GPBAR1(TGR5) is highly expressed in human gastric cancers and its activation by selective or GPBAR1/FXR dual ligands promotes epithelial mesenchymal transition and tumor spreading

Barbara Renga*, Sabrina Cipriani #, Adriana Carino * , Silvia Marchianò * , Angela Zampella † and Stefano Fiorucci *

*Dipartimento di Scienze Chirurgiche e Biomediche, Università degli Studi di Perugia, Nuova Facoltà di Medicina e Chirurgia , Sant’Andrea delle Fratte,Perugia, Italy

#Dipartimento di Medicina, Università degli Studi di Perugia ,Nuova Facoltà di Medicina e Chirurgia, Perugia, Italy †Dipartimento di Farmacia, Università di Napoli Federico II, Napoli, Italy

Background. Gastritis associated with bile acids reflux is a risk factor in the development of intestinal metaplasia in the stomach and is believed to function as an initiator of gastric carcinogenesis. Bile acids activate nuclear and G-protein coupled receptors. GPBAR1 (also known as TGR5) is expressed in esophageal adenocarcinoma cell lines and its activation by TDCA increases cell proliferation. Whether GPBAR1 is expressed in gastric cancers and exerts a functional role in disease initiation or progression is controversial

Aim of the study. In this study, we have examined the expression of GPBAR1 in human gastric adenocarcinomas and examined whether its activation modulate the phenotypes of gastric cancer cell lines in vitro and in vivo in a model of mice peritoneal implantation and dissemination.

Material and methods. 35 patients with gastric adenocarcinoma were included in this study. Authorization for samples collection was granted to SF by the CEAS, Ethical committee heal System of Umbria region (Italy). An informed written consent was obtained by each patient before surgery. Gastric cell lines used in this study were MKN45 (intestinal type cancer) and MKN74 (undifferentiated gastric cancer).

Results. GPBAR1 staining was found in all patients with intestinal metaplasia and in patients intestinal but in only 20% of the diffuse subtype of adenocarcinomas. GPBAR1 expression was found to be associated with advanced cancer (stage III and IV) and was associated with lymphatic spreading. Expression of FXR, mRNA , and GPBAR1,mRNA and protein, was detected in several cancer cell lines, with MKN45 (intestinal cell type) having the higher expression of GPBAR1 in comparison to MKN74 (indifferentiated cell type). Exposure of MKN45 cells to TLCA, oleanoic acid (selective GPBAR1 ligand) or 6-ECDCA (a dual FXR and GPBAR1 ligand) increased the expression of several genes usually associated with a metastatic phenotype including KDKN2A, HRAS, IGB3, MMP10 and MMP13 and downregulated the expression of CD44 and FAT1 (among other) (p<0.01 versus untreated cells). Exposure of MKN45 to GPBAR1 ligands, including selective GPBAR1 ligands TLCA,TDCA and oleanoic acid, or dual FXR/GPBAR1 ligands, 6-ECDA and CDCA, increased cell migration, enhanced cell adhesion to the mouse peritoneum and wound healing. These effects correlated with increased expression of N-cadherin and vimentin and were blocked by cetuximab and by MEK inhibition. By using a model of peritoneum metastasis, we found that pretreating MKN45 cells with TLCA or 6-ECDCA increases the number of peritoneal metastasis by 30% and that this effect was abrogated by cetuximab. Because cetuximab is an EGF receptor(R) inhibitor, we have then examined molecular basis for interaction of GPBAR1 with EGFR and found that exposure of MKN45 cells to TLCA, increases the phosphorylation of EGFR in a time and concentration-dependent manner. Conclusions. GPBAR1 is highly expresses in gastric cancers. Activation of GPBAR1 in gastric epithelial cells promotes the epithelial mesenchymal transition. Taken together these data support a potential role for GPBAR1 antagonists in the treatment of gastric cancers.

Figure 1: (A-D)The expression of GPBAR1 was examinated in surgical samples, obtained from patients who underwent surgery for the treatment of gastric cancer, first by immunohistochemistry. We found that GPBAR1 was expressed either in non-neoplastic and neoplastic tissues, but the expression resulted slight up-regulated, although not significantly, in the neoplastic tissue in comparison to healthy gastric mucosa. (E-G) This finding was further confirmed by quantitative RT-PCR.The expression of the receptor was significantly higher in patients with advanced disease; (F) patients with stage III and IV had significantly higher expression in comparison with patients in stage I-II (p<0.05). Such difference was not observed in non-neoplastic areas (E) and that there was no difference among the two main histological phenotypes (i.e. intestinal vs diffuse) (G). Values are normalized to GAPDH and are expressed relative to those of positive controls(non-neoplastic tissues); The relative mRNA expression is expressed as 2(-ΔΔCt). These data support the concept that development of more advanced disease associated with higher expression of GPBAR1

Figure 2: Expression levels of GPBAR1and FXR were evalueted both by ReaL-Time PCR (mRNA) (A-B) and Western Blot analysis (protein) (C) in MKN45 and MKN74 human cell lines. MKN45 cells were characterized by a slightly higher GPBAR1 mRNA expression than MKN74, FXR resulted express at very low levels in both gastric cell lines. Values are normalized to GAPDH, the relative mRNA expression is expressed as 2(-ΔΔCt). Western Blot data was normalized with α-Tubulin expression. (D-E) Transwell migration assay: MKN45 were seeded in a 6-well plate, starved and then primed with 10µM of Cholic Acid, Chenodeoxycholic Acid, Ursodeoxycholic Acid, Taurolithocholic Acid, Taurodeoxycholic Acid and 6-ECDCA (D). In an another experimental setting, cells were treated with Taurolithocholic Acid (1, 10 and 100µM), Taurodeoxycholic Acid (1, 10 and 100µM), 6-ECDCA (1, 10 and 50µM) for 72 hours (E)After the incubation period 200 µl of cell suspension (2,5x105cells/ml) were seeded in the upper chamber of the transwell. The lower chamber was filled with culture medium containing 10% FBS as a chemoattractant. The invaded cells were fixed with methanol and stained with 0.1 % Toluidine blue solution and counted under a light microscope (200×/10 fields for well). Experiments were performed in triplicate. All three ligands increased, in a dose-response manner, migration activity of gastric cells in comparison with control cells . (F) MKN45 cells were plated and treated as described and used in a cell adhesion to peritoneum experiment. On day 5, excised parietal peritoneum ( 1.6 cm2) was placed in a 24-well culture plate, filled with 1.0 ml of 1% BSA/RPMI 1640. Gastric cancer ∼cells were detached, fluorescently labeled with BCECF-AM (5 µM). 500µl of a suspension cells (5 x 105 cells/ml) were overlaid on the peritoneum in a 24-well plate, and incubated at 37°C for 60 minutes. The cells adherent to the peritoneum were lysed with 1.0 ml of TRIS 50 mM plus SDS 1%; fluorescence intensity was measured with a fluorescence spectrophotometer (Ex = 490 nm and Em = 520 nm). Experiments were conducted in triplicate. TLCA treatment of MKN45 cells significantly increased cellular adhesiveness of gastric tumor cells to murine parietal peritoneum in a dose dependent manner.

Figure 3: Scratch wound healing was used to determine migration behavior in MKN45 cell line left untreated or stimulated with CA, CDCA, TLCA, TDCA and 6E-CDCA 100 or 50 μM (A-B) In an another experimental setting, cells were treated with an increasing dose of the same compounds (1,10 and 100μM) (C). After treatments, the cell monolayers were scraped in a straight line using a p200 pipette tip in order to create a “scratch”. A wound was generated and imaged at 0 an 24 hours with a phase-contrast. Images obtained from each samples at both time points were analyzed using Image J software and migration areas were expressed in pixels. All experiments were performed at least in triplicate. Wound closure was significantly higher in treated cells, primarly with TLCA, TDCA and 6E-CDCA, and the migration increased in a dose-response manner.

Figure 4: To further characterize the effects of GPBAR1 ligation in MKN45 cells, was used a gene array designed to analyze the expression of genes related to tumor metastasis. The cDNA used for the array was obtained from MKN45 cells (1x106) plated, serum starved for 24 hours and then treated for 72 hours with 6-ECDCA 50µM, TLCA 100µM or OA 25µM. A comparative analysis of effects exerted by the three GPBAR1 agonists, demonstrated that the three agents shared a regulatory effect on a small number of genes. Array analysis was carried out with the online software RT2 Profiler PCR Array Data Analysis (http://pcrdataanalysis.sabiosciences.com/pcr/arrayanalysis.php).Up-regulated/down-regulated genes were those genes whose expressions had been altered by more than 1,9 fold.

Non tumor tissue A. Metaplastic area high expression

Low/absent expression

Tumor with high expression

Tumor with low expression

Non-neoplasticarea

I-II III-IV05

10152025303540

p=0.06

Cancer staging

GPB

AR1

m R

NA

I-II III-IV-10123456789

10 Primary tumor*p<0,5

Cancer staging

GPB

Ar1

exp

ress

ion

With Without0

10

20

30

Intestinal metaplasia

p=0.3

GPB

AR1

mRN

A

C.

D.

E.

F.

G.

h M MKN45 MKN740.0

0.5

1.0

GPB

AR

1m

RNA

rel.

expr

essi

on

HepG2 MKN45 MKN74

0.025

0.000

0.5

1.0

FXR

mR

NA re

l. ex

pres

sion

GPBAR1

α-Tubulin

MKN45 MKN74

1 10 100 1 10 100 1 10 500

100

200

300

400

NT TLCA TDCA 6-ECDCA

*

*

**

*

*

% v

s N

T ce

lls

NT CA CDCA 6-ECDCA UDCA TLCA TDCA0

100

200

300

*

*

*

**

(10m)

%

vs

NT

cells

1 10 1000

10

20

30

40

50

60

70

NT TLCA

*

*

%

vs

NT

cells

NT CA CDCA 6-ECDCA UDCA TLCA TDCA0

10

20

30

40

50

(10M)

*

*

**

%

vs

NT

cells

1 10 100 1 10 100 1 10 100 1 10 100 1 10 500

10

20

30

40

50

NT CA CDCA TLCA TDCA 6E-CDCA

**

*

* *

*

%

vs

NT

cells

NT CA CDCA 6-ECDCA UDCA TLCA TDCA0

10

20

30

40

50

(10M)

**

%

vs

NT

cells

TLCA

6-ECDCA OA

U:1D:3

U:0D:1U:5

D:6

U:2D:0

U:1D:0

U:11D:5

U:1D:0

U:18; D:11

U:7 ; D:10

U:8; D:7

CDKN2A

HRAS

ITGB3

MMP10

MMP3

CD44

FAT1

MMP13

MMP7

MTSS1

RPSA

WB: P-EGFR

WB: EGFR

NT 30’15’ 18h60’

TLCA

Figure 8: Effects of TLCA on EGFR phosphorylation in MKN45 cells. Cells treated with 100 µM of TLCA were harvested at the indicated time. Cell lysates were immunoblotted with antibodies against phosphorylated EGFR (pEGFR) and total EGFR. The exposure of MKN45 cells to TLCA, increases the phosphorylation of EGFR in a time-dependent manner.

B.

NT TLCA TDCA0.0

0.5

1.0

1.5

(100 M)

Vim

entin

mRN

A re

l.exp

ress

ion

NT TLCA TDCA0.0

0.5

1.0

1.5

(100 M)

*E-ca

derin

mR

NA

rel.e

xpre

ssio

n

NT TLCA TDCA0

1

2

3

(100 M)

*

N-c

ader

inm

RN

A re

l.exp

ress

ion

p < 0.05p < 0.05

Figure 5: We have then investigatedby a RT-PCRanalysis, whether GPBAR1 activation induces phenotypic changes consistent with acquisition of a mesenchymal phenotype. MKN45 cells were plated and after 24 hours of starvation were stimulated for 72 hours with TLCA and TDCA (100 µM). We found that exposing MKN45 cells to TLCA resulted in 50% percent reduction of E-cadherin mRNA (n= 4; P<0.05) and ≈ 100% increase of N-cadherin(n=4; P<0.05) (A-B). No change was observed in the expression of vimentin (C).

NT TLCA - TLCA - TLCA0

10

20

30 *

*

#

*

#

Cetuximab MEK Inhib

Cel

l / fi

eld

20X

Figure 6: Blockade of EGFR signaling abrogate the increased MKN45 migration activity TLCA-induced. MKN45 cells were treated with TLCA, the EFGR inhibitor cetuximab (alone or in combination with TLCA) and the MAPK inhibitor U0126. Data obtained are reported as ∆% compared to NT cells (* vs NT, # vs TLCA). (Figure 6).

Control TLCA - TLCA0

5

10

15

20

25

**

*

Cetuximab

Num

ber o

f Nod

ules

Control TLCA - TLCA0

5

10

15

20

25

**

*

Cetuximab

Num

ber o

f Nod

ules

Control TLCA - TLCA-4

-3

-2

-1

0

Cetuximab

W

eigh

t ( g

)

Figure 7: We have investigated the role in cancer progression of the GPBAR1/EGFR signaling,in a murine model of peritoneal carcinomatosis. MKN45 cells were left untreated or pretreated for with TLCA, Cetuximab or with the combination of both for 3 days and then implanted in the peritoneum of NOD-SCID mice. (A-C) MKN45 cells stimulated with TLCA significantly increased the number and the volume of peritoneal nodules in comparison to control cells, whereas the inhibition of EGF-R pathway by cetuximab almost completely abrogated MKN45 potential to form peritoneal nodules . Cetuximab reverted the increased MKN45 cells metastatic ability induced by TLCA priming

NT CA CDCA TLCA TDCA 6-ECDCA 100μM 100μM 100μM 100μM 50μM

24h

0

h

C.

B.

A.

D.

E.

F.

NT 15' 30' 60' 18H0

1

2

3

TLCA (100M)

P-EG

FR/E

GFR

B.

A.

C.

A. B. C. A. B. C.