Granzyme B-induced Cell Death Involves Induction of p53 Tumor Suppressor Gene and Its Activation in Tumor Target Cells*

In this study we investigated the involvement of p53 in cytotoxic T-lymphocyte (CTL)-induced tumor target cell killing mediated by the perforin/granzymes pathway. For this purpose we used a human CTL clone (LT12) that kills its autologous melanoma target cells (T1), harboring a wild type p53. We demonstrated initially that LT12 kills its T1 target in a perforin/granzymes-dependent manner. Confocal microscopy and Western blot analysis indicated that conjugate formed between LT12 and T1 resulted in rapid cytoplasmic accumulation of p53 and its activation in T1 target cells. Cytotoxic assay using recombinant granzyme B (GrB) showed that this serine protease is the predominant factor inducing such accumulation. Furthermore, RNA interference-mediated lowering of the p53 protein in T1 cells or pifithrin-α-induced p53-specific inhibition activity significantly decreased CTL-induced target killing mediated by CTL or recombinant GrB. This emphasizes that p53 is an important determinant in granzyme B-induced apoptosis. Our data show furthermore that when T1 cells were treated with streptolysin-O/granzyme B, specific phosphorylation of p53 at Ser-15 and Ser-37 residues was observed subsequent to the activation of the stress kinases ataxia telangiectasia mutated (ATM) and p38K. Treatment of T1 cells with pifithrin-α resulted in inhibition of p53 phosphorylation at these residues and in a significant decrease in GrB-induced apoptotic T1 cell death. Furthermore, small interference RNAs targeting p53 was also accompanied by an inhibition of streptolysin-O/granzyme B-induced apoptotic T1 cell death. The present study supports p53 induction after CTL-induced stress in target cells. These findings provide new insight into a potential role of p53 as a component involved in the dynamic regulation of the major pathway of CTL-mediated cell death and may have therapeutic implications.

In this study we investigated the involvement of p53 in cytotoxic T-lymphocyte (CTL)-induced tumor target cell killing mediated by the perforin/granzymes pathway. For this purpose we used a human CTL clone (LT12) that kills its autologous melanoma target cells (T1), harboring a wild type p53. We demonstrated initially that LT12 kills its T1 target in a perforin/ granzymes-dependent manner. Confocal microscopy and Western blot analysis indicated that conjugate formed between LT12 and T1 resulted in rapid cytoplasmic accumulation of p53 and its activation in T1 target cells. Cytotoxic assay using recombinant granzyme B (GrB) showed that this serine protease is the predominant factor inducing such accumulation. Furthermore, RNA interference-mediated lowering of the p53 protein in T1 cells or pifithrin-␣-induced p53-specific inhibition activity significantly decreased CTL-induced target killing mediated by CTL or recombinant GrB. This emphasizes that p53 is an important determinant in granzyme B-induced apoptosis. Our data show furthermore that when T1 cells were treated with streptolysin-O/granzyme B, specific phosphorylation of p53 at Ser-15 and Ser-37 residues was observed subsequent to the activation of the stress kinases ataxia telangiectasia mutated (ATM) and p38K. Treatment of T1 cells with pifithrin-␣ resulted in inhibition of p53 phosphorylation at these residues and in a significant decrease in GrB-induced apoptotic T1 cell death. Furthermore, small interference RNAs targeting p53 was also accompanied by an inhibition of streptolysin-O/granzyme B-induced apoptotic T1 cell death. The present study supports p53 induction after CTL-induced stress in target cells. These findings provide new insight into a potential role of p53 as a component involved in the dynamic regulation of the major pathway of CTL-mediated cell death and may have therapeutic implications.
Antigen-specific CD8 ϩ T cells play a crucial role in host defense against malignancies in both mouse and human models (1). In the T cell-mediated cytotoxicity process, two major pathways are involved after T cell receptor recognition of silver-major histocompatibility complex complexes expressed on target cells. The first one is an alternate pathway based on T cell receptor-induced surface expression of death receptor ligands (FasL, tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL), and TNF) on effector cells, which cross-links the corresponding receptors (Fas/CD95, TRAIL receptors, and TNF-RI/p55, respectively) on target cells (2). The second, which is undoubtedly the major pathway, is a secretory pathway involving receptor-triggered exocytosis of preformed secretory lysosomes, termed lytic granules (3,4). On the basis of findings in genetically manipulated mice, human genetic disease, and in vitro studies, the granule exocytosis pathway seems to have the dominant role in eliminating virus-infected cells and in tumor immunosurveillance (5). The cytotoxic granules contain the pore-forming protein perforin and a family of highly specific serine proteases know as granzymes. In mice and humans, A and B are the most abundant granzymes and have received the most attention, in particular granzyme B. It has been suggested that the latter induces target cell death by cleaving and activating the pro-apoptotic Bcl-2 family member Bid (6). Truncated Bid disrupts the outer mitochondrial membrane to cause release of proapoptotic factor cytochrome c and endonuclease G (7). It has also been suggested that granzyme B (GrB) 5 may induce target cell death by activating caspase 3 directly, by cleaving caspase substrates like poly ADP-ribose polymerase or inhibitor of caspase-activated DNase (CAD) to free CAD, and/or by cleaving several non-caspase substrates (8). GrB also directly disrupts the mitochondrial transmembrane potential in a caspase-and Bid-independent manner (3,9). However, despite these advances, the functional relationship between GrB and the tumor suppressor protein p53 remains unknown.
It is well established that an appropriate response to stress stimuli is crucial for preventing cellular transformation as well as for maintaining normal tissue function. The tumor suppressor protein p53 has a central role in protecting cells from a variety of stress stimuli such as DNA damage, nucleotide deple-tion, oncogene activation, or ␥-irradiation (10). The p53 protein has a short half-life and is often undetectable in normal cells. It is activated as a transcription factor through numerous post-translational modifications that allow its stabilization and accumulation in the nucleus to regulate target gene expression (11). Activated p53 induced transcription from promoters that harbor a p53 consensus binding site of genes involved in the maintenance of genetic stability and cellular homeostasis (12). Many apoptosis-related genes are regulated by p53, such as those encoding death receptors (13) and the proapoptotic Bcl-2 proteins p53-up-regulated modulator of apoptosis (Puma) (14) and Noxa (15). As an additional mode of apoptotic activity, p53 also accumulates in the cytoplasm, where it directly activates the proapoptotic protein Bax to promote mitochondrial outermembrane permeabilization (16). Moreover, it has been reported that in response to a broad spectrum of apoptotic stimuli, a fraction of wtp53 rapidly translocates to mitochondria in cell lines, in primary cells, and in vivo (17,18). Endogenous mitochondrial p53 physically interacts with the Bcl-2 family member proteins Bcl-X L and Bcl-2 and antagonizes their anti-apoptotic stabilization of the outer membrane (19). p53 possesses, therefore, a non-transcriptional function that is independent of its nuclear activity (20).
The physiological consequences of p53 activation essentially lead to cell cycle arrest, senescence, DNA repair, or apoptosis; thereby, p53 prevents cells from replicating a genetically compromised genome. Moreover, the ability of p53 to regulate the cell cycle and apoptosis has been reported to contribute to drug sensitivity induced by many anti-cancer agents (21). Nevertheless, the role of this tumor suppressor protein in the control of apoptosis mediated by cytotoxic T-lymphocyte (CTL) is not well documented. In this regard we have previously shown that p53 is a key determinant in anti-tumor CTL response as it regulates induction of Fas receptor expression, cellular FLICE/ caspase-8 inhibitory protein (cFLIP) short protein degradation, and CD95-induced activation of mitochondrial pathway in tumor cells (22,23). The present studies were designed to delineate the relationship between p53 and GrB during tumor-specific lysis. We first demonstrated that CTL-tumor target cell interaction resulted in p53 accumulation and activation. Such activation is mediated by GrB and contributes at least in part to GrB-induced apoptosis. The current study emphasizes that in addition to its role in controlling irradiation and drug responses, p53 also plays a key role in the regulation of CTLmediated apoptosis of tumor cells.
Tumor Cell Line and CTL Clone-The T1 tumor cell line was established from the primary lesion of a patient suffering from a melanoma (24) and was cultured at 37°C (5% CO 2 ) in RPMI 1640 with GlutaMax TM (Invitrogen) supplemented with 5% of fetal bovine serum (Invitrogen) and 5% Ultroser G (BioSera, France). The LT12 CTL clone was isolated from autologous tumor infiltrating lymphocytes as described previously (24) and was maintained at 37°C (5% CO 2 ) in complete medium (RPMI 1640 with GlutaMax TM ) (Invitrogen) supplemented with 1% sodium pyruvate (Invitrogen), 5% human serum (Institut Jacques Boy, Reims, France), and recombinant interleukin-2 in the presence of the autologous tumor cell line and irradiated LAZ and allogenic peripheral blood mononuclear cells.
Cell Death Analysis-T1 tumor cell sensitivity to LT12 was evaluated after interaction lasting 30 min and 1 h by 3,3Ј-dihexyloxacarbocyanine iodide (Dioc 6 (3)) and propidium iodide labeling (Molecular Probes, Eugene, OR). Inhibition of the perforin/granzymes-mediated cytotoxicity was performed using LT12 cells preincubated for 2 h with 100 nM concanamycin A (CMA) (Sigma). Cells were analyzed on a FACScalibur flow cytometer, and data were processed using Cell Quest software (BD Biosciences).
Cytotoxicity Assay-Cytotoxicity assays were performed using a standard 4-h chromium release assay. Briefly, 2 ϫ 10 3 51 Cr-labeled T1 target cells were incubated for 4 h at 37°C with effector cells (LT12) at different effector/target ratios in a final volume of 200 l in 96-well microplates. Experiments were performed in triplicate. At the end of the incubation 40 l of the supernatant was transferred into 96-well Luminaplate solid scintillation plates (Packard Instrument Co.) and, after overnight drying, counted in a Top Count ␤ counter (Packard). Data were expressed as the percentage of specific lysis at the T1/LT12 cell ratio indicated. The percentage of specific 51 Cr release (specific lysis of target cells) was calculated as (experimental release Ϫ spontaneous release)/(total release Ϫ spontaneous release) ϫ 100. Lytic units present in 10 7 effector cells were then assessed according to Pross et al. (25) using a computer program. One lytic unit was defined as the number of effector cells required for 30% lysis of 3 ϫ 10 3 target cells.
Inhibition of p53 Activation-p53 activation was inhibited by preincubating the T1 tumor cell line with 20 M pifithrin-␣ (PFT-␣) (BIOMOL Research Laboratories Inc., Plymouth, PA) for 48 h before co-culturing T1 and LT12 CTL clone or before loading with granzyme B.
Small Interference RNA (siRNA) Transfection-Two sequences of p53 siRNAs were used. The first siRNA used was designed with the Sigma-Proligo RNA interference designer tool: (siRNA p53_2) 5Ј-GUG AGC GCU UCG AGA UGU UdTdT, and the second sequence was purchased from Santa Cruz Biotechnology (NM_000546.2) (siRNA p53_JT) CGG CAU GAA CCG GAG GCC CAU dTdT. Subconfluent cells were transfected with siRNA in Opti-MEM I according to the manufacturer's instructions.
Confocal Scanning Immunofluorescence Microscopy-T1 cells cultured on coverslips were co-incubated with LT12 CTL clone (effector/target ratio 2/1) for 10, 30, or 60 min. Cells were washed with PBS, fixed with paraformaldehyde (4% w/v in PBS) for 1 h, and then permeabilized with SDS (0.1% w/v in PBS) for 10 min. Nonspecific sites were blocked with fetal bovine serum 10% in PBS for 20 min before staining with anti-p53 (DO-1) monoclonal Ab; p53 expression was detected by Alexa Fluor 546 (red) goat anti-mouse secondary Ab (Molecular Probes). Nuclear staining was performed with TO-PRO-3 (blue) (Molecular Probes). The coverslips were mounted on glass slides using a drop of Vectashield hard set (Vector Laboratories Inc., Burlingame, CA). The fluorescence was examined under an LSM 510 confocal microscope (Zeiss, Jena, Germany) as previously described (26).
Loading of Granzyme B-T1 cells were plated on coverslips at a density of 1.5 ϫ 10 5 cells per well in 6-well plates. Forty-eight hours later cells were loaded with recombinant granzyme B using sublytic doses of the pore-forming protein streptolysin-O (SLO) (Sigma, MDL number MFCD00132389). Briefly, SLO was preactivated by incubating for 30 min at room temperature in PBS containing 1 mM dithiothreitol. Cells were then washed in serum-free medium followed by dropwise addition of 150 l of RPMI containing 100 nM GrB to cell monolayers. Wells were flooded 15 min later with 1.5 ml of RPMI with 5% fetal bovine serum. Early apoptotic events were evaluated 30 min and 1, 2, or 4 h after loading GrB with Dioc 6 (3) and propidium iodide labeling.

The T1 Target Apoptotic Cell Death Induced by LT12 CTL Clone
Is Mediated by the Perforin/Granzymes Pathway-Initially we investigated the susceptibility of T1, the wild type p53 target cells, to the cytotoxicity induced by the autologous CTL clone LT12 using Dioc 6 (3) and propidium iodide labeling. As shown in Fig. 1, at an effector/target ratio of 2/1, LT12 induced between 30 and 35% apoptotic T1 cells by the times indicated. Incubation of these cells with staurosporin (1 M for 3 h) used as a positive control resulted in the induction of 50% of apoptotic cells. Moreover, preincubation of T1 target cells with CMA, an inhibitor of cytotoxic granules exocytosis by chelating free calcium, resulted in cell death being dramatically inhibited, indicating that the apoptotic death observed is mediated by the perforin/granzymes pathway.
p53 Accumulation in T1 Target Cells after Their Interaction with the LT12 CTL Clone-To gain more insights into p53 implication in the control of CTL-mediated lysis, we asked whether tumor T1 cell interaction with the CTL LT12 clone constitutes cellular stress sufficient to induce p53 activation in tumor cells. To this end we co-incubated T1 tumor target cells with LT12 CTL clone at an effector/target ratio of 2/1 for 10, 30, or 60 min. Western blot analysis consistently revealed that although low level expression of p53 was maintained in nonstressed T1 control target cells, T1/LT12 conjugation resulted in rapid p53 accumulation in T1 tumor target cells ( Fig. 2A). Using another melanoma cell line and its specific CTL clone, we obtained data confirming the p53 accumulation in target cells (supplemental Fig. S1). Because p53 activity depends on its expression level and its cell localization, p53 expression in T1 target cells was examined by confocal laser scanning microscopy. Data provided in Fig. 2B show rapid nuclear and cytoplasm p53 accumulation in T1 target cells after CTL-target conjugation. These results underline that CTL hitting of T1 target cells effectively represents significant stress sufficient to induce p53 accumulation. To determine the possible involvement of the cytotoxic granules exocytosis-dependent pathway in cytoplasmic and nuclear p53 accumulation and activity after T1 recognition by LT12, we inhibited the perforin/granzymemediated pathway using CMA. Western blot (Fig. 2C) or confocal microscopy (Fig. 2D) analysis showed that preincubation of the LT12 CTL clone with CMA resulted in the inhibition of cytoplasmic and nuclear p53 accumulation and activity at the time indicated. These data show that p53 accumulation and activation observed after T1/LT12 interaction is induced by the  NOVEMBER 9, 2007 • VOLUME 282 • NUMBER 45 perforin/granzymes pathway, further supporting the notion that it constitutes an effective stress in target cells.

p53, a Regulator of Granzyme B-induced Cell Death
The Inhibition of p53 Activity or Lowering of Its Expression Induced a Decrease in LT12-mediated T1 Target Lysis-To determine whether the cytoplasmic and nuclear p53 accumulation observed and its activation were involved in LT12-mediated lysis, we performed experiments to inhibit either p53 expression using gene silencing or its activation using PFT-␣ (27). T1 cells were transfected with siRNA targeting the p53 gene. As shown in Fig. 3A, Western blot analyses indicate that both p53 siRNA (siRNA 2 and siRNA JT) were effective in significantly lowering the p53 level (70%), whereas siRNA control (siRNA Sc, a non-targeting siRNA) had only a marginal effect. We then used the p53 inhibitor PFT-␣ that has been reported to inhibit, in vitro, p53-dependent gene transcription and to protect against a variety of genotoxic agents. As shown in Fig.  3B, when T1 tumor target cell line was preincubated with PFT-␣ before co-culture with the LT12 CTL clone, inhibition of target lysis (60% inhibition) by LT12 was observed. Furthermore, siRNA targeting p53 was also accompanied by dramatic inhibition of T1 lysis, whereas the siRNA Sc had only a slight effect.
Granzyme B-dependent Induction of p53 Phosphorylation and Accumulation in Tumor Target Cells by Granzyme B-Granule-mediated killing by cytotoxic T lymphocytes requires the combined action of the membranolytic protein perforin and granule-associated granzymes. Because the bacterial pore-forming toxin SLO was reported to have the same membranolytic and/or endosome disrupting properties as perforin, it was used in the course of these studies (28). To examine the effect of the perforin/granzyme B pathway on the induction of apoptosis in T1 cells, these cells were incubated with sublytic doses of SLO (2 g/ml) alone or in combination with human recombinant GrB (100 nM). Data shown in Fig. 4A indicate that SLO/GrB combination resulted in strong induction of apoptotic cell death, whereas SLO alone had only a slight effect.
The data produced using Western blot analysis and depicted in Fig.  4B indicate that SLO/GrB-induced apoptotic T1 cell death correlates with Bid and caspase 3 cleavage, whereas incubation of T1 cells with SLO alone has no effect on either events (Fig. 4B). To investigate the capability of GrB for inducing p53 accumulation during CTL/target tumor interaction, T1 cells were treated with SLO (2 g/ml) alone or in combination with recombinant GrB at the times indicated. The Western blot data shown in Fig. 5A indicate that SLO/GrB treatment resulted in the induction of p53 accumulation and transcriptional activity (data not shown). This accumulation was associated with the induction of apoptotic cell death in T1 cells (Fig. 4A). It should be noted, however, that treatment of T1 tumor cells with SLO alone resulted in a less significant increase

p53, a Regulator of Granzyme B-induced Cell Death
in p53 that was not accompanied by induction of apoptotic cell death (see Fig. 4A). Given that post-translational modifications such as serine and threonine phosphorylation are fundamental for p53 activation, we asked whether the concomitant SLO/GrB exposure interferes with p53 phosphorylation. Western blot analyses were performed using specific antibodies to evaluate the seven sites of p53 most commonly phosphorylated. Although no phosphorylation was detected in whole cell lysates of control cultures and in cells treated with SLO alone, specific bands corresponding only to Ser-15 and Ser-37 could be observed early on at 30 min and 1 and 2 h after exposure of T1 cells to SLO/GrB (Fig. 5B). These observations suggest that GrB can effectively induce p53 activation at least in part by a mechanism involving Ser-37 and Ser-15 phosphorylation after stress kinase activation. Given that ATM and p38K are involved in the phosphorylation of Ser-15 and Ser-37, respectively, we wondered whether SLO/GrB interfere with the activation of these stress kinases. The results illustrated in Fig. 5C indicate that whereas SLO alone had no effect, the SLO/GrB combination leads to the specific phosphorylation of ATM and p38K in T1 cells at 30 min and 1 and 2 h.
Granzyme B-mediated Target Cell Death Involves p53 Phosphorylation-Because p53 phosphorylation is essential in the regulation of its activity and to further investigate how p53 activity impacts on GrB-induced killing of T1 cells, we examined the relationship between p53 phosphorylation and GrBinduced apoptotic cell death. In this aim, to explore this, we preincubated T1 cells with the specific p53 inhibitor PFT-␣ (20 M) for 48 h. Data shown in Fig. 6A indicate that such treatment resulted in a significant decrease in GrB-induced apoptosis of T1 cells, which correlated with inhibition of Bid and caspase 3 cleavage (Fig. 6B). These data were confirmed using another melanoma cell line (supplemental Figs. 2 and 3). More interestingly, as depicted in Fig. 6C, the latter event correlated with inhibition of GrB-induced p53 phosphorylation at residues Ser-15 and Ser-37 especially at 1 and 2 h. These observations indicate that GrB-induced phosphorylation of p53 is a key event in coordinating the magnitude of apoptotic target killing.
siRNA Targeting p53 Induced Inhibition of SLO/GrB-mediated Apoptotic T1 Cell Death-To investigate the functional consequence of p53 silencing on SLO/GrB-induced apoptotic cell death in T1 cells, p53 was silenced in these cells by RNA   NOVEMBER 9, 2007 • VOLUME 282 • NUMBER 45

JOURNAL OF BIOLOGICAL CHEMISTRY 32995
interference. As shown in Fig. 7A, small interfering RNA against wild type p53 dramatically depressed p53 expression in T1 cells. We next examined the effect of SLO/GrB on the viability of siRNA-treated cells. Data depicted in Fig. 7B indicate a significant down-regulation in the percentage of apoptotic T1 cells induced by the SLO/GrB treatment.

DISCUSSION
Approaches to treatment of cancer based on the immune system have often focused on specific cytolytic effector cells such as CTLs (29). The present study was intended to provide insight into the functional relationship between GrB and p53 during tumor-specific lysis mediated by CTLs. Several phenomenological studies on T cell-mediated cytotoxicity in vitro have been extensively reviewed (30), and major advances have been made in understanding CTL-mediated apoptosis. However, increasing evidence from studies in patients and on cultured cells has highlighted the possibility that the induction of CTLs may be essential but not sufficient to the control of tumor progression (31). It is assumed that tumor cell growth in vivo is influenced not only by the ability of CTLs to recognize and

p53, a Regulator of Granzyme B-induced Cell Death
respond to the tumor but also by the susceptibility of tumor cells to host-mediated anti-tumor immune responses (32). Such susceptibility involves not only the effector and target cell features but also their reciprocal interaction, which so far remains not clearly understood. Recently we provided evidence indicating that tumor killing by autologous CTLs can be enhanced by targeting degranulation-independent mechanisms via restoration of wtp53, a key determinant of apoptotic machinery regulation (22,23). In this report we took advantage of the use of a human melanoma cell line displaying a wild type p53 and its autologous CTL clone. The latter is able to induce target lysis in a Ca 2ϩ -dependent manner involving the perforin/granzyme B pathway. GrB has been identified as the major mediator of the internal disintegration pathway of target cell damage. In this respect, mice bearing a homozygous null mutation of GrB are profoundly defective in rapid CTL-induced apoptosis (33). This serine-protease has been shown to be involved in triggering apoptosis in target cells, although the exact mechanism is not clear. In this regard, it has been reported that GrB can kill target cells even in the presence of caspase inhibitors (34), suggesting that caspase-independent pathways regulated by GrB may exist. Furthermore, Martin and co-workers (35) have recently shown that GrB directly attacks ␣-tubulin, a major component of the cell cytoskeleton.
We demonstrated in the course of these studies that the interaction between the CTL clone and its autologous target as well as the treatment of target cells with recombinant GrB induced cytoplasmic and nuclear accumulation of p53. These observations suggest that the granule exocytosis pathway contributes in the generation of stress signals to induce p53 accumulation that may be a key determinant in regulating target cell death mediated by CTL. More importantly, pharmacological suppression of p53 activity with PFT-␣ (36) or lowering of p53 using RNA interference resulted in significant inhibition of target cell lysis either by CTL or after treatment with GrB. This further confirms the crucial role of p53 in CTL-induced specific target cell death.
It is well established that the wtp53 protein is a critical transcription factor that responds to signals from a wide range of cellular stresses and allows the cell to cope with these stimuli by activating a set of target genes, facilitating adaptive and protective responses. It integrates cellular stresses, such as DNA damage and oncogenic transformation, to trigger either cell cycle arrest and senescence, DNA repair, or apoptosis (37). In the present studies, using TaqMan real-time quantitative reverse transcription-PCR analysis, we have shown that when LT12 were conjugated with autologous tumor targets, the transcription of p53 target genes (Mdm2, Noxa, PUMA, p21) was observed (supplemental Fig. S4). Delivery of GrB resulted in p53 accumulation in both the cytoplasm and nucleus and induced its phosphorylation at Ser-15 and Ser-37. This fits well with our findings demonstrating that GrB activates stress kinases including ATM and p38K involved in the phosphorylation of p53 and influencing its stabilization. As well as occurring with genotoxic signals, p53 induction may also occur in response to several stimuli (38). In this context it has been reported by Takaoka et al. (39) that the p53 gene is transcriptionally induced by interferon (IFN) ␣/␤ and that one mechanism of the anti-tumor action of IFN␣/␤ may involve p53 induction. The authors suggested that treating human cancer with interferon ␣/␤ in combination with chemotherapeutic drugs that activate p53 might be useful. It would be, therefore, of major interest to examine the ability of some chemotherapeutic drugs to potentiate the susceptibility of tumor cells to CTL or GrB. p53 peptide epitopes have been shown to be endogenously processed and presented by the human major histocompatibility class I and class II molecules of tumor cells. Given the high specificity of the CTL clone used in these studies to the melanoma-associated antigen MART-1, the possibility that wild type p53 peptide-derived epitopes contribute to the recognition of T1 cells by LT12 cells is unlikely.
Recent evidences suggest that granzyme A, the second most important granzyme, induces caspase-independent mitochondrial damage and rapid increase in reactive oxygen species (ROS). This latter process is a first step toward granzyme A-induced apoptosis that is blocked by superoxide scavengers (40). Considering evidence showing that GrB can cause rapid mitochondrial damage in the absence of Bid, Bax, and Bak (9) and recent data suggesting that stress-induced ROS generation induces a strong p53 activation (41), we can hypothesize that a rapid first step in GrB-induced apoptosis generates a ROS increase depending on GrB-induced mitochondrial damage, independently of Bid cleavage. In a second step, ROS generation may activate p53 that participates and/or regulates classi-  NOVEMBER 9, 2007 • VOLUME 282 • NUMBER 45

p53, a Regulator of Granzyme B-induced Cell Death
cal GrB-induced apoptosis, implicating Bid-induced mitochondrial outer-membrane permeabilization. In our experimental model, antioxidant N-acetyl cysteine had no effect on p53 accumulation in T1 target cells in response to CTL (data not shown), ruling out the involvement of ROS in p53 accumulation and subsequent activation. Ongoing experiments will elucidate the putative role of the Mdm2 pathway in the accumulation and activation of p53 in our experimental model. It should also be noted that several proteins other than Mdm2 regulate the stability of p53, some by influencing the interaction between Mdm2 and p53 and others by mechanisms independent of Mdm2 (42). Our results clearly point to a potential role of p53 in CTLinduced apoptosis of target harboring wild type p53. Given the fact that mitochondria is a central death regulator in response to DNA damage and is critical for p53-dependent cell death, it would seem crucial to determine how the death signal, GrB, and mitochondrial pathway are interconnected through p53 during the CTL-induced killing of target cells. Mihara et al. (19) have reported that p53 protein can directly induce permeabilization of the outer mitochondrial membrane by forming complexes with the protective Bcl-X L and Bcl-2 proteins, resulting in cytochrome c release. In parallel, p53 also accumulates in the cytoplasm, where it directly activates the proapoptotic protein Bax to induce mitochondrial release of apoptogenic factors (16). In this context, p53 seems to function like some of the proapoptotic members of the Bcl-2 superfamily called BH3-only proteins. In a recently published report, Chipuk et al. (43) suggested that Puma couples the nuclear and cytoplasmic proapoptotic functions of p53, where Puma is an enabler and p53 is an activator in a model in which Puma functions to release p53 from Bcl-X L , thereby freeing p53 to activate Bax. Moreover, in light of recent studies showing that p53 translocates to mitochondria in response to stress, Vousden (20)proposed that mitochondrial p53 could function as an enabler BH3-only protein to release an activator like Bid to interact with the antiapoptotic protein Bcl-X L . Given the fact that GrB induces target cell death by cleaving and activating the proapoptotic Bcl-2 family member Bid, which disrupts the outer mitochondrial membrane to cause release of the proapoptotic factor cytochrome c, it is tempting to speculate that p53 in our model acts by promoting GrBinduced truncated Bid release from the truncated Bid-Bcl-2/ Bcl-X L complex and also promotes cytochrome c release. It is conceivable that, to induce target killing, CTLs require both an apoptosis-sensitive phenotype and also a functional p53 pathway, which is crucial in the potentiation of target susceptibility to cell death.
Using another melanoma target cells, we demonstrated that an accumulation of p53 and its phosphorylation were observed after co-culture with the autologous CTL clone or treatment with recombinant GrB (data not shown), indicating that the reported observations are not the peculiarity of the LT12/T1 system. Accumulating evidence has been provided indicating that p53 and nuclear factor-B (NF-B) modulate each other in response to stress to stimulate gene expression and that this process is controlled by relative levels of each activated transcription factor and by competition for limiting pools of the transcriptional co-activators p300 and CBP (cAMP-response element-binding protein (CREB)-binding protein) (44). This interaction could, therefore, have many implications regulating the transcriptional decision-making mechanisms that govern cellular processes such as apoptosis. Here we demonstrated that although NF-B can be activated after effector/target conjugation or target treatment with recombinant GrB, inhibition of its activation has neither an influence on p53 accumulation nor on the killing of target cells (data not shown). Our findings suggest that, although raising the question about the potential molecular link between GrB and p53 accumulation, targeting the p53 pathway by GrB into tumor target cells may represent a novel biological end point for CTL-mediated apoptosis. Better understanding of the molecular insights into regulation of CTL-induced tumor cell death and its relationship with the p53 pathway may provide novel approaches to defining the sensitivity or resistance of tumor cells to anti-tumor therapy and new targets for rational therapeutic immune interventions.