IGF-II regulates metastatic properties of choriocarcinomacells through the activation of the insulin receptor

L.E. Diaz1, Y-C. Chuan2, M. Lewitt2, L. Fernandez-Perez3, S. Carrasco-Rodrıguez1,M. Sanchez-Gomez1,4 and A. Flores-Morales2

1Hormone Laboratory, Department of Chemistry, Universidad Nacional de Colombia, Bogota, Colombia; 2Department of Molecular

Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden; 3Department of Clinical Sciences, Molecular Endocrinology Group,

University of Las Palmas de G.C., Canary Institute for Cancer Research, RTICCC, Spain

4Correspondence address. Tel: þ57-1-316-5000 ext. 14466; Fax: þ57-1-316-5220; E-mail: mysanchezd@unal.edu.co

Choriocarcinoma is a highly malignant tumor that can arise from trophoblasts of any type of gestational event but most often from

complete hydatidiform mole. IGF-II plays a fundamental role in placental development and may play a role in gestational

trophoblastic diseases. Several studies have shown that IGF-II is expressed at high levels in hydatidiform moles and choriocarci-

noma tissues; however, conflicting data exist on how IGF-II regulates the behaviour of choriocarcinoma cells. The purpose of this

study was to determine the contribution of the receptors for IGF-I and insulin to the actions of IGF-II on the regulation of

choriocarcinoma cells metastasis. An Immuno Radio Metric Assay was used to analyse the circulating and tissue levels of

IGF-I and IGF-II in 24 cases of hydatidiform mole, two cases of choriocarcinoma and eight cases of spontaneous abortion at

the same gestational age. The JEG-3 choriocarcinoma cell line was used to investigate the role of IGF-II in the regulation of

cell invasion. We found that mole and choriocarcinoma tissue express high levels of IGF-II compared to first trimester placenta.

Both IGF-I and IGF-II regulate choriocarcinoma cell invasion in a dose dependent manner but through a different mechanism.

IGF-II effects involve the activation of the InsR while IGF-I uses the IGF-IR. The positive effects of IGF-II on invasion are the

result of enhanced cell adhesion and chemotaxis (specifically towards collagen IV). The actions of IGF-II but not those of

IGF-I were sensitive to inhibition by the insulin receptor inhibitor HNMPA(AM)3. Our results demonstrate that the insulin

receptor regulates choriocarcinoma cell invasion.

Keywords: choriocarcinoma; hydatidiform mole; metastasis; IGF-II; insulin receptor

Introduction

Gestational trophoblastic neoplasias comprise a group of

pregnancy-related disorders that include pre-malignant complete and

partial hydatidiform mole as well as malignant lesions such as

invasive mole, choriocarcinoma, and placental site trophoblastic

tumour (PSTT) (Altieri et al., 2003). Complete moles are diploid

and nearly always androgenetic in origin. By contrast, partial moles

are triploid, consisting of one maternal and two paternal sets of

chromosomes. After uterine evacuation, 10–20% of complete and

0–5% of partial moles undergo malignant change. For complete

and, to a lesser extent, partial moles this malignant change includes

invasive mole, choriocarcinoma and PSTT (Soper, 2006).

Choriocarcinoma is a highly malignant tumor that can arise from

trophoblasts of any type of gestational event but most often from com-

plete hydatidiform mole (Seckl et al., 2000). The pathogenesis of these

tumors is poorly understood. Due to the complete absence of maternal

genome in complete hydatidiform mole, the function of paternally

expressed genes has been studied in relation to choriocarcinoma

development (Lustig-Yariv et al., 1997; He et al., 1998; Kim et al.,

2003). IGF-II is expressed in many tissues from the paternally

derived allele. Several studies have shown that IGF-II is expressed

in hydatidiform moles and choriocarcinoma tissues although this

expression is not related to paternal imprinting (Kim et al., 2003).

Therefore, IGF-II appears to play an important role in the pathogenesis

of choriocarcinoma in addition to its well known function in the

regulation of normal placenta development (Kim et al., 2003).

IGF-I, IGF-II, the receptors for IGF-I, IGF-II and insulin as well as

several IGF binding proteins are produced by trophoblastic cells,

establishing a complex network of autocrine and paracrine actions

that is yet poorly understood (Han and Carter, 2000; Constancia

et al., 2002). Specifically, conflicting reports exist on how IGF-II

regulates the behaviour of choriocarcinoma cells (Korner, 1995;

Mckinnon et al., 2001). Initial reports implicated the Mannose 6-P/

IGF-II receptor in proliferative actions of IGF-II on choriocarcinoma

cell lines (Mckinnon et al., 2001) but later reports have clearly demon-

strated that this receptor has mainly inhibitory effects on IGF-II

actions (Li and Sahagian, 2004). Here, we have used the JEG-3 chor-

iocarcinoma cell line to study the contribution of the receptors for

IGF-I and insulin to the actions of IGF-II, specifically on cell

adhesion, invasion and proliferation. Our findings indicate that

IGF-I and IGF-II use different receptors to regulate the invasive

properties of choriocarcinoma cells although they share downstream

mechanisms involving ERK and PI3- kinase activation.

Materials and Methods

Subjects and clinical samples

Twenty-four patients (mean age 21.9+7.8 years) with complete hydatidiform

mole were diagnosed in gestational weeks eight through 20 with clinical,

# The Author 2007. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For

Permissions, please email: journals.permissions@oxfordjournals.org 567

Molecular Human Reproduction Vol.13, No.8 pp. 567–576, 2007

Advance Access publication June 6, 2007 doi:10.1093/molehr/gam039

by guest on February 3, 2014http://m

olehr.oxfordjournals.org/D

ownloaded from

histopathology, microsatellite and gonadotropin (hCG) analysis; and two

patients diagnosed with choriocarcinoma were included in the study. Placenta

from eight patients (mean age 27.4+7.5 years) with non-molar abortion

during weeks eight through 18 and confirmed negative for choriocarcinoma

or hydatidiform mole were used in the study. All patients were selected from

hospitals in Bogota, Colombia and written consent was obtained. All

procedures were approved by the local ethics committee. Blood samples

were obtained and serum was separated by centrifugation. Tissues were

obtained by curettage and maintained at 2708C until further analysis.

Materials

Recombinant human IGF-I was obtained from Genentech (San Francisco, CA),

recombinant human IGF-II from Upstate Biotechnology, Inc. (Lake Placid,

NY), recombinant human insulin from Sigma Chemicals (St Louis, MI) and

recombinant IGFBP-1 was purchased from Upstate Biotechnology, Inc.

(Lake Placid, NY). The monoclonal mouse antihuman IGF-I receptor antibody

(aIR3) and polyclonals against subunit b-IGF-IR, subunit b-insulin receptor,

Akt 1/2, phospho-Akt (Ser 473), Erk 1/2 were purchased from Santa Cruz Bio-

technology (Santa Cruz, CA). A polyclonal goat antihuman a5b1 antibody

(MAB1969) (Conforti et al., 1989) was purchased from Chemicon Inter-

national Inc. (Mississauga, ON). U0126, a specific inhibitor of MEK kinase,

was purchased from Promega (Madison, WI). LY294002, a PI3 kinase inhibitor

was obtained from Cell Signalling (Beverly, MA). HNMPA(AM)3

[Hydroxy-2-naphthalenylmethylphosphonic acid tris acetoxymethyl ester], an

inhibitor of insulin receptor (InsR) tyrosine kinase activity (Saperstein et al.,

1989), was purchased from Alexis, Biochemicals (Alexis, San Diego, CA).

Mouse monoclonal antibody (Clon 4G10) against phosphotyrosine and

IGF-II neutralizing antibody S1F2 were purchased from Upstate Biotechnol-

ogy, Inc. (Lake Placid, NY). A polyclonal rabbit antihuman anti-p-Erk 1/2

(Thr202/Tyr204) was obtained from Cell Signalling Technology (Beverly,

MA). Goat antimouse IgG (H þ L) was obtained from Southern Biotech

(USA).

Measurement of IGFs

Total IGF-I, IGF-II and IGFBP-1 concentrations were determined by immunor-

adiometric assay (IRMA) using a commercial kit DSL (Webster, TX; IGF-I

DSL-10-5600, IGF-II DSL-10-2600 and IGFBP-1 DSL-10-7800). Standards

and samples were tested in duplicate. One part frozen tissue was weighed,

thawed and homogenized in two parts 0.06 M Tris-hydrochloride (pH 7.2) in

a Polytron homogenizer (Kinematica Gmbh, Luzern, Switzerland) at 48C.

Total IGF-I was extracted from crude tissue homogenate with buffer, according

to the manufacturer’s instructions. Total protein in the homogenates was deter-

mined by the Bradford method.

Cell culture and treatments

The choriocarcinoma cell line JEG-3 was maintained as previously described

(Zhang and Shiverick, 1997). For the experiments, cells were sub-cultured in

60 mm culture for 24–48 h before the experiments were performed. Stimu-

lation with 10 nM IGF-I, 10 nM IGF-II, 10 nM insulin or 10 mg/ml IGFBP-1

was performed on cells kept in serum-free media for 12 h. LY294002

(40 mM), U0126 (20 mM) and HNMPA(AM)3 (10, 50 or 100 mM) were

added to the cells 2 h prior to stimulation.

Western blotting

JEG-3 cells were cultivated and treated as described above. Cells were har-

vested using RIPA buffer (Tris 50 mM, pH 7.4, NaCl 150 mM, Triton X-100

1%) supplemented with protease inhibitors and protein concentration was

determined with BCA Protein Assay Reagent (Pierce, USA). Equal amounts

of protein extracts were denatured and separated on a 12% SDS-PAGE gel

(Invitrogen Life Technologies, UK) and blotted onto a PVDF membrane

(Amersham, Life Science, UK). Membranes were then blocked in a

Tris-buffered saline solution with 5% BSA or 5% non-fat dry milk and

probed with primary antibody. The membranes were further probed with

HRP-conjugated goat anti-mouse/anti-rabbit to visualize the specific band.

Immunoprecipitation

Immunoprecipitations of cell extracts were performed by incubating lysates

with the indicated antibodies. The reaction mixture was gently rocked at 48Covernight. Protein G-agarose (Amersham, Life Science, UK) was used to

absorb immunocomplexes that were washed extensively with ice-cold RIPA

buffer containing 0.1% Triton X-100. The pellet was resuspended in loading

buffer, boiled for 5 min, and centrifuged at 14 000 rpm (5 000 g) for 5 min.

The supernatant was collected and subjected to western blotting.

Matrigel invasive assay

Matrigel invasion assays were performed as essentially described before

(Chuan et al., 2006). Matrigel was purchased from Beckton Dickinson (San

Jose, CA) and stored at 2208. 0.2 ml aliquot containing 200 000 cells was

seeded on the upper chamber of the Matrigel coated transwell filter. Serum-free

DMEM containing IGF-I, IGF-II or IGFBP-1 was added to the lower chamber

and incubated for 24 h at 378C in a humidified atmosphere of 5% CO2. Each

assay was carried out in triplicate on at least three different occasions.

Matrigel adhesion assay

Matrigel adhesion assay was performed as described (Zhang and Yeh, 2004),

with some modifications. JEG-3 cells were suspended in serum-free media.

The cell suspension (2.5 � 105 cells in 0.5 ml medium/well) was added to

24 wells pre-coated with Matrigel and allowed to attach by incubation at

378C for 60 min. Unattached cells were removed by repeated washing with

PBS. Attached cells were stained with 0.5% Crystal Violet for 10 min. After

washing with water, the stained cells were extracted with 0.25 ml of 10%

acetic acid and the absorbance of the dye extract was measured at 570 nm.

Cell proliferation assays

Cell proliferation was assessed using the MTT assay (Mosmann, 1983).

Measurements were done at the time of seeding (t ¼ 0) and 24 h after treat-

ment. Each plate was scanned and measured in a Spectra Shell microplate

reader (SLT, Austria) at 570 nm.

Cell motility assay

Cell migration assays were performed using modified Boyden chambers with a

6.5-mm diameter, 10-mm thickness, porous (8.0 mm) polycarbonate membrane

separating the two chambers (Transwellw; Costar, Cambridge, MA). Briefly,

the lower surface of the membrane was coated with fibronectin, collagen IV,

laminin or BSA (20 mg/ml in phosphate-buffered saline, pH 7.4 for 2 h at

378C). Excess ligand was removed, and the lower chambers were loaded

with medium DMEM with IGF-I (10 nM), IGF-II (10 nM), insulin (10 nM)

or IGFBP-1 (10 ng/ml). 100 000 serum-starved JEG-3 cells were plated on

the upper wells of transwell chambers and incubated at 378C for 24 h. Cells

on the upper surface of membranes were completely removed; the cells that

migrated on the lower surface of the filter were fixed in 4% formaldehyde,

stained with hematoxylin and counted under a microscope. Each assay was

carried out in triplicate on at least three different occasions. To verify

whether IGF-I, IGF-II, insulin or IGFBP-1-stimulated migration of the JEG-3

cells were mediated through IGF-IR, IR or a5b1 integrin activation, the

same procedures were performed in JEG-3 cells that were preincubated with

aIR3 (1 mg/ml), HNMPA(AM)3 (50 mM) or MAB1969 (5 mg/ml) for 60 min.

RNA extraction, cDNA synthesis and RT–PCR

RNA was extracted using TRIzol reagent (Invitrogen). Total RNA (5 mg) from

different samples were first treated with DNase I (Promega) and then reverse

transcribed using Superscript II (Invitrogen) in a reaction volume of 20 ml.

The synthesized cDNA was used to measure the InsR expression by

PCR. The reactions were performed in 20 ml with 1 ml of the respective

cDNA sample and 0.4 mM primers, Taq DNA polymerase, buffer and

nucleotides (Invitrogen). The primers used were: for GAPDH,

50GTGAAGGTCGGAGTCAACG30 and 50GGTGGAGACGCCAGTGGAC30

and for the InsR, 50AACCAGAGTGAGTATGAGGAT30 and

50CCGTTCCAGAGCGAAGTGCTT30.

Diaz et al.

568

by guest on February 3, 2014http://m

olehr.oxfordjournals.org/D

ownloaded from

Results

IGF-II is highly expressed in gestational trophoblasticdisease tissue

Trophoblastic cells are known to express IGF-II. Although the steady

state mRNA levels of IGF-II have been measured in different

trophoblastic diseases (Kim et al., 2003), no information is yet

available regarding the protein levels of this growth factor in these

tissues. Because we intend to study IGF-II effects on trophoblastic

cells, it was important to estimate IGF-II protein concentration

within the tissue. We used IRMA to analyse IGF-II and IGF-I

levels in choriocarcinoma, complete hydatidiform mole as well as

first trimester placenta tissue samples. As shown in Fig. 1, placenta

contains an average concentration of 1.38+0.21 nM (0.84+0.23 ng/mg protein) of IGF-I and 2.66+0.63 nM (0.43+0.18 ng/mg protein) of IGF-II. Hydatidiform moles, express signifi-

cantly (P , 0.05) higher levels of IGF-II in comparison with placenta

with levels of 7.33+0.58 nM (1.12+0.12 ng/mg protein). In

contrast, a lower amount of IGF-I (P , 0.05) was observed in moles

as compared to placenta with measured levels of 0.92+0.16 nM

(0.29+0.11 ng/mg protein). During this investigation, two choriocar-

cinoma samples were available for study. The individual levels of

IGF-II were 7.95 nM (1.23 ng/mg protein) and 9.37 nM (1.28 ng/mg

protein), which place them in the upper range of concentration

values measured in the samples. In addition, we observed high

IGF-II circulating levels in sera of patients with hydatidiform mole

compared to those with spontaneous abortions at the same gestational

age (Fig. 1B). In summary, these results indicate that mole and

choriocarcinoma tissue contain high levels of IGF-II, sufficient to

saturate all known high affinity receptors (Denley et al., 2005).

IGF-II regulates choriocarcinoma cell invasion

We next studied the role of IGF-II in the regulation of choriocarci-

noma cell invasion. JEG-3 choriocarcinoma cells were used since

they are known to express all known receptors for IGF-I and

IGF-II (Ritvos et al., 1988; O’Gorman et al., 1999). In preliminary

experiments measuring ERK1/2 phosphorylation, we determined

that maximal receptor activation is achieved with 5 nM of both

IGF-I and IGF-II, which is in line with the known Kd for these

receptors (Denley et al., 2005). The invasive properties of JEG-3

choriocarcinoma cells were measured using a matrigel cell invasion

assay. As shown in Fig. 2A, both IGF-I and IGF-II treatment induced

JEG-3 matrigel invasion in a dose dependent manner. Maximal

effects were obtained with concentrations of 10 nM for both

ligands. Similar treatment has little effect on cell proliferation

within the time frame analysed (Fig. 2B). Interestingly, despite the

fact that receptor saturating concentrations of IGF-I (Kd ¼ 0.2 nM)

are used (Denley et al., 2005), concomitant addition of IGF-II

further increases cell invasion suggesting that IGF-II uses receptors

other than IGF-IR to promote cell invasion (Fig. 2C). We also tested

the effects of IGFBP-1 alone or in combination with the type I and

type II IGFs. IGFBP-1 is known to be expressed in placenta and is

thought to negatively regulate IGF-I activity (Crossey et al., 2002)

but also to have IGF-I independent actions through direct

interactions with integrin a5b1 receptor (Jones et al., 1993).

Treatment with IGFBP-1 alone induced JEG-3 invasion and has an

additive effect to the actions of IGF-II. In contrast, no additive

effect was observed on the actions of IGF-I which is indicative of

a negative interaction between the two growth factors (Crossey

et al., 2002).

In order to better understand the nature of IGF-I, IGF-II and

IGFBP-1 actions on choriocarcinoma cells, the effects on cell

invasion of IGF-IR blocking antibody aIR3 and anti-a5b1 integrin

receptor antibody were measured. As shown in Fig. 3A, treatment

with aIR3 Ab completely inhibits the effects of IGF-I on cell

invasion and has a small although significant effect in the actions

of IGF-II. The actions of IGFBP-1 on cell invasion were not affected

by treatment with aIR3 Ab. On the other hand, treatment with

anti-a5b1 integrin reduces the actions of IGFBP-1 but was inactive

towards IGF-I or IGF-II (Fig. 3B). An isotype Ab (IgG) did not

influence basal or IGF-I/ IGF-II induced cell invasion (Fig. 3C).

These results strongly suggest that receptors other than IGF-I

receptor or integrin a5b1 are used by IGF-II to promote choriocarci-

noma cell invasion.

Figure 1: IGF-I and IGF-II protein expression in hydatidiform mole (A) Boxplots showing the absolute levels of IGF-I and IGF-II in hydatidiform mole incomparison to placenta from spontaneous abortions. (B) Box plots showing thelevels of IGF-II in sera from patients with complete hydatidiform mole in com-parison to those of similar age with spontaneous non-molar abortions. Statisti-cal analysis was performed using analysis of variance (ANOVA). Differentletters indicate statistically significant differences (P , 0.05)

IGF-II regulation of trophoblast invasion

569

by guest on February 3, 2014http://m

olehr.oxfordjournals.org/D

ownloaded from

IGF-II induces choriocarcinoma cell invasion through theactivation of the insulin receptor

The InsR, specifically the isoform A, can bind IGF-II with high

affinity (Kd ¼ 0.9 nM), Therefore, we wanted to test the effect of

IGF-I and IGF-II on IGF-IR and InsR activation. We first analysed

the expression levels of InsR isoforms in placenta tissue. Fig. 4A

shows that both InsR-A and InsR-B are expressed in placenta

(Benecke et al., 1992), hydatidiform mole and in JEG-3 trophoblastic

cell lines. We next used JEG-3 cells to study InsR and IGF-IR

activation. As shown in Fig. 4B, IGF-I treatment induced the tyrosine

phosphorylation of the IGF-I receptor, an effect that was inhibited

by blocking antibody aIR3. No induction of InsR tyrosine

phosphorylation by IGF-I was observed. On the other hand, IGF-II

treatment induced the InsR tyrosine phosphorylation with no

evidence of IGF-I receptor activation. The IGF-IR blocking antibody

had no discernible effect on IGF-II induced InsR activation. On the

other hand, treatment with InsR inhibitor HNMPA(AM)3 caused a

dose dependent inhibition of IGF-II-induced InsR tyrosine

phosphorylation. This inhibition was observed from 10 to 100 nM.

Similar concentrations also inhibited insulin-mediated activation of

the InsR. In contrast, HNMPA(AM)3 was able to inhibit IGF-I

activation of the IGF-I receptor at only the higher concentration. In

summary, both the direct measurement of IGF-II effects on receptor

activation as well as the actions of HNMPA(AM)3 demonstrate that

Figure 2: Regulation of JEG-3 trophoblastic cell invasion by IGF-I, IGF-IIand IGFBP-1. (A) Matrigel invasion assays to measure the treatment effectof increasing concentrations of IGF-I and IGF-II on JEG-3 cell invasion.(B) Effects of IGF-I (10 nM), IGF-II (10 nM), foetal calf serum-FCS (10%)and BSA (2%) on cell proliferation measured by MTT assay after 24 h treat-ment. (C) Matrigel invasion assays to measure the effects of JEG-3 cells treat-ments with different combinations of IGF-I (10 nM or 20 nM as indicated),IGF-II (10 nM or 20 nM as indicated) or IGFBP1 (10 ng/ml). In each casethe results of at least three independent replicates are shown. Statistical analysiswas performed using ANOVA. Different letters indicate statistically significantdifferences (P , 0.05) between cells treated with a single ligand and thosetreated with a combination of two growth factors

Figure 3: Effects of IGF-I receptor blocking Ab, aIR-3 (1 mg/ml) (A) andintegrin a5b1 blocking Ab MAB1969 (25 mg/ml) (B) on IGF-I, IGF-II andIGFBP1 induced cell invasion; the effects of non-immune goat IgG(1 mg/ml) on IGF-I and IGF-II induced cell invasion is shown for comparison(C) In each case the results of at least three independent replicates are shown.Statistical analysis was performed using ANOVA. Different letters indicatestatistically significant differences (P , 0.05)

Diaz et al.

570

by guest on February 3, 2014http://m

olehr.oxfordjournals.org/D

ownloaded from

Figure 4: Activation of IGF-IR and InsR by IGF-I, IGF-II and insulin in JEG-3 cells (A) RT–PCR based detection of insulin receptor isoforms A and B expressionin placenta from non-molar spontaneous abortions (lines 1–4 and 11), in complete hydatidiform mole (lines 5–10 and 12) and in JEG-3 trophoblastic cell line (line13). (B) The upper panel shows the tyrosine phosphorylation of the IGF-IR and InsR upon treatment with IGF-I (10 nM) or IGF-II (10 nM) given alone or incombination with either different doses of IR inhibitor HNMPA(AM)3 or the IGF-IR blocking Ab aIR-3 (1 mg/ml). The lower panel shows the tyrosine phosphoryl-ation of the IGF-I receptor and InsR receptor upon treatment with IGF-I (10 nM) or insulin (10 nM) given alone or in combination with either different doses of IRinhibitor HNMPA(AM)3 or the IGFR blocking Ab aIR-3 (1 mg/ml). (C) Effects of IR inhibitor HNMPA(AM)3 on IGF-I (10 nM), IGF-II (10 nM) and insulin(10 nM) induced matrigel cell invasion. Statistical analysis was performed using ANOVA. Different letters indicate statistically significant differences (P , 0.05)

IGF-II regulation of trophoblast invasion

571

by guest on February 3, 2014http://m

olehr.oxfordjournals.org/D

ownloaded from

the InsR is the key mediator of IGF-II actions in choriocarcinoma

cells. The analysis of cell invasion (Fig. 4C) shows that

HNMPA(AM)3 treatment at concentrations (50 mM) that specifically

inhibit IR activation by IGF-II also blocked ligand induced cell

invasion. In contrast, IGF-I induced cell invasion was not affected

by similar concentrations of HNMPA(AM)3 although inhibitory

effects are evident at 100 mM, parallel to inhibition of IGF-IR

activation. As expected, IGF-I induced cell invasion was inhibited

by the IGF-IR (aIR3) blocking antibody.

IGF-II regulates cell invasion through its activation ofAKT, ERK pathways and the induction of cell adhesionand motility

In order to understand the nature of IGF-II actions, we analysed signal-

ling pathways downstream of the IGFR and IR, which are potentially

important in the regulation of cell invasion. The activity of AKT was

measured by Western Blot using anti-phospho Ser 473; while active

ERK was measured using a specific Ab against Thr202/Tyr204. The

results are depicted in Fig. 5. We found that treatment with IGF-II,

as well as insulin, IGF-I and IGFBP-1 induced the activation of

ERK as well as AKT. Similar to previous findings regarding receptor

activation, IGF-IR blocking antibody, aIR3 inhibited IGF-I activation

of ERK and AKT but failed to inhibit IGF-II action on these kinases

(Fig. 5C). IGF-II stimulatory activity towards ERK and AKT was

instead sensitive to inhibition by low concentrations (10 mM) of the

IR inhibitor HNMPA(AM)3. The relevance of ERK and AKT

activation in cell invasion was next studied using matrigel invasion

assays. Treatment with PI3-kinase inhibitor LY294002 and MEK

inhibitor U0126, which reduce the activation of AKT and ERK

respectively, resulted in significant inhibition of IGFs-induced cell

invasion (Fig. 5D).

For cells to invade through a thick layer of matrigel a complex set of

biological processes have to take place. This includes the attachment of

cells to the matrigel, degradation of extracellular matrix and motility

towards specific matrix protein components. In order to understand

the nature of IGF-II actions on cell invasion, we measured both cell

adherence as well as motility towards three different extracellular

matrix proteins: collagen IV, fibrinogen and laminin (Fig. 6). We

found that IGF-II treatment induces the attachment of JEG-3 cells to

matrigel, an effect that was sensitive to inhibition by HNMPA(AM)3.

Similar effects could be observed when treating with insulin. IGF-I

has analogous effects through the activation of the IGF-I receptor.

The analysis of cell motility (Fig. 6B) showed that IGF-II treatment

resulted in a 3-fold increase in motility towards collagen IV. Similar

effects were observed after IGF-I treatment, but in addition, enhanced

motility towards fibrinogen and laminin was also observed. Similar to

what was observed when analysing cell invasion, IGF-II but not IGF-I

effects were sensitive to pharmacological inhibition with the InsR

kinase inhibitor HNMPA(AM)3 used at 50 mM. The above evidence

demonstrates qualitative differences in the way IGF-I and IGF-II

regulate cell motility.

Discussion

This study showed that the circulating and tissue levels of IGF-II are

elevated in gestational trophoblastic diseases, specifically complete

hydatidiform moles when compared to first trimester placenta from

spontaneous abortions. In addition, we have investigated the

mechanisms whereby IGF-II regulates the invasive properties of

trophoblastic cells. Our results show that IGF-II treatment, like

IGF-I, insulin and IGFBP-1, results in enhanced invasion of JEG-3

trophoblastic cells. IGF-II employs a distinct mechanism to initiate

its actions on trophoblasts, which differs from that used by

IGF-I. IGF-II effects are mainly initiated through the activation of

the InsR while IGF-I mostly uses the IGF-IR to initiate signalling.

On the other hand, both IGFs share common intracellular mechanisms

including the activation of ERK and AKT kinases, which are import-

ant for the regulation of cell invasion. The positive effects of IGF-II on

invasion are the result of enhanced cell adhesion and chemotaxis

(specifically towards collagen IV). In comparison to IGF-I, IGF-II

actions were more sensitive to inhibition by the InsR inhibitor

HNMPA(AM)3.

The role of IGF-II in normal placental development is well

established (Baker et al., 1993; Louvi et al., 1997). IGF-II is

abundantly expressed in human placenta, specifically in extravillous

cytotrophblasts invading into the endometrium (Han and Carter,

2000). Consequently, a role for IGF-II in the regulation of invasive

trophoblasts in normal placenta has been proposed. This study

showed that the protein expression levels of IGF-II are highly elevated

in gestational trophoblastic diseases compared to first trimester

placenta from spontaneous abortions. In the two cancer tissues we

obtained, high levels of IGF-II were also detected. Although, due to

the scarcity of tumour cases, no statistical analysis was performed,

the concentrations in choriocarcinoma are in the range observed for

complete hydatidiform mole. Increased IGF-II protein levels in

gestational trophoblastic disease are likely to be explained by

enhanced expression of this gene. Previous results have shown that

loss of imprinting of IGF-II gene locus occur in 43% of complete

hydatidiform moles concomitant to increased mRNA levels of the

gene (Kim et al., 2003). Therefore, enhanced local production of

IGF-II can potentially contribute to progression of gestational

trophoblastic diseases and choriocarcinoma pathology. In addition,

increased circulating levels of IGF-II in sera may contribute to the

invasive properties of choriocarcinoma cells in distant metastatic sites.

How IGF-II exerts its actions on trophoblastic cells is a matter of

great interest. At least two different receptors have been previously

proposed to initiate the actions of IGF-II on trophoblastic cells; the

IGF-II/Mannose-6 phosphate receptor (Mckinnon et al., 2001) and

the IGF-I receptor. Initial findings pointed to the IGF-II/Mannose-6

phosphate receptor based on the use of antibodies against the receptor

(Mckinnon et al., 2001), but numerous studies have demonstrated that

this receptor has mainly inhibitory actions on IGF-II (Ludwig et al.,

1996; O’Gorman et al., 1999, 2002). Other studies have postulated

that IGF-II actions are mediated by the IGF-I receptor (Pandini

et al., 2002). Here we demonstrate that IGF-II signalling in JEG-3

cells is initiated mainly through the activation of the InsR. Two

different lines of evidence support our findings. Firstly, IGF-II treat-

ment of JEG-3 cells mainly induced the tyrosine phosphorylation of

the InsR but little effect is observed in IGF-I receptor. In contrast,

IGF-I treatment induced the tyrosine phosphorylation of the IGF-I

receptor while no effect on InsR activation was observed. Secondly,

although both IGFs were able to activate ERK and AKT, IGF-II but

not IGF-I effects were inhibited by concentrations of

HNMPA(AM)3 that specifically inhibit the InsR (Fig. 5A). Pharmaco-

logical inhibition of the InsR with 50 mM HNMPA(AM)3 completely

blocked IGF-II induced cell invasion and motility towards collagen IV

but were ineffective against IGF-I. On the other hand a blocking

antibody against the IGF-I receptor has no effect on IGF-II actions

in the above described parameters but effectively inhibited all IGF-I

actions.

Mouse gene knockout studies first identified the important role of

the InsR in mediating the growth promoting effects of IGF-II in

early development (Louvi et al., 1997). The InsR exists in two

isoforms that originate by alternative splicing of exon 11 in the InsR

gene. The two isoforms are identical except 12 amino acids inserted

Diaz et al.

572

by guest on February 3, 2014http://m

olehr.oxfordjournals.org/D

ownloaded from

upstream of the third last residue of the extracellular alpha subunits of

the InsR-B isoform (Moller et al., 1989). The InsR-A isoform can bind

insulin and IGF-II with high affinity but binds IGF-I poorly (Frasca

et al., 1999; Denley et al., 2004). Therefore, it is likely that the

InsR-A is the receptor mediating IGF-II actions on trophoblast cells.

Indeed, previous reports show that InsR-A is highly expressed in

placenta as compared with normal adult tissues such as liver and

muscle, which mainly express InsR-B (Benecke et al., 1992).

Figure 5: Effect of IGF-I, IGF-II or insulin treatment on ERK and AKT activation in JEG-3 cells Phosphorylation of ERK and AKT kinases upon treatment withIGF-I (10 nM) (A and B), IGF-II (10 nM) (A and C), insulin (10 nM) (B) or IGFBP-1 (C) given alone or in combination with either different doses of (A) the InsRinhibitor, HNMPA(AM)3, (A and C) the IGF-IR blocking Ab, aIR-3 (1 mg/ml), (C) the MEK inhibitor UO126, (C) the PI3 kinase inhibitor LY94002 and (C) anIGF-II neutralizing antibody S1F2 (10 mg/ml). (D) Effects of MEK inhibitor UO126 and PI3 kinase inhibitor LY294002 on IGF-I, IGF-II induced cell invasion.Statistical analysis was performed using ANOVA. Different letters indicate statistically significant differences (P , 0.05)

IGF-II regulation of trophoblast invasion

573

by guest on February 3, 2014http://m

olehr.oxfordjournals.org/D

ownloaded from

An intriguing observation was that despite using similar

concentrations of IGF-II and insulin and the latter having slightly

higher affinity for InsR-A, IGF-II is significantly more potent than

insulin in the promotion of cell invasion and motility. Interestingly,

similar findings have been described in SKUT-1 cells in relation to

cell invasion (Sciacca et al., 2002). This suggests that different

signalling pathways are triggered by IGF-II and insulin despite

sharing a common receptor. It is important to notice that moderate

(10 nM) concentrations of IGF-II and insulin were used in the

experiments, which in the case of IGF-II will preferentially activate

Figure 6: IGF-I, IGF-II and IGFBP-1 regulation of JEG-3 trophobalstic cell adhesion and motility. (A) Effect of IGF-I (10 nM), IGF-II (10 nM) IGFBP-1 (10 ng/ml) and insulin (10 nM) treatment on JEG-3 cells adhesion to matrigel. The effects of combined treatment with IGF-I receptor blocking Ab, aIR-3 (1 mg/ml), theintegrin a5b1 blocking Ab, MAB1969 (25 mg/ml) and the InsR kinase inhibitor HNMPA(AM)3. (B) Results from assays measuring the effects of IGF-I (10 nM),IGF-II (10 nM) or insulin (10 nM) and IGFBP-1 (10 ng/ml) on JEG-3 cells motility towards specific extracellular matrix proteins: laminin, collagen IV and fibro-nectin. BSA was used as control. The effects of combined treatment with IGF-I receptor blocking Ab, aIR-3 (1 mg/ml), the integrin a5b1 blocking Ab, MAB1969(25 mg/ml) and the InsR kinase inhibitor, HNMPA(AM)3 are also shown. In each case the results of at least three independent replicates are shown. Statistical analy-sis was performed using ANOVA. Different letters indicate statistical significant differences (P , 0.05)

Diaz et al.

574

by guest on February 3, 2014http://m

olehr.oxfordjournals.org/D

ownloaded from

the InsR-A isoform while insulin will activate both InsR-A and B. It

has been long known that the different isoforms of the InsR have

distinct signalling properties despite having identical intracellular

domains (Sciacca et al., 2003; Pandini et al., 2004). The reasons for

this are not fully understood, although recent findings indicate that

both receptors localized to discrete regions in the cell membrane

and are differentially internalized in response to insulin (Uhles

et al., 2003); thereby suggesting the involvement of additional

membrane proteins in signalling. One possibility to explain the

enhanced invasive activity of IGF-II is through the activation of

hybrid insulin/IGF-I receptors that bind IGF-II but not insulin with

high affinity. We do not favour this hypothesis because: (i) we were

not able to detect IGF-I receptor activation upon IGF-II treatment

and (ii) IGF-II effects on cell invasion were inhibited by

HNMPA(AM)3 used at concentrations that specifically target the

InsR. To note is that, IGF-I can also bind hybrid receptors with high

affinity, yet we cannot detect increased InsR tyrosine phosphorylation

upon IGF-I treatment, again suggesting a minor role for hybrid recep-

tors in signalling. Clearly, the issue of how receptor hybrids contribute

to signalling requires further exploration, specially if we keep in mind

that the relative effect of HNMPA(AM)3 on hybrids is not known.

Another possibility to explain the different potency of IGF-II and

insulin in cell invasion assays is that InsR-A triggers different

signals depending on whether it is activated by IGF-II or insulin.

Indeed, support for this hypothesis comes from the study of transcrip-

tional actions of IGF-II and insulin in fibroblasts that lack IGF-I

receptor and overexpress InsR-A (Pandini et al., 2003). This study

showed that although most transcriptional actions of insulin and

IGF-II are shared, some ligand specific effects could be detected

such as the induction of integrin aVb3 and adhesion molecule

ICAM. Interestingly, IGF-II and IGF-I but not insulin are known to

bind vitronectin, a ligand for integrin aVb3 to promote cell migration

(Kricker et al., 2003). Therefore, integrin aVb3 activation by IGF-II/

vitronectin complex can bear direct relationship to the enhanced

effects of IGF-II on JEG-3 cell migration as compared to insulin.

Clearly, a better understanding of how the InsR isoforms contribute

to IGF-II signalling is required before the different potency of

IGF-II and insulin in cell motility assays can be explained.

Despite the differences in signalling, insulin, IGF-I, IGF-II as well

as IGFBP-1 have the capacity to promote cell invasion using a core set

of signalling intermediaries. This is evident by the fact that PI3 kinase

and MEK inhibitors reduce cell invasion induced by the three ligands.

Consequently, in the context of the tissue, trophoblast invasion may be

primarily determined by factors that influence the availability of IGFs

and its binding proteins. Because we have shown that IGF-II but not

IGF-I is overexpressed in placental tissue from complete hydatidiform

moles and in corresponding sera, IGF-II is likely to be more influential

than IGF-I in the regulation of the invasive properties of trophoblasts

from complete moles. Supporting this hypothesis are previous findings

showing that extravillous cytotrophoblasts, which have a very high

invasive capacity, express the highest amounts of IGF-II in normal

placenta while syncitiotrophoblasts express little IGF-II (Han and

Carter, 2000). In order to promote cell invasion, IGF-II has to regulate

complex cellular mechanisms that control adhesion, migration and

extracellular matrix proteolysis. We have shown that IGF-II controls

invasion by at least regulating cell adhesion and motility.

Interestingly, IGF-II but also IGF-I and insulin promote motility

specifically towards collagen IV but not laminin or fibronectin

suggesting a specific regulation of integrin subunits that act as

collagen receptors (Denley et al., 2006). In contrast, IGFBP-1 the

main secretory product of the decidua also promotes invasion

towards fibronectin, in line with its known function as a ligand of

the main fibronectin receptor: the a5/b1 integrin complex (Jones

et al., 1993; Gleeson et al., 2001). The specific activation of fibronec-

tin receptor by IGFBP-1 but not IGF-II also provides an explanation

for its enhancing effects of IGF-II induced invasion. Future studies

are needed to clarify which specific integrin receptors are influenced

by IGF-II as well as the role of matrix proteolysis in IGF-II-induced

trophoblast invasion.

In conclusion, we have shown that IGF-II induces trophoblast cell

invasion through the activation of the InsR. Clearly, additional

studies are needed to better characterize the role of the different

InsR isoforms in the regulation of IGF-II actions. The development

of a specific inhibitor of the IGF-II/InsR-A receptor interaction may

have important applications. It could potentially be used for the

pharmacological treatment of choriocarcinomas and the several

types of tumours known to overexpress IGF-II, while preserving the

metabolic actions of insulin.

Acknowledgements

This work has been supported by grants to M.S.-G. from COLCIENCIAS andto A.F.M. from the Swedish Research Council and Wallenberg Foundation.Y.-C.C. was supported by Cancerfonden. L.F.-P. was supported by grantsfrom Pfizer Spain CN-78/02-05045 and MEC SAF2003-02117.

References

Altieri A, Franceschi S, Ferlay J, Smith J, La Vecchia C. Epidemiology andaetiology of gestational trophoblastic diseases. Lancet Oncol2003;4:670–678.

Baker J, Liu JP, Robertson EJ, Efstratiadis A. Role of insulin-like growthfactors in embryonic and postnatal growth. Cell 1993;75:73–82.

Benecke H, Flier JS, Moller DE. Alternatively spliced variants of the insulinreceptor protein. Expression in normal and diabetic human tissues. J ClinInvest 1992;89:2066–2070.

Chuan YC, Pang ST, Cedazo-Minguez A, Norstedt G, Pousette A,Flores-Morales A. Androgen induction of prostate cancer cell invasion ismediated by ezrin. J Biol Chem 2006;281:29938–29948.

Conforti G, Zanetti A, Colella S, Abbadini M, Marchisio PC, Pytela R,Giancotti F, Tarone G, Languino LR, Dejana E. Interaction of fibronectinwith cultured human endothelial cells: characterization of the specificreceptor. Blood 1989;73:1576–1585.

Constancia M, Hemberger M, Hughes J, Dean W, Ferguson-Smith A,Fundele R, Stewart F, Kelsey G, Fowden A, Sibley C et al.Placental-specific IGF-II is a major modulator of placental and fetalgrowth. Nature 2002;417:945–948.

Crossey PA, Pillai CC, Miell JP. Altered placental development andintrauterine growth restriction in IGF binding protein-1 transgenic mice.J Clin Invest 2002;110:411–418.

Denley A, Bonython ER, Booker GW, Cosgrove LJ, Forbes BE, Ward CW,Wallace JC. Structural determinants for high-affinity binding ofinsulin-like growth factor II to insulin receptor (IR)-A, the exon 11 minusisoform of the IR. Mol Endocrinol 2004;18:2502–2512.

Denley A, Cosgrove LJ, Booker GW, Wallace JC, Forbes BE. Molecularinteractions of the IGF system. Cytokine Growth Factor Rev2005;16:421–439.

Denley A, Brierley GV, Carroll JM, Lindenberg A, Booker GW, Cosgrove LJ,Wallace JC, Forbes BE, Roberts CT, Jr. Differential activation of insulinreceptor isoforms by insulin-like growth factors is determined by the Cdomain. Endocrinology 2006;147:1029–1036.

Frasca F, Pandini G, Scalia P, Sciacca L, Mineo R, Costantino A, Goldfine ID,Belfiore A, Vigneri R. Insulin receptor isoform A, a newly recognized,high-affinity insulin-like growth factor II receptor in fetal and cancer cells.Mol Cell Biol 1999;19:3278–3288.

Gleeson LM, Chakraborty C, McKinnon T, Lala PK. Insulin-like growthfactor-binding protein 1 stimulates human trophoblast migration bysignaling through alpha 5 beta 1 integrin via mitogen-activated proteinKinase pathway. J Clin Endocrinol Metab 2001;86:2484–2493.

Han VK, Carter AM. Spatial and temporal patterns of expression of messengerRNA for insulin-like growth factors and their binding proteins in the placentaof man and laboratory animals. Placenta 2000;21:289–305.

IGF-II regulation of trophoblast invasion

575

by guest on February 3, 2014http://m

olehr.oxfordjournals.org/D

ownloaded from

He L, Cui H, Walsh C, Mattsson R, Lin W, Anneren G, Pfeifer-Ohlsson S,Ohlsson R. Hypervariable allelic expression patterns of the imprintedIGF2 gene in tumor cells. Oncogene 1998;16:113–119.

Jones JI, Gockerman A, Busby WH, Jr, Wright G, Clemmons DR. Insulin-likegrowth factor binding protein 1 stimulates cell migration and binds to thealpha 5 beta 1 integrin by means of its Arg-Gly-Asp sequence. Proc NatlAcad Sci U S A 1993;90:10553–10557.

Kim SJ, Park SE, Lee C, Lee SY, Kim IH, An HJ, Oh YK. Altered imprinting,promoter usage, and expression of insulin-like growth factor-II gene ingestational trophoblastic diseases. Gynecol Oncol 2003;88:411–418.

Korner C, Nurnberg B, Uhde M, Braulke T. Mannose 6-phosphate/insulin-likegrowth factor II receptor fails to interact with G-proteins analysis of mutantcytoplasmic receptor domains. J Biol Chem 1995;270:287–295.

Kricker JA, Towne CL, Firth SM, Herington AC, Upton Z. Structural andfunctional evidence for the interaction of insulin-like growth factors(IGFs) and IGF binding proteins with vitronectin. Endocrinology2003;144:2807–2815.

Li J, Sahagian GG. Demonstration of tumor suppression by mannose6-phosphate/insulin-like growth factor 2 receptor. Oncogene 2004;23:9359–9368

Louvi A, Accili D, Efstratiadis A. Growth-promoting interaction of IGF-II withthe insulin receptor during mouse embryonic development. Dev Biol1997;189:33–48.

Ludwig T, Eggenschwiler J, Fisher P, D’Ercole AJ, Davenport ML, EfstratiadisA. Mouse mutants lacking the type 2 IGF receptor (IGF2R) are rescued fromperinatal lethality in Igf2 and Igf1r null backgrounds. Dev Biol1996;177:517–535.

Lustig-Yariv O, Schulze E, Komitowski D, Erdmann V, Schneider T, de GrootN, Hochberg A. The expression of the imprinted genes H19 and IGF-2 inchoriocarcinoma cell lines. Is H19 a tumor suppressor gene? Oncogene1997;15:169–177.

McKinnon T, Chakraborty C, Gleeson LM, Chidiac P, Lala PK. Stimulation ofhuman extravillous trophoblast migration by IGF-II is mediated by IGF type2 receptor involving inhibitory G protein(s) and phosphorylation of MAPK.J Clin Endocrinol Metab 2001;86:3665–3674.

Moller DE, Yokota A, Caro JF, Flier JS. Tissue-specific expression of twoalternatively spliced insulin receptor mRNAs in man. Mol Endocrinol1989;3:1263–1269.

Mosmann T. Rapid colorimetric assay for cellular growth and survival:application to proliferation and cytotoxicity assays. J Immunol Methods1983;65:55–63.

O’Gorman DB, Costello M, Weiss J, Firth SM, Scott CD. Decreasedinsulin-like growth factor-II/mannose 6-phosphate receptor expressionenhances tumorigenicity in JEG-3 cells. Cancer Res 1999;59:5692–5694.

O’Gorman DB, Weiss J, Hettiaratchi A, Firth SM, Scott CD. Insulin-likegrowth factor-II/mannose 6-phosphate receptor overexpression reduces

growth of choriocarcinoma cells in vitro and in vivo. Endocrinology2002;143:4287–4294.

Pandini G, Frasca F, Mineo R, Sciacca L, Vigneri R, Belfiore A. Insulin/insulin-like growth factor I hybrid receptors have different biologicalcharacteristics depending on the insulin receptor isoform involved. J BiolChem 2002;277:39684–39695.

Pandini G, Medico E, Conte E, Sciacca L, Vigneri R, Belfiore A. Differentialgene expression induced by insulin and insulin-like growth factor-IIthrough the insulin receptor isoform A. J Biol Chem 2003;278:42178–42189.

Pandini G, Conte E, Medico E, Sciacca L, Vigneri R, Belfiore A. IGF-IIbinding to insulin receptor isoform A induces a partially different geneexpression profile from insulin binding. Ann N Y Acad Sci 2004;1028:450–456.

Ritvos O, Rutanen EM, Pekonen F, Jalkanen J, Suikkari AM, Ranta T.Characterization of functional type I insulin-like growth factor receptorsfrom human choriocarcinoma cells. Endocrinology 1988;122:395–401.

Saperstein R, Vicario PP, Strout HV, Brady E, Slater EE, Greenlee WJ,Ondeyka DL, Patchett AA, Hangauer DG. Design of a selective insulinreceptor tyrosine kinase inhibitor and its effect on glucose uptake andmetabolism in intact cells. Biochemistry 1989;28:5694–5701.

Sciacca L, Mineo R, Pandini G, Murabito A, Vigneri R, Belfiore A. In IGF-Ireceptor-deficient leiomyosarcoma cells autocrine IGF-II induces cellinvasion and protection from apoptosis via the insulin receptor isoform A.Oncogene 2002;21:8240–8250.

Sciacca L, Prisco M, Wu A, Belfiore A, Vigneri R, Baserga R. Signalingdifferences from the A and B isoforms of the insulin receptor (IR) in 32Dcells in the presence or absence of IR substrate-1. Endocrinology2003;144:2650–2658.

Seckl MJ, Fisher RA, Salerno G, Rees H, Paradinas FJ, Foskett M, NewlandsES. Choriocarcinoma and partial hydatidiform moles. Lancet2000;356:36–39.

Soper JT. Gestational trophoblastic disease. Obstet Gynecol 2006;108:176–187.

Uhles S, Moede T, Leibiger B, Berggren PO, Leibiger IB. Isoform-specificinsulin receptor signaling involves different plasma membrane domains.J Cell Biol 2003;163:1327–1337.

Zhang L, Shiverick KT. Benzo(a)pyrene, but not 2,3,7,8-tetrachlorodibenzo-p-dioxin, alters cell proliferation and c-myc and growth factor expressionin human placental choriocarcinoma JEG-3 cells. Biochem Biophys ResCommun 1997;231:117–120.

Zhang M AS, Yeh S. RRR-alpha-tocopheryl succinate inhibits human prostatecancer cell invasiveness. Oncogene 2004;23:3080–3088.

Submitted on March 12, 2007; resubmitted on April 19, 2007; accepted onApril 23, 2007

Diaz et al.

576

by guest on February 3, 2014http://m

olehr.oxfordjournals.org/D

ownloaded from