Mitochondrially targeted wild-type p53 suppresses growth of ... - Nature

Oncogene (2006) 25, 6133–6139

& 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc

SHORT COMMUNICATION

Mitochondrially targeted wild-type p53 suppresses growth of mutant p53 lymphomas in vivo G Palacios and UM Moll Department of Pathology, Stony Brook University, SUNY at Stony Brook, Stony Brook, NY, USA

The complex apoptotic functions of the p53 tumor suppressor are central to its antineoplastic activity in vivo. Besides its well-known action as a transcriptional activator of apoptotic genes, p53 exerts a direct proapoptotic role at the mitochondria via protein–protein interactions with Bcl2 family members, thus executing the shortest known circuitry of p53 death signaling. We recently reported that exclusive delivery of p53 to mitochondria exerts a significant in vivo tumor suppressor activity in p53-null lymphomas. However, it was unknown whether mitochondrially targeted p53 has suppressor activities in tumors harboring missense mutants, which constitute the vast majority of p53 alterations in human tumors. Here, we show that targeting p53 to mitochondria does confer a significant growth disadvantage in B-lymphomas expressing various point mutants of p53, resulting in efficient apoptosis induction in vitro and in vivo in mice. Oncogene (2006) 25, 6133–6139. doi:10.1038/sj.onc.1209641; published online 8 May 2006 Keywords: p53; lymphoma

apoptosis;

mitochondria;

Em-Myc

The p53 tumor suppressor is a crucial protector against cancer by eliminating cells with mutagenic alterations. The pleiotropic apoptotic function of p53 is central to its antineoplastic activity. p53 responds to death stimuli by rapid stabilization and activation and mediates apoptosis primarily via the mitochondrial pathway. p53 can act as a transcription factor of proapoptotic target genes like Puma, Noxa, Bax, Bid, p53AIP1 and others. A second mechanism involves a direct role of wild-type p53 protein at the mitochondria via protein– protein interactions with Bcl2 family members (Marchenko et al., 2000; Sansome et al., 2001; Dumont et al., 2003; Mihara et al., 2003; Erster et al., 2004; Arima et al., 2005; Erster and Moll 2005). In response to a broad range of death stimuli like genotoxic drugs, g-irradiation, hypoxia and deregulated viral and cellular oncogenes, a fraction of induced protein rapidly (within Correspondence: Dr UM Moll, Department of Pathology, Stony Brook University, SUNY at Stony Brook, BSTL9, Room 134, Stony Brook, NY 11794-8691, USA. E-mail: [email protected] Received 9 November 2005; revised 20 March 2006; accepted 29 March 2006; published online 8 May 2006

30–60 min) translocates to mitochondria in primary, immortal and transformed cells in culture (Marchenko et al., 2000; Sansome et al., 2001; Mihara et al., 2003; Nemajerova et al., 2005). p53 translocation precedes cytochrome c release, collapse of inner mitochondrial membrane potential and procaspase-3 activation (Marchenko et al., 2000). Mitochondrial p53 engages in inhibitory complexes with antiapoptotic BclXL and Bcl2 via its p53 DNA-binding domain, as shown by molecular modeling, mutational and nuclear magnetic resonance studies (Mihara et al., 2003; Leu et al., 2004; Petros et al., 2004; Moll et al., unpublished results). Mitochondrial p53 induces oligomerization of Bak, a resident Bcl2 member of the outer mitochondrial membrane (OMM), leading to OMM permeabilization and robust release of apoptotic activators like cytochrome c from the intermembraneous space (Mihara et al., 2003). Outer mitochondrial membrane permeabilization by p53 is dependent upon p53/BclXL complex formation, and transactivation-deficient missense mutants of p53 concomitantly lose or compromise their ability to interact with BclXL and promote cytochrome c release. p53 mutations may therefore represent ‘double hits’, eliminating the transcriptional as well as the direct mitochondrial function of p53 (Mihara et al., 2003). Thus, one mechanism of mitochondrial p53 involves BH3-only like actions of derepression, that is, inhibiting the antiapoptotic Bcl2 members (Chen et al., 2005). Mitochondrial p53 also liberates proapoptotic Bak from its sequestration by antiapoptotic Mcl-1 by forming a p53/Bak complex, thus promoting Bak oligomerization (Leu et al., 2004; Moll et al., unpublished results). A third cytosolic p53 arm was recently proposed and might occur in those cell types that express a soluble subfraction of BclXL in the cytosol (Chipuk et al., 2004, 2005). Here, stress-induced cytosolic p53 is thought to be initially sequestered into an inactive complex by cytosolic BclXL. (Bcl2 is exclusively membrane-bound and does not participate in this arm (Hsu et al., 1997).) In a subsequent phase, which requires p53 transactivation of its target gene PUMA, p53 is liberated from this complex by a newly formed PUMA/BclXL complex (Chipuk et al., 2005). Liberated p53 can then directly activate cytosolic monomeric Bax in a proposed ‘hit and run’ mechanism (Chipuk et al., 2004). Likely then, freed p53, perhaps after its transient encounter with Bax, can also translocate to mitochondria to act as described above.

Mitochondrially targeted wild-type p53 G Palacios and UM Moll

6134

Spleen

Thymus

Lymphoma D278G

Lymphoma V135A

Lymphoma R279P

Lymphoma Arf-/-

NIH3T3

a

Lymphoma p53-/-

In sum, mitochondrial p53 appears to have properties of a ‘super-BH3-only’ molecule, as it has a dual action of derepression of antiapoptotic and activation of proapoptotic monocyte-derived mesenchymal progenitor (MOMP) regulators (Kuwana et al., 2005). This direct mitochondrial p53 program participates in the physiologic p53 response after genotoxic damage in mice (Erster et al., 2004). In radiosensitive but not radioresistant organs from normal mice treated with a single dose of g-irradiation or etoposide, mitochondrial p53 translocation triggers rapid caspase-3 activation and apoptosis in a subpopulation of cells, which in thymus precedes the induction of PUMA, the fastest known p53 target gene relevant for this organ (Mihara et al., 2003). This suggests that the mitochondrial pathway also participates in tumor suppression in vivo.

p53 (CM5) p19 Arf PCNA

b

LTR

CMV GFP-bla LTR

1

393 Nuclp53 p53CTB (BclXL fusion) p53CTM (Bcl2 fusion)

Abrogation of p53-mediated apoptosis is a fundamental defect in human tumors and promotes chemoresistance; hence, its functional restoration is a premier therapeutic goal. For example, conventional p53 gene therapy (based on adenovirus) was recently approved by the Chinese Food and Drug Administration for patients with head and neck squamous cell carcinoma and is in late-stage clinical trials for other malignancies as well (Peng, 2005). Importantly, p53, when targeted to mitochondria and devoid of transcriptional activity, is sufficient to induce apoptosis in a transcription-independent manner. Depending on the specific targeting motif and cell type, mitochondrial targeting of p53 partly or wholly phenocopies nuclear p53 in launching effective apoptosis and colony suppression directly from the mitochondrial

Figure 1 (a) Em-myc lymphomas overexpressing endogenous mutant p53 were chosen from 18 Em-myc lymphomas by screening tumor lysates (50 mg proteins/lane) for strongly overexpressed p53 levels in Western blots. The p53 R279P, V135A and D278G mutations were confirmed by reverse transcription–polymerase chain reaction-based sequencing of Exons 4–9. Mouse NIH3T3, p53-null lymphoma, Arf-null lymphoma and normal thymus and spleen were used as controls. Immunoblots with specific antibodies for mouse p53 (CM5, Vector Laboratories, Burlingame, CA, USA) and p19Arf (AbCam Inc., Cambridge, MA, USA). Proliferating-cell nuclear antigen (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) is the loading control. Secondary antibodies were peroxidase-conjugated goat immunoglobuline (Amersham Biosciences, Piscataway, NJ, USA), detected with SuperSignal West Femto (Pierce Biotechnology Inc., Rockford, IL, USA). (b) Scheme of murine stem cell virus (MSCV)-p53 cytomegalovirus–green fluorescent protein–blasticidine constructs used in this work. p53 inserts were either nuclear human wild-type p53 (Nuclp53), human wildtype p53 fused at the COOH-terminus to the transmembrane domain of human BclXL (p53CTB) or Bcl2 (p53CTM), respectively, or human wild-type p53 fused at the NH2-terminus to the mitochondrial import leader sequence of ornithine transcarbamylase, followed by a Flag epitope tag (Lp53wt). The control virus lacked the p53 insert.

LFp53wt

Figure 2 Mitochondrially targeted p53 induces apoptosis in p53 mutant primary lymphoma cells in vitro. (a) Lp53wt induces apoptosis of various p53 mutant lymphoma cells. Fluorescence-activated cell sorting (FACS) analysis of fresh cells from different primary lymphoma isolates overexpressing the indicated endogenous mutant p53 proteins. Freshly harvested cell suspensions were plated in six-well plates on mitomycin-arrested NIH3T3 feeders and grown overnight in 45% Iscove’s medium, 45% Dulbecco’s medium, 10% fetal bovine serum. Cells were centrifuged three times at 3000 r.p.m. in 6 h intervals with equal aliquots from the same batches of conditioned medium containing either empty vector or Lp53wt retroviruses, respectively, followed by selection in blasticidine (50 mg/ml) for 24 h and FACS analysis. The left panels are size scatter plots in which the cell populations with low forward and high side scatter on the left exhibit typical morphological signs of apoptosis. Encircled areas represent live cells. The right panels are FACS analyses of live cells from the corresponding left panels to determine the proportion of remaining green fluorescent protein (GFP)-positive cells after viral expression. (Before selection, the initial proportions of GFP-positive cells in each pair were about 10% for vector and 25% for Lp53wt (not shown).) (b and c) p53CTB induces apoptosis in p53 mutant lymphoma cells, as shown by TdTmediated dUTP nick-end labeling (TUNEL) and FACS assays in two representative experiments. The results were very similar for p53CTM (not shown). p53 R279P mutant lymphoma cells were infected by centrifugation with viral supernatants with the indicated retroviruses as in (a). Infected cultures were selected (b) or sorted to 99% purity, followed by 12 h recovery (marked as 1d in c). Twentyfour hours after infection (b) or 1 and 5 days after infection (c), cultures were subjected to FACS analysis to determine the remaining GFP-positive cells. In parallel, aliquots were cytospun onto coated slides, fixed and stained for TUNEL (Roche; Hoechst counterstaining). Apoptosis was quantitated using the imaging software AxioVision v4.3. (d) Different mitochondrial targeting motifs of p53 yield different apoptotic efficiencies. Mutant p53 R279P lymphoma cells were infected with the indicated retroviruses as in (a) and allowed to recover for 12 h, followed by FACS analysis for GFP-positive cells. Cultures were adjusted to 40% green cells with nonlabeled parental cells. Thirty-six hours after the last centrifugation with viral supernatants, without additional DNA damage, aliquots were cytospun onto coated slides, fixed and stained for TUNEL. Bars represent means 7 s. e. of at least five independent experiments. All bars are significantly different from each other (Po0.01). Oncogene


6135

b

p53-R279P

Vector

In light of these results, the next important step is to establish whether mitochondrially targeted p53 also has in vivo suppressor activity in mutant p53 tumors, thus reflecting the vast majority of p53 alterations in human tumors (Olivier et al., 2002). As in our study on p53-null lymphomas, we exploited the well-characterized transplantable Em-myc transgenic mouse model of lymphomagenesis (Harris et al., 1988). These mice overexpress

Hoechst

Tunel Parental

800

a

1000

platform in p53-null cancer cell lines in vitro (Marchenko et al., 2000; Mihara et al., 2003). We recently showed that this direct mitochondrial p53 program also exerts a significant tumor suppressor activity in vivo. In mice with B-cell lymphomas of p53-null or ARF-null genotypes, exclusive delivery of p53 to the OMM of tumor cells caused efficient apoptosis and growth suppression within the animals (Talos et al., 2005).

600

GFP-Positive 0.03% M2

400

GFP-Positive M1

M2

Vector

R1

0

Live cells 37%

1000

0

200

400 600 FSC-H

800

Vector

GFP-Positive 47%

M1

1000

M2

p53-R279P

800

Lp53wt

p53-CTB

600

Lp53wt

p53-CTB

GFP-Positive 28%

M1

400

GFP-Positive M1

200

SSC-H

M1

Parental

200

SSC-H

Vector

M2

M2

R1

Live cells 0.2%

Nuclp53

0 0

200

400 600 FSC-H

800

M1

1000

M2

800

Vector

GFP-Positive

99

M2

1000

p53-V135A

600

Lp53wt

M1

15

0 1000

400 600 FSC-H

800

%GFP+

13

11 %TUNEL+

Vector

R1 200

25

45

36

Live cells 0.9% 0

p53CTB

Nuclp53

1000

d

p53-D278G

800

Vector

60

600

Vector

400

50 M1

GFP-Positive M2

40

200

R1

% TUNEL+

SSC-H

M2

69

28

1d

400

GFP-Positive

200

SSC-H

800

Lp53wt

68

5d

800

99

1d

400 600 FSC-H

99

5d

1000

200

% cells

0

Live cells 65% 0

100 90 80 70 60 50 40 30 20 10 0

1d

M1

5d

400

600

c

R1

0

Live cells 18%

1000

0

200

400 600 FSC-H

800

1000

p53-D278G

Lp53wt

30

800

20

600

Lp53wt

400

10

200

M1

Live cells 0.2% 0

200

400 600 FSC-H

800

GFP-Positive M2

0 Parental

R1

0

SSC-H

GFP-Positive 7%

p53-V135A

Vector

200

SSC-H

Nuclp53

1000

Vector

p53CTB

Lp53wt

Nuclp53

Retroviruses

1000

Oncogene


6136

the c-myc oncogene in their B-cell lineage and develop Burkitt-type pre-B- and B-cell lymphomas within 4–6 months of age. Test genes can be efficiently introduced by retroviral transfer into primary lymphoma isolates ex vivo and injected into tail veins of normal syngeneic recipients for lymphoma reconstitution. Tumor formation and aggressiveness depends on disabling the p53 pathway (Eischen et al., 1999). Tumor apoptosis occurs via the mitochondrial pathway (Schmitt et al., 2000). Three spontaneously arising lymphomas, harboring different p53 missense mutations (R279P, V135A and D287G, sequence confirmed) were isolated from different transgenic mice. These tumors arose in p53 þ / mice that have mutated their only p53 allele, or in p53 þ / þ that have lost one p53 allele and mutated the other. In any event, they do not contain a p53 wild-type allele in addition to their mutated one. Each tumor overexpressed the corresponding mutant p53 protein, whereas p19Arf levels were also elevated (Figure 1). To determine if mitochondrially targeted p53 exerts apoptotic and tumor suppressor activity in these tumors, we used a series of recombinant MSCV viruses (murine stem cell virus) expressing three different versions of mitochondrially targeted p53, together with green fluorescent protein (GFP) and a selectable marker (called Lp53wt, p53CTM and p53CTB) that we previously generated and characterized (Mihara et al., 2003; Talos et al., 2005). Briefly, Lp53wt is a fusion of wtp53 with the prototypical mitochondrial import leader of ornithine transcarbamylase, whereas p53CTB and p53CTM are fusions of wtp53 with the transmembrane domains of BclXL and Bcl2, respectively. All are expressed at levels comparable to that of stress-induced endogenous wtp53, localize to and act exclusively at mitochondria but not the nucleus, and are devoid of transcriptional activity in sensitive reporter assays (Mihara et al., 2003; Talos et al., 2005). Control viruses lacked p53 inserts (vector) or expressed conventional nuclear wtp53 (Nuclp53). Primary lymphoma isolates were infected and selected to determine apoptosis in vitro 2 days later. Mitochondrially targeted p53

effectively induced apoptosis in all mutant p53 lymphoma isolates. As shown in Figure 2a, Lp53wt retrovirus but not control virus depleted GFP-expressing live cells and increased the amount of dead cells, as determined by fluorescence-activated cell sorting (FACS) analyses. Similarly, p53CTB induced apoptosis, as determined by TdT-mediated dUTP nick end labeling staining (TUNEL) and GFP expression, albeit to a lesser degree and slower kinetics than nuclear p53 (Figure 2b and c). Interestingly, we consistently observed a higher apoptotic activity of Lp53wt than of p53CTB and p53CTM (which equaled each other) in multiple cell types (Figure 2d and unpublished data). To assess whether the observed tumor cell apoptosis in vitro translates to tumor suppression in vivo at natural sites, we performed competitive lymphoma reconstitution experiments. Forty percent of freshly infected, GFP-positive mutant p53 (R279P) lymphoma cells (after blasticidine selection or sterile FACS sorting, confirmed by GFP at 36 h after infection) were mixed with 60% GFP-negative parental p53 (R279P) lymphoma cells and injected intravenously into syngeneic non-transgenic recipients (1 106 cells per mouse). Four weeks later, all tumors from a given recipient mouse were pooled and their residual GFP positivity was determined by FACS to measure responsiveness. As seen in Figure 3a and b and Supplementary Table 1, all mitochondrial versions of p53 were able to significantly induce apoptosis (e.g. Figure 3c) and suppress tumor growth of infected mutant p53 tumor cells compared to empty virus. Although p53CTB and p53CTM were partially suppressive (76% down from vector), Lp53wt achieved almost the same degree of suppression as nuclear p53 (94 versus 99%, respectively), thus recapitulating in vitro results (compare Figure 3a (black bars) with Figure 2d). As expected, Nuclp53 with its full range of pleiotropic actions, including all activating and repressing transcriptional functions, had the strongest suppressor activity in mutant lymphoma cells. This was also reflected by the fact that residual Lp53wt and Nuclp53 proteins were undetectable at day 28 in five of five reconstituted

Figure 3 Mitochondrially targeted p53 suppresses growth in mutant p53 harboring lymphomas in vivo. (a) For competitive lymphoma reconstitution in vivo, 40% of green fluorescent protein (GFP)-positive primary lymphoma cells harboring mutant p53 R279P (black bars) (determined by fluorescence-activated cell sorting (FACS) 36–48 h after transduction with the indicated retroviruses and 2–4 h before intravenous injection) were mixed with 60% of unlabeled parental cells. Cells (1x106) in 50 ml phosphate-buffered saline were injected into the tail vein of normal syngeneic 12- to 14-week-old C57BL/6 recipients (Jackson Laboratories, Bar Harbor, Maine, USA) (approved by Stony Brook University Animal Care and Use Committee). After 28–30 days, the majority of mice had developed visible lymphomas (mainly cervical and inguinal, but also axilar, perianal, visceral and intramuscular) and the remainder showed symptoms of sickness and all were killed. All tumors from a given animal were pooled. To evaluate lymphoma apoptosis, half of the pooled tumorous lymph node mass of each animal was minced in red blood cell lysis buffer, collected and immediately subjected to FACS analysis to determine the percentage of remaining GFP-positive cells. (The remaining samples were used for immunoblotting (d)). The other half of each tumor pool was fixed in 10% buffered formalin for histology (c). The number of mice per group is indicated above the bars. Data are reported as means7s.e. The data for this graph are tabulated in detail in Supplementary Table 1. For comparison, our previous results obtained in p53-null lymphomas using the same competitive assay are also shown (Talos et al., 2005) Lp53wt it was not done for p53-null lymphoma. (b) Summary Table of data used to construct graph in (a). (c) Tumors derived from mixtures containing p53CTB-infected cells show apoptotic activity in situ. Representative histological sections of reconstituted 40:60 lymphomas from panel a harvested at day 28. Hexatoxylin and eoxin staining, magnification 400. (d) Detectability of residual mitochondrially targeted p53 proteins in reconstituted tumors varies and is associated with their growth suppressive abilities (a). Tumors harvested at day 28 were pooled for each of five different mice per indicated retroviral group. Total cell lysates were prepared and immunoblotted for human p53 with monoclonal antibody DO-1 (Santa Cruz) (50 mg of total proteins were loaded per lane; proliferating-cell nuclear antigen as loading control). p53CTB and p53CTM are still detectable in reconstituted tumors, whereas Lp53wt and Nuclp53 (positive control) are not. Oncogene


6137

a

40

n=17

% GFP-positive tumor cells

p53 R279P

n=18

35

p53 -/-

30 25 20 15 n=6

n=12

10 5

n=4

n=13

n=8

0

Vector

p53CTB

p53CTM

*

n=11

Lp53wt

n=12

Nuclp53

Retroviruses

b

Summary of in vivo reconstituted tumors of normal mice injected with p53 R279P mutant lymphoma and p53 null cells expressing mitochondrially targeted p53

Lymphoma cells

Retrovirus

Ratio of injected lymphoma cells

Number of mice with tumors

Average tumor weight 1,4 (g)

Percent tumor cells GFP-positive at 28-30 days post-injection 1

(transduced:parental)

p53 R279P

p53 -/-

p – value 2

Separate statistical groups 3

Vector – GFP p53CTB – GFP p53CTM – GFP Lp53wt – GFP Nuclp53 – GFP

40:60 40:60 40:60 40:60 40:60

17 12 6 8 11

0.54 ± 0.47 ± 0.58 ± 0.52 ± 0.49 ±

0.11 0.13 0.16 0.11 0.11

32.0 ± 7.78 ± 7.99 ± 1.80 ± 0.33 ±

5.08 0.89 2.00 0.17 0.17

4.9x10-4 1.2x10-2 5.1x10-4 3.5x10-5

X

Vector – GFP p53CTB – GFP p53CTM – GFP Nuclp53 – GFP

40:60 40:60 40:60 40:60

18 4 13 12

1.00 ± 0.40 ± 1.00 ± 0.40 ±

0.05 0.05 0.11 0.09

30.8 ± 0.0 ± 0.4 ± 0.0 ±

4.24 0.00 0.28 0.00

5.6x10-3 4.8x10-6 3.8x10-5

X

X X X X

X X X

1

mean ± standard error p - value calculated against the Vector group. 3 significant differences at p < 0.01; yes if vertically non-aligned; no if vertically aligned 4 no significant differences 2

c

d 1

2

3

4

5

p53 Vector PCNA 1

2

3

4

5

p53 p53CBT PCNA 1

2

3

4

5

p53 p53CTM PCNA 1

2

3

4

5

p53 Lp53wt PCNA 1

2

3

4

5

p53 Nuclp53 PCNA

Oncogene


6138

tumors each. p53CTB and p53CTM proteins remained detectable, consistent with residual cells (Figure 3d). When compared to p53-null lymphomas (Talos et al., 2005), mutant p53 lymphomas were somewhat less susceptible to killing mediated by mitochondrially targeted p53 (Figure 3a, black and white bars). In sum, targeting p53 to mitochondria in mutant p53 lymphomas overexpressing an altered protein induces apoptosis and a significant growth disadvantage in vitro and in vivo. Aberrations at the level or downstream of mitochondria can interfere with the mitochondrial p53 death program. A recent report on the mechanism of differential apoptotic susceptibility in the MCF7 breast cancer cell line found no correlation between mitochondrial p53 translocation and apoptosis. Although both treatments induced endogenous p53 translocation, MCF7 cells were sensitive towards topoisomerase inhibitor etoposide but resistant towards g-irradiation, and instead underwent senescence. Ectopic p53CTM also failed to induce apoptosis in the absence of further DNA damage, whereas ectopic wtp53 did so (Essmann et al., 2005). Several explanations could account for this effect. g-irradiation might induce very high levels of p21Waf1 and/or p27Kip1 inhibitors that mediate a dominant arrest program in MCF7, or etoposideactivated wtp53 transactivates a limiting factor required for translocation and oligomerization of cytosolic Bax, which g-irradiation-activated p53 or p53CTM cannot. At any rate, our results in a broad spectrum of p53-null and mutant cell types indicate that this factor is not limiting in those cells. The consistently higher apoptotic activity of Lp53wt versus p53CTB/CTM is intriguing and might suggest additional intramitochondrial prodeath functions of p53, aside from its surface action where it regulates Bcl2 members and thus OMM permeability. This effect appears to be physiological, as a matched control fusion protein composed of the same imported leader sequence followed by the c-Rel transcription factor sequence failed to be apoptotically active above empty vector (Marchenko et al., 2000). A subfraction of stresstranslocated mitochondrial wild-type p53 is found in complexes with mthsp70 and mthsp60 matrix proteins in cells (Marchenko et al., 2000; Dumont et al., 2003). Indeed, in TPA-treated epidermal cells endogenous wild-type p53 translocates to mitochondria within 10 min (preceding nuclear accumulation) where it sequesters the major oxidative defense enzyme manganese superoxide dismutase, located in the mitochondrial matrix, which leads to mitochondrial dysfunction and apoptosis (Zhao et al., 2005). Moreover, p53 induces mitochondrial reactive oxygen species generation, which decreases the mitochondrial membrane potential Delta psi, thereby promoting apoptosis (Li et al., 1999). Consistent with this interpretation, p53CTB/CTM target exclusively to the OMM (Mihara et al., 2003; Talos et al., 2005), whereas Lp53wt, although mainly at the OMM, is partly imported into the matrix, indicated by cleavage of the leader by endopeptidases (Marchenko et al., 2000). Oncogene

One of the many advantages of this competitive Emmyc lymphoma model is its quantitative nature, despite the fact that the data derive from independently arising tumors. Compared to the in vivo suppressor activities of p53CTM and p53CTB in p53-null lymphomas we reported earlier (0.4 and 0%, respectively, at 40:60 ratio) (Talos et al., 2005), these constructs were somewhat weaker in the p53 mutant lymphomas reported here (7.8 and 8.0%, respectively). What could be the reason? As already seen in mutant p53 breast cancer cell lines (Mihara et al., 2003), the mutant p53 lymphoma cells used here have constitutively very high total cellular as well as mitochondrial levels of mutant p53, as is typical of all mutant p53 tumor cells (not shown). Although mutant p53 is unable to bind and sequester BclXL (Mihara et al., 2003), it is conceivable that it nevertheless acts as an overwhelming dominant-negative sink for targeted wild type p53, mopping it up into inactive mixed tetramers and completely disabling its interaction with Bcl2 members. However, as seen in Figure 3, this is not the case. In conclusion, this work is a proof-of-principle that demonstrates that the highly elevated levels of tumorexpressed mutant p53 protein cannot effectively block mitochondrially directed wild-type p53 from inducing tumor cell killing in vivo. Although hoped for, this fact needed to be formally demonstrated. Moreover, the wild-type p53 based therapeutic approach relies on the preserved ability of tumor cells to respond with transcriptional activation of p53 target genes. However, many human cancers have lost this pre-requisite owing to global epigenetic deregulation of their genome, leading to broadly aberrant gene silencing patterns (Herman and Baylin 2003; Baylin and Ohm 2006). In conventional p53-based gene therapy (Peng 2005; Lim et al., 2006; Roth 2006; Sauthoff et al., 2006; Yin and Goodrich 2006), there is no control whether the target tumor cells respond with cell cycle arrest rather than apoptosis, in particular if tumor cells express p21Waf1 expression, which is usually the case. In contrast, targeted p53-based gene therapy takes a direct and – importantly – the shortest known death signaling pathway by p53. The present study significantly takes forward the notion that the direct mitochondrial p53 program, which executes the shortest known circuitry of p53 death signaling, might be exploited for inducing apoptosis in mutant p53 tumor cells in vivo. It strengthens the rationale for further studies that investigate its possible therapeutic value, perhaps as a synergistic modality with existing p53based strategies or conventional anticancer therapy. Moreover, it describes research tools to gain deeper insights into the pleiotropic mechanisms of p53mediated apoptosis. Acknowledgements We thank Patricio Mena for animal care and Christine Eischen for providing two lymphoma isolates. This work was supported by grants from the National Cancer Institute and Philip Morris USA Inc./Philip Morris International to UMM.


6139 References Arima Y, Nitta M, Kuninaka S, Zhang D, Fujiwara T, Taya Y et al. (2005). J Biol Chem 280: 19166–19176. Baylin SB, Ohm JE. (2006). Nat Rev Cancer 6: 107–116. Chen L, Willis SN, Wei A, Smith BJ, Fletcher JI, Hinds MG et al. (2005). Mol Cell 17: 393–403. Chipuk JE, Bouchier-Hayes L, Kuwana T, Newmeyer DD, Green DR. (2005). Science 309: 1732–1735. Chipuk JE, Kuwana T, Bouchier-Hayes L, Droin NM, Newmeyer DD, Schuler M et al. (2004). Science 303: 1010–1014. Dumont P, Leu JI, Della Pietra III AC, George DL, Murphy M. (2003). Nat Genet 33: 357–365. Eischen CM, Weber JD, Roussel MF, Sherr CJ, Cleveland JL. (1999). Genes Dev 13: 2658–2669. Erster S, Mihara M, Kim RH, Petrenko O, Moll UM. (2004). Mol Cell Biol 24: 6728–6741. Erster S, Moll UM. (2005). Biochem Biophys Res Commun 331: 843–850. Essmann F, Pohlmann S, Gillissen B, Daniel PT, Schulze-Osthoff K, Janicke RU. (2005). J Biol Chem 280: 37169–37177. Harris AW, Pinkert CA, Crawford M, Langdon WY, Brinter RL, Adams JM. (1988). J Exp Med 167: 353–371. Herman JG, Baylin SB. (2003). N Engl J Med 349: 2042–2054. Hsu YT, Wolter KG, Youle RJ. (1997). Proc Natl Acad Sci USA 94: 3668–3672. Kuwana T, Bouchier-Hayes L, Chipuk JE, Bonzon C, Sullivan BA, Green DR et al. (2005). Mol Cell 17: 525–535.

Leu JI, Dumont P, Hafey M, Murphy ME, George DL. (2004). Nat Cell Biol 6: 443–450. Li PF, Dietz R, von Harsdorf R. (1999). EMBO J 18: 6027–6036. Lim DS, Bae SM, Kwak SY, Park EK, Kim JK, Han SJ et al. (2006). Hum Gene Ther 17: 347–352. Marchenko ND, Zaika A, Moll UM. (2000). J Biol Chem 275: 16202–16212. Mihara M, Erster S, Zaika A, Petrenko O, Chittenden T, Pancoska P et al. (2003). Mol Cell 11: 577–590. Nemajerova A, Wolff S, Petrenko O, Moll UM. (2005). FEBS Lett 579: 6079–6083. Olivier M, Eeles R, Hollstein M, Khan MA, Harris CC, Hainaut P. (2002). Hum Mutat 19: 607–614. Peng Z. (2005). Hum Gene Ther 16: 1016–1027. Petros AM, Gunasekera A, Xu N, Olejniczak ET, Fesik SW. (2004). FEBS Lett 28073: 1–4. Roth JA. (2006). Expert Opin Biol Ther 6: 55–61. Sansome C, Zaika A, Marchenko ND, Moll UM. (2001). FEBS Lett 488: 110–115. Sauthoff H, Pipiya T, Chen S, Heitner S, Cheng J, Huang YQ et al. (2006). Cancer Gene Ther doi: 10.1038/sj.cgt.7700936. Schmitt CA, Rosenthal CT, Lowe SW. (2000). Nat Med 6: 1029–1035. Talos F, Petrenko O, Mena P, Moll UM. (2005). Cancer Res 65: 9971–9981. Yin S, Goodrich DW. (2006). Int J Oncol 28: 781–785. Zhao Y, Chaiswing L, Velez JM, Batinic-Haberle I, Colburn NH, Oberley TD et al. (2005). Cancer Res 65: 3745–3750.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Oncogene