translational regulation of cyclin d1 by 15-deoxy-...

Translational Regulation of Cyclin D1 by 15-Deoxy-�12,14-Prostaglandin J2

1

Peggy A. Campo, Sonali Das, Chin-Hui Hsiang,Tim Bui, Charles E. Samuel, and Daniel S. Straus2

Biomedical Sciences Division and Biology Department, University ofCalifornia, Riverside, California 92521-0121 [P. A. C., C-H. H., T. B.,D. S. S.], and Department of Molecular, Cellular, and DevelopmentalBiology, University of California, Santa Barbara, California 93106 [S. D.,C. E. S.]

AbstractThe D-group cyclins play a key role in the progressionof cells through the G1 phase of the cell cycle.Treatment of MCF-7 breast cancer cells with thecyclopentenone prostaglandin 15-deoxy-�12,14-PGJ2

(15d-PGJ2) results in rapid down-regulation of cyclinD1 protein expression and growth arrest in the G0/G1

phase of the cell cycle. 15d-PGJ2 also down-regulatesthe expression of cyclin D1 mRNA; however, this effectis delayed relative to the effect on cyclin D1 proteinlevels, suggesting that the regulation of cyclin D1occurs at least partly at the level of translation orprotein turnover. Treatment of MCF-7 cells with 15d-PGJ2 leads to a rapid increase in the phosphorylationof protein synthesis initiation factor eukaryoticinitiation factor 2� (eIF-2�) and a shift of cyclin D1mRNA from the polysome-associated to free mRNAfraction, indicating that 15d-PGJ2 inhibits the initiationof cyclin D1 mRNA translation. The selective rapiddecrease in cyclin D1 protein accumulation isfacilitated by its rapid turnover (t1/2 � 34 min) afterinhibition of cyclin D1 protein synthesis. The half-life ofcyclin D1 protein is not significantly altered in cellstreated with 15d-PGJ2. Treatment of cells with 15d-PGJ2 results in strong induction of heat shock protein70 (HSP70) gene expression, suggesting that 15d-PGJ2

might activate protein kinase R (PKR), an eIF-2� kinaseshown previously to be responsive to agents thatinduce stress. 15d-PGJ2 strongly stimulates eIF-2�

phosphorylation and down-regulates cyclin D1expression in a cell line derived from wild-type mouseembryo fibroblasts but has an attenuated effect inPKR-null cells, providing evidence that PKR is involvedin mediating the effect of 15d-PGJ2 on eIF-2�

phosphorylation and cyclin D1 expression. In summary,

treatment of MCF-7 cells with 15d-PGJ2 results inincreased phosphorylation of eIF-2� and inhibition ofcyclin D1 mRNA translation initiation. At later timepoints, repression of cyclin D1 mRNA expression mayalso contribute to the decrease in cyclin D1 protein.

IntroductionTransit of normal mammalian cells through G1 and into theS-phase of the cell cycle requires the action of mitogens andis controlled by CDKs3 that are sequentially activated bycyclins D, E, and A (1–5). The D-type cyclins (D1, D2, and/orD3, depending on the cell type) act as sensors for the pres-ence of mitogenic factors. Treatment of quiescent cells withmitogens causes increased expression of D-type cyclins,thereby leading to the activation of CDK4 or CDK6. ActivatedCDK4 or CDK6 complexes promote progression through G1

phase in two ways. First, these complexes catalyze the firstin a series of site-specific phosphorylations of Rb thatultimately inactivate Rb as a transcriptional repressor (6).Second, these complexes sequester the CDK inhibitorsp21CIP1/WAF1 and p27KIP1, thus relieving the inhibitory effectof the CDK inhibitors on the catalytic activity of cyclin Eand cyclin A-CDK2 complexes (7). Evidence suggests thatmitogens regulate the expression/activity of cyclin D1/CDKcomplexes at several levels, including cyclin D1 gene tran-scription (8–10) and cyclin D1 translation (11), nuclear local-ization and turnover of cyclin D1 protein (12), and assemblyof active cyclin/CDK complexes (13).

Overexpression of cyclin D1 has been implicated in theetiology of a number of types of human cancer (14). In-creased expression of cyclin D1 is observed in �50% ofinvasive primary breast carcinomas by immunohistochemi-cal staining (15). In about 15–20% of primary breast cancers,the overexpression of cyclin D1 is caused by gene amplifi-cation (16). The molecular mechanism(s) for overexpressionof cyclin D1 in the other tumors is unknown, although stabi-lization of cyclin D1 mRNA (17) and stabilization of cyclin D1protein (18) have been suggested as possible mechanisms.Overexpression of cyclin D1 mRNA and protein is also ob-served in 50% or more of breast ductal carcinoma in situlesions but more rarely in proliferative disease or atypicalductal hyperplasia, providing evidence that cyclin D1 over-expression plays a role during an early stage of tumor de-velopment (19). Results obtained with animal model systems

Received 12/14/01; revised 6/17/02; accepted 6/24/02.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 1734 solely to indi-cate this fact.1 This research was supported by Department of Defense Breast CancerResearch Program Grant DAMD17-99-1-9102 (to D. S. S.) and NIH GrantAI20611 (to C. E. S.).2 To whom requests for reprints should be addressed, at Biomedical Sci-ences Division, University of California, Riverside, CA 92521-0121. Phone:(909) 787-5612; Fax: (909) 787-5504; E-mail: [email protected].

3 The abbreviations used are: CDK, cyclin-dependent protein kinase; 15d-PGJ2, 15-deoxy-�12,14-PGJ2; PG, prostaglandin; CP, cyclopentenone;PPAR, peroxisome proliferator-activated receptor; CHX, cycloheximide;eIF-2�, eukaryotic initiation factor 2�; PKR, protein kinase R; FACS,fluorescence-activated cell sorter; GAPDH, glyceraldehyde-3-phosphatedehydrogenase; PERK, PKR-like ER kinase; ER, endoplasmic reticulum;HSP, heat shock protein; GRP, glucose-regulated protein; BCS, bovinecalf serum.

409Vol. 13, 409–420, September 2002 Cell Growth & Differentiation

also suggest a role for cyclin D1 in malignant transformationof breast cancer cells (20, 21).

The PGs are a family of biologically active molecules hav-ing a diverse range of actions, depending on the PG type andcell target. Within this family, PGs of the A and J series, whichcontain a CP ring system, are potent inhibitors of cell prolif-eration in vitro and are able to suppress tumorigenicity in vivo(reviewed in Ref. 22). The antitumor activity of the CP PGsdepends on the presence of an �,�-unsaturated carbonylmoiety within the CP ring, which reacts avidly with nucleo-philes, such as sulfhydryls located on cysteine residues incellular proteins and glutathione (22). The CP PGs causegrowth arrest in G1 or cell death, depending on the PG doseand characteristics of the target tumor cells (22–24). Theability of the CP PGs to cause growth arrest or cell death ina variety of tumor cell lines has raised the possibility that theymight be useful for the treatment of human cancer (24, 25).

The antineoplastic activity of the CP PGs is thought to berelated to their ability to regulate the expression of a varietyof stress-induced and cell cycle-related genes (22). Genesexhibiting increased expression in response to the CP PGsinclude the genes encoding HSP70 (26), c-fos (27), Egr-1(27), gadd153 (28), and the CDK inhibitor p21CIP1/WAF1 (29,30). A number of genes exhibit decreased expression inresponse to the CP prostaglandins, including the genes thatencode c-myc (31), N-myc (32), insulin-like growth factor I(30, 33), cyclin D1 (29, 33–35), and CDK4 (29). Previousstudies have demonstrated that the CP PG PGA2 causesgrowth arrest of MCF-7 breast cancer cells and C6 rat gliomacells in the G1 phase of the cell cycle (29, 33). In the MCF-7cells, PGA2 causes a concerted repression of cyclin D1 andCDK4 and induction of the CDK inhibitor p21CIP1/WAF1,events that are causally related to arrest in G1 (29, 33, 34).

Because of the role that cyclin D1 overexpression plays inmalignant transformation in breast cancer and other tumors,cyclin D1 is a potential target for the rational design of newdrugs to prevent or treat cancer. In the present study, weexamined the molecular mechanism for repression of cyclinD1 protein expression by the CP PGs. The results indicatethat these compounds cause a very rapid down-regulation ofcyclin D1 protein, which precedes a decrease in cyclin D1mRNA. The most active compound is 15d-PGJ2. The mech-anism for rapid down-regulation of cyclin D1 protein is inhi-bition of cyclin D1 translation, caused at least in part byincreased phosphorylation of eIF-2�.

ResultsEffect of PGs on Cyclin D1 Protein Expression. It hasbeen shown previously that treatment of cells with the CP

Fig. 1. A, dose-response relationships for regulation of cyclin D1 proteinby 15d-PGJ2 (top), PGA2 (middle), and CP (bottom). �, location of elec-trophilic carbon atoms. B, effect of 15d-PGJ2 on cyclin D1 protein ex-pression in HeLa and NIH-3T3 cells. Cultures were treated with DMSOvehicle or 10 �M 15d-PGJ2 for 2 h. Protein extracts were prepared and

subjected to Western blotting (upper panel). Results were quantified byscanning densitometry (lower panel). Bars, means of three different cul-tures; �, SE. C, dose-response for regulation of cyclin D1 protein byhydrogen peroxide (top, triplicate dishes treated with each concentration),sodium arsenite (middle, triplicate dishes treated with each concentra-tion), and UV light (bottom). MCF-7 cells were treated for 1 h with eachcompound at the indicated concentrations or with UV as described in“Materials and Methods.” Extracts were prepared, and the cyclin D1 and�-actin protein levels were quantified by Western blotting.

410 Translational Regulation of Cyclin D1 by 15d-PGJ2

prostaglandin PGA2 results in a rapid decrease in cyclinD1 protein levels (29). In our initial experiments, we com-pared the activity of PGA2 with the activity of two othercompounds, the CP PG 15d-PGJ2 and the model com-pound CP. Dose-response experiments performed withcells treated for 1 h with the three agents demonstratedthat 15d-PGJ2 was much more active in down-regulatingcyclin D1 protein levels than either PGA2 or CP in MCF-7cells (Fig. 1A). 15d-PGJ2 was active at the low dose of 3�M and maximally effective at a dose of 10 �M. The othertwo compounds also down-regulated cyclin D1 protein butat considerably higher concentrations. Quantitative anal-ysis of the Western blots using laser scanning densitom-etry yielded estimated IC50s of 5, 290, and 325 �M for15d-PGJ2, PGA2, and CP, respectively. �-Actin was usedas a control in these experiments and was not affected bythe treatment with these compounds.

15d-PGJ2 is a high-affinity ligand for the nuclear receptorPPAR� (reviewed in Ref. 22). Some biological effects of15d-PGJ2 are mediated by binding to PPAR�, and some areindependent of the receptor (22). We have observed previ-ously that MCF-7 breast cancer cells express PPAR� (36). Totest whether the effect of 15d-PGJ2 on cyclin D1 protein wasmediated by PPAR�, we examined the effect on cyclin D1protein levels of the thiazolidinedione compound rosiglita-zone (BRL49653), a high-affinity PPAR� ligand that is struc-turally distinct from 15d-PGJ2 and that lacks the chemicallyreactive �,�-unsaturated carbonyl group (37). Treatment ofcells for 1 h with rosiglitazone at concentrations up to 10 �M

did not have a significant effect on cyclin D1 protein levels(results not shown), indicating that this effect of 15d-PGJ2

may be independent of PPAR�. To investigate further thepossible involvement of PPAR� in mediating the repression

of cyclin D1 by 15d-PGJ2, we tested the effect of 15d-PGJ2

on cyclin D1 expression in HeLa cells and NIH-3T3 cells.Previous studies have demonstrated that neither cell lineexpresses detectable levels of PPAR� protein and that nei-ther exhibits biological responses known to be mediated byPPAR� (36, 38, 39). Treatment of the HeLa and NIH-3T3 cellswith 15d-PGJ2 resulted in decreased expression of cyclin D1(Fig. 1B). This result provides additional evidence demon-strating the existence of a PPAR�-independent pathway forregulation of cyclin D1 by 15d-PGJ2.

Several stress-inducing agents other than 15d-PGJ2 havebeen shown previously to decrease cyclin D1 protein levelsin cultured cells (40, 41). We next tested the effect of hydro-gen peroxide, sodium arsenite, and UV light on cyclin D1protein levels in MCF-7 cells (Fig. 1C). Each of these agentsdecreased cyclin D1 protein levels but not to the degree norat such a low dose as 15d-PGJ2. Therefore, all subsequentexperiments were performed with 15d-PGJ2.

15d-PGJ2 Blocks Progression of Serum-stimulatedQuiescent Cells through G1. We next determined the effectof 15d-PGJ2 on cell cycle progression of MCF-7 cells (Fig. 2).Cells were arrested by culture in medium with low serum andthen restimulated with fresh medium containing 10% serumwith 15d-PGJ2 or DMSO vehicle. FACS analysis was per-formed to monitor cell cycle phase. Restimulation with me-dium containing 10% serum and vehicle resulted in synchro-nous entry into S-phase. Cells stimulated with mediumcontaining 10% serum and 15d-PGJ2 remained arrested inG1 (Fig. 2).

Time Course of 15d-PGJ2-mediated Reduction in Cy-clin D1 Expression. The time course for the down-regula-tion of cyclin D1 protein expression by 15d-PGJ2 was nextdetermined by Western immunoblot analysis (Fig. 3). The

Fig. 2. Effect of 15d-PGJ2 on cell cycle reentry of MCF-7 cells. MCF-7 cells were incubated for 40 h in DMEM with 0.5% serum to synchronize cells inG0/G1. At T � 0, the medium was changed to fresh medium containing 10% serum with vehicle or 15d-PGJ2 (final concentration, 20 �M). Control and treatedcells were processed at various times and subjected to FACS analysis to determine cell cycle phase. A, FACS analysis of cells restimulated with 10% serumplus vehicle (upper) or 10% serum plus 15d-PGJ2 (lower). B, fraction of cells in S-phase at various time points. F, cells stimulated with 10% serum plusvehicle. f, cells stimulated with 10% serum plus 15d-PGJ2.

411Cell Growth & Differentiation

effect of 15d-PGJ2 (10 �M) was apparent within 30 min afteraddition of 15d-PGJ2, with a maximal effect observed at 120min. Partial recovery of cyclin D1 protein levels was observedat later time points. The partial recovery of cyclin D1 at thelater time points may have been attributable to rapid metab-olism and inactivation of 15d-PGJ2 mediated by glutathione-S-transferase (22). In support of this idea, in additional ex-periments in which a second dose of 15d-PGJ2 (10 �M) wasadded at the 2-h time point, cyclin D1 protein expression didnot recover from the repressed minimum observed at 2 h(data not shown).

Effect of 15d-PGJ2 on Cyclin D1 mRNA Abundance.Previous studies have demonstrated that PGA2 decreasesthe steady-state level of cyclin D1 mRNA as well as cyclin D1protein (29, 33, 34), although in one of these studies, theeffect on protein levels appeared to precede the effect onmRNA levels (29). Treatment of the MCF-7 cells with 15d-PGJ2 (10 �M) led to a time-dependent decrease in cyclin D1mRNA abundance (Fig. 4A). This effect was first evident at2 h, and cyclin D1 mRNA continued to decrease up to 8 hafter 15d-PGJ2 addition (Fig. 4A). The abundance of GAPDHmRNA was unaffected by 15d-PGJ2, indicating that the ef-fect of 15d-PGJ2 on cyclin D1 mRNA was specific. Impor-tantly, no significant effect of 15d-PGJ2 on cyclin D1 mRNAwas observed at 1 h after 15d-PGJ2 addition (Fig. 4B), a timepoint at which cyclin D1 protein was already decreased by�75% (Fig. 3). The more rapid effect of 15d-PGJ2 on cyclinD1 protein as compared with cyclin D1 mRNA indicated that15d-PGJ2 decreased cyclin D1 protein at least partly at astep subsequent to cyclin D1 mRNA production/turnover.

15d-PGJ2 Alters the Polysome Distribution of Cyclin D1mRNA. The rapid disappearance of cyclin D1 protein afterthe addition of 15d-PGJ2 could be attributable either toinhibition of translation of cyclin D1 mRNA or accelerateddegradation of cyclin D1 protein. To determine whether 15d-PGJ2 might decrease the translation of cyclin D1 mRNA, weexamined the amount of this mRNA present in the polysomal(i.e., actively translated) fraction in the absence or presenceof 15d-PGJ2 (Fig. 5A). In control cells, part of the cyclin D1mRNA sedimented with the polysomal fraction and part sedi-mented with the untranslated fraction near the top of thegradient (Fig. 5A, top panel). After treatment of cells with15d-PGJ2 (10 �M, 1 h), the polysome-associated mRNA wascompletely shifted to the more slowly sedimenting region ofthe gradient, indicating that cyclin D1 translational initiationwas severely inhibited (Fig. 5A, lower panel). The resultsdepicted in Fig. 5 were obtained 1 h after addition of 15d-PGJ2, at which time decreased size or abundance of cyclinD1 mRNA was not observed (Fig. 4B). Therefore, increasedcyclin D1 mRNA degradation could not account for the re-sults presented in Fig. 5A. �-Actin mRNA from 15d-PGJ2-treated cells was also shifted to a lighter fraction in thegradient, indicating that �-actin translation was also inhibited(Fig. 5B). However, the shift of �-actin mRNA was not ascomplete as with cyclin D1 mRNA; part of the �-actin mRNAwas shifted to an intermediate position in the gradient, con-sistent with the location of mRNA associated with a reducednumber rather than completely depleted of ribosomes (Fig.5B, lower panel). The overall results indicate that 15d-PGJ2

inhibits the translation of cyclin D1 and �-actin mRNA andare consistent with the possibility that the inhibitory effect of15d-PGJ2 on �-actin translation may be less severe than itsinhibitory effect on cyclin D1 translation.

Cyclin D1 Protein Half-Life in MCF-7 Cells. As thesteady state level of any protein in the cell is equivalent to theamount synthesized minus the amount degraded, it wasimportant to determine the half-life of cyclin D1 protein in

Fig. 3. Time course for the effect of 15d-PGJ2 on cyclin D1 expression.A, MCF-7 cells were treated with 10 �M 15d-PGJ2 for the indicated times,and cyclin D1 and �-actin protein levels were determined by Westernblotting. B, quantitative analysis of the cyclin D1 Western blot was per-formed using a laser densitometer. Error brackets, SE for three determi-nations (0 time point) or the range for two determinations (all other timepoints).

Fig. 4. Northern blots depicting the effect of 10 �M 15d-PGJ2 on cyclinD1 and GAPDH mRNA levels. A, effect on the quantity of cyclin D1 mRNAin cells treated for 2–8 h with 15d-PGJ2. B, effect in cells treated for 1 h.The major (4 kb) species of cyclin D1 mRNA is shown. C, vehicle-treatedcontrols; T, treated with 10 �M 15d-PGJ2 for the time periods indicated.


MCF-7 cells under our experimental conditions. To do this,CHX was added to stop protein synthesis, and the timecourse for the disappearance of cyclin D1 was determined(Fig. 6A). Cyclin D1 protein is degraded by the proteasome(12, 42–45) and calpain protease (46), and it is known to turnover rapidly. Our results indicated that the half-life of cyclinD1 in MCF-7 cells was �34 min, in concordance with pre-viously published half-lives (42, 47–49). In contrast, the con-trol protein �-actin turned over much more slowly in thesame cultures (Fig. 6A). (Linear regression analysis indicatedthat the slope of the �-actin decay curve did not differ sig-nificantly from zero over the 60-min time course.) We nextperformed pulse-chase experiments to examine directly thepossible effect of 15d-PGJ2 on cyclin D1 protein turnover.The results (Fig. 6B) indicated that 15d-PGJ2 had no signif-icant effect on cyclin D1 protein degradation. The half-life ofcyclin D1 protein in control cells was 44.8 min and in 15d-PGJ2-treated cells was 45.2 min. These half-lives are slightlylonger but in reasonable agreement with the half-life meas-ured after CHX treatment. Taken together with the resultspresented above, this result provides evidence indicatingthat the rapid effect of 15d-PGJ2 on cyclin D1 protein ex-pression is caused by inhibition of cyclin D1 translation ratherthan acceleration of cyclin D1 protein turnover.

15d-PGJ2 Generates a Predominantly CytoplasmicStress. The results presented above suggested that therapid down-regulation of cyclin D1 protein expression in cells

treated with 15d-PGJ2 (Fig. 3) could be accounted for bysevere inhibition of cyclin D1 translation (Fig. 5A), followed byrapid turnover of cyclin D1 protein (Fig. 6). One clearly un-derstood molecular mechanism for inhibition of translationalinitiation in eukaryotic cells is increased phosphorylation ofeIF-2� on serine 51, which results in inhibition of the GTPexchange factor eIF-2B, and ultimately, decreased formationof ternary initiation complexes. A number of agents thatinduce stress are known to stimulate phosphorylation ofeIF-2� and inhibit translation (50–52). These can be roughlydivided into two classes. Agents such as sodium arseniteinduce stress in the cytoplasm and activate PKR, whichphosphorylates eIF-2� (50, 51). Agents such as tunicamycinand thapsigargin induce stress in the ER and activate PERK,which also phosphorylates eIF-2� (52). To determinewhether 15d-PGJ2 at the concentration used in our experi-ments (10 �M) induced a stress response and whether thisresponse originated in the cytoplasm or ER, we examinedthe effect of 15d-PGJ2 on expression of the genes for twomolecular chaperones, HSP70 and GRP78. HSP70 is in-duced primarily by stress in the cytoplasm, whereas GRP78is induced primarily by stress in the ER. The results (Fig. 7)showed very strong induction by 15d-PGJ2 of HSP70 mRNAand relatively weak and delayed induction of GRP78 mRNA.These results are consistent with a strong cellular stressresponse to 10 �M 15d-PGJ2, with the stress emanatingprimarily from the cytoplasm rather than the ER.

Fig. 5. Polysomal profiles fromcontrol MCF-7 cells treated withDMSO vehicle (top) and treatedwith 10 �M 15d-PGJ2 (bottom).The A254 of each fraction from asucrose gradient (fraction #1 rep-resenting the top of the gradient)was determined (E), and isolatedRNA was subject to Northern blotanalysis to determine which frac-tions contained cyclin D1 (A) or�-actin (B) mRNA (F). The positionof free ribosomes was determinedby ethidium bromide staining ofthe Northern gels. Fraction #3within this region contained 40Ssubunits; fractions #4, 5, and 6contained 60S subunits and 80Sribosomes, incompletely resolvedfrom each other.


15d-PGJ2 Increases Phosphorylation of TranslationInitiation Factor eIF-2�. We next examined directly theeffect of 15d-PGJ2 on eIF-2� phosphorylation. Sodium ar-senite, which is known to activate PKR (50, 51) and whichalso down-regulates cyclin D1 (Fig. 1C), was used as apositive control. The results (Fig. 8) indicated that 15d-PGJ2

did in fact stimulate the phosphorylation of eIF-2� in MCF-7cells. This effect was evident at 5 min after addition of 15d-PGJ2 and was maximal at 15 min. The overall results are thusconsistent with increased phosphorylation of eIF-2�, withtranslational inhibition being the major mechanism for therapid down-regulation of cyclin D1 by 15d-PGJ2.

PKR Is Involved in Mediating the Effect of 15d-PGJ2 oneIF-2� Phosphorylation and Cyclin D1 Expression. Toinvestigate the possible role of PKR in mediating the induc-tion of eIF-2� phosphorylation by 15d-PGJ2, we examinedthe effect of 15d-PGJ2 on eIF-2� phosphorylation in wildtype (Pkr�/�) and PKR-null (Pkr0/0) mouse embryo fibroblastcell lines (Fig. 9). 15d-PGJ2 (10 �M) treatment significantlyincreased eIF-2� phosphorylation in wild-type fibroblastsabove the level seen in the absence of the prostaglandin. Theincrease reached a maximum after 2 h of treatment of thewild-type fibroblasts. In contrast, very little effect was ob-served in the PKR-null mutant fibroblasts under similar con-ditions (Fig. 9). These results suggest PKR involvement inmediating the response to 15d-PGJ2.

We next tested the effect of 15d-PGJ2 on cyclin D1 proteinlevels in the wild-type and PKR-null mouse fibroblast cells(Fig. 10). 15d-PGJ2 treatment significantly reduced cyclin D1

protein levels in the wild-type fibroblasts (Fig. 10). In threedifferent experiments, the mean decrease in cyclin D1 pro-tein expression was 55% after 2 h and 56% after 3 h treat-ment of wild-type Pkr�/� cells with 15d-PGJ2, as comparedwith vehicle-treated cells. By contrast, under similar condi-tions, 15d-PGJ2 treatment had relatively little effect on thecyclin D1 protein levels in PKR-null fibroblasts. In these cells,cyclin D1 was decreased by 25% after 2 h and by 15% after3 h in cells treated with 15d-PGJ2, as compared with vehicle-treated cells. The greater response of cyclin D1 to 15d-PGJ2

in wild-type cells as compared with PKR-null cells was con-sistently observed in all three experiments. These results,taken together with those presented in Fig. 9, provide evi-dence that PKR is involved in mediating the effect of 15d-PGJ2 on eIF-2� phosphorylation and cyclin D1 expression.

DiscussionOverexpression of cyclin D1 has been implicated in the eti-ology of a number of types of human cancer including breastcancer (14–20). Therefore, cyclin D1 is a potential target forthe rational design of new drugs to prevent or treat cancer.The present study confirms and extends previous resultsdemonstrating that PGA2 down-regulates cyclin D1 in tumorcells (29, 33). We report here the novel result that 15d-PGJ2

is �60 times more potent than PGA2 in eliciting this effect.15d-PGJ2 is a high-affinity ligand for nuclear receptorPPAR�, and MCF-7 cells have been shown previously toexpress PPAR� (22, 36). However, the rapid effect of 15d-

Fig. 6. A, time course showing effect of CHX on cyclin D1 (F) and �-actin (f) protein levels. CHX (final concentration, 100 �g/ml) was added at T � 0 toblock protein synthesis, and levels of cyclin D1 and �-actin protein were quantified by Western blotting. Each point represents the mean of three differentcultures � SE. The mean for the zero time point was set at 1.0. Regression analysis indicated that the slope of the cyclin D1 decay curve differed significantlyfrom zero (P � 0.0001), whereas the slope of the �-actin decay curve did not (P � 0.0932). The half-life of cyclin D1 protein was also calculated by regressionanalysis. B, effect of 15d-PGJ2 on cyclin D1 protein degradation as determined by the pulse-chase method. Cells were pulse-labeled with [35S]methioninefor 1 h, washed three times, and transferred to chase medium containing 2 mM unlabeled methionine at T � 0 (see “Materials and Methods”). Cyclin D1protein was immunoprecipitated at the indicated chase times and quantified by SDS gel electrophoresis and autoradiography. The level of labeled cyclinD1 protein at the beginning of the chase (T � 0) was set at 1.0. Each data point represents the mean of results obtained in two different experiments. Circlesand solid line, control cells; triangles and broken line, cells treated with 15d-PGJ2. The half-life of cyclin D1 protein was calculated by regression analysis.


PGJ2 on cyclin D1 does not appear to be mediated byPPAR�. In particular: (a) PPAR� is a regulator of transcription,whereas the effects that we have observed are posttran-scriptional; (b) another high-affinity PPAR� ligand, rosiglita-zone, had no effect on cyclin D1; (c) 15d-PGJ2 down-regu-lated cyclin D1 expression in HeLa and NIH-3T3 cells, whichdo not express detectable levels of PPAR�; and (d) CP,which is not a PPAR� ligand, was active.

Most, if not all, of the PPAR�-independent biological ac-tions of the CP PGs depend on the presence of the chemi-cally reactive �,�-unsaturated carbonyl within the CP ring(22, 30, 36). 15d-PGJ2 has been observed previously to bemore active than PGA2 with other biological endpoints in-cluding repression of nuclear factor-�B activation (36). 15d-

PGJ2 differs from PGA2 in having two chemically reactivecenters rather than one (Fig. 1A). Thus, it is possible that15d-PGJ2 could act as a cross-linking agent. Alternatively, itis possible that the higher bioactivity of 15d-PGJ2 is relatedto higher affinity for some yet-to-be-identified protein targetwithin the cell.

Similar to earlier results obtained with PGA2, we found that15d-PGJ2 negatively regulates cyclin D1 mRNA levels, andthis effect presumably contributes to the down-regulation ofcyclin D1 protein expression at times �2 h after compoundaddition. The repressive effect of the CP PGs on cyclin D1mRNA expression is reported to result from transcriptionalrepression (35) and/or destabilization of cyclin D1 mRNA(34). However, consistent with an earlier report (29), we foundthat 15d-PGJ2 decreased the level of cyclin D1 protein be-fore it decreased the level of cyclin D1 mRNA. The discordanttime courses for the changes in cyclin D1 protein and mRNAindicated the existence of an additional mode of regulation ata step subsequent to mRNA production/turnover. Two for-mal possibilities for this regulation were translational repres-sion or accelerated protein turnover. Our results indicate thattranslational repression is one mechanism for the regulationof cyclin D1 by 15d-PGJ2. Thus, the effect of 15d-PGJ2 oncyclin D1 resembles the effects of clotrimazole and tunica-mycin, which also appear to regulate cyclin D1 translation(40, 49, 53). In contrast, retinoic acid and osmotic shock

Fig. 7. A, Northern blot analysis showing the effect of 10 �M 15d-PGJ2on HSP70, GRP78, and GAPDH mRNA levels. C, vehicle-treated controls;T, treated with 10 �M 15d-PGJ2 for the time periods indicated. B, resultsof experiments presented in A quantified by scanning densitometry. �,vehicle controls; f, treated with 15d-PGJ2. Error brackets, SE.

Fig. 8. A, effect of 15d-PGJ2 on phosphorylation of eIF-2� in MCF-7cells. Cultures were treated with 10 �M 15d-PGJ2 (T), DMSO vehicle (C),or 100 �M sodium arsenite (A) for the indicated times. Protein extractswere prepared and subjected to Western blotting. Blots were probed withspecific antibody for phospho-eIF-2� (top) and then reprobed with anti-body for total eIF-2� (middle) and �-actin (bottom). B, results of experi-ments presented in A quantified by scanning densitometry. Error brackets,range for duplicate determinations.


have been reported to stimulate cyclin D1 protein turnover(43–45).

One of the most clearly understood mechanisms for reg-ulation of translation in eukaryotic cells is increased phos-phorylation of eIF-2� (54). Our results suggest that this mod-ification of eIF-2� is likely a principal mechanism fortranslational regulation by 15d-PGJ2. The differential regula-tion of cyclin D1 as compared with �-actin protein abun-dance (Fig. 1) is readily explainable by the more rapid turn-over of cyclin D1 protein as compared with �-actin afterinhibition of translation. In particular, the time course for thedecrease of cyclin D1 protein in cells treated with 15d-PGJ2

(Fig. 3) is very similar to the time course for its decrease aftertreatment of cells with the known translation inhibitor, CHX(Fig. 6A). In contrast, �-actin turns over very slowly afterinhibition of protein synthesis (Fig. 6A). The polysome profileresults also leave open the possibility that cyclin D1 transla-tion may be more severely repressed than �-actin transla-tion, and that this may also contribute to the observed dif-ferential regulation of the steady-state levels of the twoproteins. Differential regulation of the translation of reoviralmRNA transcripts by PKR has been described previously(55).

Although induction of HSP70 and GRP78 by CP PGs hasbeen reported previously (26, 56), to our knowledge this isthe first study in which both endpoints have been measured

with the same compound in the same experiment. Treatmentof MCF-7 cells with 15d-PGJ2 resulted in a rapid stronginduction of HSP70 gene expression. This effect was clearlyapparent at 2 h and was maximal at 4 h. In contrast, theeffect on GRP78 was weak and delayed compared with theeffect on HSP70. Thus, 15d-PGJ2 induces a stress response,and this response emanates primarily from the cytoplasmrather than the ER.

In considering various hypotheses to explain the repres-sive effect of 15d-PGJ2 on cyclin D1 translation, we notedthat stress-inducing agents such as sodium arsenite havebeen shown previously to induce phosphorylation of eIF-2�

(Refs. 50, 51; Fig. 8). Treatment of MCF-7 cells with sodiumarsenite also results in decreased levels of cyclin D1 protein(Fig. 1C). Sodium arsenite has been shown recently to acti-vate PKR (51); thus, PKR was a likely candidate enzyme thatphosphorylated eIF-2� in response to 15d-PGJ2. Our find-ings obtained with the Pkr0/0 fibroblasts, which had an at-tenuated response to 15d-PGJ2 as compared with wild-typefibroblasts, are consistent with the notion that PKR plays arole in the increase in eIF-2� phosphorylation and repressionof cyclin D1 caused by 15d-PGJ2. An alternative pathway(s)responsible for the weak residual response of the Pkr0/0 cellsto 15d-PGJ2 remains to be elucidated. Interestingly, the timecourse for the increase in eIF-2� phosphorylation in re-sponse to 15d-PGJ2 appeared to be more gradual in mouseembryo fibroblastic cells as compared with the humanMCF-7 cells. MCF-7 cells have been reported to expresshigh levels of PKR (57), and it is possible that this accountsfor the more rapid response observed in these cells. Alter-natively, the response in the two cell types may be modu-lated by other factors, such as differences in ability to me-tabolize 15d-PGJ2 (36) or differences in activities of cellularproteins that modulate PKR function such as PACT or P58,a known regulator of HSP70 (58).

Tunicamycin, which induces stress in the ER and initiatesthe unfolded protein response, has been reported previouslyto repress cyclin D1 translation (49). This agent is known toactivate another eIF-2� kinase, PERK (52), located in the ER.In analogy with the results presented here, it is possible thatthe mechanism for inhibition of cyclin D1 translation by tu-nicamycin is activation of PERK, followed by phosphoryla-tion of eIF-2� (49).

Fig. 10. Time course showing the effect of 15d-PGJ2 treatment on cyclinD1 protein expression in wild-type and PKR-null mouse embryo fibroblastcells. Wild-type (Pkr�/�) or PKR-null (Pkr0/0) cells were treated for theindicated period of time with 20 �M 15d-PGJ2. Cyclin D1 protein levelswere determined by Western immunoblot analysis.

Fig. 9. Effect of 15d-PGJ2 treatment on the phosphorylation of eIF-2� inwild-type and PKR-null mouse embryo fibroblast cells. A, mouse embryofibroblast cells, either wild type (Pkr�/�) or PKR-null (Pkr0/0), were treatedfor the indicated period of time with 10 �M 15d-PGJ2. PhosphorylatedeIF-2� or �-tubulin protein levels were determined by Western immuno-blot analysis. B, Western blots were quantified by scanning densitometry.Error brackets, SE for three independent experiments. �, control wild-type cells; f, wild-type cells treated with 15d-PGJ2; E, control PKR-nullcells; F, PKR-null cells treated with 15d-PGJ2.


Previous studies of the antineoplastic activity of the PGJseries prostaglandins have focused primarily on �12-PGJ2

(22, 59), the precursor of 15d-PGJ2. McClay et al. (59) havereported synergistic cytotoxic interaction between �12-PGJ2

and both ionizing radiation and cisplatin. This synergy wasobserved in a variety of tumor cell lines and resulted in adecrease of �10-fold in the dose of cisplatin or ionizingradiation needed to kill 50% of the tumor cells. It will be ofinterest to determine whether 15d-PGJ2 also synergizes withother antitumor agents. A limiting factor for usefulness ofmost anticancer drugs is toxicity in vivo. A recent study hasdemonstrated that 15d-PGJ2 (1 mg/kg/day i.p.) is well toler-ated in rats (60). At this dose, anti-inflammatory activity of15d-PGJ2 was observed in adjuvant-induced arthritis, withno toxic side effects (60). The low toxicity of 15d-PGJ2 in thisstudy provides a rationale for future testing of the antitumoractivity of this compound in vivo.

Materials and MethodsMaterials. Prostaglandins were obtained from CaymanChemical Co. and were supplied in methyl acetate. Beforeusing PGA2 and 15d-PGJ2 in experiments, the methyl ace-tate was evaporated under a stream of argon, and the com-pounds were redissolved in argon-purged DMSO. 2-Cyclo-penten-1-one was purchase from Aldrich Chemical Co. anddiluted into DMSO before use. Cycloheximide was pur-chased from Sigma Chemical Co., and stock solutions wereprepared in ethanol at a concentration of 10 mg/ml.

DNA clones were generously provided by the followingpeople: full-length human cyclin D1 cDNA clone (Ref. 61;David Beach, Cold Spring Harbor Laboratory, Cold SpringHarbor, NY); hamster GRP78, human HSP70, and GAPDHcDNA clones (Refs. 62, 63; Amy Lee, University of SouthernCalifornia School of Medicine, Los Angeles, CA); and human�-actin (Ref. 64; Larry Kedes, University of Southern Califor-nia School of Medicine).

Cell Culture. MCF-7 cells were obtained from the Amer-ican Type Culture Collection and maintained as monolayercultures in DMEM (Cellgro) supplemented with 10% fortifiedBCS (Cosmic calf serum; HyClone) plus penicillin (100 units/ml) and streptomycin (100 �g/ml). Permanent cell lines de-rived from embryo fibroblasts of Pkr�/� and Pkr0/0 mice (65)were generously provided by A. E. Koromilas (Lady DavisInstitute, Jewish General Hospital, Montreal, Canada) andcultured in DMEM plus 10% fetal bovine serum and anti-biotics.

Western Blot Analysis. Cells were plated at a density of700,000 cells/6-cm dish and cultured for 3 days at 37°C. Atthe beginning of each experiment, cell cultures were washedonce with PBS and incubated for 1 h in 0.5% BCS medium.15d-PGJ2 or other chemical agents were then added to thecultures, and cultures were incubated for an additional 60min unless otherwise indicated. For treatment with UV, cellswere cultured using the same procedure as for treatmentwith chemicals. The medium was then aspirated, and thecultures were irradiated with UV at the dose indicated, usinga Stratalinker UV cross-linking apparatus (Stratagene). Pre-warmed (37°C) medium (DMEM with 0.5% BCS and antibi-

otics) was then added to each culture, and incubation wascontinued for 60 min at 37°C.

For preparation of extracts for Western blot analysis, cellcultures were transferred to ice, washed once with PBS, andscraped into 1 ml of ice-cold PBS. Cells were then pelletedin a microcentrifuge and suspended in NETN extractionbuffer plus protease inhibitors [50 mM Tris (pH 7.6), 0.15 M

NaCl, 5 mM EDTA, 0.5% NP40, with 1 �l of 0.4 M benzami-dine, 1 �l of 10 mg/ml leupeptin, 1 �l of 10 mg/ml aprotinin,and 1 �l of 250 mM phenylmethylsulfonyl fluoride for every300 �l of NETN buffer]. For Western blot analysis of eIF-2�

phosphorylation, the following phosphatase inhibitors werealso included in the NETN buffer: NaF (20 mM), �-glycero-phosphate (20 mM), and sodium PPi (12 mM). Extracts wereclarified by centrifugation in a microcentrifuge, and proteinconcentration was determined using the Folin/Lowry method(66). Protein extracts (50 �g of protein) were size fractionatedby 10% SDS-PAGE and electroblotted onto nitrocellulosemembranes (NitroBind MSI, Westborough, MA), using stand-ard techniques. For detection of cyclin D1, a rabbit poly-clonal antibody (Santa Cruz Biotechnology Inc.; SC-718) wasused at a 1:400 dilution. For detection of �-actin, a rabbitaffinity-isolated, antigen-specific antibody (Sigma-AldrichCo.; A-2066) was used at a 1:800 dilution. For detection oftotal eIF-2�, a rabbit polyclonal antibody (Santa Cruz; SC-11386) was used at a 1:400 dilution. For detection of phos-phorylated eIF-2�, an antibody specific for phospho-Ser51-eIF-2� (Biosource International) was used at a 1:1000dilution (final concentration, 0.5 �g/ml). Antibody directedagainst �-tubulin was provided by L. Wilson (University ofCalifornia, Santa Barbara, CA). The secondary antibody wasa peroxidase-labeled, antirabbit IgG antibody, affinity puri-fied made in goat (Vector Laboratories, Inc. or Amersham) ata dilution of 1:10,000. Proteins were then detected with theenhanced chemiluminescence system (SuperSignal; Pierce,Rockford, IL).

Polysomal Profiles. Polysomal profiles were obtained asdescribed previously (67), with slight modifications. For prep-aration of cytoplasmic extracts, MCF-7 cells (three 15-cmcell culture plates) were treated with 15d-PGJ2 (10 �M) orvehicle for 1 h before extraction. The plates were then treatedwith CHX (100 �g/ml) for 5 min at 37°C, washed with PBScontaining CHX (100 �g/ml), and scraped into PBS � CHX.Cells were pelleted by centrifugation, swollen for 2 min inlow-salt buffer [LSB; 20 mM Tris (pH 7.5), 10 mM NaCl, 3 mM

MgCl2, 1 mM DTT, 50 units of RNAsin (Promega)], and thenlysed by the addition of lysis buffer (0.2 M sucrose, 0.1%Triton X-100, in LSB), followed by 10 strokes with a Douncehomogenizer using piston A. The nuclei were pelleted bycentrifugation in a microcentrifuge at 15,000 rpm for 30 s.The supernatant corresponding to the cytoplasmic extractwas poured into a new centrifuge tube containing 10% hep-arin (Sigma; 10 mg/ml), 3% 5 M NaCl, and 1 mM DTT. EqualA260 units of cytoplasmic extracts (�80 �g of RNA) fromvehicle or 15d-PGJ2-treated cells were applied to a 0.5–1.5M sucrose gradient (in LSB) layered over a 2 M sucrosecushion, and centrifuged at 36,000 rpm in a Beckman SW41swinging bucket rotor for 220 min at 4°C. Gradients werefractionated from the top into 1:10 volume of 10% SDS and


proteinase K. The absorbance of each fraction at 254 nmwas monitored. The RNA from each fraction was extractedwith an equal volume of phenol/chloroform, precipitated withisopropanol, and analyzed by Northern blotting to quantifycyclin D1 mRNA.

Determination of Cyclin D1 Protein Half-Life by Pulse-Chase Method. Cells were metabolically labeled with[35S]methionine for 1 h in pulse medium, methionine, andcystine-free DMEM supplemented with 0.5% BCS, antibiot-ics, 0.02 mM unlabeled methionine, and 80 �Ci/ml Trans35S-label (ICN). Cells were treated with either DMSO vehicle(control) or 10 �M 15d-PGJ2 during the pulse period. After the1-h labeling period, cultures were flooded with an excessvolume of chase medium (DMEM, containing 2 mM unlabeledmethionine and supplemented with antibiotics plus 0.5%BCS). The medium was aspirated, and cells were washedthree times with chase medium. Prewarmed chase mediumwas then added to the cells, and incubation was continued at37°C. For cells treated with 15d-PGJ2, this compound wasalso present during the chase. Cultures were harvested, andprotein extracts were prepared at 0, 20, 40, and 60 min,according to the methods described above for Western blot-ting. The protein concentration of each extract was deter-mined by the Bradford method using the Bio-Rad proteinassay reagent (Bio-Rad, Hercules, CA).

Cyclin D1 antibody was bound to protein A-Sepharosebeads by overnight incubation at 4°C. Beads were washedtwice with NETN. Aliquots of protein extract (1 mg of protein)were incubated with antibody-conjugated beads or uncon-jugated beads (negative control) for 2.5 h at 4°C. The beadswere washed four times with NETN and once with PBS.Labeled cyclin D1 bound to the beads was removed byboiling the beads in sample buffer and subjected to SDS-PAGE in 10% gels. The gels were fixed in 45% methanol and10% acetic acid for 20 min and then submerged in 1 M

sodium salicylate for 30 min. The gels were washed threetimes, dried, and subjected to autoradiography. The bandcorresponding to cyclin D1 was quantified by scanning den-sitometry.

Northern Blot Analysis. For Northern blot analysis, totalcellular RNA was extracted as described previously (30, 33).RNA (15-�g aliquots) was denatured and electrophoresed in1% agarose gels containing 2.2 M formaldehyde. The RNAwas then transferred to nylon filters as described previously(30, 33). The integrity of the 18S and 28S bands of theextracted RNA indicated that the RNA had not degradedduring the extraction procedure. Even loading of the gelswas confirmed by ethidium bromide staining. The DNAprobes (gel-purified restriction fragments of cDNA clones)were labeled by random priming with [�-32P]dCTP (30, 33).Filters were prehybridized, hybridized, and washed as de-scribed previously (30, 33). Results were quantified by scan-ning autoradiograms with an LKB UltroScan laser densitom-eter, exercising caution to stay within the linear range of thefilm (30, 33, 34, 37).

Flow Cytometry Analysis. FACS analysis was performedas published previously (33). Briefly, 8 � 105 cells wereplated in 10-cm dishes and grown to 60% confluency inDMEM plus 10% BCS. The culture medium was then aspi-

rated, cells were washed once with PBS, fresh medium with0.5% BCS was added, and incubation was continued for40 h at 37°C to synchronize the cells in the G0/G1 phase.Synchronized cells were stimulated to progress through thecell cycle by changing to fresh medium with 10% BCS withvehicle or 15d-PGJ2 (final concentration, 20 �M). Control andtreated groups were processed at the time points indicatedafter the initiation of cell cycle. Cells were trypsinized andresuspended in ice-cold PBS at a density of 2 � 106 cells/ml.Cells (1 ml of suspension) were fixed by adding 2 ml ofmethanol dropwise with constant mixing and then incubatedon ice for 30 min. Cells were then pelleted by centrifugationand resuspended in 500 �l of stain solution (10 mg of pro-pidium iodide, 0.1 ml of Triton X-100, and 3.7 mg of EDTAdissolved in 100 ml in PBS) and 500 �l of heat-treatedRNase-A solution (2 mg/ml in PBS) for 30 min at roomtemperature in the dark. Stained cells were analyzed using aBecton Dickinson Immunocytometry System for the relativecontent based on red fluorescence levels, and the distribu-tion of cells in different phases of the cell cycle was calcu-lated using CellFIT software (Becton Dickinson).

AcknowledgmentsWe thank A. Lee, D. Beach, and L. Kedes for providing various plasmids,B. Walter for help with the FACS analysis, and A. Koromilas for providingthe wild-type and PKR-null fibroblasts.

References1. Norbury, C., and Nurse, P. Animal cell cycles and their control. Annu.Rev. Biochem., 61: 441–470, 1992.

2. Reed, S. I. The role of p34 kinases in the G1 to S-phase transition.Annu. Rev. Cell Biol., 8: 529–561, 1992.

3. Morgan, D. O. Principles of Cdk regulation. Nature (Lond.), 374: 131–134, 1995.

4. Sherr, C. J., and Roberts, J. M. Inhibitors of mammalian G1 cyclin-dependent kinases. Genes Dev., 9: 1149–1163, 1995.

5. Sherr, C. J. Cancer cell cycles. Science (Wash. DC), 274: 1672–1677,1996.

6. Harbour, J. W., Luo, R. X., Dei Santi, A., Postigo, A. A., and Dean, D. C.CDK phosphorylation triggers sequential intramolecular interactions thatprogressively block Rb functions as cells move through G1. Cell, 98:859–869, 1999.

7. Sherr, C. J., and Roberts, J. M. CDK inhibitors: positive and negativeregulators of G1-phase progression. Genes Dev., 13: 1501–1512, 1999.

8. Albanese, C., Johnson, J., Watanabe, G., Eklund, N., Vu, D., Arnold, A.,and Pestell, R. G. Transforming p21ras mutants and c-Ets-2 activate thecyclin D1 promoter through distinguishable regions. J. Biol. Chem., 270:23589–23597, 1995.

9. Lavoie, J. N., L’Allemain, G., Brunet, A., Muller, R., and Pouyssegur, J.Cyclin D1 expression is regulated positively by the p42/p44MAPK andnegatively by the p38/HOGMAPK pathway. J. Biol. Chem., 271: 20608–20616, 1996.

10. Winston, J. T., Coats, S. R., Wang, Y-Z., and Pledger, W. J. Regula-tion of the cell cycle machinery by oncogenic ras. Oncogene, 12: 127–134,1996.

11. Muise-Helmericks, R. C., Grimes, H. L., Bellacosa, A., Malstrom, S. E.,Tsichlis, P. N., and Rosen, N. Cyclin D expression is controlled post-transcriptionally via a phosphatidylinositol 3-kinase/Akt-dependent path-way. J. Biol. Chem., 273: 29864–29872, 1998.

12. Diehl, J. A., Cheng, M., Roussel, M. F., and Sherr, C. J. Glycogensynthase kinase-3� regulates cyclin D1 proteolysis and subcellular local-ization. Genes Dev., 12: 3499–3511, 1998.


13. Cheng, M., Sexl, V., Sherr, C. J., and Roussel, M. F. Assembly ofcyclin D-dependent kinase and titration of p27Kip1 regulated by mitogen-activated protein kinase kinase (MEK1). Proc. Natl. Acad. Sci. USA, 95:1091–1096, 1998.

14. Hosokawa, Y., and Arnold A. Cyclin D1/PRAD1 as a central target inoncogenesis. J. Lab. Clin. Med., 127: 246–252, 1996.

15. Barnes, D. M. Cyclin D1 in mammary carcinoma. J. Pathol., 181:267–269, 1997.

16. Lammie, G. A., Fantl, V., Smith, R., Schuuring, E., Brookes, S., Micha-lides, R., Dickson, C., Arnold, A., and Peters, G. D11S287, a putativeoncogene on chromosome 11q13, is amplified and expressed in squa-mous cell and mammary carcinomas and linked to BCL-1. Oncogene, 6:439–444, 1991.

17. Lebwohl, D. E., Muise-Helmericks, R., Sepp-Lorenzino, L., Serve, S.,Timaul, M., Bol, R., Borgen, P., and Rosen, N. A truncated cyclin D1 geneencodes a stable mRNA in a human breast cancer cell line. Oncogene, 9:1925–1929, 1994.

18. Russell, A., Thompson, M. A., Hendley, J., Trute, L., Armes, J., andGermain, D. Cyclin D1 and D3 associate with the SCF complex and arecoordinately elevated in breast cancer. Oncogene, 18: 1983–1991, 1999.

19. Zhou, Q., Hopp, T., Fuqua, S. A., and Steeg, P. S. Cyclin D1 in breastpremalignancy and early breast cancer: implications for prevention andtreatment. Cancer Lett., 162: 3–17, 2001.

20. Wang, T. C., Cardiff, R. D., Zukerberg, L., Lees, E., Arnold, A., andSchmidt, E. V. Mammary hyperplasia and carcinoma in MMTV-cyclin D1transgenic mice. Nature (Lond.), 369: 669–671, 1994.

21. Yu, Q., Geng, Y., and Sicinski, P. Specific protection against breastcancers by cyclin D1 ablation. Nature (Lond.), 411: 1017–1021, 2001.

22. Straus, D. S., and Glass, C. K. The cyclopentenone prostaglandins:new insights on biological activities and cellular targets. Med. Res. Rev.,21: 185–210, 2001.

23. Gorospe, M., and Holbrook, N. J. Role of p21 in prostaglandin A2-mediated cellular arrest and death. Cancer Res., 56: 475–479, 1996.

24. Fukushima, M., Sasaki, H., and Fukushima, S. Prostaglandin J2 andrelated compounds. Mode of action in G1 arrest and preclinical results.Ann. NY Acad. Sci., 744: 161–165, 1994.

25. Sasaki, H., Niimi, S., Akiyama, M., Tanaka, T., Hazato, A., Kurozumi,S., Fukushima, S., and Fukushima, M. Antitumor activity of 13, 14-dihy-dro-15-deoxy-�7-prostaglandin-A1-methyl ester integrated into lipid mi-crospheres against human ovarian carcinoma cells resistant to cisplatin invivo. Cancer Res., 59: 3919–3922, 1999.

26. Holbrook, N. J., Carlson S, G., Choi, A. M., and Fargnoli, J. Inductionof HSP70 gene expression by the antiproliferative prostaglandin PGA2: agrowth-dependent response mediated by activation of heat shock tran-scription factor. Mol. Cell. Biol., 12: 1528–1534, 1992.

27. Choi, A. M., Tucker, R. W., Carlson, S. G., Weigand, G., and Holbrook,N. J. Calcium mediates expression of stress-response genes in prostag-landin A2-induced growth arrest. FASEB J., 8: 1048–1054, 1994.

28. Choi, A. M., Fargnoli, J., Carlson, S. G., and Holbrook, N. J. Cellgrowth inhibition by prostaglandin A2 results in elevated expression ofgadd153 mRNA. Exp. Cell Res., 199: 85–89, 1992.

29. Gorospe, M., Liu, Y., Xu, Q., Chrest, F. J., and Holbrook, N. J.Inhibition of G1 cyclin-dependent kinase activity during growth arrest ofhuman breast carcinoma cells by prostaglandin A2. Mol. Cell. Biol., 16:762–770, 1996.

30. Bui, T., and Straus, D. S. Effects of cyclopentenone prostaglandinsand related compounds on insulin-like growth factor-I and Waf1 geneexpression. Biochem. Biophys. Acta, 1397: 31–42, 1998.

31. Ishioka, C., Kanamaru, R., Sato, T., Dei, T., Konishi, Y., Asamura, M.,and Wakui, A. Inhibitory effects of prostaglandin A2 on c-myc expressionand cell cycle progression in human leukemia cell line HL-60. Cancer Res.,48: 2813–2818, 1988.

32. Marui, N., Sakai, T., Hosokawa, N., Yoshida, M., Aoike, A., Kawai, K.,Nishino, H., and Fukushima, M. N-myc suppression and cell cycle arrestat G1 phase by prostaglandins. FEBS Lett., 270: 15–18, 1990.

33. Bui, T., Kuo, C., Rotwein, P., and Straus, D. S. Prostaglandin A2

specifically represses IGF-I gene expression in C6 rat glioma cells. En-docrinology, 138: 985–993, 1997.

34. Lin, S., Wang, W., Wilson, G. M., Yang, X., Brewer, G., Holbrook, N. J.,and Gorospe, M. Down-regulation of cyclin D1 expression by prostag-landin A2 is mediated by enhanced cyclin D1 mRNA turnover. Mol. Cell.Biol., 20: 7903–7913, 2000.

35. Wang, C., Fu, M., D’Amico, M., Albanese, C., Zhou, J. N., Brownless,M., Lisanti, M. P., Chatterjee, V. K., Lazar, M. A., and Pestell, R. G.Inhibition of cellular proliferation through I�B kinase-independent andperoxisome proliferator-activated receptor �-dependent repression of cy-clin D1. Mol. Cell. Biol., 21: 3057–3070, 2001.

36. Straus, D. S., Pascual, G., Li, M., Welch, J. S., Ricote, M., Hsiang,C-H., Sengchanthalangsy, L. L., Ghosh, G., and Glass, C. K. 15-Deoxy-�12,14-prostaglandin J2 inhibits multiple steps in the NF-�B signalingpathway. Proc. Natl. Acad. Sci. USA, 97: 4844–4849, 2000.

37. Nolte, R. T., Wisely, G. B., Westin, S., Cobb, J. E., Lambert, M. H.,Kurokawa, R., Rosenfeld, M. G., Willson, T. M., Glass, C. K., and Milburn,M. V. Ligand binding and co-activator assembly of the peroxisome pro-liferator-activated receptor-�. Nature (Lond.), 395: 137–143, 1998.

38. Tontonoz, P., Hu, E., and Spiegelman, B. M. Stimulation of adipo-genesis in fibroblasts by PPAR-�2, a lipid-activated transcription factor.Cell, 79: 1147–1156, 1994.

39. Chawla, A., Barak, Y., Nagy, L., Liao, D., Tontonoz, P., and Evans,R. M. PPAR-� dependent and independent effects on macrophage-geneexpression in lipid metabolism and inflammation. Nat. Med., 7: 48–52,2001.

40. Tomida, A., Suzuki, H., Kim, H. D., and Tsuruo, T. Glucose-regulatedstresses cause decreased expression of cyclin D1 and hypophosphory-lation of retinoblastoma protein in human cancer cells. Oncogene, 13:2699–2705, 1996.

41. Hiyama, H., and Reeves, S. A. Role for cyclin D1 in UVC-induced andp53-mediated apoptosis. Cell Death Differ., 6: 565–569, 1999.

42. Diehl, J. A., Zindy, F., and Sherr, C. J. Inhibition of cyclin D1 phos-phorylation on threonine-286 prevents its rapid degradation via the ubiq-uitin-proteasome pathway. Genes Dev., 11: 957–972, 1997.

43. Langenfeld, J., Kiyokawa, H., Sekula, D., Boyle, J., and Dmitrovsky, E.Posttranslational regulation of cyclin D1 by retinoic acid: a chemopreven-tion mechanism. Proc. Natl. Acad. Sci. USA, 94: 12070–12074, 1997.

44. Spinella, M. J., Freemantle, S. J., Sekula, D., Chang, J. H., Christie,A. J., and Dmitrovsky, E. Retinoic acid promotes ubiquitination and pro-teolysis of cyclin D1 during induced tumor cell differentiation. J. Biol.Chem., 274: 22013–22018, 1999.

45. Casanovas, O., Miro, F., Estanyol, J. M., Itarte, E., Agell, N., andBachs, O. Osmotic stress regulates the stability of cyclin D1 in ap38SAPK2-dependent manner. J. Biol. Chem., 275: 35091–35097, 2000.

46. Choi, Y. H., Lee, S. J., Nguyen, P., Jang, J. S., Lee, J., Wu, M. L.,Takano, E., Maki, M., Henkart, P. A., and Trepel, J. B. Regulation of cyclinD1 by calpain protease. J. Biol. Chem., 272: 28479–28484, 1997.

47. Welcker, M., Lukas, J., Strauss, M., and Bartek, J. Enhanced proteinstability: a novel mechanism of D-type cyclin over-abundance identified inhuman sarcoma cells. Oncogene, 13: 419–425, 1996.

48. Ko, T. C., Yu, W., Sakai, T., Sheng, H., Shao, J., Beauchamp, R. D.,and Thompson, E. A. TGF-�1 effects on proliferation of rat intestinalepithelial cells are due to inhibition of cyclin D1 expression. Oncogene, 16:3445–3454, 1998.

49. Brewer, J. W., Hendershot, L. M., Sherr, C. J., and Diehl, J. A.Mammalian unfolded protein response inhibits cyclin D1 translation andcell-cycle progression. Proc. Natl. Acad. Sci. USA, 96: 8505–8510, 1999.

50. Brostrom, C. O., Prostko, C. R., Kaufmann, R. J., and Brostrom, M. A.Inhibition of translational initiation by activators of the glucose-regulatedstress protein and heat shock protein stress response systems. Role ofthe interferon-inducible double-stranded RNA-activated eukaryotic initia-tion factor 2� kinase. J. Biol. Chem., 271: 24995–25002, 1996.

51. Patel, C. V., Handy, I., Goldsmith, T., and Patel, R. C. PACT, astress-modulated cellular activator of interferon-induced double-strandedRNA-activated protein kinase, PKR. J. Biol. Chem., 275: 37993–37998,2000.


52. Harding, H. P., Zhang, Y., Bertolotti, A., Zeng, H., and Ron, D. Perk isessential for translational regulation and cell survival during the unfoldedprotein response. Mol. Cell, 5: 897–904, 2000.

53. Aktas, H., Fluckinger, R., Acosta, J. A., Savage, J. M., Palakurthi,S. S., and Halperin, J. A. Depletion of intracellular Ca2� stores, phospho-rylation of eIF2�, and sustained inhibition of translation initiation mediatethe anticancer effects of clotrimazole. Proc. Natl. Acad. Sci. USA, 95:8280–8285, 1998.

54. Samuel, C. E. The eIF-2� protein kinases, regulators of translation ineukaryotes from yeasts to humans. J. Biol. Chem., 268: 7603–7606, 1993.

55. Samuel, C. E. Reoviruses and the interferon system. Curr. Topics.Microbiol. Immunol., 233: 125–145, 1998.

56. Odani, N., Negishi, M., Takahashi, S., Kitano, Y., Kozutsumi, Y., andIchikawa, A. Regulation of BiP gene expression by cyclopentenone pros-taglandins through unfolded protein response element. J. Biol. Chem.,271: 16609–16613, 1996.

57. Kim, S. H., Forman, A. P., Mathews, M. B., and Gunnery, S. Humanbreast cancer cells contain elevated levels and activity of the proteinkinase, PKR. Oncogene, 19: 3086–3094, 2000.

58. Samuel, C. E. Antiviral actions of interferons. Clin. Microbiol. Rev., 14:778–809, 2001.

59. McClay, E. F., Winski, P. J., Jones, J. A., Jennerette, J., III, andGattoni-Celli, S. �12-Prostaglandin-J2 is cytotoxic in human malignanciesand synergizes with both cisplatin and radiation. Cancer Res., 56: 3866–3869, 1996.

60. Kawahito, Y., Kondo, M., Tsubouchi, Y., Hashiramoto, A., Bishop-Bailey, D., Inoue, K., Kohno, M., Yamada, R., Hla, T., and Sano, H.

15-Deoxy-�12,14-PGJ2 induces synoviocyte apoptosis and suppressesadjuvant-induced arthritis in rats. J. Clin. Investig., 106: 189–197, 2000.

61. Xiong, Y., Connolly, T., Futcher, B., and Beach, D. Human D-typecyclin. Cell, 65: 691–699, 1991.

62. Zhou, Y., and Lee, A. S. Mechanism for the suppression of themammalian stress response by genistein, an anticancer phytoestrogenfrom soy. J. Natl. Cancer Inst., 90: 381–388, 1998.

63. Hunt, C., and Morimoto, R. I. Conserved features of eukaryotic hsp70genes revealed by comparison with the nucleotide sequence of humanhsp70. Proc. Natl. Acad. Sci. USA, 82: 6455–6459, 1985.

64. Gunning, P., Ponte, P., Okayama, H., Engel, J., Blau, H., and Kedes,L. Isolation and characterization of full-length cDNA clones for human �-,�-, and �-actin mRNAs: skeletal but not cytoplasmic actins have anamino-terminal cysteine that is subsequently removed. Mol. Cell. Biol., 3:787–795, 1983.

65. Abraham, N., Stojdl, D. F., Duncan, P. I., Methot, N., Ishii, T., Dube,M., Vanderhyden, B. C., Atkins, H. L., Gray, D. A., McBurney, M. W.,Koromilas, A. E., Brown, E. G., Sonenberg, N., and Bell, J. C. Character-ization of transgenic mice with targeted disruption of the catalytic domainof the double-stranded RNA-dependent protein kinase, PKR. J. Biol.Chem., 274: 5953–5962, 1999.

66. Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. Proteinmeasurement with the Folin phenol reagent. J. Biol. Chem., 193: 265–275,1951.

67. Rousseau, D., Kaspar, R., Rosenwald, I., Gehrke, L., and Sonenberg,N. Translation initiation of ornithine decarboxylase and nucleocytoplasmictransport of cyclin D1 mRNA are increased in cells overexpressing eu-karyotic initiation factor 4E. Proc. Natl. Acad. Sci. USA, 93: 1065–1070,1996.


translational regulation of cyclin d1 by 15-deoxy-...

Documents