Content uploaded by Alessia Pacifico
Author content
All content in this area was uploaded by Alessia Pacifico on May 07, 2018
Content may be subject to copyright.
Fate of UVB-induced p53 mutations in SKH-hr1 mouse skin after
discontinuation of irradiation: relationship to skin cancer development
Vladislava O Melnikova
1,4
, Alessia Pacifico
1,4,5
, Sergio Chimenti
2
, Ketty Peris
3
and Honnavara N Ananthaswamy*
,1
1
Department of Immunology, The University of Texas MD Anderson Cancer Center, PO Box 301402, Unit 902, Houston, TX 77030,
USA;
2
Department of Dermatology, The University of Rome ‘tor Vergata’, Rome, Italy;
3
Department of Dermatology, The University
of l’Aquila, L’Aquila 67100, Italy
Chronic exposure to ultraviolet (UV) radiation causes
skin cancer in humans and mice. We have previously
shown that in hairless SKH-hr1 mice, UVB-induced p53
mutations arise very early, well before tumor development.
In this study, we investigated whether discontinuation of
UVB exposure before the onset of skin tumors results in
the disappearance of p53 mutations in the skin of hairless
SKH-hr1 mice. Irradiation of mice at a dose of 2.5 kJ/m
2
three times a week for 8 weeks induced p53 mutations in
the epidermal keratinocytes of 100% of the mice. UVB
irradiation was discontinued after 8 weeks, but p53
mutations at most hotspot codons were still present even
22 weeks later. During that period, the percent of mice
carrying p53
V154A/R155C
,p53
H175H/H176Y
, and p53
R275C
mutant
alleles remained at or near 100%, whereas the percentage
of mice with p53
R270C
mutation decreased by 45%. As
expected, discontinuation of UVB after 8 weeks resulted in
a delay in tumor development. A 100% of tumors carried
p53
V154A/R155C
mutant alleles, 76% carried p53
H175H/H176Y
mutants, and 24 and 19% carried p53
R270C
and p53
R275C
mutants, respectively. These results suggest that different
UVB-induced p53 mutants may provide different survival
advantages to keratinocytes in the absence of further
UVB exposure and that skin cancer development can
be delayed but not prevented by avoidance of further
exposure to UVB radiation.
Oncogene (2005) 24, 7055–7063. doi:10.1038/sj.onc.1208863;
published online 11 July 2005
Keywords: UVB carcinogenesis; p53; tumor progression;
differentiation; skin tumor
Introduction
The incidence of skin cancer exceeds the incidence of all
other human cancers combined (Urbach, 1991; Miller
and Weinstock, 1994). Epidemiologic, clinical, and
biologic studies indicate that solar ultraviolet (UV)
radiation is the major etiologic agent in skin cancer
development (Urbach, 1991, 1997; Brash et al., 1991;
Miller and Weinstock, 1994). Wavelengths in the UVB
range of the solar spectrum (290–320 nm) are absorbed
by the skin, producing erythema, burns, immuno-
suppression, DNA mutations, and nonmelanoma skin
cancers (NMSC) (Gilchrest, 1990; Young, 1990; Brash
et al., 1991; Kripke, 1991; Kanjilal et al., 1993; de Gruijl
and Forbes, 1995; Brash et al., 1996; Ananthaswamy
et al., 1997). Several studies have shown that the p53
tumor suppressor gene is susceptible to UV-induced
mutations and plays a critical role in the induction of
NMSC (Brash et al., 1991, 1996; Kanjilal et al., 1993;
Ziegler et al., 1994; Jonason et al., 1996; Ananthaswamy
et al., 1997). In response to DNA damage, the p53
protein transactivates downstream genes such as p21
Waf1/
Cip1
and gadd45 that induce cell-cycle arrest at the
G
1
–S phase to allow DNA repair (Kuerbitz et al., 1992;
Li et al., 1996). If the damage is not repaired, p53-
dependent apoptosis, or ‘cellular proofreading,’ is
triggered to eliminate severely damaged cells (Ziegler
et al., 1994; Brash, 1996; Smith and Fornace, 1997).
Upon repeated exposure to UV radiation, DNA lesions
in the p53 gene are transformed into mutations, mainly
C-TorCC-TT transitions at dipyrimidine sites,
thereby initiating the molecular process of skin carcino-
genesis (Leffell and Brash, 1996). Thousands of
p53-mutant cell clones are found in sun-exposed skin
that appears normal (Jonason et al., 1996; Ren et al.,
1996). The frequency of UV-signature mutations in
the p53 gene is high in precancerous lesions and reaches
50–90% in human basal and squamous cell carcinomas
(SCCs) and 100% in murine UV-induced skin tumors
(Kanjilal et al., 1993; Ziegler et al., 1994; Brash et al.,
1996).
UV radiation is a complete carcinogen in that it not
only initiates tumorigenesis by inducing mutations in the
p53 tumor suppressor gene but also promotes tumor
development (Epstein and Epstein, 1963; Blum, 1969).
As a tumor promoter, UV radiation induces cell
proliferation by stimulating the production of various
growth factors and cytokines as well as the activation of
their receptors (De-Metys et al., 1995; Rosette and
Received 10 February 2005; revised 6 April 2005; accepted 28 April 2005;
published online 11 July 2005
*Correspondence: HN Ananthaswamy;
E-mail: hanantha@mdanderson.org
4
These authors contributed equally to this work
5
Current address: Department of Dermatology, The University of
l’Aquila, L’Aquila, Italy
Oncogene (2005) 24, 7055–7063
&
2005 Nature Publishing Group
All rights reserved 0950-9232/05 $30.00
www.nature.com/onc
Karin, 1996; Bender et al., 1997; Kuhn et al., 1999; Jost
et al., 2000; Peus et al., 2000; Ullrich, 2000; Walterscheid
et al., 2002). In addition, cell survival and growth
following UV irradiation involves activation of NF-kB,
mitogen-activated protein kinase, and phosphoinositide-
30-kinase/Akt pathways (Coffer et al., 1995; De-Metys
et al., 1995; Rosette and Karin, 1996; Bender et al.,
1997). Repeated exposure of skin to UV radiation
results in the clonal expansion of p53-mutant cells
(Ziegler et al., 1994; Berg et al., 1996; Rebel et al., 2001;
Zhang et al., 2001). Two mechanisms are believed to
contribute to the selective expansion of p53-mutant cells:
their resistance to UV-induced apoptosis and their
proliferative advantage over normal keratinocytes in
response to stimulation with UV radiation. While
evidence supports the role of mutant p53 in enabling
apoptosis resistance in keratinocytes (Ziegler et al.,
1994; Tron et al., 1998; Zhang et al., 2001; Mudgil et al.,
2003), the possibility that mutant p53 provides kerati-
nocyte progenitors with proliferative advantages re-
mains largely unexplored, even though two phenotypes
may be related to the same underlying molecular
changes (Kuhn et al., 1999; van Hogerlinden et al.,
1999; Peus et al., 2000; Marconi et al., 2003). Never-
theless, discontinuation of UV irradiation has been
shown to result in the fast spontaneous regression of
some mutant p53 clones in mouse skin, although the
mechanisms involved in this process are unclear (Berg
et al., 1996; Rebel et al., 2001; Remenyik et al., 2003).
We have previously shown that exposure of SKH-hr1
mice to chronic UV radiation induced SCCs (Ouhtit
et al., 2000a). UV-signature CC-TT and C-T
mutations in p53 codons 154–155, 175–176, and 270 or
275 were detected in epidermis as early as after 1 week of
chronic UV irradiation (Ouhtit et al., 2000a). These
mutations were detected in the epidermis of 80% of mice
irradiated for 4 weeks and 90% of mice irradiated for 8
weeks (Ouhtit et al., 2000a). The mice began to develop
skin tumors after 8 weeks of chronic irradiation, and all
mice developed multiple skin tumors, mostly SCCs, by
week 25 of chronic UV irradiation. These results
suggested that p53 mutations arise very early and well
before skin tumor development. This finding raised an
important question, that is, what would happen to p53
mutations in the skin if UV irradiation is discontinued
after 8 weeks, and how it relates to skin cancer
development? We, therefore, investigated the fate of
UV-induced p53 mutations in mouse skin and their
relationship to skin cancer development after disconti-
nuation of UV exposure. Our results indicate that while
the incidence of some p53 mutant alleles remains at or
near 100% even 22 weeks after UV irradiation is
discontinued, other p53 mutations do get eliminated
from skin, evidently as a result of cell differentiation and
desquamation. Discontinuation of UV treatment de-
layed the time required for tumor development, but it
did not prevent tumors. All tumors harbored p53
mutations at one or more of the hotspots, and the
dynamics of the retention or loss of these mutations in
lesion-free skin after the discontinuation of UV irradia-
tion correlated with their incidence in tumors.
Results
UV-induced p53 mutations in mouse skin after
discontinuation of irradiation
Representative allele-specific polymerase chain reaction
(AS-PCR) gel electrophoresis data shown in Figure 1a
indicate that 1 day after discontinuation of UV
irradiation, p53 mutations were detected in the skin
of all 10 mice in the group studied. Specifically,
p53
V154A/R155C
,p53
H175H/H176Y
,p53
R270C
, and p53
R275C
mutant
alleles were detected in 8, 9, 9, and 10 of the 10 mice,
respectively (Figure 1a, Table 1). DNA from unirra-
diated mouse skin was not amplified by any sets of
mutant-specific primers (lane 2 on each gel). In addition,
DNA from all 10 (five at 1 day post-UV and five at 22
weeks post-UV time points) unirradiated mouse skin did
not amplify at codon 270 (Figure 1d, Table 1) or at any
other codons tested (data not shown). The absence of
p53
R270C
mutations in untreated mouse skin versus
p53
R270C
mutation at 1 day post-UV skin was statistically
significant (P¼0.01). However, genomic DNA from a
UV-induced tumor cell line known to contain p53
mutations at specific codons was amplified (Figure 1a,
positive control ( þcontrol) lane on each gel). Interest-
ingly, p53 mutations at all the hotspot codons were
still present in the epidermis of mice 22 weeks after
Table 1 Incidence of p53 mutations in SKH-hr1 mouse skin and tumors after discontinuation of UV irradiation
Time No. of mice with mutations in a particular codon/no. of mice or tumors analysed
154–155 CC-TT 175–176 CC-TT 270 C-T 275 C-T Total % of mice or tumors with mutations
1 day post-UV 8/10 9/10 9/10 10/10 100
2 weeks post-UV 10/10 9/10 9/10 8/10 100
6 weeks post-UV 10/10 10/10 9/10 8/10 100
10 weeks post-UV 9/10 10/10 8/10 7/10 100
22 weeks post-UV 9/9 9/9 5/9 8/9 100
No UV skin 0/10 0/10 0/10 0/10 0
Tumors
a
21/21 (100%) 16/21 (76%) 5/21 (24%) 4/21 (19%) 100
Groups of 10 tumor-free skin specimens collected on day 1 or weeks 2, 6, 10, and 22 after last UV exposure were analysed for mutations in
UV-hotspot codons 154–155, 175–176, 270, and 275 of p53 gene using AS-PCR method.
a
In all, 21 randomly chosen SCCs were analysed for
p53 mutations
UVB cessation, p53 mutations, and skin cancer
VO Melnikova et al
7056
Oncogene
discontinuation of UV irradiation. Specifically, at 22
weeks, the incidence of p53
V154A/R155C
,p53
H175H/H176Y
, and
p53
R275C
mutant alleles remained at 90–100% of their
incidence 1 day after UV irradiation was discontinued
(Figure 1b, Table 1), whereas the incidence of mutation
at p53 codon 270 decreased by 45% (Figure 1c, Table 1),
which was not statistically significant (P¼0.72). It
should be noted that AS-PCR is a highly sensitive
method that detects mutations regardless of the size or
number of mutant clones. A low amount of mutant p53
alleles will allow a skin sample to be scored as positive.
Consequently, this method does not reflect the actual
disappearance of cells with a mutant p53 until it has
decreased to the limit of AS-PCR sensitivity. We did
not, however, determine the number or size of mutant
p53 clones in our experiments.
Histologic changes in skin after discontinuation of UV
Epidermal hyperplasia To determine whether disconti-
nuation of irradiation affects epidermal hyperplasia
induced by chronic UV, we examined hematoxylin and
eosin (H&E)-stained skin sections of mice at various
time points after discontinuation of UV exposure. As
expected, chronic UV irradiation for 8 weeks induced
epidermal hyperplasia (Figure 2). However, termination
of UV treatment led to a noticeable decrease in
epidermal hyperplasia as early as 7 days later (Figure 2).
At 14 days post-UV irradiation and thereafter, the
epidermis looked quite normal and resembled the skin
of unirradiated mice. Analogous to the gross H&E data,
immunohistochemical analysis also indicated the pre-
sence of numerous (59.577.4 per 100 nucleated cells)
proliferating cell nuclear antigen (PCNA)-positive cells
throughout the epidermis of mouse skin irradiated with
UV light for 8 weeks (Figure 2, 1 day post-UV).
Correspondingly, PCNA-positive cells were fewer in
1 day after
UV discontinuation
p53V154A/R155C
p53R270C
p53R270C
p53H175H/H176Y
p53R275C
p53V154A/R155C
2 weeks
6 weeks
22 weeks
Normal skin
p53R270C
water
no UV
12345678910
+control
water
no UV
12345678910
+control
a
b
c
d
2 weeks
6 weeks
22 weeks
water
no UV
12345678910
+control
water
no UV
12345678910
+control
c
Figure 1 AS-PCR detection of p53 mutations in mouse skin. (a)
At 1 day after discontinuation of chronic UV irradiation (3 per
week for 8 weeks). (b, c) Different time points after UV disconti-
nuation. Lanes marked from 1 to 10 represent genomic DNA from
treated skin samples. Positive control ( þcontrol) is genomic DNA
from four different established mouse skin tumor cell lines
containing p53 mutations at codons of interest. DNA from an
unirradiated mouse skin (lane marked No UV) did not amplify at
any of the codons of interest. (d) DNA from 10 unirradiated mice
(five at 1 day post-UV time point, lanes 1–5 and five at 22 weeks
post-UV time point, lanes 6–10) did not amplify at codon 270 or at
other codons (data not shown)
H&E PCNA
no UV
1 day
post UV
7 days
post UV
14 days
post UV
Figure 2 H&E stains and immunohistochemical analysis for
PCNA protein expression in nonirradiated skin and in skin taken
1, 7, and 14 days after discontinuation of UV treatment
UVB cessation, p53 mutations, and skin cancer
VO Melnikova et al
7057
Oncogene
number (42.4712.6 per 100 cells, Po0.1 by Student’s
t-test) at 7 days post-UV exposure. At day 14 after UV
radiation discontinuation, the number of PCNA-posi-
tive cells was significantly lower when compared to day
1 after UV radiation discontinuation (32.576.8 per 100
cells, Po0.5), and was equal to that seen in unirradiated
skin (32.073.5 per 100 cells, P>0.1). In addition, most
of the PCNA-positive cells were already localized in the
basal layer of the epidermis 14 days after termination of
UV exposure, a pattern seen in unirradiated mouse skin
(Figure 2). Thus, both H&E and PCNA expression data
indicate that after discontinuation of UV irradiation,
epidermal hyperplasia subsides, at least in part, because
of a gradual decrease in the number of proliferating
keratinocytes.
Apoptosis and terminal differentiation To determine
whether the decreased epidermal hyperplasia seen in the
mouse skin after discontinuation of UV irradiation was
mediated by apoptotic death or terminal differentiation
followed by keratinocyte desquamation, we analysed the
skin of mice at various time points after discontinuation
of UV exposure for the presence of TUNEL-positive
cells and expression of keratin 10 and loricrin. TUNEL
assay data indicated the absence of TUNEL-positive
cells in mouse skin irradiated with UV light for 8 weeks.
This was expected because we showed previously that
chronic UV irradiation results in apoptosis resistance
(Ouhtit et al., 2000a). However, TUNEL-positive cells
were also absent in the epidermis of mice at all time
points after UV irradiation discontinuation (data not
shown), suggesting that decreased epidermal hyperplasia
was not due to apoptosis. Further, expression patterns
of the keratinocyte terminal differentiation markers
keratin 10 and loricrin were similar in nonirradiated skin
and in irradiated skin after discontinuation of UV
irradiation. Keratin 10 expression was observed
throughout the epidermis but not in the basal cell layer
(Figure 3). The expression of loricrin was localized in the
upper layer of the epidermis (Figure 3). Denucleated
cells in the upper epidermis expressed the highest levels
of keratin 10 and loricrin, which is quite similar to the
pattern seen in unirradiated epidermis (Figure 3).
Finally, we observed an increase in the thickness of
the stratum corneum layer, which was particularly
evident at weeks 2 (Figures 2 and 3) and 4 (data not
shown) after UV radiation discontinuation. Together
with the PCNA expression data, these results suggest
that epidermal hyperplasia induced by chronic UV
irradiation occurs owing to an increase in the number of
proliferating cells rather than because of a severe
inhibition of terminal differentiation. Discontinuation
of UV exposure causes a decrease in the number of
proliferating cells, a gradual elimination of excess
keratinocytes owing to desquamation, and the
re-establishment of normal epidermal thickness.
Expression of p53 after discontinuation of UV irradiation
We showed previously that UV-irradiated mouse skin
expresses high levels of p53 protein (Ouhtit et al.,
2000a). To determine whether discontinuation of UV
irradiation affects expression of p53 protein, we
conducted the immunohistochemical analysis of mice
skin for p53 expression at various time points after dis-
continuation of UV radiation. Strong nuclear immuno-
reactivity with the CM5 anti-p53 antibody, attributed
to the expression of mutant p53 protein, was observed
in the epidermis of the mice at all time points. In
particular, 1 day after discontinuation of UV irradia-
tion, keratinocytes in both the upper and basal layers of
the epidermis expressed high levels of p53 (Figure 4). In
addition, even though there was a decrease in epidermal
hyperplasia at day 14 after UV discontinuation and
thereafter, p53-positive cells were present in all layers of
epidermis, including basal cells, transient proliferating
Keratin 10 Loricrin
no UV
1 day
post UV
7 days
post UV
14 days
post UV
Figure 3 Immunohistochemical analysis for keratin 10 and
loricrin protein expression in nonirradiated skin and in skin taken
1, 7, and 14 days after discontinuation of UV treatment
no UV 1 day post UV
7 days post UV 14 days post UV
Figure 4 Immunohistochemical analysis for p53 protein expres-
sion in unirradiated skin and in skin taken 1, 7, and 14 days after
discontinuation of UV treatment
UVB cessation, p53 mutations, and skin cancer
VO Melnikova et al
7058
Oncogene
cells, and cells undergoing terminal differentiation
(Figure 4), suggesting that some keratinocytes with
mutant p53 can still undergo terminal differentiation. In
contrast, no p53 expression was observed in unirra-
diated mouse skin (Figure 4).
Development of skin tumors after discontinuation of UV
treatment
To determine whether discontinuation of UV irradiation
after 8 weeks of exposure abrogates or reduces the
incidence of skin tumors, we monitored 50 SKH-hr1
mice for tumor development. Even though UV irradia-
tion was discontinued, all the mice developed skin
tumors; however, the time to 100% tumor incidence was
significantly delayed. A small lesion (o2 mm in dia-
meter) was first observed in a mouse at 2 weeks after
discontinuation of UV irradiation, and by week 52, all
mice developed skin tumors (Figure 5). In contrast, none
of the 10 nonirradiated mice developed skin tumors. The
combined tumorigenicity data for continuous (Ouhtit
et al., 2000a) and discontinuous UV irradiation experi-
ments show differences in the kinetics of tumor
development (Figure 5). It should be noted that the
tumorigenicity data were obtained several years apart as
part of different studies, and a direct comparison is only
shown for reference. Nevertheless, the time needed for
50% of mice to develop tumors under the continuous
irradiation protocol was about 14 weeks, compared to
34 weeks under the UV radiation discontinuation
protocol. In addition, while 100% of mice continually
irradiated with UV light developed skin tumors of at
least 3 mm in diameter by week 24, it took significantly
longer (52 weeks) for tumors to develop in the UV
radiation discontinuation experiment. In both proto-
cols, the mean number of 3-mm or larger tumors was
approximately equal: 3.071.1 and 3.370.4 per mouse
for continuous and discontinuous UV protocols, respec-
tively. Finally, analogous to the continuous UV irradi-
ation protocol, most of the tumors that developed after
UV radiation discontinuation were diagnosed as SCC
(Figure 6). In addition, analysis of 21 SCCs that deve-
loped in mice after discontinuation of UV irradiation
revealed that 100% of the tumors carried p53
V154A/R155C
mutant alleles, 76% carried p53
H175H/H176Y
mutants, and
24 and 19% carried p53
R270C
and p53
R275C
mutants,
respectively (Table 1).
Discussion
We showed previously that UV-induced p53 mutations
arise very early in mice, well before the onset of skin
cancer (Ananthaswamy et al., 1997; Ouhtit et al.,
2000a), which suggests that this event may initiate the
process of multistage carcinogenesis. In hairless mouse
skin, UV-induced p53 mutations could be detected by
AS-PCR as early as 1 week after the first UV radiation
exposure, with 80–90% of animals incurring p53
mutations after 8 weeks of UV treatment (Ouhtit
et al., 2000a). Clones of keratinocytes carrying mutant
p53 have also been detected in mouse skin within 2–3
weeks after UV treatment (Berg et al., 1996; Rebel
et al., 2001; Remenyik et al., 2003). In this study, we
investigated the fate of p53 mutations and the develop-
ment of skin cancer after discontinuation of UV
treatment. We found that discontinuation of UV
irradiation before the onset of skin tumor development
did not drastically decrease the overall frequency of p53
mutations. While p53
V154A/R155C
and p53
H175H/H176Y
muta-
tions persisted in keratinocytes at or near 100% of the
8-week incidence for as long as 22 weeks after UV
radiation discontinuation, the incidence of p53
R270C
mutations decreased by 45% at 22 weeks post-UV
irradiation (Table 1), which was not considered statis-
0 8 16 24 32 40 48 56
0
25
50
75
100
Time (weeks)
No. of mice with tumors (%)
continuous
UV
discontinued
UV
Figure 5 Time course of tumor development in SKH1-hr mice
exposed to continuous UV radiation (open symbols) or disconti-
nuous UV radiation (closed symbols). For continuous UV
treatment, each point represents data from 20 mice; for disconti-
nuous UV treatment, each point represents data from 50 mice.
Lesions >3.0 mm were counted as established tumors in both the
continuous UV irradiation experiment (data reproduced from
Ouhtit et al., 2000a) and the discontinuous UV experiment
Figure 6 H&E staining of a typical SCC, which developed in an
SKH-hr1 mouse after discontinuation of UV treatment
UVB cessation, p53 mutations, and skin cancer
VO Melnikova et al
7059
Oncogene
tically significant (P¼0.72). Previous studies have
shown that the number and size of p53-positive cell
clones in mouse epidermis decrease within 2 weeks after
the discontinuation of chronic UV treatment (Berg
et al., 1996; Remenyik et al., 2003). This acute regression
phase may be followed by a significantly slower second
phase (Berg et al., 1996). Berg et al. (1996) found that
more than 50% of p53-positive patches induced in Skh-
hr1 mouse skin by exposure to 39 kJ/m
2
of total UV
radiation over a period of 30 days disappeared within
the first 2 weeks after discontinuation of UV irradiation.
The rate of disappearance of p53-positive patches was
significantly higher (85%) when the total dose of UV
radiation was decreased to 22 kJ/m
2
(Berg et al., 1996).
However, in both UV regimens, p53-positive clones
were still detectable for as long as 8 weeks after the last
UV radiation exposure (Berg et al., 1996). Similarly,
Remenyik et al. (2003) found that approximately 50%
of p53-positive keratinocyte clones in C57Bl/6 mouse
skin disappeared within 2 weeks after discontinuation of
UV treatment (37 or 47 kJ/m
2
total UV). In our
experiments, there was a slight decrease in the incidence
of p53
R270C
mutations at 22 weeks after UV radiation
discontinuation, but it was not statistically significant.
However, our data do not dispute the existence of the
acute regression phase because the highly sensitive
AS-PCR method we used to detect mutations reveals
the presence of p53 mutations regardless of the size
or number of mutant clones. Thus, it is possible that
even though p53 mutations were present in 100% of
UV-irradiated mice at 22 weeks after discontinuation of
UV irradiation, the number of p53 mutant clones or
colonies may have decreased during the same period.
We did not, however, determine the number or size of
mutant p53 clones in our experiments.
Remenyik et al. (2003) demonstrated that regression
of precancerous p53-positive clones occurs owing to
mechanisms other than antigen-specific immunity, since
it proceeds with similar kinetics in the skin of immune-
deficient Rag1
/
mice and their wild-type counterparts.
In addition, they found that precancerous p53-positive
cells have normal morphology and that patches of these
cells are not infiltrated with lymphocytes or any other
immune cells (Remenyik et al., 2003). Two other
possible mechanisms for elimination of initiated kerati-
nocytes are apoptotic death and normal cell turnover.
Our histologic and immunohistochemical studies indi-
cate that chronic UV irradiation for 8 weeks induced
epidermal hyperplasia and that discontinuation of UV
irradiation resulted in a decrease in epidermal hyper-
plasia. This decrease in epidermal hyperplasia after UV
irradiation discontinuation could have occurred through
keratinocyte differentiation and desquamation rather
than apoptotic elimination. This conclusion is supported
by the fact that (1) the stratum corneum increased in
thickness during the first 14 days after discontinuation
of UV irradiation (Figures 2 and 3); (2) TUNEL-
positive keratinocytes were absent after UV irradiation
discontinuation; and (3) some p53-positive terminally
differentiated keratinocytes were localized in the upper
layers of epidermis (Figure 4).
Despite discontinuation of UV irradiation, all mice
developed skin tumors, but the kinetics of tumor
development in the current study were quite different
from those in the continuous UV irradiation experi-
ment. For example, the time required for 50% of mice to
contract tumors (t
50
) was 34 weeks in the present
discontinuous UV protocol versus 14 weeks in the
continuous UV study. Secondly, the time required for
tumor development in 100% of mice increased from 24
weeks in the continuous UV irradiation experiment to
52 weeks in the UV discontinuation experiment.
Analogous to this finding, de Gruijl and van der Leun
(1991) showed that limiting the duration of UV
treatment from continuous to 35 (53 kJ/m
2
total UV
dose) or 19 days (29 kJ/m
2
total UV dose) delayed t
50
from 20 weeks to 40 and 92 weeks, respectively. de
Gruijl and van der Leun (1991) have further adapted a
mathematical model that relates tumor occurrence to
the daily dose of UV radiation and the time needed for
animals to develop tumors, thus separating UV- and
time-dependent phases of skin carcinogenesis. The
60-kJ/m
2
total UV dose administered to mice over a
period of 8 weeks in our experiments yielded a t
50
value
of 34 weeks, which correlates with the results of de
Gruijl and van der Leun (1991).
There is evidence to indicate that acute UV exposure
induces the expansion of p53-mutant cells by stimulating
their proliferation while inducing apoptosis in normal
keratinocytes (Ziegler et al., 1994; Zhang et al., 2001;
Mudgil et al., 2003). However, under continuous UV
irradiation, the selective proliferative advantage of
mutant cells over normal cells could contribute to the
clonal expansion of p53-mutant cells because continuous
UV treatment quickly induces apoptosis resistance and
stimulates hyperproliferation as an adaptive response
(Ouhtit et al., 2000a). The selective retention of some
mutant p53 alleles seen in our experiments suggests that
when a constant supply of UV-induced proliferative
stimuli is interrupted by UV irradiation discontinuation,
a population of keratinocytes containing specific p53
mutations still undergoes terminal differentiation, while
other keratinocytes with different p53 mutations survive
and proliferate. This proliferative advantage may be due
to the increased expression of certain growth factors
or growth-related cytokines or their cognate receptors.
Interestingly, different human tumor-associated p53
mutants have been shown to exhibit oncogenic features,
such as promoting cell growth and tumorigenicity, when
compared to p53-null phenotypes (Dittmer et al., 1993;
Sigal and Rotter, 2000). Some p53 mutants have also
been shown to transactivate promoters of epidermal
growth factor receptor and basic fibroblast growth
factor genes and potentiate the expression of vascular
endothelial growth factor (Kieser et al., 1994; Ueba
et al., 1994; Ludes-Meyers et al., 1996). The suggestion
that hyperproliferative and apoptosis-resistant pheno-
types may both result from mutant p53’s ‘gain of
function’ is supported by the finding that resistance to
drug-induced apoptosis in the presence of various
tumor-associated p53 mutants relies on their transcrip-
tional capacity and on potentiation of the c-myc
UVB cessation, p53 mutations, and skin cancer
VO Melnikova et al
7060
Oncogene
expression (Matas et al., 2001). Although it remains to
be proven whether UV-induced p53 mutant proteins
produce similar ‘gain of function’ phenotypes, our
previous studies have shown that mutant p53 protein
is localized in the cell nucleus and phosphorylated at
critical N- and C-terminal residues in all the UV-
induced mouse skin tumors (Melnikova et al., 2003),
potentially fulfilling some of the requirements for being
transcriptionally active ‘gain of function’ mutants (Lin
et al., 1995; Lanyi et al., 1998; Matas et al., 2001).
Even though skin tumors are clonal in origin, most of
them harbor multiple mutations in the p53 gene
(Kanjilal et al., 1993). The presence of multiple p53
mutations has been reported in human and UV-induced
mouse skin cancers and head and neck cancers (Chung
et al., 1993; Kanjilal et al., 1993, 1995). In addition, we
demonstrated previously that UV-irradiated mouse skin
also harbors multiple p53 mutations (Ananthaswamy
et al., 1997; Ouhtit et al., 2000a). It is possible that a cell
containing one p53 mutation acquire secondary and
tertiary mutations in the same or different allele due to
repeated exposure to UVB. Nonetheless, the data
presented herein suggest that different p53 mutants
confer different degrees of survival and/or proliferative
advantage to keratinocyte progenitors in the absence of
further UV exposure. Supporting this assumption, the
retention of p53 mutations in the skin after disconti-
nuation of UV irradiation correlated with their high
incidence in skin tumors (Table 1). Conversely, p53
R275C
mutants were detected in eight of nine (90%) skin
specimens at 22 weeks after UV radiation discontinua-
tion, but only four of 21 (19%) tumors carried p53
R275C
mutant alleles, which was considered statistically sig-
nificant (P¼0.04). However, the approximate mathe-
matical rate constant for the process of mutation
disappearance is similarly high for p53
R275C
and
p53
R270C
mutants, suggesting that a decrease in p53
R275C
incidence in lesion-free skin could be observed if
monitored for a longer period of time.
In summary, our results indicate that mice exposed to
chronic UV radiation retain a population of epidermal
keratinocytes containing p53-mutant alleles long after
discontinuation of UV treatment. Correlation between
the dynamics of retention of mutant p53 alleles in the
skin after discontinuation of UV radiation and their
frequency in tumors suggests that different UV-induced
p53 mutants may provide keratinocyte progenitors with
different degrees of survival and/or proliferation ad-
vantages in the absence of further UV exposure. Finally,
our results show that despite discontinuation after
8 weeks, UV irradiation results in 100% skin tumor
incidence, although the kinetics of tumor occurrence is
greatly delayed. In terms of human relevance, our results
suggest that early life exposure to UV may introduce
p53 gene mutations in epidermal keratinocytes as well
as keratinocyte progenitors. While some p53-mutated
keratinocytes may be eliminated via differentiation and
epidermal desquamation, others, perhaps the initiated
progenitor cells, may still persist and eventually give
rise to skin tumors even in the absence of further UV
exposure. However, this process will be greatly pro-
moted by UV exposure. Thus, the cancer development
can be delayed but not abrogated upon further avoid-
ance of exposure to UV radiation.
Materials and methods
Mice
Female, 8-week-old SKH-hr1 mice were obtained from
Charles River Laboratories (Wilmington, MA, USA) and
housed in cages in a room with controlled temperature and
humidity and an alternating 12-h light and dark cycle. The
room was lit with yellow fluorescent lamps in ceiling fixtures
with plastic diffusers to eliminate all ambient UV radiation.
The animals were maintained in facilities approved by the
Association for Assessment and Accreditation of Laboratory
Animal Care International, in accordance with current
regulations and standards of the National Institutes of Health.
All animal procedures were reviewed and approved by the
Institutional Animal Care and Use Committee. The mice were
fed ad libitum with a commercial diet and water.
UV irradiation
UV irradiation was performed as described previously (Ouhtit
et al., 2000a, b). Briefly, five mice at a time were placed in a
standard cage with Plexiglas dividers and exposed three times
per week to 2.5 kJ/m
2
of UVB radiation (290–320 nm) from a
bank of six Kodacel-filtered FS40 lamps (Westinghouse
Electric Corp., Bloomfield, NJ, USA) for 8 weeks. The
Kodacel filter (127-mm thick TA422 cellulose triacetate film;
Eastman Kodak, Rochester, NY, USA) removed all UV
wavelengths below 290 nm. The fluence rate was monitored
weekly, and any decrease in fluence was compensated for by a
corresponding increase in irradiation time.
Post-UV procedures and observations
UV irradiation was discontinued after 8 weeks of treatment,
and the mice were randomly divided into two groups. The first
group of 50 mice was monitored for skin tumor development.
The second group of 80 mice was divided into eight subgroups,
and the 10 mice in each subgroup were killed at 1, 4, 7, 10, and
14 days and 6, 10, and 22 weeks after discontinuation of UV
exposure. The skin samples were collected and analysed for
p53 mutations. As a control, a group of 10 unirradiated mice
were monitored for tumor development. In addition, two
groups of five unirradiated mice were each killed at 1 day and
22 weeks post-UV exposure, and their skin was used as
controls in p53 mutation and other assays.
Isolation of skin samples
A22 cm area of dorsal skin was excised from each mouse
that was killed and cut into two pieces. One piece was
immediately fixed in 4% buffered formaldehyde for paraffin-
embedded sectioning. The other piece was floated dermis-side
down in buffered 0.5 Methylenediaminetetraacetic acid solu-
tion, pH 7.4, for 1 h at 371C to separate the epidermis from the
dermis. The epidermal tissues were snap-frozen in liquid
nitrogen and stored at 801C until analysis.
Tumors and pathologic analysis
All 50 mice in the group being monitored for skin tumor
development were visually examined three times per week.
Lesions larger than 3 mm in diameter that persisted for more
UVB cessation, p53 mutations, and skin cancer
VO Melnikova et al
7061
Oncogene
than 2 weeks were considered skin tumors and recorded for
each mouse. When the tumors reached approximately 10 mm
in diameter, the mice were euthanized with CO
2
gas, and the
tumors were excised and fixed in 4% formalin and then
embedded in paraffin. Tissue sections were cut, stained with
H&E, and examined for histopathologic characteristics by a
certified veterinary pathologist.
Allele-specific polymerase chain reaction
Epidermal or tumor DNA was isolated using the phenol–
chloroform method and analysed using AS-PCR for CC-TT
mutations at codons 154/155 and 175/176 in exon 5 and for
C-T mutations at codons 270 and 275 in exon 8 of the p53
gene, as described previously (Ananthaswamy et al., 1997; Hill
et al., 1999; Ouhtit et al., 2000a). The mutant-specific forward
primers used were 50-CCTCCAGCTGGGAGCCGTGCTT-30
and 50-TCGTGAGACGCTGCCCCCATT-30for mutations at
codons 154/155 and 175/176, respectively. The reverse primer
used for amplification of codons 154/155 and 175/176 was
50-GCCTGCGTACCTCTCTTTGC-30.C-T mutations at
codons 270 and 275 were detected using forward mutant-
specific primers 50-GGACGGGACAGCTTTGAGGTTT-30
and 50-GTGTTTGTGCCTGCCT-30, respectively. The reverse
primer used to detect both mutations was 50-GCCTGCGTA
CCTCTCTTTGC-30. PCR reactions were performed with
360 ng of genomic DNA, 2 mMof each primer labeled with
3000 Ci/nmol [g-
32
P]ATP, 2.5 mMdNTPs, 0.2 U Taq DNA
polymerase (Promega) in a buffer containing 1 mMMgCl
2
,in
a final volume of 25 ml. Following initial denaturation step
(941C, 4 min), 35 cycles (denaturation at 941C, 1 min; annea-
ling for 1 min at 581C (codons 154–155 mutations), 571C
(codons 175–176), 691C (codon 270) or 651C (codon 275); and
extension (721C, 1 min)) were carried on a DNA thermal cycler
(Perkin-Elmer/Cetus). An aliquot (7 ml) of the PCR product
was mixed with sequencing stop solution (3 ml) and electro-
phoresed on 8% (codons 154 and 175) or 6% (codons 270 and
275) polyacrylamide gel at 150 V for 30 min. The gel was dried
and visualized on autoradiographic film.
Immunohistochemistry
Sections (5 mm) of paraffin-embedded tissues were analysed for
the expression of p53, PCNA (a marker of hyperplasia),
keratin 10, and loricrin (markers of differentiation for
epidermal keratinocytes) using immunohistochemical analysis,
as described previously (Hill et al., 1999; Ouhtit et al.,
2000a, b). After deparaffinization, the sections were treated
with target retrieval solution (DAKO, Carpinteria, CA, USA),
washed three times with phosphate-buffered saline (PBS), and
incubated in H
2
O
2
/methanol/PBS solution (1 : 50 : 50) for
15 min to block endogenous peroxidase activity. After three
washes in PBS with 0.5% Tween-20, the sections were
preincubated for 10 min in 10% normal goat serum in PBS
and then incubated overnight at 41C with rabbit polyclonal
anti-mouse p53 antibody (cat. # NCLp53-CM5; NovoCastra
Laboratories Ltd, Newcastle upon Tyne, UK; dilution factor
1 : 200), rabbit polyclonal PCNA antibody (clone FL-261, cat.
# sc-7907; Santa Cruz Biotechnology, Santa Cruz, CA, USA;
dilution factor 1 : 100), or rabbit polyclonal antibodies against
mouse keratin 10 or loricrin (cat. # PRB-159P and PRB-145P,
respectively; Covance, Berkley, CA, USA; dilution factor
1 : 500). After three washes in PBS plus 0.5% Tween-20, the
sections were incubated for 1 h at room temperature in
corresponding horseradish peroxidase-linked secondary anti-
body solution (Vectastain Elite ABC kit; Vector Laboratories,
Burlingame, CA, USA). After a wash in PBS, the staining was
performed using the Vectastain Elite ABC kit with diamino-
benzidine as the chromagen, as recommended by the
manufacturer. Counterstaining was performed with hematox-
ylin. As a negative control, tissue sections were stained with
secondary antibody only.
Statistical analysis
Immunohistochemical data were analysed with unpaired
Student’s t-test and p53 mutation data were analysed by
two-sided Fisher’s Exact test. P-values of o0.05 were
considered to be statistically significant.
Acknowledgements
We thank Alexander Gorny for technical assistance and Dawn
Chalaire for editorial assistance. This work was supported by
National Cancer INSTITUTE Grant CA 46523 and U01
CA105345, National Institute of Environmental Health
Sciences Center Grant ES07784, and The University of Texas
MD Anderson Cancer Center institutional core grant CA
16672.
References
Ananthaswamy HN, Loughlin SM, Cox P, Evans RL, Ullrich
SE and Kripke ML. (1997). Nat. Med.,3, 510–514.
Bender K, Blattner C, Knebel A, Iordanov M, Herrlich P and
Rahmsdorf HJ. (1997). J. Photochem. Photobiol. B,37, 1–17.
Berg RJW, van Kranen HJ, Rebel HG, de Vries A, van Vloten
WA, van Kreijl CF, van der Leun JC and de Gruijl FR.
(1996). Proc. Natl. Acad. Sci. USA,93, 274–278.
Blum HF. (1969). The Biologic Effects of Ultraviolet Irradiation
Urbach F (ed). Pergamon Press: Oxford, UK, pp. 543–549.
Brash DE. (1996). Nat. Med.,2, 525–526.
Brash DE, Rudolph JA, Simon JA, Lin A, McKenna GJ,
Baden HP, Halperin AJ and Ponten J. (1991). Proc. Natl.
Acad. Sci. USA,88, 10124–10128.
Brash DE, Ziegler A, Jonason AS, Simon JA, Kunala S
and Leffell DJ. (1996). J. Invest. Dermatol. Symp. Proc.,1,
136–142.
Chung KY, Mukhopadhyay T, Kim J, Casson A, Ro JY,
Goepfert H, Hong WK and Roth JA. (1993). Cancer Res.,
53, 1676–1683.
Coffer PJ, Burgering BM, Peppelenbosch MP, Bos JL and
Kruijer W. (1995). Oncogene,11, 561–569.
de Gruijl FR and Forbes PD. (1995). Bioessays,17, 651–660.
de Gruijl FR and van der Leun JC. (1991). Cancer Res.,51,
979–984.
De-Metys P, Urso B, Christoffersen CT and Shymko RM.
(1995). Ann. NY Acad. Sci.,766, 388–401.
Dittmer D, Pati S, Zambetti G, Chu S, Teresky AK, Moore M,
Finlay C and Levine AJ. (1993). Nat. Genet.,4, 42–46.
Epstein JH and Epstein WL. (1963). J. Invest. Dermatol.,39,
455–460.
Gilchrest BA. (1990). Ann. Rev. Med.,41, 199–210.
Hill LL, Ouhtit A, Loughlin SM, Kripke ML, Ananthaswamy
HN and Owen-Schaub LB. (1999). Science,285, 898–900.
Jonason AS, Kunala S, Price GJ, Restifo RJ, Spinelli HM,
Persing JA, Leffell DJ, Tarone RE and Brash DE. (1996).
Proc. Natl. Acad. Sci. USA,93, 14025–14029.
Jost M, Kari C and Rodeck U. (2000). Eur. J. Dermatol.,10,
505–510.
UVB cessation, p53 mutations, and skin cancer
VO Melnikova et al
7062
Oncogene
Kanjilal S, Pierceall WE, Cummings KK, Kripke ML and
Ananthaswamy HN. (1993). Cancer Res.,53, 2961–2964.
Kanjilal S, Strom S, Clayman G, Weber R, El-Naggar A,
Cummings KK, Hill, LA, Kapur V, Spitz M, Kripke M and
Ananthaswamy HN. (1995). Cancer Res.,55, 3604–3609.
Kieser A, Weich HA, Brandner G, Marme D and Kolch W.
(1994). Oncogene,9, 963–969.
Kripke M. (1991). J. Dermatol.,18, 429–433.
Kuerbitz SJ, Plunkett BS, Walsh WV and Kastan MB. (1992).
Proc. Natl. Acad. Sci. USA,89, 7491–7495.
Kuhn C, Hurwitz SA, Kumar MG, Cotton J and Spandau
DF. (1999). Int. J. Cancer,80, 431–438.
Lanyi A, Deb D, Seymour RC, Ludes-Meyers JH, Subler MA
and Deb S. (1998). Oncogene,16, 3169–3176.
Leffell DJ and Brash DE. (1996). Sci. Am.,275, 52–53,
56–59.
Li G, Mitchell DL, Ho VC, Reed JC and Tron VA. (1996).
Am. J. Pathol.,148, 1113–1123.
Lin J, Teresky AK and Levine AJ. (1995). Oncogene,10,
2387–2390.
Ludes-Meyers JH, Subler MA, Shivakumar CV, Munoz RM,
Jiang P, Bigger JE, Brown DR, Deb SP and Deb S. (1996).
Mol. Cell. Biol.,16, 6009–6019.
Marconi A, Vaschieri C, Zanoli S, Giannetti A and Pincelli C.
(2003). J. Invest. Dermatol.,113, 920–927.
Matas D, Sigal A, Stambolsky P, Milyavsky M, Weisz L,
Schwartz D, Goldfinger N and Rotter V. (2001). EMBO J.,
20, 4163–4172.
Melnikova VO, Santamaria AB, Bolshakov SV and Ana-
nthaswamy HN. (2003). Oncogene,22, 5958–5966.
Miller DL and Weinstock MA. (1994). J. Am. Acad.
Dermatol.,30, 774–778.
Mudgil AV, Segal N, Andriani F, Wang Y, Fusenig NE and
Garlick JA. (2003). J. Invest. Dermatol.,121, 191–197.
Ouhtit A, Gorny A, Muller HK, Hill LL, Owen-Schaub
LB and Ananthaswamy HN. (2000a). Am. J. Pathol.,157,
1975–1981.
Ouhtit A, Muller HK, Davis WD, Ullrich SE, McConkey D and
Ananthaswamy HN. (2000b). Am. J. Pathol.,156, 201–207.
Peus D, Vasa RA, Meves A, Beyerle A and Pittelkow MR.
(2000). Photochem. Photobiol.,72, 135–140.
Rebel H, Mosnier O, Berg RJW, Westerman-de Vries A, van
Steeg H, van Kranen HJ and de Gruijl FR. (2001). Cancer
Res.,61, 977–983.
Remenyik E, Wikonkal NM, Zhang W, Paliwal V and Brash
DE. (2003). Oncogene,22, 6369–6376.
Ren ZP, Hendrum A, Potten F, Nister M, Ahmadian A,
Lundeberg J, Uhlen M and Ponten J. (1996). Oncogene,12,
765–773.
Rosette C and Karin M. (1996). Science,274, 1194–1197.
Sigal A and Rotter V. (2000). Cancer Res.,60, 6788–6793.
Smith ML and Fornace Jr AJ. (1997). Proc. Natl. Acad. Sci.
USA,94, 12255–12257.
Tron VA, Trotter MJ, Tang L, Krajewska M, Reed JC, Ho
VC and Li G. (1998). Am. J. Pathol.,153, 579–585.
Ueba T, Nosaka T, Takahashi JA, Shibata F, Florkiewicz RZ,
Vogelstein B, Oda Y, Kikuchi H and Hatanaka M. (1994).
Proc. Natl. Acad. Sci. USA,91, 9009–9013.
Ullrich SE. (2000). Biochemical Modulation of Skin Reactions
Kydonieus AF and Wille JJ (eds). CRC Press: Boca Raton,
FL, pp. 281–300.
Urbach F. (1991). Dermatol. Clin.,9, 751–755.
Urbach F. (1997). J. Photochem. Photobiol. B,40, 3–7.
van Hogerlinden M, Rozell BL, Ahrlund-Richter L and
Toftgard R. (1999). Cancer Res.,59, 3299–3303.
Walterscheid JP, Ullrich SE and Nghiem DX. (2002). J. Exp.
Med.,195, 171–179.
Young AR. (1990). Semin. Dermatol.,9, 25–31.
Zhang W, Remenyik E, Zelterman D, Brash DE and
Wikonkal NM. (2001). Proc. Natl. Acad. Sci. USA,98,
13948–13953.
Ziegler A, Jonason AS, Leffell DJ, Simon JA, Sharma HW,
Kimmelman J, Remington L, Jacks T and Brash DE. (1994).
Nature,372, 773–776.
UVB cessation, p53 mutations, and skin cancer
VO Melnikova et al
7063
Oncogene
