Epidermal Growth Factor Induces Cyclin D1 in aHuman Prostate Cancer Cell Line

Jaime E. Perry, Michael E. Grossmann, and Donald J. Tindall*

Department of Urology Research, Mayo Clinic and Foundation, Rochester, Minnesota

BACKGROUND. The human prostate carcinoma cell line, LNCaP, proliferates under stimu-lation by a limited number of mitogenic signals, which include members of the growth factorand steroid hormone families. Androgens and epidermal growth factor (EGF) are among theLNCaP cell mitogens. We tested the hypothesis that these mitogens stimulate LNCaP cellproliferation at least in part through the induction of cyclin D1, a protein requisite for cell cycleprogression, which is expressed in the G1 phase of the cell cycle.METHODS. LNCaP cells were grown in serum-free medium with 10 ng/ml or 100 ng/mlEGF, 0.1 nM or 1.0 nM mibolerone (a potent androgen agonist), or vehicle (distilled water or0.01% ethanol). Expression of cyclin D, mRNA, and protein were assessed by Northern andWestern blot analyses. Transcription regulation was assessed by nuclear runoff assay.RESULTS. Western analyses demonstrated that EGF stimulated cyclin D1 protein expression4-fold over 12 hr. Northern analyses showed a 4-fold increase in mRNA expression, peakingwithin 4 hr of EGF stimulation. There were no effects on cyclin D1 protein or mRNA expres-sion with mibolerone treatments. We further explored the mechanism of cyclin D1 induction.LNCaP cells stimulated for 1 hr with EGF demonstrated a 2-fold increase in cyclin D1 mes-sage, as assayed by nuclear runoff transcription assay. In addition, we demonstrated theinvolvement of the protein kinase C pathway in mediating the EGF induction of cyclin D1.CONCLUSIONS. We conclude that one of the mechanisms by which growth factors such asEGF may stimulate prostate cell proliferation is through the direct induction of cyclin pro-teins, which are necessary for entry of cells into mitosis. Prostate 35:117–124, 1998.© 1998 Wiley-Liss, Inc.

KEY WORDS: cell cycle; cyclin; prostatic cancer; EGF; androgen

INTRODUCTION

In the normal prostate gland, growth factors de-rived from the stromal compartment are thought tostimulate prostate epithelial proliferation [1]. Some ofthe stromally-derived growth factors are secreted un-der the stimulation of androgens [2]. Therefore, theepithelial cells may receive an indirect androgenicstimulus mediated through these growth factors. It isthis interaction between androgens and growth fac-tors which is likely to maintain normal prostate func-tion. Androgens appear to have different functionswithin the two basic compartments of the prostategland [3]. Prostatic epithelial cells are stimulated tosecrete prostate-specific products by the interaction ofandrogens with the androgen receptor [4–6]. The de-velopment of the prostate gland, however, requires

the action of androgens on the stromal compartment(urogenital mesenchyme) [7,8]. Within the prostaticstromal cells, androgens, via the androgen receptor,induce growth factors such as keratinocyte growthfactor (KGF) [9,10]. Keratinocyte growth factor isknown to act on the prostatic epithelium to stimulateproliferation [11].

Contract grant sponsor: NIH; Contract grant numbers: DK47592,HD09140.Current address for Jaime E. Perry is the Department of ClinicalSciences, University of Wisconsin-La Crosse, 1725 State Street, LaCrosse, WI 54601.*Correspondence to: Donald J. Tindall, Ph.D., Department of Urol-ogy Research, The Mayo Clinic, Guggenheim 1737, 200 First St. SW,Rochester, MN 55905. E-mail: tindall@mayo.eduReceived 25 July 1997; Accepted 15 October 1997

The Prostate 35:117–124 (1998)

© 1998 Wiley-Liss, Inc.

Prostatic cancers appear to overcome the need forexogenous or indirect signals. In such pathologicstates, the functional interaction between androgensand growth factors may be interrupted or altered, re-sulting in abnormal prostatic epithelial cell growth[12]. One model for the proliferative effects of andro-gens on prostatic cancer cells is the LNCaP cell line,which was derived from a human prostate cancer me-tastasis to the lymph node [13]. LNCaP cells respondto androgens [14] and growth factors such as epider-mal growth factor (EGF) [15–18], b-fibroblast growthfactor (b-FGF) [19], and transforming growth factor-alpha (TGF-a) [20,21]. Evidence that androgens andgrowth factors interact in their stimulation of LNCaPproliferation was derived from observations that theexpression of amphiregulin [22], TGF-a [23], EGF[24,25], and a-FGF [26] are modulated by androgens.The proliferative response to androgens in prostatecancers is mediated through a functional androgen re-ceptor [27], while that to growth factors is thought tobe through second-messenger systems [28,29].

The earliest committed step in the initiation of thecell cycle leading to cell proliferation is the formationof a holoenzyme complex in the G1 phase of the cycle,composed of cyclin D [30], an associated cyclin-dependent kinase, a cyclin-dependent kinase inhibi-tory protein, and proliferating cell nuclear antigen(PCNA) [31], which is required for DNA synthesis[32,33]. We have demonstrated that androgens mayindirectly stimulate cell growth by influencing the ex-pression of cell cycle-associated proteins [34], albeit byposttranscriptional mechanisms. The goal of the ex-periments here was to test the hypothesis that growthfactors induce cellular proliferation by more directmeans. We describe experiments in which we usedgrowth factors to stimulate proliferative responses inLNCaP cells. The expression of cyclin D1 is the endpoint we have measured, as this protein is expressedearly in the cell cycle and is considered to be a key stepin committing cells to division.

MATERIALS AND METHODS

Cells and Culture Conditions

LNCaP cells (American Type Culture Collection,Rockville, MD) were maintained in RPMI-1640 withglutamine and sodium bicarbonate. The medium wassupplemented with the antibiotics penicillin andstreptomycin, which were added to a final concentra-tion of 50 mg/ml and fetal calf serum (Biofluids, Rock-ville, MD) to 5%. Cells were used between passages25–40 after attaining about 70% confluence. Cells wereincubated in serum-free medium supplemented with0.1% BSA for 2 days prior to treatment. Treatments

were added with a change of serum-free medium con-taining 0.1% BSA. Treatments included EGF, mibo-lerone, thapsigargin, bisindolylmaleimide, cyclohexi-mide, and TPA (all from Sigma Chemical Co., St.Louis, MO).

Flow Cytometric Analysis

Flow cytometry was performed on LNCaP cells asfollows: cells were collected, washed, and dispersed inPBS. Dispersion of the cells into a single-cell suspen-sion was accomplished by drawing the cells in PBS 10times through a 26-gauge, blunt-ended needle. An ali-quot of the cells was set aside for cell counts, using aCoulter counter (Coulter Electronics, Hialeah, FL). Therest of the cells were fixed in 1% paraformaldehyde inPBS with 30 mg/ml lysolecithin (ICN Biochemicals,Cleveland, OH) added to permeabilize the cells. Thecells were washed in buffer A (3% BSA in PBS with0.05% Tween-20) and incubated overnight with anti-cyclin D1 (Upstate Biotechnology, Lake Placid, NY)diluted 1:200 in buffer A. After washing in buffer A,the cells were incubated for 1 hr with an FITC-conjugated second antibody (GAR-FITC; 1:100, Orga-non Technika/Cappel, West Chester, PA) diluted inbuffer A, washed, and treated with RNase (1 mg/ml;Sigma Chemical Co.) for 15 min at 37°C. Cells weresuspended in buffer A with 0.1 mg/ml final concen-tration of propidium iodide, and analyzed on a FAC-Star Plus flow cytometer (Becton Dickinson, Sun-nyvale, CA) at a wavelength of 488 nm with a 620-nmpass filter.

Western Blot Analysis

Western blot analysis was performed on cellscounted and homogenized in homogenization buffer(10 mM Tris-Cl, pH 7.4, 0.25 M sucrose, 1 mM EDTA,5 mg/ml aprotinin, 5 mg/ml pepstatin, and 5 mg/mlleupeptin). Proteins were separated in 10% polyacryl-amide gels by SDS-PAGE and transferred to nitrocel-lulose. The blots were blocked and exposed overnightto either anti-cyclin D1 (1:2,000), anti-cdk-4, anti-cdk-5,anti-p21, anti-p16, or antibodies which had been pre-absorbed with corresponding peptide. All antiseraand peptides were obtained from Upstate Biotechnol-ogy. Specific bands were visualized by enhanced che-miluminescence (ECL method, Amersham Corpora-tion, Arlington Heights, IL). Blots were scanned on aScanJet IIcx (Hewlett-Packard, Palo Alto, CA) in orderto quantitate band densities.

Northern Blot Analysis

Northern blot analysis was performed on totalRNA isolated from LNCaP cells lysed in 3 M lithium

118 Perry et al.

chloride and 6 M urea. The DNA was sheared with anUltra-Turrax polytron (24,000 rpm; Janke and Kunkel,Staufen, Germany) and incubated overnight at 4°C.RNA was separated by centrifugation at 107,000g. Thepellet containing the RNA was dissolved in 0.1% SDSand 0.2 mM EDTA and extracted with phenol/chloroform (1:1) and chloroform. The RNA was pre-cipitated with 0.16 M sodium acetate in ethanol. Puri-fied, total RNA (20 mg/lane) was separated in a 1%agarose gel which contained 2% formaldehyde, using1 × MOPS buffer (0.02 M 3-N-morpholino-propansulfonic acid, 0.005 M sodium acetate, and0.001 M EDTA) and transferred to Hybond-N blottingmembrane (Amersham Corporation). Blots werecrosslinked using an automatic UV Stratalinker(Stratagene, La Jolla, CA), prehybridized for 1 hr at42°C in prehybridization solution (45% deionizedformamide), 5 × SSC, 0.5% SDS, 10% Denhardt’s solu-tion, 10 mM phosphate buffer, 15% dextran sulfate,and 100 mg/ml herring sperm DNA. Randomly [32P]-labeled probes prepared from cDNAs to cyclin D1 orGAPDH were prepared by boiling 125 ng of the cDNAin 12 ml water for 3 min before the addition of 4.5 mloligonucleotide labeling mixture (0.2 mM dCTP, 0.2mM dTTP, 0.2 mM dGTP, 50 mM Tris-HCl, pH 7.8, 5mM MgCl2, and 10 mM 2-mercaptoethanol), 1 unitKlenow, and 3 ml a-dATP-32P. Labeled cDNA was de-natured by boiling, 1 × 106–7 cpm/ml was added to theblots in prehybridization solution and incubated over-night at 42°C in a shaking waterbath. Blots werewashed twice in 1 × SSC containing 0.25% SDS at 42°Cand once at 55°C before being exposed. Radioactivityin each lane was quantitated using a PhosphorImager(Molecular Dymanics, Sunnyvale, CA). Cyclin D1

mRNA data were normalized to GAPDH.For mRNA half-life experiments, cells were treated

for 4 hr with or without 100 ng/ml EGF prior to ad-dition of actinomycin D (final concentration of 10 mg/ml), or a-amanitin (final concentration of 1 mM, suffi-cient to inhibit both RNA polymerases I and II), toprevent new RNA synthesis. Cells were collected atvarious time points after the addition of the RNApolymerase inhibitor (0–30 hr) and prepared forNorthern blot analysis as above. Gels were photo-graphed prior to blotting; the 28S ribosomal RNAbands were used to normalize loading of lanes on theprobed blots.

Nuclear Runoff Transcription Assay

The cDNA slot blots were prepared by denaturing5 mg cyclin D1, GAPDH, or pBL DNA with a 1/9volume of 10 N NaOH for 30 min at 20°C in a total

volume of 100 ml. The reaction was diluted 1:3 with 6× SSC and applied to Hybond-N blotting membrane(Amersham). Slot blots were rinsed and crosslinkedwith the aid of a UV Stratalinker (Stratagene). Blotswere prehybridized for 4 hr in prehybridization solu-tion (10 mM TES, pH 7.5, 10 mM EDTA, 0.2% SDS, 0.3M NaCl, 250 mg/ml tRNA, and 1 × Denhardt’s solu-tion).

Nuclei from cells treated for 1 hr with or without10, 100, or 500 ng/ml EGF were prepared as follows.Approximately 5 × 106 cells per treatment were col-lected for each labeling reaction. Cells were washedwith ice-cold PBS, and then with cold NP-40 buffer (10mM Tris-HCl, pH 7.4, 10 mM NaCl, and 3 mM MgCl2).Cells were lysed in NP-40 lysis buffer (NP-40 bufferwith 0.5% Nonidet NP-40) for 20 min, with periodicvigorous shaking to loosen cells. Cells were centri-fuged (5 min, 500g, 4°C) and washed once with NP-40lysis buffer. Cells were resuspended in sucrose bufferA (0.32 M sucrose, 3 mM CaCl2, 2 mM Mg acetate, 0.1mM EDTA, 10 mM Tris-HCl, pH 8.0, 5 mM DTT, 0.1%Triton, 0.6 mM PMSF, and 0.6 mM bacitracin) and leftfor 5 min on ice before being homogenized on ice witha Dounce (VWR Scientific, West Chester, PA) homog-enizer (30 strokes). The homogenate was then dilutedwith two volumes of sucrose buffer B (2.05 M sucrose,5 mM Mg acetate, 0.1 mM EDTA, 10 mM Tris-HCl, pH8.0, 1.5 mM DTT, 0.2 mM PMSF, and 0.2 mM bacitra-cin) and layered over a 1/4 volume cushion of sucrosebuffer B in ultracentrifuge tubes. The homogenate wasspun for 25 min at 100g (4°C). Pellets were resus-pended in 400 ml reaction buffer (10 mM Tris-HCl, pH8.0, 5 mM MgCl2, 0.3 M KCl, 1 mM ATP, 1 mM CTP,1 mM UTP, and 5 mM DTT) to which 100 mCi a-32P-GTP was immediately added. Reaction mixtures wereincubated for 30 min at 30°C with periodic gentleshaking. Labeled RNA probes were isolated usingTRIZOL LS reagent (Gibco-BRL, Gaithersburg, MD).Two volumes of TRIZOL LS reagent were added tothe samples, followed by a 5-min incubation. The sam-ples were vigorously vortexed for 1 min each after theaddition of 200 ml chloroform, then left for 15 min atroom temperature before microfuging for 15 min at 4°C.The upper (aqueous) phase was retrieved and the RNAwas precipitated by the addition of 0.5 ml isopropylalcohol. The samples were incubated overnight at−20°C and the RNA was pelleted by microfugation for10 min at 4°C. The pellets were washed in 70% etha-nol. The pellets were resuspended in prehybridizationsolution (see above) in a 55–60°C waterbath. Probeswere denatured by boiling and added to strips of slotblot. Incubation continued for 72 hr at 65°C. Stripswere washed in a gradient of SSC from 6× to 2× beforeexposure. Radioactivity in each slot was quantitatedusing a PhosphorImager (Molecular Dymanics).

Regulation of Cyclin D1 by EGF 119

RESULTS

We tested the hypothesis that EGF and mibolerone(a nonmetabolizable, synthetic androgen agonist)stimulate proliferation of LNCaP cells by increasingthe expression of cyclin D1. Figure 1 demonstrates theeffects of EGF on the number of cyclin D1-immunopositive cells as measured by flow cytometricanalysis. The number of LNCaP cells staining positivefor cyclin D1 was significantly increased in cellstreated with EGF at both 2 and 4 days following theinitiation of treatment. There was no significant in-crease in cyclin D1-immunopositive cells at either timepoint with mibolerone treatment.

The hypothesis was further analyzed by Westernblot analysis of cyclin D1 immunoreactivity in LNCaPcell homogenates in which the flow cytometry datawere corroborated. Figure 2 shows immunoblots ofLNCaP cell homogenates from cells treated with EGFfor various times over the course of 24 hr. Cyclin D1expression was unchanged in the mibolerone-treatedcells (data not shown) and in the untreated cells, whileEGF induced a fourfold increase in cyclin D1 immu-noexpression, which peaked 18 hr following stimula-tion.

Western blot analyses were also performed to de-termine the effects of EGF on the expression of other

proteins (cdk-4, cdk-5, p16, p21, and cyclin E) whichare known to be expressed in the G1 phase of the cellcycle. As seen in Figure 3, cdk-4 expression was in-duced by treatment of LNCaP cells with EGF. Theincrease in cdk-4 immunoexpression was not visibleuntil 12 hr following treatment, and peaked with atwofold increase in cdk-4 expression at 18 hr. The ef-fect of EGF on the expression of cdk-5, the cyclin-dependent kinase inhibitors p16 and p21, and cyclin Ewere also tested using Western blot analysis. Theseproteins were not affected by either treatment within24 hr (data not shown).

In order to begin investigating the mechanism bywhich EGF was acting to induce the expression of cy-clin D1, we first chose to look at the effect of EGF onsteady-state levels of cyclin D1 mRNA in LNCaP cells.Figure 4 demonstrates a twofold increase in steady-state levels of cyclin D1 mRNA in LNCaP cells follow-

Fig. 1. Flow cytometric analysis of the EGF induction of cyclinD1 in LNCaP cells. LNCaP cells were treated for 0, 2, or 4 dayswith 0.1 nM mibolerone, 100 ng/ml EGF, or vehicle. Cells werecollected, dispersed, and fixed in 1% formaldehyde with 30 µg/mllysolecithin before being incubated overnight with anti-cyclin D1

antiserum at 1:200 with or without the addition of a specific block-ing peptide. Cells were incubated with an FITC-conjugated secondantibody, treated with 1 mg/ml RNase, and suspended in 0.1 mg/mlpropidium iodide. The number of LNCaP cells staining positive forcyclin D1 was significantly increased at both 2 and 4 days with EGFtreatment (ANOVA; P ø 0.05; n = 6). Mibolerone had no signifi-cant effect at any time point tested.

Fig. 2. Western blot analysis of cyclin D1 immunoreactivity inLNCaP cells with EGF treatment. LNCaP cells were treated with100 ng/ml EGF for 0–24 hr. Cells were collected and homog-enized. Ten micrograms of protein from each sample were loadedinto each lane of a 10% polyacrylamide gel and separated by SDS-PAGE. Proteins were transferred to nitrocellulose, and specificbands were visualized after overnight exposure to anti-cyclin D1

antiserum (diluted 1:2,000), using enhanced chemiluminescence.Blots were scanned and band densities quantified.

120 Perry et al.

ing stimulation with EGF. Cyclin D1 message peaked6 hr after stimulation with EGF. This experiment wasalso performed on LNCaP cells stimulated with mibo-lerone. No effect of mibolerone was observed on thesteady-state level of cyclin D1 mRNA (data notshown). Cyclin D2 message was not detectable byNorthern blot analysis in LNCaP cells (data notshown).

We next focused on establishing the mechanism ofEGF induction of cyclin. The possibility that EGF actsto increase the half-life of cyclin D1 message wastested. LNCaP cells were treated with either actino-mycin D or a-amanitin to inhibit the synthesis of newRNA after cells had been stimulated with EGF or ve-hicle. In Figure 5, the half-life of the cyclin D1 messagewith and without EGF treatment can be seen. Afternormalization of the mRNA for loading using 28S ri-bosomal RNA, the half-life of the message from EGF-treated and untreated cells was calculated by regres-sion analysis. The half-life of the cyclin D1 mRNAfrom control cells was calculated to be 20.1 hr, and thatfor the EGF-treated cells was 20.3 hr. These valueswere not significantly different.

Nuclear run-on transcription assays were used todetermine if the EGF-mediated increase in cyclin D1expression was due to an increase in transcription ofthe gene. A representative example of our findings isshown in Figure 6, which demonstrates a twofold in-crease in cyclin D1 transcription following 1 hr of EGFtreatment. Cycloheximide was able to block the EGF-mediated induction of cyclin D1 message as measuredby Northern blot analysis (Fig. 7), suggesting that thetranscriptional effect may be a secondary response tothe growth factor.

The hypothesis that EGF mediates its effects on the

expression of cyclin D1 via the protein kinase C path-way was tested using Northern blot analysis. If theprotein kinase C pathway is involved in transducingthe EGF signal for proliferation, the phorbol ester TPAshould induce the expression of cyclin D1, and theeffect of EGF should be blocked with the specific pro-tein kinase C inhibitor bisindolylmaleimide (BIM)[35]. The induction of cyclin D1 with TPA and a doseresponse to EGF treatment are shown in Figure 8.Phorbol ester stimulated a fourfold increase in cyclinD1 expression, while maximal response to EGF wasachieved with 50 ng/ml. In Figure 9, the specific pro-tein kinase C inhibitor BIM is shown to inhibit theEGF-stimulated increase in cyclin D1 mRNA.

Further evidence of a role for calcium in the induc-tion of cyclin D1 by EGF is shown in Figure 8. Thecalcium ATPase inhibitor, thapsigargin, inhibited theEGF-stimulated increase in cyclin D1 expression.Thapsigargin acts to inhibit the release of calcium

Fig. 3. Western blot analysis of cdk-4 immunoreactivity in LN-CaP cells with EGF treatment. LNCaP cells were treated with 100ng/ml EGF for 0–24 hr. Cells were collected and homogenized.Ten micrograms of protein from each sample were loaded intoeach lane of a 10% polyacrylamide gel and separated by SDS-PAGE.Proteins were transferred to nitrocellulose, and specific bandswere visualized after overnight exposure to anti-cdk-4 antiserum(diluted 1:2,000), using enhanced chemiluminescence. Blots werescanned and band densities quantified.

Fig. 4. Northern blot analysis of cyclin D1 steady-state mRNAafter EGF treatment. LNCaP cells treated for 0–12 hr with 100ng/ml EGF were prepared for Northern blot analysis. Total RNA wasisolated using the lithium chloride method and separated in adenaturing agarose gel before transfer to Hybond-N blottingmembrane. Each lane contains 5 µg RNA. Randomly-labeled cyclinD1 and GAPDH cDNA probes were used to visualize the mRNAbands. Arrows indicate the position of the 4.5-kb cyclin D1 mRNAband. Cyclin D1 data were normalized to GAPDH.

Regulation of Cyclin D1 by EGF 121

from stores in the endoplasmic reticulum. This leadsto a failure of the membrane calcium pumps, causingcalcium to flood the cells. Although thapsigargin isknown to induce apoptosis with long-term treatment[36], the relatively short treatment in this experiment(4 hr) did not have any visible effect on cell viability.

DISCUSSION

We have demonstrated that EGF directly influencescellular proliferation in the LNCaP model system by

Fig. 6. Nuclear runoff transcription assay of LNCaP cell nucleiafter treatment with or without EGF. LNCaP cell nuclei wereisolated after a 1-hr treatment in the presence or absence of 100ng/ml EGF. 32P-labeled GTP and CTP were added to the nuclei andincubated for 30 min before RNA was isolated from the cells usingthe Trizol reagent. RNA was then used as a probe on slot blotscontaining denatured cDNA probes of cyclin D1, GAPDH, or vec-tor. The radioactivity in each lane was quantified. Cyclin D1 tran-scripts were normalized to GAPDH after subtraction of the vec-tor. This experiment was repeated three times.

Fig. 7. Summary of data from Northern blot analysis of LNCaPcells treated with cycloheximide in combination with EGF or TPA.LNCaP cells were treated with either cycloheximide (10 µg/ml),EGF (100 ng/ml), or TPA (1 nM), or pretreated for 30 min withcycloheximide, before the addition of EGF or TPA for 4 hr. RNAwas isolated and prepared for Northern blot analysis of cyclin D1

and GAPDH. The fold induction of cyclin D1 mRNA is presentedgraphically under each of the treatment conditions.

Fig. 5. Northern blot analysis of cyclin D1 half-life in the pres-ence or absence of EGF. LNCaP cells treated for 4 hr in thepresence or absence of 100 ng/ml EGF were subjected to treat-ment with actinomycin D at a final concentration of 10 µg/ml. Cellswere collected at time points after addition of the mRNA synthe-sis inhibitor and prepared for Northern blot analysis. The 28Sribosomal band served as a loading control.

Fig. 8. Summary of data from Northern blot analysis of LNCaPcells treated with increasing doses of EGF or TPA with or withoutthapsigargin. LNCaP cells were treated with doses of EGF (0–100ng/ml) or TPA (1 nM) in the presence or absence of 10 nM thap-sigargin for 4 hr. RNA was isolated and prepared for Northernblot analysis of cyclin D1 and GAPDH. The fold induction of cyclinD1 mRNA is presented graphically under each of the treatmentconditions.

122 Perry et al.

transcriptional regulation of cyclin D1. We showedthat EGF induces cyclin D1, which is part of the com-plex expressed in the G1 phase of the cell cycle andconsidered to be the commitment step for entry into Sphase [30]. The mechanism of this induction appearsto be at the transcriptional level, although posttran-scriptional mechanisms may also be involved. Datawere also presented which suggest that the mecha-nism by which cyclin D1 is induced may involve theprotein kinase C pathway, possibly via alterations inintracellular calcium levels.

That cyclin expression is induced by growth factorshas been documented in other cell types. A variety ofgrowth factors induce the expression of both cyclin Dprotein and mRNA in fibroblast cell lines, with timecourses similar to those described here [37,38]. A simi-lar mechanism of cell growth stimulation has beendemonstrated in mammary carcinoma cells with insu-lin, again with time courses similar to our findings[39].

The half-life of the cyclin D1 message in the LNCaPcells appears to be considerably longer than that re-ported in other cells. The typical half-life of a cyclinmessage is on the order of less than 1 hr [40], yet ourdata indicate a half-life of several hours. One theorysuggests that overexpression of cyclin proteins maycontribute to cancer progression. It has been shownthat within the MDA MB-453 breast cancer cell line,the cyclin D1 gene is truncated, effectively extendingthe half-life of the cyclin message within the cells [40].While we do not see any evidence of truncated mes-sage (we routinely observed the full-length, 4.5-kb cy-clin D1 transcript), the possibility exists that in theLNCaP cell line, additional mutation(s) may have oc-

curred which extend the half-life of the cyclin mes-sage.

We described a modest transcriptional induction(twofold) of cyclin D1 with EGF, but also demon-strated that pretreatment of the cells with cyclohexi-mide blocks the increase in steady-state mRNA in-duced by treatment of the cells with either EGF orTPA. The transcriptional induction of cyclin D1 mayrely on the presence of labile factors within the nucleiof the cells, e.g., the early growth response factor Mycwhich has been shown to be linked with the expres-sion of cyclin D1 [28]. Pretreatment of the cells withcycloheximide may block the production of such fac-tors and therefore prevent the growth factor-inducedinduction of the cyclin. This is in contrast to the find-ings of Sewing et al. [37], who were not able to blockcyclin D1 steady-state mRNA expression in a humandiploid fibroblast cell line with cycloheximide.

Our results are consistent with the findings of oth-ers in the involvement of calcium as an intermediate inthe growth factor induction of cyclin D1. The ability ofTPA to stimulate cyclin D1 mRNA expression has beendemonstrated in fibroblasts [37], corroborating the hy-pothesis that growth factors may induce cyclinsthrough the protein kinase C pathway. This pathwayis further implicated in the induction of cyclin D1 inthe prostate by our demonstration that a specific pro-tein kinase C inhibitor is able to abolish EGF induc-tion. Regulation of cyclin D1 expression in cells mayalso involve changes in intracellular calcium levels, orin calcium modulatory second messengers, perhapsby changing the phosphorylation state of the cyclin D1molecule [29].

ACKNOWLEDGMENTS

The authors thank Dr. Charles Sherr (Departmentof Tumor Biology, Howard Hughes Medical Center,St. Jude Children’s Research Hospital, Memphis,TN) for his generous gift of the cDNAs for cyclins D1and D2.

REFERENCES

1. Cunha GR, Alarid ET, Turner T, Donjacour AA, Boutin EL,Foster BA: Normal and abnormal development of the male uro-genital tract: Role of androgens, mesenchymal-epithelial inter-actions and growth factors. J Androl 1992;13:465–475.

2. Wilson JD: Syndrome of androgen resistance. Biol Reprod 1992;46:168–173.

3. Perry JE, Tindall DJ: Androgen regulation of gene expression inprostate cancer. In Pasqualini JR, Katzenellenbogen BS (eds):‘‘Molecular and Cellular Endocrinology, Volume I, HormoneDependent Cancer,’’ New York: Marcel Dekker, 1996:343–356.

4. Young CY-F, Montgomery BT, Andrews PE, Qiu S, Bilhartz DL,Tindall DJ: Hormonal regulation of prostate-specific antigenmessenger RNA in human prostatic adenocarcinoma cell lineLNCaP. Cancer Res 1991;50:3748–3752.

Fig. 9. Summary of data from Northern blot analysis of LNCaPcells treated with the protein kinase C inhibitor BIM. LNCaP cellswere treated with 100 ng/ml EGF, 10 mM BIM, both, or neither for0–6 hr. RNA was isolated and prepared from cells collected at 0,1, 3, or 6 hr of treatment for Northern blot analysis of cyclin D1

and GAPDH. The fold induction of cyclin D1 mRNA is presentedgraphically under each of the treatment conditions.

Regulation of Cyclin D1 by EGF 123

5. Young CY-F, Andrews PE, Montgomery BT, Tindall DJ: Tissue-specific and hormonal regulation of human prostate-specificglandular kallikrein. Biochemistry 1992;31:818–824.

6. Murtha P, Tindall DJ, Young CY-F: Androgen induction of ahuman prostate-specific kallikrein, hKLK2: Characterization ofan androgen response element in the 58 promoter region of thegene. Biochemistry 1993;32:6459–6464.

7. Cunha GR: Androgenic effects upon prostatic epithelium aremediated via trophic influences from stroma. In Kimball FA,Buhle AE, Carter DB (eds): ‘‘New Approaches to the Study ofBenign Prostatic Hyperplasia,’’ New York: A.R. Liss, 1984:81–102.

8. Cunha GR: Mesenchymal-epithelial interactions during andro-gen-induced development of the prostate. Prog Clin Biol Res1985;171:15–24.

9. Yan G, Fukabori Y, Nikolaropoulos S, Wang F, McKeehan WL:Herparin-binding keratinocyte growth factor is a candidate stro-mal to epithelial andromedin. Mol Endocrinol 1992;6:2123–2128.

10. Fukabori Y, Yan G, Yamanaka H, McKeehan WL: Rapid induc-tion of keratinocyte growth factor (FGF-7) and b-actin after ex-posure of prostate stromal cells to androgen. In Vitro Cell DevBiol Anim 1994;30:745–746.

11. Cunha GR: Growth factors as mediators of androgen actionduring male urogenital development. Prostate [Suppl 6]1996:22–25.

12. MacDonald A, Habib FK: Divergent responses to epidermalgrowth factor in hormone sensitive and insensitive human pros-tate cancer cell lines. Br J Cancer 1992;65:177–182.

13. Horoszewicz JS, Leong SS, Chu TM, Wajsman ZL, Friedman M,Papsidero L, Kim U, Chai LS, Kakati S, Arya SK, Sandberg AA:The LNCaP cell line–A new model for studies on human pros-tatic carcinoma. In Murphy GD (ed): ‘‘Models for Prostate Can-cer,’’ New York: A.R. Liss, 1980:115–132.

14. Horoszewicz JS, Leong SS, Kawinski E, Karr JP, Rosenthal H,Chu TM, Mirand EA, Murphy GP: LNCaP model of humanprostatic carcinoma. Cancer Res 1983;43:1809–1818.

15. Knabbe C, Kellner U, Schmahl M, Voight KD: Growth factors inhuman prostate cancer cells: Implications for improved treat-ment of prostate cancer. J Steroid Biochem Mol Biol 1991;40:185–192.

16. Henttu P, Vihko P: Growth factor regulation of gene expressionin human prostatic carcinoma cell line LNCaP. Cancer Res 1993;53:1051–1058.

17. Shuurmans ALG, Bolt J, Mulder E: Androgens stimulate bothgrowth rate and epidermal growth factor receptor activity in thehuman prostate tumor cell LNCaP. Prostate 1988;12:55–63.

18. Eaton CL, Davies P, Harper M, France T, Rushmere N, GriffithsK: Steroids and the prostate. J Steroid Biochem Mol Biol 1986;40:175–183.

19. Nakamoto T, Chang CS, Li AK, Chodak GW: Basic fibroblastgrowth factor in human prostate cancer cells. Cancer Res 1992;52:571–577.

20. Kim JH, Sherwood ER, Sutkowski DM, Lee C, Kozlowski JM:Inhibition of prostatic tumor cell proliferation by suramin: Al-terations in TGF alpha-mediated autocrine growth regulationand cell cycle distribution. J Urol 1991;146:171–176.

21. Wilding G: Transforming growth factors in human prostate can-cer. In Karr JP (ed): ‘‘Molecular and Cellular Biology of theProstate.’’ New York: Plenum Press, 1991:185–200.

22. Sehgal I, Bailey J, Hitzemann K, Pittelkow M, Maihle NJ: Epi-dermal-growth factor receptor-dependent stimulation of amphi-regulin expression in androgen-stimulated human prostate can-cer cells. Mol Biol Cell 1994;5:339–347.

23. Liu XH, Wiley HS, Meikle AW: Androgens regulate prolifera-tion of human prostate cancer cells in culture by increasingtransforming growth factor-alpha (TGF-alpha) and epidermalgrowth factor (EGF)/TGF-alpha receptor. J Clin EndocrinolMetab 1993;77:1472–1478.

24. Sciarra F: Sex steroids and epidermal growth factor in benignprostatic hyperplasia (BPH). Ann NY Acad Sci 1995;761:66–78.

25. Schuurmans ALG, Bolt J, Voorhorst M, Blankenstein RA, Mul-der E: Regulation of growth and epidermal growth factor re-ceptor levels of LNCaP prostatic tumor cells by different ste-roids. Int J Cancer 1988;42:917–922.

26. Harris SE, Hall JA, Rong ZX, Chui I-M, Harris M: Structure andfunction of the acidic fibroblast growth factor I gene in the an-drogen-dependent prostate cancer cell line LNCaP and hamsterDDT1 cells. J Cell Biochem 1992;16:87.

27. Takahashi H, Furusato M, Allsbrook WC Jr, Nishii H, Wakui S,Barrett C, Boyd J: Prevalence of androgen receptor gene muta-tions in latent prostatic carcinomas from Japanese men. CancerRes 1995;55:1621–1624.

28. Daksis JI, Lu RY, Facchini LM, Marhin WW, Penn LJZ: Mycinduces cyclin D1 expression in the absence of de novo proteinsynthesis and links mitogen-stimulated signal transduction tothe cell cycle. Oncogene 1994;9:3635–3645.

29. Sewing A, Muller R: Protein kinase A phosphorylates cyclin D1at three distinct sites within the cyclin box and at the C-terminus. Oncogene 1994;9:2733–2736.

30. Sherr CJ: D-type cyclins. Trends Biochem Sci 1995;20:187–190.31. Xiong Y, Zhang H, Beach D: D-type cyclins associate with mul-

tiple protein kinases and the DNA replication and repair factorPCNA. Cell 1992;71:505–514.

32. Prelich G, Tan CK, Kostura M, Mathews MB, So AG, DowneyKM, Stillman B: Functional identity of proliferating cell nuclearantigen and a DNA polymerase-d auxiliary protein. Nature1987;326:517–520.

33. Prelich G, Kostura M, Marshak DR, Mathews MB, Stillman B:The cell cycle regulated proliferating cell nuclear antigen is re-quired for SV40 DNA replication in vivo. Nature 1987;326:471–475.

34. Perry JE, Tindall DJ: Androgens regulate the expression of pro-liferating cell nuclear antigen posttranscriptionally in the hu-man prostate cancer cell line, LNCaP. Cancer Res 1996;56:1539–1544.

35. Toullec D, Pianetti P, Coste H, Bellvergue P, Grand-Perret T,Ajakane M, Baudet V, Boissin P, Boursier E, Loriolle F, DuhamelL, Charon D, Kirilovsky J: The bisindolylmaleimide GF 109203Xis a potent and selective inhibitor of protein kinase C. J BiolChem 1991;266:15771–15781.

36. Gong Y, Blok LJ, Perry JE, Lindzey JK, Tindall DJ: Calciumregulation of androgen receptor expression in the human pros-tate cancer cell line LNCaP. Endocrinology 1995;136:2172–2178.

37. Sewing A, Burger C, Brusselbach S, Schalk C, Lucibello FC,Muller R: Human cyclin D1 encodes a labile nuclear proteinwhose synthesis is directly induced by growth factors and sup-pressed by cyclic AMP. J Cell Sci 1993;104:545–554.

38. Winston JT, Pledger WJ: Growth factor regulation of cyclin D1mRNA expression through protein synthesis-dependent and in-dependent mechanisms. Mol Biol Cell 1993;4:1133–1144.

39. Musgrove EA, Hamilton JA, Lee CSL, Sweeney KJE, WattsCKW, Sutherland RL: Growth factor, steroid, and steroid an-tagonist regulation of cyclin expression associated with changesin T-47D human breast cancer cell cycle progression. Mol CellBiol 1993;13:35577–3587.

40. Lebwohl DE, Muisse-Helmericks R, Sepp-Lorenzino L, Serve S,Timaul M, Bol R, Borgen P, Rosen N: A truncated cyclin D1 geneencodes a stable mRNA in a human breast cancer cell line. On-cogene 1994;9:1925–1929.

124 Perry et al.