The pathological response to DNA damage does not
contribute to p53-mediated tumour suppression
M. A. Christophorou1, I. Ringshausen1, A. J. Finch1, L. Brown Swigart1& G. I. Evan1
The p53 protein has a highly evolutionarily conserved role in
metazoans as ‘guardian of the genome’, mediating cell-cycle arrest
animals with substantial somatic regenerative capacity, such as
vertebrates, p53 is an important tumour suppressor—an attribute
thought to stem directly from its induction of death or arrest in
mutant cells with damaged or unstable genomes. Chemotherapy
and radiation exposure both induce widespread p53-dependent
DNAdamage. This triggers potentially lethal pathologies2that are
generally deemed an unfortunate but unavoidable consequence of
the role p53 has in tumour suppression. Here we show, using a
mouse model in which p53 status can be reversibly switched in
vivo between functional and inactive states3, that the p53-
mediated pathological response to whole-body irradiation, a
prototypical genotoxic carcinogen, is irrelevant for suppression
of radiation-induced lymphoma. In contrast, delaying the restor-
ation of p53 function until the acute radiation response has
subsided abrogates all of the radiation-induced pathology yet
preserves much of the protection from lymphoma. Such protec-
tion is absolutely dependent on p19ARF—a tumour suppressor
induced not by DNA damage, but by oncogenic disruption of the
p53-deficient mice show a marked resistance to apoptosis of
‘radiosensitive’ tissues following systemic genotoxic injury4but
have a greatly increased risk of cancer5,6. In the homozygous
p53ERTAMknockin (p53KI/KI) mouse, each copy of the endogenous
which is completely dependent on the ligand 4-hydroxytamoxifen
(4-OHT)forallmeasurable p53functions.Without 4-OHT,p53KI/KI
mice are functionally p53-deficient but are rapidly and reversibly
toggled to p53 wild-type status following systemic 4-OHT adminis-
tration3. It is important to note that the 4-OHT ligand does not, in
itself, activate p53ERTAMbut, rather, renders it competent to become
activated by appropriate upstream signals. We used p53KI/KImice
to explore the cause-and-effect relationship between the acute
pathological response to systemic genotoxic injury and suppression
of the consequent tumours.
Gamma (g)-radiation is a prototypical carcinogen that acts by
directly inducing DNA damage, principally single- and double-
strand breaks. Exposure of C57BL/6 mice with a wild-type p53
genotype (p53þ/þ) to 2.5-Gy g-radiation induces widespread p53-
dependent cell death in radiosensitive tissues, but the mice survive
and are significantly protected from radiation-induced cancer. In
contrast, irradiation of p53-deficient mice (p532/2) induces negli-
gible pathology but 100% of animals are dead from lymphoma by24
weeks (ref.7). Toascertainwhen p53 function isrequired tosuppress
radiation-induced lymphoma, p53KI/KImice were treated with
4-OHT for six days to reinstate p53 function and then exposed to
2.5-Gy whole-body g-radiation (the “concurrent p53 restoration”
cohort). A second, control cohort was treated with vehicle control
instead of 4-OHT (the “no p53 restoration” cohort). Irradiation at
2.5Gy was chosen after careful titration as a radiation dose that
induces a p53-dependent, but not a p53-independent, apoptotic
response in radiosensitive tissues3. As shown previously, irradiation
of p53-competent mice3, which precipitates widespread apoptosis of
primary lymphoid organs and intestinal epithelium3and sustained
Figure 1 | The p53-dependent response to acute DNA damage induces
multiple pathologies but no tumour suppression. a, TUNEL (TdT-
p53KI/KIanimals exposed to 2.5Gy of whole-body g-radiation in the
presence of 4-OHTor vehicle (oil) control. p53þ/þand p532/2control
animals treated with 4-OHTshow that radiation-induced apoptosis is
dependent on p53 function and not 4-OHT. Bone marrow was collected 4h
after irradiation. TUNEL-positive apoptotic cells are shown in green, with
Hoechst nuclear counterstain in red. Original magnification, £20.
b, Leukocyte counts of triplicate p53KI/KIanimals exposed to 2.5Gy of
g-radiation in the presence of 4-OHTor vehicle control. Animals were
irradiated on day 1. Error bars represent ^1s.d. c, Kaplan–Meier survival
curves of p53KI/KImice in ‘concurrent restoration’ and ‘no restoration’
cohorts. Mice were treated with a single dose of 2.5Gy of g-radiation at 5–6
weeks of age.
1Cancer Research Institute and Department of Cellular & Molecular Pharmacology, Comprehensive Cancer Center, University of California, San Francisco, California 94143, USA.
Vol 443|14 September 2006|doi:10.1038/nature05077
© 2006 Nature Publishing Group
leukocytopenia (Fig. 1a, b). In contrast, control vehicle-treated
p53KI/KImice show negligible lymphoid or intestinal apoptosis3,
only transient leukocytopenia (Fig. 1a, b), and no long-term tissue
pathology, consistent with their uninterrupted p53-deficient status8.
Thus, at this radiation dose functional p53 mediates all measurable
radiation-induced pathologies in p53KI/KImice. As whole-body
irradiation is a potent carcinogen, both cohorts of mice were then
monitored for tumour incidence. Despite their dramatic pathologi-
cal response to acute irradiation, p53KI/KImice in which p53 was
functional concurrently with irradiation enjoyed no protection from
function had never been restored (Fig. 1c).
Following acute, single-dose radiation exposure, global DNA
damage in tissues is rapidly resolved, as indicated by the disappear-
ance of DNA-damage-related markers such as phosphorylated
gH2AX (Supplementary Fig. 1) and phosphorylated ATM (data
not shown). p53-activating signals decay with an identical kinetic
and are undetectable after 48–72h. The decay rate is identical in
p53KI/KIand p53þ/þanimals (ref. 3 and Supplementary Fig. 2),
indicating that p53ERTAMis as effective as wild-type p53 in resolving
acute DNA damage. Consistent with this rapid decay, restoration of
p53 function eight days after irradiation induces no detectable
apoptosis in radiosensitive tissues (Fig. 2) nor any pathology. To
test whether restoration of p53 competence at this delayed time
suppresses the emergence of radiation-induced lymphoma, p53KI/KI
mice were exposed to 2.5-Gy g-irradiation in the absence of 4-OHT,
and p53 function was then restored eight days later and maintained
for six days. Thus, the total period of p53 function in this “delayed
restoration” cohort was identical to that in the “concurrent restor-
ation” cohort. A schematic of these different regimens is shown in
Fig. 3a, and Kaplan–Meier survival curves for the “no p53 restor-
ation” versus the “delayed p53 restoration” cohorts in Fig. 3b.
Delayed restoration of p53 offered significant protection
(P ¼ 0.0007) from radiation-induced lymphomagenesis, with a
median delay in death of 99 days (see also Supplementary Fig. 3).
Together, these data demonstrate that—despite culling a large
number of cells within the DNA-damaged populations from which
subsequent lymphomas arise—the acute, p53-dependent pathology
induced by whole-body genotoxic injury has a negligible role in the
suppression of the eventual tumours induced by that insult. In
contrast, delaying the restoration of p53 until after the global,
acute DNA-damage response has subsided abrogates all radiation-
induced pathology yet remains markedly tumour suppressive. Of
note, only a brief period of transient p53 functional restoration is
sufficient for substantial (,100-day) delay in tumour onset.
The effectiveness of delayed p53 restoration in suppressing
lymphomagenesis indicates that long after the p53-activating
DNA-damage response has decayed in the bulk of cells, persistent
p53-activating signals must remain in the small number of progeni-
tor lymphoma cells engendered by the initial radiation insult.
Figure 2 | Delaying restoration of p53 function abrogates radiation
pathology. a–c, TUNEL assay of apoptosis in paraffin-embedded sections
of thymus (a), spleen (b) and small intestinal epithelium (c) from p53KI/KI
animals exposed to 2.5Gy of g-radiation at 5–6 weeks of age. p53 function
was restored by systemic 4-OHTadministration either 6h or 8days after
4-OHT. TUNEL-positive apoptotic cells are shown in green, with Hoechst
nuclear counterstain in red. Original magnification, £2.5.
Figure 3 | Delaying restoration of p53 function until after the response to
acute DNA damage has attenuated elicits marked tumour suppression.
a, Schematic of experimental design for tumorigenesis experiments.
b, Kaplan–Meier analysis for the survival of p53KI/KImice in the ‘delayed
restoration’ cohort compared with the ‘no restoration’ cohort. Mice were
treated with a single dose of 2.5Gy of g-radiation at 5–6 weeks of age.
Log-rank analysis determined that the ‘delayed restoration’ cohort shows a
cohort (P ¼ 0.0007).
© 2006 Nature Publishing Group
NATURE|Vol 443|14 September 2006
Recently, it was reported that early human cancerous lesions show
evidence of persistent DNA damage9,10, possibly as a result of
telomere erosion or aberrant replication structures triggered by
uncontrolled proliferation. Although telomere erosion is not exten-
sive in inbred mouse tissues or tumours11,12, and so cannot be the
trigger for the delayed p53 activation we see, it remains possible
that radiation injury generates unresolvable DNA damage that
persistently signals to p53. Alternatively, p53 is also potently13–17
and persistently3activated by oncogenic mutations that disrupt
cell-cycle control (for example, activation of Myc, Ras (Hras1) or
E2F, or loss of Rb1), via induction of the tumour suppressor p19ARF
(refs 18–23). Notably, p19ARFis not induced by DNA damage, and
activates p53 directly, in the absence of any ancillary DNAdamage or
stress signals24. As oncogene activation is likely to be an obligate
attribute of incipient lymphoma cells, p19ARFis a plausible trigger of
p53 activity in such cells. Therefore, we tested directly whether
p19ARFis required for the tumour surveillance mediated by delayed
p53 restoration. p19ARF-null, p53KI/KI(p53KI/KI;ARF2/2) mice were
irradiated as described above, and eight days later p53 function was
reinstated for six days; the mice were then monitored for lymphoma.
p19ARFis the crucial conduit of p53-mediated tumour suppression in
incipient cancer cells, and also indicates that any persistent DNA
damage signals—if they exist—are insufficient to activate p53 in the
absence of p19ARF. To confirm that p19ARFstatus exerts no significant
impact on DNA-damage responses, as published24, p53 function
was restored in p53KI/KI;ARF2/2and p53KI/KI;ARFþ/þmice and
the animals exposed to 2.5Gy of g-irradiation. Radiation induced
Bbc3)) expression and cellular apoptosis in all radiosensitive tissues of
observed DNA-damage responses were p53-dependent. The pivotal
in all of the p53KI/KIcohorts. p53ERTAMis spontaneously activated on
administration of 4-OHT to tumours that express p19ARFbut not
tumour cells, despite all of the additional lesions and potential
genomic damage they may have accumulated, we see no evidence
for any p53-activating signals other than p19ARF.
In summary, our observations illustrate two key points. First, the
acute pathological response to systemic genotoxic injury is irrelevant
to suppression of eventual tumours induced by that injury. Second,
restoring p53 function after the acute DNA-damage response has
subsided engages substantial tumour suppression that is dependent
on the p19ARFpathway—a pathway activated not by DNA damage
but by aberrant cell proliferation25. These data raise the possibility
that p53 sits at the unhappy confluence of two signals generated by
DNA damage. Directly, DNA damage triggers widespread and
associated with systemic genotoxic injury, including extensive apop-
tosis of radiosensitive tissues. However, this widespread DNA
damage contributes negligibly to tumour suppression. Indirectly,
that same DNA damage triggers rare oncogenic mutations in a few
surviving or misrepaired cells that, as a consequence, induce p19ARF
and elicit sustained p53 activation. It is this non-pathological,
indirect pathway that seems to be crucial for the suppression of
DNA-damage-induced lymphomagenesis (Supplementary Fig. 5).
Recently, several studies have called into question the widely held
view that the DNA-damage response is integral to the actions of p53
as a tumour suppressor. Thus, replacement of endogenous p53 by a
mutant that cannot be phosphorylated at the key Ser18 residue (the
mouse homologue of human p53 Ser15) by DNA-damage-transdu-
cing kinases (Atm, Atror Chk2) generates micethat areincompetent
for DNA-damage-induced apoptosis yet fully protected from can-
cers26. Moreover, absence of Atm has no impact on the ability of p53
to suppress tumours that are caused by cell-cycle-deregulating
oncogenes in a transgenic brain tumour model27. Notably, these
investigations and our own are all in vivo studies, which skirt
the complication of stress-induced p53 activation that confounds
analyses of cells in vitro. It is also noteworthy that the close p53
relative p73 is potently activated by DNAdamage and shares most of
the target genes and pro-apoptotic properties of p5328. Unlike p53,
however, p73 is not activated by p19ARFand is an unconvincing
The possibility that the acute DNA damage response may be
dispensable for p53-mediated tumour suppression has intriguing
implications. First, although DNAdamage is very efficient at trigger-
ing p53-dependent cell death, as evident from the widespread
apoptosis of radiosensitive tissues following irradiation, it is highly
inefficient at generating tumour cells: mice lacking p53 are immune
to much of the pathological cell death induced by irradiationyet still
take months to develop rare, clonal tumours. Thus, only a tiny
number of damaged cells that p53 kills would ever have evolved into
tumours had they survived. Widespread activation of p53 following
Figure 4 | p19ARFmediates p53-dependent suppression of radiation-
induced lymphoma but not p53-dependent DNA-damage responses.
a, Kaplan–Meier survival plots of p53KI/KI;ARF2/2mice exposed to a single
dose of 2.5Gy whole-body g-radiation at 5–6 weeks of age in the absence of
p53 function (‘no restoration’) or the presence of p53 function but after the
acute response to DNA damage has ceased (‘delayed restoration’).
b, Real-time quantitative polymerase chain reaction (PCR) analysis of total
thymic RNA from p53KI/KI;ARFþ/þand p53KI/KI;ARF2/2mice exposed to
2.5Gy of g-radiation in the presence of 4-OHTor vehicle control. Thymic
shown relative to that of b-glucuronidase. Error bars represent ^1s.d. for
triplicate mice for each condition and genotype. c, TUNEL staining for
apoptosis in radiosensitive tissues of p53KI/KI;ARF2/2mice exposed to
2.5Gy of g-radiation in the presence of 4-OHTor vehicle control.
p53KI/KI;ARFþ/þanimals are included as controls. Tissues were harvested
5h after irradiation. TUNEL-positive apoptotic cells are shown in green,
with Hoechst nuclear counterstain in red. Original magnification, £20.
NATURE|Vol 443|14 September 2006
© 2006 Nature Publishing Group
DNAdamageistherefore,atbest,anunwieldytoolwithwhichtocull Download full-text
a few potential tumour cells. In contrast, because p19ARFis induced
only in those rare cells that, as a consequence of DNA damage,
acquire oncogenic mutations25, its expression is highly specific to
those few cells set on a neoplastic trajectory. Second, the p53-
dependent DNA-damage response mediates much of the life-threa-
tening pathology and side effects that accompany radiation exposure
and chemotherapy. Restoring p53 function only at a later time
abrogates all of the pathology yet, by focusing p53 activity only on
tumour suppression. Such observations suggest that transient phar-
macological inhibition of p53 during, or shortly after, acute geno-
toxic injury may be helpful in ameliorating the pathology, as
suggested2, without compromising subsequent tumour suppression.
Perhaps, most perversely, it is even possible that the linkage between
p53 and DNAdamage actuallydrives erosion of the efficacy of p53 as
a tumour suppressor by imposing a lifelong selective pressure to
inactivate p53 in cells sustaining such damage. Once lost, p53 is then
unavailable to act as a tumour suppressor. Clearly, the extent towhich
the various attributes of p53—good and bad—can be separated by
adroit manipulation is an area worthy of further investigation.
Mouse breeding and handling. p53KI/KImice were bred and genotyped as
described previously3. p53 function was restored by daily intraperitoneal injec-
tion of Tamoxifen in peanut oil vehicle. Where appropriate, 5–6-week-old mice
were exposed to 2.5-Gy ionizing radiation from a137Cs source. Irradiated mice
were monitored thrice-weekly for general health, well-being and tumour
harvested and fixed in zinc-buffered formalin, paraffin embedded, sectioned,
and stained with either haematoxylin and eosin or appropriate antibodies for
classification. Blood was collected from the orbital sinus or plexus and haema-
tology profiles were determined immediately using a Hemavet 850 machine
Real-timequantitative PCR analysis.RNAwas extractedwith a QiagenRNeasy
Mini kit. Assay of RNA concentration and quality, DNase treatments, reverse
transcription, and quantification were as described previously3.
Immunoblot analysis. Immunoblot analysis was as described previously3.
p19ARFproteinwas detectedwith rat monoclonal anti-p19ARFprimary antibody
(clone C3, Novus) and b-actin with a mouse monoclonal anti-b-actin antibody
(clone AC-15, Sigma).
Received 1 May; accepted 13 July 2006.
Published online 6 September 2006.
Lane, D. P. p53, guardian of the genome. Nature 358, 15– -16 (1992).
Gudkov, A. V. & Komarova, E. A. The role of p53 in determining sensitivity to
radiotherapy. Nature Rev. Cancer 3, 117– -129 (2003).
Christophorou, M. A. et al. Temporal dissection of p53 function in vitro and in
vivo. Nature Genet. 37, 718– -726 (2005).
Lain, S. & Lane, D. Improving cancer therapy by non-genotoxic activation of
p53. Eur. J. Cancer 39, 1053– -1060 (2003).
Jacks, T. et al. Tumor spectrum analysis in p53-mutant mice. Curr. Biol. 4, 1– -7
Donehower, L. A. et al. Mice deficient for p53 are developmentally normal but
susceptible to spontaneous tumours. Nature 356, 215– -221 (1992).
Kemp, C. J., Wheldon, T. & Balmain, A. p53-deficient mice are extremely
susceptible to radiation-induced tumorigenesis. Nature Genet. 8, 66– -69 (1994).
Komarova, E. A. & Gudkov, A. V. Could p53 be a target for therapeutic
suppression? Semin. Cancer Biol. 8, 389– -400 (1998).
Gorgoulis, V. G. et al. Activation of the DNA damage checkpoint and genomic
instability in human precancerous lesions. Nature 434, 907– -913 (2005).
10. Bartkova, J. et al. DNA damage response as a candidate anti-cancer barrier in
early human tumorigenesis. Nature 434, 864– -870 (2005).
11. Prowse, K. R. & Greider, C. W. Developmental and tissue-specific regulation of
mouse telomerase and telomere length. Proc. Natl Acad. Sci. USA 92,
4818– -4822 (1995).
12. Kipling, D. & Cooke, H. J. Hypervariable ultra-long telomeres in mice. Nature
347, 400– -402 (1990).
13. Wu, X. & Levine, A. J. p53 and E2F-1 cooperate to mediate apoptosis. Proc. Natl
Acad. Sci. USA 91, 3602– -3606 (1994).
14. Wagner, A. J., Kokontis, J. M. & Hay, N. Myc-mediated apoptosis requires
wild-type p53 in a manner independent of cell cycle arrest and the ability of
p53 to induce p21waf1/cip1. Genes Dev. 8, 2817– -2830 (1994).
15. Serrano, M., Lin, A. W., McCurrach, M. E., Beach, D. & Lowe, S. W. Oncogenic
ras provokes premature cell senescence associated with accumulation of p53
and p16INK4a. Cell 88, 593– -602 (1997).
16. Lowe, S. W., Cepero, E. & Evan, G. Intrinsic tumour suppression. Nature 432,
307– -315 (2004).
17. Hermeking, H. & Eick, D. Mediation of c-Myc-induced apoptosis by p53.
Science 265, 2091– -2093 (1994).
18. Zindy, F. et al. Myc signaling via the ARF tumor suppressor regulates p53-
dependent apoptosis and immortalization. Genes Dev. 12, 2424– -2433 (1998).
19. Palmero, I., Pantoja, C. & Serrano, M. p19ARFlinks the tumour suppressor p53
to Ras. Nature 395, 125– -126 (1998).
20. de Stanchina, E. et al. E1A signaling to p53 involves the p19ARFtumor
suppressor. Genes Dev. 12, 2434– -2442 (1998).
21. Bates, S. et al. p14ARFlinks the tumour suppressors RB and p53. Nature 395,
124– -125 (1998).
22. Quelle, D. E., Zindy, F., Ashmun, R. A. & Sherr, C. J. Alternative reading frames
of the INK4a tumor suppressor gene encode two unrelated proteins capable of
inducing cell cycle arrest. Cell 83, 993– -1000 (1995).
23. Mao, L. et al. A novel p16INK4A transcript. Cancer Res. 55, 2995– -2997 (1995).
24. Kamijo, T. et al. Tumor suppression at the mouse INK4a locus mediated by the
alternative reading frame product p19ARF. Cell 91, 649– -659 (1997).
25. Zindy, F. et al. Arf tumor suppressor promoter monitors latent oncogenic
signals in vivo. Proc. Natl Acad. Sci. USA 100, 15930– -15935 (2003).
26. Sluss, H. K., Armata, H., Gallant, J. & Jones, S. N. Phosphorylation of serine 18
regulates distinct p53 functions in mice. Mol. Cell. Biol. 24, 976– -984 (2004).
27. Liao, M. J., Yin, C., Barlow, C., Wynshaw-Boris, A. & van Dyke, T. Atm is
dispensable for p53 apoptosis and tumor suppression triggered by cell cycle
dysfunction. Mol. Cell. Biol. 19, 3095– -3102 (1999).
28. Harms, K., Nozell, S. & Chen, X. The common and distinct target genes of the
p53 family transcription factors. Cell. Mol. Life Sci. 61, 822– -842 (2004).
29. Soengas, M. S. & Lowe, S. W. p53 and p73: seeing double? Nature Genet. 26,
391– -392 (2000).
Supplementary Information is linked to the online version of the paper at
Acknowledgements We thank F. Rostker and the Genome and Pathology Cores
at UCSF for technical assistance, J. Hwang for statistical analysis, J. H. Mao,
F. McCormick, C. O’Shea, K. Shannon and the entire Evan laboratory for critical
discussions. We thank T. Littlewood and D. Martin Zanca for showing that the
p53ERTAMknockin mouse was feasible. This work was supported by grants from
the National Cancer Institute and National Institute on Aging.
Author Information Reprints and permissions information is available at
www.nature.com/reprints. The authors declare no competing financial interests.
Correspondence and requests for materials should be addressed to G.I.E.
NATURE|Vol 443|14 September 2006
© 2006 Nature Publishing Group