9
ORIGINAL ARTICLE A truncated phosphorylated p130Cas substrate domain is sufficient to drive breast cancer growth and metastasis formation in vivo Joerg Kumbrink 1,2,3 & Ana de la Cueva 1 & Shefali Soni 1,4 & Nadja Sailer 1,5 & Kathrin H. Kirsch 1 Received: 25 September 2015 /Accepted: 20 January 2016 # International Society of Oncology and BioMarkers (ISOBM) 2016 Abstract Elevated p130Cas (Crk-associated substrate) levels are found in aggressive breast tumors and are associated with poor prognosis and resistance to standard therapeutics in pa- tients. p130Cas signals majorly through its phosphorylated substrate domain (SD) that contains 15 tyrosine motifs (YxxP) which recruit effector molecules. Tyrosine phosphor- ylation of p130Cas is important for mediating migration, in- vasion, tumor promotion, and metastasis. We previously de- veloped a Src*/SD fusion molecule approach, where the SD is constitutively phosphorylated. In a polyoma middle T-antigen (PyMT)/Src*/SD double-transgenic mouse model, Src*/SD accelerates PyMT-induced tumor growth and promotes a more aggressive phenotype. To test whether Src*/SD also drives metastasis and which of the YxxP motifs are involved in this process, full-length and truncated SD molecules fused to Src* were expressed in breast cancer cells. The functionality of the Src*/SD fragments was analyzed in vitro, and the active proteins were tested in vivo in an orthotopic mouse model. Breast cancer cells expressing the full-length SD and the func- tional smaller SD fragment (spanning SD motifs 610) were injected into the mammary fat pads of mice. The tumor pro- gression was monitored by bioluminescence imaging and cal- iper measurements. Compared with control animals, the com- plete SD promoted primary tumor growth and an earlier onset of metastases. Importantly, both the complete and truncated SD significantly increased the occurrence of metastases to multiple organs. These studies provide strong evidence that the phosphorylated p130Cas SD motifs 610 (Y236, Y249, Y267, Y287, and Y306) are important for driving mammary carcinoma progression. Keywords p130Cas . Breastcancer . Substratedomain . Src*/ SD . Metastasis . Bioluminescence imaging . Mouse tumor model Introduction p130 Crk-associated substrate (p130Cas) 1 is a scaffold protein that integrates large multi-protein complexes in response to growth factors, hormones, and integrin signaling [1, 2]. p130Cas contains a central substrate domain (SD) where 15 YxxP repeats serve as phosphorylation sites for tyrosine ki- nases. Through its interaction with Src family members, focal adhesion kinase (FAK) and phosphoinositide 3-kinase (PI3K), 1 The abbreviations used are the following: aa, amino acid; abs, antibod- ies; BLI, bioluminescence imaging; DOX, doxycycline; FAK, focal ad- hesion kinase; IB, immunoblotting; KM, kinase inactive; MFP, mammary fat pad; p130Cas, p130 Crk-associated substrate; Src*, attenuated kinase activity; SD, substrate domain; WCE, whole cell extract. Electronic supplementary material The online version of this article (doi:10.1007/s13277-016-4902-8) contains supplementary material, which is available to authorized users. * Kathrin H. Kirsch [email protected] 1 Department of Biochemistry, Boston University School of Medicine, 72 E. Concord Street, Boston, MA 02118, USA 2 Department of Medicine III, University Hospital Grosshadern, University of Munich, Munich, Germany 3 Present address: Institute of Pathology, University of Munich, Munich, Germany 4 Present address: The Leona M. and Harry B. Helmsley Charitable Trust, New York, NY 10169, USA 5 Department of Biochemistry, University of Munich, Munich, Germany Tumor Biol. DOI 10.1007/s13277-016-4902-8

A truncated phosphorylated p130Cas substrate domain is ... · ORIGINAL ARTICLE A truncated phosphorylated p130Cas substrate domain is sufficient to drive breast cancer growth and

  • Upload
    others

  • View
    6

  • Download
    0

Embed Size (px)

Citation preview

Page 1: A truncated phosphorylated p130Cas substrate domain is ... · ORIGINAL ARTICLE A truncated phosphorylated p130Cas substrate domain is sufficient to drive breast cancer growth and

ORIGINAL ARTICLE

A truncated phosphorylated p130Cas substrate domainis sufficient to drive breast cancer growth and metastasisformation in vivo

Joerg Kumbrink1,2,3& Ana de la Cueva1 & Shefali Soni1,4 & Nadja Sailer1,5 &

Kathrin H. Kirsch1

Received: 25 September 2015 /Accepted: 20 January 2016# International Society of Oncology and BioMarkers (ISOBM) 2016

Abstract Elevated p130Cas (Crk-associated substrate) levelsare found in aggressive breast tumors and are associated withpoor prognosis and resistance to standard therapeutics in pa-tients. p130Cas signals majorly through its phosphorylatedsubstrate domain (SD) that contains 15 tyrosine motifs(YxxP) which recruit effector molecules. Tyrosine phosphor-ylation of p130Cas is important for mediating migration, in-vasion, tumor promotion, and metastasis. We previously de-veloped a Src*/SD fusion molecule approach, where the SD isconstitutively phosphorylated. In a polyoma middle T-antigen(PyMT)/Src*/SD double-transgenic mouse model, Src*/SDaccelerates PyMT-induced tumor growth and promotes a moreaggressive phenotype. To test whether Src*/SD also drivesmetastasis and which of the YxxP motifs are involved in thisprocess, full-length and truncated SD molecules fused to Src*were expressed in breast cancer cells. The functionality of theSrc*/SD fragments was analyzed in vitro, and the active

proteins were tested in vivo in an orthotopic mouse model.Breast cancer cells expressing the full-length SD and the func-tional smaller SD fragment (spanning SD motifs 6–10) wereinjected into the mammary fat pads of mice. The tumor pro-gression was monitored by bioluminescence imaging and cal-iper measurements. Compared with control animals, the com-plete SD promoted primary tumor growth and an earlier onsetof metastases. Importantly, both the complete and truncatedSD significantly increased the occurrence of metastases tomultiple organs. These studies provide strong evidence thatthe phosphorylated p130Cas SD motifs 6–10 (Y236, Y249,Y267, Y287, and Y306) are important for driving mammarycarcinoma progression.

Keywords p130Cas .Breastcancer .Substratedomain .Src*/SD .Metastasis . Bioluminescence imaging .Mouse tumormodel

Introduction

p130 Crk-associated substrate (p130Cas)1 is a scaffold proteinthat integrates large multi-protein complexes in response togrowth factors, hormones, and integrin signaling [1, 2].p130Cas contains a central substrate domain (SD) where 15YxxP repeats serve as phosphorylation sites for tyrosine ki-nases. Through its interaction with Src family members, focaladhesion kinase (FAK) and phosphoinositide 3-kinase (PI3K),

1 The abbreviations used are the following: aa, amino acid; abs, antibod-ies; BLI, bioluminescence imaging; DOX, doxycycline; FAK, focal ad-hesion kinase; IB, immunoblotting; KM, kinase inactive; MFP, mammaryfat pad; p130Cas, p130 Crk-associated substrate; Src*, attenuated kinaseactivity; SD, substrate domain; WCE, whole cell extract.

Electronic supplementary material The online version of this article(doi:10.1007/s13277-016-4902-8) contains supplementary material,which is available to authorized users.

* Kathrin H. [email protected]

1 Department of Biochemistry, Boston University School of Medicine,72 E. Concord Street, Boston, MA 02118, USA

2 Department of Medicine III, University Hospital Grosshadern,University of Munich, Munich, Germany

3 Present address: Institute of Pathology, University of Munich,Munich, Germany

4 Present address: The Leona M. and Harry B. Helmsley CharitableTrust, New York, NY 10169, USA

5 Department of Biochemistry, University of Munich,Munich, Germany

Tumor Biol.DOI 10.1007/s13277-016-4902-8

Page 2: A truncated phosphorylated p130Cas substrate domain is ... · ORIGINAL ARTICLE A truncated phosphorylated p130Cas substrate domain is sufficient to drive breast cancer growth and

it recruits the adaptor proteins Crk and Nck to its phosphory-lated SD. These complexesmodulate signal transduction path-ways that control cell survival, proliferation, adhesion, migra-tion, and invasion [1, 2]. All of these cellular programs arefrequently deregulated during tumorigenesis [3]. Recently,functionally distinct amino-terminal variants of p130Cas wereidentified that exert different binding activities to FAK, there-by suggesting how p130Casmay be involved in the regulationof multiple pathways [4]. Elevated expression and activity ofp130Cas promote tumor progression in several cancer typesincluding lung, pancreas, prostate, and mammary cancers [1,2, 5, 6] and resistance to the anti-tumor drugs adriamycin(Doxorubicin) and tamoxifen in breast cancers [7, 8].

We previously established a decoy approach (Src*/SD) tomodulate p130Cas SD downstream signaling [1, 9, 10]. Src*/SD is a chimeric molecule composed of the c-Src kinase do-main with reduced activity on cellular substrates [(Src*, tyro-sine 416 to phenylalanine (Y416F)] fused to the p130Cas SD.Src* constitutively phosphorylates the SD within Src*/SDindependent of upstream signals. Thereby, the phosphorylatedSD in Src*/SD competes with endogenous p130Cas forinteracting proteins. As a result, the activity of severalp130Cas-modulated signal transduction pathways and cellularprograms is altered [1, 9, 10].

Earlier, we investigated the effects of the Src*/SDmoleculeon breast tumor development in vivo by crossingmouse mam-mary tumor virus (MMTV)-Src*/SD mice and MMTV-polyoma middle T-antigen (PyMT) mice [11]. PyMT miceare a commonly used breast cancer model, because they de-velop mammary gland tumors with short latency and the tu-mor progression reflects that of human disease [12]. Thedouble-transgenic PyMT/Src*/SD mice displayed significant-ly accelerated tumor formation and developed more aggres-sive lesions compared with single-transgenic PyMT animals[11]. However, during the study period, no metastases wereobserved with this model system.

Therefore, we aimed to answer the question whether theSrc*/SD molecule also drives metastasis formation in vivoand which part of the phosphorylated SD may be importantfor this process by applying an orthotopic mouse model. Here,we present that the full-length SD and a fragment containingthe SD motifs 6–10 [amino acids (aa) 207–315] significantlyenhanced tumor progression and metastasis formation. Ourstudies indirectly show that the phosphorylated SD motifs6–10 (Y236, Y249, Y267, Y287, and Y306) of endogenousp130Cas may be important for tumor progression.

Material and methods

Expression constructs The human p130Cas SD fragments(Fig. 1b) were generated by PCR with Pfu polymerase(Stratagene) using p130Cas cDNA as template and subcloned

Cla I/Not I into the doxycycline-inducible retroviral expres-sion vector pCXbsr containing HA-tagged Src* or SrcKM.Primer sequences are provided in Table S1. All constructswere verified by sequencing. The rat SD construct has beendescribed [10]. QIAGEN Plasmid Maxi Kit was used forDNA preparation.Cell culture and retroviral transductions Cell lines wereobtained as follows: TAM-R, estrogen receptor (ER) positive,tamoxifen resistant MCF-7-derivative (Robert Nicholson,Cardiff, UK); LM2, clone 4175, triple-negative, MDA-MB-231-derivative, constitutively expressing luciferase (JoanMassague [13]). Culture conditions: TAM-R (phenol red-free RPMI-1640, 5 % charcoal-stripped FBS, and 4 mM glu-tamine); LM2 (DMEM, 5 % FBS). Media were supplementedwith antibiotics/sodium pyruvate and the cells were main-tained as described [14]. Retroviral transductions were carriedout as described [14, 15].Protein analysis and cell fractionation Whole cell proteinextraction, fractionation, and immunoblotting were performedas described [11]. Antibodies: mAbs: β-Actin (Sigma);PARP1 (clone Y17, recognizes 116 kDa full length,Millipore). Polyclonal: HA (12CA5, Roche AppliedScience); CasB (directed against the SD of p130Cas, AmyH. Bouton, Charlottesville); phospho-tyrosine (p-Tyr; PY-99,sc-7020, Santa Cruz Biotechnology).

B

A

HA

Src* SD (YxxP)15P P P P P P

P

aa 109 419

Src*/SD

SrcKM SD (YxxP)15

aa 109 419

Src /SDKM

HA

YQVPYLVP YDVP YAVP phosphorylation by Src, v-Crk binding

Crk-SH2 bindingCrk-, Nck- and Src-SH2 binding Crk- and Src-SH2 binding

aa 109 200 300 400

1 2 3 64 97 8 105 1412 13 15111 3 64 97 8 105 1412 13 1511

>500 1900-5 >250 >4018

2061092

207 3153

419

4

207 259

260 315

6316

4195

200 439 130

516157rat SD

hum

an

1human

PhosphoSitePlus

YDIP

Fig. 1 Expression constructs used in this study. a Schematic of the HA-tagged Src/SD fusion proteins. Src* attenuated Src kinase domain, SrcKM

inactive Src kinase domain, SD p130Cas substrate domain, Pphosphorylation, aa amino acids. b Sketch of the relative positions andphosphorylation of the YxxP motifs in the p130Cas SD. Motifs 1–15 andinteraction partners are indicated. The symbols for the motif types andinteractors are explained at the bottom of the figure. The number ofstudies detecting phosphorylation of each motif by proteomics isindicated in the gray box (PhosphoSitePlus, e.g., motif 1 was foundphosphorylated in 18 studies). The SD regions present in the generatedconstructs are displayed and boundaries are marked by aa numbers

Tumor Biol.

Page 3: A truncated phosphorylated p130Cas substrate domain is ... · ORIGINAL ARTICLE A truncated phosphorylated p130Cas substrate domain is sufficient to drive breast cancer growth and

Densitometric analysis of gel bands was performed usingImage Studio Lite Version 5.2 (LI-COR Biosciences). Equaladjustments of contrast and brightness of gel pictures andscanned films were applied to all parts of the image usingAdobe Photoshop CS2 version 9.0. Densitometric values of theSrc*/SD cytoplasmic (S) and membrane/pellet (P) bands wererelated to the corresponding actin band values. The resultingvalues were related to the matching S values (set to one).

Mammary fat pad injections and survival surgery Animalexperiments were performed and the mice euthanized accord-ing to guidelines of and approved by the Institutional AnimalCare and Use Committee (IACUC) at Boston University. Atthe day of injection, LM2 cells (passaged 1 day earlier) weretrypsinized and washed with regular growth medium and thenwith cold PBS. 1×106 cells were diluted in 25 μl cold PBS,mixed with 25 μl growth factor-reduced Matrigel (FisherScientific), and kept on ice. The mixture (50 μl) was injectedwith a 27.5-gauge needle into the fourth inguinal fat pad ofanesthetized (3 % isoflurane) 8-week-old NOD/MrkBomTac-Prkdcscid mice (Taconic). Proper injection of the cells intothe mammary fat pad (MFP) was confirmed by biolumines-cence imaging (BLI). Src*/SD induction in the injected cellswas achieved by continuous DOX administration (2 mg/ml) indrinking water. Primary tumor growth and metastases weremonitored by bioluminescence imaging (BLI) and calipermeasurements as described [11].Bioluminescence imaging and analysis The Xenogen IVISSpectrum Instrument and Living Image Software Version 3.2(Caliper LifeSciences) were used for imaging, analysis, andquantification of BLI signals. Animals were injected intraperito-neally (IP) with 150 mg/kg D-luciferin (Gold Biotechnology)diluted in Dulbecco’s phosphate-buffered saline (DPBS withoutmagnesium and calcium, Fisher Scientific) 20 min prior to invivo imaging. Up to five mice at a time were anesthetized(1–3 % isoflurane), placed in a warmed imaging chamber withcontinuous isoflurane exposure (1–2 %), and imaged for up to1 min. For ex vivo imaging, 150 mg/kg D-luciferin was injectedIP immediately prior to necropsy. Tissues of interest were ex-cised, placed individually into 12-well plates with 300 μg/mlD-luciferin in DPBS, and imaged for up to 2 min.

Results

Identification of regions in the SD of p130Caswith potential function during tumor progression

The rat Src*/SD (Src*/rat SD) decoy molecule (Fig. 1a) hasbeen previously shown by us to promote primary tumorgrowth and a more severe phenotype in the PyMT breastcancer mouse model in vivo [11]. Based on these studies,we were interested in which tyrosine motifs are especiallyimportant for driving tumor development. Previous work from

our laboratory and by others has demonstrated that not all ofthe YxxP SDmotifs are phosphorylated by Src or required forCrk binding [16, 17] (Fig. 1b). A PhosphoSitePlus analysis[18] indicates that the central and carboxy-terminal sites 6–15,primarily of the YDVP sequence, are more frequently phos-phorylated than the amino-terminal motifs 1 (YLVP), 3, 4, and5 (YQVP sequences) (Fig. 1b). Motif 2 (YQVP) phosphory-lation was reported in >500 studies. This suggests that certainmotifs are more important than others for p130Cas function.

To identify the SD region that mediates important p130Casfunctions during tumor progression, the Src*/SD approachwas used as a tool. For this study, the human full-length SD(construct 1) and the truncated SD fragments 2 to 6 (Fig. 1b)were fused in frame to the attenuated Src kinase domain(Y416F; Src*) or the inactivated Src kinase domain(K295M; SrcKM) as controls. Stable DOX-inducible transduc-tants were established for the Src*/1 to Src*/6 as well as thecorresponding SrcKM controls and control (c; empty vector)constructs in metastatic LM2 breast cancer cells. LM2 cellsare derived from triple-negative MDA-MB-231 cells [13]. Allof the ectopic proteins were expressed, and as expected, theSrc*/SD constructs were tyrosine phosphorylated, whereasthe corresponding SrcKM/SD were not (Figs. 2a and S1).Because signaling changes by expression of Src*/SD arestrongly associated with induction of significant morphologicchanges in other breast cancer cell lines [10], the morphologyof the LM2 transductants was analyzed.

No alterations in cell morphology were observed in theSrcKM controls (Fig. S1A and not shown) and Src*/4, 5, and6 transductants (Fig. S1B). In contrast, induced Src*/1 orSrc*/3 expression induced cell rounding (Fig. 2b). After48 h of DOX treatment, 72 % of Src*/1 and 51 % of Src*/3cells rounded up whereas significantly less (P ≤ 0.01)rounded-up cells were observed in control cells (24 %). Thesame constructs were also active in a different breast cancercell line (data not shown). This suggests that YxxP motifs 6–10 within aa 207–315 and present in Src*/3 are important forp130Cas function.

In our PyMT/Src*/SD mouse model, partial localiza-tion of Src*/SD to cellular membranes by PyMT acceler-ated tumor development [11]. Therefore, the cellular lo-calization of the SD proteins was examined. Whole cellextracts of Src*/1 and Src*/3 LM2 transductants wereanalyzed by cell fractionation and immunoblotting(Fig. 2c). TAM-R cells (rat SD) were included as controlbecause we previously presented that Src*/SD expressionis almost entirely restricted to the cytoplasm in these cells[11]. TAM-R cells expressed the phosphorylated SD con-struct preferentially in the cytoplasm (26 % of cytoplas-mic levels were found at membranes) (Fig. 2d). LM2 cellsexpressed slightly higher Src*/1 and Src*/3 levels in thecytoplasmic than in the membrane fraction. Importantly, asubstantial portion of phosphorylated Src*/1 (60 %) and

Tumor Biol.

Page 4: A truncated phosphorylated p130Cas substrate domain is ... · ORIGINAL ARTICLE A truncated phosphorylated p130Cas substrate domain is sufficient to drive breast cancer growth and

Src*/3 (53 %) was detected at membranes reflecting inpart the localization pattern found in PyMT/Src*/SD cellsindicated above.

The full-length and a truncated phosphorylated p130CasSD promote metastasis formation

In the PyMT/Src*/SD mouse model, we observed acceleratedprimary tumor growth [11]. However, we could not detectmetastases within the study period. To investigate whetherthe full-length and the truncated Src*/SD protein can promotemetastasis formation when substantial amounts of these phos-phorylated SDmolecules are located at membranes (as shownfor the LM2 transductants), an orthotopic mouse model wasapplied. LM2 cells have a high potential to form metastaseswhen injected into the mammary fat pad of mice [13]. LM2cells expressing the inducible Src*/1, Src*/3, or control con-structs were injected into the mammary fat pad of NOD-Scidmice. Src*/1 and Src*/3 were utilized for these in vivo studiesbecause out of all constructs tested in vitro, they were the onlyones that induced changes in cell morphology. The rationalefor this selection was that previously, we have shown thattransductants that displayed these morphological changes alsohad altered signaling and cell behavior [10]. Primary tumorgrowth and metastasis formation were monitored by BLI andcaliper measurements.

Representative expression and phosphorylation of the SDchimeras in the primary tumors were confirmed by immuno-blotting (Fig. 3a). Of note, SD phosphorylation is highly var-iable but present in all tumors. In all groups, the injected cellsestablished primary tumors and in part formed spleen metas-tases at later time points (Fig. 3b, BLI quantification inFig. 3c, d). Compared with control cells, the primary tumorgrowth in mice injected with Src*/1 cells was significantlyaccelerated (Fig. 3c). Primary tumors reached a volume of500 mm3 on average within 29 days (±5.5, P≤0.001) and37 days (±3.4, P ≤ 0.05) in Src*/1 and Src*/3 animals,

C

*/3 +DOX

*/1 -DOX */1 +DOX

*/3 -DOX

control +DOX

A LM2CasB

100 kDa

55 kDa

c */3 */1c */3 */1

p-Tyr

B

p130Cas

actin

S P S P S P S P*/1 */3 c rat SD

100 kDa

55 kDa

40 kDa

135 kDa

LM2 TAM-R

1

3

rat SD

p130Cas

actin

1

3

rat SD 100 kDa

55 kDa

40 kDa

135 kDa

Cas

B

p-Ty

r

40 kDa

actin 40 kDa

actin

D

0

20

4060

80100

contr

ol */1 */3

+DOX

% ro

und

up c

ells

****

020406080

100120140

S P S P S P*/1 */3 SD

TAM-RLM2

rel.

phos

phor

ylat

ion

(pTy

r)

(%

of S

frac

tion)

020406080

100120

S P S P S P*/1 */3 SD

TAM-RLM2

rel.

expr

essi

on (C

asB

)

(% o

f S fr

actio

n)

*

�Fig. 2 Expression of the full-length Src*/1 and the truncated Src*/3,containing an SD fragment (amino acids 207–315 of p130Cas), inducesmorphological changes in LM2 cells. a–d LM2 cells stably transducedwith the indicated inducible expression constructs were treated for 48 hwith DOX or left untreated. * Src*, KM SrcKM, C control (empty vector).a Expression (CasB) and phosphorylation (p-Tyr) in WCE (30 μg) wasconfirmed by IB (actin, control). b Morphology of LM2 transductants(100,000 cells/well in 6-well plates) was analyzed by microscopy. Scale,100 μm. Lower left panel, rounded up and flat/spread cells were countedafter 48 h of DOX treatment in three fields from three independentxperiments per transductant (each field contained ∼70–180 cells) andthe results were averaged. Percent rounded up cells of all cells andstandard deviation are shown. P values were calculated using Student’st test (*P ≤ 0.05). c WCE from the indicated LM2 and TAM-Rtransductants were fractionated into cytoplasmic (S) and crudemembrane/pellet (P) fractions as described previously and analyzed byIB with CasB, p-Tyr, and actin abs. c control (empty vector). dDensitometric analysis of the experiment presented in c. Averageprotein expression (CasB) and phosphorylation (p-Tyr) of the indicatedtransduced constructs in the P fraction relative to the S fraction (set to one)and standard deviation from three (CasB) and two (p-Tyr) representativeexperiments are shown. P values were calculated using Student’s t test(*P ≤ 0.05)

Tumor Biol.

Page 5: A truncated phosphorylated p130Cas substrate domain is ... · ORIGINAL ARTICLE A truncated phosphorylated p130Cas substrate domain is sufficient to drive breast cancer growth and

respectively, compared with 41 days (±4.3) in control animals(Table 1). While only 36 % of control mice developed metas-tases by 67 days, metastases were evident in 54 % of Src*/1and 82 % Src*/3 mice (Fig. 4a and Table 1). Src*/1 animalscould not be observed for longer than 50 days because oftumor burden or other pathologies that required euthanasia.Moreover, metastases were formed earlier in Src*/1 (average

of 28 ± 8.8 days, P≤ 0.05) and Src*/3 (42 ± 16.4 days, notstatistically significant) than in control animals (47±16.9 days). At the end of the study period or when animalshad to be euthanized due to tumor burden, ex vivo imaging ofselected organs was performed (Fig. 4b and Table 2) to assayfor metastasis formation. In Src*/1 and Src*/3 animals, thetumor cells formed multiple metastases at several organs

day 7 day 14

*/1 (animal 2-5)day 22

*/3 (animal 1-8)day 17 day 26day 7

day 29 day 36day 22day 7

control (animal 2-5)B

CasBactin

p-Tyr6 7 8 LM

2-*/3

40 kD

a55

kDa55

kDa

A */31 53 4 6 7 - LM

2-*/1

40 kDa

100 kDaCasBactin

p-Tyr 100 kDa

*/1

1 2 3 5control

100 kDa

55 kDa

7 LM2-*

/1

LM2-*

/3

1 2 3 5control

7 LM2-*

/1

LM2-*

/3

p-Ty

r

E

05.00E+081.00E+091.50E+092.00E+092.50E+093.00E+093.50E+094.00E+09

0 7 14 21 28 35 42

tota

l Flu

x (p

hoto

ns/s

)

day

*/3 1-8*/1 2-5control 2-5

C

40 kDa

Cas

B

actin

05.00E+071.00E+081.50E+082.00E+082.50E+083.00E+083.50E+08

tota

l Flu

x (p

hoto

ns/s

)

cont

rol 2

-5*/1

2-5

*/3 1

-8

D control (n=11)

*/1 (n=11)

*/3 (n=11)

*

*

**

0200

400

600

800

1000

1200

1400

1600

1800

prim

ary

tum

or v

olum

e (m

m )3

22 29 36 43 49day

Fig. 3 The p130Cas SD drivesprimary breast cancer growth andmetastasis. Mice were injectedwith LM2 transductants [*/1, */3,or control (empty vector)] into themammary fat pad and analyzedby bioluminescence imaging(BLI) (b). a Expression (CasB)and phosphorylation (p-Tyr) ofthe transgene in primary tumorlysates (40 μg) were analyzed byIB (actin, control). Lysates ofLM2-*/1 and */3 cells were usedas positive control, and in theupper panels, only 20 μg wasloaded onto the gel due to the highexpression in this positive controlset. Numbers above the IBs areinternal animal numbers. Thedash indicates empty lane. bIn vivo BLI at indicated days postcell injection (exposure time 1 s).Representative images for eachcohort are shown. Color bars withdifferent signal intensities aredisplayed. Black circles (solidline, primary tumor; dashed line,metastases) mark the areasquantified in c and d. c Timecourse of primary tumor growthas analyzed by BLI signals of theindicated areas in animals in b.The measurements ended byanimal protocol after the firstdetection of metastases. dQuantification of BLI intensitiesof metastases at first detection ofanimals in b. e Comparison of theprimary tumor growth in miceinjected with the indicated LM2cells. n number of mice at thebeginning of the study. Primarytumor volumes were determinedby caliper measurements asdescribed. P values werecalculated using Student’s t test(*P ≤ 0.05; **P ≤ 0.01)

Tumor Biol.

Page 6: A truncated phosphorylated p130Cas substrate domain is ... · ORIGINAL ARTICLE A truncated phosphorylated p130Cas substrate domain is sufficient to drive breast cancer growth and

(spleen, lungs, liver, kidneys) in a single animal, whereas incontrol animals, majorly smaller single metastases to thespleen or lungs were observed.

These results show that the Src*/SD approach, including ashorter fragment of the SD, promotes breast cancer progres-sion in this orthotopic in vivo model. Expression of the trun-cated SD fragment led to a considerably higher number ofdetected metastases (more than twice of the animals devel-oped metastases compared with control mice). The full-length SD enhanced primary tumor growth and led to a sig-nificantly earlier metastasis onset and increased metastasisformation. This suggests that the full-length SD mediates astronger tumor-promoting activity, including earlier dissemi-nation. Nevertheless, these studies also indirectly indicate thatthe phosphorylated SD motifs 6–10 (aa 207–315 in wild-typep130Cas) of endogenous p130Cas may be important for tu-mor progression when a portion of the protein is found atcellular membranes.

Discussion

Our studies demonstrate that the phosphorylated p130Cas SDis important for p130Cas function during tumor progressionin vivo. In an orthotopic mouse model, Src*/SD proteins (full-length SD as well as a SD fragment containing motifs 6–10 ofthe p130Cas SD) promoted tumor progression. The full-length SD accelerated primary tumor growth and mediated asignificantly earlier metastasis onset. Both the full-length andthe truncated forms significantly increased metastasisformation.

We recently uncovered that partial localization of Src*/SDto cellular membranes by the polyoma middle T-antigen(PyMT) accelerates PyMT-induced tumor growth and pro-motes a more aggressive phenotype in vivo [11]. In thePyMT murine breast cancer model, this was attributed to thebinding of the attenuated Src kinase domain to the PyMTtransgene, which is a membrane-anchored viral protein. Thefindings that the Src*/SD fusion protein is partially located tomembranes in LM2 cells and drives tumor progression in anorthotopic breast cancer model suggest that other mechanismsof membrane targeting exist for this system. Importantly, viathe constitutively phosphorylated SD fragment, the chimeramight integrate membrane-associated active signaling

complexes. This is further supported by our previous studiesdescribing that at membranes, the decoy activates extracellularsignal-regulated kinase (ERK) survival signaling [11] whereasin the cytoplasm, it inhibits it [10]. The fact that only a portionof the Src*/SD molecules is located at membranes in the LM2transductants indicates that the tumor-promoting activity isdominant in this model system.

In the PyMT/Src*/SD mouse model, no metastasis forma-tion was detected during the study period [11] althoughp130Cas is known to drive metastases [1, 2]. Therefore, inan orthotopic model, we interrogated whether the decoy mol-ecule, when partially located at membranes, also promotesmetastases and which part of the phosphorylated SD is impor-tant for this process. LM2 transductants, with a substantialportion of the Src*/SDmolecules at membranes, were injectedinto the mammary fat pad of NOD-Scid mice. Similar to thePyMT model, primary tumor growth was enhanced. In addi-tion, we could clearly demonstrate that the Src*/SD mole-cules, containing the full-length or a truncated SD (comprisingaa 207–315 of wild-type p130Cas), drive metastasis forma-tion. Both Src*/SD proteins induced a considerably more se-vere phenotype as indicated by the detection of multiple me-tastases in several organs in single animals. Importantly, thecontrol animals developed majorly smaller single metastasesonly to the spleen or lungs. Furthermore, the full-length chi-mera greatly accelerated primary tumor growth and metastasisformation, whereas the effects of the truncated form were notas prominent. However, the shorter fragment induced metas-tases in more animals (82 % of the mice) than the full-lengthform (54 %). This may in part be explained by the fact thatanimals expressing the full-length SD developed pathologieswithin 50 days after tumor cell injection that required eutha-nasia whereas the animals expressing the truncated proteincould be observed for up to 67 days. Taken together, theseresults suggest that additional sequences in the full-length

Table 1 Tumor growth andmetastasis detection aftermammary fat pad injections ofLM2 transductants in mice

Cohort n Tumor volume (500 mm3) Metastasis(within 67 days)

Average metastasisdetection

Control 11 41 days (±4.3) 36 % 47 days (±16.9)

Src*/1 13 29 days (±5.5, P ≤ 0.001) 54 % 28 days (±8.8, P ≤ 0.05)Src*/3 11 37 days (±3.4, P ≤ 0.05) 82 % 42 days (±16.4, n.s.)

n.s. not significant

�Fig. 4 The p130Cas SD drives metastasis. Mice were injected with theindicated LM2 cells [*/1, */3, or control (empty vector)] into themammary fat pad and analyzed by bioluminescence imaging (BLI). aPercent metastases-free animals were calculated based on detections inlive animals as well as from ex vivo studies at the end of the study period.b Ex vivo BLI of the indicated organs. Different exposure times wereused as stated. Color bars with diverse signal intensities are displayed.DPBS+ luc (luciferin), negative control

Tumor Biol.

Page 7: A truncated phosphorylated p130Cas substrate domain is ... · ORIGINAL ARTICLE A truncated phosphorylated p130Cas substrate domain is sufficient to drive breast cancer growth and

SD not present in the truncated molecule might be responsiblefor a stronger tumor-promoting activity. In addition to motifs6–10, the motifs 11–15 are phosphorylated by c-Src andbound by Crk [17]. However, Crk binds with a lower affinityto the c-terminal SD motifs 11–15 than to the central tyrosinemotifs 6–10 [17]. Nevertheless, testing of truncated SDs con-taining motifs 1–5 or 11–15 with the Src*/SD approach didnot significantly affect the cellular behavior of two breast can-cer cell lines (Fig. S1 and not shown).

The finding that the truncated Src*/SD decoy drives in vivotumor metastases provides new information that this region ofendogenous p130Cas may be important for mammary cancerprogression. This region contains the tyrosine motifs 6–10 ofthe human SD (Y236, Y249, Y267, Y287, and Y306).Phosphorylation of the SD is essential for the dynamic assem-bly and disassembly of focal adhesions [19]. Each of thesemotifs was found to be phosphorylated in more than 190 pro-teomic studies as curated at PhosphoSitePlus [18], whereasmost of the other SDmotifs were significantly less often phos-phorylated. Other studies demonstrated that the motifs 6–10are central for Crk, c-Src, and Nck binding and migrationin vitro [16, 17], which are important processes for cellulartransformation and tumor progression [2, 5].

The importance of these results is further indicated becausep130Cas SD phosphorylation is a central step in the regulationof cell adhesion, migration, and invasion as well as prolifera-tion and survival [6, 19–21]. In the past few years, p130Caswas also identified as a mechanosensor in response to extra-cellular matrix-integrin engagement [22, 23]. Upon mechani-cal stretching, the SD tyrosines are exposed and accessible tophosphorylation by kinases leading to the activation ofmechanotransduction signaling pathways [23]. Thereby, cellscan respond and adapt to their environment by activating var-ious cellular programs that may also affect the cells’ surround-ings. These facts demonstrate the importance of p130Cas SDphosphorylation in tissue and cell homeostasis, and it is notsurprising that deregulation of p130Cas expression and phos-phorylation is involved in various pathogeneses. Elevatedp130Cas levels in primary breast tumors correlate with in-creased rate of relapse and with poor response to tamoxifentreatment [24]. Increased p130Cas expression in comparison

(1 m exposed, day 50)control

heart lungs kidneys

liver DPBS + lucspleen

(10 s exposed, day 50)*/3

heart lungs kidneys

liver DPBS + lucspleen

0 22 26 32 36 43 45 50 67

A

day0

10

20

30

40

50

60

70

80

90

100

% m

etas

tase

s-fr

ee a

nim

als

control (n=11)

*/1 (n=13)

*/3 (n=11)

(30 s exposed, day 36)*/1

heart lungs kidneys

liver DPBS + lucspleen

B

Table 2 Ex vivo detection of metastasis in indicated organs aftermammary fat pad injections of LM2 transductants in mice (endpointmaximum 67 days)

Organ Control (n = 5) Src*/1 (n= 3) Src*/3 (n= 4)

Heart 0/5 0/3 0/4

Lungs 3/5 3/3 4/4

Kidneys 0/5 2/3 4/4

Liver 1/5 3/3 3/4

Spleen 1/5 3/3 3/4

Tumor Biol.

Page 8: A truncated phosphorylated p130Cas substrate domain is ... · ORIGINAL ARTICLE A truncated phosphorylated p130Cas substrate domain is sufficient to drive breast cancer growth and

to primary tumors was also detected in tumor cells isolatedfrom pleural effusions of mammary carcinoma patients [25].In addition, a recent study described that 75 % of the patientswith triple-negative breast cancer expressed high p130Caslevels [26]. Moreover, p130Cas contributes to the progressionof several additional cancers such as gliomas, prostate cancer,and leukemias [27, 28]. Therefore, targeting p130Cas mayopen new avenues for the treatment of certain malignancies.Initial testing by in vivo downregulation of p130Cas byintranipple injection of short-interfering (si)RNA resulted ina reduction of tumor growth in BALB/c-HER2/neu mice [29].Although speculative, data presented here suggest thattherapeutics that specifically prevent p130Cas transloca-tion to membranes and/or block YDxP motif 6–10downstream signaling might be considered for futuretreatment options.

In summary, this study identified a region in thep130Cas substrate domain that when phosphorylatedand found at membranes may be essential for promotingtumor progression.

Acknowledgments We gratefully acknowledge Joan Massague andRobert I. Nicholson for cell lines and Amy H. Bouton for the CasBantibody. We thank Matthew D. Layne for critical reading of the manu-script. We greatly appreciate the help ofManish Bais and TomBalonwithbioluminescence imaging and of Kim Bayer in establishing the animalprocedures. All bioluminescence imaging was performed at the IVISImaging Core of Boston University School of Medicine. This work wassupported by the National Cancer Institute grant CA106468, by theNational Center for Advancing Translational Sciences through the BU-CTSI grant U54TR001012, both from the National Institute of Health,and the Susan G. Komen for the Cure Breast Cancer Foundation grantKG101208.

Compliance with ethical standards Animal experiments were per-formed and the mice euthanized according to guidelines of and approvedby the Institutional Animal Care and Use Committee (IACUC) at BostonUniversity.

References

1. Kumbrink J, Kirsch KH. Targeting Cas family proteins as a noveltreatment for breast cancer, breast cancer—current and alternativetherapeutic modalities. Esra Gunduz and Mehmet Gunduz (Ed),InTech. 2011;ISBN: 978-953-307-776-5:37-62.

2. Tikhmyanova N, Little JL, Golemis EA. CAS proteins in normaland pathological cell growth control. Cell Mol Life Sci. 2010;67(7):1025–48.

3. Hanahan D, Weinberg RA. Hallmarks of cancer: the next genera-tion. Cell. 2011;144(5):646–74.

4. Kumbrink J, Soni S, Laumbacher B, Loesch B, Kirsch KH.Identification of novel Crk-associated substrate (p130Cas) variantswith functionally distinct focal adhesion kinase binding activities. JBiol Chem. 2015;290(19):12247–55.

5. Cabodi S, Pilar Camacho-LealM, Di Stefano P, Defilippi P. Integrinsignalling adaptors: not only figurants in the cancer story. Nat RevCancer. 2010;10(12):858–70.

6. Nikonova AS, Gaponova AV, Kudinov AE, Golemis EA. CASproteins in health and disease: an update. IUBMB Life.2014;66(6):387–95.

7. Dorssers LC, van der Flier S, Brinkman A, van Agthoven T,Veldscholte J, Berns EM, et al. Tamoxifen resistance in breast can-cer: elucidating mechanisms. Drugs. 2001;61(12):1721–33.

8. Ta HQ, Thomas KS, Schrecengost RS, Bouton AH. A novel asso-ciation between p130Cas and resistance to the chemotherapeuticdrug adriamycin in human breast cancer cells. Cancer Res.2008;68(21):8796–804.

9. Kirsch KH, Kensinger M, Hanafusa H, August A. A p130Cas ty-rosine phosphorylated substrate domain decoy disrupts v-crk sig-naling. BMC Cell Biol. 2002;3:18.

10. Soni S, Lin BT, August A, Nicholson RI, Kirsch KH. Expression ofa phosphorylated p130(Cas) substrate domain attenuates the phos-phatidylinositol 3-kinase/Akt survival pathway in tamoxifen resis-tant breast cancer cells. J Cell Biochem. 2009;107(2):364–75.

11. Zhao Y, Kumbrink J, Lin BT, Bouton AH, Yang S, Toselli PA, et al.Expression of a phosphorylated substrate domain of p130Cas pro-motes PyMT-induced c-Src-dependent murine breast cancer pro-gression. Carcinogenesis. 2013;34(12):2880–90.

12. Lin EY, Jones JG, Li P, Zhu L, Whitney KD, Muller WJ, et al.Progression to malignancy in the polyoma middle T oncoproteinmouse breast cancer model provides a reliable model for humandiseases. Am J Pathol. 2003;163(5):2113–26.

13. Minn AJ, Gupta GP, Siegel PM, Bos PD, Shu W, Giri DD, et al.Genes that mediate breast cancer metastasis to lung. Nature.2005;436(7050):518–24.

14. Kumbrink J, Kirsch KH. Regulation of p130(Cas)/BCAR1 expres-sion in tamoxifen-sensitive and tamoxifen-resistant breast cancercells by EGR1 and NAB2. Neoplasia. 2012;14(2):108–20.

15. Kumbrink J, Kirsch KH. p130Cas acts as survival factor duringPMA-induced apoptosis in HL-60 promyelocytic leukemia cells.Int J Biochem Cell Biol. 2013;45(3):531–5.

16. Huang J, Hamasaki H, Nakamoto T, Honda H, Hirai H, Saito M, etal. Differential regulation of cell migration, actin stress fiber orga-nization, and cell transformation by functional domains of Crk-associated substrate. J Biol Chem. 2002;277(30):27265–72.

17. Shin NY, Dise RS, Schneider-Mergener J, Ritchie MD, KilkennyDM, Hanks SK. Subsets of the major tyrosine phosphorylation sitesin Crk-associated substrate (CAS) are sufficient to promote cellmigration. J Biol Chem. 2004;279(37):38331–7.

18. Hornbeck PV, Chabra I, Kornhauser JM, Skrzypek E, Zhang B.PhosphoSite: a bioinformatics resource dedicated to physiologicalprotein phosphorylation. Proteomics. 2004;4(6):1551–61.

19. Machiyama H, Hirata H, Loh XK, Kanchi MM, Fujita H, Tan SH,et al. Displacement of p130Cas from focal adhesions links actomy-osin contraction to cell migration. J Cell Sci. 2014;127(Pt 16):3440–50.

20. Cunningham-Edmondson AC, Hanks SK. p130Cas substrate do-main signaling promotes migration, invasion, and survival of estro-gen receptor-negative breast cancer cells. Breast Cancer (DoveMedPress). 2009;1:39–52.

21. Donato DM, Ryzhova LM, Meenderink LM, Kaverina I, HanksSK. Dynamics and mechanism of p130Cas localization to focaladhesions. J Biol Chem. 2010;285(27):20769–79.

22. Camacho Leal MP, Sciortino M, Tornillo G, Colombo S, DefilippiP, Cabodi S. p130Cas/BCAR1 scaffold protein in tissue homeosta-sis and pathogenesis. Gene. 2015;562(1):1–7.

23. Janostiak R, Pataki AC, Brabek J, Rosel D. Mechanosensors inintegrin signaling: the emerging role of p130Cas. Eur J Cell Biol.2014;93(10-12):445–54.

24. van der Flier S, Brinkman A, Look MP, Kok EM, Meijer-vanGelder ME, Klijn JG, et al. Bcar1/p130Cas protein and primarybreast cancer: prognosis and response to tamoxifen treatment. JNatl Cancer Inst. 2000;92(2):120–7.

Tumor Biol.

Page 9: A truncated phosphorylated p130Cas substrate domain is ... · ORIGINAL ARTICLE A truncated phosphorylated p130Cas substrate domain is sufficient to drive breast cancer growth and

25. Konstantinovsky S, Smith Y, Zilber S, Tuft SH, Becker AM, NeslandJM, et al. Breast carcinoma cells in primary tumors and effusionshave different gene array profiles. J Oncol. 2010;2010:969084.

26. Tornillo G, Elia AR, Castellano I, SpadaroM, Bernabei P, Bisaro B,et al. p130Cas alters the differentiation potential of mammary lumi-nal progenitors by deregulating c-Kit activity. Stem Cells.2013;31(7):1422–33.

27. Barrett A, Evans IM, Frolov A, Britton G, Pellet-Many C, YamajiM, et al. A crucial role for DOK1 in PDGF-BB-stimulated glioma

cell invasion through p130Cas and Rap1 signalling. J Cell Sci.2014;127(Pt 12):2647–58.

28. Zheng Y, Asara JM, Tyner AL. Protein-tyrosine kinase 6 promotesperipheral adhesion complex formation and cell migration by phos-phorylating p130 CRK-associated substrate. J Biol Chem.2012;287(1):148–58.

29. Cabodi S, Tinnirello A, Bisaro B, Tornillo G, Pilar Camacho-LealM, Forni G, et al. p130Cas is an essential transducer element inErbB2 transformation. FASEB J. 2010;24(10):3796–808.

Tumor Biol.