is he olume, a is he leng h o he
majo axis o he umo , and b is he leng h o i s mino
axis. Tumo s we e su gically emo ed when an a e age
olume o 1000 mm
3
was eached as p e iously de-
sc ibed. Mice we e main ained o e alua e he impac o
dasa inib in mice o e all su i al, o a maximum o
210 days, unless hey showed any signs o disease p o-
g ession. Fo ha , mice we e eu hanized acco dingly o
he es ablished Endpoin s. Upon ex ac ion, P ima y u-
mo s we e ixed in 10% bu e ed o malin and hen em-
bedded in pa a in, sec ioned, and s ained wi h
haema oxylin and eosin. Immunohis ochemis y o E-
and P-cadhe in was pe o med in umo s sec ions.
Immunohis ochemis y
A se ies o 467 p ima y in asi e b eas ca cinomas
we e e ie ed om he Pa hology Depa men , Hos-
pi al Xe al-Cíes, Vigo, Spain, diagnosed be ween
1978 and 1992. The umo s ha e been cha ac e ized
o clinical and pa hological ea u es as p e iously
desc ibed [11].
IHC was pe o med as p e iously desc ibed [6]. The
memb ane s aining ex ension was e alua ed on a scale
o 0 o 3 (0–0% o 10%; 1–10% o 25%; 2–25% o 50%;
3–50% o 100%). Fo E-cadhe in and o al S c, sco es o
0 and 1 we e conside ed nega i e and immuno eac i i y
o 2 and 3 was sco ed as posi i e. Fo P-cadhe in all
umo s p esen ing an unequi ocal memb anous s aining,
in a leas 10% o he neoplas ic cells (≥1), we e sco ed
as posi i e. Fo pSRC(Ty 416) umo s p esen ing
memb ane o cy oplasmic s aining, in a leas 10% o he
neoplas ic cells (≥1), we e sco ed as posi i e. Some sam-
ples could no be assessed due o co e alling o lack o
enough biological ma e ial o analyze.
S a is ical analyses
All o he s a is ical analyses we e pe o med by G aph
Pad P ism e sion 5.0c so wa e (G aph Pad So wa e,
San Diego, CA, USA), unless o he wise s a ed. P alues
less han 0.05 we e conside ed s a is ically signi ican .
Fo he AFM analysis, S uden ’s - es s we e used o
de e mine s a is ically signi ican di e ences. F equency
his og ams we e pe o med in O igin (O iginLab, No h-
amp on, MA, USA) and Gaussian cu es we e applied.
Quan i a i e pa ame e s o In e nuclea p o iles (no -
malized o a cons an leng h o 100 a bi a y uni s) in
P-cadhe in cells we e analyzed using a Mann-Whi ney
es wi h a Bon e oni co ec ion.
Fo he in i o xenog a assays, ANOVA es was
used o e alua e he di e ences in umo size, conside -
ing a con idence in e al o 95%. Su i al cu es we e
es ima ed by he Kaplan-Meie ’s me hod, using he
log- ank es o assess signi ican di e ences o mice
o e all su i al.
Conce ning he unc ional in i o assays, all we e pe -
o med independen ly and in iplica e.
Fo s a is ical analysis o he immunohis ochemis y
esul s, con ingency ables and chi-squa e es s we e
pe o med by SPSS 15.0 so wa e package o Win-
dows (SPSS, Inc., USA), o es ima e he ela ionship
be ween s aining pa e ns o P-cadhe in and pSRC
(Ty 416).
All s a is ical es s we e wo-sided.
Ribei o e al. Cell Communica ion and Signaling (2018) 16:75 Page 5 o 16

Fig. 1 (See legend on nex page.)
Ribei o e al. Cell Communica ion and Signaling (2018) 16:75 Page 6 o 16
Resul s
P-cadhe in/CDH3 o e exp ession is signi ican ly
associa ed wi h SRC ac i a ion in human b eas cance
cells
In o de o con i m he associa ion be ween P-cadhe in
and SRC ac i a ion, we s a ed by analysing he exp es-
sion o CDH3 (P-cadhe in codi ying gene) and SRC asso-
cia ed genes (SRC, YES1, FYN, LYN, LCK, CSK)ina
public da abase o b eas cance cell lines belonging o
di e en molecula sub ypes (Fig. 1a) [CCLE da abase
[20]]. In e es ingly, cell lines showing he highes le els
o CDH3 also p esen a signi ican ly inc eased exp ession
o YES, SRC, LYN and FYN. Acco dingly, mos o hese
cell lines we e classi ied as Basal A, which a e he ones
ha be e esemble he BLBC sub ype (Fig. 1b-c).
We hen alida ed he associa ion be ween P-cadhe in
and SRC ac i a ion using wo es ablished in i o b eas
cance cell models, whe e P-cadhe in exp ession was
manipula ed. P-cadhe in o e exp essing cells showed
inc eased le els o pSRC(Ty 416), while P-cadhe in
silencing led o a signi ican dec ease o pSRC(Ty 416)
le els (Fig. 1d and Addi ional ile 1: Figu e S1A and B).
No al e a ions in he o al le els o SRC we e ound by
Wes e n blo . E-cadhe in and p120c n we e also e alu-
a ed in bo h models and he le el o exp ession and
localiza ion o bo h p o eins we e analysed by he luo -
escence in e nuclea p o iles [21]. In con as o wha
was obse ed o E-cadhe in exp ession, which did no
al e upon gene ic manipula ion o P-cadhe in, a de-
c ease o p120c n memb ane in ensi y was obse ed in
P-cadhe in o e exp essing cells (Fig. 1e- ). Mo eo e , a
signi ican inc ease in Rac1 ac i i y, a downs eam a ge
o cy oplasmic p120c n, was obse ed in cells wi h
P-cadhe in o e exp ession, as well as he silencing o
P-cadhe in in BT20 cells was accompanied by a signi i-
can dec ease in Rac1 ac i a ion (Fig. 1g). Cell-cell adhe-
sion s eng h was also clea ly comp omised in
P-cadhe in o e exp essing cells, as demons a ed by
(See igu e on p e ious page.)
Fig. 1 P-cadhe in o e exp ession associa es wi h SRC ac i a ion co ela ing wi h an agg essi e beha io . aGene exp ession p o ile o SRC amily
membe s om BCC lines e ie ed om he CCLE da abase [20]. Unsupe ised hie a chical clus e ing based on classical cadhe ins exp ession
(CDH1,CDH2,CDH3), subdi ide he BCC lines in o hei molecula sub ype (Luminal, Basal A and Basal B). bBox-plo ep esen a ion o CDH1,
CDH2 and CDH3 gene exp ession alues in he di e en b eas cance molecula sub ypes. cBox-plo o SRC associa ed genes exp ession in he
Luminal and Basal A BCC sub ypes. P- alues indica e signi ican ly di e ences obse ed in gene exp ession p o ile om Luminal and Basal A BCC
lines. dWes e n blo analysis o P-cadhe in, E-cadhe in, p120c n, pSRC(Ty 416), o al SRC and β-ac in p o eins a e P-cadhe in o e exp ession and
P-cadhe in silencing, espec i ely. eDual immuno luo escence con ocal images using an an i-p120c n an ibody (g een), F-ac in ( ed) and DAPI
(blue). In e nuclea p o ile o p120c n and E-cadhe in, ob ained h ough he compu a ional analysis o immuno luo escence images
om MCF-7/AZ BCC model. P- alue indica e signi ican di e ences obse ed in he memb ane mean luo escen in ensi y o p120c n
be ween MCF-7/AZ.Mock e sus MCF-7/AZ.Pcad cells. gQuan i ica ion o he ac i e Rac1-GTP, measu ed by a G-Lisa assay a e P-cadhe in
o e exp ession and P-cadhe in silencing, espec i ely. hA e age alues o Wo k (J), ep esen ing he cell-cell adhesion s eng h, using
AFM Fo ce Spec oscopy analysis. S uden ’s - es was used o de e mine s a is ically signi ican di e ences, and P- alues a e indica ed in he igu e.
iRep esen a i e Images o ime-lapse mic oscopy o BCC sphe oids embedded in collagen ype I. Scale ba = 50 μm. jBox-plo quan i ica ion o he
ex ension (μm) o in asi e p o usions. kBox-plo quan i ica ion o he numbe o in asi e p o usions. lRep esen a ion o he pe cen age o cases
posi i e o bo h pSRC(Ty 416) and P-cadhe in exp ession. P- alue was calcula ed by Chi-squa e es in o de o es ima e he ela ionship be ween
s aining pa e ns o P-cadhe in and pSRC(Ty 416)
Table 1 Cha ac e iza ion o he umo se ies, conse ning umo
pa hological ea u es and molecula ma ke s
N%
His ological g ade 1 75 18.0
2 117 28.1
3 204 49.0
Missing 20 4.8
ERαNega i e 140 33.7
Posi i e 274 65.9
Missing 2 0.5
HER2 Nega i e 352 84.6
Posi i e 60 14.4
Missing 4 1.0
EGFR Nega i e 395 95.0
Pos i e 21 5.0
Missing 0 0.0
CKS Nega i e 357 85.8
Posi i e 59 14.2
Missing 0 0.0
CK14 Nega i e 394 94.7
Pos i e 22 5.3
Missing 0 0.0
Vimen in Nega i e 345 82.9
Pos i e 69 16.6
Missing 2 0.5
CD44 Nega i e 205 49.3
Posi i e 210 50.5
Missing 1 0.2
Sub ype Luminal 303 72.8
He 2 29 7.0
T iple-Nega i e 84 20.2
Missing 0 0.0
Ribei o e al. Cell Communica ion and Signaling (2018) 16:75 Page 7 o 16
Fig. 2 (See legend on nex page.)
Ribei o e al. Cell Communica ion and Signaling (2018) 16:75 Page 8 o 16
a omic o ce mic oscopy (AFM) measu emen s (Fig. 1h),
as well as a signi ican inc ease in he p o usi e and
in asi e beha io (Fig. 1i-kand Addi ional ile 1: Figu e
S1C) o b eas cance cells was p omo ed, by he e alu-
a ion o 3D sphe oids embedded in collagen ype I
ma ix by ime-lapse mic oscopy.
These indings, oge he wi h he ac ha P-cadhe in
a ec s he E-cadhe in/p120c n complex memb ane
s abiliza ion, and ha i s jux amemb ane domain is c u-
cial o i s p o-in asi e unc ion [11], led us o p esume
ha P-cadhe in could be modi ying cell-cell adhesion
and in asion capaci y h ough p120c n delocaliza ion
occu ing downs eam o SRC ac i a ion signaling.
Finally, o alida e in i o he associa ion be ween
pSRC(Ty 416) and P-cadhe in, a se ies o 416 p ima y
b eas umo s was s udied by immunohis ochemis y
(Table 1). A signi ican associa ion was ound be ween
he p esence o bo h p o eins, whe e 44% (18 in 41
cases) o he umo s wi h pSRC(Ty 416) we e also posi-
i e o P-cadhe in (Fig. 1l).
SRC ac i i y supp ession wi h dasa inib inhibi s he in
i o unc ional ac i i y induced by P-cadhe in
o e exp ession
Taking in o accoun he esul s showing a posi i e asso-
cia ion be ween P-cadhe in exp ession and SRC ac i a-
ion, we specula ed i he downs eam SRC signalling
could be an e ec i e a ge o poo p ognos ic
P-cadhe in-o e exp essing b eas cance . This s a egy
would o e come he lack o FDA-app o ed d ugs
di ec ly a ge ing P-cadhe in.
Thus, we examined he e ec o SRC inhibi ion in he
p e iously de ined in i o unc ional ac i i y induced by
P-cadhe in exp ession, by ea ing cells wi h 100 nM o
dasa inib (concen a ion ha does no a ec he me abolic
s a e and iabili y o b eas cance cells –Addi ional ile 1:
Figu e S1D). B eas cance cells ea ed wi h dasa inib
showed a dec ease o pSRC(Ty 416) exp ession (Fig. 2a
and Addi ional ile 1: Figu e S1E); mo eo e , his ea -
men led o a dec ease in cell mig a ion (Fig. 2b) and cell
in asion (Fig. 2c), associa ed wi h a dec ease o MMP2
ac i i y and sP-cad exp ession in he condi ioned media
(Fig. 2d), as well as wi h a dec eased mammosphe e o m-
ing e iciency (MFE) (Fig. 2e).
Addi ionally, cell-cell adhesion was highly p omo ed in
P-cadhe in o e exp essing cells ea ed wi h dasa inib, as
obse ed by he b igh ield images and quan i ied by
AFM in bo h cell lines a e SRC inhibi ion (Fig. 2 ).
Addi ionally, when we e alua ed he in i o 3D in asion
abili y o P-cadhe in o e exp essing cells a e dasa inib
ea men , we obse ed a signi ican dec ease in he
numbe and leng h o p o usi e and in asi e s uc u es
in o he collagen ype I ma ix (Fig. 2g-i). As a con-
ol, we ha e pe o med dasa inib ea men in
P-cadhe in low b eas cance cells (MCF-7/AZ.mock
and BT20 siC ) and, as expec ed, he impac o he
dasa inib ea men is e y modes , since he pa hway
is no as ac i a ed as in P-cadhe in o e exp essing
cells (Addi ional ile 1:Figu eS2A-C)[22].
Al oge he , he esul s suppo ha SRC ac i i y sup-
p ession by dasa inib is able o p e en he unc ional e -
ec s induced by P-cadhe in o e exp ession in b eas
cance cells.
Dasa inib ea men p omo es he eco e y o cell-cell
adhesion unc ion by s abilizing E-cadhe in/p120c n
complex a he cell memb ane
We ha e p e iously shown ha P-cadhe in o e exp es-
sion in an E-cadhe in wild- ype con ex weakens cell-cell
adhesion by dis up ing he E-cadhe in/ca enins complex
[11]. Taking his in o accoun , we e alua ed he e ec o
dasa inib a his molecula le el. As obse ed in Fig. 3a,
dasa inib p omo ed an inc ease in memb ane exp ession
o p120c n, as obse ed by i s in e nuclea p o ile
(Fig. 3b), which was associa ed wi h a dec ease in he
downs eam ac i a ion o Rac 1 (Fig. 3c). Once again,
E-cadhe in exp ession and localiza ion was no al e ed
(Fig. 3b), al hough dasa inib ea men p omo ed a
signi ican eco e o he E-cadhe in/p120c n complex,
mainly a he cell memb ane (Fig. 3d and Addi ional
ile 1: Figu e S3A and B). Bo h esul s show ha ac i-
a ion o SRC signaling in e e es wi h E-cadhe in/
(See igu e on p e ious page.)
Fig. 2 Dasa inib ea men inhibi s he in i o unc ional ac i i y induced by P-cadhe in exp ession. aWes e n blo ing o pSRC(Ty 416), o al S c,
P-cadhe in, E-cadhe in and p120c n in P-cadhe in o e exp essing cells a e dasa inib ea men (100 nM) o 48 h. P o ein le els o β-ac in we e
analyzed and used as he loading con ol. bRep esen a i e expe imen om a wound healing mig a ion assay, in bo h P-cadhe in o e exp essing
BCC (MCF-7/AZ.Pcad and BT20), ea ed wi h 100 nM o dasa inib o DMSO o 24 h. cFold change in he numbe o in asi e cells, e alua ed
by he ma igel in asion assay o bo h MCF-7/AZ.Pcad and BT20 ea ed wi h DMSO o 100 nM o dasa inib. dZymog aphy o MMP2 ac i i y
and Wes e n blo ing o sP-cad, using he condi ioned medium om cells ea ed wi h DMSO o 100 nM o dasa inib o 48 h, in bo h
MCF-7/AZ.Pcad and BT20 BCC models. eMammosphe e o ming assay was pe o med o bo h cell models. The P- alues indica e he
s a is ically signi ican di e ence be ween DMSO and dasa inib ea ed cells, in bo h P-cadhe in BCC models. A e age alues o Wo k
(J), ep esen ing he cell-cell adhesion s eng h, o bo h BCC cance cell models ea ed wi h DMSO and dasa inib, using AFM Fo ce Spec oscopy
analysis. gBox-plo quan i ica ion o he leng h (μm) and hnumbe o in asi e p o usions o MCF-7/AZ.Pcad sphe oids, ea ed wi h DMSO o
100 nM dasa inib, o 24 h. P- alues < 0.05 we e conside ed s a is ically signi ican . iRep esen a i e Images o ime-lapse mic oscopy o BCC sphe oids
embedded in collagen ype I, o MCF-7/AZ.Pcad ea ed wi h DMSO and dasa inib. Scale ba = 50 μm
Ribei o e al. Cell Communica ion and Signaling (2018) 16:75 Page 9 o 16
21. Sanches JM, Figuei edo J, Fonseca M, Du aes C, Melo S, Esmenio S, Se uca
R. Quan i ica ion o mu an E-cadhe in using bioimaging analysis o in si u
luo escence mic oscopy. A new app oach o CDH1 missense a ian s. Eu J
Hum Gene . 2015;23:1072–9.
22. Ribei o AS, Ca alho FA, Figuei edo J, Ca alho R, Mes e T, Mon ei o J,
Guedes AF, Fonseca M, Sanches J, Se uca R, e al. A omic o ce mic oscopy
and g aph analysis o s udy he P-cadhe in/SFK mechano ansduc ion
signalling in b eas cance cells. Nanoscale. 2016;8:19390–401.
23. Huang F, Ree es K, Han X, Fai child C, Pla e o S, Wong TW, Lee F, Shaw P,
Cla k E. Iden i ica ion o candida e molecula ma ke s p edic ing sensi i i y
in solid umo s o dasa inib: a ionale o pa ien selec ion. Cance Res. 2007;
67:2226–38.
24. Lomba do LJ, Lee FY, Chen P, No is D, Ba ish JC, Behnia K, Cas aneda S,
Co nelius LA, Das J, Doweyko AM, e al. Disco e y o N-(2-chlo o-6-me hyl-
phenyl)-2-(6-(4-(2-hyd oxye hyl)- pipe azin-1-yl)-2-me hylpy imidin-4-
ylamino) hiazole-5-ca boxamide (BMS-354825), a dual S c/Abl kinase
inhibi o wi h po en an i umo ac i i y in p eclinical assays. J Med Chem.
2004;47:6658–61.
25. Wang XD, Ree es K, Luo FR, Xu LA, Lee F, Cla k E, Huang F. Iden i ica ion o
candida e p edic i e and su oga e molecula ma ke s o dasa inib in
p os a e cance : a ionale o pa ien selec ion and e icacy moni o ing.
Genome Biol. 2007;8:R255.
26. Pa edes J, Co eia AL, Ribei o AS, Milanezi F, Cameselle-Teijei o J, Schmi FC.
B eas ca cinomas ha co-exp ess E- and P-cadhe in a e associa ed wi h
p120-ca enin cy oplasmic localisa ion and poo pa ien su i al. J Clin
Pa hol. 2008;61:856–62.
27. Pa edes J, Milanezi F, Reis-Filho JS, Lei ao D, A hanazio D, Schmi F.
Abe an P-cadhe in exp ession: is i associa ed wi h es ogen-independen
g ow h in b eas cance ? Pa hol Res P ac . 2002;198:795–801.
28. Se els A, Canel M, B un on VG, F ame MC. S c/FAK-media ed egula ion o
E-cadhe in as a mechanism o con olling collec i e cell mo emen :
insigh s om in i o imaging. Cell Adhes Mig . 2011;5:360–5.
29. Idoux-Gille Y, Nassou M, Lakis E, Bonini F, Theille C, Du Manoi S, Sa agne
P. Slug/Pcad pa hway con ols epi helial cell dynamics in mamma y gland
and b eas ca cinoma. Oncogene. 2018;37:578–88.
30. Bay ak a S, Gluck S. Molecula ly a ge ed he apies o me as a ic iple-
nega i e b eas cance . B eas Cance Res T ea . 2013;138:21–35.
31. Finn RS. Ta ge ing S c in b eas cance . Ann Oncol. 2008;19:1379–86.
32. He old CI, Chada am V, Pe e son BL, Ma com PK, Hopkins J, Kimmick GG,
Fa a o J, Hamil on E, Welch RA, Bacus S, Blackwell KL. Phase II ial o
dasa inib in pa ien s wi h me as a ic b eas cance using eal- ime
pha macodynamic issue bioma ke s o S c inhibi ion o escala e dosing.
Clin Cance Res. 2011;17:6061–70.
33. Somlo G, A zo i F, S auss LC, Geese WJ, Spech JM, G adisha WJ, Rybicki A,
Sy O, Vahda LT, Co es J. Dasa inib plus capeci abine o ad anced b eas
cance : sa e y and e icacy in phase I s udy CA180004. Clin Cance Res. 2013;
19:1884–93.
34. He land A, an de Mee AD, Fi zGe ald EA, Pa k TE, Sleeboom JJ, Ingbe DE.
Dis inc con ibu ions o as ocy es and Pe icy es o Neu oin lamma ion
iden i ied in a 3D human blood-b ain ba ie on a Chip. PLoS One.
2016;11:e0150360.
35. Mo on JP, Ka im SA, G aham K, Timpson P, Jamieson N, A hineos D, Doyle
B, McKay C, Heung MY, Oien KA, e al. Dasa inib inhibi s he de elopmen
o me as ases in a mouse model o panc ea ic duc al adenoca cinoma.
Gas oen e ology. 2010;139:292–303.
36. Summy JM, Gallick GE. S c amily kinases in umo p og ession and
me as asis. Cance Me as asis Re . 2003;22:337–58.
37. Yea man TJ. A enaissance o SRC. Na Re Cance . 2004;4:470–80.
38. Canel M, Se els A, F ame MC, B un on VG. E-cadhe in-in eg in c oss alk in
cance in asion and me as asis. J Cell Sci. 2013;126:393–401.
Ribei o e al. Cell Communica ion and Signaling (2018) 16:75 Page 16 o 16