Oncogene (2001) 20, 4601 ± 4612 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc ORIGINAL PAPERS Tissue speci®c expression of target suggests a key role for KILLER/DR5 in p53-dependent in vivo

Timothy F Burns1, Eric J Bernhard6 and Wa®k S El-Deiry*,1,2,3,4,5

1Laboratory of Molecular Oncology and Cell Cycle Regulation, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, PA 19104, USA; 2Howard Hughes Medical Institute, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, PA 19104, USA; 3Department of Medicine, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, PA 19104, USA; 4Department of Genetics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, PA 19104, USA; 5Department of Pharmacology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, PA 19104, USA; 6Department of Radiation Oncology, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, PA 19104, USA

The p53 tumor suppressor plays a key role in the cell's Keywords: p53; apoptosis; g irradiation; KILLER/ response to genotoxic stress and loss of this `guardian of DR5; c-FLIP-s; real-time RT ± PCR the genome' is an important step in carcinogenesis. The ability of p53 to induce apoptosis through transactivation of its target genes is critical for its function as tumor suppressor. We have found that overexpression of p53 in Introduction human cell lines resulted in apoptosis as measured by PARP cleavage. Furthermore we observed cleavage of The p53 tumor suppressor has been shown to play a both caspase 9 and caspase 8 after overexpression of p53 key role in mediating apoptosis induced by a variety of and found that p53-dependent apoptosis was inhibited by cellular and extracellular signals including genotoxic either cellular (c-Flip-s, Bcl-XL) or pharmacological stress and inappropriate proliferative signals (reviewed inhibitors of caspase 8 or caspase 9 respectively. These in Burns and El-Deiry, 1999). Mutation or loss of p53 results indicate that p53 is mediating apoptosis through has been observed in over 50% of all tumors and both the mitochondrial and death receptor pathways. To almost every tumor type (Hollstein et al., 1991). elucidate the relevant p53 target genes and examine the Furthermore, it has been estimated that the p53 caspase pathways utilized in vivo, we treated p53+/+ pathway is disrupted by mutation or inhibition of its and age matched p537/7 mice with 5 Gy ionizing function in the vast majority of tumors (Vogelstein and radiation or 0.5 mg/animal dexamethasone and har- Kinzler, 1992). Hereditary loss of p53 results in Li- vested tissues at 0, 6 and 24 h. We examined the mRNA Fraumeni syndrome which is characterized by a greater expression of p21, bax, KILLER/DR5, FAS/APO1 and than 50% incidence of neoplasia by the age of 30 EI24/PIG8 using TaqMan real time quantitative RT ± (Malkin et al., 1990). Although animals in which the PCR in the spleen, thymus and small intestine. Although p53 locus has been deleted develop normally, 75% of the basal mRNA levels of these genes did not depend on the p53-null animals develop tumors by six months of the presence of p53, we observed a p53-dependent age and all the p53-null animals develop tumors and induction of all these targets in response to g-irradiation die by 10 months of age (Donehower et al., 1992). Loss and a p53-independent regulation for p21 and KILLER/ of p53 or disruption of the p53 pathway is clearly a DR5 in response to dexamethasone. Furthermore, we critical step in carcinogenesis. have demonstrated that the relative induction of these p53 mediates many of its key functions through the p53 target genes is tissue speci®c. Despite observing transactivation of its target genes. Although p53 can, otherwise similar levels of death in these tissues, our in some situations, induce apoptosis in the absence of ®ndings suggest that in some cases apoptosis mediated transactivation (Haupt et al., 1995, 1996) or de novo through p53 occurs by redundant pathways or by a mRNA or synthesis (Caelles et al., 1994), most `group e€ect' while in other tissues one or few targets tumor-derived mutants of p53 are defective in DNA may play a key role in p53-dependent apoptosis. binding and transactivation, supporting a critical role Surprisingly, KILLER/DR5 is the dominantly induced for transactivation in p53's ability to suppress transcript in both the spleen and small intestine neoplasia. In response to DNA damage, cells can suggesting a potentially important role for this p53 either undergo G1 cell cycle arrest (Kastan et al., 1991) target in vivo. Oncogene (2001) 20, 4601 ± 4612. or apoptosis (Clarke et al., 1993; Lowe et al., 1993). p53-dependent G1 arrest is primarily mediated by p21WAF1, a cyclin dependent kinase inhibitor (El- Deiry et al., 1993). The target genes and pathways *Correspondence: WS El-Deiry Received 22 June 2000; revised 20 March 2001; accepted 26 March which mediate p53-dependent apoptosis are less well 2001 understood. In vivo apoptosis and KILLER/DR5 TF Burns et al 4602 Several p53 target genes have been proposed to irradiation induced apoptosis in two tissues (spleen and mediate p53-dependent apoptosis. These target genes thymus) which undergo p53-dependent apoptosis after tend to fall into two major classes based on the caspase irradiation (Hakem et al., 1998). Furthermore, both pathway they mediate their e€ects through (see Apaf-1 and caspase 9 were shown to be required for Thornberry and Lazebnik, 1998) Bax, a proapoptotic p53-dependent apoptosis after oncogene overexpression member of the Bcl-2 family, is the best characterized in MEFs (Soengas et al., 1999). The role of caspase 8 mediator of p53-dependent apoptosis (Miyashita and in p53-dependent apoptosis is unclear since similar Reed, 1995). Bax mediates cell death through the irradiation studies have not been performed in the induction of Cytochrome c release from the mitochon- caspase 87/7 animals (Varfolomeev et al., 1998). dria which results in the activation of caspase 9 (for However given the known p53 targets which act review see Adams and Cory, 1998). Although bax has through caspase 8 it is likely that caspase 8 plays a been shown to be regulated by p53, it is not required in role in mediating p53-dependent apoptosis. In this most forms of p53-mediated apoptosis. Bax7/7 study we examined the role of both caspase 8 and 9 in thymocytes and colonic epithelia are not defective in p53-dependent apoptosis in two human cancer cell gamma irradiation-induced apoptosis while p537/7 lines and investigated the expression of several p53 thymocytes and colonic epithelia are severely impaired target genes involved in apoptosis in an in vivo model (Knudson et al., 1995; Pritchard et al., 1999). However, of p53-dependent apoptosis. We have found that both Bax does seem to be required for p53-dependent caspase 8 and caspase 9 are cleaved during p53- apoptosis in the developing central nervous system dependent apoptosis and that speci®c inhibitors of the after g-irradiation (Chong et al., 2000) and contributes activity of each, block p53-dependent apoptosis. to DNA damage-induced apoptosis in E1A expressing Furthermore, we have examined the expression of the ®broblasts (McCurrach et al., 1997). Clearly there are p53 target genes: p21, bax, E124/PIG8, FAS/APO1 other genes which mediate p53-dependent apoptosis. and KILLER/DR5 after irradiation of p53-wild-type Several other p53 target genes have also been shown to and-null animals. We have found a striking tissue- function through the mitochondrial death pathway speci®c expression pattern for these apoptotic target including a recently discovered pro-apoptotic bcl-2 genes. Our ®ndings suggest that KILLER/DR5 may family member, Noxa (Oda et al., 2000). EI24/PIG8 play a key role in mediating p53-dependent apoptosis isolated independently from mouse and human cells as in vivo, for example in the spleen and small intestine a p53 target gene also appears to mediate its e€ects where its transcriptional induction following g-irradia- through the mitochondrial pathway (Lehar et al., 1996; tion predominates among the known pro-apoptotic Polyak et al., 1997; Gu et al., 2000). targets examined. These results suggest a role for both Two p53 targets have been shown to mediate cell the death receptor and mitochondrial pathways in death through the death receptor pathway. Fas/APO1, mediating p53-dependent apoptosis. a member of the tumor necrosis factor (TNF) receptor family, was found to be regulated by p53 by both transcriptional and transcription-independent mechan- Results isms (Owen-Schaub et al., 1995; Bennett et al., 1998). However, Fas/APO1 does not appear to be required p53 mediates apoptosis through both the mitochondrial for p53-dependent apoptosis in response to DNA and death receptor pathways damage (Fuchs et al., 1997). KILLER/DR5 is a proapoptotic member of the TRAIL (TNF-related In order to examine the downstream caspases through apoptosis-inducing ligand) receptor family and was which p53 induces cell death, we infected human identi®ed in a subtractive hybridization screen for p53 cancer cell lines lacking wild-type p53 with an target genes (Wu et al., 1997). Both Fas/APO1 and adenovirus expressing wild-type p53 or a control KILLER/DR5 transduce their death signals through adenovirus containing the LacZ gene. In both the the formation of a death-inducing signaling complex colon carcinoma cell line, SW480 and the breast (DISC) which includes the receptor, an adaptor carcinoma cell line, SKBR3 overexpression of p53 molecule and caspase 8 (for review, see Ashkenazi resulted in apoptosis by 48 h as evidenced by cleavage and Dixit, 1998). Ligation of the receptor leads to of the caspase 3 substrate, PARP (Figure 1a). To activation of caspase 8 and subsequent activation of elucidate the downstream caspases which are cleaved caspase 3 which results in apoptosis. Caspase 8 has during p53-mediated apoptosis, we examined the also been shown to cleave the pro-apoptotic bcl-2 cleavage of caspase 8 and caspase 9 after Ad-p53 family member, Bid, which leads to the activation of infection. Western blot analysis revealed that both caspase 9 through the mitochondrial pathway (Li et al., caspase 8 and caspase 9 were cleaved during p53- 1998). Therefore activation of the death receptor dependent apoptosis (Figure 1b). Furthermore, the pathway can induce its signal through both mitochon- cleavage of caspase 8 appeared to precede the cleavage dria-dependent and-independent pathways. of caspase 9 suggesting that activation of caspase 8 Several lines of evidence have suggested that the may be independent of caspase 9 activation. mitochondrial pathway is required for p53-dependent To determine the relative contributions of caspase 8 apoptosis. Studies using the caspase 97/7 animals and caspase 9 to p53-induced cell death, we over- have shown that caspase 9 is required for gamma- expressed p53 in SW480 cell lines that stably over-

Oncogene In vivo apoptosis and KILLER/DR5 TF Burns et al 4603

Figure 1 Infection of colon carcinoma cell line SW480 and breast carcinoma cell line SKBr3 with Ad-p53 results in caspase cleavage and apoptosis. (a) Western blot analysis for PARP cleavage after adenovirus infection. SW480 and SKBr3 cell lines were infected with an MOI of 50 with either Ad-LacZ or Ad-p53 and harvested at 0, 24 and 48 h after infection. Actin was used to con®rm that an equivalent amount of protein was loaded in each lane. (b) Western blot analysis of caspase 8 and 9 cleavage after adenovirus infection. (c) FACS analysis for percentage of cells with a sub-G1 content after adenoviral infection with Ad-LacZ or 5 Ad-p53 of SW480 stably overexpressing pLIB (control vector), pLIB-c-Flip-s or pLIB-Bcl-XL. 2.5610 cells were infected at an MOI of 50 with either Ad-LacZ or Ad-p53 and cells were harvested at 48 h after infection for FACS analysis. The Sub-G1 content of 20 000 cells was examined for each sample. This experiment was performed in triplicate. (d) FACS analysis for active caspase 3 in adrenal carcinoma cell line SW13 after transfection with p53 in the presence or absence of irreversible caspase 8 inhibitor or irreversible caspase 9 inhibitor. 5.06106 cells were transfected with 1.8 mg of pCEP4-p53 or control vector and 0.200 mg of pEGFP- Spectrin. Four hours after transfection either the irreversible caspase 8 inhibitor, Z-IETD-FMK or irreversible caspase 9 inhibitor, Z-LEHD-FMK or DMSO was added to the media at a ®nal concentration of 20 mM. Cells were harvested at 24 h. The percentage of cells expressing activated caspase 3 for 10 000 GFP positive cells were examined for each sample. This experiment was performed in triplicate. The chart depicts percentage of transfected cells which express activated caspase 3

expressed either c-FLIP-s or Bcl-XL which are speci®c inhibited by either a caspase 8 or caspase 9 inhibitor inhibitors of caspase 8 and caspase 9 activation (Figure 1d). These results further support a role for the respectively (Irmler et al., 1997; Kluck et al., 1997; death receptor and mitochondrial pathways in mediat- Zou et al., 1997; Tschopp et al., 1998; Adams and ing p53-dependent apoptosis. Cory, 1998). Either overexpression of c-Flip-s or Bcl- X was able to partially inhibit p53-mediated death L Similar levels of ionizing radiation-induced p53-dependent after Ad-p53 infection (Figure 1c). Furthermore we apoptosis in vivo examined the ability of speci®c irreversible inhibitors of caspase 8 (Z-IETD-FMK) and caspase 9 (Z-LETD- To explore the role of p53 targets genes in activating FMK) to inhibit p53-dependent apoptosis (Figure 1d). the caspase cascade and inducing cell death, we used Overexpression of p53 results in apoptosis in the an in vivo system of p53-mediated apoptosis. The study adrenal carcinoma cell line, SW13, which could be of p53-mediated apoptosis in vivo allows one to

Oncogene In vivo apoptosis and KILLER/DR5 TF Burns et al 4604 examine the e€ects of endogenous p53 on its target expression levels of p53 target genes normalized to genes when p53 is expressed at a physiological level. endogenous GAPDH mRNA levels. The result of four We decided to determine the relative levels of apoptosis independent trials revealed that in the spleen, thymus in the thymus and spleen which have previously been and small intestine, there was no statistically signi®cant shown to undergo a p53-dependent apoptosis in di€erence in the basal expression level of the ®ve p53- response to ionizing radiation (Lowe et al., 1993, target genes in p53+/+ and p537/7 animals (data Midgley et al., 1995). We treated p53+/+ and age not shown). We next examined the expression of these matched p537/7 mice with 5 Gy of ionizing radiation ®ve p53-target genes after treatment with dexametha- or dexamethasone and harvested tissues at 0, 6, and sone or 5 Gy of ionizing radiation in p53 wild-type and 24 h after treatment. We con®rmed that ionizing p53 null animals. radiation results in a p53-dependent apoptosis in the Previous studies and our data above (Figure 2) have thymus and spleen as PARP cleavage was only shown that the spleen undergoes p53-dependent observed in tissues from p53+/+ animals (Figure 2). apoptosis after g-irradiation (Midgley et al., 1995). Furthermore we observed similar levels of apoptosis as We examined the expression of p21WAF1 and found measured by PARP cleavage in thymus and spleen as expected it was strongly induced after 5 Gy in the (Figure 2). Because similar levels of apoptosis were p53+/+ animals by 6 h and was still elevated 24 h present in these tissues, we decided to determine after treatment but was unchanged in p537/7 animals whether the pattern of known p53 (Figure 3a). Interestingly, p21WAF1 was also induced target genes involved in apoptosis were similar after over fourfold by dexamethasone treatment at 6 h in a treatment with ionizing radiation in vivo. p53 independent manner (Figure 3a). However, the dexamethasone treatment did not cause cell death in the spleen (Figure 2). We next examined the expression p53 target genes involved in apoptosis are expressed in a levels of bax, Fas/APO1, KILLER/DR5 and EI24/ p53-dependent and tissue-specific manner following PIG8. Although these four apoptosis-inducing genes exposure to g irradiation in vivo were upregulated in a p53-dependent manner after g- We analysed the mRNA expression level of ®ve p53 irradiation they di€ered greatly in their extent of target genes in three tissues which undergo p53- induction. While the mRNA levels of bax and EI24 dependent apoptosis. We chose to analyse p21WAF1, were elevated approximately fourfold at 6 h, the bax, Fas/APO1, KILLER/DR5 and E124/PIG8. mRNA level of mouseKILLER/DR5 (MK) was p21WAF1 is a known cyclin-dependent kinase in- increased almost 30-fold at 6 h after g-irradiation. hibitor involved in p53-dependent growth arrest while The p53-dependent induction after irradiation of the other four targets have been implicated in p53- mouseKILLER/DR5 agrees with previous published dependent apoptosis (El-Deiry et al., 1993; Zhao et al., reports that the KILLER/DR5 promoter contains p53 2000). Bax and EI24/PIG8 have previously been shown binding sites and that KILLER/DR5 is induced in a to mediate cell death through the caspase 9 and the p53-dependent manner after DNA damage in human mitochondrial pathway while Fas/AP01 and KILLER/ and mouse tumor cell lines (Wu et al., 1997, 1999; DR5 induce death through caspase 8 (reviewed in Takimoto and El-Deiry, 2000). In contrast, Fas/APO1 Burns and El-Deiry, 1999). To determine the mRNA was only weakly induced (2.4-fold) by g-irradiation in expression levels of the ®ve targets we developed a the spleen (Figure 3b). Western blot analysis for p21 TaqMan real-time quantitative RT ± PCR assay with and bax con®rmed that these targets were induced in a

Figure 2 Ionizing radiation induces a p53-dependent apoptosis in the spleen and thymus. Western analysis of PARP cleavage after treatment in p53+/+ and p537/7 animals. Two animals per condition received 0.5 mg dexamethasone or 5 Gy total body irradiation and tissues were harvested at 0, 6, and 24 h after treatment. Left panel: spleen right panel: Thymus. Westerns are from individual (unpooled) tissues at each condition. Actin was used to con®rm that an equivalent amount of protein was loaded in each lane

Oncogene In vivo apoptosis and KILLER/DR5 TF Burns et al 4605

Figure 3 Expression of p53 target genes in the spleen after irradiation or dexamethasone treatment in p53+/+ and p537/7 animals. (a) p21 mRNA expression in p53+/+ animals and p537/7 animals after irradiation or dexamethasone treatment. (b) mRNA levels of four p53 target genes (upper left: EI24/PIG8; upper right: mouseKILLER/DR5 (MK); lower left: bax; lower right: Fas/APO1). mRNA expression levels were determined by TaqMan real time quantitative RT ± PCR assay (see Materials and methods). Results from individual tissues are shown. All expression levels are normalized to GAPDH in each well. Fold induction is de®ned as the fold increase for each sample relative to p53+/+ untreated animals. (c) Western blot analysis for bax and p21 after irradiation or dexamethasone treatment. Actin was used to con®rm that an equivalent amount of protein was loaded in each lane. (d) Oligonucleotide microarray analysis of total RNA from the spleen of irradiated and untreated p53+/+ and p537/7 animals for p21, mdm2, FAS/APO1, bax, and ei24/pig8. The chart depicts the fold induction of p21, mdm2, FAS/APO1, bax, and ei24/pig8 6 h after 5 Gy irradiation relative to the wild type untreated animals

Oncogene In vivo apoptosis and KILLER/DR5 TF Burns et al 4606 p53-dependent manner (Figure 3c). Furthermore dependent apoptosis in the spleen as compared to microarray analysis con®rmed that p21, bax, FAS/ other targets such as bax or FAS/APO1. APO1, EI24/PIG8 as well as mdm2 exhibited a similar We next analysed the expression pattern of pattern of induction relative to each other as observed p21WAF1, bax, Fas/APO1, KILLER/DR5 and EI24/ by Taqman RT ± PCR assay (Figure 3d, data not PIG8 in the thymus after g-irradiation. The mRNA shown). These results suggest that KILLER/DR5 may expression of p21WAF1 was dramatically increased by play a more signi®cant role in mediating p53- 6 h and remained elevated at 24 h after g-irradiation in

Oncogene In vivo apoptosis and KILLER/DR5 TF Burns et al 4607 a p53-dependent manner (Figure 4a). As observed in bax con®rmed that the target were induced in a the spleen, p21 mRNA was also induced by dex- p53-dependent manner, however, Bax protein did not amethasone treated in a p53-independent manner seem to be as signi®cantly induced compared to the (Figure 4a). Several studies have demonstrated that induction observed at the mRNA level (Figure 4c). These dexamethasone treatment results in a p53-independent results suggest that several target genes may be making a apoptosis (Lowe et al., 1993). Interestingly, KILLER/ signi®cant contribution to the p53 induced cell death in DR5 was also induced by dexamethasone at 24 h after the thymus. treatment in a p53-independent manner (Figure 4b). Finally we examined small intestine in which cells of This agrees with previous studies in human cancer cell epithelial origin have been shown to undergo p53- lines where it was observed that dexamethasone induced dependent apoptosis after exposure to g-radiation apoptosis correlated with KILLER/DR5 induction (Pritchard et al., 1999). Although p53-dependent regardless of p53 status (Meng and El-Deiry, 2001). As apoptosis was observed in the small intestine after observed in the spleen, all ®ve p53 target genes were ionizing radiation, KILLER/DR5 was the only p53 induced in the thymus by ionizing radiation in a p53- target gene whose mRNA levels signi®cantly increased dependent manner. However, unlike the spleen, all four after g-irradiation (Figure 5a,b). KILLER/DR5 of the apoptotic targets examined were increased by at mRNA was induced 10-fold at 6 h after irradiation least sevenfold at 6 h (Figure 4b). KILLER/DR5 was and remained elevated at 24 h. Western blot analysis again induced almost 30-fold at 6 h after irradiation and for p21 and bax con®rmed these results as p21 was remained elevated 24 h after treatment. Bax mRNA was only weakly induced and Bax protein levels did not also signi®cantly increased (15.7-fold) and remained signi®cantly change (Figure 5c). These results suggest a elevated at 24 h. Microarray analysis con®rmed these potentially unique role for KILLER/DR5 in mediating ®ndings as p21, bax, FAS/APO1, EI24/PIG8 as well as p53-dependent apoptosis in the small intestine follow- mdm2 exhibited a similar pattern of induction relative to ing exposure to g-radiation. each other as observed by Taqman RT ± PCR assay (Figure 4d, data not shown). It is possible that some post-transcriptional mechanism may further regulate the Discussion levels of Bax protein. Western blot analysis for p21 and To examine the mediators of p53-dependent apoptosis we have used both tissue culture and in vivo models of p53-dependent cell death. Previous studies have shown an important and in some cases essential role for the mitochondria and caspase 9 in p53 induced apoptosis (Hakem et al., 1998; Soengas et al., 1999). The role of caspase 8 in p53-dependent death is not well understood. Two target genes of p53 studied here, KILLER/DR5 and FAS/APO1, signal through cleavage of caspase 8 and when overexpressed lead to massive apoptosis (Wu et al., 1997, Owen-Schaub et al., 1995). In this study, we have shown that caspase 8 is cleaved at the same time or earlier than caspase 9 after p53 overexpression, indicat- ing that p53 is not activating caspase 8 through caspase 9 and 3 indirectly. Furthermore, both caspase 8 and caspase 9 appear to contribute to the observed p53- mediated cell death as either inhibitors of caspase 8 and caspase 9 activity respectively blocked p53-mediated Figure 4 Expression of p53 target genes in the thymus after apoptosis. Since this model of p53-dependent apoptosis irradiation or dexamethasone treatment in p53+/+ and p537/7 depended on overexpression of p53 above physiological animals. (a) p21 mRNA expression in p53+/+ animals and levels, we next examined a model of endogenous p53- p537/7 animals after irradiation or dexamethasone treatment. (b) dependent apoptosis in vivo. mRNA levels of four p53 target genes (upper left: EI24/PIG8; upper right; mouseKILLER/DR5 (MK); lower left: bax; lower Our in vivo studies suggest that p53-dependent right: Fas/APO1). mRNA expression levels were determined by apoptosis may occur through di€erent mediators TaqMan real time quantitative RT-PCR assay (see Materials and depending on the tissue type. Furthermore, our methods). Results from individual tissues are shown. All ®ndings suggest that in some cases apoptosis mediated expression levels are normalized to GAPDH in each well. Fold through p53 may occur by redundant pathways or by a induction is de®ned as the fold increase for each sample relative to p53+/+ untreated animals. (c) Western blot analysis for bax and `group e€ect' while in other tissues one or few targets p21 after irradiation. Actin was used to con®rm that an equivalent may play a key role in p53-dependent apoptosis. There amount of protein was loaded in each lane. (d) Oligonucleotide is some support for a single target gene being essential microarray analysis of total RNA from the thymus of irradiated as studies in the CNS revealed a key role for bax in and untreated p53+/+ and p537/7 animals for p21, mdm2, FAS/APO1, bax, and ei24/pig8. The chart depicts the fold mediating p53-dependent apoptosis (Chong et al., induction of p21, mdm2, FAS/APO1, bax, and ei24/pig8 6 h after 2000). The tissue type di€erences we observed in the 5 Gy irradiation relative to the wild type untreated animals expression of p53 target genes cannot be explained by

Oncogene In vivo apoptosis and KILLER/DR5 TF Burns et al 4608

Figure 5 Expression of p53 target genes in the small intestine after irradiation or dexamethasone treatment in p53+/+ and p537/7 animals. (a) p21 mRNA expression in p53+/+ animals and p537/7 animals after irradiation or dexamethasone treatment. (b) mRNA levels of four p53 target genes (upper left: EI24/PIG8; upper right: mouseKILLER/DR5 (MK); lower left: bax, lower right: Fas/APO1). mRNA expression levels were determined by TaqMan real time quantitative RT ± PCR assay (see Materials and methods). Results from individual tissues are shown. All expression levels are normalized to GAPDH in each well. Fold induction is de®ned as the fold increase for each sample relative to p53+/+ untreated animals. (c) Western blot analysis for bax and p21 after irradiation. Actin was used to con®rm that an equivalent amount of protein was loaded in each lane

Oncogene In vivo apoptosis and KILLER/DR5 TF Burns et al 4609 di€erences in the amount of apoptosis in each tissue as another p53 target gene not tested in this study may be we observed similar levels of death after irradiation in responsible for mediating the caspase 9 dependent the spleen and thymus. In the thymus we found that death. One candidate gene is the recently discovered bax, EI24/PIG8, FAS/APO1 and KILLER/DR5 in p53 target, Noxa, which is a BH3 containing molecule irradiated thymus are signi®cantly induced, suggesting expressed in the thymus (Oda et al., 2000). However, that p53 mediates its cell death through several based on the evidence presented, Noxa is upregulated redundant targets. Because p53 induces all of these by p53 either during cell cycle arrest or apoptosis genes strongly in the thymus one would predict that suggesting that it may be necessary but not sucient the loss of a single target would not attenuate p53- for cell death induction. Another explanation for the dependent apoptosis. Previous studies agree with this dissociation between a requirement for bax versus prediction as either bax or FAS/APO1 but not p53 caspase 9 is that in these tissues the cleavage of Bid were dispensable in the thymus for DNA damage- downstream of caspase 8 activation, for example by induced cell death (Knudson et al., 1995; Fuchs et al., KILLER/DR5, leads to the eventual activation of 1997) However, it is still possible that a single target caspase 9. Several studies have shown that both Fas gene may be essential for apoptosis, as knockout and TRAIL treatment can lead to Bid cleavage and animals of many p53 target genes including KILLER/ activation of caspase 9 (Li et al., 1998; Yamada et al., DR5 and EI24/PIG8 have not yet been tested in g- 1999). Furthermore, there is at least one in vivo radiation experiments. example in which Fas requires the mitochondrial In contrast to the thymus, the spleen and small pathway for inducing apoptosis (Yin et al., 1999). intestine exhibit a more restrictive expression pattern of The requirement of KILLER/DR5 and caspase 8 in p53 target genes during apoptosis. In spleen, Fas/APO, mediating p53-dependent apoptosis in the thymus and EI24/PIG8 and bax were either induced weakly or not spleen can be tested by irradiating the Bid7/7 or at all. KILLER/DR5 was strongly induced and was the caspase 87/7 animals and observing whether a cell dominant induced apoptotic p53 target gene in these death defects exists in the thymus or spleen after tissues. In the small intestine, an example of p53- irradiation. However, even if Bid7/7 tissues are not dependent death in epithelial tissue, KILLER/DR5 was de®cient in g-radiation-induced apoptosis, there may be the only p53 target gene examined in this study which redundant or yet to be discovered connections between was induced. These results suggests that in these two the cell membrane and mitochondrial death pathways. tissues, one or few p53 target genes may be essential In the small intestine, there is no published evidence for p53-dependent death. There are no published that caspase 9 is required for p53-dependent apoptosis reports that either FAS/APO1 or bax is required for and bcl-2 knockout animals have only a small increase apoptosis after ionizing radiation in the spleen. in apoptosis after irradiation, suggesting that the However, the weak induction of these target genes mitochondrial pathway may not be as important in and EI24 suggests that they probably are not playing a the small intestine (Pritchard et al., 1999). Interestingly, key role in p53-dependent apoptosis in these tissues. KILLER/DR5 was the only p53 target gene examined Previous studies in the small intestine have shown that that was signi®cantly upregulated in the small intestine bax is not required for p53-dependent apoptosis following exposure to g-radiation. These results suggest (Pritchard et al., 1999). This result agrees with our that KILLER/DR5 may play a unique role in ®ndings that bax is not signi®cantly induced in the mediating p53-dependent apoptosis in the small small intestine. Our evidence suggests a key role for intestine. It remains to be seen whether KILLER/ KILLER/DR5 in mediating p53-dependent apoptosis DR5 is required and whether it is dependent upon the in the spleen and especially the small intestine. Whether mitochondrial pathway in radiation-induced death of KILLER/DR5 is an essential gene for mediating p53- the small intestine. Irradiation studies of the KILLER/ dependent death or one of a few critical targets DR5 and Bid knockout animals and examination of remains to be determined by examining the KILLER/ the resulting apoptosis in the small intestine would in DR5 knockout animals for defects in radiation induced the future address these questions. apoptosis. It is possible that the requirement for one or Given the potentially unique role of KILLER/DR5 multiple target genes in a particular tissue may be in mediating p53-dependent death after ionizing dependent upon the balance of pro- or anti-apoptotic radiation in vivo, it is not surprising that mutations molecules in that particular tissue. of KILLER/DR5 in head and neck and cancer and in Although de®nitive evidence for an essential role for non-small cell lung have been found (Pai et al., most p53 targets in vivo is lacking, several animal 1998; Lee et al., 1999) Further support for a potential studies have shown a critical role for caspase 9 is role for KILLER/DR5 in human cancer, comes from mediating DNA damage induced apoptosis in both the previous reports which demonstrated that the anti- thymus and spleen (Hakem et al., 1998). Our data apoptotic TRAIL decoy receptor, TRID/TRAIL-R3 is suggests that in these tissues KILLER/DR5 may play overexpressed in primary tumors of the gastrointestinal an important role in mediating p53-dependent apopto- tract (Sheikh et al., 1999). Because KILLER/DR5 is sis. Since caspase 9 but not bax has been shown to be uniquely induced by p53 in the small intestine after g- required for radiation-induced death in both the irradiation it may be necessary for primary tumors of thymus and spleen, there are several possible explana- the gastrointestinal tract to upregulate TRID/TRAIL- tions for our observations. One explanation is that R3 to prevent p53-dependent apoptosis (Sheikh et al.,

Oncogene In vivo apoptosis and KILLER/DR5 TF Burns et al 4610 1999; Meng et al., 2000). It is possible that in some The Bcl-XL and c-FLIP-s (short form) (Irmler et al., 1997) cases loss of KILLER/DR5 may substitute for loss of genes were ampli®ed from a human placenta retroviral library p53 or at least may increase a tumor cell's tolerance to (Clontech) and cloned into the Not1 site of pLIB (Clontech) wild-type p53. respectively and fully sequenced prior to use in transfection In summary, we have demonstrated in two human studies. tumor cell lines that both caspase 8 and caspase 9 are cleaved during p53-dependent apoptosis and that Transfections and fluorescence-activated cell sorting (FACS) inhibitors of death receptor or mitochondrial death analysis signaling inhibit p53-mediated cell death. Our studies in Transfections were carried out as previously described (Wu et vivo have revealed striking tissue speci®c patterns of al., 1997). In brief, before transfection, 56105 cells/6 cm well expression for p53 target genes. Our data suggests that were seeded. The cells were transfected with lipofectamine p53-mediated apoptosis occurs in some tissues such as Plus (Life Technologies Inc.) according to the manufacturer's the thymus through redundant pathways or by a speci®cations. Four hours after transfections the media was `group e€ect' while in other tissues one or few targets changed and an irreversible caspase inhibitor, Z-IETD-FMK may play a key role in p53-dependent apoptosis. (caspase 8) or Z-LEHD-FMK (caspase 9) (R & D Systems) was added to the media at a ®nal concentration of 20 mM. Finally we have found evidence that KILLER/DR5 Cells were harvested 24 h after transfection and prepared for may be playing a key role in mediating p53-dependent detection of active-caspase 3 by ¯ow cytometry as previously apoptosis in vivo and may have a unique role among described (Ozoren et al., 2000). For adenoviral experiments, transcriptional targets of p53, causing cell death in cells were harvested 48 h after infection and ®xed and stained response to g-irradiation in the small intestine. with propidium iodide as previously described (Ozoren et al., 2000). Cell sorting was performed on a Coulter Epics Elite counter.

Materials and methods Western blotting Animals and treatments Total cellular protein was harvested in 16Laemmli sample bu€er and quantitated using the Amido Black Staining Healthy 6 ± 7-week old female p53+/+ and p537/7 method as previously described (Meng et al., 1998). A total animals were obtained from Jackson Laboratories (Bay of 80 mg of protein per lane was loaded on either a 10% Harbor, Maine, USA). Two animals in each experimental (PARP) or 15 % SDS ± PAGE gel and electrophoresed and group received either 0.5 mg/animal of dexamethasone electroblotted as previously described (El-Deiry et al., 1994). (Elkins-Sinn Inc, Cherry Hill, NJ, USA) or sterile 1X PBS The following antibodies were used in these experiments: by single bolus intraperitoneal injection. Total body irradia- rabbit polyclonal anti-PARP (Boehringer Mannheim), mouse tion was performed using a 137Cesium g-source at a dose rate monoclonal anti-caspase 8 (Cell Signalling Technologies), of 1.532 Gy/min. At 0, 6 and 24 h, the mice were euthanized rabbit polyclonal anti-caspase 9 (Imgenex), mouse mono- using an approved Institutional Animal Care and Use clonal anti-actin (Santa Cruz), rabbit polyclonal anti-p21 Committee Protocol, which followed recommendations of (Santa Cruz), and rabbit polyclonal anti-bax (Santa Cruz). the Panel on Euthanasia of the American Veterinary Medical Association. Tissues were harvested and snap frozen in liquid nitrogen and kept at 7808C until used. Taqman real time quantitative RT ± PCR assay for mRNAs of p21, bax, EI24, Fas/APO1, mKILLER/DR5, and GAPDH Cell culture and adenoviral/retroviral infection Taqman RT ± PCR assay was conducted according to the manufacturer's speci®cations (PE Applied Biosystems). In The culture conditions for SW480, HCT 116 and SkBr3 were brief, oligonucleotides (probes) for TaqMan RT ± PCR were maintained as previously described (Wu et al., 1997). The labeled with FAM(6-carboxy¯uorescin) (p21, bax, EI24, Fas/ adrenal carcinoma cell line, SW13 was a kind gift from R APO1, and mKILLER/DR5) or VIC (GAPDH) and 3' prime Sheikhattar (Wistar Institute). Ad-LacZ and Ad-p53 were quencher, TAMRA (6-carboxyl-N,N,N',N'-tetramethylrhoda- obtained from B Vogelstein (Johns Hopkins University). All mine). The following primer and probe sequences where used adenoviruses were propagated, titered, and ampli®ed as for each gene respectively: mp21primers ± 5'-CGAGAACGG- previously described (El-Deiry et al., 1993). The retroviral TGGAACTTTGAC-3',5'-TCCCAGACGAAGTTGCCCT- vector pLIB (Clontech), pLIB-EGFP (Clontech), pLIB-c- 3', probe: 6FAM-TCGTCACGGAGACGCCGCTG-TAM- Flip-s, or pLIB-Bcl-XL was cotransfected with pVSVG and RA; bax primers: 5'-GGAGCAGCTTGGGAGCG-3',5'- pCgP (Dr Michael Malim, University of Pennsylvania) into AAAAGGCCCCTGTCTTCATGA-3', probe: 6FAM-CGG- 293T cells using the CaPO4 method. The resulting super- GCCCACCAGCTCTGAACA-TAMRA; FAS/APO1 pri- natant was used to infect the SW480 cell line (36105 cell/well mers: 5'-TCGGAGATGCTATTAGTACCTTGAGTAT-3', plated 48 h before infection) at a titer necessary to obtain 5'-GATCTGGGCTGTCCTGCCT-3', probe: 6FAM-TGGT- greater than 90% infectivity as previously described (Fouch- GCTTGCTGGCTCACAGTTAAGAGT-TAMRA; mKIL- ier et al., 1997). LER/DR5 primers: 5'-ATAAAAAGAGGCTGTGAACG- GG-3',5'-GGTCCAAGAGAGACGA-3', probe: 6FAM-TC- Plasmids CCGAAAGTGCGAACTCTGTGCA-TAMRA; EI24/PIG8 primers: 5'-GAAGGATGTGGCGAAGGGA-3',5' CCA- The wildtype p53 expression vector pCEP4-p53 (El-Deiry et CAAGACCGTCTACCTGCA-3', probe: 6FAM-CAGCCC- al., 1993) and control vector were generously supplied by Bert TGAGCAGCTCGTCCTCTG-TAMRA. Whenever possible Vogelstein (Johns Hopkins University). The pCMVEGFP- probe and primer sets were designed over intron/exon Spectrin vector was previously described (Kalejta et al., 1997). boundaries to prevent ampli®cation of genomic DNA. All

Oncogene In vivo apoptosis and KILLER/DR5 TF Burns et al 4611 primer pairs were tested to ensure that the proper size Oligonucleotide microarray analysis fragment was generated. GAPDH primer and Vic labeled probe were obtained from PE Applied Biosystems. All Total RNA was isolated from individual tissues as previously primers and probes were designed with the use of Primer described (El-Deiry et al., 1993) and ampli®ed, labeled, and Express Version 1.0 (PE Applied Biosystems). Total RNA hybridized to the Murine 11K gene chips as previously was isolated from individual tissues as previously described described (Lockhart et al., 1996; Wodicka et al., 1997). For (El-Deiry et al., 1993) and 1 mg was used for reverse each RNA sample the average intensity of all genes on the transcription and ampli®cation using TaqMan Reverse chip was adjusted to 1500 to allow for comparison and Transcription Reagents according to the manufacturer's subsequent analysis (Lockhart et al., 1996; Wodicka et al., protocol (PE Applied Biosystems). Total RNA samples were 1997). Analysis was performed using the A€ymetrix Micro- electrophoresed on a MOPS-formaldehyde gel and the array Suite 4.0 according to the manufacturer's protocol. pattern of the 28S and 18S ribosomal bands were observed to verify the quality of the total RNA preparation (El-Deiry et al., 1993). A master mix of TaqMan reagents was prepared and 10 ng of each RT sample was used in the Taqman PCR reaction. Each tube contained both a gene probe and primers and a GAPDH control probe and primer. Each sample was done in quadruplicate. The increase in ¯uorescence (D Rn) Acknowledgments was proportional to the concentration of template in the This work was presented at the 91st Annual American PCR. The PCR cycle number at threshold represents CT. Association for Cancer Research meeting in San Francisco, Reactions in which reverse transcriptase was not added to the CA, USA on April 2nd, 2000, and at the 10th International RT rxn were used to control for genomic contamination. Any p53 Workshop in Monterey, CA, USA on April 8th, 2000. contribution of genomic contamination to the CT value for a The authors thank David T Dicker for excellent technical sample was ignored since the CT value of the (7)RT rxn was assistance with ¯ow cytometry. We thank the MIT at least 10 cycles less than for samples in which reverse Biopolymers facility for hybridization of our A€ymetix transcriptase was added. The standard curve method was gene chips. TF Burns was supported in part by an NIH- used to quantitate amounts of each gene relative to the MSTP training grant to the University of Pennsylvania. GAPDH amount in each reaction according to the This work was supported by NIH grants CA 75454 (WS manufacturer's protocol (PE Applied Biosystems). Reactions El-Deiry) and CA 75138 (EJ Bernhard and WS El-Deiry). were carried out in 96-well plates using the ABI PRISM 7700 WS El-Deiry is an Assistant Investigator of the Howard Sequence Detection System (PE Applied Biosystems). Hughes Medical Institute.

References

Adams JM and Cory S. (1998). Science, 281, 322 ± 325. Hakem R, Hakem A, Duncan GS, Henderson JT, Woo M, Ashkenazi A and Dixit VM. (1998). Science, 281, 1305 ± Soengas MS, Elia A, de la Pompa JL, Kagi D, Khoo W, 1308. Potter J, Yoshida R, Kaufman SA, Lowe SW, Penninger Bennett M, MacDonald K, Chan S-W, Luzio JP, Simari R JM and Mak TW. (1998). Cell, 94, 339 ± 352. and Weissberg P. (1998). Science, 282, 290 ± 293. Haupt Y, Rowan S, Shaulian E, Vousden K and Oren M. Burns TF and El-Deiry WS. (1999). J. Cell. Physiol., 181, (1995). Genes Dev., 9, 2170 ± 2183. 231 ± 239. Haupt T, Barak Y and Oren M. (1996). EMBO J., 15, 1596 ± Caelles C, Helmberg A and Karin M. (1994). Nature, 370, 1606. 220 ± 223. Hollstein M, Sidransky D, Vogelstein B and Harris CC. Chong MJ, Murray MR, Gosink EC, Russel HRC, (1991). Science, 253, 49 ± 53. Srinivasen A, Kapsetaki M, Korsmeyer SJ and McKinnon Irmler M, Thome M, Hahne M, Schneider P, Hofmann K, PJ. (2000). PNAS, 97, 889 ± 894. Steiner V, Bodmer J-L, Schroter M, Burns K, Mattmann Clarke AR, Purdie CA, Harrison DJ, Morris RG, Bird CC, C,RimoldiD,FrenchLEandTschoppJ.(1997).Nature, Hooper ML and Wyllie AH. (1993). Nature, 362, 849 ± 388, 190 ± 195. 852. Kalejta RF, Shenk T and Beavis AJ. (1997). Cytometry, 29, Donehower LA, Harvey M, Slagle BL, McArthur MJ, 286 ± 291. Montgomery Jr CA, Butel JS and Bradely A. (1992). Kastan MB, Onyekwere O, Sidransky D, Vogelstein B and Nature, 356, 215 ± 221. Craig RW. (1991). Cancer Res., 51, 6304 ± 6311. El-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons Kluck RM, Bossy-Wetzel E, Green DR and Newmeyer DD. R, Trent JM, Lin D, Mercer E, Kinzler KW and (1997). Science, 275, 1132 ± 1136. Vogelstein B. (1993). Cell, 75, 817 ± 825. Knudson CM, Tung KSK, Tourtellotte WG, Brown GAJ El-Deiry WS, Harper JW, O'Connor PM, Velculescu VE, and Korsmeyer SJ. (1995). Science, 270, 96 ± 99. Canman CE, Jackman J, Pietenpol JA, Burrell M, Hill LeeSH,ShinMS,KimHS,LeeHK,ParkWS,KimSY,Lee DE, Wang YS, Wiman KG, Mercer WE, Kastan MB, JH, Han SY, Park JY, Oh RR, Jang JJ, Han JY, Lee JY Kohn KW, Elledge SJ, Kinzler KW and Vogelstein B. and Yoo NJ. (1999). Cancer Res., 59, 5683 ± 5686. (1994). Cancer Res, 54, 1169 ± 1174. Lehar SM, Nacht M, Jacks T, Vater CA, Chittenden T and Fouchier RAM, Meyer BE, Simon JHM, Fischer U and Guild BC. (1996). Oncogene, 20, 1181 ± 1187. Malim M. (1997). EMBO J., 16, 4531 ± 4539. Li H, Zhu H, Xu CJ and Yuan J. (1998). Cell, 94, 491 ± 501. Fuchs EJ, McKenna KA and Bedi A. (1997). Cancer Lockhart DJ, Dong H, Byrne MC, Follettie MT, Gallo MV, Research, 57, 2550 ± 2554. Chee MS, Mittmann M, Wang C, Kobayashi M, Horton Gu Z, Flemington C, Chittenden T and Zambetti GP. (2000). H and Brown EJ. (1996). Nature Biotechnol., 14, 1675 ± Mol. Cell. Biol., 20, 233 ± 241. 1680.

Oncogene In vivo apoptosis and KILLER/DR5 TF Burns et al 4612 Lowe SW, Schmitt EM, Smith SW, Osborne BA and Jacks T. Soengas MS, Alarcon RM, Yoshida H, Giaccia AJ, Hakem (1993). Nature, 362, 847 ± 852. R, Mak TW and Lowe SW. (1999). Science, 284, 156 ± 159. Malkin D, Li FP, Strong FC, Fraumeni JFJ, Nelson CE, Takimoto R and El-Deiry WS. (2000). Oncogene, 19, 1735 ± KimDH,KasselJ,GrykaMA,Bischo€FZandTainsky 1743. MA. (1990). Science, 250, 1233 ± 1238. Thornberry NA and Lazebnik Y. (1998). Science, 281, 312 ± McCurrach ME, Connor TMF, Knudson CM, Korsmeyer 316. SJ and Lowe SW. (1997). PNAS, 94, 2345 ± 2349. Tschopp J, Irmeler M and Thome M. (1998). Curr. Opin. Meng RD, Shih H, Prabhu NS, George DL and El-Deiry Immumol., 10, 552 ± 558. WS. (1998). Clin. Cancer Res., 4, 251 ± 259. Varfolomeev EE, Schuchmann M, Luria V, Chiannikulchai Meng RD, McDonald ER, Sheikh MS, Fornace Jr AJ and N, Beckmann JS, Mett IL, Rebrikov D, Brodianski VM, El-Deiry WS. (2000). Mol. Ther., 1, 130 ± 144. Kemper OC, Kollet O, Lapidot T, So€er D, Sobe T, Meng RD and El-Deiry WS. (2001). Exp. Cell Res., 262, Avraham KB, Goncharov T, Holtmann H, Lonai P and 154 ± 169. Wallach D. (1998). Immunity, 9, 267 ± 276. Midgley CA, Owens B, Briscoe CV, Thomas DB, Lane DP Vogelstein B and Kinzler KW. (1992). Cell, 70, 523 ± 526. and Hall PA. (1995). J. Cell. Sci., 108, 1843 ± 1848. WodickaL,DongH,MittmannM,HoM-HandLockhart Miyashita T and Reed JC. (1995). Cell, 80, 293 ± 299. DJ. (1997). Nature Biotechnol., 15, 1359 ± 1367. Oda E, Ohki R, Murasawa H, Nemoto J, Shibue T, Wu GS, Burns TF, McDonald III ER, Jiang W, Meng R, Yamashita T, Tokino T, Taniguchi T and Tanaka N. KrantzID,KaoG,GanD-D,ZhouJ-Y,MuschelR, (2000). Science, 288, 1053 ± 1057. Hamilton SR, Spinner NB, Markowitz S, Wu G, El-Deiry Owen-Schaub LB, Zhang W, Cusack JC, Angel LS, Santee WS. (1997). Nature Genet., 17, 141 ± 143. SM, Fujiwara T, Roth JA, Deisseroth AB, Zhang W-W, Wu GS, Burns TF, Zhan Y, Alnemri ES and El-Deiry WS. Kruzel E and Radinsky R. (1995). Mol. Cell. Biol., 15, (1999). Cancer Res., 59, 2770 ± 2775. 3032 ± 3040. Yamada H, Tada-Oikawa S, Uchida A and Kawanishi S. Ozoren N, Kunhong K, Burns TF, Dicker DT, Moscioni D (1999). Biochem. Biophys. Res. Commun., 265, 130 ± 133. and El-Deiry WS. (2000). Cancer Res., 60, 6259 ± 6265. YinXM,WangK,GrossA,ZhaoY,ZinkelS,KlockeB, PaiSI,WuGS,OzorenN,WuL,JenJ,SidranskyDandEl- Roth KA and Korsmeyer SJ. (1999). Nature, 400, 886 ± Deiry WS. (1998). Cancer Res., 58, 3513 ± 3518. 891. Pritchard DM, Potten CS, Korsmeyer SJ, Roberts S and Zhao R, Gish K, Murphy M, Yin Y, Notterman D, Ho€man Hickman JA. (1999). Oncogene, 18, 7287 ± 7293. WH, Tom E, Mack DH and Levine AJ. (2000). Genes Polyak K, Xia Y, Zweier JL, Kinzler KW and Vogelstein B. Development, 14, 981 ± 993. (1997). Nature, 389, 300 ± 305. Zou H, Henzel WJ, Liu X, Luschg A and Wang X. (1997). Sheikh MS, Huang Y, Fernandez-Salas EA, El-Deiry WS, Cell, 90, 405 ± 413. Friess H, Amundson S, Yin J, Meltzer SJ, Holbrook NJ, Fornace Jr AJ. (1999). Oncogene, 18, 4153 ± 4159.

Oncogene