Protein kinase C delta inhibits Caco-2 cell proliferation by selective changes in cell cycle and cell death regulators

ORIGINAL ARTICLE

Protein kinase C delta inhibits Caco-2 cell proliferation by selective

changes in cell cycle and cell death regulators

SR Cerda1, R Mustafi1, H Little1, G Cohen1, S Khare2, C Moore1, P Majumder1 andM Bissonnette1

1Department of Medicine, Division of Gastroenterology, University of Chicago, Chicago, IL, USA and 2Department of Medicine,Division of Gastroenterology, Loyola University Medical Center, Maywood, IL, USA

PKC-d is a serine/threonine kinase that mediates diversesignal transduction pathways. We previously demon-strated that overexpression of PKC-d slowed the G1progression of Caco-2 colon cancer cells, acceleratedapoptosis, and induced cellular differentiation. In thisstudy, we further characterized the PKC-d dependentsignaling pathways involved in these tumor suppressoractions in Caco-2 cells overexpressing PKC-d using aZn2þ inducible expression vector. Consistent with a G1arrest, increased expression of PKC-d caused rapid andsignificant downregulation of cyclin D1 and cyclin Eproteins (50% decreases, Po0.05), while mRNA levelsremained unchanged. The PKC agonist, phorbol12-myristate 13-acetate (TPA, 100 nM, 4 h), inducedtwo-fold higher protein and mRNA levels of p21Waf1, acyclin-dependent kinase (cdk) inhibitor in PKC-d trans-fectants compared with empty vector (EV) transfectedcells, whereas the PKC-d specific inhibitor rottlerin (3 lM)or knockdown of this isoenzyme with specific siRNAoligonucleotides blocked p21Waf1 expression. Concomi-tantly, compared to EV control cells, PKC-d upregulationdecreased cyclin D1 and cyclin E proteins co-immunopre-cipitating with cdk6 and cdk2, respectively. In addition,overexpression of PKC-d increased binding of cdkinhibitor p27Kip1 to cdk4. These alterations in cyclin-cdksand their inhibitors are predicted to decrease G1 cyclinkinase activity. As an independent confirmation of thedirect role PKC-d plays in cell growth and cell cycleregulation, we knocked down PKC-d using specific siRNAoligonucleotides. PKC-d specific siRNA oligonucleotides,but not irrelevant control oligonucleotides, inhibited PKC-d protein by more than 80% in Caco-2 cells. Moreover,PKC-d knockdown enhanced cell proliferation (B1.4-2-fold, Po0.05) and concomitantly increased cyclin D1 andcyclin E expression (B1.7-fold, Po0.05). This was aspecific effect, as nontargeted PKC-f was not changed byPKC-d siRNA oligonucleotides. Consistent with acceler-ated apoptosis in PKC-d transfectants, compared to EVcells, PKC-d upregulation increased proapoptotic regu-

lator Bax two-fold at mRNA and protein levels, whileantiapoptotic Bcl-2 protein was decreased by 50% at apost-transcriptional level. PKC-d specific siRNA oligonu-cleotides inhibited Bax protein expression by more than50%, indicating that PKC-d regulates apoptosis throughBax. Taken together, these results elucidate two criticalmechanisms regulated by PKC-d that inhibit cell cycleprogression and enhance apoptosis in colon cancer cells.We postulate these antiproliferative pathways mediate animportant tumor suppressor function for PKC-d in coloniccarcinogenesis.Oncogene (2006) 25, 3123–3138. doi:10.1038/sj.onc.1209360;published online 23 January 2006

Keywords: PKC; cyclins; p21Waf1; Bcl-2; Bax; coloncancer cells

Introduction

Colorectal cancer is the third leading cause of cancer-related deaths in the United States (Greenlee et al.,2000). The molecular events in colon cancer involve astepwise accumulation of activating mutations in proto-oncogenes, such as K-ras, and inactivating mutations oftumor suppressor genes, such as APC (Fearon andVogelstein, 1990), together with nonmutational changesin many effectors including protein kinase C (PKC). ThePKC family regulates colonic crypt cell fate, withproliferation at the base, progressive maturation alongthe ascending crypt axis and apoptosis at the colonicluminal surface (Shanmugathasan and Jothy, 2000). ThePKC family is comprised of at least 11 isoforms(Nishizuka, 1995). Members of this family are classifiedinto three groups, distinguished by differences insequence homology and cofactor requirements foractivation (Newton, 1997). While all isoforms requireacidic phospholipids (phosphatidylserine), the conven-tional PKCs (a, bI, bII, and g) are activated by calcium(Ca2þ ) and 1,2-diacyl-sn-glycerol (DAG), whereas thenovel class of PKCs (d, e, y, and Z) are Ca2þ -independent, but DAG-dependent. The atypical PKCs(l, z, and i) are not regulated by Ca2þ or DAG (Liu andHeckman, 1998). In addition to these structural andregulatory differences, PKC isoforms exhibit distinct

Received 20 May 2005; revised 28 November 2005; accepted 28November 2005; published online 23 January 2006

Correspondence: Dr SR Cerda, Department of Medicine, TheUniversity of Chicago, MC 4076, 5841 S. Maryland Avenue, Chicago,IL 60637, USA.E-mail: [email protected]

Oncogene (2006) 25, 3123–3138& 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00

www.nature.com/onc

patterns of tissue expression and subcellular localiza-tion, supporting the current belief that these kinasesmediate unique cell-specific functions (Wetsel et al.,1992; Newton, 1997).

In normal colonic mucosa, the expression levels ofseveral PKC isoforms follow an increasing gradientfrom the base of the crypt to the mucosal surface,paralleling crypt cell maturation (Black, 2001). WhilePKC signaling regulates postmitotic events duringnormal intestinal epithelial cell maturation (Saxonet al., 1994), derangements in PKC isoforms are alsothought to contribute to colonic malignant transforma-tion. For example, PKC a, bI, d and z are decreased atdifferent stages of neoplastic progression (Weinstein,1991; McGarrity and Peiffer, 1994; Kahl-Rainer et al.,1994, 1996; Brasitus and Bissonnette, 1998; Verstovseket al., 1998; Black, 2001). These changes in specificisoforms suggest that PKC signal dysregulations con-tribute to neoplastic transformation. Consistent with apotential tumor suppressor role for PKC-d, a number ofstudies have reported that several proto-oncogenestransform cells, at least in part, by causing the loss ofPKC-d activity (Geiges et al., 1995; Li et al., 1996, 1998;Lu et al., 1997; Zang et al., 1997). Conversely, over-expression of PKC-d inhibited the transformed pheno-type of src-overexpressing rat colonic epithelial cells,causing an arrest in cellular proliferation (Perletti et al.,1999). In vitro overexpression of PKC-d has beenassociated with decreased growth and enhanced differ-entiation and cell death in several noncolonic cell lines(Gschwendt, 1999; Kikkawa et al., 2002; Brodie andBlumberg, 2003; Jackson and Foster, 2004). Our ownstudies, moreover, have established that PKC-d regu-lates the growth phenotype in Caco-2 colon cancer cells(Cerda et al., 2001). This human colon cancer cell linehas been extensively studied as a model of intestinalepithelial cell biology (Pinto et al., 1983; Evers et al.,1996; Abraham et al., 1998; Ding et al., 1998; Scaglione-Sewell et al., 1998). Using stable transfectants withcDNA coding for full-length PKC-d, we examined thebiological consequences of alterations in the expressionof PKC-d on the neoplastic phenotype of these cellsusing two different Zn2þ -inducible metallothioneineukaryotic expression vectors (Cerda et al., 2001). Weshowed that PKC-d upregulation in Caco-2 cells, acondition more closely resembling nontransformedcolonocytes, attenuated the transformed phenotype.Effects of PKC-d upregulation included inhibitedanchorage-dependent and -independent growth, G1 cellcycle slowing, increased differentiation, and enhancedapoptosis. Based on our studies in these cells, theobserved loss of PKC-d in colon cancers would appearto provide a clear growth advantage for transformingcolonocytes and contribute to the dysregulated growthcharacterizing colonic carcinogenesis (Black, 2001).

In agreement with our prior studies showing PCK-dslowed G1 progression, there is increasing evidence thatPKC regulates the cell cycle through multiple pathways(Arita et al., 1998; Toyoda et al., 1998; Besson andYong, 2000; Graham et al., 2000; Zhu et al., 2000;Shanmugam et al., 2001; Wakino et al., 2001; Acevedo-

Duncan et al., 2002; Kikkawa et al., 2002; La Portaet al., 2002; Lee et al., 2002; Lin et al., 2002; Brodie andBlumberg, 2003; Soh and Weinstein, 2003; Avazeriet al., 2004; Clark et al., 2004; Gavrielides et al., 2004;Leontieva and Black, 2004; Atten et al., 2005). Innontransformed rat intestinal epithelial crypt cells, PKCsignaling mediates cell cycle exit through rapid down-regulation of G1 cyclins and increased expression ofCip/Kip cyclin-dependent kinase inhibitors (Frey et al.,1997, 2000; Clark et al., 2004). Studies in noncoloniccellular systems have demonstrated that PKC-d inhibitsproliferation by suppressing G1 cyclin expression andincreasing expression of cyclin-dependent kinase (cdk)inhibitors (Weinstein, 1991; Fukumoto et al., 1997;Toyoda et al., 1998; Ashton et al., 1999; Shanmugamet al., 2001; Page et al., 2002; Nakagawa et al., 2005).However, the effects of this isoform on the growthphenotype depend upon the cellular context. In rat 3Y1fibroblasts, for example, PKC-d promotes cell cycleprogression in late G1, in addition to negativelyregulating M phase entry (Kitamura et al., 2003). Incontrast, in rat thyroid cells, this isoenzyme stimulatesapoptosis by initiating G1 phase cell cycle progressionwhile arresting cells in S phase (Santiago-Walker et al.,2005). The role of PKC-d in cell death has also beeninvestigated in many cellular systems (Kikkawa et al.,2002; Brodie and Blumberg, 2003; Wang et al., 2003;Jackson and Foster, 2004). Studies have demonstratedthat PKC-d promotes cell death in many noncolonic celllines, including keratinocytes (Denning et al., 1998),neutrophils (Pongracz et al., 1999), cerebral granule cells(Villalba, 1998), HL-60 and prostate cancer cells(Savickiene et al., 1999; Yin et al., 2005), but not inbreast cancer cells (McCracken et al., 2003; Nabha et al.,2005). In studies in colonic cells, PKC signaling path-ways are also involved in cell cycle and cell deathregulation (Goldstein et al., 1995; Qiao et al., 1996;Sauma et al., 1996; Abraham et al., 1998; Perletti et al.,1998, 1999; Andre et al., 1999; Assert et al., 1999; Welleret al., 1999; Umar et al., 2000; Cerda et al., 2001; Linet al., 2002; McMillan et al., 2003; Di Mari et al., 2005;Meyer et al., 2005). Several studies, including our owninvestigations, have demonstrated a tumor suppressorrole for PKC-d in colon cancer cells, including cell cyclearrest and enhanced apoptosis (Weller et al., 1999;Cerda et al., 2001; Lin et al., 2002; McMillan et al.,2003; Perletti et al., 2004, 2005). There is, however,at least one report that inhibition of this isoenzymeis sufficient to promote cell death (Lewis et al., 2005).Moreover, recent studies indicate that PKC-d canmediate antiapoptotic signals in some colon cancer cells(Wang et al., 2004, 2006). Thus, PKC-d signaling candrive growth promotion or inhibition, and enhance orinhibit apoptosis in cell context-specific manners.

We previously showed that PKC-d causes a cell cyclearrest in G1 and enhances apoptosis in Caco-2 coloncancer cells (Cerda et al., 2001). While the downstreamtargets that mediate these tumor suppressor actions ofPKC-d are currently not well recognized in colon cancercells, we speculated that they involve key regulatoryelements of cell cycle control and apoptosis. To further

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3124

Oncogene

elucidate the mechanisms controlling these processes, inthis study, we directly examined G1 cell cycle and celldeath regulators that might contribute to the antipro-liferative actions of PKC-d in Caco-2 cells. These cellsdo not contain mutations in c-K-ras or c-src, upstreamregulators of PKC that might otherwise confoundmechanistic interpretations of effects observed followingPKC-d overexpression. We identified PKC-d mediatedalterations in several important regulators involved incell cycle and cell death controls in Caco-2 cells, whichwould be expected to attenuate the malignant phenotypeof these cells. These changes include downregulation ofcyclin D1 and cyclin E expression, and upregulation ofcyclin cdk inhibitor, p21Waf1. These alterations wereaccompanied by decreased cyclins and increased cdkinhibitors co-associating with G1 cdks. Changes inapoptosis regulators included decreases in antiapoptoticBcl-2 and increases in proapoptotic Bax. Using siRNAoligonucleotide strategies, as an independent strategy,we also demonstrate a direct role for PKC-d in thedownregulation of cyclin D1 and cyclin E expression,and upregulation of Bax. These results, and a discussionof their implications in colonic carcinogenesis, form thebasis for the current study.

Results

In previous studies we demonstrated overexpression ofPKC-d inhibited proliferation, enhanced differentiationand accelerated apoptosis in Caco-2 colon cancer cells(Cerda et al., 2001). The inhibited proliferation inducedby PKC-d involved both a selective cell cycle arrest inG1 and enhanced apoptosis (Cerda et al., 2001). In thisstudy, we further characterized the signaling pathwaysinvolved in the tumor suppressor actions of thisisoenzyme in Caco-2 cells transfected with PKC-d underthe control of a Zn2þ inducible metallothionein promo-ter.

Activation of PKC-d inhibits cellular proliferation andcauses cell cycle slowing in G0/G1 and increased cell deathin Caco-2 cellsTo extend our previous studies and further elucidate therole PKC-d plays in control of cell cycle and cell death incolon cancer cells, we initially investigated the functionof this isoenzyme in control of Caco-2 colonic cellproliferation. As shown in Figure 1, PKC-d over-expression caused a significant 30% decrease in cellproliferation relative to empty vector (EV) transfectedcells (Po0.05, n¼ 6). We then treated cells with 12-O-tetradecanoylphorbol 13-acetate (TPA), a phorbol ester,known to activate classical and novel PKC isoforms(Castagna et al., 1982; Ryves et al., 1991), or withBistratene A (BisA), a PKC-d specific activator (Watterset al., 1998). Treatment of EV transfected Caco-2 cellswith TPA or with BisA for 48 h, significantly inhibitedcellular proliferation by 33 and 29%, respectively,compared with DMSO (vehicle treated) controls(Po0.05, n¼ 6). When Caco-2 cells overexpressing

PKC-d were treated with these agonists, cellularproliferation was further significantly inhibited at 24and 48 h. By 48 h, PKC-d agonists TPA and BisAinhibited cellular proliferation by 50 and 64%, respec-tively, compared with untreated EV controls (Po0.01,n¼ 6). These agonists caused similar effects on cellproliferation as assessed by BrdU incorporation (datanot shown). As an independent confirmation of thedirect role PKC-d plays in Caco-2 cell proliferation, weknocked down this isoenzyme using siRNA oligonu-cleotides that specifically targeted human PKC-d (Irieet al., 2002). These siRNAs have successfully inhibitedPKC-d mRNA by more than 80% in both HEK-293and HeLa cells (Irie et al., 2002) and also inhibitedPKC-d in Caco-2 cells (see below, Figure 6). As shownin Figure 2, knockdown of PKC-d with specific, but notirrelevant random sequence oligonucleotides (control),increased Caco-2 cell proliferation by 140 and 225%, asdetermined by DNA labeling and mitochondrial activityrespectively, compared to control cells (Po0.05). This isconsistent with our previous results that PKC-dupregulation (and kinase activation) inhibited Caco-2cell proliferation.

We previously demonstrated that PKC-d overexpres-sion in the absence of Caco-2 cell treatments causeda 20% increase in the number of Caco-2 cells in theG0/G1 phase of the cell cycle (Cerda et al., 2001). In thecurrent study, we examined the effects of TPA or BisAon transfected Caco-2 cells. Caco-2 transfectants weretrypsinized, and replated at lower confluency to induce alarge fraction to enter S phase. As shown in Table 1, cellcycle analysis by flow cytometry revealed that within15 h of treatment, both TPA and BisA significantlyincreased the number of PKC-d transfected Caco-2 cells

0

0.2

0.4

0.6

0.8

1

1.2

EV EV TPA EVBISA

Cel

l Pro

lifer

atio

n

24hr

48hr

‡

‡

*

PKC-δ PKC-δ PKC-δTPA BISA

* **

†

†

Figure 1 PKC-d inhibits Caco-2 cell proliferation. Caco-2 and EVtransfectants, plated on 96-well microtiter tissue culture plates,were treated with 100 nM PKC agonist, phorbol 12-myristate 13-acetate (TPA), 100 nM Bistratene A (BISA), or DMSO (vehicle),for 24 h (closed bars) and 48 h (open bars) in the presence of 175mM

Zn2þ . The y-axis depicts mean absorbance values (Abs 450 nm) foreach treatment, normalized to empty vector (EV) controls at 24and 48 h, respectively; x-axis shows indicated treatment groups.Values are means7s.d. of six identically treated wells. Data arefrom a representative experiment repeated two times. *Po0.05,compared with time matched vehicle-treated EV controls. wPo0.05and zPo0.01, compared with time-matched TPA- or BISA-treatedEV cells.

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3125

Oncogene

in G0/G1 to 54.073.4 and 54.672.2%, respectively,compared with 45.273.0 and 46.37.2% for EVcontrols (Po0.05, n¼ 4). Concomitantly, there was asignificant decrease in the number of cells in S phase,indicating that PKC-d slows G1-S transition in the cellcycle. By 24 h of phorbol ester treatment, TPA increasedthe number of EV cells in S phase, compared with PKC-d transfectants (compare 48.471.4% for EV cells with40.0273.6% for PKC-d cells). Although in some cellsphorbol esters, which can activate multiple isoforms ofPKC, promote growth by accelerating G1-S transi-

tions (Livneh and Fishman, 1997; Fishman et al., 1998;Black, 2000), we attribute TPA-induced G1 slowing inour Caco-2 transfectants to the activation of PKC-d,since the nontargeted isoforms did not differ inexpression among parental, EV, and PKC-d transfec-tants (Cerda et al., 2001). By 48 h of phorbol estertreatment, there was decreased proliferation in EV andPKC-d transfectants that we speculate arises fromdownregulation of growth-promoting PKC isoformsthat result in an overall decrease in cellular proliferation(Figure 1). We then confirmed and extended ourprevious studies that showed PKC-d overexpressionincreased programmed cell death in Caco-2 cells (Cerdaet al., 2001). As shown in Table 1, PKC-d transfectedcells exhibited significantly increased nuclear condensa-tion and fragmentation, compared with EV controls(6.570.9% compared with 4.070.5% for EV cells;Po0.05, n¼ 4). Consistent with increased activation ofPKC-d in Caco-2 cells overexpressing this isoenzyme,treatment with PKC agonists BisA or TPA furtherenhanced apoptosis in these cells, compared to EV cells(10.070.9 and 12.270.3%, respectively for PKC-d, and5.170.8 and 7.8570.2% for EV cells, Po0.05).Together, these studies suggest that PKC-d inhibitscellular proliferation in Caco-2 cells via two distinctpathways, slowing cell cycle progression in G0/G1 andincreasing the apoptotic rate of these cells.

PKC-d decreases in cyclin D1 and cyclin E, contributingto the G1 arrest in Caco-2 cellsThe cell cycle is regulated by the coordinated actions ofcyclin-dependent kinases, phosphatases and cell cycleinhibitors (King et al., 1996). The major cyclin-cdkcomplexes involved in cell cycle progression from the G1to the S phase are cyclin D-cdk4, 6 and cyclin E-cdk2(Sherr and Roberts, 1999). We next asked whetheralterations in these important G1 phase regulators mightmediate the PKC-d-dependent G1 slowing. In othercells, PKC-isoforms, -a, -d, -e, and -Z, are known toregulate G1 cyclin expression (Fima et al., 2001; Sohand Weinstein, 2003). To analyse the expression of thesecyclins, whole-cell lysates were prepared from EV and

0

0.5

1

1.5

2

2.5

3

CTL (50 nM) 50 nM 100 nM PKC- δ siRNA

Cel

l Pro

lifer

atio

n (

Wst

-1)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

Cel

l Pro

lifer

atio

n (

Brd

U)

*

*

‡

Figure 2 PKC-d knockdown increases proliferation in CaCo-2cells. CaCo-2 cells, plated on 96-well microtiter tissue cultureplates, were treated with 50 or 100 nM of PKC-d specific RNAi 21-nt oligonucleotide duplex, or scrambled control (50 nM), usingsiPORT transfection reagent. Cell proliferation was measured after48 h by mitochondrial activity (open bars; Wst-1 assay) and BrdUincorporation (closed bars). The y-axis depicts mean absorbancevalues, Wst-1 (y1, Abs 450 nm) and BrdU (y2, Abs 370 nm), foreach treatment normalized to scrambled treated controls (CTL).The x-axis shows treatment groups. Each value is calculated from10 identically treated wells (including controls). Control cells weretreated with 50 or 100 nM scrambled oligonucleotide. There wereno differences in proliferation between 50 and 100nM scrambledoligonucleotide. Data are from a representative experimentrepeated two times. *,zPo0.05, compared with appropriatecontrols (50 or 100 nM oligonucleotide treated).

Table 1 Effects of PKC-d activation on cell cycle and cell death distributions in Caco-2 Cells

15 h 24 h 48 h

%G0/G1 %S %G2/M %G0/G1 %S %G2/M % Apoptosis

EV 43.5373.3 40.3672.3 16.1171.1 43.6672.2 41.8973.2 14.4572.2 4.070.5EV TPA 45.2873.0 40.6575.2 14.0772.2 36.6470.8 48.4271.4* 14.9472.3 5.170.8EV BisA 46.3174.2 35.3675.4 18.3471.2 42.8272.8 37.7972.0 19.3971.6 7.8570.2*PKC-d 42.3270.1 38.3171.6 19.3771.7 43.3171.7 36.7570.7 19.9472.0 6.570.9*PKC-d TPA 54.0473.4*,w 31.42+0.7*,w 14.5472.7 45.4274.4 40.273.6 14.3871.3 10.0270.8*,w

PKC-d BisA 54.6372.2*,w 30.12+0.2*,w 15.2572.4 48.1271.0 32.96+1.5 18.9271.41 12.2170.3*,w

Caco-2 cells, transfected with empty vector (EV) or full-length human PKC-d, were treated for 15–24 h with 100 nM TPA, 100 nM BisA, or DMSO(vehicle) in the presence of zinc. Cells were fixed and analysed for their distribution in the cell cycle using a FacScan flow cytometer and CellQuestsoftware. Distribution of DNA in the cell cycle was determined using Modfit LT software. Apoptotic rates were measured after a period of 48 h.Treated cells were fixed, and nuclei were stained by DAPI and visualized by microscopy. Rates of apoptosis are expressed as a percentage ofapoptotic nuclei, by counting 1000 random cells in duplicate platings per experiment. Values are reported as means of four samples7s.d. *Po0.05,compared with EV transfectants treated with DMSO (vehicle). wPo0.05, compared with PKC-d cells treated with DMSO.

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3126

Oncogene

PKC-d transfectants treated with phorbol esters for 0–8 h. Changes in PKC-d phosphorylation were assessedas surrogate measures of PKC-d kinase activation.Using phospho-specific PKC-d antibodies and Westernblotting, we probed phosphorylation at the Thr505

(activation loop) and Ser643 (autophosphorylation) sitesof PKC-d, which are important for kinase activation (Liet al., 1997; Parekh et al., 1999; Srivastava et al., 2002;Rybin et al., 2003; Steinberg, 2004; Seki et al., 2005).Western blot analysis of whole-cell extracts in Figure 3demonstrates upregulation of PKC-d protein in PKC-dtransfectants compared with EV cells. Futhermore, TPAdecreased levels of PKC-d expression (Figure 3, upperpanel, compare lanes 5–8 for PKC-d with 1–4 for EVcells). As shown in Figure 3, PKC-d transfectants alsoexhibited increased levels of phospho-Ser643 and phos-pho-Thr505 PKC-d, compared with EV transfectants(Figure 3, lane 5 vs 1). These levels diminished upontreatment with TPA but remained high, compared withundetectable levels in EV transfectants (Figure 3, lanes6–8 vs 2–4). These results are in agreement with ourprevious findings of increased PKC-d kinase activity inPKC-d transfectants (Cerda et al., 2001). As shown inFigure 4, in cells overexpressing PKC-d, Western blotanalysis of whole-cell extracts demonstrated that expres-sion levels of cyclin D1 and cyclin E proteins weredecreased by approximately 60 and 50%, respectively,

as compared to EV controls (note: cyclin D1 and cyclinE at zero time point, lanes 1 (EV) and 5 (PKC-d)). Incontrast to PKC-d transfectants, where cyclin D1remained unchanged and cyclin E went down, TPAtreatment enhanced cyclin D1 and cyclin E expression1.5- and 2.0-fold, respectively, in EV cells (Figure 4, lane1 vs 3 and 4). This is consistent with phorbol estersaccelerating G1-S transitions in EV cells, comparedwith PKC-d transfectants (Table 1). We speculate thatincreased PKC-d signaling suppresses cyclin D1 andcyclin E in PKC-d transfectants, over-riding cyclinupregulation by nontargeted PKC isoforms, as seen inTPA-treated EV cells.

In contrast to PKC-d dependent decreases in cyclinD1 and cyclin E proteins, however, there were nodifferences in cyclin D1 and cyclin E mRNA expressionlevels between untreated EV control and PKC-d over-expressing Caco-2 cells, as assessed by quantitative real-time PCR (Figure 5). These results suggest that thechanges induced in cyclin D1 and cyclin E proteinexpression levels by PKC-d involve post-transcriptionalmechanisms. Treatment with TPA for 3 h enhancedcyclin D1 and cyclin E mRNA expression 1.8- and 2.3-fold in EV cells, respectively, and 1.6- and 1.7-fold inPKC-d transfectants (Figure 5). Thus, we speculate thatone of the mechanisms by which PKC-d inhibits G1 cellcycle progression is by suppressing expression of cyclinD1 and E post-transcriptionally. Additionally, treat-ment with TPA partially downregulated cyclin E mRNAlevels in PKC-d transfectants compared with EV cells,1.7- vs 2.3-fold respectively, suggesting that upregulatedPKC-d also suppresses TPA induction of cyclin E at atranscriptional level.

To more specifically address the effects of PKC-d inregulation of these cyclins, we knocked down this

+ TPA+ TPA

0 2 4 8 0 2 4 8

EV PKC-δ

Cyclin E

Cyclin D1

β-actin

1 2 3 4 5 6 7 8

Time (hrs)

36

50

42

kDa

Figure 4 PKC-d downregulates cyclin D1 and cyclin E in Caco-2cells. Whole-cell lysates were prepared from EV and PKC-dtransfected cells treated with 100 nM TPA or DMSO (vehicle) forthe indicated times (0–8 h). Proteins (20mg/lane) were probed byWestern blotting for cyclin D1, cyclin E and b-actin expression. Arepresentative blot of two independent platings is shown.

0

0.5

1

1.5

2

2.5

3

CTL TPA CTL TPA

mR

NA

Exp

ress

ion

(AU

)

EV

PKC-δ

*

*

**

Cyclin D1

†

Cyclin E

Figure 5 Levels of cyclin D1 and cyclin E mRNA are not alteredby PKC-d in Caco-2 cells. Preconfluent Caco-2 cells, transfectedwith empty vector (EV) or full-length human PKC-d (PKC-d),treated with 100 nM TPA or DMSO (vehicle), were assessed forcyclin D1 and cyclin E mRNA by real-time PCR. Real-time PCRanalysis depicts levels of cyclin D1 and cyclin E mRNA expressionin arbitrary units, normalized to GAPDH. Data are means7s.d. oftwo independent experiments in triplicate. *Po0.05, comparedwith their respective transfectants controls. wPo0.05, comparedwith TPA-treated EV cells.

+ TPA + TPA

0 2 4 8 0 2 4 8 Time (hrs)PKC-δ 78 kDa

p-Ser643PKC-δ

p-Thr505PKC-δ

β-actin 42 kDa

7654321 8

EV PKC-δ

Figure 3 PKC-d overexpression increases phosphoPKC-d. Whole-cell lysates were prepared from EV and PKC-d transfected cells,treated with 100 nM TPA or DMSO (vehicle) for the timesindicated (0–8 h). Proteins (20mg/lane) were probed by Westernblotting for PKC-d, phosphoPKC-d (Ser643/676) and phos-phoPKC-d (Thr505) expression and b-actin as loading control. Arepresentative blot of two independent platings is shown.

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3127

Oncogene

isoenzyme in Caco-2 cells overexpressing PKC-d, usingsiRNA oligonucleotides that specifically target expressionof human PKC-d (Irie et al., 2002). As shown in Figure 6,PKC-d knockdown with specific siRNA oligonucleotides,but not irrelevant random sequence oligonucleotides(control), inhibited expression of PKC-d by more than80%. PKC-d knockdown concomitantly caused a1.670.3- and 1.870.6-fold increase in cyclin D1 andcyclin E expression, respectively (Figure 6, compare lanes1 and 2 with lanes 3 and 4), consistent with our previousresults that upregulation of PKC-d inhibited the expres-sion of these cyclins. This was a specific effect, as highlyhomologous but nontargeted PKC-z was not changed byPKC-d siRNA oligonucleotides. b-Actin was also un-affected by PKC-d siRNA and equivalent b-actin bandsconfirmed comparable protein loading.

p21Waf1 is a downstream mediator of PKC-d in Caco-2cellsDuring cell division, cyclin-dependent kinases (cdks) aresequentially activated through increased levels of cyclinsand decreased levels of cdk inhibitors (King et al., 1996).We next evaluated the role of cdk inhibitors, p21Waf1 andp27Kip1, in the PKC-d induced cell cycle arrest in Caco-2cells. As shown in Figure 7a, p21Waf1 protein wasexpressed at very low levels in both EV and PKC-dtransfectants (note: p21Waf1 at zero time point, lanes 1and 5). By 4 h the phorbol ester, TPA, inducedexpression of this cyclin dependent kinase (cdk)inhibitor (Figure 7a, compare lanes 1 with 3 and lane5 with 7). This TPA induction of p21Waf1 protein wassignificantly greater in PKC-d transfectants (B2-fold),compared to EV controls (Figure 5a, lanes 7 and 8 vs 3and 4). Increased p21Waf1 protein was also accompaniedby significantly increased p21Waf1 mRNA, as assessed byquantitative real-time PCR (Figure 7b). Within 3 h oftreatment, TPA increased p21Waf1 mRNA expressionlevels B5-fold in PKC-d transfectants, compared with

TPA treated EV control cells (Po0.05, n¼ 6), indicatingthat PKC-d drives increases in p21Waf1 mRNA expres-sion. There were no significant differences in expressionlevels of p27Kip1, but there were important differences inp27Kip1–cdk4 co-association (see below for changes inp27Kip1 co-association). Taken together, these datasuggest that reductions induced by PKC-d in cyclin D1and cyclin E involve post-transcriptional mechanisms,whereas increased p21Waf1 expression involves transcrip-tional or mRNA stabilization mechanisms.

To further assess if p21Waf1 is a downstream mediatorof PKC-d in Caco-2 cells, we employed rottlerin, anisoform-specific PKC inhibitor (Gschwendt et al., 1994).Rottlerin is a specific inhibitor of PKC-d in the rangeof 3–6 mM, whereas higher concentrations (30–100 mM)are required to inhibit other PKC isoforms (Gschwendtet al., 1994). As shown in Figure 8a, the cyclin-dependent kinase inhibitor p21Waf1 is induced within4 h by the PKC agonist phorbol 12-myristate 13-acetate(TPA, 100 nM) (Figure 8a, compare lanes 1 and 2).Rottlerin pretreatment for 1 h blocked p21Waf1 inductionby TPA, suggesting that p21Waf1 is a downstreammediator of PKC-d (Figure 8a, compare lanes 2 and

PKC-ζ

PKC-δ

1 2 3 4

Cyclin D1

Cyclin E

β-actin

SS PKC-δ siRNAsiRNA

PKC-δ CF

kDa

78

40

36

50

76

42

Figure 6 PKC-d knockdown increases cyclin D1 and cyclin Eexpression in Caco-2 cells. Caco-2 cells overexpressing PKC-d weretransfected with 100 nM (lanes 3, 4) PKC-d specific RNAi 21-ntduplex, or a scrambled sequence (SS; lanes 1, 2), using siPORTtransfection reagent. After 72 h whole-cell lysates (30mg protein)were assayed for PKC-d, PKC-d catalytic fragment (PKC-dCF),cyclin D1, cyclin E, and PKC-z and b-actin as loading control.Representative blots of two independent experiments in duplicateare shown.

+ TPA

0 2 4 8 0 2 4 8

+ TPA

Time (hrs) kDa

p21 21

β-actin 42

1 2 3 4 5 6 7 8

EV PKC-δ

0

2

4

6

8

10

12

a

b

EV PKC-δ

p21

mR

NA

(A

U)

-TPA +TPA

*

‡

Figure 7 PKC-d increases protein and mRNA levels of cyclin-dependent kinase inhibitor p21Waf1 in Caco-2 cells. (a) Whole-celllysates were prepared from EV and PKC-d transfected cells treatedwith 100 nM TPA or DMSO (vehicle) for the times indicated(0–8 h). Proteins (50 mg/lane) were probed by Western blotting forp21Waf1 and b-actin expression. A representative blot of threeindependent platings is shown. (b) Real-time PCR analysis forlevels of p21Waf1 mRNA expression in arbitrary units. EV andPKC-d transfected cells were treated with 100 nM TPA or DMSO(vehicle) for 3 h. RNA was extracted and real-time PCR performedas described in ‘Materials and methods’. Values were normalized toGAPDH. Data are means7s.d. of two independent experiments intriplicate. *Po0.05 and zPo0.01, compared with DMSO-treatedrespective transfectants.

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3128

Oncogene

4). Because several studies have questioned the mechan-ism of action and specificity of Rottlerin (Davies et al.,2000; Tillman et al., 2003), we carried out additionalexperiments using siRNA technology to knockdownPKC-d expression and confirm that PKC-d induces thiscdk inhibitor. Using this strategy, Nakagawa et al.(2005) recently demonstrated that phorbol ester inducedG1 phase arrest in lung adenocarcinoma cells isselectively mediated by a PKC-d dependent inductionof p21waf1. As shown in Figure 8b, PKC-d knockdownwith specific siRNA oligonucleotides, but not irrelevantoligonucleotides with random sequences (SSsiRNA),inhibited p21Waf1 induction by phorbol esters (Figure 8b,compare lanes 1 and 2 with 3 and 4). These experimentsare broadly in agreement with the results of the rottlerinexperiments.

Altered levels of cyclin D1, cyclin E, p27Kip1and p21Waf1

co-associating with cyclin-dependent kinases (cdk-2, -4,and -6) contribute to the PKC-d induced G1 arrest inCaco-2 cellsTo further elucidate the mechanisms involved in the G1arrest by PKC-d, we analysed cell cycle regulatoryproteins co-associating with cdk-2, -4, and -6. G1-specific kinases are not regulated by changes in kinaseexpression (Sherr, 1996). In agreement with thesereports, in whole-cell lysates, we confirmed that therewere no changes in cdk levels. As shown in Figure 9a,PKC-d transfectants had 50% decreased cyclin E in

cdk2 immunoprecipitates compared to EV control cells.Levels of cyclin D1 co-associating with cdk6 immuno-precipitates were 60% lower in PKC-d transfectants,compared to EV cells (Figure 9c). In contrast, cyclin D1co-associating with cdk4 did not change (Figure 9b).Although we observed no alterations in total expressionof p27Kip1 in lysates, overexpression of PKC-d increasedbinding of p27Kip1 to cdk4 by 30% (Figure 9b), but didnot alter p27Kip1 association with cdk2 or cdk6(Figure 9a and c). Taken together, decreased levels ofcyclin E and cyclin D1 co-associating with cdks-2 and-6, respectively, and increased levels of p27Kip1

co-associating with cdk-4, are expected to reduce kinaseactivities of these cdks and inhibit G1-S progression.

Complexes between p21Waf1 and cdk-4 and -6 werealso analysed in Caco-2 transfectants. As shown inFigure 10, 6 h of TPA treatment increased p21Waf1 co-associating with cdk-4 and -6 in both EV and PKC-dransfectants (note: p21Waf1–cdk-4 and -6 complexes atzero time and 6 h time points, lanes 1 vs 2 and 4 vs 5).Moreover, TPA caused p21Waf1–cdk4 and p21Waf1–cdk6complexes to persist for up to 15 h in Caco-2 cellsoverexpressing PKC-d, but not in EV controls (comparelane 6 with lane 3).

Taken together, our results suggest that PKC-ddecreases G1 cell cycle activators, cyclin D1 and cyclinE, and increases G1 cell cycle inhibitors, p21Waf1 andp27Kip1 co-associating with G1 cdks. These changes,moreover, are expected to decrease the kinase activitiesof cylin E-cdk2 and cyclin D1-cdk4 and cyclin D1-cdk6,and thereby mediate the G1 cell cycle slowing inducedby PKC-d overexpression in Caco-2 cells.

PKC-d accelerates apoptosis by downregulating Bcl-2and inducing Bax expressionIn prior studies, we have shown (Cerda et al., 2001) andconfirmed in the current studies (Table 1) that PKC-d

SS

β-actin

PKC-δ

p21Waf1

β-actin

p21Waf1

1 2 3 4

1 2

a

b

3 4

PKC-δ siRNAsiRNA kDa

78

PKC-δ CF 40

21

42

+TP

A

+TP

A

21 kDa

42 kDa

Rottlerin − − + +

Figure 8 PKC-d mediates p21Waf1 induction by TPA in Caco-2cells. (a) Caco-2 transfectants were pretreated with 3 mM rottlerin orvehicle for 1 h and then treated with 100 nM TPA or DMSO(vehicle) for 3 h. Proteins (50mg/lane) were probed for p21Waf1

expression by Western blotting. b-Actin was probed as loadingcontrol. (b) Caco-2 cells overexpressing PKC-d were transfectedwith 100 nM (lanes 3, 4) PKC-d specific RNAi 21-nt duplex, or100 nM scrambled sequence (SS; lanes 1,2), using siPORTtransfection reagent. After 72 h, cells were treated with 100 nM

TPA for 3 h and whole-cell lysates (50mg protein) were assayed forPKC-d, PKC-dCF, and p21Waf1 expression, as well as b-actinexpression. A representative blot of two independent platings isshown.

EV PKC-δ EV PKC-δ

EV PKC-δ

Cyclin E

IP:cdk2

p27kip1

Cyclin D1

p27kip1

Cyclin D1

p27kip1

IP:cdk4

IP:cdk6

a b

c

Figure 9 PKC-d overexpression decreases the cyclins andincreases p27Kip1 co-associating with cdks. (a–c) Cdk2, 4, and 6(panels a, b, and c, respectively) were immunoprecipitated fromtotal cell extracts (200mg) from synchronized empty vector (EV)and PKC-d transfected cells. Indicated co-associating proteins weredetected by Western blotting. Western blots that are representativeof two independent experiments in duplicate are shown. Note thedecreased cyclin E co-associating with cdk-2, decreased cyclin D1co-associating with cdk-6, and increased p27Kip1 co-associating withcdk-4 in PKC-d transfectants.

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3129

Oncogene

overexpression increased programmed cell death inCaco-2 cells. Because the increased ratio of proapoptoticto antiapoptotic proteins is an important determinant ofa cell’s proclivity to apoptose, we characterized altera-tions in expression levels of members of the Bcl-2 familyof proteins, key regulators of cell survival pathways(Potten et al., 1997). As shown in Figure 11, PKC-dtransfectants exhibited a significant 50% decrease inexpression of antiapoptotic protein, Bcl-2, and asignificant 100% increase in expression of proapoptoticprotein, Bax, as compared to EV control cells (note: Baxand Bcl-2 at zero time point, lanes 1 and 5). Consistentwith an enhancement of apoptosis observed in phorbolester-treated cells (Cerda et al., 2001), TPA furtherdecreased Bcl-2 expression B1.5-fold by 8 h of treat-ment in Caco-2 cells overexpressing PKC-d, in contrastto EV controls where expression of Bcl-2 was slightlyincreased (Figure 11, compare lanes 8 and 4). Althoughproapoptotic Bak expression was remarkably increasedupon phorbol ester treatment, there was no significantdifference in expression of this protein between EV andPKC-d Caco-2 transfectants, suggesting that this proa-poptotic member does not play a key role in PKC-d

induced Caco-2 cell death (Figure 11, compare lanes 2–4with 6–8). In summary, PKC-d overexpression changesthe apoptotic rate of Caco-2 colon cancer cells byincreasing the ratio of proapoptotic Bax to antiapopto-tic Bcl-2.

Real-time PCR analysis demonstrated that BaxmRNA was significantly increased two-fold in PKC-dtransfectants, whereas Bcl-2 mRNA was unchanged(Figure 12). Compared with EV cells, Bax mRNA levelsremained significantly higher in PKC-d transfectantsupon TPA treatment (Figure 12). This is consistent withincreased transcription of proapoptotic Bax in PKC-dtransfectants. These results suggest that PKC-d-depen-dent decreases in Bcl-2 involve a post-transcriptionalmechanism, whereas transcriptional mechanisms,and/or possibly mRNA stabilization, are involved inthe PKC-d mediated increase in Bax expression.

PKC-d knockdown with specific siRNA oligonucleo-tides, but not irrelevant control sequences, inhibited Baxprotein expression by more than 50% (Figure 13,compare lanes 3 and 4 with 1 and 2), further supportingour conclusion that PKC-d enhances apoptosis throughBax upregulation.

Discussion

The colonic epithelium requires continuous renewal toreplace cell loss (Shanmugathasan and Jothy, 2000). Theprogenitor stem cells are localized in the crypt base.Daughter cells soon cease to divide and mature as theymigrate upward. Colonocytes eventually apoptose ascells reach the colonic lumen. These colonic differentia-tion and cell death paradigms are regulated by multiplesignaling pathways, including members of the PKC

TPA

0 6 15 0 6 15 Time (hrs)

- + + - + +

IP:cdk4WB:p21Waf1

IP:cdk6WB:p21Waf1

1 2 3 4 5 6

EV PKC-δ

Figure 10 PKC-d causes a persistent co-association of p21Waf1

protein with cdk 4 and 6. Empty vector (EV) and PKC-dtransfected cells were treated with 100 nM TPA or DMSO (vehicle)for the indicated times (0–15h). Cdk4 and 6 were immunopreci-pitated from total cell extracts (200mg). Co-associating p21Waf1 wasdetected by Western blotting. Arrow denotes p21Waf1

(MR¼ 21 kDa). Western blots are representative of two indepen-dent experiments in duplicate. Note p21Waf1–cdk4 and p21Waf1-6complexes were more temporally sustained in PKC-d transfectants.

0 2 4 8 0 2 4 8

EV PKC-δ

Bax

Bcl-2

β-actin

Bak

1 2 3 4 5 6 7 8

+ TPA

Time (hrs)

21

26

23

42

+ TPAkDa

Figure 11 PKC-d increases the ratio of Bax/Bcl-2 to favor deathpromotion in Caco-2 cells. Whole-cell lysates were prepared fromempty vector (EV) and PKC-d transfected cells, treated with100 nM TPA or DMSO (vehicle) for the indicated times (0–8 h).Proteins (50mg/lane) were probed by Western blotting for Bax, Bcl-2, and Bak protein expression. Blots shown are representative oftwo independent experiments with comparable results.

0

0.5

1

1.5

2

2.5

3

3.5

CTL TPA CTL TPA

mR

NA

exp

ress

ion

(A

U)

EV

PKC-δ

Bcl -2Bax

*

*

Figure 12 PKC-d alters Bax, but not Bcl-2 mRNA levels in Caco-2 cells. Preconfluent Caco-2 cells, transfected with empty vector, orfull-length human PKC-d, treated for 3 h with 100 nM TPA orDMSO (vehicle), were assessed for Bcl-2 and Bax mRNA by real-time PCR. Real-time PCR analysis for levels of Bcl-2 and BaxmRNA are expressed in arbitrary units, normalized to GAPDH.Data are means of two independent experiments in replicates ofthree. *Po0.05, compared with EV treated with respective agonist.

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3130

Oncogene

family (Weinstein, 1991; Klein et al., 2000). Spatial andtemporal changes in PKC-d expression and membraneassociation in normal rat intestinal and colonic epithe-lium suggest this isoform regulates normal differentia-tion and apoptosis in the GI tract (Saxon et al., 1994;Verstovsek et al., 1998; Klein et al., 2000).

Colonic carcinogenesis is characterized by the pro-gressive loss of homeostatic regulators (Lipkin, 1971;Moss et al., 1996; Shanmugathasan and Jothy, 2000).Alterations in several PKC isoforms occur duringcolonic malignant transformation, including loss ofPKC-d in neoplastic progression (Weinstein, 1991;Craven and DeRubertis, 1992, 1994; Davidson et al.,1994; Kahl-Rainer et al., 1994, 1996; McGarrity andPeiffer, 1994; Wali et al., 1995). Caco-2 cells, derivedfrom a human colon cancer, also show downregulationof PKC-d (Cerda et al., 2001; Banan et al., 2002). Thesecells are widely employed to study mechanisms ofcolonocyte growth, maturation and cell death (Pintoet al., 1983; Evers et al., 1996; Gartel et al., 1996;Abraham et al., 1998; Ding et al., 1998). Caco-2 cellsexpress multiple PKC isoforms, which are thought todeliver distinct mitogenic and/or apoptotic signals incolon cancer cells (Sauma et al., 1996; Scaglione-Sewellet al., 1998; Chen et al., 1999; Weller et al., 1999; Perlettiet al., 2005; Wang et al., 2006). Our previous studies inCaco-2 transfectants, engineered to upregulate PKC-d,demonstrated that PKC-d inhibited cell growth, enhan-ced differentiation and accelerated apoptosis (Cerdaet al., 2001).

In the present study, we have successfully used thesetransfected Caco-2 cells, as well as Caco-2 cells withPKC-d knockdown by sequence specific siRNA, todissect several growth-regulating pathways mediated bythis isoform. Specifically, in this study, we havedemonstrated that PKC-d overexpression in Caco-2cells downregulated cyclin D1 and cyclin E, andupregulated G1 cdk co-associations with cyclin-cdkinhibitors, p21Waf1 and p27Kip1. PKC-d also decreasedantiapoptotic Bcl-2 and increased proapoptotic Bax. To

our knowledge, this is the first study demonstrating thatPKC-d downregulates cyclin D1 and cyclin E expressionin colon cancer cells. In addition, we have shown thatPKC-d alters Bcl-2 and Bax expression in directionspredicted to enhance apoptosis in Caco-2 cells. Thecurrent studies have thus extended our previous findingsby elucidating several important underlying mechanismsby which PKC-d inhibits growth and enhances apop-tosis in colon cancer cells.

Since PKC-d is increased in postmitotic differentiat-ing colonocytes, we previously speculated and subse-quently confirmed that this kinase inhibited Caco-2cellular proliferation (see Figure 1 and Cerda et al.,2001). Previously, we have demonstrated that shortincubations of Caco-2 cells with phorbol esters(o60 min) did not downregulate PKC-d (Khare et al.,1999). In the current study, we found that short-termtreatment of Caco-2 cells with phorbol esters (o60 min),caused similar growth inhibition as long-term treatment(data not shown). Thus in our system, we infer thatPKC-d activation, not downregulation, is sufficient toinhibit proliferation. This conclusion is in agreementwith our fundamental hypothesis that PKC-d subservesa tumor suppressor role. Loss of PKC-d, as occurs incolonic tumorigenesis, would then be predicted toincrease proliferation. Using siRNA to specificallyinhibit expression of this enzyme, we found that PKC-dknockdown enhanced Caco-2 cell proliferation,consistent with our previous results that PKC-dupregulation (and activation) inhibited Caco-2 cellproliferation (see Figure 2 and Cerda et al., 2001).These antiproliferative effects of PKC-d might reflectphase-specific or nonspecific cell cycle inhibition and/oracceleration of cell death. In the present study, weinitially investigated PKC-d induced changes in the cellcycle. We demonstrated that upregulation of PKC-dslowed the G1-S transition upon agonist stimulationin Caco-2 cells (Table 1). In agreement with ourfindings, in many noncolonic cells, PKC-d is a negativeregulator of G1-S transition (Jackson and Foster,2004). In contrast, in other cell types PKC-d promotesproliferation, indicating that PKC-d can mediate oppos-ing effects determined by the cell context (Jackson andFoster, 2004).

Since PKC-d upregulation inhibited G1 to S transi-tion, we next examined several regulators of G1/S cellcycle control. We found PKC-d overexpression down-regulated cyclin D1 and cyclin E protein expression inCaco-2 cells through post-transcriptional (cyclins D1and E) and transcriptional (cyclin E) mechanisms(Figures 4 and 5). This downregulation was a specificeffect of PKC-d since knockdown of this isoform bysequence specific siRNAs increased the protein expres-sion of these cyclins (Figure 6). This downregulation isin contrast to the TPA-induced increase in protein andmRNA expression of cyclin E, and to a lesser extentcyclin D1, seen in EV cells. We speculate cyclin D1 andE upregulation in EV cells is mediated by activation ofother (non-PKC-d) phorbol ester-sensitive PKC iso-forms that enhance cell cycle progression (Table 1). TPAin fact increased S phase in EV but not in PKC-d

SS siRNA PKC-δ siRNA kDa

PKC-δ 78

PKC-δ CF 40

Bax21

β-actin 42

4321

Figure 13 PKC-d knockdown decreases Bax expression in Caco-2cells. Caco-2 cells overexpressing PKC-d were transfected with100 nM (lanes 3, 4) PKC-d specific RNAi 21-nt duplex, or 100 nM

scrambled control (SS; lanes 1,2), using siPORT transfectionreagent. After 72 h, whole-cell lysates (30mg protein) were assayedfor PKC-d, PKC-dCF, Bax, and b-actin. Blots representative oftwo independent experiments in duplicate are shown.

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3131

Oncogene

transfectants. We believe these opposing effects in EV vsPKC-d transfectants occurred since PKC-d mediates acell cycle slowing by increasing p21Waf1 and decreasingcyclin D1 and E (Figures 4, 7 and 8).

Cyclins D1 and E are overexpressed in a variety oftumors, including colon carcinomas (Bartkova et al.,1994; MacLachlan et al., 1995; Arber et al., 1996;Hosokawa and Arnold, 1998; Donnellan and Chetty,1999; Jang et al., 1999; Milde-Langosch et al., 2000).Numerous studies have confirmed the importance ofthese cyclins in promoting G1/S transit (Bartek andLukas, 2001; Jones and Kazlauskas, 2001), and theircritical roles in cancer cell proliferation (Liu et al., 1995;Arber et al., 1997; Donnellan and Chetty, 1999). Loss ofcyclin D1, moreover, has been demonstrated to inhibitthe growth and tumorigenicity of colon cancer cells(Arber et al., 1997). Our studies are in agreement withprevious findings in which PKC-d overexpressioninhibited proliferation of vascular smooth muscle andcapillary endothelial cells by suppressing cyclin D1 andcyclin E expression (Fukumoto et al., 1997; Ashtonet al., 1999). In contrast to our findings in Caco-2 cells,however, PKC-d has been shown to downregulate cyclinD1 expression at the transcriptional level in severalnoncolonic cell types (Page et al., 2002; Soh andWeinstein, 2003). These differences emphasize theimportance of understanding the control of cell cycleregulators by PKC-d in the context of colon cancer cellsto elucidate its tumor suppressor role in the colon.

In addition to changes in cyclins D1 and E, G1progression is regulated by coordinated actions ofcyclin-dependent and cyclin-independent cdk inhibitors,including members of the CIP/KIP family (p21Waf1,p27Kip1 and p57Kip2) and INK4 (p16, p19) proteins, aswell as the retinoblastoma (pRb) and p53 tumorsuppressor proteins (Akiyama et al., 1992; Dulic et al.,1994; King et al., 1996). In intestinal epithelial cells,PKC agonist-mediated cell cycle arrest in the G1 phasewas accompanied by accumulation of hypophosphory-lated pRb and induction of p21Waf1 and p27Kip1 (Freyet al., 1997). Consistent with a G1 cell cycle slowingobserved in PKC-d transfected Caco-2 cells, we foundthat PKC-d upregulation reduced levels of cyclin D1-cdk6 and cyclin E-cdk2 complexes, and increasedp21Waf1 and p27Kip1co-associating with cdk4 (Figures 9and 10). These changes are predicted to decrease cyclin-cdk kinase activities and thereby contribute to the cellcycle slowing observed in PKC-d transfectants.

Compared to low p21Waf1 expression in proliferatingcells, levels of this cyclin-cdk inhibitor are high innonproliferating and terminally differentiated cells ofthe mouse colon (Gartel et al., 1996). Loss of p21Waf1 canoccur as early as dysplastic aberrant crypt foci, theearliest detectable histological abnormalities and puta-tive precursors of colon cancer (Polyak et al., 1996). Wehave demonstrated that PKC-d upregulates p21Waf1 inCaco-2 cells since TPA treatment of PKC-d transfec-tants enhanced p21Waf1 expression and caused asustained increase in p21Waf1–cdk4 association, whereasrottlerin, a specific PKC-d inhibitor, suppressed thisinduction (Figures 7, 8 and 10). Furthermore, knock-

down of this isoform by sequence specific siRNAsblocked induction of the cyclin-cdk inhibitor (Figure 8).Our findings are in agreement with studies in noncoloniccell systems that have demonstrated a requirement forPKC-d in p21Waf1 induction by phorbol esters (Zezulaet al., 1997; Arita et al., 1998; Shanmugam et al., 2001;Nakagawa et al., 2005; Yokoyama et al., 2005).Although TPA is a general activator of phorbol estersensitive PKC isoforms, we attribute these effects toPKC-d in these transfectants since the endogenousnontargeted PKC isoforms did not differ in expressionin EV or PKC-d transfectants, compared to parentalCaco-2 cells (Cerda et al., 2001). PKC-d upregulationalso enhanced the steady-state mRNA levels of this cdkinhibitor in Caco-2 cells, indicating that PKC-d regu-lates p21Waf1 by increasing transcription and/or stabiliz-ing p21Waf1 mRNA (Figure 7). This finding is consistentwith the established role of transcription in regulatingp21Waf1 expression in Caco-2 cells (Evers et al., 1996;Gartel et al., 1996, 2000a). Loss of PKC-d intumorigenesis might, therefore, contribute to the de-creased p21Waf1 expression observed in colonic carcino-genesis.

Recent studies by Perletti et al. (2005) have demon-strated that PKC-d requires the tumor suppressor p53 toinhibit growth and induce differentiation of HCT116colon cancer cells by a p21Waf1-dependent mechanism.Depending on the cell context, however, p21Waf1 can alsoinduce growth arrest by p53-independent pathways (el-Deiry et al., 1993; Zeng and el-Deiry, 1996; Cayrol et al.,1998). In the case of Caco-2 cells, which are known tolack p53 (Djelloul et al., 1997), we conclude that PKC-dinhibits cell cycling by a p53-independent p21Waf1-dependent pathway (Figures 7 and 8). Several investi-gators have also demonstrated p21Waf1 is induced duringCaco-2 cell differentiation (Evers et al., 1996; Gartelet al., 1996, 2000b; Abraham et al., 1998). Since p53 ismutated or lost in many cancers, including nearly halfof all colonic tumors (Baker et al., 1989), identifyingpathways that activate p21Waf1 by p53-independentmechanisms have potentially important therapeuticimplications. In this regard, p21Waf1 induction byincreased PKC-d signaling is a novel pathway thatmight be exploited in future chemopreventive andchemotherapeutic strategies. In this regard, in prelimin-ary studies, we have found that chemopreventive fishoils (o-3 fatty acids) increased PKC-d expression,whereas tumor promoting o-6 fatty acids downregu-lated the expression of this isoform (Cerda et al., AGAproceedings 2005). We anticipate future investigations inthis area will identify new avenues for more selective andeffective ways to preserve PKC-d, and thereby restrainthe growth of transforming colonoytes.

In addition to the role of PKC-d as a negativeregulator of cell cycle progression, we have previouslydemonstrated that this isoform enhanced the rate ofCaco-2 cell apoptosis (Cerda et al., 2001). To gainfurther insights into this proapoptotic effect, weinvestigated the molecular mechanisms governingPKC-d-mediated cell death in Caco-2 cells. Antiapop-totic Bcl-2 and proapoptotic proteins Bak and Bax are

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3132

Oncogene

key mediators of apoptosis in colon cancer cells(Bronner et al., 1995; Ruemmele et al., 1999; Mandalet al., 2001). Current models suggest that ratios ofantiapoptotic to proapoptotic members of the Bcl-2family of proteins regulate the apoptotic potential of acell (Boise et al., 1995). We found that PKC-dupregulation suppressed Bcl-2 and induced Bax expres-sion, thereby altering the ratio of these cell deathregulators to favor apoptosis in Caco-2 cells (Figures11–13). Our studies, thus, suggest that changes in Bcl-2and Bax contribute to the proapoptotic signaling ofPKC-d. These results are in agreement with previousstudies that suggested this PKC isoform is a positiveregulator of apoptosis in other colon cancer cells (Welleret al., 1999; McMillan et al., 2003; Perletti et al., 2005).Similar effects were recently reported by Sitailo et al.(2004) in human keratinocytes, in which PKC-d-inducedapoptosis was mediated by Bax upregulation.

PKC-d decreased Bcl-2 by a post-transcriptionalmechanism, whereas this isoform increased Bax expres-sion by mechanisms involving transcriptional activation,and/or mRNA stabilization. Moreover, PKC-d knock-down with specific siRNA oligonucleotides inhibitedBax protein expression, further confirming that PKC-dupregulates Bax. Expression patterns of Bcl-2 and Baxare spatially distinct. While Bcl-2 is expressed only incolonic crypt stem cells, where it could protectprogenitor cells from DNA-damage induced apoptotic

death, Bax expression is limited to maturing epithelialcells destined for programmed cell death at the colonicluminal surface (Krajewski et al., 1994; Potten et al.,1997). Upregulation of Bcl-2 has been suggested to playa role in colorectal tumorigenesis, especially in the earlyphases of the adenoma-carcinoma sequence, and in-creased Bcl-2 expression is associated with a worseprognosis in colon cancer (Bosari et al., 1995; Bronneret al., 1995; Thompson, 1995). In contrast, Bax-positivetumors have a significantly better prognosis and greater5-year survival rates compared to Bax-negative tumors(Ogura et al., 1999). Thus, PKC-d, by decreasing Bcl-2and increasing Bax, would be predicted to serve a tumorsuppressor role in colonic carcinogenesis. Bcl-2 andPKC-d also play antagonistic roles in other cells. Forexample, in murine myeloid cells, PKC-d blockedantiapoptotic functions of Bcl-2 overexpression,whereas in vascular smooth muscle cells, upregulatedBcl-2 inhibited PKC-d induced apoptosis (Barrett et al.,2002; Goerke et al., 2002). These results indicate PKC-dand Bcl-2 mediate opposing cell survival paradigms incell context-specific manners.

In conclusion, our experiments have uncovered twoimportant molecular mechanisms governing cell cycleand cell death that contribute to the tumor suppressorfunction of PKC-d in human colonic carcinogenesis.Our findings are summarized in Figure 14. Firstly, PKC-dsuppresses G1-S progression by inhibiting levels of

Figure 14 Proposed model for PKC-d antiproliferative functions in Caco-2 cells. PKC-d signaling suppresses G1–S progression byinhibiting the expression levels of G1 cell cycle activators, cyclin D1 and cyclin E, and upregulating G1 cell cycle inhibitor p21Waf1 aswell as p27Kip1–cdk4 complex. PKC-d signals also inhibit cell survival by downregulating antiapoptotic Bcl-2 and increasingproapoptotic Bax. These tumor suppressor pathways inhibit mitogenic signaling and enhance cell death that result in decreased cellularproliferation of colon cancer cells.

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3133

Oncogene

G1 cell cycle activators, cyclin D1 and cyclin E, andupregulating G1 cell cycle inhibitor p21Waf1 and enhan-cing p27Kip1 co-association with cdk4. Secondly, PKC-dinhibits cell survival by downregulating antiapoptoticBcl-2 and increasing pro-apoptotic Bax. We concludethat PKC-d inhibits mitogenic signaling and enhancescell death in colon cancer cells via these pathways. Theobserved loss of this tumor suppressor in transformingcolonocytes would be predicted, therefore, to provide aclear growth advantage, contributing to the dysregu-lated growth characterizing colonic carcinogenesis. Ourfuture efforts will focus on identifying PKC-d substratesthat mediate these tumor suppressor effects.

Materials and methods

MaterialsTissue culture reagents were obtained from Fisher Scientific(Springfield, NJ, USA), and media supplements were fromGIBCO BRL-Life Technologies (Gaithersburg, MD, USA).The mouse monoclonal antibody to human PKC-d proteinused for Western blotting was from Signal Transduction Labs(Lexington, KY, USA). Protein A/G PLUS-Agarose beads,mouse monoclonal antibody to Bcl-2, and rabbit polyclonalantibodies to human Bax, cyclin D1, cyclin E, PKC-d and -zwere from Santa Cruz Biotechnology (Santa Cruz, CA, USA).Other primary antibodies used include anti-p21Waf1 and -p27Kip1

(monoclonal, Transduction Labs.); anti-human Bak (mono-clonal, PharMingen); antiphosphoPKC-dSer643/676 and-phosphoPKC-dThr505 (polyclonal, Cell Signaling); andanti-b-actin (monoclonal, Sigma). Immobilon-P (polyvinyli-dene difluoride) membranes were purchased from Millipore(Bedford, MA, USA). The peroxidase-conjugated sheep anti-mouse and donkey anti-rabbit antibodies and enhancedchemiluminescence (ECL) Western blotting protein detectionkit were supplied by Amersham (Arlington Heights, IL, USA).Instruments and reagents used for immunoblotting and gelelectrophoresis, and Chelex 100, were from Bio-Rad Labora-tories (Richmond, CA, USA). Kodak (Rochester, NY, USA)supplied the X-OMAT AR film. The phorbol ester, 12-O-tetradecanoylphorbol 13-acetate (TPA), was obtained fromSigma Chemical (St Louis, MO, USA). The PKC-d-specificactivator, BisA, was kindly provided by Dr Diane Watters(Royal Brisbane Hospital, Australia). The PKC-d-specificinhibitor, Rottlerin, was obtained from Calbiochem (SanDiego, CA, USA). For RNA isolation, RNeasy and DNA-freet kits were purchased from Quiagen (Santa Clarita, CA,USA). Custom PCR primers were ordered from InvitrogenLife Technologies (Carlbad, CA, USA). Taqman probes weresupplied by Synthegen (Houston, TX, USA). Other PCRreagents, including Moloney murine leukemia virus reversetranscriptase were from Applied Biosystems (Foster City, CA,USA). HotStarTaqt DNA polymerase was supplied byQuiagen. The Cepheid Smart Cycler was used for PCR(Sunnyvale, CA, USA). In total, 21 nucleotide sense andantisense siRNAs were chemically synthesized by Ambion Inc.(Austin, TX, USA). All other reagents used in this study werefrom Sigma Chemical or Fisher Scientific, unless otherwiseindicated, and were of the highest purity available.

Preparation of zinc-depleted mediumZinc-depleted medium was prepared by a modification of themethod by Palmiter (1994), as previously described by ourlaboratory (Cerda et al., 2001).

Cell cultureCaco-2 cells were obtained from ATCC and low passage cells(passages 21–31) were cultured at 371C in a humidifiedatmosphere of 5% CO2–95% air, as previously described byour laboratory (Scaglione-Sewell et al., 1998). PKC-d induciblestable transfectants were prepared as previously described(Cerda et al., 2001). In the presence of Zn2þ , PKC-dtransfectants upregulated PKC-d protein B4-fold and demon-strated B2-fold increase in PKC-d-specific kinase activity,compared with EV transfected control cells (Cerda et al.,2001). Cells are routinely maintained in standard Dulbecco’sminimum essential medium (DMEM) containing 4.5 g/lglucose, 10mM HEPES (pH 7.4), 2 mM L-glutamine, 50 U/mlpenicillin G, 50 mg/ml streptomycin, 2 mg/ml gentamicin, 1%essential and nonessential amino acids, and 20% fetal bovineserum (FBS). Stable transfectants were maintained in zinc-depleted medium containing 400 mg/ml of G418 (Cerda et al.,2001). To induce expression of PKC-d in these clones, EV andPKC-d transfected Caco-2 cells were cultured in Chelex-treated DMEM supplemented with 175 mM Zn2þ , a nontoxicconcentration of zinc that activates the metallothioneinpromoter. Maximal induction of this gene is achieved by24–48 h.

Western blottingTotal cell lysates were prepared from treated cells and boiledfor 3min in 2� Laemmli SDS buffer (Laemmli, 1970). Proteinwas assayed by the Bio-Rad Detergent Compatible proteinassay using BSA as the standard. Equal amounts of protein(15–50mg/lane) were separated by sodium dodecyl sulfate–polyacrylamide electrophoresis (SDS–PAGE), as described byLaemmli (1970), and transferred onto Immobilon-P mem-branes (Millipore, Bedford, MA, USA). Nonspecific bindingof antibodies was blocked by incubating blots at roomtemperature for 3 h with 5% non-fat dry milk in TBST (Tris-buffered saline in 0.05% Tween-20). After blocking, blots wereprocessed for immunoblotting by overnight incubation at 41Cwith primary antibodies in TBST-5% milk. For phospho-specific antibodies, TBST-5% BSA was used to blocknonspecific binding of antibodies and TBST-3% BSA wasused for immunoblotting procedures. After incubation, blotswere rinsed and incubated with appropriate (monoclonal vspolyclonal) peroxidase-conjugated secondary antibodies. Pro-teins were detected using an ECL system followed byxerography on X-OMAT AR film, as recommended by themanufacturer. The xerograms were scanned with a JX-3F6scanner (Sharp Electronics, Mahwah, NJ, USA), and quanti-fied using IP lab gel software (Signal Analytics, Vienna, VA,USA). Exposure times were adjusted to ensure a linearresponse.

Measurement of the abundance of G1 cyclin-cdk proteincomplexes, cyclin D1–cdk 4,6 and cyclin E–cdk 2Immunoprecipitation of kinase complexes using specificantibodies were performed as previously described by ourlaboratory (Scaglione-Sewell et al., 2000). Briefly, EV andPKC-d cDNA transfected Caco-2 cells were seeded at a densityof 5� 105 cells in Falcon 60mm tissue culture dishes andgrown in zinc-depleted medium for 2 days. Induction of PKC-d expression was initiated by replacing medium with Chelex-treated DMEM medium containing 175mM Zn2þ . Following a48 h induction, treated cells were washed with PBS and lysed in300 ml lysis buffer containing 50mM HEPES (pH 7.5), 150mM

NaCl, 1mM DTT, 2.5mM EGTA, 0.1% Tween-20, 10%glycerol, 0.1 mM PMSF, 10 mg/ml leupeptin, 20 U/ml aproti-nin, 10mM b-glycerophosphate, 0.1 mM sodium orthovana-

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3134

Oncogene

date, and 1mM NaF. Cells were lysed on ice for 30 min,sonicated briefly, and centrifuged at 10 000� g for 15min toremove insoluble debris. Lysates (50 mg protein) were pre-cleared with 1mg of normal rabbit serum and 20ml of proteinA/G sepharose and incubated overnight with 5ml antibodies tocdk2, cdk4, or cdk6 in a rocking platform at 41C. Protein A/GPLUS-Agarose beads (30 ml) were added to lysates andincubated for an additional 3 h at room temperature in arocking platform. After four washes in lysis buffer, agarosebeads were resuspended in 2� Laemmli buffer and boiled for5min to release immunoprecipitated proteins. Proteins wereseparated by SDS–PAGE on a 12 or 15% resolving gels, andelectroblotted to polyvinylidene difluoride membranes. Com-parable levels of kinase abundance were confirmed by probinglysates for cdk expression levels.

mRNA isolation and quantification by real-time PCRTotal RNA was isolated from treated cells using the RNAeasyKit (Quiagen), according to the manufacturer’s protocol.RNA was dissolved in DEPC-treated water and genomic DNAremoved by DNase-I using the DNA-freet kit. RNA wasstored in RNase free conditions at �801C until further use. Areal-time PCR-based assay using Taqman probes was used toquantify the levels of mRNA as previously described withseveral modifications (Schmittgen et al., 2000). Briefly, RNAwas reverse transcribed in a 20 ml reaction containing 1 mg ofRNA, 5 mM MgCl2, 1� PCR Buffer II, 4mM dNTPs(deoxyribonucleotide triphosphates: dATP, dGTP, dCTP,dTTP), 20 U of RNase inhibitor, 2.5 mM random hexamers,and 50 U of Moloney murine leukemia virus reverse tran-scriptase. Samples were reverse-transcribed in a PTC-100thermal cycler at 421C for 60 min and the temperature thenincreased to 951C for 5 min to denature the reverse transcrip-tase. For real-time PCR, each reaction contained 5ml of a 1:5dilution of the cDNA, 1� PCR Buffer II, 5.5 mM MgCl2,200mM dNTPs, 200 nM of forward and reverse primers, 100 nM

of Taqman probe, 0.75 U of HotStarTaqt DNA polymerase,and water added to adjust the final volume to 25ml. Thereactions were incubated at 951C for 15min to activate theDNA polymerase followed by 45 cycles of 15 s at 951C and 60 sat 601C. To normalize for total RNA, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as internalcontrol gene in our samples. In preliminary experiments, PCRproducts of the expected sizes were visualized on an agarosegel and their predicted sequences confirmed by DNA sequen-cing. All primers were designed using computer software tooptimize amplification of specific cDNA sequences andchecked for binding against nontargeted genes using theavailable genomic sequences in the database. Sequences forprimers and Taqman probes used for real-time PCR were asfollows (50 to 30): cyclin D1: TGCGAGGAACAGAAGTGCG(forward primer), AGCTGCAGGCGGCTCTT (reverse pri-mer), TCCTGTCGCTGGAGCCCGTGAA (Taqman Probe);cyclin E: CTCCAGGAAGAGGAAGGCAA (forward primer),TCGATTTTGGCCATTTCTTCA (reverse primer), CGTGACCGTTTTTTTGCAGGATCC (Taqman Probe); p21: GACTCTCAGGGTCGAAAACG (forward primer), GGATTAGGGCTTCCTCTTGG (reverse primer), CATGACAGATTTCTACCACTCCAAACGCC (Taqman Probe); Bax: GTCGCCCTTTTCTACTTTGC (forward primer), GGAGGAAGTCCAATGTCCAG (reverse primer), CCAAGGTGCCGGAACTGATCAGAAC (Taqman Probe); Bcl-2: GGATTGTGGCCTTCTTTGAG (forward primer), GCCGGTTCAGGTACTCAGTC (reverse primer), CGGTTGACGCTCTCCACACACATG(Taqman Probe). cDNA samples from EV and PKC-d cloneswere synthesized in parallel reactions, along with negative

controls for which reverse transcriptase was omitted from thereaction. PCR reactions were run in triplicate, and mRNAlevels were expressed as relative changes after normalization toGAPDH mRNA abundance.

Preparation and transfection of short interfering RNAs(siRNAs) targeting PKC-dRNA interference (RNAi), the targeted mRNA degradationinduced by double-stranded RNA (dsRNA), was used as acomplementary technique to our overexpression systems toanalyse PKC-d function. We used siRNAs targeting humanPKC-d, which have been confirmed to successfully inhibitPKC-d mRNA by more than 80% (Irie et al., 2002), followingthe methods described for siRNAs applicable to culturedmammalian cells (Elbashir et al., 2001). Briefly, 21 nucleotideRNAs, containing 30overhangs of 20-deoxythymidine, werechemically synthesized by Ambion Inc. PKC-d siRNAsequences, corresponding to nucleotides 603–621 after thestart codon, were as follows: 50-CGAGAAGAUCAUCGGCAGATT-30 and 50-UCUGCCGAUGAUCUUGUCGTT-30.Duplexes of siRNAs were generated by incubating 20 mM senseand antisense strands in annealing buffer (100mM potassiumacetate, 30 mM HEPES-KOH pH 7.4, and 2mM magnesiumacetate) for 1 min at 901C to denature, followed by a 1 hincubation at 371C to renature. EV and PKC-d overexpressingCaco-2 cells were transfected with 100 nM PKC-d specificRNAi 21-nt duplex or a scrambled control, using siPORTtransfection reagent (Ambion). After 72 h, cells were lysed andproteins (20 mg) assayed for PKC-d other PKC isoforms, andb-actin. We achieved more than 80% downregulation of PKC-d expression under these conditions whereas other PKCisoforms were unchanged. Lysates were used to probe forprotein expression by Western blotting.

Formazan and BrdU-based cell proliferation assaysEV and PKC-d transfectants were plated onto 96-wellmicrotiter tissue culture plates in zinc-depleted DMEMmedium containing 400 mg/ml G418 at a density of5� 103 cells/well. Following a 24-h period of zinc pretreatment,cells in replicates of six were treated with 100 nM of either thePKC agonist, phorbol 12-myristate 13-acetate (TPA), or thePKC-d-specific activator, BisA, in the presence of 175 mM Zn2þ

for an additional 48 h. Groups treated with DMSO alone wereused as vehicle controls. For RNAi experiments, Caco-2 cellswere treated with 50 or 100 nM PKC-d-specific RNAi orscrambled control oligonucleotides using siPORT transfectionreagent. Effects of PKC agonists or siRNA on cell growthwere detected using two separate colorimetric assays for thequantification of cell proliferation; the formazan assay and theBrdU cell proliferation assay (Roche Molecular Biochemicals,IN, USA). The formazan assay is based on the cleavage of thetetrazolim salt Wst-1 by mitochondrial dehydrogenases inviable cells. The BrdU colorimetric immunoassay is aquantitative cell proliferation assay based on the measurementof BrdU incorporation during DNA synthesis. After treat-ments, 10ml/well of Wst-1 reagent or 20ml/well of BrdU wereadded to each well, followed by an incubationof 0.5–4 h at371C. The formazan dye produced by metabolically active cellswas quantified by measuring absorbance of samples on ascanning multiwell spectrophotometer (ELISA reader) using a450 nm filter. For the BrdU assay, cells were fixed and theDNA denatured. Anti-BrdU-peroxidase immune complexeswere detected by substrate reaction and quantified in anELISA reader at 370 nm. We found that Wst-1 and BrdUincorporation gave comparable results for cell proliferation.

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3135

Oncogene

Cell cycle analysis by flow cytometryEV and PKC-d cDNA transfected Caco-2 cells were seeded insix-well tissue culture dishes at a density of 1� 105 cells/well inzinc-depleted medium and allowed to attach overnight. PKC-dinduction was initiated by replacing medium in triplicate wellswith Chelex-treated DMEM medium containing 175 mM Zn2þ

for a period of 48 h. This was followed by treatment of cellswith 100 nM of either TPA or 100 nM BisA, or DMSO controlfor 6–24 h in the presence of zinc. At 6, 15 and 24 h oftreatment, the distribution cells in the cell cycle were assessedaccording to the method of Darzynkiewicz (1994), aspreviously described by our laboratory (Cerda et al., 2001).

Assessment of apoptosisEV and PKC-d transfectants were seeded at a density of1.5� 105 cells in Nunc polystyrene slideflasks (Nunc, Naper-ville, IL, USA) and grown in zinc-depleted medium for twodays. Preconfluent cells were then treated for a period of 48 hin Chelex-treated DMEM supplemented with 175mM Zn2þ inthe presence of 100 nM of either TPA or 100 nM BisA, orvehicle (DMSO, 0.05%). Cells were washed twice with HBSSand fixed for 10min at room temperature with 4% parafor-maldehyde in PBS. Fixative was removed by aspiration andthe monolayer washed twice in PBS. DNA was stained for1 min with 2mg/ml 40,6-diamidino-2-phenylindole dihy-drochloride (DAPI; Molecular Probes, Eugene, OR, USA) atroom temperature. Excess DAPI stain was removed and themonolayer was extensively washed with water and a coverslipmounted with Vectashieldt mounting medium (Vector La-boratories, Burlingame, CA, USA). Stained nuclei were viewedat � 200 using an Olympus BH-2 fluorescence microscope(Olympus Optical Co., Ltd, Japan) with the UV cube inposition. Apoptotic nuclei, as assessed by nuclear condensa-

tion and fragmentation, were quantified in 1000 random cellsin duplicate platings per sample for each experiment, aspreviously described by our laboratory (Cerda et al., 2001).Rates of apoptosis were expressed as a mean percentage7s.d.(apoptotic nuclei/total nuclei� 100).

Statistical analysisNumerical data are expressed as means7s.d. Statisticalsignificance of the differences between samples was determinedby unpaired Student’s t-test. Data were considered signifi-cantly different at the level of Po0.05.

Abbreviations

PKC, protein kinase C; PKC-d, PKC-delta; EV, empty vector;SDS–PAGE, sodium dodecyl sulfate–polyacrylamide gelelectrophoresis; FACS, fluorescence-activated cell sorter;BisA, Bistratene A; TPA, 12-O-tetradecanoylphorbol 13-acetate; IP, immunoprecipitation; cdk, cyclin dependentkinase; SS, scrambled sequence; DAG, diacylglycerol; PC,phosphatidylcholine; PI, phophatidylinositol; PLC, phospho-lipase C; RTK, receptor tyrosine kinase; GPCR, G proteincoupled receptors.

Acknowledgements

This research was supported in part by National Institutes ofHealth Grant 2R01CA36745 (MB), the American CancerSociety Illinois Division Grant 04-14 (SRC), and the DigestiveDisease Research Core Center Grant P30DK42086 at theUniversity of Chicago.

References

Abraham C, Scaglione-Sewell B, Skarosi SF, Qin W,Bissonnette M, Brasitus TA. (1998). Gastroenterology 114:503–509.

Acevedo-Duncan M, Patel R, Whelan S, Bicaku E. (2002). CellProlif 35: 23–36.

Akiyama T, Ohuchi T, Sumida S, Matsumoto K, ToyoshimaK. (1992). Proc Natl Acad Sci USA 89: 7900–7904.

Andre F, Rigot V, Remacle-Bonnet M, Luis J, Pommier G,Marvaldi J. (1999). Gastroenterology 116: 64–77.

Arber N, Doki Y, Han EK, Sgambato A, Zhou P, Kim NHet al. (1997). Cancer Res 57: 1569–1574.

Arber N, Hibshoosh H, Moss SF, Sutter T, Zhang Y, Begg Met al. (1996). Gastroenterology 110: 669–674.

Arita Y, Buffolino P, Coppock DL. (1998). Exp Cell Res 242:381–390.

Ashton AW, Watanabe G, Albanese C, Harrington EO, WareJA, Pestell RG. (1999). J Biol Chem 274: 20805–20811.

Assert R, Kotter R, Bisping G, Scheppach W, Stahlnecker E,Muller KM et al. (1999). Int J Cancer 80: 47–53.

Atten MJ, Godoy-Romero E, Attar BM, Milson T, Zopel M,Holian O. (2005). Invest New Drugs 23: 111–119.

Avazeri N, Courtot AM, Lefevre B. (2004). J Cell Sci 117:4969–4978.

Baker SJ, Fearon ER, Nigro JM, Hamilton SR, Preisinger AC,Jessup JM et al. (1989). Science 244: 217–221.

Banan A, Fields JZ, Farhadi A, Talmage DA, Zhang L,Keshavarzian A. (2002). J Pharmacol Exp Ther 303:17–28.

Barrett CM, Lewis FL, Roaten JB, Sweatman TW, Israel M,Cleveland JL et al. (2002). Mol Cancer Ther 1: 469–481.

Bartek J, Lukas J. (2001). FEBS Lett 490: 117–122.Bartkova J, Lukas J, Strauss M, Bartek J. (1994). Int J Cancer58: 568–573.

Besson A, Yong VW. (2000). Mol Cell Biol 20: 4580–4590.Black JD. (2000). Front Biosci 5: D406–D423.Black JD. (2001). Gastroenterology 120: 1868–1872.Boise LH, Gottschalk AR, Quintans J, Thompson CB. (1995).

Curr Top Microbiol Immunol 200: 107–121.Bosari S, Moneghini L, Graziani D, Lee AK, Murray JJ,

Coggi G et al. (1995). Hum Pathol 26: 534–540.Brasitus TA, Bissonnette M. (1998). Gastroenterology 115:

225–227.Brodie C, Blumberg PM. (2003). Apoptosis 8: 19–27.Bronner MP, Culin C, Reed JC, Furth EE. (1995). Am J

Pathol 146: 20–26.Castagna M, Takai Y, Kaibuchi K, Sano K, Kikkawa U,

Nishizuka Y. (1982). J Biol Chem 257: 7847–7851.Cayrol C, Knibiehler M, Ducommun B. (1998). Oncogene 16:

311–320.Cerda SR, Bissonnette M, Scaglione-Sewell B, Lyons MR,

Khare S, Mustafi R et al. (2001). Gastroenterology 120:1700–1712.

Chen A, Davis BH, Bissonnette M, Scaglione-Sewell B,Brasitus TA. (1999). J Biol Chem 274: 35505–35513.

Clark JA, Black AR, Leontieva OV, Frey MR, Pysz MA,Kunneva L et al. (2004). J Biol Chem 279: 9233–9247.

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3136

Oncogene

Craven PA, DeRubertis FR. (1992). Cancer Res 52: 2216–2221.

Craven PA, DeRubertis FR. (1994). Dig Dis Sci 39: 481–489.Darzynkiewicz Z. (1994). Cell Biology: A Laboratory Handout.

Academic Press: New York, pp 261–271.Davidson LA, Jiang YH, Derr JN, Aukema HM, Lupton JR,

Chapkin RS. (1994). Arch Biochem Biophys 312: 547–553.Davies SP, Reddy H, Caivano M, Cohen P. (2000). Biochem

J 351: 95–105.Denning MF, Wang Y, Nickoloff BJ, Wrone-Smith T. (1998).

J Biol Chem 273: 29995–30002.Di Mari JF, Mifflin RC, Powell DW. (2005). Gastroenterology128: 2131–2146.

Ding QM, Ko TC, Evers BM. (1998). Am J Physiol 275:C1193–C1200.

Djelloul S, Forgue-Lafitte ME, Hermelin B, Mareel M,Bruyneel E, Baldi A et al. (1997). FEBS Lett 406: 234–242.

Donnellan R, Chetty R. (1999). FASEB J 13: 773–780.Dulic V, Kaufmann WK, Wilson SJ, Tlsty TD, Lees E, Harper

JW et al. (1994). Cell 76: 1013–1023.Elbashir SM, Harborth J, Lendeckel W, Yalcin A, Weber K,

Tuschl T. (2001). Nature 411: 494–498.el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R,

Trent JM et al. (1993). Cell 75: 817–825.Evers BM, Ko TC, Li J, Thompson EA. (1996). Am J Physiol271: G722–7.

Fearon ER, Vogelstein B. (1990). Cell 61: 759–767.Fima E, Shtutman M, Libros P, Missel A, Shahaf G, Kahana

G et al. (2001). Oncogene 20: 6794–6804.Fishman DD, Segal S, Livneh E. (1998). Int J Oncol 12:

181–186.Frey MR, Clark JA, Leontieva O, Uronis JM, Black AR,

Black JD. (2000). J Cell Biol 151: 763–778.Frey MR, Saxon ML, Zhao X, Rollins A, Evans SS, Black JD.

(1997). J Biol Chem 272: 9424–9435.Fukumoto S, Nishizawa Y, Hosoi M, Koyama H, Yamakawa

K, Ohno S et al. (1997). J Biol Chem 272: 13816–13822.Gartel AL, Goufman E, Najmabadi F, Tyner AL. (2000a).

Oncogene 19: 5182–5188.Gartel AL, Najmabadi F, Goufman E, Tyner AL. (2000b).

Oncogene 19: 961–964.Gartel AL, Serfas MS, Gartel M, Goufman E, Wu GS, el-

Deiry WS et al. (1996). Exp Cell Res 227: 171–181.Gavrielides MV, Frijhoff AF, Conti CJ, Kazanietz MG.

(2004). Curr Drug Targets 5: 431–443.Geiges D, Marks F, Gschwendt M. (1995). Exp Cell Res 219:

299–303.Goerke A, Sakai N, Gutjahr E, Schlapkohl WA, Mushinski

JF, Haller H et al. (2002). J Biol Chem 277: 32054–32062.Goldstein DR, Cacace AM, Weinstein IB. (1995). Carcino-

genesis 16: 1121–1126.Graham MA, Rawe I, Dartt DA, Joyce NC. (2000). Invest

Ophthalmol Vis Sci 41: 4124–4132.Greenlee RT, Murray T, Bolden S, Wingo PA. (2000). CA

Cancer J Clin 50: 7–33.Gschwendt M. (1999). Eur J Biochem 259: 555–564.Gschwendt M, Muller HJ, Kielbassa K, Zang R, Kittstein W,

Rincke G et al. (1994). Biochem Biophys Res Commun 199:93–98.

Hosokawa Y, Arnold A. (1998). Genes Chromosomes Cancer22: 66–71.

Irie N, Sakai N, Ueyama T, Kajimoto T, Shirai Y, Saito N.(2002). Biochem Biophys Res Commun 298: 738–743.

Jackson DN, Foster DA. (2004). FASEB J 18: 627–636.Jang SJ, Park YW, Park MH, Lee JD, Lee YY, Jung TJ et al.

(1999). Eur J Surg Oncol 25: 157–163.

Jones SM, Kazlauskas A. (2001). FEBS Lett 490: 110–116.Kahl-Rainer P, Karner-Hanusch J, Weiss W, Marian B.

(1994). Carcinogenesis 15: 779–782.Kahl-Rainer P, Sedivy R, Marian B. (1996). Gastroenterology110: 1753–1759.

Khare S, Bissonenette M, Scaglione-Sewell B, Wali RK, SitrinMD, Brasitus TA. (1999). Am J Physiol 276: G993–G1004.

Kikkawa U, Matsuzaki H, Yamamoto T. (2002). J Biochem(Tokyo) 132: 831–839.

King RW, Deshaies RJ, Peters JM, Kirschner MW. (1996).Science 274: 1652–1659.

Kitamura K, Mizuno K, Etoh A, Akita Y, Miyamoto A,Nakayama K et al. (2003). Genes Cells 8: 311–324.

Klein IK, Ritland SR, Burgart LJ, Ziesmer SC, Roche PC,Gendler SJ et al. (2000). Cancer Res 60: 2077–2080.

Krajewski S, Krajewska M, Shabaik A, Miyashita T, WangHG, Reed JC. (1994). Am J Pathol 145: 1323–1336.

La Porta CA, Porro D, Comolli R. (2002). Melanoma Res 12:297–307.

Laemmli UK. (1970). Nature 227: 680–685.Lee YJ, Soh JW, Dean NM, Cho CK, Kim TH, Lee SJ et al.

(2002). Cell Growth Differ 13: 237–246.Leontieva OV, Black JD. (2004). J Biol Chem 279: 5788–5801.Lewis AE, Susarla R, Wong BC, Langman MJ, Eggo MC.

(2005). Cell Signal 17: 253–262.Li W, Jiang YX, Zhang J, Soon L, Flechner L, Kapoor V et al.

(1998). Mol Cell Biol 18: 5888–5898.Li W, Michieli P, Alimandi M, Lorenzi MV, Wu Y, Wang LHet al. (1996). Oncogene 13: 731–737.

Li W, Zhang J, Bottaro DP, Pierce JH. (1997). J Biol Chem272: 24550–24555.

Lin SY, Liang YC, Ho YS, Tsai SH, Pan S, Lee WS. (2002).Mol Carcinog 35: 21–28.

Lipkin M. (1971). Cancer 28: 38–40.Liu JJ, Chao JR, Jiang MC, Ng SY, Yen JJ, Yang-Yen HF.

(1995). Mol Cell Biol 15: 3654–3663.Liu WS, Heckman CA. (1998). Cell Signal 10: 529–542.Livneh E, Fishman DD. (1997). Eur J Biochem 248: 1–9.Lu Z, Hornia A, Jiang YW, Zang Q, Ohno S, Foster DA.

(1997). Mol Cell Biol 17: 3418–3428.MacLachlan TK, Sang N, Giordano A. (1995). Crit Rev

Eukaryot Gene Exp 5: 127–156.Mandal M, Olson DJ, Sharma T, Vadlamudi RK, Kumar R.

(2001). Gastroenterology 120: 71–78.McCracken MA, Miraglia LJ, McKay RA, Strobl JS. (2003).

Mol Cancer Ther 2: 273–281.McGarrity TJ, Peiffer LP. (1994). Dig Dis Sci 39: 458–463.McMillan L, Butcher SK, Pongracz J, Lord JM. (2003). Br J

Cancer 88: 748–753.Meyer E, Vollmer JY, Bovey R, Stamenkovic I. (2005). Cancer

Res 65: 4261–4272.Milde-Langosch K, Bamberger AM, Methner C, Rieck G,

Loning T. (2000). Int J Cancer 87: 468–472.Moss SF, Liu TC, Petrotos A, Hsu TM, Gold LI, Holt PR.

(1996). Gastroenterology 111: 1425–1432.Nabha SM, Glaros S, Hong M, Lykkesfeldt AE, Schiff R,

Osborne K et al. (2005). Oncogene 24: 3166–3176.Nakagawa M, Oliva JL, Kothapalli D, Fourier A, Assoian

RK, Kazanietz MG. (2005). J Biol Chem 280: 33926–33934.Newton AC. (1997). Curr Opin Cell Biol 9: 161–167.Nishizuka Y. (1995). FASEB J 9: 484–496.Ogura E, Senzaki H, Yamamoto D, Yoshida R, Takada H,

Hioki K et al. (1999). Oncol Rep 6: 365–369.Page K, Li J, Corbit KC, Rumilla KM, Soh JW, Weinstein IBet al. (2002). Am J Respir Cell Mol Biol 27: 204–213.

Palmiter RD. (1994). Proc Natl Acad Sci USA 91: 1219–1223.

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3137

Oncogene

Parekh D, Ziegler W, Yonezawa K, Hara K, Parker PJ. (1999).J Biol Chem 274: 34758–34764.

Perletti G, Marras E, Dondi D, Osti D, Congiu T, Ferrarese Ret al. (2005). Int J Cancer 113: 42–53.

Perletti G, Marras E, Osti D, Felici L, Zaro S, de Eguileor M.(2004). J Cell Mol Med 8: 563–569.

Perletti GP, Concari P, Brusaferri S, Marras E, Piccinini F,Tashjian Jr AH. (1998). Oncogene 16: 3345–3348.

Perletti GP, Marras E, Concari P, Piccinini F, Tashjian Jr AH.(1999). Oncogene 18: 1251–1256.

Pinto M, Robine-Leon S, Appay M-D, Kedinger M, TriadouN, Dussaulx E et al. (1983). Biol Cell 47: 323–330.

Pongracz J, Webb P, Wang K, Deacon E, Lunn OJ, Lord JM.(1999). J Biol Chem 274: 37329–37334.

Polyak K, Hamilton SR, Vogelstein B, Kinzler KW. (1996).Am J Pathol 149: 381–387.

Potten CS, Wilson JW, Booth C. (1997). Stem Cells 15: 82–93.Qiao L, Koutsos M, Tsai LL, Kozoni V, Guzman J, Shiff SJet al. (1996). Cancer Lett 107: 83–89.

Ruemmele FM, Dionne S, Qureshi I, Sarma DS, Levy E,Seidman EG. (1999). Cell Death Differ 6: 729–735.

Rybin VO, Sabri A, Short J, Braz JC, Molkentin JD, SteinbergSF. (2003). J Biol Chem 278: 14555–14564.

Ryves WJ, Evans AT, Olivier AR, Parker PJ, Evans FJ.(1991). FEBS Lett 288: 5–9.

Santiago-Walker AE, Fikaris AJ, Kao GD, Brown EJ,Kazanietz MG, Meinkoth JL. (2005). J Biol Chem 280:32107–32114.

Sauma S, Yan Z, Ohno S, Friedman E. (1996). Cell GrowthDiffer 7: 587–594.

Savickiene J, Gineitis A, Stigbrand T. (1999). Cell Death Differ6: 698–709.

Saxon ML, Zhao X, Black JD. (1994). J Cell Biol 126: 747–763.Scaglione-Sewell B, Abraham C, Bissonnette M, Skarosi SF,

Davidson NO, Wali RK et al. (1998). Cancer Res 58:1074–1081.

Scaglione-Sewell BA, Bissonnette M, Skarosi S, Abraham C,Brasitus TA. (2000). Endocrinology 141: 3931–3939.

Schmittgen TD, Zakrajsek BA, Mills AG, Gorn V, Singer MJ,Reed MW. (2000). Anal Biochem 285: 194–204.

Seki T, Matsubayashi H, Amano T, Shirai Y, Saito N, SakaiN. (2005). Genes Cells 10: 225–239.

Shanmugam M, Krett NL, Maizels ET, Murad FM, RosenST, Hunzicker-Dunn M. (2001). Cancer Lett 172: 43–53.

Shanmugathasan M, Jothy S. (2000). Pathol Int 50: 273–279.Sherr CJ. (1996). Science 274: 1672–1677.

Sherr CJ, Roberts JM. (1999). Genes Dev 13: 1501–1512.Sitailo LA, Tibudan SS, Denning MF. (2004). J Invest

Dermatol 123: 434–443.Soh JW, Weinstein IB. (2003). J Biol Chem 278: 34709–34716.Srivastava J, Procyk KJ, Iturrioz X, Parker PJ. (2002).

Biochem J 368: 349–355.Steinberg SF. (2004). Biochem J 384: 449–459.Thompson CB. (1995). Science 267: 1456–1462.Tillman DM, Izeradjene K, Szucs KS, Douglas L, Houghton

JA. (2003). Cancer Res 63: 5118–5125.Toyoda M, Gotoh N, Handa H, Shibuya M. (1998). Biochem

Biophys Res Commun 250: 430–435.Umar S, Sellin JH, Morris AP. (2000). Am J Physiol

Gastrointest Liver Physiol 278: G765–74.Verstovsek G, Byrd A, Frey MR, Petrelli NJ, Black JD.

(1998). Gastroenterology 115: 75–85.Villalba M. (1998). Neuroreport 9: 2381–2385.Wakino S, Kintscher U, Liu Z, Kim S, Yin F, Ohba M et al.

(2001). J Biol Chem 276: 47650–47657.Wali RK, Frawley Jr BP, Hartmann S, Roy HK, Khare S,

Scaglione-Sewell BA et al. (1995). Cancer Res 55: 5257–5264.Wang Q, Wang X, Evers BM. (2003). J Biol Chem 278:

51091–51099.Wang Q, Wang X, Zhou Y, Evers BM. (2006). Int J Cancer118: 326–334.

Wang X, Wang Q, Hu W, Evers BM. (2004). Oncogene 23:1885–1895.

Watters D, Garrone B, Coomer J, Johnson WE, Brown G,Parsons P. (1998). Biochem Pharmacol 55: 1691–1699.

Weinstein IB. (1991). Princess Takamatsu Symp 22: 277–283.Weller SG, Klein IK, Penington RC, Karnes Jr WE. (1999).

Gastroenterology 117: 848–857.Wetsel WC, Khan WA, Merchenthaler I, Rivera H, Halpern

AE, Phung HM et al. (1992). J Cell Biol 117: 121–133.Yin L, Bennani-Baiti N, Powell CT. (2005). J Biol Chem 280:

5533–5541.Yokoyama G, Fujii T, Tayama K, Yamana H, Kuwano

M, Shirouzu K. (2005). Biochem Biophys Res Commun 327:720–726.

Zang Q, Lu Z, Curto M, Barile N, Shalloway D, Foster DA.(1997). J Biol Chem 272: 13275–13280.

Zeng YX, el-Deiry WS. (1996). Oncogene 12: 1557–1564.Zezula J, Sexl V, Hutter C, Karel A, Schutz W, Freissmuth M.

(1997). J Biol Chem 272: 29967–29974.Zhu GH, Wong BC, Slosberg ED, Eggo MC, Ching CK, Yuen

ST et al. (2000). Gastroenterology 118: 507–514.

Mechanisms of action of tumor suppressor PKC-dSR Cerda et al

3138

Oncogene