potential autocrine role of insulin-like growth factor ii...

[CANCER RESEARCH 52. 3182-3188, June 1, 1992]

Potential Autocrine Role of Insulin-like Growth Factor II during Suramin-inducedDifferentiation of HT29-D4 Human Colonie Adenocarcinoma Cell Line1

Gilbert J. Pommier,2 FranÃ§oiseL. Garrouste, l'ut ima El Atiq, Monique Roccabianca, Jacques L. Marvaldi,

and Maryse M. Remacle-BonnetLaboratoire d'Immunopathologie, FacultÃ©de MÃ©decine[G. J. P., F. L. G., F. E. A.], and Institut de Chimie Biologique, CNRS-URA 202, FacultÃ©des Sciences [M. R.,

J. L. M., M. M. R. B.J, Marseille, France

ABSTRACT

Siiramin, a drug that binds to several types of growth factors, has beenpreviously shown to induce the enterocyte-like differentiation of III29-

IM human colonie adenocarcinoma cells, suggesting that growth factorsare involved in such a process. Undifferentiated III29-1)4 cells releaseinsulin-like growth factor II (IGF-II) into the culture medium that is

totally complexed to heterogeneous IGF binding proteins (IGFBP) expressing high affinities for this growth factor ( A,'du= 0.02 IIMand A',,/,=

1.4 DM). These complexes do not allow IGF-II to bind to II I 29-1)4 cellsurface type I IGF receptors, as evidenced by using 125I-IGF-II-IGFBP

complexes. However, the addition of 40-100 Â¿tg/mlsuramin, i.e., concentrations identical to the ones that are able to induce III 29-1)4 celldifferentiation, induces the release of IGF-II from IGF-II-IGFBP complexes, thereby allowing IGF-II to bind to the cell surface receptors. Atsuch concentrations, suramin is indeed unable to alter IGF-II binding to111 29-1)4 cells, a capacity that is observed only for concentrations higher

than 200 Â«ig/ml.Thus, suramin might have the unusual capacity to allowthe establishment of an IGF-II autocrine loop involved in HT29-D4 celldifferentiation. Consistent with this hypothesis is the fact that exoge-nously applied IGF-I (2.5 Â«tg/ml)or agonist monoclonal antibody aIR-3

(2.5 Mg/ml). which can bypass IGFBP present in the culture medium,induces part of HT29-D4 cell differentiation that is characterized by an

important carcinoembryonic antigen release and the induction of numerous intercellular cysts with microvilli.

INTRODUCTION

Previous reports have shown that the human colonie adenocarcinoma cell line HT29 releases into the culture medium anumber of growth factors (1,2). We have recently reported therelease of IGF-II3 entirely complexed to three forms of IGFBP

(Mr 27,000, 28,000, and 31,000 molecular forms) (3), whichhave been recently identified as isoforms of the IGFBP-4 class(4). In addition, high-affinity type I IGF receptors, but not typeII, have been identified at the surface of both HT29 and HT29-D4 cells (5), a clone derived from the parental HT29 cell line(6). Elevated levels of mRNA for IGF-II and/or secretion ofelevated amounts of this peptide as compared to control cellsor tissues have been reported in a wide variety of tumors (for areview, see Refs. 7 and 8) including human colorectal tumors(9, 10). Moreover, in several cell systems, IGFs, IGFBP, andtype I IGF receptors have been shown to be involved duringcell differentiation (11-16), although the mechanisms responsible are not well understood.

HT29-D4 cells constitute an interesting model for studying

Received 11/5/91; accepted 3/20/92.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1Supported by grants from the Association pour la Recherche sur le Cancerand Ligue Nationale FranÃ§aisecontre le Cancer.3To whom requests for reprints should be addressed, at Laboratoire d'Immu

nopathologie, FacultÃ©de MÃ©decine,27, Boulevard Jean Moulin, 13385 MarseilleCedex 5, France.

3The abbreviations used are: IGF, insulin-like growth factor; IGFBP, IGFbinding protein; CEA, carcinoembryonic antigen; DMEM, Dulbecco's modifiedEagle's medium; PCS, fetal calf serum; BSA, bovine serum albumin; PBS,phosphate-buffered saline; aIR-3, monoclonal antibody to type I IGF receptor.

the relationships that link the endogenous secretion of growthfactors and the differentiation state of colonie cancer cells.HT29-D4 cells can indeed be induced to differentiate by twoways: (a) the removal of glucose and its replacement by galactose in the culture medium (6); (b) the addition of suramin inthe glucose-containing medium in the presence or absence ofPCS (17, 18). This latter observation strongly suggests thatgrowth factors secreted by HT29-D4 cells themselves may beinvolved, at least in part, in the mechanisms that control theprocess of induction of HT29-D4 cell differentiation. Suramin,a polysulfonated naphthylurea, is indeed known to interferewith growth factor-receptor interaction, primarily by bindingto the growth factor itself rather than to its receptor (19-22).In addition, suramin, previously used to treat trypanosomiasisand onchocerciasis (23), has been recently found to have anantineoplastic activity in humans (24, 25).

In order to clarify how suramin could induce HT29-D4 celldifferentiation by interacting with cell-released IGF-II complexed to IGFBP, we first characterized the binding parametersof these secreted IGFBPs. Furthermore, we determined howsuramin could interfere with IGF-II interaction with eitherIGFBP in the culture medium or type I IGF receptors at thecell surface. Moreover, we explored the potential autocrine roleof this growth factor in the induction of HT29-D4 celldifferentiation.

MATERIALS AND METHODS

Materials. Recombinant human IGF-I and IGF-II were purchasedfrom Bachem. These peptides were radioiodinated using the chloramin-T method to a specific activity of 150-200 jiCi/Mg. aIR-3 was purchasedfrom Oncogene Science. Suramin was a gift from Specia France.Sephadex G-75 and G-200 were purchased from Pharmacia. Bio-GelP-100 was from Bio-Rad. RPMI containing 20 mM Ar-2-hydroxyethyl-piperazine-A/'-2-ethanesulfonic acid, DMEM, Hanks' balanced salt so

lution, FCS, and other cell culture reagents were from Flow Laboratories. Tissue culture flasks and multiwell plates were purchased fromFalcon. Multitray units were from Nunc. All other reagents were ofanalytical grade.

Cell Lines and Culture Conditions. The HT29-D4 human colonadenocarcinoma cell line (6) was routinely grown in DMEM containing25 mM glucose and 10% FCS. Large-scale cultures of HT29-D4 cellswere maintained in a 6000-cm2 multitray unit as described elsewhere

(1).Collection of HT29-D4 Cell Conditioned Medium. Serum-free HT29-

D4 conditioned medium was collected as previously reported (3).Briefly, subconfluent HT29-D4 cell monolayers were washed threetimes with Hanks' balanced salt solution and then incubated with

serum-free DMEM supplemented with 5 mM glutamine as a maintenance medium. The first collection of conditioned medium, made after24 h of maintenance culture, was discarded, since I CS derived growthfactors were still expected to be present. Further collections were made,and cultures were refed with fresh DMEM at 48 h intervals for 14 days.Cell debris was removed by centrifugation at 2500 rpm for 20 min;HT29-D4 conditioned medium containing 5 ^M phenylmethylsulfonyl-fluoride was stored at -20Â°C.

3182

Research. on September 9, 2019. © 1992 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

AUTOCRINE ROLE OF IGF-II IN SURAMIN-INDUCED DIFFERENTIATION

Preparation of IGFBP from HT29-D4 Conditioned Medium. Theconditioned medium (500 ml) was thawed, concentrated 50-fold usingan Amicon B15 system (M, cutoff, 15,000), then passed through a 2.5x 90 cm Sephadex G-200 column, equilibrated with 0.05 M NH4HCO3(pH 8.0) at 4Â°C,and calibrated with appropriate standards (26). Fractions (2.7 ml) were assayed for binding of either '"I-IGF-I or '"I-IGF-

II as described below. Active fractions corresponding to an apparentmolecular mass range of 30-50 kilodaltons (IGFBP) were pooled,lyophilized, and reconstitued in 7 ml l M acetic acid. After clarificationby ultracentrifugation at 100,000 x g for 120 min, the supernatant wasapplied to a 1.6 x 60 cm Bio-Gel P-100 column and eluted with l Macetic acid in order to dissociate bound endogenous IGF-II fromIGFBP. Fractions from 30 kilodaltons to the F0 were pooled, lyophilized, and then frozen and stored at -50Â°Cfor further analysis.

Preparation of '"I-IGF-IGFBP Complexes. Aliquots of IGFBP(about 200 ng protein) were incubated overnight with either 125I-IGF-Ior I25I-IGF-II (10 ng, about 4-5 x 10" cpm) in PBS containing 0.1%BSA and passed over a Sephadex G-75 column (1.6 x 50 cm) elutedwith the same buffer. The radioactivity of each one fraction was measured in a gamma counter. Fractions corresponding to the first elutedradioactive peak were pooled and used in the further experiments asradioactive IGF-IGFBP complexes. To assess the absence of free 125I-

IGF in these preparations they were routinely checked, by the bindingassay described in the next paragraph, to see that they were unable toinduce specific '"I-IGF binding on HT29-D4 cells.

IGF Binding Assay to HT29-D4 Cell Surface Receptors. This wasdone as described elsewhere (3). Briefly, HT29-D4 cells were culturedin 16-mm tissue culture wells as described above. At confluent density,cell culture trays were placed on ice 30 min before and throughout theassay. Cell monolayers were washed three times in serum-free DMEMat 4Â°Cand then incubated in a total volume of 0.2 ml of bindingmedium (RPMI containing 20 mM /V-2-hydroxyethylpiperazine-/V'-2-

ethanesulfonic acid, 0.1% BSA, and 0.1 ^M potassium iodide, andadjusted to pH 7.0) for 2 h at 4Â°C(equilibrium binding conditions)with '25I-IGF-I, '25I-IGF-II, or '"I-IGF-IGFBP complex (6-10 x 10"

cpm) with or without various concentrations of suramin. At the end ofthe incubation, cells were washed three times with cold PBS containing0.1% BSA and then lysed in the same buffer containing 1% Triton X-100 and 10% glycerol, and cell-associated radioactivity was counted ina gamma counter. Nonspecific binding, determined as the radioactivitybound in the presence of 2 Mg/ml of unlabeled IGF-I or IGF-II, wassubtracted from total binding to obtain specific binding. Experimentpoints were estimated in triplicate. SD was always less than 5%.

Free available IGF-I present in the culture medium of HT29-D4 cellstreated by IGF-I in excess for 24 h was assayed after molecular sievingof the conditioned medium on Sephadex G-75 as previously described(3).

IGF Binding Assay to IGFBP. These assays were performed essentially as previously described (26) using an albumin (2%)-treated charcoal (5%) suspension in PBS for separating bound from free tracer. Inbinding studies, IGFBP (about 10 Â¿igprotein for IGF-I and l Â¿igproteinfor IGF-II tracer binding) was incubated in triplicate with variousamounts of either '"I-IGF-I or '"I-IGF-II in a total volume of 0.4 mlof PBS, pH 7.4, containing 0.2% BSA for 16h at 4Â°C.Nonspecific

binding (2 Mg/ml of either cold IGF-I or IGF-II) was subtracted fromall data. The SD was never greater than 7%. Binding data were analyzedby using the EBDA/LIGAND computer program (27, 28) distributedby Biosoft. The model which best described the results was determinedby F test comparison of the sum of squares of the various modelsexamined and was used to derive the values for binding capacities anddissociation constants (KÂ¿)of IGFBP. In suramin-induced dissociationexperiments, 12SI-IGF-IGFBP complexes (about 6.0 x 10" cpm) were

incubated in triplicate with or without various amounts of suramin ina total volume of 0.4 ml of PBS, pH 7.4, containing 0.2% BSA for 2 hat 4Â°C.Results were expressed as the percentage of 125I-IGFremaining

bound to IGFBP, i.e., the radioactivity measured in the supernatant,with respect to the bound radioactivity measured in the absence ofsuramin.

CEA Measurement. HT29-D4 cells were plated (1.5-2.0 x 10" cells)in 25-cm2 flasks and cultured as described above. Two days later, IGF-

I (2.5 Mg/ml) or aIR-3 (2.5 Mg/ml) was added in the culture medium.

This supplemented medium was renewed every day. The daily collectedcell culture medium was centrifuged for 10 min at 2500 rpm and storedat -20Â°C.The amount of CEA was determined with the commercially

available Abbott CEA-EIA Monoclonal One-Step kit.Electron Microscopy. HT29-D4 cells, treated or not by IGF-I as

described above, were fixed in situ with 2.5% glutaraldhehyde in 0.2 Msodium cacodylate buffer, pH 7.4, for 2 h, washed overnight using thesame buffer containing 7.5% saccharose, postfixed in 1% osmiumtetroxide, dehydrated in ethanol, and embedded in Epon and observedwith a Jeol 100C electron microscope.

RESULTS

Binding Parameters of IGFBP. The binding characteristics ofIGFBP released into the culture medium by the undifferentiatedHT29-D4 cells were examined in saturation binding experiments with either '"I-IGF-I or '"I-IGF-II as a ligand. TheScatchard plot derived from '"I-IGF-I binding data analyzed

by the LIGAND computer program is shown in Fig. 1. Thebest fit of the data was consistent with a single class of bindingsite (Kd = 3.6 nivi;Amas= 6 fmol bound/Mg IGFBP). When I25I-IGF-II was used as a ligand, the Scatchard plot was curvilinear(Fig. 2). Computer analysis of the binding data by LIGANDpredicted that they could be resolved into a specific high-affinity, low-capacity IGF-II binding site (Kda = 0.02 nM; Bmn= 1.2 fmol/Mg IGFBP) and a specific low-affinity, high-capacityIGF-II binding component (Ka/, = 1.4 HM; Ã„max= 80 fmol/^gIGFBP). Although these values probably represent the resultantof composite binding parameters, it follows that IGFBP isolatedfrom HT29-D4 conditioned medium had an affinity for IGF-IImanyfold higher than for IGF-I.

Effect of Suramin on IGF Binding. To determine whethersuramin could induce HT29-D4 cell differentiation by interfering with the binding of endogenous released IGF-II to cell-associated type I IGF receptors, we first analyzed the drugeffect on IGFs binding to HT29-D4 cells. Fig. 3 shows thatsuramin inhibited the specific binding of both human recombinant '"I-IGF-I and I2SI-IGF-II with a 50% inhibitory concentration of approximately 0.15 mM, i.e., ~200 ng/m\ . In addition, Fig. 4 shows that suramin was also able to efficientlydissociate either I25I-IGF-I or 125I-IGF-II from preformed I25I-

0.04-

X 0.02-Q

IGF-I

KM _. l (MI M

40 60

BOUND! i,M )

Fig. I. Scatchard analysis of '"I-IGF-I binding to IGFBP secreted by HT29-D4 cells. IGFBP was incubated with various concentrations (1.0 x 10~" to 2.7 xIO'' M) of '"I-IGF-I for 16 h at 4Â°Cand then analyzed for specific IGFBP-associated radioactive binding as described in "Materials and Methods." The data

plotted represent the result of a representative experiment made in triplicate andanalyzed by LIGAND. , computer-generated best fit for a one-site bindingmodel.

3183




BOUND(pM)

Fig. 2. Scatchard analysis of '"I-IGF-II binding to IGFBP secreted by HT29-D4 cells. IGFBP was incubated with various concentrations (1.0 x 10~" to 2.7 x10-' M) of '"I-IGF-II for 16 h at 4'C and then analyzed for specific IGFBP-associated radioactive binding as described in "Materials and Methods." The data

plotted represent the result of a representative experiment made in triplicate andanalyzed by LIGAND. , computer-generated best fit for a two-site bindingmodel; , two binding sites.

3o

10" 10a

SURAMIN IpM)

Fig. 3. Suramin inhibition of '"I-IGFs binding to HT29-D4 cell surface typeI IGF receptors. HT29-D4 cell monolayers were incubated with suramin at theindicated concentrations with either '"I-IGF-I {â€¢)or '"I-IGF-II (D) for 2 h at4'C. Suramin was added immediately prior to adding the labeled material. Specificcell-associated binding was determined as described in "Materials and Methods"

and expressed as a percentage of binding achieved in the absence of the drug.Data represent the mean of three duplicate experiments. SD was always less than

IGF-IGFBP complexes using IGFBP isolated from HT29-D4conditioned medium. The 50% inhibitory concentration was0.04 and 0.05 mM, i.e., about 60 and 75 Me/ml, respectively.

These experiments show that the amount of suramin requiredfor the dissociation of IGF-IGFBP complexes was about 3times lower than the amount able to inhibit IGF binding to cellsurface receptors. We therefore attempted to determine whetherthis differential effect of suramin could allow the specific binding of radiolabeled IGF-II, complexed to IGFBP, to the HT29-D4 cell surface type I IGF receptors. HT29-D4 cells weretherefore incubated at 4Â°Cwith '"I-IGF-II-IGFBP complexesand various amounts of suramin, and then specific 125I-IGF-II

cell binding was determined. In agreement with our previousresults (3) showing that IGFBP bound all secreted IGF-II thuspreventing its binding to cell surface type I IGF receptors, no

specific I25I-IGF-II binding to cells was observed when HT29-D4 cells were incubated with ' "I-IGF-II-IGFBP alone (Fig. 5).

However, the addition of suramin allowed the specific bindingof I25I-IGF-II from the radiolabeled complexes to the cell

surface (Fig. 5). This was observed only in a definite range ofsuramin concentrations between 0.02 and 0.1 mM with anoptimum at 0.04 mM, i.e., ~60 ng/m\. These suramin concentrations are identical to the ones found to be able to dissociateIGF-II from IGF-II-IGFBP complexes without preventing IGFbinding to cell surface type I IGF receptors (compare Figs. 3and 4). At this optimum suramin concentration, about 50% ofI25I-IGF-II prebound to IGFBP was released, i.e., -30,000 cpm.Thus, the percentage of I25I-IGF-II bound to cells in theseexperimental conditions (~8% of released free IGF-II) was of

the same order of magnitude as the one obtained when about0.1 ng I25I-IGF-II (-30,000 cpm) was mixed with HT29 cells.

SURAMINIpM)

Fig. 4. Effect of suramin on the dissociation of'"I-IGFs from IGFBP isolatedfrom HT29-D4 conditioned medium. Complexes of '"I-IGF-IGFBP prepared asdescribed in "Materials and Methods" were incubated with suramin at theindicated concentrations for 2 h at 4'C and then analyzed for I2'I-IGF-I (â€¢)or'"I-IGF-II (D) specific binding which remained bound to IGFBP as described in"Materials and Methods." Data expressed as a percentage of the binding achieved

in the absence of the drug represent the mean of three duplicate experiments. SDwas always less than 5%.

10 103

SUR AMIMI pM I

Fig. 5. Induction of '"I-IGF II binding from '"I-IGF-II-IGFBP complexes toHT29-D4 cell surface. HT29-D4 cell monolayers were incubated with '"I-IGF-II-IGFBP complexes (about 6.O x 10* cpm ) and suramin at the indicatedconcentrations for 2 h at 4'C. Specific cell-associated binding was determined asindicated in "Materials and Methods" and expressed as a percentage of the total

radioactivity added. Data represent the mean of three triplicate experiments. SDwas always less than 5%.

3184




Fig. 6. CEA release during the tentative induction of HT29-D4 cell differentiation by IGF-I and Â«IR-3.HT29-D4 cells were plated in 25-cnr flasks andgrown in standard culture conditions. Two days later, the culture medium wasreplaced by either standard DMEM (A), DMEM containing 2.5 Mg/ml Â«IR-3(D), or 2.5 jig/ml IGF-I (â€¢).The medium was reneweded every day. and cellswere subcultured two times (arrows) and then further cultured in the sameexperimental conditions as above. The amount of CEA was measured in each 24-h medium collection as described in "Materials and Methods." Points, mean of

triplicate determinations from three separate experiments. SD was less than 5%.*, IGF-l-treated cells versus control cells (P < 0.01 ).

It was also verified that HT29-D4 cells did not release IGFBPinto the medium in the binding experimental conditions used,i.e., 2 h at 4"C. A repartitioning of the labeled IGF between

released IGFBP and the HT29-D4 cell surface can therefore beexcluded.

Induction of Differentiation Markers in HT29-D4 Cells Cultured in the Presence of Exogenously Added IGF-I or Â«IR-3.The results reported above suggest that suramin could be ableto dissociate IGF-II-IGFBP complexes, thus allowing free IGF-II to interact with HT29-D4 cell surface type I IGF receptorsto create an autocrine loop. If such a mechanism is involved, atleast in part, in the ability of suramin to induce HT29-D4 celldifferentiation (17, 18), a similar triggering must be obtainedby adding IGF-II to the culture medium. However, we haveshown above that HT29-D4 cell-secreted IGFBP had both ahigh binding capacity and a high affinity for IGF-II. In contrast,IGF-I, which was poorly bound by IGFBP, should have a bettercapacity than IGF-II to deliver a biological action via the cellsurface type I IGF receptors. Moreover, Â«IR-3,a monoclonalantibody against the type I IGF receptor (29), might induce asimilar triggering, since it was shown to be an agonist antibodyespecially able to activate the type I IGF receptor-associatedtyrosine kinase (30).

HT29-D4 cells growing in standard culture conditions weretherefore exposed to an excess of either IGF-I (2.5 ng/mi) orÂ«IR-3(2.5 Â¿ig/ml)48 h after seeding; this supplemented medium was then renewed daily in order to maintain high concentrations of both glucose and either IGF-I or Â«IR-3.Since theHT29-D4 cell differentiation process was found to be highlycorrelated with CEA apical cell surface expression and release(31), the amount of CEA released into the culture medium wasquantified every day. As shown in Fig. 6, this amount increasedsharply 3 days after the cells had reached confluency. At theoptimum (day 13), a 10- and 3-fold CEA concentration increasewas measured in the culture medium of IGF-I-treated (63.6 ng/ml/24 h) and Â«IR-3-treated (20.8 ng/ml/24 h) cells, respectively, as compared to untreated control cells (6.5 ng/ml/24 h).CEA release strongly decreased after two successive passages.The most important fall was observed after the first passage (4and 1.5 times for IGF-I-treated and Â«IR-3-treated cells, respectively). However, CEA release remained significantly higher intreated cells (7.7 and 6.1 ng/ml/24 h, respectively) with respect

to undifferentiated control cells (3.5 ng/ml/24 h) even after thefirst two passages (day 24) (P < 0.01) (Fig. 6). It was routinelyverified that the culture medium of IGF-I-treated cells stillcontained free IGF-I after a 24-h culture period. At the lightmicroscope level, semithin sections of confluent HT29-D4 cells(day 13) showed multilayers 2-4 cells thick (Fig. 7). At thistime, we observed numerous vacuoles in IGF-I-treated cells(Fig. IB) which were absent in control cells (Fig. 1A). Thesevacuoles are in fact intercellular cysts and were shown byelectron microscopy to be small foci of differentiation. Thecellular membranes facing the lumen of the cysts indeed displayed well-organized microvilli with typical cytoskeletal rootlets related to the axis of the microvilli (Fig. 8). At day 13, theultrastructural observation did not evidence a polarized mono-layer with a well-organized apical brush border as previouslyreported for either suramin- or galactose-differentiated HT29-

D4 cells (6, 17, 18). However, at day 24, semithin sections ofIGF-I-treated cultures exhibited rare areas which were a cellmonolayer, whereas few intercellular cysts were seen. By transmission electron microscopy, such cells were polarized with anapical brush border facing the medium and tight junctions (notshown).

DISCUSSION

We have demonstrated recently that the HT29 human coloniecarcinoma cell line releases IGF-II into the culture medium (3).This growth factor is not free but totally complexed to severalmolecular forms of IGFBP (3) that have been identified asÂ¡soformsof the IGFBP-4 class (4). In addition, low amounts ofsoluble truncated type II IGF receptors (about 500-fold lessthan IGFBP) are also found in the cell conditioned medium(26). HT29-D4 cells express type I IGF receptors, whereasneither membrane type II IGF receptors nor membrane-linkedIGFBP are detectable (5). It is generally acknowledged thatboth IGF-I and IGF-II signaling is mediated by the type I IGFreceptor (7, 32). Moreover, IGFs have been shown to play anessential role in cell differentiation in several cell models (lilo). We have therefore searched for a central role for autocrinesecretion of IGF-II during the HT29-D4 cell differentiationprocess. Moreover, the previous finding that such a process canbe induced by suramin in the presence or absence of FCS (17,18) also suggests that this drug might act at the level ofautocrine growth factors.

In recent years, suramin has been used as a useful laboratory

Fig. 7. Morphology of HT29-D4 cells cultured in standard culture medium(A) or in medium supplemented with 2.5 Mg/ml IGF-I (B) for 13 days. T, top ofthe culture; K, empty vacuoles; ears. 10 f<m.

3185




Fig. 8. Transmission electron micrograph of HT29-D4 cells cultured in thepresence of 2.5 Â»jg/mlIGF-I for 13 days. Empty vacuoles, like those seen in Fig.IB, are intercellular cysts (1C) with a membrane facing the lumen displaying well-organized microvilli (m) with typical cytoskeletal rootlets (r); IS, intercellularspace; bar, 2 urn.

tool in experimental oncology to inhibit the binding of severalgrowth factors (e.g., epidermal growth factor, platelet-derivedgrowth factor, fibroblast growth factor, transforming growthfactor ÃŸ)to their receptors and to dissociate receptor-boundgrowth factors (19-22). These properties of suramin have beenused to shunt extracellular autocrine growth factor loops in avariety of cultured neoplastic cells (1, 19, 20). Autocrine/paracrine loops have been reported to be a characteristic of thetumor phenotype, since they confer a partial or complete autonomy on the cell (33, 34). However, several recent observations lead to the emerging picture that autocrine growth factorstimulation is not necessarily a transforming event and thatautocrine/paracrine stimulation may be a physiological mechanism used to exercise control over growth and differentiationof normal cells in many tissues, especially in the intestinalepithelium (35-38). This may be consistent with a model inwhich the disappearance of a biological response of cells to aspecific growth factor they normally release might lead todefects in the regulation of the delicate balance between cellproliferation and cell differentiation.

In this study, we investigate what could be the mechanismsthat account for HT29-D4 cell differentiation induced by suramin. An attractive hypothesis is that IGF-II secreted by HT29-D4 cells may be involved in the state of cell differentiationthrough its interaction with the type I IGF receptors displayedat the cell surface. However, free IGF-II is not available forsignaling via these receptors, since this growth factor is tightlybound to IGFBP, preventing its binding to receptors (3). Also,neither IGF-II nor IGFBP was found to be bound to the surface

of HT29-D4 cells (5), suggesting that an IGF autocrine loop isnot operative in these cells. In the same way, we show here thatpreformed I25I-IGF-II-IGFBP complexes do not allow I2SI-IGF-II binding to the HT29-D4 cell surface (Fig. 5). Such a patternappears to be due to the sequestration of IGF-II in the culturemedium by high amounts of IGFBP expressing a high affinityfor IGF-II, and this in contrast to its affinity for IGF-I (Figs. 1and 2).

The data presented here show that suramin is able to dissociate IGF-II-IGFBP complexes at concentrations between 40and 100 Mg/ml (Fig. 4), thereby allowing the release of freeIGF-II into the culture medium. It is quite possible that suraminis also able to induce an IGF-II release from soluble type IIIGF receptors. These drug concentrations are identical to theones previously found to induce an optimum enterocyte-likedifferentiation of HT29-D4 cells (17, 18). In contrast, they aresignificantly lower than the ones required to antagonize freeIGF-II binding to HT29-D4 cell surface type I IGF receptors,i.e., 200-400 Â¿Â¿g/ml(Fig. 3). Thus, it is possible that free IGF-II released from IGFBP under well-defined suramin concentrations interacts with the membrane type I IGF receptors totrigger a potential autocrine biological effect, e.g., the inductionof a cell differentiation pathway. If such a hypothesis is correct,it would be possible to induce HT29-D4 cell differentiation byadding a ligand able to directly interact with the cell surfacetype I IGF receptors, i.e., bypassing IGF-II high-affinity IGFBPin the culture medium. This was attempted by adding an excessof recombinant IGF-I, since secreted IGFBPs are found toexpress a low affinity for this ligand. Under such an experimental design, HT29-D4 cells release high amounts of CEA (about2.3 ng/106 cells/24 h at the optimum), a recognized marker ofHT29-D4 cell differentiation (31). Â«IR-3is also a partial agonist in HT29-D4 cells, since with Â«IR-3CEA release is aboutone-third the levels obtained with IGF-I. This result is inagreement with a recent report indicating that this antibody isable to signal via the stimulation of the type I IGF receptorkinase (30). It is noteworthy that the amount of CEA releasedby IGF-I-treated or aIR-3-treated HT29-D4 cells decreaseswith both passages and culture time. However, it is conceivablethat under our experimental conditions, the type I IGF receptors are down-regulated by the high concentrations of exoge-nously added IGF-I or aIR-3. Concurrently with the CEArelease, the ultrastructural observation of IGF-I-treated HT29-D4 cells shows the first morphological signs of cell differentiation. The numerous intercellular cysts seen are considered asinitiation sites for the differentiation and the reorganization ofthe epithelium in a cell monolayer (39-41). However, we rarely

observed tight junctions in these structures, which explains thefact that CEA, which is exclusively released at the apical sideof the differentiated HT29-D4 cells (31), can still be measuredin the culture medium. However, the perfect cellular organization observed with growing cells either in the absence of glucose(6) or in the presence of suramin (17, 18) is never observed.

In HT29-D4 cells, the IGF-I-type I IGF receptor interactionmay represent only part of the mechanism by which thesecolonie cells control their differentiation. Other, multiple cellular events may have occurred which, primarily in combinationwith IGF-I, are involved in the partial cell differentiation process here observed. On the other hand, the incapacity of IGF-I-treated HT29-D4 cells to reach a fully differentiated state maybe related to the necessity to deliver yet unidentified additionalbiological signals and/or to suppress the constitutive activationof metabolic pathways that prevent cells from fully differentiat-

3186



AUTOCRINE ROLE OF IGF-ll IN SURAMIN-INDUCED DIFFERENTIATION

ing. In particular, the roles of the other growth factors alsosecreted by HT29-D4 cells, e.g., transforming growth factor aand ÃŸ(2), may be addressed, since they may act by eitherstimulating or inhibiting cell proliferation and/or celldifferentiation.

Finally, several classes of IGFBP have been identified thatmodulate IGF actions in a positive or negative fashion. TheseIGFBPs express relative differences in affinity for IGF-I andIGF-II that may determine the bioavailability of the IGFs totheir target cells (42-45). Thus, the factors that control thelevel and type of IGFBP secreted into the conditioned mediumin vitro as well as into the extracellular spaces in vivo may beimportant modulators of the balance between inhibitory orenhancing IGF effects on the colonie epithelial cells. Whetherthe molecular forms of IGFBP secreted by cancer colonieepithelial cells and by their nonmalignant counterparts aredistinct is yet to be determined.

In summary, suramin, a drug that has been widely used as alaboratory tool to shunt growth factor autocrine loops, exhibitshere an unusual opposite property allowing such a self-triggering. This potential is linked to the differential effects deliveredby this drug on IGF-II interactions with either the secretedIGFBP or the cell-associated type I IGF receptors. Furthermore, the treatment of undifferentiated HT29-D4 cells with anexcess of either IGF-I or oIR-3 to mimic the putative autocrineloop results in the induction of an intermediate stage of celldifferentiation that is associated with the appearance of typicalluminal membranes with microvilli and an important CEArelease, a known marker of HT29-D4 cell differentiation.

ACKNOWLEDGMENTS

We thank R. Ranee for her excellent technical assistance.

REFERENCES

1. Culouscou, J. M., Garrouste, F., Remacle-Bonnet. M.. Bettetini, D., Mar-valdi, J., and Pommier, G. Autocrine secretion of a colorectum-derivedgrowth factor by HT-29 human colon carcinoma cell line. Int. J. Cancer.,Â«.â€¢895-901.1988.

2. Anzano. M. A., Rieman. D., Prichett. W.. Bowen-Pope, D. F.. and Greig, R.Growth factor production by human colon carcinoma cell lines. Cancer Res.,49: 2898-2904, 1989.

3. Culouscou. J. M., Remacle-Bonnet, M.. Garrouste, F., Fantini, J.. Marvaldi,J., and Pommier. G. Production of insulin-like growth factor II (IGFII) anddifferent forms of IGF-binding proteins by HT-29 human colon carcinomacell line. J. Cell. Physiol.. 143:405-415, 1990.

4. Culouscou, J. M.. and Shoyab, M. Purification of a colon cancer cell growthinhibitor and its identification as an insulin-like growth factor bindingprotein. Cancer Res.. 51: 2813-2819, 1991.

5. Remacle-Bonnet. M. M.. Culouscou, J. M., Garrouste, F. L.. Rabenandra-sana, C., Marvaldi, J. L., and Pommier. G. J. Expression of type I, but nottype II insulin-like growth factor (IGF) receptor on both undifferentiated anddifferentiated HT29 human colon carcinoma cell line. J. Clin. Endocrino!.Metab., in press, 1992.

6. Fantini. J., Abadie, B.. Tirard, A., Remy, L.. Ripert, J. P.. El Batari, A., andMarvaldi. J. Spontaneous and induced dome formation by two clonal cellpopulations derived from a human adenocarcinoma cell line, HT-29. J. CellSci., S3: 235-249. 1986.

7. Humbel, R. E. Insulin-like growth factors I and II. Eur. J. Biochem.. 790:445-462, 1990.

8. Daughaday, W. H., and Rotwein. P. Insulin-like growth factors I and II.Peptide, messenger ribonucleic acid and gene structures, serum, and tissueconcentrations. Endocr. Rev., 10: 68-91, 1989.

9. Tricoli. J. V., Rail, L. B., Karakousis, C. P., Herrera, L., Petrelli, N. J.. Bell,G. I., and Shows, T. B. Enhanced levels of insulin-like growth factor messenger RNA in human colon carcinomas and liposarcomas. Cancer Res., 46:6169-6180, 1986.

10. Lambert, S., Vivario, J.. Boniver, J., and Gol-Winkler, R. Abnormal expression and structural modification of the insulin-like growth factor-II gene inhuman colorectal tumors. Int. J. Cancer. 46: 405-410, 1990.

11. Holly, J. M. P., and Wass, J. A. H. Insulin-like growth factors; autocrine,paracrine or endocrine? New perspectives of the somatomedin hypothesis in

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.3187

the light of recently developments. J. Endocrinol., 122:611-618, 1989.Fontana, J. A., Burrows-Mezu, A., Clemmons, D. R.. and LeRoith. D.Retinoid modulation of insulin-like growth factor-binding proteins and inhibition of breast carcinoma proliferation. Endocrinology, 128: 1115-1122,1991.Rosenthal, S. M., Brunetti, A.. Brown. E. J., Mamula, P. W., and Goldfme.I. D. Regulation of insulin-like growth factor (IGF) I receptor expressionduring muscle cell differentiation. Potential autocrine role of IGF-II. J. Clin.Invest., 87: 1212-1219, 1991.Pahlman, S., Meyerson, G., Lingren, E., Schalling, M., and Johansson, I.Insulin-like growth factor I shifts from promoting cell division to potentiatingmaturation during neuronal differentiation. Proc. Nati. Acad. Sci. USA, 88:9994-9998, 1991.Tollefsen, S. E., tajara, R.. McCusker, R. H., Clemmons, D. R., andRotwein. P. Insulin-like growth factors (IGF) in muscle development. Expression of IGF-I, the IGF-1 receptor, and an IGF binding protein duringmyoblast differentiation. J. Biol. Chem.. 264: 13810-13817. 1989.Florini, J. R., Magri, K. A., Ewton. D. Z... James, P. L., Grindstaff, K., andRotwein, P. S. "Spontaneous" differentiation of skeletal myoblasts is dependent upon autocrine secretion of insulin-like growth factor-II. J. Biol.Chem., 266: 15917-15923, 1991.Fantini, J.. Rognoni, J. B., Roccabianca, M., Pommier, G., and Marvaldi, J.Suramin inhibits cell growth and glycolytic activity and triggers differentiation of human colic adenocarcinoma cell clone HT29-D4. J. Biol. Chem.,264: 10282-10286, 1989.Fantini, J., Verrier, B.. Robert, C., Pic, P., PichÃ³n, J., Mauchamp. J., andMarvaldi, J. Suramin-induced differentiation of the human colic adenocarcinoma cell clone HT29-D4 in serum-free medium. Exp. Cell Res., 189:109-117, 1990.Betsholtz, C., Johnsson, A.. Heldin, C. H., and Westermark, B. Efficientreversion of simian sarcoma virus-transformation and inhibition of growthfactor-induced mitogenesis by suramin. Proc. Nati. Acad. Sci. USA, 83:6440-6444, 1986.Huang, S. S., and Huang, J. S. Rapid turnover of the platelet-derived growthfactor receptor in sis-transformed cells and reversal by suramin. Implicationsfor the mechanism of autocrine transformation. J. Biol. Chem., 263: 12608-12618. 1988.Coffey. R. J., Leof. E. B., Shipley, G. D., and Moses, H. L. Suramin inhibitionof growth factor receptor binding and mitogenicity in AKR-2B cells. J. Cell.Physiol., 132: 143-148, 1987.Pollak. M.. and Richard, M. Suramin blockade of insulinlike growth factorI-stimulated proliferation of human osteosarcoma cells. J. Nati. Cancer Inst.,82: 1349-1352, 1990.Hawking. F. Suramin: with special reference to onchocerciasis. Adv. Phar-macol. Chemother., /5: 289-322. 1978.Armand. J. P., and Cvitkovic, E. Suramin: a new therapeutic concept. Eur.J. Cancer. 26: 417-419. 1990.La Rocca. R. V., Stein, C. A., and Myers, C. E. Suramin: prototype of a newgeneration of antitumor compounds. Cancer Cells (Cold Spring Harbor), 2:106-115, 1990.Garrouste, F.. Remacle-Bonnet. M.. Culouscou, J. M., Marvaldi, J., andPommier, G. Type-Il insulin-like growth-factor receptor in conditioned medium from HT-29 human colon carcinoma cell line. Int. J. Cancer, 47: 760-764, 1991.Munson, P. J.. and Rodbard. D. LIGAND: a versatile computerized approachfor the characterization of ligand binding systems. Anal. Biochem.. 707:220-239, 1980.McPherson, G. A. Analysis of radioligand binding experiments: a collectionof computer programs for the IBM PC. J. Pharmacol. Methods. 14: 213-228, 1985.Flier, J. S., Usher, P., and Moses, A. C. Monoclonal antibody to the type Iinsulin-like growth factor (IGF-I) receptor blocks IGF-I receptor-mediatedDNA synthesis: clarification of the mitogenic mechanisms of IGF-I andinsulin in human fibroblasts. Proc. Nati. Acad. Sci. USA. 83:664-668. 1986.Steele-Perkins, G., and Roth. R. A. Monoclonal antibody Â»1R-3inhibits theability of insulin-like growth factor II to stimulate a signal from the type Ireceptor without inhibiting its binding. Biochem. Biophys. Res. Commun..///: 1244-1251, 1990.Fantini. J.. Rognoni. J. B., Culouscou. J. M., Pommier. G., Marvaldi. J..and Tirard. A. Induction of polarized apical expression and vectorial releaseof carcinoembryonic antigen (CEA) during the process of differentiation ofHT29-D4 cells. J. Cell. Physiol., 141: 126-134, 1989.Czech, M. P. Signal transmission by the insulin-like growth factors. Cell, 59:235-238. 1989.Browder, T. M., Dumbar, C. E., and Nienhuis, A. W. Private and publicautocrine loops in neoplastic cells. Cancer Cells (Cold Spring Harbor), /: 9-17. 1989.Lang. R. A., and Burgess, A. W. Autocrine growth factors and tumourigcnictransformation. Immunol. Today, //: 244-249. 1990.Markowitz, S. D., Molkentin, K., Gerbic, C., Jackson. J., Stellato, T.. andWillson, J. K. V. Growth stimulation by coexpression of transforming growthfactor-Â«and epidermal growth factor-receptor in normal and adenomatoushuman colon epithelium. J. Clin. Invest., 86: 356-362, 1990.Barnard, J. A., Beauchamp. R. D., Coffey, R. J., and Moses, H. L. Regulationof intestinal cell growth by transforming growth factor type tÃ.Proc. Nati.Acad. Sci. USA. 86: 1578-1582. 1989.Suemori. S.. Ciacci. C., and Podolsky, D. K. Regulation of transforming




growth factor expression in rat intestinal epithelial cell line. J. Clin. Invest., 42. Busby. W. H., Klapper, D. G., and Clemmons. D. R. Purification of a 31000-87: 2216-2221, 1991. dalton insulin-like growth factor binding protein from human amniotic fluid.

38. Cross, M., and Dexter, T. M. Growth factors in development, transformation. Isolation of two forms with with different biologic actions. J. Biol. Chem.,and tumorigenesis. Cell, 64: 271-280, 1991. 26i: 14203-14210, 1988.

39. Remy, L., Marvaldi, J., Rua, S., Secchi, J., and Lechene de la Porte, P. The 43- DeMellow, J S. M., and Baxter, R. C. Growth hormone-dependent insulin-role of intracellular lumina in the repolarization process of a colonie adeno- llke 8iÂ°wÂ«h*gf <IG? b.lndln8 Protein both inhibits and potentiates IGF-I

.... ... . . , ,f, ,, â€ž¿�... ,. ., .,â€ž, ,â€ž- lno. stimulated DNA synthesis in human skin fibrob asts. Biochem. Biophys.carcinoma cell Ime. Virchows Arch. (CelKPatho 1), 46: 297-305, 1984. Res Commu ,5Â¡. 199_204 ,9gg40. Le BIVIC,A., Hirn, M., and Regg.o, H. HT-29 cells are an ,n v,,w model for 44 E|gjn R G ÃŸusbyw H and clemmons_ D R An Â¡nsuiin.likegrowlh

the generation of cell polarity in epithelia during embryonic differentiation. fac,or â€ž¿�GF)bindi tein enhances the biologie responses to IGF-I. Proc.Proc. Nati. Acad. Sci. USA, 85: 136-140, 1988. Nati. Acad. Sci. USA- 84: 3254-3258, 1987.

41. Godefroy, O., Huet, C., Blair, L. A. C, Sahuquillo-Mermo, C.. and Louvard, 45. Mohan, S., Bautista, C. M., Wergedal, J., and Baylink. D. J. Isolation ofD. Differentiation of a clone isolated from the HT29 cell line: polarized inhibitory insulin-like growth factor (IGF) binding protein from bone celldistribution of histocompatibility antigens (HLA) and of transferrin recep- conditioned medium: a potential local regulator of IGF action. Proc. Nati,tors. Biol. Cell, 63: 41-55, 1988. Acad. Sci. USA, 86: 8338-8342, 1989.

3188



1992;52:3182-3188. Cancer Res Gilbert J. Pommier, Françoise L. Garrouste, Fatima El Atiq, et al. Adenocarcinoma Cell LineSuramin-induced Differentiation of HT29-D4 Human Colonic Potential Autocrine Role of Insulin-like Growth Factor II during

Updated version

http://cancerres.aacrjournals.org/content/52/11/3182

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/52/11/3182To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



potential autocrine role of insulin-like growth factor ii...

Documents