Oncogene
SEARCH     advanced search my account e-alerts subscribe register
Journal home
Advance online publication
Current issue
Archive
Press releases
For authors
For referees
Contact editorial office
About the journal
For librarians
Subscribe
Advertising
naturereprints
Contact NPG
Customer services
Site features
NPG Subject areas
Access material from all our publications in your subject area:
Biotechnology Biotechnology
Cancer Cancer
Chemistry Chemistry
Dentistry Dentistry
Development Development
Drug Discovery Drug Discovery
Earth Sciences Earth Sciences
Evolution & Ecology Evolution & Ecology
Genetics Genetics
Immunology Immunology
Materials Materials Science
Medical Research Medical Research
Microbiology Microbiology
Molecular Cell Biology Molecular Cell Biology
Neuroscience Neuroscience
Pharmacology Pharmacology
Physics Physics
Browse all publications
 
28 March 2002, Volume 21, Number 14, Pages 2141-2153
Table of contents    Previous  Article  Next   [PDF]
Original Paper
Ganciclovir-induced apoptosis in HSV-1 thymidine kinase expressing cells: critical role of DNA breaks, Bcl-2 decline and caspase-9 activation
Maja T Tomicic1, Rudolf Thust2 and Bernd Kaina1

1Division of Applied Toxicology, Institute of Toxicology, Medical Faculty, University of Mainz, Obere Zahlbacher Str. 67,D-55131 Mainz, Germany

2Institute for Antiviral Chemotherapy, Medical Faculty, University of Jena, Winzerlaer Str. 10,D-07745 Jena, Germany

Correspondence to: B Kaina, Division of Applied Toxicology, Institute of Toxicology, Medical Faculty, University of Mainz, Obere Zahlbacher Str. 67,D-55131 Mainz, Germany. E-mail: kaina@mail.uni-mainz.de

Abstract

Although ganciclovir (GCV) is most often used in suicide anticancer gene therapy, the mechanism of GCV-induced cell killing and apoptosis is not fully understood. We analysed the mechanism of apoptosis triggered by GCV using a model system of CHO cells stably transfected with HSV-1 thymidine kinase (HSVtk). GCV-induced apoptosis is due to incorporation of the drug into DNA resulting in replication-dependent formation of DNA double-strand breaks and, at later stages, S and G2/M arrest. GCV-provoked DNA instability was likely to be responsible for the observed initial decline in Bcl-2 level and caspase-9/-3 activation. Further decline in the Bcl-2 level was due to cleavage of the protein by caspase-9, as demonstrated by use of caspase inhibitors and transfection with trans-dominant negative caspase expression vectors. Bcl-2 cleavage resulted in the appearance of a pro-apoptotic 23 kDa Bcl-2 fragment and in excessive cytochrome c release, dephosphorylation of BAD, cleavage of PARP and finally DNA degradation. Since Fas/CD95 and caspase-8 were only slightly activated we conclude GCV-induced apoptosis to occur in this cell system mainly by activating the mitochondrial damage pathway. This process is independent of p53 for which the cells are mutated. Caspase-9 mediated cleavage of Bcl-2 accelerates the apoptotic process and may explain the high potential of GCV to induce apoptosis. Data are also discussed as to implications for HSVtk gene therapy utilizing GCV.

Oncogene (2002) 21, 2141-2153 DOI: 10.1038/sj/onc/1205280

Keywords

apoptosis; Bcl-2 cleavage; caspase-9; ganciclovir; DNA breakage; suicide gene therapy

Introduction

The concept of HSVtk/GCV gene therapy is based on transduction of tumor cells with a 'suicide gene' encoding the thymidine kinase of herpes simplex virus type 1 (HSVtk) or by administration of HSVtk stably transfected cells followed by systemic treatment with ganciclovir (GCV). The HSVtk expressing cells become specifically sensitive to GCV and die by apoptosis (Moolten, 1986). Moreover, not only cells expressing the transgene but also neighboring non-transduced tumor cells can undergo apoptosis. This so-called 'bystander killing' is due to intercellular transfer of the activated pro-drug (i.e. GCV-triphosphate) via gap junctions and is restricted to dividing cells (Hamel et al., 1996; Mesnil et al., 1996). The most profound results thus far have been collected in clinical trials on the assessment of HSVtk/GCV gene therapy for the treatment of brain tumors (Oldfield et al., 1993; Klatzmann et al., 1996; Izquierdo et al., 1996). However, the therapy suffers from various limitations such as low target specificity and low transfection/transduction efficiency of target cells. Also, because of lack of knowledge of GCV action, the optimal protocol of GCV administration is not fully established. Despite this, GCV is still being used in anti-cancer gene therapy. To improve HSVtk/GCV therapy, a more detailed insight into the molecular mechanism of GCV-induced cell killing is required which is a prerequisite for optimized application of GCV.

Apart from utilization in suicide gene therapy, GCV is frequently used as a classical antiviral agent in the therapy of human cytomegalovirus (HCMV) diseases in immunocompromised individuals, such as cancer or AIDS patients and transplant recipients (Crumpacker, 1996). The wide application of the drug raises the question as to its genotoxic potential. GCV was found to be a potent recombinogenic and chromosomal breakage-inducing agent in metabolically competent cells expressing HSVtk (Thust et al., 2000a). Regarding the biochemical and molecular mechanism involved in GCV-induced genotoxicity, evidence has been provided that GCV is incorporated into the cellular DNA, thus causing sister-chromatid exchanges (SCEs) and chromosomal aberrations, reproductive cell death and apoptosis (Rubsam et al., 1998; Wei et al., 1998; Thust et al., 2000a,b). Whether the genotoxic effects and apoptosis induced by GCV are inter-related is an interesting issue that waits to be clarified.

In this work, we investigated the mechanism of GCV-induced apoptosis in order to explore cellular processes evoked by GCV both in herpes virus-infected cells and in antiviral or HSVtk suicide gene therapy. To this end we used Chinese hamster ovary (CHO) cells stably transfected with HSV-1 thymidine kinase (HSV-TK). We chose CHO cells as a model system for these investigations because they have most widely been used for cyto- and genotoxicity studies and are well-characterized in terms of mutagenic, cytotoxic and genotoxic responses to a broad range of genotoxic agents (for alkylating drugs see Kaina et al., 1997). We should also note that CHO cells are, contrary to most tumor cell lines, karyotypically quite stable and represent a homogenous cell population facilitating chromosomal and cell cycle studies. The cells are mutated for p53 (Lee et al., 1997). With this cell system the molecular mechanism of p53-independent apoptosis due to non-repaired alkylation and UV-C induced DNA damage has recently been analysed (Ochs and Kaina, 2000; Dunkern et al., 2001). Here we show that GCV-induced apoptosis is due to incorporation of GCV into the cellular DNA. GCV-induced apoptosis is a replication-dependent process which is associated with Bcl-2 decline and amplified by caspase-9-driven cleavage of Bcl-2. The positive amplification loop of Bcl-2 degradation provokes GCV-exposed HSVtk-transduced tumor cells to accelerate cell killing and may explain the high efficiency of GCV to induce apoptosis. To our knowledge, this is the first report demonstrating GCV to cause DNA breakage that triggers caspase-9-driven cleavage of Bcl-2 resulting in p53-independent apoptosis. Since resistance of tumor cells is often associated with abnormal production of Bcl-2 protein (Gallo et al., 1999; Feinmesser et al., 1999) and mutations in the gene encoding caspase-9 have been reported (Srinivasula et al., 1999; Seol and Billiar, 1999), the data bare implications for HSVtk/GCV gene therapy.

Results

Hypersensitivity of CHO-HSV-TK cells to the cytotoxic effect of GCV

The killing response of HSVtk-transfected CHO cells and control cells (transfected with the neo gene only) treated for one cell cycle (i.e. 14 h) with GCV is shown in Figure 1A. A concentration of 0.5 and 1 muM GCV reduced survival (colony formation) to 60 and less than 1% respectively. The CHO-neo control was completely insensitive to GCV in this concentration range. Cytotoxicity caused by GCV in HSVtk-transfected cells was mainly due to apoptosis, which was quantified and compared to the induction of necrosis (see Materials and methods) by flow cytometry. When exponentially growing cells were treated with GCV for the duration of one cell cycle (14 h), the frequency of apoptosis increased as a function of dose (Figure 1B). Time-course experiments revealed that the induction of apoptosis was a late response, starting between 24-36 h after treatment with GCV. The frequency of necrosis remained nearly unaffected (Figure 1C). Induction of apoptosis by GCV in CHO-HSV-TK cells was confirmed by gel electrophoresis showing the typical internucleosomal fragmentation pattern (Figure 1D).

GCV causes cell cycle delay in the second post-exposure cycle due to S- and G2/M-arrest

Since GCV-induced apoptosis in CHO-HSV-TK cells is a late response starting not earlier than one day of treatment with GCV, it indicates that post-exposure cell cycle events are involved. To elucidate this process in more detail we followed cell cycle progression upon GCV exposure. Growth curves (data not shown) and flow cytometry (Figure 2A) revealed that exponentially growing cells exposed to GCV (for 14 h) were not immediately blocked in division but started to accumulate after the first post-exposure cycle (>12 h post-exposure) in S and G2/M phase. Thus, 24 h after GCV treatment ~40 and 25% of cells were arrested in S and G2/M, respectively, whereas in the untreated sub-confluent control the percentage of cells in S and G2/M was 10 and 16%, respectively. At 24 h after GCV exposure, cells started to undergo apoptosis, as shown in the histograms by an increase in the sub-G1 peak. The time-course of appearance of the sub-G1 population was consistent with the induction of apoptosis determined by annexin V assay (see Figure 1C). After a post-incubation time of 36 and 48 h, cells accumulated predominantly in G2/M and stopped division. The cell cycle distribution of control cells (CHO-neo) was unaffected under the same conditions of treatment and no apoptotic changes were detected (data not shown). Overall, the data show that during the GCV exposure cycle and the first post-exposure cycle, CHO-HSV-TK cells are not significantly delayed in cell cycle progression. They were, however, severely blocked thereafter, in the second post-exposure cell cycle, where they started to die by apoptosis.

Induction of DNA double-strand breaks by GCV precedes induction of apoptosis

Previously we showed that GCV becomes incorporated into DNA. The incorporated GCV however was not genotoxic by its own, but appeared to generate critical DNA lesions in the post-exposure cell cycle triggering genotoxicity (such as SCEs and aberrations) and apoptosis (Thust et al., 2000a). These secondary DNA lesions were hypothesized to be DNA double-strand breaks (DSBs) that were previously shown to trigger genotoxicity (Obe et al., 1985) and apoptosis (Lips and Kaina, 2001). To prove whether GCV treatment results in DNA breakage, we determined the frequency of DSBs by single cell gel electrophoresis (SCGE). As shown in Figure 2B, DSBs were induced as early as 18 h after GCV exposure, reaching a maximum at 36 h after treatment. Since significant increase of apoptosis was observed at times >36 h after GCV treatment, the induction of DSBs clearly preceded apoptosis following GCV exposure.

Caspase activation and cleavage of PARP

To analyse the involvement of caspases in GCV-induced apoptosis, we determined the time-course of caspase activation upon GCV exposure. As shown in Figure 3A, a 1.6-fold increase of caspase-9 activity was detected already 12 h after GCV treatment. This activity strongly increased 24 h after treatment and remained at a high level up to 60 h thereafter. Induction of caspase-3 was first noticed 24 h after GCV exposure and thus occurred significantly later than the induction of caspase-9. High level induction of caspase-3 was observed 48 h later, remaining constant up to 60 h. Caspase-8 was activated early (12 h post-incubation) and remained constantly enhanced (1.9-fold) throughout the whole post-exposure incubation period.

To assess the role of individual caspases in triggering apoptosis, GCV-exposed cells were treated with specific caspase inhibitors. The general caspase inhibitor (zVAD) efficiently blocked apoptosis by 75%. The same strong blockage effect was observed with the caspase-9 inhibitor (zLEHD), while caspase-3 (zDEVD) and caspase-8 (zIETD) inhibitors were rather ineffective in blocking apoptosis (Figure 3B). The effect of the inhibitors on GCV-induced caspase activity was determined in control experiments, confirming the inhibitors to be active and specific at the concentration used (100 muM), i.e. they did not have undesired cross-inhibitory effects on other caspases (Figure 3C). Thus, zDEVD and zIETD did not exert a cross-reactive effect on any examined caspases, whereas zLEHD reduced caspase-3 activity by ~20% indicating that, in CHO-HSV-TK cells after GCV treatment, activation of caspase-3 was under the control of caspase-9. Inhibition of caspase-9 activity did not lead to inhibition of caspase-8, implicating caspase-8 to be an initiator caspase in our cell system. All assayed caspase activities were significantly down-modulated by zVAD. Overall, the data indicate that caspase-9 plays a decisive regulatory role in GCV-induced apoptosis in CHO cells expressing HSV-TK.

A specific endogenous substrate of caspases is PARP. Cleavage of PARP was observed as early as 14 h after GCV exposure and continued at later times (Figure 4A). Since caspase-9 activity was detectable already 12 h after GCV treatment and clearly preceded the rise in caspase-3 activity, it was obvious that at that time point cleavage of PARP was mediated by caspase-9. This was verified by utilization of caspase inhibitors, showing that zVAD and zLEHD prevented formation of the 85 kDa PARP cleavage fragment whereas zDEVD was barely effective (Figure 4B). This supports the hypothesis that caspase-9 rather than caspase-3 is an executive caspase during GCV-triggered apoptosis.

Decline in Bcl-2 level is not due to changes in transcription or de novo protein synthesis

To elucidate whether the mitochondrial damage pathway is involved in GCV-induced apoptosis, the expression of proteins of the Bcl-2 family was assessed. As shown in Figure 5A, the level of Bcl-2 protein decreased 24-72 h after GCV treatment, which was not observed in control (CHO-neo) cells. The timing of Bcl-2 decline was related to the induction of apoptosis (Figure 5B). The amount of bcl-2 mRNA did not change, as shown by RT-PCR analysis of RNA isolated from control and GCV-treated cells (Figure 5C). Also, treatment of GCV-exposed cells with inhibitors of protein synthesis (cycloheximide+anisomycin) did not affect the expression level of Bcl-2 (data not shown). This supports the view that Bcl-2 decline is a post-translational phenomenon (see below and Discussion). Bcl-xL and Bax were expressed during the whole post-exposure period without significant changes either in CHO-HSV-TK or in control cells. BAD was expressed as a double-band with some increase in expression of the unphosphorylated pro-apoptotic form and simultaneous decrease in the expression of the phosphorylated anti-apoptotic form, 48-72 h after GCV exposure of CHO-HSV-TK cells, which was not observed in the control (Figure 5D).

CD95 receptor-mediated apoptosis is a minor pathway during GCV-induced apoptosis in CHO-HSV-TK cells

Since caspase-8, which is downstream of Fas, was found to be (slightly) activated after GCV treatment (see Figure 3A) and its inhibition (by zIETD) resulted in ~25% reduction of apoptosis, the expression of Fas (CD95/APO-1) receptor (FasR) and Fas ligand (FasL) was examined. There was no significant change in the expression of FasR, whereas a twofold induction of FasL was observed during the time period of 12-72 h after GCV treatment (Figure 5E). This may be taken to indicate that the Fas system only plays a minor role in GCV-induced apoptosis. It has been reported that the activation of this pathway is regulated by p53 (Muller et al., 1998). We should note that CHO cells are mutated for p53 (Lee et al., 1997) and the cells we were working with do not show p53-driven promoter activation (Dunkern and Kaina, unpublished data).

Initial release of cytochrome c is independent of caspase activation

To further prove that mitochondrial damage is associated with GCV-induced apoptosis, we determined the release of cytochrome c (cyt c) from mitochondria into the cytosol (Figure 6A,B). cyt c release occurred early (i.e. 12 h post-exposure). It paralleled caspase-9 activation (see Figure 3A) and was detectable before the decline in the Bcl-2 level. 72 h after exposure to GCV, the cytosolic fraction was highly enriched with cyt c (Figure 6B). Interestingly, neither zVAD nor zLEHD blocked the release of cyt c (Figure 6C) which indicates that the initial cyt c release was independent of caspase activation.

Cleavage of Bcl-2 by caspase-9

To analyse whether Bcl-2 is degraded by caspases upon GCV exposure, we performed Western blot experiments using two different cross-reactive antibodies: anti-Bcl-2 mAb and pAb (both epitopes corresponding to amino acids 1-205 of human Bcl-2). The mAb detected the intact Bcl-2 protein of 26 kDa (Bcl-2-p26), whose quantity significantly decreased after GCV exposure (see Figure 5A). The pAb gave only a weak signal for the intact Bcl-2 protein, but detected a Bcl-2 fragment of 23 kDa (Bcl-2-p23). This fragment appeared first 24 h after GCV treatment and accumulated up to 50-fold in the further post-incubation period. Bcl-2-p23 was not observed in CHO-neo control cells (Figure 7A). Since caspases were activated within this time period, it was reasonable to assume that the Bcl-2 p23 fragment resulted from proteolytic cleavage of Bcl-2 by caspases. To prove this, the expression level of both Bcl-2 p26 and p23 was determined in GCV-treated cells in the presence and absence of caspase inhibitors (Figure 7B). Interestingly, the generation of Bcl-2-p23 was almost completely abolished by either zVAD or zLEHD. The caspase-3 inhibitor (zDEVD) was clearly less effective in preventing the cleavage. This indicates that in GCV-treated CHO-HSV-TK cells the proteolytic cleavage of Bcl-2 is mainly accomplished by caspase-9. The appearance of Bcl-2-p23 24 h after GCV exposure and its subsequent accumulation correlated with the kinetics of the induction of apoptosis (see Figure 1C) indicating cleavage of Bcl-2 by caspase-9 to contribute significantly to GCV-induced apoptosis.

Trans-dominant inhibition of caspase-9 but not of caspase-3 abrogates cleavage of Bcl-2 and overall apoptosis following exposure to GCV

To further elucidate the importance of caspase-9 in GCV-induced apoptosis, we modulated caspase-9 activity by transient transfection of either wild-type or a dominant-negative caspase-9 expression construct in CHO-HSV-TK cells. We also transfected the cells with a dominant-negative caspase-3 construct and compared the effect of expression with the induction of apoptosis by GCV. Wild-type caspase-9 (here designated as casp9-wt) was in this experiment only moderately expressed (Figure 8A) and did not significantly increase GCV-induced apoptosis (Figure 8B). On the other hand, high expression of dominant-negative caspase-9 (here designated as casp9-DN; for Western blot analysis, see Figure 8A) led to a significant decrease in the induction of apoptosis (Figure 8B). A strong expression of dominant-negative caspase-3 (designated as casp3-DN; see Figure 8A, low panel) resulted in only marginal reduction of apoptosis (Figure 8B). This was in accordance with the weak reduction of apoptosis observed after treatment with zDEVD. The modulation of caspase activities upon transient transfection of the transdominant negative vectors (noted above) is shown in Figure 8C, demonstrating a significant reduction of endogenous caspase-9 and -3 activities upon transfection and GCV treatment.

To determine whether the expression of casp9-DN and casp3-DN exerts an effect on Bcl-2 cleavage, we analysed Bcl-2 expression in transiently transfected CHO-HSV-TK cells upon GCV treatment. The amount of Bcl-2-p23 was significantly reduced in cells in which caspase-9 activity was down-modulated by transfection of casp9-DN but not in cells transfected with casp3-DN (Figure 8D). These results support the conclusion that caspase-9 mediates GCV-induced apoptosis via cleavage of Bcl-2; it is therefore an important execution caspase in this process. To further assess the involvement of caspase-9 in Bcl-2 cleavage, we imitated the apoptotic process evoked by GCV in CHO-HSV-TK cells. For this purpose we overexpressed the active form of caspase-9 (casp9-wt) and determined the cleavage of Bcl-2 in the presence and absence of the caspase-3-like inhibitor zDEVD. As shown in Figure 8E, the transiently expressed active wild-type caspase-9 cleaved Bcl-2, both in the absence and presence of zDEVD. Under the conditions of total caspase-3 inhibition, the activity of caspase-7 was also blocked (data not shown) which makes it unlikely that it participated in cleavage of Bcl-2. Thus this data confirm that Bcl-2 is cleaved by caspase-9 in vivo without the requirement of caspase-3.

Discussion

GCV has been used for a decade in the therapy of cytomegalovirus infections, which are a life-threatening complication in immunocompromised patients. Upon metabolization to the corresponding triphosphate, GCV is incorporated into the virus DNA thus blocking its replication. GCV becomes also incorporated into the genomic DNA of mammalian cells during replication (Rubsam et al., 1998; Thust et al., 2000a; Tomicic and Kaina, unpublished data). Unlike acyclovir (ACV), which is an absolute chain terminator, GCV only poorly inhibits replication of cellular DNA in the exposure cell cycle (Thust et al., 2000a). GCV incorporation into DNA does not lead to genotoxic alterations per se. Chromosomal changes are rather a consequence of secondary DNA alterations arising from replication of a template containing GCV in the post-exposure cell cycles (Thust et al., 2000a).

The mechanism of GCV-induced cell death downstream of the induced DNA damage is far less understood. GCV-induced cytotoxicity is largely due to apoptosis. Apoptotic cell death increased with the dose of GCV, whereas the frequency of necrotic cells (as measured by annexin V/propidium iodide double-staining) remained fairly constant below 10%. When exponentially growing cells were exposed to GCV for a period of one cell cycle, apoptotic cells appeared first at ~24 h after treatment with the drug and significantly increased further at later time points (36-72 h). Cell cycle analysis revealed that cells passed both the incorporation and the first post-exposure cell cycle without much delay, whereas they were severely inhibited in the next cell cycle (i.e. the second post-exposure cycle). At this stage the cells became arrested in S and G2/M phase and were triggered to die. The induction of S and G2/M arrest seems to be a general feature of GCV-treated cells. It was also observed in other systems, such as human glioblastoma and murine melanoma cells upon GCV treatment (Rubsam et al., 1998; Wei et al., 1998). It has been proposed that the irreversible G2/M-arrest is associated with cytoskeletal reorganization leading to cell death (Halloran and Fenton, 1998). We favor the hypothesis that the cell cycle arrest is a late effect of GCV-induced DNA damage. GCV incorporated into DNA leads to replication-dependent induction of DNA double-strand breaks (DSBs) that were observed in the post-exposure replication cycle. It is pertinent to conclude that these DSBs finally trigger apoptosis because: (i) DSBs generated within cellular DNA by electroporation of restriction enzyme efficiently trigger apoptosis but not necrosis (Lips and Kaina, 2001); (ii) DSBs are likely the most critical lesion inducing apoptosis upon ionizing radiation (Radford and Murphy, 1994; Lips and Kaina, 2001); and (iii) DSBs are formed upon treatment of cells with chemical mutagens and UV-C light and are supposed to trigger apoptosis (Ochs and Kaina, 2000; Dunkern and Kaina, unpublished data). Since DSBs formation requires replication of DNA containing GCV, the data provide at the same time an explanation for the finding that GCV-induced apoptosis is replication-dependent (Hamel et al., 1996).

It is important to note that CHO-9 cells possess mutated p53 (Orren et al., 1995; Lee et al., 1997; Hu et al., 1999) and do not show transactivation of a p53-driven promoter (Dunkern and Kaina, unpublished data). Also, p21 was constitutively expressed in the cells without any further rise after GCV treatment (data not shown). It thus appears that apoptosis induced by GCV in CHO-HSV-TK cells does not require p53/p21. The involvement of a p53/p21-independent pathway in GCV-induced apoptosis is of special interest in view of the application of HSVtk/GCV suicide gene therapy for tumors that do not express functional p53. It has been shown that p53 can activate the gene encoding CD95 (Fas/Apo-1) receptor in response to DNA damage by anticancer drugs (Muller et al., 1998). In our cell system, the expression level of the CD95 receptor was unaltered, and ~twofold induction of the CD95 ligand was observed which was accompanied by ~twofold induction of caspase-8 activity. We should also note that we were unable to stimulate induction of apoptosis in CHO cells using an agonistic anti-Fas mAb, or to significantly inhibit the activation of caspase-8 by transfection of dominant-negative FADD (Dunkern et al., 2001). Therefore, Fas receptor-triggered pathways seem to play only a marginal role in this cell system. Activation of caspase-8 was shown to be dependent on CD95 ligand-mediated oligomerization of CD95 receptor (Srinivasula et al., 1996). CD95 receptor/ligand-independent caspase-8 activation and apoptosis have also been described (Wesselborg et al., 1999; Ferreira et al., 2000). It has been reported that GCV induces apoptosis via caspase-8 activation in p53 proficient HSVtk-expressing neuroblastoma cells, due to p53-mediated translocation of the CD95 receptor from the cytosol to the membrane without affecting the CD95 ligand (Beltinger et al., 1999). Involvement of FasR/FasL pathway in GCV-induced apoptosis was also demonstrated in different tumor cell lines, all of which possessed functional p53 (Wei et al., 1999; Beltinger et al., 2000). Since in our cell system expression of FasR was unaltered and caspase-8 was initially activated but not further induced, we assume that Fas/caspase-8-driven apoptosis plays a minor role in GCV treated cells not expressing wild-type p53. Overall it appears that, depending on the cell type, GCV can induce apoptosis both via a p53-dependent and p53-independent pathway.

A hallmark of GCV-induced p53-independent apoptosis in HSV-TK expressing cells was decline in Bcl-2 that correlated with the appearance of apoptotic cells. Decrease in the Bcl-2 protein level was not due to a change in bcl-2 transcription or mRNA stability. Also, inhibitors of protein synthesis did not affect the decline of Bcl-2 (data not shown), indicating that the regulation of Bcl-2 activity after GCV exposure is dependent on post-translational events. We also observed that Bax and Bcl-xL expression was not altered upon GCV exposure. This is, for instance, in contrast to increased levels of Bax in rat glioma cells upon GCV treatment (Craperi et al., 1999). In metabolically competent CHO-HSV-TK cells, BAD was expressed in the pro-apoptotic hypophosphorylated form. In the CHO-neo control, which was not affected by GCV, BAD was hyper-phosphorylated, which is known to be mediated by the protein kinase Akt (Datta et al., 1997).

To elucidate in more detail the mechanism of Bcl-2 decline and the involvement of caspases in GCV-triggered apoptosis, we investigated the effect of caspase inhibitors. The caspase-3 inhibitor zDEVD only partially prevented Bcl-2 and PARP cleavage and insignificantly reduced apoptosis. This indicates that in CHO-HSV-TK cells caspase-3 does not play a major role in mediating cell killing. On the other hand, the caspase-9 inhibitor zLEHD and the broad caspase inhibitor zVAD were efficient in suppressing GCV-induced PARP/Bcl-2 cleavage and apoptosis. This suggests that caspase-9 activation is an early event in GCV-induced cell killing, PARP cleavage and apoptosis. It also demonstrates that Bcl-2 is mainly cleaved by caspase-9 and, to a lesser extent, by caspase-3, thus causing the decline in the cellular Bcl-2 level upon GCV exposure. This conclusion was verified by transfection of dominant-negative (DN) caspase-9 and caspase-3 into CHO-HSV-TK cells. Expression of DN-caspase-9 strongly prevented Bcl-2 cleavage and apoptosis after GCV treatment, whereas DN-caspase-3 exerted only a slight effect. This supports the critical role of caspase-9 in GCV-induced Bcl-2 cleavage. If caspase-9 is predominantly involved in Bcl-2 cleavage, overexpression of the constitutively active form of caspase-9 should be able to imitate some of the effects brought about by GCV. This was indeed the case, since the transiently transfected active form of human wild-type caspase-9 cleaved Bcl-2 in HSV-TK expressing cells without GCV treatment, which occurred even in the presence of the inhibitor of caspase-3.

A direct interaction between caspase-3 and Bcl-2 has been shown in several cell systems (Cheng et al., 1997; Grangirard et al., 1998; Fujita and Tsuruo, 1998). However, the involvement of caspase-9 in Bcl-2 cleavage has not been demonstrated before. Caspase-9-mediated cleavage of Bcl-2 resulted in a cleavage product of 23 kDa (Bcl-2-p23), which accumulated in HSV-TK expressing cells upon treatment with GCV. The Bcl-2 cleavage fragment was shown to localize in the mitochondrial membrane and to promote the release of cyt c from mitochondria into the cytosol, thus being actively involved in apoptosis (Kirsch et al., 1999).

Cleavage of Bcl-2 by caspases occurs at a defined cleavage site that was first characterized for caspase-3 in the human and rat protein (Cheng et al., 1997). To determine the putative caspase cleavage site of the hamster Bcl-2 protein, we cloned the hamster bcl-2 cDNA and showed that potential cleavage sites for caspases are highly conserved between hamster and other mammalian species (Tomicic et al., 2000). We also observed that hamster Bcl-2 protein is cleaved in vitro by caspase-9 (Tomicic and Kaina, 2001) resulting in the same 23 kDa fragment that has been detected in vivo. This supports our conclusion that caspase-9 is majorly involved in cleavage of Bcl-2 upon GCV treatment.

One of the earliest apoptotic features observed in CHO-HSV-TK cells after GCV treatment was release of cyt c. It occurred before significant decline of Bcl-2 became detectable. Since the initial cyt c release was not blocked by caspase inhibitors, which was also reported by others for chemical agent-induced apoptosis (Bossy-Wetzel et al., 1998; Sun et al., 1999), cleavage of Bcl-2 by caspase-9 is likely not to be responsible for the very early cyt c release but rather for excessive leakage of mitochondria at later times after GCV exposure. Based on these and previous results, the overall scenario of GCV-induced p53-independent cell death is hypothesized as follows (see Figure 9). Incorporation of GCV into DNA leads to the formation of secondary DNA lesions within the subsequent (post-exposure) DNA replication cycle due to errors in replication or repair of GCV-induced DNA damage. The critical ultimate lesions are supposed to be DSBs which trigger, by a yet unknown mechanism, the early release of cyt C from mitochondria, probably through a slight initial decrease in the mitochondrial membrane level of Bcl-2 changing the Bax:Bcl-2 ratio. Binding of cyt C to Apaf-1 activates caspase-9 which in turn activates downstream executive caspases (Li et al., 1997; Slee et al., 1999). Moreover, active caspase-9 cleaves Bcl-2. The product of this cleavage, a pro-apoptotic Bcl-2 fragment (Bcl-2-p23), promotes further cyt c release and caspase-9 activation. This positive amplification loop appears to be a specific feature of GCV-induced apoptosis in HSV-TK expressing cells and may explain the exceptional high efficiency of GCV to trigger apoptosis independently of p53. It remains to be seen whether this process is a general characteristic of p53 mutated tumor cells which respond to GCV by activating the mitochondrial damage pathway.

The finding that GCV must be incorporated into DNA and cells need to replicate further in order to become apoptotic suggests that slowly replicating tumor cells are likely not accessible to GCV if the drug is administered only for a short period of time. Since DNA breaks appear to be an ultimate trigger of GCV-induced apoptosis it would also be interesting to see whether a combination of GCV and a DNA breaking agent such as bleomycin or ionizing radiation (using a dose not blocking replication) enhances the killing effect. Finally, the finding that caspase-9 is essential for GCV-induced apoptosis in p53-deficient cells may be useful to improve suicide HSVtk/GCV gene therapy. Thus, transfer of caspase-9 along with the HSVtk gene may enhance the therapeutic effect of GCV by increasing apoptosis in tumor cells transduced with HSVtk or targeted by metabolic cooperation. This protocol might be of particular benefit for tumor cells that exhibit intrinsic resistance due to overexpression of anti-apoptotic proteins (e.g. Bcl-2, Bcl-xL) or due to a defect in caspase-9 expression or its regulation. An enhanced killing response of tumor cells might allow to reduce the dose of GCV to be administered which may diminish systemic toxicity of the agent, allowing patients a better recovery after the therapy.

Materials and methods

Reagents and antibodies

Ganciclovir (9-[1,3-dihydroxy-2-propoxymethyl]-guanine, CymeveneTM) was a product of Syntex Arzneimittel GmbH (Aachen, Germany). zVAD, zDEVD, zLEHD and zIETD were irreversible cell permeable fluoromethyl ketone (fmk) modified caspase inhibitors and purchased from Calbiochem, R&D Systems and Enzyme Systems Products. Mouse anti-Bcl-2 monoclonal antibody (mAb) as well as rabbit anti-Bcl-2, anti-Bax, anti-Fas, anti-PARP, anti-cytochrome c and anti-ERK2 polyclonal antibody (pAb) were from Santa Cruz, and mouse anti-BAD mAb, as well as rabbit anti-Bcl-xL pAb were from Transduction Laboratories. Mouse anti-FasL mAb used was either from Santa Cruz or Transduction Laboratories. Mouse anti-Flag M2 and anti-AU1 mAb were products of Stratagene, and BabCo, respectively. Horseradish peroxidase-coupled secondary anti-mouse and anti-rabbit IgG antibodies were from Amersham.

Cell lines and transfection experiments

CHO-9 cells were routinely cultivated in DMEM/Ham's F12 medium supplemented with 5% inactivated (30 min, 56°C) fetal bovine serum (Gibco Life Technologies). The cells were co-transfected with pMCI-tk and pSV2neo expression vectors as described (Thust et al., 2000a). For transient transfection, 2´105 CHO-HSV-TK cells were transfected with 2 mug of the test vector using lipofectine according to the manufacturer's instructions (Gibco Life Technologies). Otherwise, the cells were transiently transfected using effectine protocol (Qiagen). The test vectors, ICE-LAP6-FLAG (casp9-wt), ICE-LAP6mt-FLAG (casp9-DN) and AU1-Yama-mt (casp3-DN) were already described (Duan et al., 1996). After 16-20 h transfection, the fresh medium containing serum and GCV (1 muM) was added. Sixty hours later, the cells were subjected to annexin V staining to determine apoptosis.

Semi-quantitative RT-PCR

Total RNA was used as template for both the cDNA synthesis and PCR according to TitanTM One Tube RT-PCR System (Roche). Oligonucleotide primers for amplification of 580 bp bcl-2 fragment were 5'-GGAAGGATGGCGCAAGCCGGGAG-3' (forward) and 5'-CCCAGCCTCCGTTATCCTGGATC-3' (reverse). The RT-PCR program used was as described (Tomicic et al., 2000).

Clonogenic survival and determination of apoptosis

Reproductive cell death was assayed by measuring colony formation with and without GCV treatment as described (Thust et al., 2000a). Drug-induced apoptosis was quantitatively and qualitatively determined. For quantification of apoptosis and necrosis, cells were stained with annexin V-FITC (Pharmingen) and propidium iodide (PI), and subjected to flow cytometry, as described (Vermes et al., 1995). For qualitative determination, internucleosomal fragmentation was determined as described (Ioannou and Chen, 1996).

Cell cycle analysis

Cell cycle analysis was performed as described (Nicoletti et al., 1991). Briefly, after 14 h-exposure to GCV, cells were trypsinized and washed twice with cold PBS. Cell pellets were re-suspended in PBS and fixed by the addition of 70% ethanol overnight at -20°C and processed for flow cytometry. Cells within different cell cycle phases and the fraction of cells exhibiting a DNA content lower than G1 were quantified using CellQuest (Becton Dickinson).

Neutral single cell gel electrophoresis

The originally described procedure for the neutral single cell gel electrophoresis (SCGE) for detection of DNA double-strand breaks (DSBs) (Klaude et al., 1996) was slightly modified. At various times after GCV treatment, sub-confluent cells were trypsinized, washed with cold PBS, and kept on ice until assayed. Cells were embedded in 0.1% low melting point-agarose, and microscope slides were immersed in ice-cold lysis solution. Slides were incubated in the lysis buffer (2.5 M NaCl, 100 mM EDTA, 10 mM Tris, 1% Na-laurylsarcosine (pH 7), 1% Triton X-100, 10% DMSO were added freshly) for 1 h at 4°C. After lysis, electrophoresis (25 V) was carried out for 15 min at 4°C (90 mM Tris, 90 mM boric acid, 2 mM EDTA). The fixed and ethidium bromide-stained slides were analysed using fluorescence microscope. Analysis of DNA migration (related to the induction of DNA DSBs) was performed by image analysis system (Kinetic Imaging Ltd.; Komet 4.0.2; Optilas) using defined standard and expressed as Olive tail moment (Olive et al., 1991).

Determination of caspase activity

Caspase Colorimetric Assay (R&D Systems) was used according to the manufacturer's protocol. Briefly, at various post-exposure times, cells were trypsinized, counted, and collected by centrifugation. The cell pellet was lyzed on ice and centrifuged. The supernatant was transferred and kept on ice. The enzymatic reaction was carried out in 96-well microplate (405 nm, 37°C, 1-2 h) with addition of reaction buffer and appropriate caspase substrate supplied with the kit.

Preparation of cell extracts and immunoblotting

Whole-cell extracts were prepared by lysis in ice-cold sample buffer (25 mM Tris-HCl (pH 6.8), 5% glycerol, 2.5% 2-mercaptoethanol; fresh PMSF), followed by sonification (Branson sonifier, 30 KHz, 3´10 s) on ice. The mitochondrial and cytosolic extracts were isolated by differential centrifugation as described (Wang and Studzinski, 1997). Protein extracts were quantified (Bradford, 1976) and fractions of 10-40 mug protein were separated by 10-12% SDS-PAGE, electroblotted onto nitrocellulose membrane (Schleicher and Schuel, Germany), that were incubated with antibodies as described (Tomicic et al., 1997). The used antibodies were mostly diluted (1 : 500) in 5% non-fat dry milk, 0.2% T-PBS. Protein-antibody complexes were visualized by enhanced chemiluminiscence (ECL Kit, Amersham).

Acknowledgements

We are grateful to Dr Vincenz (Univesity of Michigan, USA) for the generous gift of the ICE-LAP6-Flag, ICE-LAP6-mt-Flag and AU1-Yama-mt constructs. We thank Dr Markus Christmann for helpful discussion and critical reading of the manuscript. This work was supported by the Deutsche Forschungsgemeinschaft (grants KA 724/7-1 and 7-3 to B Kaina and TH 670/1-2 to R Thust) and the Stiftung Rheinland-Pfalz.

References

Beltinger C, Fulda S, Kammertoens T, Meyer E, Uckert W, Debatin K-M. (1999). Proc. Natl. Acad. Sci. USA, 96: 8699-8704. MEDLINE

Beltinger C, Fulda S, Kammertoens T, Uckert W, Debatin K-M. (2000). Cancer Res., 60: 3212-3217. MEDLINE

Bossy-Wetzel E, Newmeyer DD, Green DR. (1998). EMBO J., 17: 37-49. Article MEDLINE

Bradford MM. (1976). Anal. Biochem., 72: 248-254. Article MEDLINE

Cheng E H-Y, Kirsch DG, Clem RJ, Ravi R, Kastan MB, Bedi A, Ueno K, Hardwick JM. (1997). Science, 278: 1966-1968. Article MEDLINE

Craperi D, Virat JM, Nisson MF, Mathieu J, Baudier J, Benabid AL, Verna JM. (1999). Hum. Gene Ther., 10: 679-688. Article MEDLINE

Crumpacker CS. (1996). N. Engl. J. Med., 335: 721-729. MEDLINE

Datta SR, Dudek H, Tao X, Masters S, Fu H, Gotoh Y, Greenberg ME. (1997). Cell, 91: 231-241. MEDLINE

Duan H, Orth K, Chinnaiyan AM, Poirier GG, Froelich CJ, HE W-W, Dixit VM. (1996). J. Biol. Chem., 271: 16720-16724. MEDLINE

Dunkern TR, Fritz G, Kaina B. (2001). Oncogene, 20: 6026-6038. MEDLINE

Feinmesser M, Halpern M, Fenig E, Tsabari C, Hodak E, Sulkes J, Brenner B, Okon E. (1999). Human Path., 30: 1367-1372.

Ferreira CG, Tolis C, Span SW, Peters GJ, van Lopik T, Kummer AJ, Pinedo HM, Giaconne G. (2000). Clin. Cancer Res., 6: 203-212. MEDLINE

Fujita N, Tsuruo T. (1998). Biochem. Biophys. Res. Commun., 246: 484-488. Article MEDLINE

Gallo O, Chiarelli I, Boddi V, Bocciolini C, Bruschini L, Porfiri B. (1999). Int. J. Cancer, 84: 573-579. Article MEDLINE

Grangirard D, Studer E, Monney L, Belser T, Fellay I, Borner C, Michel MR. (1998). EMBO J., 17: 1268-1278. Article MEDLINE

Halloran PJ, Fenton RG. (1998). Cancer Res., 58: 3855-3865. MEDLINE

Hamel W, Magnelli L, Chiarugi VP, Israel MA. (1996). Cancer Res., 56: 2697-2702. MEDLINE

Hu T, Miller CM, Ridder GM, Aardema MJ. (1999). Mutat. Res., 426: 51-62. MEDLINE

Ioannou YA, Chen FW. (1996). Nucl. Acids Res., 24: 992-993.

Izquierdo M, Martin V, de Felipe P, Izquierdo JM, Peres-Higueras A, Cortes ML, Paz JF, Isla A, Blazques MG. (1996). Gene Ther., 3: 491-495. MEDLINE

Kaina B, Ziouta A, Ochs K, Coquerelle T. (1997). Mutat. Res., 381: 227-241. MEDLINE

Kirsch DG, Deseff A, Chau BN, Lim D-S, de Souza-pinto NC, Hansford R, Kastan MB, Lazebnik YA, Hardwick JM. (1999). J. Biol. Chem., 274: 21155-21161. Article MEDLINE

Klatzmann D, Phillipon J, Valery A, Bensimon G, Salzmann J-L. (1996). Hum. Gene Ther., 7: 109-126.

Klaude M, Eriksson S, Nygren J, Ahnstrom G. (1996). Mutat. Res., 363: 89-96. MEDLINE

Lee H, Larner JM, Mamlin JL. (1997). Gene, 184: 177-183. Article MEDLINE

Li P, Nijhawan D, Budihardjo I, Srinivasula SM, Ahmad M, Alnemri ES, Wang X. (1997). Cell, 91: 479-489. MEDLINE

Lips J, Kaina B. (2001). Carcinogenesis, 22: 579-585. Article MEDLINE

Mesnil M, Piccoli C, Tiraby G, Willecke K, Yamasaki H. (1996). Proc. Natl. Acad. Sci. USA, 93: 1831-1835. MEDLINE

Moolten FL. (1986). Cancer Res., 46: 5276-5281. MEDLINE

Muller M, Wilder S, Bannasch D, Israeli D, Lehlbach K, Li-Weber M, Friedmann SL, Galle PR, Stremmel W, Oren M, Krammer PH. (1998). J. Exp. Med., 188: 2033-2045. MEDLINE

Nicoletti I, Migliorati G, Pagliacci MC, Grignani F, Riccardi C. (1991). J. Immunol. Meth., 139: 271-279.

Obe G, Palitt F, Tanzarella C, Degrassi F, De Salvia R. (1985). Mutat. Res., 150: 359-368. MEDLINE

Ochs K, Kaina B. (2000). Cancer Res., 60: 5815-5824. MEDLINE

Oldfield EH, Ram Z, Culver KW, Blaese RM, DeVroom HL, Anderson WF. (1993). Hum. Gene Ther., 4: 39-69. MEDLINE

Olive PL, Wlodek D, Banath JP. (1991). Cancer Res., 51: 4671-4676. MEDLINE

Orren DK, Petersen LN, Bohr VA. (1995). Mol. Cell. Biol., 15: 3722-3730. MEDLINE

Radford IR, Murphy TK. (1994). Int. J. Radiat. Biol., 65: 229-239. MEDLINE

Rubsam LZ, Davidson BL, Shewach DS. (1998). Cancer Res., 58: 3873-3882. MEDLINE

Seol DW, Billiar TR. (1999). J. Biol. Chem., 274: 2072-2076. Article MEDLINE

Slee EA, Harte MT, Kluck RM, Wolf BB, Casiano CA, Newmeyer DD, Wang HG, Reed JC, Nicholson DW, Alnemri ES, Green DR, Martin SJ. (1999). J. Cell Biol., 144: 281-292. MEDLINE

Srinivasula SM, Ahmad M, Fernandes-Alnemri T, Litwack G, Alnemri ES. (1996). Proc. Natl. Acad. Sci. USA, 93: 14486-14491. Article MEDLINE

Srinivasula SM, Ahmad M, Guo Y, Zhan Y, Lazebnik Y, Fernandes-Alnemri T, Alnemri ES. (1999). Cancer Res., 59: 999-1002. MEDLINE

Sun X-M, MacFarlane M, Zhuang J, Wolf BB, Green DR, Cohen GM. (1999). J. Biol. Chem., 274: 5053-5060. Article MEDLINE

Thust R, Tomicic M, Klöcking R, Voutilainen N, Wutzler P, Kaina B. (2000a). Cancer Gene Ther., 7: 107-117.

Thust R, Tomicic M, Klöcking R, Wutzler P, Kaina B. (2000b). Mutagenesis, 15: 177-184.

Tomicic MT, Eschbach E, Kaina B. (1997). Mutat. Res., 383: 155-165. MEDLINE

Tomicic MT, Christmann M, Kaina B. (2000). Biochem. Biophys. Res. Commun., 275: 899-903. Article MEDLINE

Tomicic MT, Kaina B. (2001). Biochem. Biophys. Res. Commun., 281: 404-408. Article MEDLINE

Vermes I, Haanen C, Steffens-Nakken H, Reutlingsberger C. (1995). J. Immunol. Meth., 184: 39-51.

Wang X, Studzinski GP. (1997). Exp. Cell Res., 235: 210-217. Article MEDLINE

Wei SJ, Chao Y, Hung YM, Lin WC, Yang DM, Shih YL, Chang LY, Whang-Peng J, Yang WK. (1998). Exp. Cell Res., 241: 66-75. Article MEDLINE

Wei SJ, Chao Y, Shih YL, Yang DM, Hung YM, Yang WK. (1999). Gene Ther., 6: 420-431. MEDLINE

Wesselborg S, Engels IH, Rossmann E, Los M, Schulze-Osthoff K. (1999). Blood, 93: 3053-3063. MEDLINE

Figures

Figure 1 Cell survival and apoptosis after exposure of CHO-HSV-TK cells to GCV. (A) Reproductive cell death was determined by colony formation. Two independent experiments in duplicates were performed. Fraction of survived colonies is expressed as percentage of control. (B) Induction of apoptosis as a function of GCV concentration was quantified 72 h after treatment. (C) Induction of apoptosis as a function of time was measured after exposure of exponentially growing cells for 14 h to 1 muM GCV. Apoptosis and necrosis were quantified by annexin V and propidium iodide staining via flow cytometry. (D) Internucleosomal fragmentation after exposure to 1 muM GCV. HSV-TK cells show a typical DNA laddering pattern (48 and 72 h post-exposure). CHO-neo cells and the untreated HSV-TK control show no fragmentation. As a control of DNA damage-induced apoptotic fragmentation, CHO-neo cells were exposed to 15 muM MNNG as a positive control of DNA fragmentation

Figure 2 Cell cycle progression and induction of DNA double-strand breaks after exposure to GCV. (A) The flow histograms showing cell cycle progression of CHO-HSV-TK cells with the sub-G1 DNA content at different post-exposure times were determined by flow cytometry. The sub-G1 fraction representing the apoptotic population is indicated by 'Apo'. Abscissa, relative DNA content; ordinate, relative cell number. Untreated cells (control) and cells exposed to 1 muM GCV for 14 h were assayed at the indicated time points (0-72 h) after drug exposure. (B) Induction of DNA double-strand breaks as a function of post-exposure time, as revealed by neutral SCGE. The relative comet length ('Olive tail movement') was defined as the product of the quantity of DNA in the comet tail and the distance between the mass centers of the comet head and comet tail. Fifty comets on each slide were randomly evaluated. The diagram shows the mean of three independent experiments±s.d.

Figure 3 Determination of caspase activity and modulation of apoptosis by caspase inhibitors. (A) Cytosolic extracts of CHO-HSV-TK cells were assayed for caspase-3, -8 and -9 activity as a function of time after exposure to 1 muM GCV. The relative caspase activity is represented as fold-increase of the untreated control. (B) Apoptosis was modulated by caspase inhibitors. CHO-HSV-TK cells were pre-treated for 10 h with 40 and 100 muM of the appropriate inhibitor and thereafter exposed to 1 muM GCV for 60 h and subjected to annexin V staining. zVAD, pan caspase inhibitor; zDEVD, zIETD and zLEHD, caspase-3, -8 and -9 inhibitor, respectively. Data (relative apoptosis expressed in % of GCV control) are the mean of three independent experiments±s.d. (C) Effect of various caspase inhibitors on caspase activity. Caspase inhibitors were administered at a concentration of 100 muM and were shown to act specifically. Significant undesired cross-inhibitory effects were not detected

Figure 4 Expression of PARP upon GCV treatment. (A) cleavage of PARP (p113) as a function of time after exposure of cells to 1 muM GCV. The relative increase of the PARP fragment (p85) was quantified by densitometry (Bioimage). (B) Inhibition of PARP cleavage in the presence of caspase inhibitors. CHO-HSV-TK cells were pre-incubated with an appropriate caspase inhibitor for 10 h and exposed to 1 muM GCV for 60 h. The membranes were stained with Poencaue R to confirm equal loading

Figure 5 Expression of apoptosis-regulating proteins after exposure to GCV. (A) Incubation with anti-Bcl-2 mAb (epitope 1-250) revealed unique signal of intact Bcl-2 (p26). For equal loading, the membrane was stained with Poencaue R. (B) Inverse correlation of Bcl-2 decrease and induction of apoptosis as a function of post-exposure time. (C) RT-PCR amplification of the 580 bp bcl-2 fragment after exposure to 1 muM GCV; gapdh (glyceraldehyde phosphodehydrogenase) was used as a loading control. (D) Expression of Bax, Bcl-xL and BAD in extracts of CHO-HSV-TK cells and the CHO-neo control after exposure to 1 muM GCV. Expression of BAD in extracts of GCV-treated CHO-HSV-TK cells; lower band, unphosphorylated form; upper band (designated as BAD-p), phosphorylated form. (E) Expression of Fas receptor (FasR) and Fas ligand (FasL) as a function of post-exposure time. The membranes were incubated with anti-ERK2 antibody as a loading control

Figure 6 Expression of cytochrome c (cyt c) and effect of caspase inhibitors on its release upon exposure to GCV. (A) Decrease in the amount of cyt c in the mitochondrial fraction (mCytc) and (B) increase in the amount of cyt c in the cytosol (cCytc), measured in parallel to (A). (C) Release of cyt c from mitochondria after exposure to 1 muM GCV in the presence of the broad caspase inhibitor (zVAD) and the specific caspase-9 inhibitor (zLEHD). For control of loading, the membranes were either incubated with anti-ERK2 or stained with Poencaue R

Figure 7 Cleavage of Bcl-2 and modulation of the cleavage by caspase inhibitors. (A) Incubation with anti-Bcl-2 pAb (epitope 1-250); upper band, a weak signal for intact Bcl-2 p26; lower band, a strong signal for the 23 kDa Bcl-2 cleavage product (Bcl-2 p23). CHO-neo cells treated with GCV were negative for the cleavage. (B) Effect of caspase inhibitors on Bcl-2 cleavage in the presence of 1 muM GCV, as shown by immunoblotting with anti-Bcl-2 mAb and pAb. CHO-HSV-TK cells exposed to 15 muM MNNG were used for comparison. The membranes were stained with Poencaue R to confirm equal loading

Figure 8 Modulation of apoptosis by transient transfection of caspase-9 and -3 expression constructs. (A) Expression of the human Flag-tagged wild-type caspase-9 (casp9-wt), dominant-negative caspase-9 (casp9-DN), and AU1-tagged dominant-negative caspase-3 (casp3-DN) was determined by Western blot analysis. (B) Modulation of GCV-triggered apoptosis (60 h after exposure to 1 muM GCV) by transient transfection with different expression constructs. Control, pcDNA3/pcDNA3.1-transfected cells. The expression of casp9-DN significantly reduced apoptosis. Expression of casp3-DN had only an insignificant effect. (C) Expression of caspase activity after transient expression of trans-dominant negative constructs and GCV exposure. (D) Modulation of Bcl-2 cleavage carried out in the transfectants treated (+) and not treated (-) with 1 muM GCV. Reduction of Bcl-2 cleavage was detected only in cells expressing casp9-DN. ERK2 was used as a loading control. Data of two separate experiments are shown. (E) Transient transfection of CHO-HSV-TK cells with the active form of human wild-type caspase-9 (casp9-wt) in the presence and absence of the caspase-3 inhibitor (zDEVD), without exposure to GCV

Figure 9 A model of induction of apoptosis by GCV. Incorporation of GCV-TP into DNA leads to generation of secondary DNA lesions within the subsequent (post-exposure) replication cycle due to errors in replication or repair of GCV-induced DNA damage. The critical, ultimate lesions are supposed to be DNA double-strand breaks (DSBs) that accumulate and lead to the formation of chromosomal aberrations and S- and G2/M-arrest. The early release of cytochrome c (cyt c) from mitochondria is probably due to a slight initial decrease in the mitochondrial membrane level of Bcl-2. Cyt c (in a complex with Apaf-1, dATP and pro-caspase-9) activates caspase-9, which in turn cleaves downstream executive caspases and other substrates such as PARP. Moreover, active caspase-9 and, to a lesser extent, caspase-3, cleave Bcl-2. The product of this cleavage, a pro-apoptotic Bcl-2 p23 fragment, promotes additional cyt c release and further caspase-9 activation, resulting in an amplification loop of Bcl-2-driven apoptosis

Received 6 September 2001; revised 17 December 2001; accepted 19 December 2001
28 March 2002, Volume 21, Number 14, Pages 2141-2153
Table of contents    Previous  Article  Next    [PDF]