Inhibition of Caspase Activity Does Not Preventthe Signaling Phase of Apoptosis in Prostate

Cancer Cells

Samuel R. Denmeade,1* Xiaohui S. Lin,2 Bertrand Tombal,3 andJohn T. Isaacs1,2

1Johns Hopkins Oncology Center, Johns Hopkins University School of Medicine,Baltimore, Maryland

2James Buchanan Brady Urological Institute, Johns Hopkins University School of Medicine,Baltimore, Maryland

3Department of Urology and Physiology, Catholic University of Louvain Medical School,Brussels, Belgium

BACKGROUND. Caspases are a family of cysteine proteases capable of characteristicallycleaving after an aspartic acid residue. Various members of the caspase family (e.g., caspases8 and 9) have been implicated as critical initiators in the signaling phase, while others (e.g.,caspases 3, 6,and 7) have been implicated in the effector or execution phase of apoptosis.Thapsigargin (TG) is capable of inducing cell proliferation-independent apoptosis of prostatecancer cells. This study was undertaken to determine if caspase inhibition can prevent TG- or5-fluorodeoxyuridine (5-FrdU)-induced apoptosis in prostate cancer cells.METHODS. Caspase activity was evaluated by Western blot analysis of the cleavage ofretinoblastoma (Rb) protein, a caspase substrate during TG-induced death of prostate cancercells. In addition, hydrolysis of caspase-specific fluorescent peptide substrates was assayed inlysates from TG-treated cells. Clonogenic survival assays were performed following treatmentof rat AT3 and human TSU-Pr1 prostate cancer cell lines with TG and 5-FrdU in the presenceand absence of peptide caspase inhibitors. AT3.1 cells transfected with the crmA gene, en-coding a viral protein with caspase-inhibitory activity, were also tested for clonogenic sur-vival following TG and 5-FrdU exposure.RESULTS. During treatment with TG, Rb is first dephosphorylated and then proteolyticallycleaved into 100-kDa and 40-kDa forms, indicative of caspase activity. A 6–8-fold increase inclass II (i.e., caspases 3, 7, and 10) hydrolysis of the caspase substrate Z-DEVD-AFC wasobserved after 24 hr of TG or 5-FrdU. AT3 cells expressing crmA (i.e., an inhibitor of caspases1, 4, and 8) were not protected from apoptosis induced by TG or 5-FrdU. The caspase inhibi-tors Z-DEVD-fmk (i.e., an inhibitor of caspases 3, 7, and 10) and Z-VAD-fmk (i.e., a generalcaspase inhibitor) were also unable to protect TSU and AT3 cells from apoptosis induced byTG or 5-FrdU.CONCLUSIONS. Caspase activation may play a role in the downstream effector phase of theapoptotic cascade; however, in this study, caspase inhibition did not prevent the signalingphase of apoptosis induced by two agents with distinct mechanisms of cytotoxicity, TG or

Abbreviations: TG, thapsigargin; 5-FrdU, 5-fluorodeoxyuridine;cmk, chloromethylketone; fmk, fluoromethylketone; AFC, 7-amino-4-trifluorocoumarin.Grant sponsor: CaPCure.*Correspondence to: Samuel R. Denmeade, Johns Hopkins Oncol-ogy Center, Johns Hopkins School of Medicine, 600 N. Wolfe St.,Baltimore, MD 21287.

The Prostate 39:269–279 (1999)

© 1999 Wiley-Liss, Inc.

5-FrdU. These results suggest that caspase inhibition by recently described endogenouscaspase inhibitors should not lead to development of resistance to TG. A strategy for targetingTG’s unique cytotoxicity to metastatic prostate cancer cells is currently under development.Prostate 39:269–279, 1999. © 1999 Wiley-Liss, Inc.

INTRODUCTION

Although specific steps can vary between differentcell types and with different inducing agents, apopto-sis comprises two distinct stages: an initial, potentiallyreversible, “decision to die” or signaling phase, and anirreversible “killing” or execution phase [1]. The sig-naling phase can be activated either by sufficient in-jury to the cell by various exogenous agents (e.g., ra-diation, chemicals, or viruses) or by changes in thelevels of a series of endogenous signals (e.g., hor-mones and growth/survival factors) [2]. In contrast tothe variable nature of the biochemical cascade in-volved in the signaling phase, the key features of theexecution phase are common to cells undergoing ap-optosis and characteristically involve fragmentation ofthe nucleus and the cell into apoptotic bodies [3]. Al-though the common key features of the executionphase are independent of the nature of the inducingstimuli, they may be mediated by the activation ofmultiple enzymes (e.g., endonucleases, poly(ADP-ribosyl) polymerase, caspases, nuclear scaffold prote-ases, cytochrome c, or transglutaminases) that varybetween different cell types and different inducingagents [4,5].

While an understanding of the general apoptoticprocess is rapidly progressing [6–9], resolving thecritical biochemical steps in the apoptosis induced byparticular agents in specific cell types is needed if thispathway is to be of practical value for the develop-ment of effective therapies for specific cancers. Thisissue is particularly critical for prostate cancers be-cause currently there is no treatment that significantlyprolongs survival in men with metastatic prostate can-cer [10].

Androgen ablation therapy, while of substantialpalliative benefit, eventually fails because the meta-static prostate cancer within an individual patient isheterogeneously composed of clones of both andro-gen-dependent and -independent cancer cells [11].Due to these androgen-independent prostatic cancercells, the patient is no longer curable using the numer-ous chemotherapeutic agents that have been testedover the past 30 years [10]. Standard antiproliferativechemotherapeutic agents may be ineffective againstandrogen-independent prostatic cancers because thesecancers have a low proliferative rate. Berges et al. [12]demonstrated that the median daily proliferative rateof prostate cancer cells within lymph node or bonemetastases was <3.0% per day. Newer agents are

needed that target the >95% of prostate cancer cellswithin a given metastatic site that are not immediatelyproliferating.

Previously we demonstrated that thapsigargin(TG), a potent and selective inhibitor of the sarcoplas-mic/endoplasmic reticulum Ca2+ ATPase (SERCA)pump [13], induces programmed cell death (PCD)/apoptosis of metastatic, androgen-independent hu-man and rodent prostatic cancer cells in a prolifera-tion-independent manner [14]. TG is also able to in-duce PCD in proliferatively quiescent (i.e., G0) humanprostate cancer cells in primary culture without re-cruitment into the proliferative cell cycle [15]. TGtreatment induces an epigenetic reprogramming ofprostate cancer cells undergoing apoptosis. Within 24hr of TG exposure, increased expression of calmodulin[16], a-prothymosin [16], c-myc [16], and gadd153 [15]occurs at the same time, with decreased expression ofother proteins such as H-ras and cyclin D [16]. Whenprostate cancer cells are exposed to TG, an initial 4–6-fold increase in intracellular calcium [Ca2+]i (i.e., 50nM increasing to 200–300 nM) is observed within 1 hrof treatment [14]. Recently, we described a methodthat allows for longitudinal, kinetic analysis of [Ca2+]ichanges in cells treated with TG [17]. This study dem-onstrated that following this initial rise in [Ca2+]i, cal-cium levels return to baseline within 3–12 hr. A secondrise in [Ca2+]i to 10–15 mM then occurs asynchronouslyin individual cells between 16–48 hr [18]. This micro-molar rise in [Ca2+]i occurs prior to the morphologicchanges and DNA fragmentation in cells undergoingapoptosis [17]. The TG-induced micromolar rise in[Ca2+]I initiates the apoptotic signaling cascade thatincludes cytochrome c release from mitochondria andendonuclease-mediated DNA fragmentation [18]. Ad-ditional studies have demonstrated that by loweringextracellular calcium during the period of the second-ary [Ca2+]i rise, the secondary rise is prevented andthe cells do not die [18].

The role of protease activity in TG-induced apop-tosis has not yet been defined. TG-induced death of ahuman breast cancer line [19] and a human monocyteline [20] was recently demonstrated, and this TG-induced cell death was associated with induction ofcaspase activity. Within the last few years, a largenumber of studies have documented the involvementof the caspase family of proteases in apoptosis in-duced by other agents (reviewed by Thornberry andLazebnik [9], Cohen [21], Thornberry [22], and Villa et

270 Denmeade et al.

al. [23]). Members of the caspase family have beendescribed that function as either initiators of the sig-naling phase or as downstream effectors of the execu-tion phase of the apoptotic process [23]. A role forcaspases in the apoptosis of prostate cancer cells wasalso recently demonstrated. Marcelli et al. [24] de-scribed activation of caspase-7 during lovastatin-induced apoptosis of LNCaP cells. In a further study,this group also demonstrated that apoptosis can beactivated by using a replication-deficient adenoviralvector to overexpress caspase-7 in LNCaP cells [25]. Inaddition, Bowen et al. [26] demonstrated that apopto-sis induction in human prostate cancer cell linesDU145 and TSU-Pr1 by okadaic acid is associated withcaspase-3 activation. Caspase-3 activation also occursin apoptosis of TSU-Pr1 cells following radiation ex-posure but is not observed in DU145 cells, which donot undergo apoptosis following radiation [26].DU145 cells are known to be negative for expression ofthe retinoblastoma protein (Rb), and Bowen et al. [26]went on to demonstrate that DU145 cells that havebeen genetically engineered to reexpress Rb can bemade sensitive to radiation-induced apoptosis. Inter-estingly, however, in these Rb-expressing DU145 cells,radiation-induced apoptosis is not associated with anycaspase activation [26].

Expression of Rb is also required for prostate cancercells to undergo apoptosis by phorbol ester activationof protein kinase C [27]. In this phorbol ester-inducedmodel, activation of caspases was also implemented[27]. In contrast to the requirement for Rb expressionfor radiation and phorbol ester-induced apoptosis ofprostate cancer cells, Furuya et al. demonstrated thatTG-induced apoptosis occurs in all types of prostatecancer cells regardless of Rb or p53 status [14]. Forexample, TG induces apoptosis in both Rb-negativeDU145 cells and p53-negative TSU cells [14].

Recently, Srikanth and Kraft [28] demonstrated thatLNCaP cells expressing crmA, a viral encoded serpinprotein that inhibits caspases (i.e., caspases 1, 4, 6, and8) [21], were resistant to apoptosis induced by topoi-somerase I inhibitors, alkylating agents, phorbol ester,TNF-alpha, and staurosporine. This observation isparticularly important, since Deveraux et al. [29] dem-onstrated that there are nonviral, endogenous intra-cellular proteins termed “inhibitors of apoptosis”(IAPs) that can inhibit caspase activity. In addition, itwas demonstrated that when activated by various sig-naling pathways, various ubiquitously expressed pro-tein kinases (e.g., PKB or Akt) can phosphorylatecaspases and thereby also inhibit their enzymatic ac-tivity [30]. Thus, if TG induction of apoptosis is criti-cally dependent on caspase activity, then there are avariety of methods to become resistant to such deathbased on inhibition of caspase activity. Because we are

attempting to develop a TG-based therapy for meta-static prostate cancer, it will be critical to determinewhat role caspase activation may play in the TG-induced apoptosis of prostate cancer cells. The goal ofthe present study was to clarify the requirement forcaspase activation in TG-induced apoptosis of prostatecancer cells.

MATERIALS AND METHODS

Materials

Ac-Tyr-Val-Ala-Asp-7-amino-4-trifluoromethyl-coumarin (AFC) (Ac-YVAD-AFC), Z-Asp-Glu-Val-Asp-AFC (Z-DEVD-AFC), cell permeant Z-Asp-Glu-Val (OMe)-Asp-fluoromethyl ketone (Z-DEVD-fmk),and Z-Val(OMe)-Ala-Asp-fmk (Z-VAD-fmk) werefrom Enzyme Systems Products (Livermore, CA). Ac-Tyr-Val-Ala-Asp-chloromethyl ketone (Ac-YVAD-cmk) and Ala-Pro-Phe-cmk (APF-cmk) were fromBachem (Torrance, CA). Thapsigargin, Na-Tosyl-phenylalanine chloromethyl ketone (TPCK) and Na-Tosyl-lysine chloromethyl ketone (TLCK) were fromCalbiochem (CA). 5-fluorodeoxyuridine (5-FrdU) andall other reagents were from Sigma Chemical Co. (St.Louis, MO). CrmA vectors consisting of crmA cDNAand antisense cDNA cloned into a B-actin ST neo Bvector were transfected into AT3 cells as previouslydescribed [31]. Clones expressing crmA were identi-fied by Western blotting using primary rabbit anticrmA antibodies, as previously described [31]. CrmAvectors and crmA antibody were a gift from Dr. J.Yuan (Massachusetts General Hospital, Boston, MA).

Cell Lines

AT3.1 and TSU-Pr1 cell lines were maintained inRPMI-1640 medium (M.A. Bioproducts, Walkersville,MD) supplemented with 10% fetal calf serum (Hy-Clone, Logan, UT), 1 mM glutamine, and 100 IU/ml ofpotassium penicillin G and streptomycin sulfate (M.A.Bioproducts) at 37°C equilibrated with 5% CO2/95%air. The characteristics of these cell lines were de-scribed previously [32,33].

Western Blot Analysis of Rb

Preparation of cell lysates was performed as de-scribed previously [15]. Rb immunoblotting was per-formed as described previously [34]. Briefly, superna-tants from cell lysates were incubated with the mousemonoclonal anti-Rb antibody bound to agarose beads(i.e., clone C36-Rb (Ab-1)-agarose, Oncogene Science).This monoclonal antibody recognizes both phosphor-ylated and unphosphorylated forms of Rb protein[34]. Specific protein detection on the blot was per-

Inhibition of Caspase Activity and Apoptosis 271

formed via the ECL system (Amersham) according tothe manufacturer’s instructions, and the proteins werevisualized by exposure to X-ray film.

Caspase Enzymatic Activity Assay

Lysates of control or treated AT3.1 cells were pre-pared by placing equal numbers of cells (i.e., 106) intolysis buffer consisting of 25 mM HEPES, pH 7.5, 2 mMMgCl2, 1 mM EGTA, 5 mM DTT, 1 mM PMSF, and 10mg/ml leupeptin. Cells were broken in a dounce ho-mogenizer, then frozen and thawed twice followed bycentrifugation at 30,000g, and the supernatant wasused for assay of caspase activity. The caspase sub-strates were then added to a final concentration of 50mM. Hydrolysis of the caspase substrates was fol-lowed kinetically (ex 400 nm and em 505 nm) in acuvette using a spectrophotometer from Photon Tech-nologies International (PTI) (New Brunswick, NJ)equipped with a D-104 photomultiplier tube (PTI). Todetermine specific caspase 1 activity, the caspase1-specific inhibitor Z-YVAD-cmk (100 mM) wasadded to a matched sample, and the activity was sub-tracted from that of the sample containing lysate andsubstrate alone. Similarly, to determine the caspase3-specific activity, the inhibitor Z-DEVD-fmk (100mM) was utilized. Relative enzyme activities werethen determined from the ratio of treated cells at in-dicated times to control cells.

Clonogenic Survival Assays

The rat prostate cancer cell line AT3.1 cells and hu-man prostate cancer cell line TSU-Pr1 were treated forindicated times with either TG or 5-FRDU in the ab-sence or presence of protease inhibitors. Cell mediacontaining floating cells and adherent cells removedby trypsin/EDTA were combined and centrifuged,and cell numbers were determined using a Coultercounter (Hilaeah, FL). Cells were then plated in 60-mm dishes (i.e., 100 cells/dish for AT3 and 200 cells/dish for TSU) and after 6 days, plates were washedwith phosphate-buffered saline (PBS) and stainedwith 5% crystal violet/25% methanol. Colonies werecounted using a colony counter. Percentage of clono-genic survival was determined from the ratio of colo-nies in treated groups to colonies from vehicle-treatedcontrols.

RESULTS AND DISCUSSION

Effect of TG and 5-FrdU on Clonogenic Survival,Cell Viability, and Cell Cycle Progression

To determine the kinetics of cytotoxicity of 5-FrdUand TG, the rat prostate cancer cell line AT3.1 was

treated for indicated amounts of time, and cell viabil-ity (e.g., Trypan blue exclusion) and clonogenic sur-vival were determined (Fig. 1). After 24 hr of 5-FrdU(100 nM) treatment, >95% of the cells were irreversiblycommitted to loss of their clonogenic ability, while cellviability remained at >98% of controls (Fig. 1B). ForTG (500 nM)-treated cells, a similar irreversible com-mitment to loss of clonogenic ability (i.e., >90%) re-quired 48 hr of exposure, a time when no decrease inthe percentage of viable cells was demonstrated (Fig.1A). For both 5-FrdU and TG treatments, even thoughcommitted to loss of clonogenic ability, more than 72hr of continuous exposure was required before 50% ofthe cells underwent lysis into apoptotic bodies. Theseresults defined the minimal exposure times for 5-FrdU(i.e., 24 hr) and TG (i.e., 48 hr) needed to commit AT3rat prostate cancer cells to loss of clonogenic ability.

Ki-67 is a nonhistone nuclear protein expressed inall phases of the cell cycle (i.e., G1, S, G2, and mitosis)

Fig. 1. Comparison of time course of loss of clonogenic abilityvs. decrease in number of viable cells by Trypan blue exclusionassay. A: AT3 cells exposed to 500 nM TG. B: AT3 cells exposedto 100 nM 5-FrdU (n = 5 for each time point ± standard error).

272 Denmeade et al.

[35]. Only proliferatively quiescent cells in G0 arenegative for Ki-67 nuclear expression. In a previousstudy, monoclonal antibodies against the Ki-67 anti-gen (i.e., M1B1) were used in immunohistochemicalstaining to demonstrate that in human prostate cancerTSU-Pr1cells, >96 ± 2% of cells are arrested in G0 (i.e.,are Ki-67-negative) after 24–36-hr exposure to 500 nMTG. Since M1B1 does not recognize the rodent equiva-lent of Ki-67, in a further study, flow cytometric analy-sis of AT3.1 cells following TG exposure was used todemonstrate that 90 ± 7% of cells were in G0/G1 fol-lowing 24-hr exposure to 500 nM TG [36]. Using amore recently described monoclonal antibody (i.e.,M1B5) [35] that recognizes the rodent homologue ofKi-67, immunohistochemical staining demonstratedthat >96% of AT3 cells were negative for Ki-67 (i.e., thecells were in G0) following 24-hr exposure to TG. Incontrast, when AT3 cells were treated for 24 hr with100 nM 5-FrdU, <5% of the cells were negative forKi-67 expression using the M1b5 antibody. These re-sults demonstrate that TG initially induces growth ar-rest of AT3 cells into G0 and then activates these G0cells to undergo apoptosis in a proliferation-indepen-dent manner. In contrast, 5-FrdU does not arrest AT3cells in G0 but allows their continued entrance intoand progression through the cell cycle. Thus, 5-FrdUinduces the apoptotic death of these cells during cellproliferation.

Effect of TG and 5-FrdU on the Phosphorylationand Proteolysis of Rb

Because TG but not 5-FrdU induces a G0 arrest inproliferating AT3 cells, differential regulation of cellcycle-dependent proteins should be observable in cellstreated with TG vs. cells treated with 5-FrdU. The ret-inoblastoma protein (Rb) is an example of a proteinthat is phosphorylated in cells entering (i.e., in G1) andprogressing through the cell cycle but which becomesdephosphorylated when cells are in G0 [37]. On thebasis of the data in Figure 1, AT3 cells were treatedfrom 0–24 hr with 5-FrdU (100 nM) or from 0–48 hrwith TG (500 nM), and Western blot analysis of Rbprotein expression was performed using an antibodythat recognizes both the phosphorylated (top band,top half, Fig. 2) and dephosphorylated (middle band,top half, Fig. 2) forms of Rb. In the top half of Figure2, cells were treated from 1–24 hr with 5-FrdU (i.e., thetime needed to irreversibly commit cells to loss of theirclonogenic ability by >90%); Rb remains in the phos-phorylated state (i.e., top band, top half, Fig. 2), con-sistent with the fact that >95% of cells remain in thecycle.

In contrast, in AT3 cells treated with 500 nM TG(bottom half, Fig. 2), the phosphorylated Rb (i.e., top

band, bottom half, Fig. 2) is the main form of the pro-tein that can be visualized until 12 hr of TG exposure.At 12 hr of TG exposure, the unphosphorylated Rb(i.e., middle band, bottom half, Fig. 2) is now the majorform of the protein. By 24 hr of TG exposure, the totalamount of Rb protein is decreased. To better visualizethe remaining forms of Rb, the blot in Figure 2 wasreexposed to film for a longer time period to bettervisualize the Rb bands at the 24- and 36-hr time pointsof TG treatment (see inset, bottom right, Fig. 2). Withlonger exposure of the blot, two new bands were de-tected. Recently, it was demonstrated that Rb iscleaved early in apoptosis, with loss of its carboxyterminal 42 amino acids [38]. The lowest band of ∼100kDa that can be seen just below the dephosphorylatedRb (i.e., middle band, Bottom half, Fig. 2) correspondsto this slightly lower molecular weight carboxy-terminal truncated form of Rb. In addition, furthercleavage of Rb into 43-kDa and 30-kDa fragments wasalso previously described [39,40]. In the present study,the ∼40-kDa band was detected by 24 hr and increasedin intensity by 36-hr treatment with TG (bottom half atleft, Fig. 2). The time course of cleavage of Rb into

Fig. 2. Western blot analysis of Rb expression, using Rb anti-body-recognizing phosphorylated and unphosphorylated forms ofRb. Hash lines denote three different molecular-weight forms ofRb. Top bands (i.e., ppRb) are the phosphorylated form, middlebands are the unphosphorylated form, and bottom bands are theRb protein lacking 42 amino-acid carboxy-teminus. Above: Timecourse of Rb expression in AT3 cells during exposure to 0.1 µM5-FrdU. Below: Time course of Rb expression in AT3 cells duringexposure to 500 nM TG. Inset: Lanes corresponding to 24- and36-hr time points during TG exposure were cut from membraneand reexposed to film for longer time periods to visualize lower,40-kDa Rb cleavage product.

Inhibition of Caspase Activity and Apoptosis 273

100-kDa and 40-kDa fragments corresponds to the in-duction of growth arrest and occurs at the time whenTG-treated AT3 cells begin to irreversibly commit toundergo apoptosis, as evidenced by their loss of clo-nogenic ability (Fig. 1B).

Recent studies demonstrated that Rb cleavage dur-ing apoptosis is secondary to proteolytic processing bycaspases [20,38–40]. Caspases 3 and/or 7 have beenproposed to cleave the carboxy terminal fragment ofRb [20]. Additional cleavage into the 43-kDa and 30-kDa forms has been suggested to be due to additionalmembers of the caspase family [20]. Thus, the resultsin Figure 2 suggest that caspases are activated duringTG-induced apoptosis. In contrast, AT3 cells treatedwith 5-FrdU do not undergo growth arrest, and Rbdephosphorylation is not observed. Cleavage of Rb isalso not observed (Fig. 2, top half), even at time pointsup to 72 hr (data not shown), demonstrating that in5-FrdU-induced apoptosis, proteolysis of Rb bycaspases does not occur, possibly because Rb remainsin a phosphorylated, proteolysis-resistant state [41].

Kinetics of Caspase Activation During ApoptosisInduced by TG or 5-FrdU

The caspase family includes more than 13 proteinswith cysteine protease activity characteristically ableto cleave after an aspartic acid [42]. Based upon sub-strate specificity, three classes of caspase have beendefined. Only the class I caspases (i.e., caspases 1, 4,and 5) can efficiently hydrolyze the YVAD-AFC sub-strate and are irreversibly inhibited by Z-YVAD-cmk[21,22]. In contrast, the substrate DEVD-AFC is effi-ciently hydrolyzed only by the class II caspases (i.e.,caspases 3, 7, and 10) and irreversibly inhibited byZ-DEVD-fmk [21,22].

To evaluate if either TG- or 5-FrdU-induced apop-tosis is associated with caspase activation, AT3 cellswere treated with either 500 nM TG or 100 nM 5-FrdUand assayed at several time points for hydrolysis ofcaspase class I or II substrates (Fig. 3). To determinethe amount of total proteolysis specifically attributableto caspase activity, only the appropriate caspase-specific inhibitors (i.e., Z-YVAD-cmk for class I andZ-DEVD-fmk for class II) were added to matchedsamples containing the corresponding caspase sub-strate.

Following either TG or 5-FrdU exposure, no statis-tically significant increase in hydrolysis of the classI-specific caspase substrate Ac-YVAD-AfC was ob-served, even though baseline levels were measurablein untreated cells (Fig. 3A). In contrast, when the class2 caspase-specific substrate Z-DEVD-AFC was uti-lized, a statistically significant 6–8-fold increase in ac-tivity compared to untreated cells was detected fol-

lowing 24-hr exposure to either TG or 5-FrdU (Fig.3B). These results demonstrate that class II caspase(e.g., caspase 3 and/or 7) activity is increased in cellsinduced to undergo apoptosis by both TG and 5-FrdU.The increase in this activity occurs at a time when AT3cells begin to irreversibly commit to the loss of theirclonogenic ability following drug exposure, but beforethe cells have progressed to lose the ability to excludeTrypan blue.

Fig. 3. Assay of caspase enzymatic activity in AT3 cell lysatesduring exposure to 500 nM TG or 100 nM 5-FrdU. A: Relativehydrolysis of class I (i.e., caspases 1, 4, and 5) caspase substrateAc-YVAD-AFC. B: Relative hydrolysis of class II (i.e., caspases 3, 7,and 10) caspase substrate Z-DEVD-AFC. Activity in untreatedcontrol cell lysates was normalized to 1 and relative activitiesrepresent ratio of treated to control (n = 3 ± standard error, donein duplicate).

274 Denmeade et al.

Effect of Caspase Inhibition on ApoptosisFollowing Exposure to TG or 5-FrdU

A large number of recent studies have suggestedthat apoptosis following exposure to either cytotoxicagents, radiation, or growth factor withdrawal can bekinetically slowed or even prevented by caspase inhi-bition [21–23]. To demonstrate apoptosis inhibition,these studies measured decreases in either dye exclu-sion, TUNEL end-labeling of fragmented genomicDNA, staining of apoptotic nuclei with DNA-depen-dent fluorochromes to visualize chromatin condensa-tion, or electrophoretic analysis of DNA fragmenta-tion in cells cotreated with caspase inhibitors for onlya relatively short observation period (i.e., hours to 1day). However, the ability of caspase inhibitors to re-tard the kinetics of detection of these markers of ap-optosis induction does not provide information aboutwhether apoptosis is actually prevented or simply ki-netically delayed by caspase inhibitors.

To determine this last point, clonogenic survivalassays were performed on AT3 cells. Based on the datain Figure 1, in these clonogenic survival assays, cellswere treated with either TG for 48 hr or 5-FrdU for 24hr in the absence or presence of caspase inhibitors. Thecells were then washed free of all drugs, and the sub-sequent ability to survive and form clonogenic colo-nies was determined in drug-free media. If specificcaspase activity is required to initiate apoptosis byeither TG or 5-FrdU, caspase inhibition during the ir-reversible commitment period (i.e., 48 hr for TG, and24 hr for 5-FrdU in AT3 cells) should be sufficient toprevent the completion of apoptosis when the cells aresubsequently placed in a drug-free media. In contrast,if these caspase activities are involved in the effectorphase and not in the signaling phase (i.e., in the ini-tiation of commitment to cell death), then the presenceof caspase inhibitors only during the signaling phasewill not prevent subsequent death (i.e., loss of clono-genic survival) when these committed cells are placedin drug-free media. Thus, clonogenic survival assayswere utilized to determine if caspase activity is essen-tial for the signal phase of apoptosis to occur or is onlyone of a number of effector pathways associated withapoptosis.

To determine the effect of caspase inhibition onsubsequent clonogenic survival following TG or5-FrdU treatment, two different peptide caspase in-hibitors were utilized. First, the class II (i.e., caspases 3,7, and 10) selective inhibitor Z-DEVD-fmk was chosenbecause, of all the caspases, caspases 3 and 7 are thebest-characterized caspase members involved in pros-tate cancer cell apoptosis [24–26]. Caspase 3 remainsone of the best-characterized members of the caspasefamily and is thought to be one of the key effectors of

apoptosis. When activated, caspase 3 is able to cleavea variety of cellular proteins including PARP [43,44],Rb [38–40], DNA-PKcs [45], and gelsolin [46].

To broadly inhibit other members of the caspasefamily that may also be activated during the signalingphase of apoptosis, the general caspase inhibitor Z-VAD-fmk was utilized, since it has been demonstratedto inhibit all of the human caspases [21–23]. Caspasesinhibited by Z-VAD-fmk include the recently de-scribed caspase component of the vertebrate “apopto-some” consisting of cytochrome-c, apoptosis-activating factor-1 (Apaf-1), and procaspase-9 [47].Through interactions with Apaf-1 and cytochrome c,caspase 9 is activated, and this activation appears to bean important component in apoptosis induced by che-motherapeutic agents [48]. Caspase 9 can also activatedownstream effector caspases such as caspase 3[48,49]. In addition, another caspase-activating proteintermed “apoptosis-inducing factor” (AIF) was re-cently described [50]. Like cytochrome c, AIF is alsoreleased from mitochondria during apoptosis, and theactivity of AIF is also inhibited by the general caspaseinhibitor z-VAD-fmk [50].

To verify that both of the peptidyl caspase inhibi-tors (i.e., Z-DEVD-fmk and Z-VAD-fmk) actually wereable to enter cells and inhibit caspase activity, AT3cells were incubated with the inhibitors for 2 hr andcell lysates were prepared. Caspase activity in theseinhibitor-treated cell lysates was determined using thecaspase substrate Z-DEVD-AFC and was compared toactivity in untreated control cells. In cells pretreatedwith the caspase inhibitors, caspase activity was com-pletely inhibited, demonstrating that the inhibitorshad entered cells as expected. To determine whetherthe inhibition of caspases can increase the survival ofAT3 cells following exposure to TG or 5-FrdU, cellswere preincubated with each of the caspase inhibitorsand then exposed to drug for the indicated amount oftime, followed by clonogenic survival assay. No sig-nificant decrease in clonogenic survival was observedwhen AT3 cells were treated with up to 200–400 mM ofthe inhibitors alone (data not shown). At a concentra-tion of 100 mM, neither the Z-DEVD-fmk nor the Z-VAD-fmk inhibitor was able to significantly increaseclonogenic survival of AT3 cells following exposure500 nM TG for 48 hr (Fig. 4A). A similar result wasobserved following 24-hr exposure of AT3 cells to 100nM 5-FrdU in combination with even higher inhibitorconcentrations (Fig. 4B).

In addition to the peptide-based caspase inhibitors,AT3 cells were also transfected with the crmA gene todetermine if expression of this intracellular, protein-based caspase inhibitor could effect clonogenic sur-vival following TG or 5-FrdU. When ectopically ex-pressed, the crmA gene has been shown to prevent

Inhibition of Caspase Activity and Apoptosis 275

apoptosis in a variety of systems. The product of thecrmA gene is a protein that is an effective inhibitor ofcaspases 1, 4, 6, and 8, but is a poor inhibitor ofcaspases 3, 7, and 10 (reviewed by Cohen [21]). Re-cently, LNCaP cells transfected with crmA were dem-onstrated to be resistant to apoptosis induced by avariety of cytotoxic agents in vitro, and the crmAtransfectants were also resistant to androgen ablationin vivo [28]. To determine whether crmA could inhibitTG-induced apoptosis, crmA vectors carrying thecrmA gene in the sense and antisense orientation weretransfected into AT3 cells, and crmA expressing cloneswere selected based on expression detected by West-ern blot analysis (Fig. 5, inset). The crmA sense andantisense clones were then treated with TG and5-FrdU, and clonogenic survival was compared tosimilarly treated wild-type AT3 cells (Fig. 5, inset). Nosignificant difference in clonogenic survival was seen

in either of the crmA sense clones when compared toeither control AT3 or antisense cells following expo-sure to TG or 5-FrdU (Fig. 5).

The combined results using the peptide-basedcaspase inhibitors and crmA transfectants, when com-bined with early results documenting increasedcaspase activity in treated cells, indicate that whilecaspases may play a role in the effector arm of apop-tosis induced by TG or 5-FrdU, caspase activity is notessential for the signaling phase of apoptosis in AT3rat prostate cancer cells exposed to these two agents.

These results appear contrary to the previouslycited studies discussed in the Introduction in whichhuman prostate cancer cells were suggested to requirecaspase activation for apoptosis. Therefore, to testwhether the negative findings in the rat AT3.1 prostatecancer cells were species-specific, similar caspase in-hibitor studies were performed on TSU-Pr1 humanprostate cancer cells, because caspase activation waspreviously documented in TSU cells following radia-tion exposure [26]. In similar analysis to that presentedin Figure 1, it was determined that longer exposure(i.e., 72 hr) of TSU cells to 500 nM TG is required toachieve a >90% loss of clonogenic ability (data notshown). On the basis of this finding, TSU cells wereexposed to 500 nM TG for 72 hr in the absence orpresence of 100 mM concentration of either Z-VAD-fmk or Z-DEVD-fmk, and then clonogenic survival as-says were performed. These studies demonstratedthat caspase inhibition was unable to increase clono-genic survival of TSU cells following TG exposure(Fig. 6). The combined results with the TSU and AT3cell lines demonstrate that, although TG-inducedcaspase activation does occur, TG-treated cells un-

Fig. 4. Clonogenic survival assays performed on AT3 cells fol-lowing treatment with drug ± caspase or protease inhibitors.A: AT3 cells treated for 48 hr with 500 nM TG ± inhibitors andthen placed into survival assays. B: AT3 cells treated for 24 hr with100 nM 5-FrdU ± inhibitors and then placed into survival assays.Percent clonogenic survival is ratio of colonies observed in treatedvs. untreated controls (n = 5 ± standard error, done in duplicate).

Fig. 5. AT3 cells transfected with crmA sense or antisense con-structs were treated with 500 nM TG (48 hr) or 100 nM 5-FrdU(24 hr) and assayed for clonogenic survival compared to wild-typeAT3 cells. Percent clonogenic survival is ratio of colonies observedin treated vs. untreated controls (n = 5 ± standard error, done induplicate). Inset: Western blot analysis of crmA expression. S,AT3 cells with crmA sense construct; WT, wild-type AT3; AS,AT3 cells with antisense construct.

276 Denmeade et al.

dergo apoptosis even when caspase activity is inhib-ited.

Role of Other Apoptosis-Associated Proteases inTG-Induced Cell Death

In previous studies using thymocytes, a Ca2+-regulated serine protease activity has been describedwith considerable selectivity for nuclear lamin hydro-lysis [51]. In thymocytes exposed to TG, treatmentwith both the caspase inhibitor Z-YVAD-cmk andwith an inhibitor of the nuclear scaffold multicatalyticserine protease, APF-cmk (i.e., 1 mM), blocked laminB1 degradation [51]. In the same study, when thymo-cyte nuclei were incubated with Ca2+ and a panel ofprotease inhibitors, only the nuclear scaffold proteaseinhibitor APF-cmk or the protease inhibitor TPCKblocked lamin degradation, histone H1 cleavage, andDNA fragmentation [51]. In a second study, TPCKwas able to inhibit low molecular weight DNA frag-mentation in U-937 exposed to the topoisomerase Iinhibitor camptothecin [52]. In a further study, Lazeb-nik et al. [53] initially described a caspase-like enzymein a cell-free system that was a lamin proteinase thatwas inhibited by both Z-YVAD-cmk and the serineprotease inhibitor TLCK. This caspase was later iden-tified as caspase 6 [54].

To determine if the serine protease inhibitors TPCKand TLCK could increase clonogenic survival follow-ing TG treatment, AT3 or TSU cells were exposed tothe protease inhibitors for 3 hr prior to 48 (AT3) or 72(TSU) hr exposure to 500 nM TG (Figs. 3A, 4). In bothcell lines, the protease inhibitors alone had no effect onclonogenic survival vs. controls; nor were these inhibi-

tors able to change the clonogenic survival of TG-treated cells. TSU cells were also treated with thenuclear scaffold protease inhibitor APF-cmk and, at1-mM concentration, this inhibitor was also unable tosignificantly increase the clonogenic survival of TG-treated TSU cells (data not shown). In summary, thesedata demonstrate that, while the activity of bothcaspases and other apoptosis-associated proteasesmay have an important role in the effector arm ofapoptosis, none of these activities are critical in theinduction of apoptosis by TG in prostate cancer cells.

CONCLUSIONS

Previous studies demonstrated that TG is able toinduce the apoptotic death of all types of rodent andhuman prostate cancer cells, independent of their Rbor p53 status [14]. This is in contrast to a variety ofother treatments such as androgen ablation, radiation,phorbol ester, staurosporine, or lovastatin [24,26,27].In addition, TG-induced apoptosis does not requireprostate cancer cells entering or progressing throughthe proliferative cell cycle (i.e., it can effectively killproliferatively quiescent G0 cells) [15]. As demon-strated in the present study, the initiation (i.e., signal-ing) phase of apoptosis induced by TG is not pre-vented by inhibition of caspase activity. This latterpoint is significant, since it has been demonstratedthat when activated by various signaling pathways,various ubiquitously expressed protein kinases (e.g.,PKB or Akt) can phosphorylate caspases and therebyinhibit their enzymatic activity [30]. Besides phos-phorylation-based inhibition, caspases can also be in-hibited by the expression of a series of intracellularproteins (e.g., IAPs) [29]. Thus, agents that initiate ap-optosis via a caspase-dependent pathway are suscep-tible to development of resistance via caspase inhibi-tion due to either direct interaction with protein in-hibitors, or secondary to phosphorylation-inducedactivated protein kinases. Since, as demonstrated inthe present study, initiation of apoptosis in prostatecancer cells by TG occurs even when caspase ability isinhibited, this could explain why, when prostate can-cer cells were treated in culture with repeated cyclicexposure to TG, resistance to TG-induced apoptosisdid not occur either quickly or at a high frequency(data not shown).

In a previous study, it was demonstrated that ifprostate cancer cells were genetically engineered tooverexpress the survival gene, bcl-2, then a higherconcentration of TG was required to activate apoptosisin these cells [35]. Interestingly, of all of the unma-nipulated prostate cancer cell lines available, TSU-Pr1cell lines express the highest level of bcl-2 protein(data not shown) and yet these cells, as demonstrated

Fig. 6. Clonogenic survival assays performed on TSU cells fol-lowing treatment with TG for 72 hr ± caspase or protease inhibi-tors. Percent clonogenic survival is ratio of colonies observed intreated vs. untreated controls (n = 5 ± standard error, done induplicate).

Inhibition of Caspase Activity and Apoptosis 277

in the present study and others [14], are just as sensi-tive to TG-induced apoptosis as AT3.1 cells which donot express bcl-2 in culture [35].

Thus, while TG is a nearly ideal agent for inducingapoptosis of prostate cancer cells (i.e., high potencyand efficiency, cell proliferation independent, caspase,Rb, and p53-independent), it is not selective in its kill-ing ability. As an approach to target TG’s cytotoxicityspecifically to metastatic prostate cancer, we are de-veloping a prodrug form of TG that can be selectivelytargeted to sites of metastatic prostate cancer whileavoiding systemic toxicity. The prodrug will consist ofa primary amine containing a TG derivative coupledvia its amino group to form a peptide bond to a water-soluble peptide carrier [55]. Several primary amine-containing TG analogs have been identified that caninduce sustained elevations in intracellular calciumand activate programmed cell death in human pros-tate cancer cells [56]. The peptide carrier has been de-signed so that the peptide linkage between the amino-thapsigargin derivative and the peptide is efficientlyand specifically cleavable only by the proteolytic ac-tivity of prostate-specific antigen (PSA). PSA is a ser-ine protease produced as a unique differentiationproduct of normal and malignant prostate epithelialcells. Recently, a 6-amino acid PSA-specific peptidecarrier was identified [57] that can be used to targetcytotoxic agents to PSA-producing metastatic prostatecancer sites [58]. Currently, the TG analog coupled tothe PSA-specific peptide is being tested for activityagainst PSA-producing and -nonproducing prostatecancer cells.

ACKNOWLEDGMENTS

This work was supported by a CaPCure Award (toJ.T.I.).

REFERENCES

1. Thompson CB. Apoptosis in the pathogenesis and treatment ofdisease. Science 1995:267;1456–1462.

2. Isaacs JT. Role of programmed cell death in carcinogenesis. En-viron Health Perspect 1993:101;27–34.

3. Kerr JFR, Wyllie AG, Currie AR. Apoptosis: a basic biologicalphenomenon with wide ranging implications in tissue kinetics.Br J Cancer 1972:26;239–257.

4. Eastman A. Survival factors, intracellular signal transduction,and the activation of endonucleases in apoptosis. Semin CancerBiol 1995:6;45–52.

5. Zhivotovsky B, Gahm A, Ankarcrona M, Nicotera P, Orrenius S.Multiple proteases are involved in thymocyte apoptosis. ExpCell Res 1995:221;404–412.

6. Evan G, Littlewood T. A matter of life and cell death. Science1998;281:1317–1321.

7. Ashekenazi A, Dixit V. Death receptors: signaling and modula-tion. Science 1998;281:1305–1308.

8. Green DR, Reed JC. Mitochondria and apoptosis. Science 1998;281:1309–1312.

9. Thornberry NA, Lazebnik Y. Caspases: enemies within. Science1998;281:1312–1316.

10. Yagoda A, Petrylak D. Cytotoxic chemotherapy for advancedhormone-resistant prostate cancer. Cancer 1993:71;1098–1109.

11. Carter HB, Isaacs JT. Experimental and theoretical basis for hor-mone treatment of prostatic cancer. Semin Urol 1988;6:262–268.

12. Berges RR, Vukanovic J, Epstein JI, CarMichel M, Cisek L,Johnson DE, Veltri RW, Walsh PC, Isaacs JT. Implications of cellkinetic changes during the progression of human prostatic can-cer. Clin Cancer Res 1995;1:473–480.

13. Thastrup O, Cullen PJ, Drøbak BK, Hanley MR, Dawson AP.Thapsigargin, a tumor promoter, discharges intracellular Ca2+

stores by specific inhibition of the endoplasmic retirulum Ca2+-ATPase. Proc Natl Acad Sci USA 1990;87:2466–2470.

14. Furuya Y, Lundmo P, Short AS, Gill DL, Isaacs JT. The role ofcalcium, pH, and cell proliferation in the programmed (apop-totic) death of androgen-independent prostatic cancer cells in-duced by thapsigargin. Cancer Res 1994;54:6167–6175.

15. Lin XS, Denmeade SR, Cisek L, Isaacs JT. Role of growth arrestin programmed (apoptotic) death of prostate cancer cells bythapsigargin. Prostate 1997;33:201–207.

16. Furuya Y, Isaacs JT. Proliferation-dependent vs. independentprogrammed cell death of prostatic cancer involves distinctgene regulation. Prostate 1994;25:301–309.

17. Tombal B, Denmeade SR, Isaacs JT. Assessment and validationof a microinjection method for kinetic analysis of [Ca2+]i in in-dividual cells undergoing apoptosis. Cell Calcium 1999;25:19–28.

18. Denmeade SR, Tombal B, Isaacs JT. Apoptotic pathways in pros-tate cancer. In: Mattson M, editor. Programmed cell death: rolein disease, pathogenesis, and prevention. Greenwich: JAI Press,Inc., 1999.

19. Qi M-X, He H, Hongying Z, Distelhorst CW. Baculovirus p35and Z-VAD-fmk inhibit thapsigargin-induced apoptosis ofbreast cancer cells. Oncogene 1997;15:1207–1212.

20. Browne SJ, MacFarlane M, Cohen GM, Paraskeva C. The ade-nomatous polyposis coli protein and retinoblastoma protein arecleaved early in apoptosis and are potential substrates forcaspases. Cell Death Differ 1998;5:206–213.

21. Cohen GM. Caspases: the executioners of apoptosis. Biochem J1997;326:1–6.

22. Thornberry NA. The caspase family of cysteine proteases. BrMed Bull 1996;53:478–490.

23. Villa P, Kaufmann SH, Earnshaw WC. Caspases and caspaseinhibitors. TIBS 1997;22:00–00.

24. Marcelli M, Cunnigham GR, Haidacher SJ, Padayatty SJ, SturgisL, Kagan C, Denner L. Caspase-7 is activated during lovastatin-induced apoptosis of the prostate cancer cell line LNCaP. Can-cer Res 1998;58:76–83.

25. Marcelli M, Cunningham GR, Walkup M, He Z, Sturgis L, Ka-gan C, Mannuci R, Nicoletti I, Teng B, Denner L. Signalingpathway activated apoptosis of the prostate cancer cell line LN-CaP: overexpression of caspase-7 as a new gene therapy strat-egy for prostate cancer. Cancer Res 1999;59:382–390.

26. Bowen C, Spiegel S, Gelmann EP. Radiation-induced apoptosismediated by retinoblastoma protein. Cancer Res 1998;58:3275–3281.

27. Zhao X, Gschwend JE, Powell CT, Foster RG, Day KC, Day ML.Retinoblastoma-protein-dependent growth signal conflict andcaspase activity are required for protein kinase C-signaled ap-optosis of prostate epithelial cells. J Biol Chem 1997;272:22751–22757.

28. Srikanth SD, Kraft AS. Inhibition of caspases by cytokine re-

278 Denmeade et al.

sponse modifier A blocks androgen ablation-mediated prostatecancer cell death in vivo. Cancer Res 1998;58:834–839.

29. Deveraux QL, Roy N, Stennicke R, Van Arsdale T, Zhou Q,Srinivasula SM, Alnemri ES, Salvesen GS, Reed JC. IAPs blockapoptotic events induced by caspase-8 and cytochrome c bydirect inhibition of distinct caspases. EMBO J 1998;17:2215–2223.

30. Cardone MH, Roy N, Stennicke HR, Salvesen GS, Franke TF,Stanbridge E, Frisch S, Reed JC. Regulation of cell death prote-ase caspase-9 by phosphorylation. Science 1998;282:1318–1321.

31. Miura M, Friedlander RM, Yuan J. Tumor necrosis factor-induced apoptosis is mediated by crmA-sensitive cell deathpathway. Proc Natl Acad Sci USA 1995;92:8318–8322.

32. Martikainen P, Kyprianou N, Tucker RW, Isaacs JT. Pro-grammed death of non-proliferating androgen-independentprostatic cancer cells. Cancer Res 1991;51:4693–4700.

33. Iizumi T, Yazaki T, Kanoh S, Kondo I, Koiso K. Establishment ofa new prostatic carcinoma cell line (TSU-Pr1). J Urol 1987;137:1304–1306.

34. Furuya Y, Walsh JC, Lin X, Nelson WG, Isaacs JT. Androgenablation induced programmed death of prostatic glandular can-cer cells does not involve recruitment into a defective cell cycleor p53 induction. Endocrinology 1995;136:1898–1906.

35. Gerlach C, Golding M, Larue L, Alison MR, Gerdes J. Ki-67immunoexpression is a robust marker of proliferative cells inthe rat. Lab Invest 1997;77:697–698.

36. Furuya Y, Krajewski S, Epstein JI, Reed JC, Isaacs JT. Expressionof bcl-2 and the progression of human and rodent prostaticcancers. Clin Cancer Res 1996;2:389–398.

37. Knudsen ES, Wang JY. Dual mechanism for the inhibition of E2Fbinding to RB by cyclin-dependent kinase-mediated RB phos-phorylation. Mol Cell Biol 1997:17:5771–5783.

38. Chen W-D, Otterson GA, Lipkowitz S, Khleif SN, Coxon AB,Faye FJ. Apoptosis is associated with cleavage of a 5 kDa frag-ment from RB which mimics dephosphorylation and modulatesE2F binding. Oncogene 1997;14:1243–1248.

39. Janicke RU, Walker PA, Lin XY, Porter AG. Specific cleavage ofthe retinoblastoma protein by an ICE-like protease in apoptosis.EMBO J 1996;15:6969–6978.

40. An B, Dou QP. Cleavage of retinoblastoma protein during ap-optosis: interleukin-1B-converting enzyme-like protease as can-didate. Cancer Res 1996;56:438–442.

41. Fan G, Ma X, Kren BT, Steer CJ. The retinoblastoma gene prod-uct inhibits TGF-beta 1 induced apoptosis in primary hepato-cytes and human HuH-7 hepatoma cells. Oncogene 1996;12:1909–1919.

42. Thornberry NA, Rano TA, Peterson EP, Rasper DM, Timkey T,Garcia-Calvo M, Houtzager VM, Nordstrom PA, Roy S, Vail-lancourt JP, Chapman KT, Nicholson DW. A combinatorial ap-proach defines specificities of members of the caspase familyand granzyme B. J Biol Chem 1997;272:17907–17911.

43. Lazebnik YA, Kaufmann SH, Desnoyers S, Poirier GG, Earn-shaw WC. Reconstitution of the apoptotic cascade in vitro: piv-otal role for prICE, a protease resembling interleukin-1B con-verting enzyme and demonstration that poly(ADP ribose) poly-merase is a substrate for this enzyme during apoptosis. Nature1994;371:346–347.

44. Kaufmann SH, Desnoyers S, Ottaviano Y, Davidson NE, PoirierGG. Specific proteolytic fragmentation of poly (ADP-ribose)polymerase: an early marker of chemotherapy-induced apopto-sis. Cancer Res 1993;53:3976–3985.

45. Song Q, Lees-Miller SP, Kumar S, Zhang N, Chan DW, SmithGCM, Jackson SP, Alnemri ES, Litwack G, Khanna KK, LavinMF. DNA-dependent protein kinase catalytic subunit: a targetfor an ICE-like protease in apoptosis. EMBO J 1996;15:3238–3246.

46. Kothakota S, Azuma T, Reinhard C, Klippel A, Tang J, Chu K,McGarry TJ, Kirschner MW, Koths K, Kwiatkowski DJ, Willi-ams LT. Caspase-3-generated fragment of gelsolin: effector ofmorphological change in apoptosis. Science 1997;278:294–298.

47. Li P, Nijhawan D, Budihardjo I, Srinivasula SM, Ahmad M,Alnemri ES, Wang X. Cytochrome c and dATP-dependent for-mation of Apaf-1/caspase-9 complex initiates an apoptotic pro-tease cascade. Cell 1996;91:479–489.

48. Hakem R, Hakem A, Duncan GS, Henderson JT, Woo M, Soen-gas MS, Elia A, de la Pompa JL, Kagi D, Khoo W, Potter J,Yoshida R, Kaufman SA, Lowe SW, Penninger JM, Mak TW.Differential requirement for caspase 9 in apoptotic pathways invivo. Cell 1998;94:339–352.

49. Kuida K, Haydar TF, Kuan C-Y, Gu Y, Taya C, Karasuyama H,Su MS, Rakic P, Flavell RA. Reduced apoptosis and cytochromec-mediated caspase activation in mice lacking caspase 9. Cell1998;94:325–327.

50. Susin SA, Zamzami N, Castedo M, Daugas E, Wang HG, GeleyS, Fassy F, Reed JC, Kroemer G. The central executioner of ap-optosis: multiple connections between protease activation andmitochondria in Fas/APO-1/CD95-and ceramide-induced ap-optosis. J Exp Med 1997;186:25–37.

51. McConkey DJ. Calcium-dependent, interleukin 1B-convertingenzyme inhibitor-sensitive degradation of lamin B1 and DNAfragmentation in isolated thymocyte nuclei. J Biol Chem 1996;271:22398–22406.

52. Alain-Theophile S, Bertrand R. Distinct steps in DNA fragmen-tation pathway during camptothecin-induced apoptosis in-volved caspase-, benzyloxycarbonyl- and N-tosyl-L-phenylalanylchloromethyl ketone-sensitive activities. CancerRes 1998;58:3066–3072.

53. Lazebnik YA, Takahashi A, Moir RD, Goldman RD, Poirier GG,Kaufmann SH, Earnshaw WC. Studies of the lamin proteinasereveal multiple parallel biochemical pathways during apoptoticexecution. Proc Natl Acad Sci USA 1995;92:9042–9046.

54. Takahashi A, Musy P-Y, Martins LM, Poirer GG, Moyer RW,Earnshaw WC. CrmA/SPI-2 inhibition of an endogenous ICE-related protease responsible for lamin A cleavage and apoptoticnuclear fragmentation. J Biol Chem 1996;271:32487–32490.

55. Denmeade SR, Isaacs JT. Enzymatic-activation of prodrugs byprostate-specific antigen: targeted therapy for metastatic pros-tate cancer. Cancer J Sci Am [Suppl] 1998;4:15–21.

56. Christensen SB, Andersen A, Kromann H, Treiman M, TombalB, Denmeade SR, Isaacs JT. Thapsigargin analogues for target-ing programmed death of androgen-independent prostatic can-cer cells. Bioorg Med Chem 1998; in press.

57. Denmeade SR, Lou W, Lovgren J, Malm H, Lilja H, Isaacs JT.Specific and efficient peptide substrates for assaying the proteo-lytic activity of prostate-specific antigen. Cancer Res 1997;57:4924–4930.

58. Denmeade SR, Nagy A, Gao J, Lilja H, Schally A, Isaacs JT.Enzymatic-activation of a doxorubicin-peptide prodrug byprostate-specific antigen. Cancer Res 1998;58:2537–2540.

Inhibition of Caspase Activity and Apoptosis 279