Focal adhesion kinase dependent activation of the PI3 kinase pathway by the functional soluble form of neurotensin receptor-3 in HT29 cells

Page 1

Fb

FTC

a

ARRAA

KSNHCC

1

deitouommgrc2ehpc

b

1h

The International Journal of Biochemistry & Cell Biology 45 (2013) 952– 959

Contents lists available at SciVerse ScienceDirect

The International Journal of Biochemistry& Cell Biology

journa l h o me page: www.elsev ier .com/ locate /b ioce l

ocal adhesion kinase dependent activation of the PI3 kinase pathwayy the functional soluble form of neurotensin receptor-3 in HT29 cells

abienne Massa, Christelle Devader, Sophie Béraud-Dufour, Frédéric Brau,hierry Coppola, Jean Mazella ∗

NRS, Institut de Pharmacologie Moléculaire et Cellulaire, UMR 7275, Université de Nice Sophia Antipolis, 660 route des Lucioles, 06560 Valbonne, France

r t i c l e i n f o

rticle history:eceived 14 September 2012eceived in revised form 21 January 2013ccepted 31 January 2013vailable online xxx

eywords:

a b s t r a c t

The neurotensin (NT) receptor-3 (NTSR3), also called sortilin, is thought to display several functionsincluding a role as a receptor or a co-receptor, in the sorting to plasma membrane and to lysosomes, and inthe regulated secretion. The aim of this study was to investigate the function of the soluble form of NTSR3(sNTSR3) released from several cell lines including colonic cancer cells. The human adenocarcinomaepithelial cell line HT29 has been used to monitor the release, the binding and internalization of sNTSR3by radioreceptor assays and confocal microscopy. The modulation of the intracellular signaling pathways

oluble sortilineurotensinT29ell signalingancer

by the protein has been investigated by using Fura-2 fluorescence calcium imaging microscopy andWestern blots analysis. We demonstrated that sNTSR3 specifically binds and internalizes into HT29 cells.This binding, independent from the transactivation of the epidermal growth factor receptor, leads to theincrease of intracellular calcium concentration and to the activation of a FAK/Src-dependent activation ofthe PI3 kinase pathway. In conclusion, sNTSR3 released from the membrane bound NTSR3 is a functionalprotein able to activate intracellular pathways involved in cell survival but probably not in cell growth.

. Introduction

The growth of cancer cells is a multifactorial event that remainsifficult to clearly characterize and understand. Indeed, numerousxtracellular stimuli are responsible for the activation of severalntracellular pathways leading in all cases to abnormal prolifera-ion of neoplastic cells. In addition to these problems, disseminationf some cancer cells from the tumor is a process also poorlynderstood but crucial since it is responsible for developmentf metastasis (for review, see Thiery, 2002). Among extracellularolecules that activate gastrointestinal cancer tissues, several hor-onal peptides acting on peptide-positive-receptor tumors act as

rowth factors and then are the targets for diagnosis by the use ofadiolabeled analogs and for therapy by the development of spe-ific peptide targeted toxins or of selective antagonists (Pini et al.,008). The neuropeptide neurotensin (NT) belongs to this periph-ral hormonal family that fullfills the function of gastrointestinalormone able to regulate normal and neoplastic cells derived fromancreas (Yamada et al., 1995), prostate (Sehgal et al., 1994), or

olon (Maoret et al., 1999).

NT is involved in a number of important biological processesoth in the brain and in the periphery (for reviews see Vincent

∗ Corresponding author. Tel.: +33 4 93 95 77 61; fax: +33 4 93 95 77 08.E-mail address: [email protected] (J. Mazella).

357-2725/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.biocel.2013.01.020

© 2013 Elsevier Ltd. All rights reserved.

et al., 1999; Myers et al., 2009). These processes include dopaminetransmission (Kitabgi et al., 1989), analgesia (Dobner, 2006),hypothermia (Popp et al., 2007) and hormonal activity regulation(Rostene and Alexander, 1997). The effects of NT are triggered fol-lowing its interaction with three different receptors (NTSRs). NTSR1and NTSR2 are both seven transmembrane (TM) domain G protein-coupled receptors (GPCR) whereas NTSR3 is a single TM domaintype I receptor which shows 100% homology with the sorting pro-tein, sortilin (Mazella, 2001). The effects of NT on the growth ofgastrointestinal cancer cells have been correlated with an increasein the expression of NTSR1 (Souaze et al., 2006; Zhao et al., 2007)and a transactivation of EGF receptors in nontransformed humancolonocytes NCM460 cells (Zhao et al., 2004). The EGF receptortransactivation induced by NT would be mediated by NTSR1 acti-vation leading to a PKC-dependent increase in metalloproteasesactivity responsible for the release of EGFR ligands (for review seeCarraway and Plona, 2006). However, the NTSR3 has also beenshown to be expressed in several human cancer cell lines on whichNT exerts its proliferative effects (Dal Farra et al., 2001). Moreover,in the human colonic adenocarcinoma cell line HT29, the intracellu-lar signaling of NT has been shown to be modulated by the existenceof a functional interaction between NTSR1 and NTSR3. This complex

is responsible for the NT-induced phosphoinositide (PI) accumu-lation and phosphorylation of MAP kinases (Martin et al., 2002).Interestingly, NTSR3 is also shedded from HT29 and Lovo cells bymetalloproteases inducing the release of a soluble protein (sNTSR3)

Page 2

f Bioch

(cnHmwpdac

2

2

gifb1pFDTAAEtCAF(

2

su

ePnfPabp

2H

drb1

hcvc(a(

between 1 and 10 s on an EMCCD camera (Cascade 512, Roper Sci-

F. Massa et al. / The International Journal o

Navarro et al., 2002) whose biological function remains to be elu-idated. In this study, we sought to examine the role of sNTSR3 onumerous biological responses of the colonic epithelial cancer cellsT29. Our data show that NTSR3 can be released from the plasmaembrane following various stimuli and that the resulting sNTSR3as able, by itself and upon specific binding to cells, to activate thehosphorylation of Akt, Src and FAK. Taken together, our resultsemonstrate for the first time that the soluble form of NTSR3 is

functional protein that can regulate the signaling of the humanolonic adenocarcinoma cell line HT29.

. Materials and methods

.1. Materials

NT was purchased from Peninsula Laboratories. Epidermalrowth factor (EGF) was from Upstate Millipore. Dulbecco’s mod-fied Eagle’s medium (DMEM) was from Life Technologies Inc. andetal calf serum from Lonza. Gentamicin, 1–10-phenanthroline,ovine serum albumin (BSA), mowiol, paraformaldehyde, phorbol-2 myristate 13 acetate (PMA), carbachol, PGE2, mammalianrotease and phosphatase inhibitor cocktails were from Sigmarance. FAK inhibitor II was from Calbiochem. Lysotracker Red-ND-99 and NHS-Bodipy-493 were from Molecular Probes.he soluble recombinant human NTSR3/sortilin protein (Ser78-sn755), resuspended in PBS-BSA 0.1%, was from R&D systems.ntibodies against the phosphorylated Erk1/2 or total forms ofrk 1/2 and Akt, EGFR and PKC� were from Santa Cruz Labora-ory Inc. The rabbit polyclonal phospho-Akt antibody was fromell Signaling. The antibody against phosphorylated Src was frombcam, antibodies against the phosphorylated or total forms ofAK were from Upstate Biotech. Alexa-595 Weat Germ AgglutininWGA) was from BD Pharmingen.

.2. Cell culture and release of soluble NTSR3 experiments

The human cell lines HT29 and Lovo were maintained in DMEMupplemented with 10% FBS and 50 �g/ml gentamicin at 37 ◦Cnder 5% CO2.

To recover the soluble form of sNTSR3, cells were rinsed andquilibrated in the Earle’s buffer then treated with 2.5 �g/mlMA, or 1 �M PGE2 or 100 �M carbachol for 6 h at 37 ◦C. Super-atants were recovered and precipitated using 10% TCA. Proteins

rom the pellets were transfered onto nitrocellulose after SDS-AGE and the NTSR3 was detected using an antibody directedgainst the luminal domain of the protein (1/1000) and revealedy the Lumilight enhanced chemiluminescence method (Milli-ore).

.3. Iodination of sNTSR3 and EGF and binding experiments onT29 cells

125I-EGF (600 Ci/mmol) was prepared and purified as previouslyescribed (Comens et al., 1982). sNTSR3 was iodinated using chlo-amine T and the iodinated product was separated from free iodiney gel filtration onto Sephadex G50 fine equilibrated and eluted in00 mM Tris–HCl, pH 7.5, containing 0.2% BSA.

Binding experiments were carried out on HT29 cellomogenates as previously described (Botto et al., 1998). Briefly,ell homogenates (50 �g of protein) were incubated in a totalolume of 100 �l of incubation buffer (50 mM Tris–HCl, pH 7.5,

ontaining 0.2% BSA) at 25 ◦C for 30 min with 0.4 nM 125I-sNTSR3200 Ci/mmol) or with 0.2 nM 125I-EGF (600 Ci/mmol) in thebsence or in the presence of increasing concentrations of sNTSR3from 1 × 10−10 to 5 × 10−7 M) or NT or EGF (both from 1 × 10−10

emistry & Cell Biology 45 (2013) 952– 959 953

to 1 × 10−6 M). Binding experiments were terminated by filtra-tion under reduced pressure through Sartorius filters (SM11107,0.2 �m pore size). The tube and the filter were rapidly washedtwice with 2 ml of incubation buffer. The radioactivity retained bythe filter was counted in a Packard g-counter.

Internalization experiments were performed on whole HT29cells in a Earle’s buffer (140 mM NaCl, 5 mM KCl, 2 mM MgCl2,0.8 mM CaCl2, 25 mM Tris–HEPES, pH 7.4, 0.1% bovine serum albu-min) as already described (Botto et al., 1998).

2.4. Preparation of Bodipy-green-sNTSR3 and internalizationexperiments

50 �g of sNTSR3 (R&D) diluted in 150 �l of 60 mM Phos-phate buffer, pH 7.5, were mixed overnight at room temperaturewith 170 �g of NHS-Bodipy-493 diluted in 40 �l acetonitrile. Theincubation medium was centrifuged and the Bodipy-sNTSR3 wasseparated from free NHS-Bodipy by gel filtration onto SephadexG50 fine equilibrated and eluted in 20 mM phosphate buffer, pH7.5 containing 20% acetonitrile.

Cells were plated on glass coverslips coated with 1 mg/ml poly-l-lysine. Bodipy-493-sNTSR3 (10−8 M) was incubated for 30 min at37 ◦C with HT29 cells in the absence or in the presence of 0.45 Msucrose or with 1 �M Lysotracker Red-DND-99 (Molecular Probes).At the end of incubation, cells were washed twice with the equil-ibration buffer (Earle’s buffer) then once with the equilibrationbuffer containing 0.5 M NaCl, pH 4.0 for 2 min to remove extra-cellular bound ligand, and mounted in Fluoprep before confocalimaging analysis. In some cases, sequestrated Bodipy-493-sNTSR3was extracted from cells directly with SDS-PAGE solubilizing buffer,loaded on a 10% polyacrylamide gel and transferred onto nitrocel-lulose. The amount of Bodipy-493-sNTSR3 remaining intact wasquantified using the ProEXPRESS Proteomic Imaging System fromPerkin.

Confocal microscopy was performed with a Laser ScanningConfocal Microscope (TCS SP5, Leica, Rueil Malmaison, France)equipped with a DMI6000 inverted microscope, using a PlanApo 63x/1.4 NA oil immersion objective. The fluorescent mark-ers were respectively and sequentially excited by the 488 nm and561 nm wavelengths of an argon and Diode Pumped Solid-Statelaser. Fluorescence was detected through 500–540 nm (Bodipy-493) and 590–650 nm (Alexa-594) spectral windows. Images wereacquired as single transcellular optical section and averaged over12 scans/frame.

2.5. Measurement of cytosolic calcium variation

Cells were incubated for 30 min at 37 ◦C in the Fura-2AM loadingsolution consisted of standard extracellular saline (SES) contain-ing: 135 mM NaCl, 5 mM KCl, 1.2 mM MgCl2, 2 mM CaCl2, 10 mMHepes/NaOH pH 7.4, 1 mM bicarbonate and 5 mM glucose, with0.1% BSA and fura-2AM (10 �M). The loading solution was removedand cells were equilibrated in fresh SES for 15 min. Cells werevisualized under an inverted epi-fluorescence microscope (AxioOb-server, Carl Zeiss, France) using a Fluar 40 × 1.3 oil immersionobjective. The intracellular Fura-2AM was sequentially excited at340 and 380 nm with a xenon arc-lamp through a high-speedmulti-filter wheel. For each excitation wavelength, the fluores-cence emission was discriminated by a same 400 LP dichroic mirrorand a 510/40 bandpass filter. Fluorescence images were acquired

entific, Evry, France). Calcium images analysis were made usingMetaMorph, Metafluor (Universal Imaging). Fura-2 fluorescenceintensities were expressed as changes relative to the initial fluo-rescence ratio (F340/380).

Page 3

9 f Bioch

2

psfprtcw2Tfksomt1h

2

mwoew5pIoipocatEalbt

2

hNosctce

2

bpt

54 F. Massa et al. / The International Journal o

.6. Immunocytochemistry

HT29 cells, plated on glass coverslips coated with 1 mg/mloly-l-lysine, were preincubated for 10 min in phosphate-bufferedaline (PBS) and incubated or not (control) with sNTSR3 (10−8 M)or indicated times. Then, cells were fixed 20 min with 4%araformaldehyde in PBS at room temperature. Coverslips wereinsed twice with PBS and incubated with the quenching solu-ion 50 mM NH4Cl in PBS for 10 min. After 20 min in PBSontaining 3% horse serum (HS) and 0.1% Triton-X100, cellsere labeled with a mouse monoclonal anti-PKC� (1/400) for

h at room temperature in PBS containing 0.5% HS and 0.1%riton-X100. Cells were rinsed three times in PBS and incubatedor 45 min at room temperature with a FITC-conjugated don-ey anti-mouse antibody (1/400) in PBS containing 0.5% horseerum and 0.1% Triton-X100. After two washes with PBS andne with water, coverslips were mounted on glass slides withowiol for confocal microscopy examination. PKC� transloca-

ion observed by confocal imaging was analyzed using ImageJ.4.3.67 software (W.S. Rasband, National Institute of Health,ttp://www.rsb.info.nih.gov.gate1.inist.fr/ij/).

.7. Detection of phosphorylated Akt, Erk1/2, FAK, Src and EGFR

Cells (70% confluence) were incubated overnight in serum freeedium. Cells were then incubated for various times at 37 ◦Cith 10−8 M sNTSR3 or EGF for stimulation of Erk1/2, FAK, Src

r Akt. In some cases, cells were pre-incubated in the pres-nce of FAK inhibitor II (10 �M) for 1 h. Cells were then lysedith a buffer containing 50 mM Tris/HCl pH 7.5, 100 mM NaCl,

mM EDTA, 0.5% Na-deoxycholate, 1% NP40, 0.1% SDS in theresence of phosphatase and protease inhibitor cocktails (1/100).

dentical amounts of solubilized protein (40 �g) were analyzedn SDS-PAGE, electroblotted onto nitrocellulose and subjected tommunoblotting using an antibody directed against the activehosphorylated forms of Erk1/2, Src, Akt or EGFR. For detectionf phosphorylated FAK, solubilized proteins were previously pre-ipitated with the anti-FAK antibodies before to be analyzed withn anti-phospho-Tyr antibody. Results were standardized withinhe same blot using the anti-tubulin antibody or the total anti-rk1/2 or anti-FAK or Anti-Akt antibodies. Analysis of proteins wasssessed on digital images using the Image J software. The stimu-ation was calculated from the ratio between the density of eachand and the density of the band obtained in the control condi-ion.

.8. Cell proliferation

Exponentially growing cells from non-confluent cultures werearvested and seeded into a flat-bottom 96-well plate (Corning,Y, USA) at a density of 5 × 103 cells per well, in the presencef 10% FBS. After 24 h, cells were serum-starved and treated withNTSR3 (10 nM) or EGF (10 nM) or 10% FBS. After 96 h of treatment,ells number was determined using the fluorometric cell prolifera-ion Assay Kit (CellTiter-Blue® from Promega). A separate standardurve was used to convert fluorescence units to cell numbers. Eachxperiment was performed in triplicate.

.9. Statistics

Each median value is obtained from 6 to 12 determinations,ox plots show medians, 25% and 75% percentiles, and 5% and 95%ercentiles. Significance was determined using the Kruskal–Wallisest: *p < 0.05 **p < 0.01 and ***p < 0.005.

emistry & Cell Biology 45 (2013) 952– 959

3. Results

3.1. sNTSR3 in HT29 cells: release, binding properties and absenceof interaction with EGFR

We already observed that in extracellular medium collectedfrom HT29 cells the NTSR3 antibody detected a protein with amolecular weight (92 kDa) slightly lower than that observed incrude cell homogenates (Navarro et al., 2002). Interestingly, block-ing internalization process by hyperosmolar sucrose enhanced thebasal level of sNTSR3 release (Fig. 1A), suggesting either that accu-mulation of the protein at the cell surface increased the amountof protein that can be cleaved or that sucrose activated the shed-ding process. The amount of this soluble form of NTSR3 (sNTSR3)released by “shedding” was increased when cells were incubatedwith PMA, a protein kinase (PKC)-dependent activator of met-alloproteases (Navarro et al., 2002). To determine whether PKCactivators could activate the shedding of NTSR3, we first measuredthe amount of sNTSR3 recovered in the extracellular medium afterincubation of cells either with NT, carbachol, or PGE2, known toactivate PKCs (Warhurst et al., 1994). Fig. 1A showed that both car-bachol and PGE2, but not NT, activated the release of sNTSR3. So,these results showed that PKC effectors can activate the sheddingof NTSR3 from cancer cells.

To determine whether sNTSR3 may behave as a ligand able tobind to HT29 cell membranes, we iodinated the protein to prepare aradioactive ligand. Competition experiments in which the bindingof 125I-sNTSR3 to HT29 cell homogenates were carried out in thepresence of increasing concentrations of non-radioactive sNTSR3or NT, showed that sNTSR3 competed for 125I-sNTSR3 binding withan IC50 of 8 nM and that NT inhibited 35% of the binding at 1 �M(Fig. 1B). This experiment demonstrated that sNTSR3 was indeedable to specifically bind to HT29 cell membranes, but not on theNT receptors. Since sNTSR3 was released by a mechanism simi-lar to those releasing EGFR ligands, we wondered whether sNTSR3and EGF could be competitive ligands. Then, competition exper-iments were also realized with increasing concentrations of EGF.We observed that EGF did not interact with sNTSR3 binding sites(Fig. 1B). Finally, we verified that the binding of 125I-EGF performedin the presence of increasing concentrations of EGF or sNTSR3 wasonly efficiently competed by EGF with an expected affinity of 1 nM(Comens et al., 1982) but not by sNTSR3 which can displace only30% of the binding of 125I-EGF at a concentration of 1 �M (Fig. 1C).

To confirm the absence of action of sNTSR3 through EGF receptor(EGFR), we evaluated the effects of long term incubation of cellswith sNTSR3 (24 h) on EGF-induced stimulation of Erk1/2 and cellproliferation. As illustrated in Fig. 1D, preincubation of cells withsNTSR3 for 24 h did not modify the EGF-induced phosphorylationof Erk1/2 after 5 min or 24 h. In the same way, sNTSR3 was unableto activate or modulate the EGF-induced phosphorylation of EGFR(Fig. 1E). As expected from the latest observation, the action of EGFon HT29 cell growth was not affected by sNTSR3 (EGF + sNTSR3)which was by itself devoid of activity on cell proliferation (Fig. 1F).

3.2. Internalization properties of sNTSR3 into HT29 cells

In order to determine whether sNTSR3 is internalized, we havetested the ability of the acid-NaCl wash to remove all the surface-bound ligand (Fig. 2A). HT29 cells were incubated with 125I-sNTSR3(0.5 nM) at 37 ◦C with or without preincubation for 30 min with0.45 M sucrose, a general inhibitor of endocytosis. The bindingreactions were terminated at the indicated times by removing the

free ligand and washing. Cell-associated radioactivity was countedeither directly or after treatment with 0.5 M NaCl, pH 4, for 2 min at37 ◦C. The amount of 125I-sNTSR3 bound to cells in the presence ofthe inhibitor represented only about 50% of the binding observed

Page 4

F. Massa et al. / The International Journal of Biochemistry & Cell Biology 45 (2013) 952– 959 955

Fig. 1. Characterization of sNTSR3 properties in HT29 cells. (A) Cells were serum-starved, pre-incubated or not with sucrose (0.45 M for 1 h) and incubated in the absence(Cont) or in the presence of NT (10−8 M) or carbachol (100 �M) or PGE2 (1 �M) or PMA (2.5 �g/L) for 6 h at 37 ◦C. Supernatants were collected and total proteins wereprecipitated and subjected to Western blot using antibody directed against NTSR3. (B) Cell homogenates (50 �g) were incubated in 50 mM Tris–HCl, pH 7.5, containing 0.2%BSA at 25 ◦C for 30 min with 0.4 nM 125I-sNTSR3 (200 Ci/mmol) in the absence or in the presence of increasing concentrations of sNTSR3 (closed circles), EGF (open circles)or NT (stars). (C) Cell homogenates (50 �g) were incubated as above with 0.2 nM 125I-EGF (600 Ci/mmol) in the absence or in the presence of increasing concentrations ofsNTSR3 (open circles) or EGF (closed circles). Data are expressed as the percent of binding obtained in the absence of unlabeled ligands and are mean from three independentexperiments performed in triplicate. (D) Cells were serum-starved and incubated or not with sNTSR3 (10−8 M) for 24 h, in the absence or in the presence of EGF (10−8 M) for5 min or 24 h. Cell lysates were subjected to Western blotting using antibodies directed against phospho-Erk1/2 or against total Erk1/2. The results are representative from3 independent experiments. E- HT29 cells were treated with sNTSR3 (10−8 M) or EGF (10−8 M) or both for 5 min. Equal amount of cell proteins were separated on SDS-PAGEand immunoblotted with antibodies directed against phospho-EGFR (Tyr 1173) or total EGFR. The results are representative from three independent experiments. (F) Cellswere serum-starved and incubated in the absence (SVF 0%) or in the presence of sNTSR3 (10−8 M) or EGF (10−8 M) or sNTSR3 with EGF for 96 h at 37 ◦C. Cells number wasd media7

i1

6fcaibfitewi

ifa

etermined using the fluorometric cell proliferation Assay Kit (CellTiter-Blue). Each5% percentiles, and 5% and 95% percentiles. **p < 0.01.

n the control experiment. When HT29 cells were incubated with25I-sNTSR3 without sucrose, acid-NaCl treatment removed about0% of the 125I-sNTSR3 bound to the cells corresponding to the sur-ace bound ligand (Fig. 2A), the radioactivity remaining into theells (about 40%) corresponded to the amount of internalized ligandt equilibrium. When HT29 cells were preincubated with sucrose,nternalization was completely blocked since all the radioactivityound to the cells was removed by the acid-NaCl wash. We con-rm that sNTSR3 can bind to HT29 cells and that about 40% ofhe bound ligand is sequestrated at 37 ◦C. We conclude from thesexperiments that the acid-NaCl washing technique is an efficientay to eliminate surface-bound ligand and that sucrose block the

nternalization of sNTSR3.

To further determine the amount of surface bound and internal-

zed ligand, we first incubated HT29 cells with 125I-sNTSR3 (0.5 nM)or 1 h at 4 ◦C, the excess of ligand was removed by washing at 4 ◦Cnd cells were incubated at 37 ◦C for indicated times. At the end of

n value is obtained from 6 to 12 determinations, box plots show medians, 25% and

the incubation, cells were washed with 0.5 M NaCl, pH 4, for 2 minat 37 ◦C. The radioactivity recovered in the washing buffer (the sur-face bound ligand) and from the cell lysate (internalized ligand)was counted (Fig. 2B). In this case the maximal amount of seques-trated ligand was 20% at the equilibrium (60 min) (Fig. 2B). Weconfirmed that part of the bound ligand is internalized whateverthe experimental approach.

To visualize sNTSR3 internalization in HT29 cells, whole cellswere incubated for 30 min at 37 ◦C in the presence of 10 nM Bodipy-sNTSR3, washed with the acid-NaCl buffer to remove extracellularbound ligand and examined by confocal microscopy. HT29 cellsincubated in these conditions exhibited intracellular clusters ofbound fluorescent ligand (Fig. 2C). This Fluo-sNTSR3 labeling was

specific since it was entirely competed by an excess (1 �M) ofnon-fluorescent sNTSR3 (Fig. 2C). Intracellular Fluo-sNTSR3 label-ing was also totally abolished when the incubation was carried outin the presence of 0.45 M sucrose which by itself did not inhibit

Page 5

956 F. Massa et al. / The International Journal of Bioch

Fig. 2. Internalization properties of sNTSR3 into HT29 cells. (A) HT29 cells werepreincubated in the absence (circles) or in the presence (squares) of 0.45 M sucrosefor 30 min at 37 ◦C. Then cells were incubated for indicated times at 37 ◦C with0.5 nM 125I-sNTSR3 in the absence (total binding) or in the presence (non-specificbinding) of 1 �M sNTSR3. At the end of incubations, cells were washed twice withthe binding buffer and immediately recovered to be counted in a g counter (closedsymbols) or incubated with the acid-NaCl buffer for 2 min before being washedtwice with the binding buffer and recovered for counting (open symbols). Data areexpressed as percent of maximal specific binding (total minus non specific binding)obtained after 60 min in the absence of both sucrose and acid-NaCl washing (closedcircles). (B) HT29 cells were incubated with 0.5 nM 125I-sNTSR3 in the absence (totalbinding) or in the presence (non specific binding) of 1 �M sNTSR3 for 60 min at 4 ◦C.After removing the excess of ligand, cells were incubated at 37 ◦C for indicated times.At the end of incubations, cells were washed with the acid-NaCl buffer for 2 min andthe radioactivity present in the washing buffer and in the cell lysate was counted.Data are expressed as percent of the total binding (radioactivity of the washingplus radioactivity of the cell lysate). (C) HT29 cells were incubated for 30 min at37 ◦C with 10 nM Bodipy-sNTSR3 (green) in the absence or in the presence of 1 �MsNTSR3 or 0.45 M sucrose. At the end of the incubation, cells were washed twicewith acid-NaCl buffer and the fluorescence remaining inside cells was visualized byconfocal microscopy. Scale bar: 10 �m. (D) HT29 cells were incubated for 30 minat 37 ◦C with 10 nM Bodipy-sNTSR3 (green) in the presence of 1 �M LysotrackerRed-DND-99. Cells were washed and visualized as above. Scale bar: 10 �m. (E)

emistry & Cell Biology 45 (2013) 952– 959

the binding of sNTSR3 to HT29 cells (Fig. 2A) indicating that theinternalization process of sNTSR3 was clathrin dependent.

To examine the fate of internalized sNTSR3, we performed incu-bation of Fluo-sNTSR3 in the presence of DES-Red LysotrackerTM

and observed that in these conditions (45 min at 37 ◦C), part ofsequestrated Fluo-ligand was recovered into lysosomes (Fig. 2Dand E). Then, we assessed the integrity of internalized Fluo-sNTSR3after various times of incubation in HT29 cells. For this purpose,sequestrated Fluo-sNTSR3 was extracted from cells and visualizedusing the ProEXPRESS proteomic Imaging System after SDS-PAGE.The point 5 min was considered as the minimum time to recover asufficient amount of internalized Fluo-sNTSR3 and then was usedas the control to evaluate the amount of remaining intact Fluo-sNTSR3 after 30 and 45 min. As shown in Fig. 2E (insert), the amountof intact Fluo-sNTSR3 decreased up to 45 min. At this time, about60–70% of Fluo-sNTSR3 was degraded, thus confirming its localiza-tion into lysosomes.

3.3. Signaling of sNTSR3 in HT29 cells

3.3.1. sNTSR3 increases intracellular calcium concentration andinduces pKC ̨ translocation

Numerous effectors modify the intracellular calcium concen-tration and activate PKCs. Therefore, we have chosen to investigatethese classical pathways first. As shown in Fig. 3A, sNTSR3 (10 nM)induced a significant increase of intracellular calcium. Interest-ingly, short time repetition of sNTSR3 application resulted in theloss of response suggesting desensitization of the system (Fig. 3A).However a 10 min lap-time was enough to recover an almost fullresponse. Then, we tested both the inactive NT analog NT(1–8) andNT known to activate intracellular calcium in HT29 cells (Bozouet al., 1989). NT(1–8) (100 nM) was unable to increase the calciumratio whereas NT (10 nM) strongly increased the response (Fig. 3A),even immediately after the application of sNTSR3. Several repeatedapplications of NT also decreased the response as expected from adesensitization process already described for the peptide (Bozouet al., 1989). This was the first experiment demonstrating thatsNTSR3 was a functional protein able to increase intracellular cal-cium concentration in HT29 cells.

Since PKCs are known to be involved in the regulation of calciuminflux and to translocate to the plasma membrane upon activa-tion, we tested the ability of sNTSR3 to modulate the intracellularlocalization of the representative PKC�. As shown in Fig. 3B and C,the pattern of the labeling of PKC� was modified after incubationof HT29 cells with 10 nM sNTSR3 for 15 min. This was visualizedby the increase of PKC� location close to the plasma membranestained using Fluo-WGA and analyzed by plot profile of both label-ing (Fig. 3C).

3.3.2. FAK-dependent stimulation of the PI3 kinase pathway bysNTSR3 in HT29 cells

Since PI3 kinase is known to regulate calcium signaling, we won-dered whether sNTSR3 might modulate Akt activity. When HT29cells were incubated with 10 nM sNTSR3, the phosphorylation ofAkt was rapidly and transiently enhanced between 2 and 10 min(Fig. 4A), with a maximal stimulation by a factor of 3 after 5 min of

incubation as illustrated by the mean of 8 independent determina-tions in Fig. 4B (p < 0.05). Note that incubation of the buffer alone(PBS, BSA 0.1%) during the same periods of time was without effecton Akt phosphorylation (Fig. 4A, mock).

Magnification of images from C, arrows: co-labeling, scale bar: 10 �m. Insert:sequestrated Bodipy-sNTSR3 was extracted from cells directly with SDS-PAGE solu-bilizing buffer and subjected to Western blot and visualized using the ProEXPRESSProteomic Imaging System from Perkin.

Page 6

F. Massa et al. / The International Journal of Biochemistry & Cell Biology 45 (2013) 952– 959 957

Fig. 3. sNTSR3 stimulates intracellular calcium and PKC� translocation. (A) HT29cells were incubated with sNTSR3 (10−8 M), or NT (10−8 M), or NT(1–8) (1 �M).The intracellular calcium amount was determined as described in Section 2. Dataare mean of 40 recorded cells from one representative experiment out of threeindependent experiments. (B) Immunocytochemistry was performed on HT29 cellslabeled with Alexa-594 WGA (red) to visualize the plasma membrane, using anti-body against PKC� (green), after stimulation with sNTSR3 (10−8 M) for 15 min oridentical volume of PBS-BSA 0.1%. Scale bar, 10 �m. The results are representa-tive from 3 independent experiments. (C) Analysis of the labeling density of PKC�as

aoa1FTtaptp

Fig. 4. Signaling of sNTR3 in HT29 cells. (A) HT29 cells were serum-starved and stim-ulated with sNTSR3 (10−8 M) or an equivalent volume of PBS-BSA 0.1% (Mock) forindicated times. Cell lysates were subjected to Western blotting using antibodiesdirected against phospho-Akt or against total Akt. (B) Results were standardizedusing total anti-Akt and are expressed as median value obtained from 6 to 12determinations, box plots show medians, 25% and 75% percentiles, and 5% and95% percentiles. **p < 0.01, ns: non-significant. (C) HT29 cells were serum-starvedand stimulated with sNTSR3 (10−8 M) for indicated times. Cells were solubilizedas describes in Section 2. Solubilized cell extracts were immunoprecipitated byusing the anti-FAK antibody. Proteins precipitated were analyzed by SDS-PAGE,transferred onto nitrocellulose and immunoblotted using the anti-phospho-tyrosineantibody and the anti-FAK antibody. For Src activation, cells were serum-starvedand stimulated with sNTSR3 (10−8 M) for indicated times. Cell lysates were sub-jected to Western blotting using antibodies directed against phospho-Src or againsttotal Erk1/2. The results are representative from three independent experiments.(D) Results of FAK phosphorylation were standardized using total anti-FAK andexpressed as median value obtained from 6 to 12 determinations, box plots showmedians, 25% and 75% percentiles, and 5% and 95% percentiles. ***p < 0.005. (E) HT29cells were serum-starved, pre-incubated or not with FAK Inhibitor II (10 �M for1 h) and incubated in the absence or in the presence of sNTSR3 (10−8 M) for 5 min.Cell lysates were subjected to Western blotting using antibodies directed againstphospho-Akt or against total Akt. The results are representative from three indepen-

nd WGA through the cells (white bars in B) for the control (left panels) and thetimulated (right panels) conditions.

In order to identify upstream signaling pathways leading to thectivation of the PI3 kinase, we analyzed the effects of sNTSR3n the complex focal adhesion kinase (FAK)-Src known to playn important role in this pathway. Incubation of HT29 cells with0 nM of sNTSR3 enhanced the level of phosphorylation of bothAK and Src at 5 min that was maintained up to 60 min (Fig. 4C).hese effects were strongly significant as shown by the represen-ation illustrated in Fig. 4D for phospho-FAK (n = 8, p < 0.005). The

ctivation of FAK was likely involved in the effect of sNTSR3 on Akthosphorylation since the use of the FAK inhibitor II totally blockedhe response of sNTSR3 (Fig. 4E). Finally, to ensure that the phos-horylation of Akt was not dependent only on the cell line used, we

dent experiments. (F) Lovo cells were serum-starved and stimulated with sNTSR3(10−8 M) for indicated times. Cell lysates were subjected to Western blotting usingantibodies directed against phospho-Akt or against total Akt.

verified that sNTSR3 (10 nM) was also able to activate Akt as mea-sured by its phosphorylation in the human Lovo colon carcinomacell line (Fig. 4F).

These results confirmed that the soluble form of NTSR3 was anactive molecule that can efficiently stimulate the PI3 kinase path-way in the HT29 cell line by a mechanism dependent on FAK.

4. Discussion

This work provides the first evidence that the soluble form ofNTSR3 is an active protein likely able to play an important rolein the epithelial HT29 cell line through the PI3 kinase pathway.

Page 7

9 f Bioch

Nr22silw

ai(Et(wodoab

slobaltbtbotHmratl4vtc

dAaooawtpts

itibtit22

58 F. Massa et al. / The International Journal o

TSR3, also called sortilin, is dedicated to several functions from aole as a receptor or a co-receptor (Martin et al., 2003; Nykjaer et al.,004), to a role in the sorting to plasma membrane (Mazella et al.,010), to lysosomes (Lefrancois et al., 2003) or to the regulatedecretion (Chen et al., 2005). However, although the shedding ofts extracellular domain has been already described in several cellines including colonic cancer cells (Navarro et al., 2002), nothing

as known about a possible function of this protein.We demonstrated that sNTSR3 binds to HT29 homogenates with

relatively good affinity of about 10 nM, a concentration recoveredn the medium following constitutive and PKC-dependent sheddingNavarro et al., 2002). This binding is selective since neither NT norGF competed with iodinated sNTSR3. We observed a 35% inhibi-ion with 1 �M NT likely due to the ability of NT to bind sNTSR3Navarro et al., 2002). On the other hand, we verified that sNTSR3as not able to efficiently compete with iodinated EGF since we

bserved an affinity of about 1 �M for sNTSR3 to EGFR. This resultemonstrates that EGFR is not the receptor that triggers the effectsf sNTSR3 since the results of the present work were obtained using

concentration of 10 nM sNTSR3, a concentration not sufficient toind and activate EGFR.

The amount of sNTSR3 internalized was 40% of the total boundNTSR3 when association and internalization of the iodinatedigand was performed at 37 ◦C. When the association was carriedut at 4 ◦C followed by the washing of the excess of ligand andy the incubation at 37◦ to follow the internalization process, themount of sNTSR3 sequestration into HT29 cells was 20%. In theatter condition, since the excess of ligans is removed, the radioac-ivity bound to the cell surface either internalizes or dissociatesecause the equilibrium is lost. In these conditions, the internaliza-ion process is in competition with the dissociation of the surfaceound ligand characterized by a decrease of the apparent amountf sequestrated iodinated sNTSR3 from 40 to 20%. On another hand,he fact that the sNTSR3 protein was efficiently internalized inT29 cells and partly recovered into lysosomes indicates that theembrane protein which recognize sNTSR3 behaves as a functional

eceptor whose presence at the cell surface can be modulated by clathrin-dependent mechanism. Although a large part of seques-rated sNTSR3 was degraded as confirmed by its co-localization intoysosomes, the amount of ligand remaining intact (about 30%) after5 min at 37 ◦C suggests that sNTSR3 could be sorted to recyclingesicles or to other particular intracellular compartments in ordero display some cellular functions. Further identification of theseompartments should give evidence to validate this hypothesis.

Since sNTSR3 specifically binds to HT29 cells, we attempted toefine intracellular pathways that are activated by the protein.lthough sNTSR3 recognizes EGFR with a poor affinity (IC50 ofbout 1 �M), we first verified if sNTSR3 could mimic the actionf EGF on the MAP kinases Erk1/2 and EGFR phosphorylation, andn cell growth. sNTSR3 was unable by itself to stimulate Erk1/2, toctivate EGFR nor to induce HT29 cell proliferation. The proteinas also unable to modulate the effects of EGF on the activa-

ion of MAP kinases, on the phosphorylation of EGFR and on cellroliferation. These results definitively confirm that EGFR is nothe receptor involved or modulated in the cellular functions ofNTSR3.

The first experimental data revealing that sNTSR3 is functionals given by its ability to increase intracellular calcium concen-ration and to induce plasma membrane translocation of PKC�n HT29 cells. Interestingly, this stimulation can be desensitizedy repeated incubations with sNTSR3, the system needing upo 10 min to resensitize. This frequently observed phenomenon

s generally due to both sequestration and uncoupling of func-ional receptors including G-protein coupled receptors (Evron et al.,012) and low-density lipoprotein lipase receptor family (Hussain,001).

emistry & Cell Biology 45 (2013) 952– 959

In addition, we evidenced that sNTSR3 significantly activates thePI3 kinase pathway by phosphorylation of Akt and we observedthat this activation was dependent on FAK phosphorylation andinvolved also Src (Fig. 4). The activation of such a pathway is impli-cated in a variety of distinct survival mechanisms of cancer cells(for review, see Buchheit et al., 2012).

We investigated the possible function of sNTSR3 since the mech-anism of action of this protein resembles to the activation of EGFRligands (i.e.: shedding from the plasma membrane, release in thecirculation and recognition of a membrane receptor). However, thereceptor for sNTSR3 remains to be identified. This receptor is prob-ably neither EGFR nor a receptor for NT since neither EGF nor NTis able to compete for the binding of iodinated sNTSR3 (Fig. 1). Wecannot exclude the possibility that sNTSR3 might bind to an EGFor a NT receptor, on a site which is not inhibited by the corre-sponding ligands. However, NT is unable to activate the shedding ofsNTSR3 and application of the soluble protein does not desensitizethe NT-induced response on intracellular calcium concentrationsuggesting that sNTSR3 functions independently from NT.

In conclusion, our results indicate that sNTSR3 can be releasedfrom cancer cells by various effectors able to activate PKCs andmatrix metalloproteases and that this release leads to the activa-tion of survival pathways of target cells. These activations couldbe involved in the regulation of tumor cell growth by increasingor decreasing the growth factors effects or dissemination of cancercells.

Acknowledgement

This work was supported by the Centre National de la RechercheScientifique.

References

Botto JM, Chabry J, Sarret P, Vincent JP, Mazella J. Stable expression of the mouselevocabastine-sensitive neurotensin receptor in HEK 293 cell line: binding prop-erties, photoaffinity labeling, and internalization mechanism. Biochemical andBiophysical Research Communications 1998;243:585–90.

Bozou JC, Rochet N, Magnaldo I, Vincent JP, Kitabgi P. Neurotensin stimulatesinositol trisphosphate-mediated calcium mobilization but not protein kinaseC activation in HT29 cells. Involvement of a G-protein. Biochemical Journal1989;264:871–8.

Buchheit CL, Rayavarapu RR, Schafer ZT. The regulation of cancer cell death andmetabolism by extracellular matrix attachment. Seminars in Cell and Develop-mental Biology 2012;4:402–11.

Carraway RE, Plona AM. Involvement of neurotensin in cancer growth: evi-dence, mechanisms and development of diagnostic tools. Peptides 2006;27:2445–60.

Chen ZY, Ieraci A, Teng H, Dall H, Meng CX, Herrera DG, et al. Sortilin controls intra-cellular sorting of brain-derived neurotrophic factor to the regulated secretorypathway. Journal of Neuroscience 2005;25:6156–66.

Comens PG, Simmer RL, Baker JB. Direct linkage of 125I-EGF to cell surface recep-tors. A useful artifact of chloramine-T treatment. Journal of Biological Chemistry1982;257:42–5.

Dal Farra C, Sarret P, Navarro V, Botto JM, Mazella J, Vincent JP. Involvement ofthe neurotensin receptor subtype NTR3 in the growth effect of neurotensin oncancer cell lines. International Journal of Cancer 2001;92:503–9.

Dobner PR. Neurotensin and pain modulation. Peptides 2006;27:2405–14.Evron T, Daigle TL, Caron MG. GRK2: multiple roles beyond G protein-coupled recep-

tor desensitization. Trends in Pharmacological Sciences 2012;33:154–64.Hussain MM. Structural, biochemical and signaling properties of the low-density

lipoprotein receptor gene family. Frontiers in Bioscience 2001;6:D417–28.Kitabgi P, Herve D, Studler JM, Tramu G, Rostene W, Tassin JP. Neurotensin/dopamine

interactions. Encephale 1989;15(Spec No.):91–4.Lefrancois S, Zeng J, Hassan AJ, Canuel M, Morales CR. The lysosomal trafficking

of sphingolipid activator proteins (SAPs) is mediated by sortilin. EMBO Journal2003;22:6430–7.

Maoret JJ, Anini Y, Rouyer-Fessard C, Gully D, Laburthe M. Neurotensin and anon-peptide neurotensin receptor antagonist control human colon cancer cell

growth in cell culture and in cells xenografted into nude mice. InternationalJournal of Cancer 1999;80:448–54.

Martin S, Navarro V, Vincent JP, Mazella J. Neurotensin receptor-1 and -3 complexmodulates the cellular signaling of neurotensin in the HT29 cell line. Gastroen-terology 2002;123:1135–43.

Page 8

f Bioch

M

M

M

M

N

N

P

P

R

S

F. Massa et al. / The International Journal o

artin S, Vincent JP, Mazella J. Involvement of the neurotensin receptor-3 in theneurotensin-induced migration of human microglia. Journal of Neuroscience2003;23:1198–205.

azella J. Sortilin/neurotensin receptor-3: a new tool to investigate neurotensinsignaling and cellular trafficking? Cellular Signalling 2001;13:1–6.

azella J, Petrault O, Lucas G, Deval E, Beraud-Dufour S, Gandin C, et al. Spadin, asortilin-derived peptide, targeting rodent TREK-1 channels: a new concept inthe antidepressant drug design. PLoS Biology 2010;8:e1000355.

yers RM, Shearman JW, Kitching MO, Ramos-Montoya A, Neal DE, Ley SV. Cancer,chemistry, and the cell: molecules that interact with the neurotensin receptors.ACS Chemical Biology 2009;4:503–25.

avarro V, Vincent JP, Mazella J. Shedding of the luminal domain of the neurotensinreceptor-3/sortilin in the HT29 cell line. Biochemical and Biophysical ResearchCommunications 2002;298:760–4.

ykjaer A, Lee R, Teng KK, Jansen P, Madsen P, Nielsen MS, et al. Sortilin is essentialfor proNGF-induced neuronal cell death. Nature 2004;427:843–8.

ini A, Falciani C, Bracci L. Branched peptides as therapeutics. Current Protein andPeptide Science 2008;9:468–77.

opp E, Schneider A, Vogel P, Teschendorf P, Bottiger BW. Time course ofthe hypothermic response to continuously administered neurotensin. Neuro-

peptides 2007;41:349–54.

ostene WH, Alexander MJ. Neurotensin and neuroendocrine regulation. Frontiersin Neuroendocrinology 1997;18:115–73.

ehgal I, Powers S, Huntley B, Powis G, Pittelkow M, Maihle NJ. Neurotensin is anautocrine trophic factor stimulated by androgen withdrawal in human prostate

emistry & Cell Biology 45 (2013) 952– 959 959

cancer. Proceedings of the National Academy of Sciences of the United States ofAmerica 1994;91:4673–7.

Souaze F, Viardot-Foucault V, Roullet N, Toy-Miou-Leong M, Gompel A, Bruyneel E,et al. Neurotensin receptor 1 gene activation by the Tcf/beta-catenin pathwayis an early event in human colonic adenomas. Carcinogenesis 2006;27:708–16.

Thiery JP. Epithelial-mesenchymal transitions in tumour progression. NatureReviews Cancer 2002;2:442–54.

Vincent JP, Mazella J, Kitabgi P. Neurotensin and neurotensin receptors. Trends inPharmacological Sciences 1999;20:302–9.

Warhurst G, Fogg KE, Higgs NB, Tonge A, Grundy J. Ca(2+)-mobilising agonists poten-tiate forskolin- and VIP-stimulated cAMP production in human colonic cellline, HT29-cl. 19A: role of [Ca2+]i and protein kinase C. Cell Calcium 1994;15:162–74.

Yamada M, Ohata H, Momose K, Yamada M, Richelson E. Pharmacological char-acterization of SR 48692 sensitive neurotensin receptor in human pancreaticcancer cells, MIA PaCa-2. Research Communications in Molecular Pathology andPharmacology 1995;90:37–47.

Zhao D, Zhan Y, Koon HW, Zeng H, Keates S, Moyer MP, et al. Metalloproteinase-dependent transforming growth factor-alpha release mediates neurotensin-stimulated MAP kinase activation in human colonic epithelial cells. Journal of

Biological Chemistry 2004;279:43547–54.

Zhao D, Zhan Y, Zeng H, Koon HW, Moyer MP, Pothoulakis C. Neurotensin stimu-lates expression of early growth response gene-1 and EGF receptor throughMAP kinase activation in human colonic epithelial cells. International Journal ofCancer 2007;120:1652–6.