Vol. 7, 1 705-i 712, December 1996 Cell Growth & Differentiation 17�

p2Vas�mediated Decrease of the Retinoblastoma Protein inFibroblasts Occurs through Growth Factor-dependentMechanisms’

Laura Kivinen, Erja Tiihonen, Tarja Haapajarvi, andMarikki Laiho2Haartman Institute, Department of Virology, University of Helsinki, P. 0.

Box 21 (Haartmaninkatu 3), FIN-00014 Helsinki, Finland

Abstract

Stable coexpression of the human retinoblastomaprotein (PRB) cDNA and EJ c-Ha-ras oncogene inmunne fibroblasts leads to loss of pRB expression withconcomitant transformation of the cells (1). We showhere that conditional expression of p21� in mousefibroblasts expressing human pRB leads to a rapiddecrease of pRB expression at both protein and mRNAlevels. The decrease of pRB mRNA is blocked bycycloheximide, suggesting the requirement of ongoingprotein synthesis. p21�3 expression leads also todecreases of c-myc and tissue metalloproteinaseinhibitor-2 mRNAs, whereas cyclin-dependent kinase 4,cyclin , E2F-1 , and omithine decarboxylase areunaffected. The decrease in pRB is accompanied byprogressive morphological transformation of the cells.The effect of p21�8 on pRB expression was serum andgrowth factor dependent. A shift of the cells to lowserum (0.2% FCS) abolished the effects of p21’� onpRB, but this effect was reconstituted by the additionof growth factors epidermal growth factor, fibroblastgrowth factor-2, transforming growth factor �31 , andplatelet-derived growth factor to the cells. The resultssuggest a complex interaction between p2V�, pRB,and growth factors in the control of cell growth. p21�uappears to drive the cell cycle by deregulation of keycell cycle regulators, the functions of which in lowserum become redundant or require the presence ofgrowth factors positively driving the cell cycle.

Introduction

p2imS is mutated at high frequency in human cancer (re-viewed in Ref. 2). The mutant p2i� confers its transformingproperty apparently by deregulation of cell signaling cas-cades. Residing in plasma membrane, the normal p2i’� isactivated to its GTP-bound form by tynosine phosphorylation

Received 7/15/96; accepted 10/24/96.The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1 734 solely to mdi-cate this fact.1 Supported by the Academy of Finland, the Sigrid Juselius Foundation,and the Finnish Cancer Organizations.2 To whom requests for reprints should be addressed. Phone: 358-9-4346509; Fax: 358-9-4346491 ; E-mail: Marikki.Laiho@Helsinki.Fl.

of several growth factor receptors and their interactionthrough SOS and Gnb2 proteins with ras (reviewed in Refs. 3and 4). Shortly after activation, the normal p2lrns is hydro-lyzed to the GDP-bound form turning the signaling effect off,

whereas the mutant forms remain active. The GTP-boundform of ras activates subsequently several effector systems,including the raf-MEK-MAP kinase cascade (reviewed in Ref.

5). Autonomous activation of the kinase cascade by rasmutation positively drives the cell cycle and leads to induc-tion of early-response genes. The ras signaling cascade is,however, modulated by other signaling events initiated by itsprincipal, but not exclusive activators, the receptor tyrosinekinases (reviewed in Ref. 6). In the cell growth cycle, p2i�2activity is central and appears to be required during multiplepoints in the progression of the G1 phase of the cell cycle (7).Ras function is also required for gene expression and mat-uration of Xenopus oocytes (3), insulin-derived growth fac-ton-I-stimulated cell cycle progress in late G1 (8), and fornerve growth factor-mediated differentiation of PCi 2 cells(9).

Expression of mutant p21� in immortalized fibroblasts

leads to cell transformation and increased metastatic capac-ity, which is associated with increased production of proteo-lytic enzymes, cathepsins, and type IV collagenase, do-creased expression of proteinase inhibitors such as TIMPs3,and extracellular matrix genes such as fibronectin, type Icollagen, and syndecan-i (10-16). p2i� increases the ex-pression of calcium-binding proteins calcyclin and os-teopontin (17), ODC, and junD (18, 19). Mutant p21� tran-

scniptionally activates promoters for fos (20, 21), multidrugresistance gene- 1 (22), TGF-j31 (23), transin (1 9), and colla-genase (21). Furthermore, posttnanslational modification ofc-jun by p2i’� increases AP-i binding activity (24), despitethat under continuous presence of ras the c-fos levels arelow (20).

Transgenic animal models have indicated that ras expres-sion causes either hyperplasia (25), concomitant with muta-tion of p53 tumor suppressor gene (26), on malignanciestogether with other somatic changes (25). Furthermore, thepresence of mutant ras in the cells is suggested to lead togenomic instability portrayed by frequent chromosomal ab-ennations (27). Similarly, expression of ras for even briefperiods in mammary tumor cells, despite its subsequentremoval, maintains the metastatic capacity of the cells, sug-

3 The abbreviations used are: TIMP, tissue inhibitor of metalloprotemnase;ODC, ornithine decarboxylase; TGF, transforming growth factor; pRB,retinoblastoma protein; IPTG, isopropyl-1 -thio-f3-o-galactopyranoside;cdk, cyclin-dependent kinase; MAP, mitogen-activated protein; EGF,epidermal growth factor� FGF, fibroblast growth factor� PDGF, platelet-derived growth factor; GAPDH, glyceraldehyde-3-phosphate dehy-drogenase.

17� p21’� Decreases RB Expression

4 K. Pitk#{226}nenand L Kivinen, unpublished observations.

gesting permanent changes in gene expression or genomicintegrity (28). Therefore, mutant p2lrns appears to derail thecellular homeostasis at multiple points by creating inadver-tant positive growth signals and gene regulation and, addi-tionally, by promoting genomic instability.

We have analyzed the ability of mutant ras to affect cellgrowth control by interacting with tumor suppressor genes.The tumor suppressor gene p53 suppresses the transform-ing activity of mutant ras (29, 30) and similarly, albeit transi-tionally, does the human retinoblastoma gene product (pRB;Ref. 1). We have found previously that stable coexpression ofpRB and mutant p21’� leads to progressive morphologicaltransformation of cells and loss of pRB expression. Cellslosing pAB expression maintained the RB expression vec-tons, whereas neither pRB mRNA nor protein were detecta-ble (1). The decrease was specific for wild-type pRB becauseno change occurred in the expression of mutant pRB simi-larly coexpressed in the cells with p21’�. These results sug-gested that p2irns is able to modulate pRB expression andthus relieve cells from the growth control exerted by pRB. Inthe current work, we have established conditions to analyzethe interplay between p21� and pRB by using conditionalexpression of p2i�3 in mouse cells expressing high levels ofpRB. We find that the activation of p2i’� expression leads toa decrease of pRB expression in a serum and growth factor-dependent manner.

Results

Conditional Expression of p21�5 Leads to a Decrease ofpRB Expression and Cell Transformation. A lactose ana-logue (IPTG)-inducible expression vector driving the expres-sion of mutant ras p21 l2Gly-val (pSVlacOras; Ref. 31) wasused to study the interaction between p2i� and pRB. Theexpression of p21� is driven by the SV4O promoter underthe influence of the prokaryotic lac-operon. Simultaneousexpression of the lactose repressor vector (laclneo) inhibitsgene transcription unless a lactose analogue (IPTG) is pre-sented to the cells. Both constructs were transfected toNIH3T3 cells and NIH3T3 RB4.6 cells stably expressing thehuman retinoblastoma cDNA at high levels (32), and cellcolonies were selected for additional studies by their sus-ceptibility to transformation in the presence of IPTG (Fig. 1B).Despite profound transformed morphology of the p21�-expressing cells cultivated in the presence of IPTG, therewas neither a change in their doubling time nor in the distni-bution to the different compartments of the cell cycle (Fig. 1Aand data not shown).

The effect of p2i� on the expression of pRB was studiedin two different clones (RB4.6ras5 and RB4.6ras6) stablyoverexpressing human pRB. Induction of p2i� by IPTG for3 days led to a decrease of pRB expression as detected byimmunoblotting analyses (Fig. 2A). The decrease in pRBprotein levels did not affect its phosphorylation pattern, asindicated by a decrease of both faster and slower migratingpRB forms. There was neither a change in pRB protein levelsin the parental RB4.6 cells nor NIH3T3 cells expressing theIaclneo vector only, when incubated in the presence of IPTG,

indicating that the observed change was specific for p2i�5expression. As control, the regulation of extracellular matrix

proteins in IPTG-treated NIH3T3ras and RB4.6ras cloneswas studied. Changes previously attributed to p2lrns, suchas down-regulation of fibronectin expression (1 3), took placesimilarly in both RB4.6ras and NlH3T3ras clones upon p21�induction (data not shown). Furthermore, the levels of severalextracellular matrix proteins were unaffected by p21�SS ox-pression, indicating the specificity of the system. The induc-tion of p2l� expression was pronounced in the selectedclones and occurred within 24 h with very little if any basalexpression in low cell passages (Fig. 2). The decrease inpRB, as detected by immunoblotting, became evident within2 days and was prominent 3-4 days after turning on p2l�expression (Fig. 2B). Furthermore, turning off the continuous

synthesis of p2i�as by washing off IPTG restored the expres-sion of pRB within 24 h (Fig. 2B, Lane 3 + 1).

The change in pRB expression was found to take place atthe mRNA level as detected by Northern analysis (Fig. 3). Inthe presence of p21’�, pRB mRNA levels decreased 1.7-3.4-fold in RB4.6ras clones. Because pRB expression fromthe cDNA is constitutively driven by SV4O promoter andtransient transfections of RB4.6 cells using SV4O promoter-chloramphenicol acetyltransferase gene showed no changein the chloramphenicol acetyltransferase activity upon p21�induction (data not shown), it is possible that p2i� exerts itseffects on pRB at posttranscniptional levels.

p21� Affects the Expression of c-myc and TIMP-2 butnot Cell Cycle-related Genes. Northern hybridizationanalyses were used to study the putative target genes ofp21� and those modulated by pRB expression. p21�sexpression upon IPTG induction was found to decreasemRNAs for c-myc and TIMP-2 1 .8-2.2- and 1 .6-2.6-fold,respectively, in both RB4.6ras clones (Fig. 3). In contrast,after ras induction there was no change in the mRNA levelsfor cdk 4 and cyclin Dl , positively driving the cell cycle andregulating the phosphorylation of pRB, transcription factorE2F-1 (data not shown), or ODC shown previously to beunder the control of p2Vas (Ref. 19; Fig. 3). The latterfinding is probably explained by the fact that the basallevels for ODC are considerably higher in the parentalRB4.6 cells as well as RB4.6ras clones in contrast towild-type NIH3T3 cells. Because the levels of ODC varyconsiderably in several other pRB-expressing clones andcannot be induced in cells expressing pRB conditionally,4the latter phenomenon is attributed to clonal variationinstead of a positive regulation of ODC by pRB. Similarly,the basal levels for c-myc and cyclin Di were higher in theclone RB4.6 and its derivatives. Whereas the increase inc-myc may also reflect clonal variation, cyclin Dl expres-sion is constitutively increased in RB-overexpressing cells

(33). Interestingly, no decrease of c-myc was observedafter induction of ras in clones that did not overexpresspRB, suggesting that the change is related to the level ofRB expression (data not shown).

To confirm that the observed decrease in pRB mRNAlevels required ongoing protein synthesis and p2l� expres-sion, RB4.6ras6 cells were incubated with IPTG and cyclo-

A25

20

�:15

c�q11�

I-

5

0

IPTG I- +i I - +1 I - +1 1 - +1 I- +1

Clone neo4 ras8 parental ras5 ras6

3T3

B

C,I-0.

3T3 RB4.6

Cell Growth & Differentiation 1707

Fig. 1. Effects of conditional expres-sion of p21� on cell proliferation andphenotype of NIH3T3ne0 and RB4.6cells. A, cell doubling times (Tv�). Cells(1O�) expressing neomycin or p2l�expression vectors, as indicated,were plated, and the cell numberswere determined after 1 , 2, and 4days. T112 of the cells in the absence(Li) and in the presence (�) of 10 m�IPTG are shown. T112 and SO are themean of three independent experi-ments. B, phenotype of the cellclones. Cells were cultured for 3 dayswithout (a-d) or with 10 m�i IPTG (e-h), and photomicrographs were takenwith 100-fold optical magnification. aand e, 3T3neo4; b and f, RB4.6; c andg, RB4.6ras5; d and h, RB4.6ras6.

e.

C, � � � :‘�-,-,

�+ � � ;‘c’

::‘; � ::,�

heximide for 2 days, and mRNAs were isolated and sub-

jected to Northern analysis. Cycloheximide was found to

prevent the decrease in pRB mRNA (Fig. 4), suggesting that

protein synthesis is required.

The Effect of p2l� � pRB Is Serum Dependent andCan Be Reconstituted by Growth Factors in Low Serum.To characterize the putative modulators of p21� on pRB

expression, RB4.6 cells and RB4.6ras cells were incubated

with IPTG and decreasing concentrations of serum. It was

found that the decrease of pRB by p21�s was abolished in

0.2% FCS (Fig. 5). Under low serum concentrations, the cells

were transformed by p2lras and in contrast to the parental

RB4.6 cells, were unable to growth arrest and maintained

high proliferative index, as measured by 5-bromodeoxyuri-

dine immunostaining (data not shown; Ref. 32). The bevel of

basal pRB expression was slightly decreased in low serumconcentrations. The inability of p2lras to affect pRB levels in

low serum suggested to us the involvement of serum-derived

growth factors as mediators of the p21’� effect. RB4.6ras

cells were, therefore, incubated in the presence of 0.2%

FCS, various growth factors, and IPTG. Analysis of the celllysates by immunoblotting indicated that EGF, FGF-2,TGF-pl and PDGF were able to reconstitute the decrease ofpRB (Fig. 6A). The addition of growth factors in low serum did

not change the percentage of 5-bromodeoxyuridine-incor-

porating cells, indicating that the observed changes in pRB

were independent of growth rate of the cells (data not

shown).

The growth factor-mediated decrease of pRB by p21 ras in

low serum was studied by Northern analysis. In cells incu-

bated in 0.2% FCS and IPTG, there was no decrease of pRBmRNA levels as opposed to a 1 .8-fold decrease in cells

grown in 10% FCS (Fig. 6, B and C). pRB mRNA bevels,however, were decreased by 6-fold in cells incubated with

TGF-�1 in low serum and IPTG. Similarly, although not quiteas pronounced effects were seen by EGF and FGF-2 (Fig. 6,B and C). CDK4 levels were, however, unaltered in all con-ditions studied (Fig. 6, B and C).

A

3T3neo4 parental ras5 ras6

+“ - +“ - +“- +IPTG

pRB -

B

3T3neo4 parental ras5 ras6

IPTG ‘- +“ - +,,- +,, .

hAS-

C’fT� -

IPTG(d) - 1 2 3 4 3+1

pRB-�

p2lAS

. .. �

Ixic 4 -

�- -

A

B

CydlnDl -

c�4pDH-

#{149}hRB

0 c-myc

N cdk 4

D11MP-2

F�C�iclin Dl

I’m’DiscussionWe find that induction of mutant p21�s expression down-regulates the protein and mRNA expression of human pRB.These events take place in cells constitutively expressinghuman pRB at high levels, require ongoing protein synthesis,and are serum and growth factor dependent. The ras effectswere directed at the exogenously highly expressed pRB andendogenous genes such as c-myc, TIMP-2, and fibronectin,whereas several cell cycle-dependent genes such as CDK4,cyclin Dl , E2F-1 , and ODC were unaffected. Of these, ras

has been observed previously to regulate TIMP (1 5) and

ODC, as well as certain early-response genes and multiplegenes involved in cell adhesion.

Ras exerts transcriptional, posttranscriptional, and post-translational control over gene expression. Posttranscrip-tional regulation by ras of fibronectin mRNA occurs throughaltered nuclear processing or altered stability of the pro-cessed nuclear mRNA (13), whereas the cytoplasmic mRNAfor ODC is stabilized (1 8). A decrease in cell surface expres-sion of syndecan-1 , on the other hand, occurs posttransla-tionally, and is associated with cell transformation (1 6). In thecase of ras-mediated decrease of exogenous pRB, it is un-likely that ras could affect transcription, since RB cDNA is

under the influence of SV4O promoter, neither splicing norpolyadenylation, because an already processed message isused and a poly(A)� addition signal is provided in the vector.

Therefore, ras is likely to affect either the stability of nuclear

or cytoplasmic RB mRNA or its transport to the cytoplasm.The effect of p2Vas on pRB protein is mediated within 24 h

after enhanced levels of p21ras begin to accumulate. Further-

1703 p21� Decreases RB Expression

- - � - �

Fig. 2. Effect of p21s on pRB expression. A, immunobbotting analysis.Cells were cultured with (+) or without (-) IPTG (10 mM) for 3 days, andcell lysates were prepared and analyzed by either 7.5% SOS-PAGE (150�.Lg of each bysate) for pRB or 12.5% SOS-PAGE (250 �ig of each lysate)for p21’�. B, kinetic analysis of the p21’�-mediated decrease of pRB.RB4.6ras5 cells were cuitured in the presence of IPTG (10 mM) for theindicated number of days. Abtematively, the cells were grown for 3 days inthe presence of IPTG, followed by washes and an additional incubation for1 day without IPTG (Lane 3 + 1). Lysates (200 �zg) were analyzed by9-15% gradient SOS-PAGE, and immunobbots were probed with mono-cbonal anti-pRB and monocbonab anti-p21� antibodies. The hyper- andhypophosphorylated forms of pRB are not discernible because ofthe highpobyacrybamide concentration of the gel. Migration of pRB and p2l� areindicated to the left.

�i -

� � � � - � �- -

ww - - � w w

flP�-2 - � � �- � - �-..

G4PDH-

4-

wci,3-

C.)

#{149}D

#{149}0

0LI.

2-

1-

0

w w -- w �

I �__iI II II I

3T3neo4 parental ras5 ras6

3T3RB4.6

Fig. 3. Effects of p21s on the expression of cell cycle-regulated genesand TIMP-2. A, cells were cultured for 3 days without (-) or with (+) IPTG(1 0 mM), and poly(A)� RNA (3 �xg) was isolated and analyzed by Northernblotting. Three different filters with the same mRNAs were hybridized withlabeled cONA inserts as indicated. The GAPOH signal for each filter isshown. B, quantitation of the Northern analyses. The signals were quan-titated and normalized against GAPOH levels. The fold decrease of eachsignal of cells grown in the presence of IPTG is shown. In clone 3T3neo4,the human pRB and c-myc bevels were below detection level.

IPTG - + - +

CHX - - + +

F�S(%)

hRB-

10 5 1 0.2

=: �

IPTG ‘- +“- +,,- +,,- +

pRB - � � �

Fig. 5. Effect of p21� on pRB is serum dependent. RB4.6ras5 cellswere plated in 1 0% FCS. The following day, the cells were washed twicewith serum-free medium, and medium containing increasing concentra-tions of FCS with (+) or without (-) IPTG (10 mM) was added; cells werethen incubated for 2 days. Cell lysates (150 �g) were analyzed by 9-15%gradient SOS-PAGE, followed by Western blotting with pRB antibodies.Migration of pRB is indicated to the left.

1I �JI

2

Fig. 4. Cycloheximide abolishes the decrease of RB mRNA. RB4.6ras6cells were cultured with (Lanes 2) or without (Lanes 1) cycloheximide (10.tglml) in the presence or absence of IPTG (1 0 mM) for 2 days as indicated,

and poly(A) � RNAs were prepared. The Northern filters were probed withRB and GAPOH probes. Quantitations of the signals indicated a 1 .5-folddecrease of RB mRNA by p21S and no change in the presence ofcycloheximide.

Cell Growth & Differentiation 1709

GAPDH -

more, turning off the expression of p21ras restores the ex-pression of pRB similarly within 24 h, indicating direct and

rapid effects of ras on RB expression. These findings cor-

roborate our earlier observation that cells stably transfected

with pRB and p2Vas eventually lose pRB expression and

transform (1). Regulation of c-myc by ras has not been do-scribed before. The finding is unexpected, because c-mycappears to be required for the positive regulation of the cellcycle. In this case, the observed decrease of c-myc by ras

was directly related to high levels of pRB in the cells. This

would suggest that c-myc levels are under positive control of

pRB.The effect of p21ras on pRB was dependent on the pres-

ence of serum or growth factors. In low serum cell culture,

p2Vas lost its ability to down-regulate pRB, whereas mor-

phobogical transformation of the cells still occurred. Further-

more, the cells maintained high proliferative index, even in

low serum, indicating a general independence of the cells of

the growth factors required for cell proliferation. However, in

spite of this, growth factors were essential for p21�-medi-

ated down-regulation of pRB. All growth factors tested were

able to reconstitute the ras effect on pRB in low serum. The

cells were particularly sensitive to TGF-pl , where even 5 �M

concentration was able to restore the effect.4

Interestingly, of the growth factors tested, EGF, FGF-2,

and PDGF have all been strongly implicated in the activation

of p21’� to its GTP-bound form (reviewed in Ref. 6; Ref. 34)

and, therefore, the MAP kinase cascade pathway (reviewed

in Ref. 5; Ref. 35). The effects of TGF-f31 on p21� activation

are, however, more complex. In mink lung epithelial cells, it

either stimulates ras activation (36) or inhibits ras activation

of TGF-[31 -arrested cells stimulated to enter back into the

cycle (37). Similarly, its actions in colon carcinoma cells are

dual. In two sublines of colon carcinoma cells differing in

their growth-regulatory responses to TGF-�1 , TGF-131 acti-

yates p2Vas in conjunction with growth inhibition but not

when it acts as a mitogen (38). Expression of p21�s in epi-thelial cells generally suppresses the growth-inhibitory action

of TGF-/31 (39-41), whereas TGF-�31 stimulates the growth

of mouse fibroblasts both in the presence and absence of ras

(42). Although TGF-f31 is not known to use the MAP kinasepathway for its intracellular signal transduction, a new mem-ber of MAP kinase kinase pathway, TAK-1 , has been impli-

cated in mediation of TGF-�1 signals (43). TGF-�1 also ac-tivates the expression of components of the AP-1 complex,

which is also one of the main known targets of the MAP

kinase pathway. Interestingly, the farnesyltransferase sub-

unit a, required for p21� carboxyterminal farnesylation andthus its membrane localization and biological activity, has

been shown to associate with the TGF-�1 type I receptor.

Upon TGF-/31 type II receptor association with type I recep-

tor, farnesyltransferase is phosphorylated and released and

is presumably active to farnesylate ras (44, 45). Thus, all of

the studied growth factors have the capacity to either acti-vate ras or elicit signals mimicking the activated ras pathway

and, therefore, to activate common downstream elements.

The identification of the key molecules activated by thegrowth factors responsible forthe decrease of pRB by p21’�

and the mechanistics of the decrease appear to be of great

interest.

The contribution of p21� on cell growth and ultimately, in

many cases, cell transformation appears to occur at multiple

levels, the targets of which are only partially known. Uncon-

trolbed intracellular signaling pathways have been considered

as major constituents of cell transformation, leading to al-tered nuclear events. The capacity of p2lras to affect the

levels of proteins responsible for cellular phenotype such as

fibronectin and collagens, and proteins responsible for con-

trol of cell cycle progression such as pRB, underscore its

central importance in cell growth control pathways.

Materials and MethodsCell Culture and Transfection. Mouse NIH3T3 (ATCC CRL 1658) fibro-blasts and its derivative NIH3T3 RB4.6 stably expressing high levels of the

human retinoblastoma cONA (32) were transfected by lipofection (Lipo-

fectAMINE; Life Technologies, Inc.) with plasmids pSVlacOras andpHj3INLSneo (kindly provided by N. Denko, University of Cincinnati Med-ical Center, Cincinnati, OH) or pHf3INLSneo alone. Plasmid pSVlacOras

consists of SV4O promoter and lac-operator enhancer elements driving

expression of human c-Ha-ras(VaI-1 2) oncogene, and plasmidpH�lNLSneo contains human (3-actin promoter driving the expression oflac repressor gene and a neomycin resistance gene (31). NIH3T3 cells

transfected with pH(3lNLSneo-vector were selected with 0.6 mg/mI G41 8

Ac�c . EGF FGF-2 TGF- 61 PDGF

I P1’G � ii II II II I- + - + - + - + - +

pRB-

p2lras-

B

�-�-

- � - - =� �

GF l0%F�S 0.2%FcS EGF FGF-2 TGF-131

cck4-

IPTG l +H. +ll . +H. +11. +1

i.,,. � � � , � Ti: � :�

� � � � � � - � -�-

- .� � � � .�- -� � -_

C

C)(I)Cs

C)

0C)

V

V

0U.

I I I II I I I ______1o’YcFcs o.2�cs �c�: FGF-2 TGF. �

1710 p21S Decreases RB Expression

tflB-

G�PDH -

Fig. 6. Growth factors recon-stitute p21ras�mediated de-crease of pRB in low serum. A,Western blothng analysis ofpRB. Clone RB4.6ras6 cellswere plated in 10% FCS. Thefollowing day, the medium waschanged to medium contain-ing 0.2% FCS without (-) orwith (+) IPTG, growth factorsEGF (30 ng/mI), FGF-2 (7 ng/ml), PDGF (5 ng/mI), andTGF-�31 (250 nM) as indicated,and the cells were incubatedfor 2 days. Cell lysates (150and 250 �.tg for pRB andp21ras, respectively) were ana-lyzed by 7.5 and 12.5% SOS-PAGE for pRB and p21�s,respectively, followed byWestern blotting with the re-spective antibodies. The mi-gration of pRB and p21� areindicated to the left. B, North-em blotting analysis of RB.Clone RB4.6ras6 cells wereplated and cultured as mdi-cated above in 10% FCS or in0.2% FCS in the presence ofthe indicated growth factorsand IPTG. PoIy(A) � RNAs (3�Lg) were analyzed by Northernblotting, and the filters werehybridized with labeled cONAinserts for RB, COK4, andGAPDH as shown. C, The sig-nals were quantitated and nor-mabized against GAPOH. Thefold decreases in each caseare shown.

(Life Technologies, Inc.) for 14 days, after which clones were isolated byring-cloning. NIH3T3 and NIH3T3 RB4.6 (RB4.6) cells transfected with

pHf3INLSneo and pSVlacOras were first grown in the presence of G41 8 for14 days, after which they were transferred to 0.34% soft agar containing

10% FCS (Life Technologies, Inc.), 10% tryptose-phosphate broth, 0.6mg/mI G418, and 20 mM IPTG (Promega Corp.) in 2x OMEM and plated

on top of 0.5% agar containing 1 0% FCS, 1 0% tryptose-phosphate broth,

2x OMEM, and 10 mM IPTG. After 2 weeks, colonies with transformedmorphology were picked and expanded in the absence of IPTG. Clonesused in the experiments were NIH3T3 Iacbneo4 (3T3neo4), NIH3T3RB4.6lacOras5 (RB4.6ras5), and NIH3T3 RB4.6lacOras6 (RB4.6ras6).

NIH3T3 cells and its transfectants were maintained in OMEM containing10% newbom calf serum (Life Technologies, Inc.), and RB4.6 cells and its

transfectants were maintained in OMEM containing 10% FCS. All trans-fected clones were maintained in the presence of0.4 mg/mI G418. Growth

factors (EGF, human recombinant FGF-2, and human PDGF) were ob-tamed from Calbiochem. TGF-f31 was purified from human platelets.

Determination of Cell Doubling Times. Cells (1 0�) were plated on1 00-mm culture dishes with or without 10 m�i IPTG. At days 1 , 2, and 4after plating, the cells were trypsmnized and counted using Coulter Coun-

ter. The cell doubling time was determined as the number of hoursrequired for the cells to double during their exponential growth phase.

Cell Growth & Differentiation 1711

8. Lu, K., and Campisi, J. Ras proteins are essential and selective for theaction of insulin-like growth factor 1 late in the G1 phase of the cell cycle

Immunoblotting Analyses. Cells were washed twice with cold 25 m�Tris-HCI buffer (pH 8.0) containing 150 m� NaCI (TBS) and lysed withbuffer containing 50 mM Tris-HCI (pH 8.0), 150 m� NaCI, 1 % Triton X-100,

1 % NP4O, 100 lU/mb aprotinin, and 1 m�i phenylmethylsulfonyl fluoride for

p21�as or buffer containing 25 m� Tris-HCI (pH 8.0), 120 m� NaCI, 0.5%NP4O, 4 mM NaF, 100 p.r�i Na3VO4, 100 KIU/ml aprotinin, 1 m� phenyl-

methybsulfonyb fluoride, and 10 �zg/ml beupeptin for pRB at 4#{176}Cfor 20 mm.Lysates were collected and cleared by centrifugation for 1 5 mm at 4#{176}C,

and the supernatants were recovered. Protein concentrations were de-termined by Bradford protein assay, and 150-250 �g of lysates wereebectrophoretically resolved in 7.5, 12.5, or 9-15% gradient SOS-PAGE.

Equal loading was verified by Coomassie Brilliant Blue staining of parts ofthe gels. Proteins were transferred to Immobibon P membranes (Millipore).To detect pRB, the membranes were probed with monoclonal anti-RBantibody (PMG3-245; PharMingen) and peroxidase conjugated rabbit an-

timouse antibody (Dako). To detect p21’s-’, the membranes were probedwith anti-p21�#{176} antibody (Ab-1 ; Oncogene Science), followed by biotiny-

bated rabbit antimouse antibody (Dako) and peroxidase-conjugatedstreptavidin (Dako). The final detection step was carried out with en-

hanced chemiluminescence (ECL; Amersharn Corp.) according to the

manufacturer’s protocol.mRNA Isolation and Hybridization. PoIy(A)� mRNA was isolated

from the cells using obigo(dT) cellulose (Calbiochem), separated in 1%agarose gels containing formaldehyde, and transferred to Hybond-C extra

membrane (Amersham Corp.) in 20x SSC (3 M NaCI, 0.3 PA sodium citrate,pH 7.0). RNA was detected by probing with the indicated cONA insertlabeled with [a-321#{176}]dCTPby random priming (Ready-To-Go; Pharmacia).

The cONAs were obtained from the following sources: human RB(pSG5RB), J. DeCaprio (Dana-Faber Cancer Institute, Boston, MA); COK4

(pCMJ3cdk4), C. Sherr (St. Jude’s Childrens Hospital, Memphis, TN);cyclin 01 (pHsCVCD1 -H123), 0. Beach (Cold Spring Harbor Laboratory,

Cold Spring Harbor, NV); TIMP-2 (537-bp cDNA), U. Saariabho-Kere (Uni-versity Hospital of Helsinki, Helsinki, Finland); mouse ODC (pOOC16), 0.A. J#{228}nne(University of Helsinki); E2F-1 (pSP72-pRBAP-1), W. Kaelin

(Dana-Faber Cancer Institute); and c-myc, K. Alitalo (University of Hel-sinki). Hybridizations were carried out in 50% formamide at 42#{176}C.As a

control for RNA loading, the filters were probed with 32P-labeled GAPDHcONA, quantitated by scanning the films with Mikrotek scanner and an-

alyzed with NIH Image 1.47 program and compared to the GAPDH levels.

Acknowledgments

We thank N. Denko for plasmids pSVlacOras and pH�lNLSneo, K. Alitalofor c-myc, 0. Beach for pHsCVCO1 -H123, J. DeCaprio for pSG5RB, 0. A.

Jftnne for pOOC16, W. Kaelin for pSP72-pRBAP-1, C. Sherr for

pCMJ3cdk4, and U. Saarialho-Kere for TIMP-2. E. HOltt#{226}and J. Keski-Oja, Haartman Institute, are kindly acknowledged for critical review of the

manuscript.

References

1 . Kivinen, L, Pitk#{228}nen,K., and Laiho, M. Human retinobbastoma geneproduct prevents c-Ha-ras oncogene mediated cellular transformation of

mouse fibroblasts. Oncogene, 8: 2703-271 1 , 1993.

2. Bos, J. L ras oncogenes in human cancer: a review. Cancer Res., 49:4682-4689, 1989.

3. Medema, R. H., and Bos, J. L The robe of p21� in receptor tyrosinekinase signaling. Crit. Rev. Oncogen., 4: 615-661, 1993.

4. Hall, A. A biochemical function for Rae-at last. Science (WashingtonDC), 264: 1413-1414, 1994.

5. Waskiewicz, A. J., and Cooper, J. A. Mitogen and stress responsepathways: MAP kinase cascades and phosphatase regulation in mam-mabs and yeast. Curr. Opin. Cell Biol., 7: 798-805, 1995.

6. Pawson, T. Protein modules and signalling networks. Nature (Lond.),373: 573-580, 1995.

7_ Dobrowolski, S., Harter, M., and Stacey, 0. W. Cellular ras activity isrequired for passage through multiple points of the GO/Gi phase in

BALB/c 3T3 cells. Mob. Cell. Biob., 14: 5441-5449, 1994.

in BALB/c murine fibroblasts. Proc. NatI. Acad. Sci. USA, 89: 3889-3893,

1992.

9. Hagag, N., Halegoua, S., and Viola, M. Inhibition of growth factor-

induced differentiation of PC12 cells by microinjection of antibody to ras

p21 . Nature (Lond.), 319: 680-682, 1986.

10. Tuck, A. B., Wilson, S. M., Khokha, R., and Chambers, A. F. Different

patterns of gene expression in ras-resistant and ras-sensitive cells. J. NatI.

Cancer Inst., 83: 485-491 , 1991.

1 1 . Sistonen, L, Keski-Oja, J., Ulmanen, I., HOItt#{226},E., Wikgren, B-J., andAlitalo, K. Dose effects of transfected c-Ha-ras�2 oncogene in trans-formed cell clones. Exp. Cell Res., 168: 518-530, 1987.

12. Schalken, J. A., Ebeling, S. B., Isaacs, J. T., Treiger, B., Bussemakers,

M. J. G., de Jong, M. E. M., and Van den Ven, W. J. M. Down modulationof fibronectin messenger RNA in metastasizing cancer cells revealed by

differential hybridization analysis. Cancer Res., 48: 2042-2046, 1988.

13. Chandler, L A., Ehretsman, C. P., and Bourgeois, S. A novel mech-

anism of Ha-ras oncogene action: regulation of fibronectin mRNA bevels

by a nuclear transcriptional event. Mol. Cell. Biol., 14: 3085-3093, 1994.

14. Slack, J. L, Parker, M. I., Robinson, V. R., and Bornstein, P. Regu-lation of collagen I gene expression by ras. Mol. Cell. Biol., 12: 4714-

4723, 1992.

15. Chambers, A. F., Colella, R., Denhardt, 0. T., and Wilson, S. M.

Increased expression of cathepsins L and B, and decreased activity oftheir inhibitors, in metastatic ,as-transforrned NIH 3T3 cells. Mol. Car-

cinog., 5: 238-245, 1992.

1 6. Kirjavainen, J., Leppa, S., Hynes, N. E., and Jalkanen, M. Translational

suppression of syndecan-1 expression in Ha-ras transformed mouse

mammary epithelial cells. Mob. Biol. Cell, 4: 849-858, 1993.

1 7. Chambers, A. F., and Tuck, A. B. Ras-responsive genes and tumor

metastasis. Crit. Rev. Oncogen., 4: 95-1 14, 1993.

18. Hobtt#{228},E., Sistonen, L, and Alitabo, K. The mechanisms of omithinedecarboxylase deregulation in c-Ha-ras oncogene-transformed NIH 3T3

cells. J. Biol. Chern., 263: 4500-4507, 1988.

19. Sistonen, L, HOltt#{226},E., M#{228}kel#{226},T., Keski-Oja, J., and Alitalo, K. The

cellular response to induction of the p21 cHa�8S oncoprotein includesstimulation ofjun gene expression. EMBO J., 8: 815-822, 1989.

20. Stacey, 0. W., Watson, T., Kung, H. F., and Curran, T. Microinjectionof transforming ras protein induces c-fos expression. Mob. Cell. Biol., 7:523-527, 1987.

21 . SchOnthal, A., Herrlich, P., Rahmsdorf, H. J., and Ponta, H. Require-ment for fos gene expression in the transcriptional activation of collagen-ase by other oncogenes and phorbol esters. Cell, 54: 325-334, 1988.

22. Chin, K-V., Usda, K., Pastan, I., and Gottesman, M. M. Modulation ofactivity of the promoter of the human MDR1 gene by Ras and p53.Science (Washington DC), 255: 459-462, 1992.

23. Geiser, A. G., Kim, S-J., Roberts, A. B., and Sporn, M. B. Character-ization of the mouse transforming growth factor-pl promoter and activa-

tion by the Ha-ras oncogene. Mob. Cell. Biol., 11: 84-92, 1991.

24. Binetruy, B., Smeal, T., and Karmn, M. Ha-Ras augments c-Jun activity

and stimulates phosphorylation of its activation domain. Nature (Lond.),351: 122-127, 1991.

25. Sinn, E, Muller, W., Pattengale, P., Tepler, I., Wallace, R., and Leder,

P. Coexpression of MMTV/v-Ha-ras and MMTV/c-myc genes in trans-genic mice: synergistic action of oncogenes in vivo. Cell, 49: 465-475,

1987.

26. Lu, X., Hee Park, S., Thompson, T. C., and Lane, 0. P. ras-inducedhyperplasia occurs with mutation of p53, but activated ras and myctogether can induce carcinoma without p53 mutation. Cell, 70: 153-161,

1992.

27. Denko, N. C., Giaccia, A. J., Stringer, J. R., and Stambrook, P. J. The

human Ha-ras oncogene induces genomic instability in murine fibroblasts

within one cell cycle. Proc. NatI. Acad. Sci. USA, 91: 5124-5128, 1994.

28. Schlatter, B., and Waghorne, C. G. Persistence of Ha-ms-inducedmetastatic potential of SP1 mouse mammary tumors despite loss of the

Ha-ms shuttle vector. Proc. NatI. Acad. Sci. USA, 89: 9986-9990, 1992.

1712 p21’�#{176}Decreases RB Expression

29. Ebiyahu, 0., Michabovitz, 0., Ebiyahu, S., Pinhasi-Kimhi, 0., and Oren,M. Wild-type p53 can inhibit oncogene-mediated focus formation. Proc.NatI. Acad. Sci. USA, 86: 8763-8767, 1989.

30. Finlay, C. A., Hinds, P. W., and Levine, A. J. The p53 proto-oncogenecan act as a suppressor of transformation. Cell, 57: 1083-1093, 1989.

31 . Liu, H-S., Scrabbe, H., Villaret, 0. B., Ueberman, M. A, and Stambrook,

P. J. Control of Ha-ms-mediated mammalian cell transformation by Esche-richia coli regulatory elements. Cancer Res., 52: 983-989, 1992.

32. Pitk#{228}nen,K., Kivinen, L, DeCaprio, J. A., and Laiho, M. Expression of

the human retinobbastoma gene product in mouse fibroblasts: effects oncell proliferation and susceptibility to transformation. Exp. Cell Res., 207:

99-106, 1993.

33. MUller, H., Lukas, J., Schneider, A, Warthoe, P., Bartek, J., Eilers, M.,

and Strauss, M. Cyclmn Dl expression is regulated by the retinoblastomaprotein. Proc. NatI. Acad. Sci. USA, 91: 2945-2949, 1994.

34. Huang, J., Mohammadi, M., Rodrigues, G. A., and Schlessinger, J.Reduced activation of RAF-1 and MAP kinase by a fibroblast growth

factor receptor mutant deficient in stimulation of phosphatidylinositolhydrolysis. J. Biol. Chem., 270: 5065-5072, 1995.

35. Satoh, T., Nakafuku, M., and Kaziro, V. Function of Ras as a molecular

switch in signal transduction. J. Biol. Chem., 267: 24149-24152, 1992.

36. Mulder, K. M., and Moms, S. L Activation of p21� by transforminggrowth factor (3 in epithelial cells. J. Biol. Chem., 267: 5029-5031 , 1992.

37. Howe, P. H., Dobrowolski, S. F., Reddy, K. B., and Stacey, 0. W.

Release from G1 growth arrest by transforming growth factor �31 requires

cellular ras activity. J. Biol. Chem., 268: 21448-21452, 1993.

38. Van, Z., Winawer, S., and Friedman, E. Two different signal transduc-tion pathways can be activated by transforming growth factor �31 in

epithelial cells. J. Biol. Chem., 269: 13231-13237, 1994.

39. Filmus, J., Zhao, J., and Buick, R. N. Overexpression of H-ras

oncogene induces resistance to the growth-inhibitory action of trans-

forming growth factor beta-i (TGF-f3i) and alters the number and type

of TGF-pl receptors in rat intestinal epitheliab cells. Oncogene, 7:521-526, 1992.

40. Game, S. M., Huelsen, A., Patel, V., Donnelly, M., Veudall, W. A.,

Stone, A., Fusenig, N. E., and Prime, S. S. Progressive abrogation ofTGF-�i and EGF growth control is associated with tumour progression

in ras-transfected human keratmnocytes. Int. J. Cancer, 52: 461-470,

1992.

41 . Longstreet, M., Miller, B., and Howe, P. H. Loss of transforming

growth factor (31 (TGF-j3i)-induced growth arrest and p34cdc2 regulation

in ras-transfected epitheliab cells. Oncogene, 7: 1549-1556, 1992.

42. Benzakour, 0., Merzak, A., 000ghe, V., Pironin, M., Lawrence, 0., and

Vigier, P. Transforming growth factor-(3 stimulates mitogenically mouse

NIH3T3 flbroblasts and those cells transformed by the EJ-Hras oncogene.

Growth Factors, 6: 265-275, 1992.

43. Yarnaguchi, K., Shirakabe, K., Shibuya, H., Irie, K., Oishi, I., Ueno, N.,Taniguchi, T., Nishida, 5, and Matsumoto, K. Identification of a member

of the MAPKKK family as a potential mediator of TGF-j3 signal transduc-

tion. Science (Washington DC), 270: 2008-201 1 , 1995.

44. Kawabata, M., Imamura, T., Miyazono, K., Engeb, M. E., and Moses,

H. L Interaction of the transforming growth factor-p type I receptor with

farnesyl-protein transferase-a. J. Biol. Chem., 270: 29628-29631 , 1995.

45. Wang, T., Danielson, P. 0., Li, B-y., Shah, P. C., Kim, S. 0., and

Donahoe, P. K. The p2i� farnesyltransferase a subunit in TGF-f3 andactivin signaling. Science (Washington DC), 271: 1 120-1 122, 1996.