Oh et al 6/4/07 7:29 PM 1

Histological and proteomic analysis of reversible H-RasV12G expression in transgenic mouse

skin

Won-Jun Oh#, Vikas Rishi#, Steven Pelech$, and Charles Vinson#*

# Laboratory of Metabolism, NCI, CCR, NIH, MD 20892, USA, $Kinexus Bioinformatics

Corporation and the Department of Medicine, University of British Columbia, Vancouver,

British Columbia, Canada

Running Title: Proteomic targets of H-RasV12G de-induction

*Address correspondence to: Charles Vinson, Bldg. 37, Rm. 3128, National Cancer Institute,

National Institutes of Health, Bethesda, MD 20892, Tel: 301-496-8753; Fax: 301-496-8419; E-

mail: Vinsonc@dc37a.nci.nih.gov

Published by Oxford University Press 2007. Carcinogenesis Advance Access published June 5, 2007 by guest on June 7, 2013

http://carcin.oxfordjournals.org/D

ownloaded from

Oh et al 6/4/07 7:29 PM 2

Abstract

We have used a two-transgene tetracycline system to reversibly express oncogenic H-RasV12G in

mouse skin and primary keratinocytes culture using the bovine keratin 5 promoter. Induction of

H-RasV12G expression in skin at 30 days after birth causes epidermal basal cell hyperplasia, an

eruption of keratinous cysts, and loss of hair follicles by 3 weeks. Subsequent H-RasV12G de-

induction for 3 days results in massive apoptosis in the non H-RasV12G expressing stroma as well

as in the suprabasal cells of the epidermis. Several procaspases such as CASP-3, -1? , -5, -12

disappeared while the pro-apoptotic proteins AIF, Bax, and Fas ligand were induced in H-

RasV12G de-induction skin. This process is followed by a wave of cell division at 14 days as hair

follicles regrew, returning to near normal histology and skin appearance by 30 days. Using

Kinetworks™ multi- immunoblotting screens, the phosphorylation status of 37 proteins and

expression levels of 75 protein kinases in the skin were determined in three samples: 1) wild-

type skin, 2) hyperplastic H-RasV12G expressing skin, 3) skin where H-RasV12G expression was

suppressed for 7 days. Following H-RasV12G induction, 16 kinases were increased over 2-fold,

and 2 kinases were reduced over 50%. This included increased phosphorylation of both known

downstream H-RasV12G targets and unknown H-RasV12G targets. After H-RasV12G suppression,

many but not all protein changes were reversed. These results from skin and primary

keratinocytes are organized to reflect the molecular events that cause the histological changes

observed. These proteomic changes identify markers that may mediate the oncogenic addiction

paradigm.

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 3

Introduction

Expression in transgenic mice of either oncogenic H-Ras or K-Ras results in both cellular

proliferation and tumorigenesis [1] producing a variety of mouse models for human disease

including follicular adenoma [2], melanoma [3] and Costello syndrome [4]. Several

investigators have constitutively expressed Ras or its oncogenic forms in the skin and observed

both hyperproliferation of the epidermis and subsequent formation of squamous cell carcinomas

[5] demonstrating that activated Ras is sufficient to produce a malignant phenotype. Expression

of oncogenic Ras in different compartments of the skin produces different results. Suprabasal

epithelial cell expression using the keratin 10 promoter causes hyperplasia and papillomas which

slowly progress into carcinomas [6] while expression in the hair follicle compartment that

contains skin stem cells using the keratin 5 promoter caused papillomas which readily convert

into carcinomas [5].

Mutations in each Ras gene have been detected at high frequency in many human cancers,

with frequencies ranging from 10 to 90% [7], with different cancers types having reproducible

mutations in specific Ras genes. Mouse models of skin chemical carcinogenesis indicate that

oncogenic mutations in Ras plays a central role in skin tumor development [8] suggesting that

the oncogenic properties of Ras are similar in mice and humans.

There are three Ras proto-oncogenes in mammals [1] which have GTPase activity and

function to activate kinases [9] whose downstream targets are transcription factors culminating in

changes in gene expression [10]. Oncogenic Ras is able to transform cells in culture by both

promoting cell growth and inhibiting anoikis [11], a term for cell death induced in tissue culture

when cells fail to make proper cell-cell and cell-stromal interactions [12]. The inhibition of

anoikis is critical for the ability of Ras transformed cells to grow in soft agar [13].

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 4

Ras-signaling pathways underlie cellular transformation as demonstrated by transfection

of either Ras or its downstream targets [14]. However, in epidermal keratinocytes, conflicting

data have been presented indicating that Ras promotes proliferation and suppress differentiation

[15] or does the opposite, such as cell cycle arrest and induction of apoptosis [16].

More recently, oncogenic H-RasV12G has been reversibly expressed in melanocytes to

show that it is continually needed to maintain the malignant phenotype [17]. Similar

observations were made for the reversible expression of Myc [18] and a mutant EGF receptor

[19]. These observations have been conceptua lized to suggest that cancer cells develop a

dependence or “addiction” to the continued activity of over expressed oncogenes for

maintenance of their malignant phenotype [20,21]. However, the molecular correlates of this

poorly understood phenomenon have not been reported.

To identify such molecular correlates, we have reversibly expressed H-RasV12G in the

basal layer of the mouse epidermis using the K5 promoter. Here we show that the expression of

H-RasV12G induces hyperplasia in the skin by 3 weeks and subsequent de- induction result in a

return to normal skin and histology by 30 days. To help determine the mechanism of tumor

regression, we have examined changes in the concentration of over 100 phosphoproteins and

proteins that occur in the skin when H-RasV12G expression is activated and subsequently

suppressed. The reversibility by H-RasV12G induction and de- induction also was also observed in

primary keratinocytes. These proteomic changes that occur as H-RasV12G is de-induced may be

potential therapeutic targets.

Experimental procedures

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 5

Transgenic mice and cell culture. The transgenic mouse containing the tetracycline

transactivator under the control of the bovine K5 promoter (K5-tTA) was a kind gift from Dr.

Adam Glick [22]. The tetO-H-RasV12G transgenic mouse [17] were obtained from NCI Mouse

Repository (Frederick, MD). Offspring from the cross between K5-tTA and tetO-H-RasV12G

mice were genotyped using the PCR primers for H-RasV12G transgene: 5'-

GGTCCACTTCGCATATTAAGG and 5'-GCCGGCGGTATCCAGGATGTCCAAC, and for

tTA transgene: 5'- AACAACCCGTAAA-CTCGCCC and 5’-GCAACCTAAAGTAAAAT-

GCCCCAC.

Primary keratinocytes were isolated from wild-type or K5-tTA:tetO-H-RasV12G (K5:H-RasV12G)

transgenic newborn littermates (1day old) as described [23].

Regulated H-RasV12G expression in mouse skin. To generate K5:H-RasV12G double transgenic

mice, heterozygous K5-tTA mice were crossed to homozygous tetO-H-RasV12G mice in the

presence of 200 mg/kg doxycycline feed (Bioserv, Baltimore, MD). To induce H-RasV12G

expression, 30 day old mice were removed from doxycycline food. After 3 weeks when K5:H-

RasV12G mice showed a phenotype, they were feed food containing 200 mg/kg doxycycline to

suppress expression of the H-RasV12G transgene.

Skin histology, measurement of cell proliferation, and apoptosis. Skin tissues were fixed in

10% NBF overnight, transferred to 95% Et-OH and embedded in paraffin. Six-micrometer

sections were cut and stained with hematoxylin and eosin. To label dividing cells, mice were

sacrificed 60 min after a single intraperitoneal injection of a sterile solution of BrdU (Sigma; 10

mg/ml in PBS, 50 mg/kg of body weight). Skin sections were treated for 30 min at 37°C with

2N HCl in PBS containing 0.5% Triton X-100, rinsed in sodium tetraborate buffer (0.1 M, pH

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 6

8.5) and stained using an anti-BrdU antibody (DAKO diagnostics). Apoptosis was determined in

tissue sections using an ApopTag kit (Chemicon, Temecula, CA).

Western blot analysis. Protein extracts were made from whole skin or primary keratinocytes

using RIPA lysis buffer containing 1% SDS, 1% NP-40, and 0.5% Na-deoxycolate and cocktail

of protease inhibitors. Western blots of lysates were performed on 15% SDS-PAGE and probed

with ? -146 anti H-Ras monoclonal antibodies (Quality Biotech, New Jersey), anti-Bax

polyclonal antibodies and anti-VEGF polyclonal antibodies (Santa Cruz Biotechnology),

differentiation-related antibodies (K1, K10, involucrin, and loricrin; Covance, Richmond, CA),

anti-PKC? , and anti-phospho PKC? (Santa Cruz Biotechnology). Protein bands were visualized

using the ECL chemiluminescence detection kit (GE Healthcare, Piscataway, NJ).

Immunohistochemistry. Serial 6 ? m-thick sections were prepared and immunohisto-chemistry

was done using the ABC and DAB kits from Vector Laboratories. Skin was immunostained with

goat anti-CD31 (1:200; Santa Cruz Biotechnology), and rabbit anti-cytokeratin antibodies

(1:2000; DAKO diagnostics). Histological sections were photographed using an Olympus AX70

microscope (Olympus Optical) with analysis software (Soft Imaging System).

Apoptosis, protein kinase, and phospho-protein screening. Whole skin lysate from mice (0.5

mg) was prepared as described previously [24]. The Kinetworks™ apoptosis (KAPS-1.0),

phospho-site (KPSS-1.3), and protein kinase (KPKS-1.2) screens were performed by Kinexus

Bioinformatics (Vancouver, BC) [25].

Results

Regulated expression of activated H-RasV12G in the skin basal epidermis was achieved by

producing a mouse containing two transgenes. One transgene expresses the tetracyc line

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 7

transactivator under the control of the bovine keratin 5 gene promoter (K5-tTA) that is active in

the basal epidermis of the skin both in utero and in the adult. The second transgene contains the

oncogenic H-RasV12G gene driven by a minimal promoter containing seven DNA binding sites

for the tetracycline transactivator (tetO-H-RasV12G). Crosses between K5-tTA and tetO-H-

RasV12G mice did not produce any double transgenic pups indicating that K5 driven expression of

activated H-RasV12G in utero is lethal. However, if the mothers were fed doxycycline (Dox) in

the food to suppress H-RasV12G expression during development, double transgenic K5:H-RasV12G

mice were born at the expected Mendelian frequencies. These double transgenic mice develop

like their wild-type littermates when continually fed Dox in the food.

H-RasV12G expression was induced at 30 days by feeding K5:H-RasV12G double

transgenic mice food without Dox. All the double transgenic but none of the single transgenic

mice (either K5 or tetO-H-RasV12G mice) developed a wide variety of epithelial alterations,

ranging from benign skin hyperplasia to dysplasia, alopecia, keratinous cysts (Figure 1B, 2A)

and swelling of the salivary glands (data not shown) which are consistent with other groups

findings [5] [26]. However, we did not observe any macro-abnormality of other organs (i.e., oral

mucosa, tongue, esophagus, uterine cervix, and fore stomach) as reported by Vitale-Cross L. et al

[5]. Massive angiogenesis was also observed in the skin (Figure 1F). 21 days after H-RasV12G

expression was induced, expression was suppressed by using food containing 200 mg/kg Dox.

This allowed us to examine the dependence of the skin phenotype on continual H-RasV12G

expression. The expression of H-RasV12G protein was suppressed in the skin after 3 days (Figure

1A). After 10 days, mouse skin showed intermediate phenotype. Finally, the skin appearance

returned to near normal after 30 days (Figure 1C-E). Compare to H-RasV12G expressing skin

(Figure 2A; left panel), dramatic histological changes in the skin were observed at 3 day after

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 8

suppression of H-RasV12G protein (Figure 2A; right panel) and the skin histology became normal

after 30 days (Figure 2C). However, in the presence of Dox when H-RasV12G is not expressed,

WT (Supplementary figure) and K5:H-RasV12G (Figure 2B) skins were histologically

indistinguishable, indicating the tight control of oncogenic H-Ras protein expression on

doxycycline.

Histological analysis of skin during reversible H-RasV12G expression: growth and death

Three types of histological analyses of the skin revealed cellular events involved in both

the H-RasV12G dependent skin hyperplasia and the subsequent regression following H-RasV12G

de-induction. These analyses included; 1) Hematoxylin and eosin-stained sections to examine

general tissue histology, 2) BrdU labeling and immunohistological staining to identify cells

undergoing cell proliferation, and 3) TUNEL analysis to identify cells undergoing apoptosis.

The dorsal skin of H-RasV12G expressing (Figure 2C-E) and wild-type (Supplementary figure)

mice was examined at 8 time points. Four time points were 0, 3, 7, and 21 days of H-RasV12G

expression. Four additional time points examined the consequences of H-RasV12 suppression for

3, 7, 14, and 30 days after being expressed for 21 days. The quantification of cell growth and

apoptosis during this time course is presented for both the K5:H-RasV12G and wild-type mice

(Figure 2F-G).

The induction of H-RasV12G expression resulted in hyperproliferation of the epidermis

that was evident as early as 3 days. At 7 day, the hyperplasia is more pronounced in both the

basal layer of the epidermis and in the hair follicles (Figure 2C), something not observed in wild-

type skin (Supplementary figure). At 21 day, the epidermis was hyperplastic and developed

keratinous cysts. BrdU positive cells were abundant in basal cells and in cells of the cyst walls

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 9

(Figure 2D). There were no apoptotic cells observed at either 3 or 7 day although there was a

slight increase in TUNEL positive cells at 21 day (Figure 2E).

The de- induction of H-RasV12G after 21 days of expression resulted in a dramatic shift

from cell growth to apoptosis (Figure 2F-G). Three days after H-RasV12G de- induction, BrdU

incorporation was not detected while apoptosis, as evident by TUNEL staining, occurred

primarily in the stromal cells that do not express H-RasV12G. In some sections, the entire region

of stroma underwent apoptosis. In addition, apoptotic cells also appeared in the suprabasal cells

of the epidermis. On 14 day of H-RasV12G de-induction, apoptosis was lower while proliferating

cells became detectable in the hair follicles. At this time, many of hair follicles were in the

anagen phase as a part of skin recovery.

To identify if the apoptotic cells are epidermal in origin, serial sections of skin were

stained using a pan-keratin antibody to detect epidermis cells and a CD31 antibody to detect

endothelial cells. Within three days of H-RasV12G de-induction, apoptotic cells were observed

that are pan-keratin negative (Figure 3A, C; left panel). They are near CD31 positive cells

adjacent to blood vessels (Figure 3A, B; left panel). In addition, vascular endothelial growth

factor (VEGF) protein was strongly expressed in H-RasV12G expressing skin, but completely

disappeared within one week of H-RasV12G de-induction (Figure 3D). The significant apoptotic

rate of cells lining tumor vessels suggest that continued H-RasV12G expression is required for the

critical host-tumor symbiotic interaction that sustains stable tumor vasculature and is consistent

with the reports that K- and H-RasV12G can stimulate VEGF expression [27]. Apoptotic cells

were also detected in epidermal cells expressing keratin protein (Figure 3A, C; right panel).

Increase in proapoptotic proteins without effecting the expression of differentiation

markers in H-RasV12G suppressing skin

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 10

It has been reported previously that tumor regression upon oncogene de- induction is due

to apoptosis and differentiation [28]. To identify molecular pathways involved in the induction

of apoptosis in skin after H-RasV12G de-induction, the Kinetworks™ Apoptosis Protein Screen

(KAPS-1.0) that determines protein levels of 25 candidate apoptosis proteins (Figure 4) was used.

The induction of H-RasV12G for 3 weeks increased concentrations of several proteins

including the procaspases (proCASP1? , proCASP3, CASP5 int, and proCASP12) and decreased

concentrations of others including CytoC, Fas and CASP5 p20, a matured form of procaspase 5

(Figure 4D). The increase in procaspase level in H-RasV12G expressing human ovarian cancer

model has been described earlier [29].

When H-RasV12G is de- induced, the four procaspases that were most strongly induced by

H-RasV12G expression decreased to below wild-type levels. In contrast, several proteins that are

proapoptotic were induced including AIF, Fas, and FasL (Figure 4D). Bax protein, that induces

the releases of mitochondrial AIF and is a member of Bcl-2 protein family [30], increased over

three fold in mice after H-RasV12G de-induction (Figure 4E).

To evaluate if the observed apoptosis coincided with cell differentiation as occurs in

normal skin [31], we examined the levels of differentiation markers such as keratin 1, keratin 10,

involucrin, and loricrin. These markers increased with H-RasV12G expression but decreased after

H-RasV12G suppression indicating that the observed apoptosis was not part of the normal

differentiation program (Figure 4F).

Phosphoprotein changes in skin after reversible H-RasV12G expression

Ras act via downstream effectors cascades to alter cellular growth and apoptosis [32]. To

identify kinases and protein phosphorylation changes that accompany H-RasV12G activation and

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 11

subsequent suppression, we used four Kinetworks™ multi- immunoblotting screens to examine

the phosphorylation of 37 proteins and expression levels of 75 protein kinases.

We first utilized a KinetworksTM Phospho-Site Screen (KPSS-1.3) to identify changes in

the phosphorylation status of 37 protein phosphorylation sites after H-RasV12G expression in the

three skin samples described above (Figure 5A-C). These results are summarized in Figure 5D

organized to show the most striking increases in protein phosphorylation at the top and the

largest decreases at the bottom of the bar graph. It should be appreciated that the most reliable

results are provided when the intensity of the ECL signals from the immunoblots are higher.

Figure 5D reveals tha t several known downstream effectors of H-RasV12G including RSK1 (10-

fold), STAT3 (9-fold), and MEK1 (3.5-fold) have increased phosphorylation levels following H-

RasV12G expression. Other proteins, not previously associated with Ras activation, such as NR1

and nucleophosmin (B23), had over a 2-fold increase in their phosphorylation levels. Decreased

phosphorylation is observed with adducin ? and ?, CREB1, p38MAPK and JNK. CREB does

not show increased phosphorylation even though it is a direct target of RSK1, which is induced

10-fold [33].

Following H-RasV12G de- induction, the phosphorylation states of many proteins returned

to normal (Raf1, MEK1, and B23). However, several proteins had increased phosphorylation

(adducin? /?, S6K? ??, PKC? /? /?, GSK3? /? ) and may mediate the increased apoptosis observed

as the hyperplasic state reverts to a normal.

Protein kinase changes in skin after reversible H-RasV12G expression

To evaluate the consequences of H-RasV12G induction and de-induction on protein kinase

expression levels, we employed a Kinetworks™ protein kinase screen, which detected 48 out of

a possible 75 protein kinases in the mouse skin samples. Following H-RasV12G induction, 16

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 12

kinases increased over 2-fold while only two decreased more than 2-fold. Seven kinases

increased 3-fold or more following H-RasV12G expression (PAK3, CDK1/4, BMX, S6K? ?? , Csk,

Raf1 and MEK1) (Figure 5E).

When H-RasV12G is de- induced, only three (Pak3, BMX, and S6K? ?? ) of the seven

kinases that increase 3-fold revert to near normal while CDK4 concentration actually increases.

Several additional kinases also increase with H-RasV12G de-induction (ZIPK, CaMK4, CaMKK,

FAK, and PKC??? ) (Figure 5E).

The protein kinase screen (Figure 5E) and the phosphoprotein screen (Figure 5D) contain

signals for 16 common proteins allowing one to determine if the fraction of protein that is

phosphorylated changes in the different samples. The stoichiometry of phosphorylation (the

ratio of total phospho-site signal divided by the total protein signal) was substantially reduced

with H-RasV12G de-induction for Raf1 (by 54%), MEK1 (by 76%), ERK2 (by 55%), GSK3? ?(by

54%), and PKC? (by 55%). In contrast, H-RasV12G de- induction increases the stoichiometry of

phosphorylation for S6K? /? ?T389 (by greater than 1000%) and Rsk 1/3 T359+S363/T356+S360

(by 330%), but there was a compensatory reduction in the total protein levels of these kinases at

the same time.

Reversible morphological changes following H-RasV12G suppression in primary

keratinocytes.

To determine if the reversibility observed in adult mice skin after H-RasV12G suppression

was also observed in cell culture, we prepared primary keratinocytes cultures from the K5:H-

RasV12G new-born pups. 2 days after induction of H-RasV12G expression in primary keratinocytes,

distinct morphological changes were observed including swollen cytoplasm with vacuoles. 9

days of H-RasV12G expression causes dramatic morphological changes (Figure 6A; upper panel).

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 13

These morphological changes were totally reversed within 7 days after suppression of H-RasV12G

expression (Figure 6A; lower panel). We also found that expression of H-RasV12G protein was

dramatically decreased within 5 days after addition of doxycycline (Figure 6B). To confirm the

skin screening results, we examined the levels of PKC? in the keratinocytes. Compare to H-

RasV12G expressing keratinocytes, the phosphorylation of PKC? protein dramatically increased

>5-fold on 5 day after H-RasV12G suppression and then decreased on 7 day (Figure 6B).

However, compare to H-Ras expressing keratinocytes, total PKC? protein level was not

significantly different in the H-RasV12G suppressed keratinocytes (Figure 6B).

Discussion

We present for the first time a histological and proteomic examination of the reversible

expression of oncogenic H-RasV12G in the mouse epidermis under keratin 5 promoter control.

This study shows that the H-RasV12G induced epidermal hyperplasia is reversible after H-RasV12G

de-induction, a process that involves apoptosis in both the stroma and the suprabasal cells of the

epidermis. The massive induction of apoptosis in the stroma within 3 day of H-RasV12G

suppression occurred near blood vessels in an “all or none” type pattern. These results indicate

that H-RasV12G expression is needed in a non-cell autonomous (paracrine) manner to maintain the

non-H-RasV12G expressing cells. Apoptosis also occurred in the suprabasal cells with individual

cells being TUNEL positive. To initiate a molecular characterization of this poorly understood

phenomenon of cell death induced by oncogene withdrawal, we have identified changes in

protein expression and phosphorylation status in the skin following H-RasV12G de-induction.

These proteins are potential biomarkers and molecular mediators of oncogenic addiction.

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 14

Oncogenic dependence or oncogenic addiction is a term used to describe the observation

that tumors initiated by oncogene expression are dependent on their continual expression for

tumor stability [20] [21]. When the oncogene is de-induced, the tumors swiftly regress mainly

by apoptosis or differentiation. For example, H-RasV12G induced mouse melanomas regressed

via apoptosis when H-RasV12G expression was suppressed [17] or partially regress when K-Ras is

suppressed in lung cancers [34]. A similar reversal of the tumor phenotype via apoptosis is

observed when the expression of the c-Myc oncogene [18] or mutant EGF receptor [19] is de-

induced. However, this phenomenon has never been described in the epidermis.

The induction of apoptosis in the stroma where H-RasV12G is not expressed suggests a

paracrine function for cells that express H-RasV12G. One of these proteins could be the secreted

VEGF that is upregulated by H-RasV12G induction in skin and substantially downregulated when

H-RasV12G is suppressed (Figure 3). Down regulation of VEGF following H-RasV12G

suppression is also observed in melanocytes [17]. In that study, forced expression of VEGF was

not sufficient to reverse the induction of apoptosis when H-RasV12G was de- induced suggesting

that additional H-RasV12G targets are critical for the induction of apoptosis in melanomas.

Using the Kinetworks apoptosis protein screen (KAPS-1.0) (Figure 4), expression

profiles of number of apoptotic proteins were studied. Several procaspases (1? , 3, 5, and 12)

that were induced following H-RasV12G induction decreased or disappeared in the skin after H-

RasV12G de-induction as the skin switches from cell growth to apoptosis (Figure 4). Several

proapoptotic proteins increased including Bax and AIF involved in the intrinsic apoptosis

pathway and FAS and FAS ligand involved in the extrinsic apoptosis pathway [35]. An increase

in AIF level suggests that a caspase- independent mechanism may also play a role to induce

apoptosis. However, unlike the previous reports showed that marked differentiation in oncogene

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 15

inactivation [18] [36], H-RasV12G suppressing skin did not show any increase in differentiation

(Figure 4), suggesting that differentiation may not be a contributing factor in reversible

phenotype.

The molecular changes that accompany tumor regression when the activating oncogene is

de-induced have not previously been described. To begin to address this issue, we investigated

protein changes in the skin after H-RasV12G induction and subsequent de- induction. We

examined the phosphorylation status of 37 phosphoproteins and the expression levels of 75

protein kinases. The proteomic analysis examined the entire skin and describes a global change

in protein concentrations while our histological analysis identified apoptosis in two distinct cell

populations following H-RasV12G suppression, the stroma and the suprabasal epidermis. Future

studies will address where in the skin these proteomic changes occur.

Targets of H-RasV12G activity that induce cell proliferation are readily identifiable in the

hyperplastic skin. These include phosphorylation of several well-known downstream effectors

of Ras signaling via Raf1, including Raf1 itself, MEK1, RSK1, and STAT3. There was also a

decrease in the phosphorylation of MEK6 and its target p38 MAP kinase as well as JNK MAP

kinase following H-RasV12G induction (Figure 5) consistent with repression of the inhibitory roles

that these protein kinases normally play in cell proliferation and their proapoptotic functions. In

contrast, ERK phosphorylation was not changed much in H-RasV12G induced skin (Figure 5D).

These results have precedence. Mao and co-workers [37] have shown that phophorylated ERK

proteins are barely detectable in skin papilloma with the oncogenic H-Ras protein. We speculate

that our results may indicate an imbalance of phophotases and oncogene signaling. We also

measured the expression levels of 25 different phosphatases using a Kinetworks™ protein

phosphatase screen (KPPS-1.2; unpublished work), H-RasV12G induced greater than a 2-fold

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 16

elevation of the catalytic and regulatory A’2 subunit levels of protein-serine phosphatase 4 (PP4)

and the dual specificity MAP kinase phosphatase 2 (MKP2). With H-RasV12G suppression, there

was a complete loss of PP4-A’2 expression and a reduction of PP4-C levels to even below the

wild-type values. By contrast, H-RasV12G suppression also caused even more marked increases in

the expressions of MKP2. The finding that the concentrations of diverse protein phosphatases

were also markedly altered with the gain of Ras function as well as its subsequent repression

adds an extra dimension to the complexity in interpreting the observed changes in protein

phosphorylation. The induction of the dual specificity MAP kinase phosphatase (MKP2) may

partly account for why the phosphorylation levels of ERK1 and ERK2 were not significantly

affected in the skin of H-RasV12G expressing and repressed animals, despite the increases in

expression and activating phosphorylation of the upstream kinase MEK1.

When H-RasV12G is de- induced, signaling pathways upstream of the apoptotic machinery

(discussed above) also changes. These changes have previously been shown to have either anti-

apoptotic or proapoptotic consequences. Anti-apoptotic signals are increased phosphorylation of

the?? ? and??? ??subunits of S6K which increase ribosomal translation. An additional anti-apoptotic

signal is the increase in FAK protein, a kinase involved in organizing stress fibers in cell culture

[38].

Several proapoptotic changes are evident. These include decreased phosphorylation of

Akt1, a survival factor, that has also been shown to respond to VEGF to promote endothelial cell

survival [39] thus contributing to apoptosis in the stroma. In addition, an increase in p38 MAPK

phosphorylation level [40] and ZIP kinase protein after H-RasV12G de-induction is consistent with

previously reported work [41].

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 17

Proapoptotic signals that increase include phosphorylation of PKC? ?? ?and PKC?. PKC’s

promote apoptosis in human gastric cancer cells [42] and mouse keratinocytes [43]. Our studies

indicate PKC? activation after H-RasV12G suppression could also contribute to cell killing that is

often required before the recovery process. The result showing that the phosphorylation of

PKC? decreased in H-RasV12G expressing skin (Figure 5D) was consistent with the previous

reports that downregulation of PKC protein level and activity are seen in cells expressing

activated Ras genes [44-48]. The PKC phosphorylation targets adducin? ???also increases which

weaken cell-cell interactions resulting in apoptosis [49]. Furthermore, the phenotypic

reversibility and an increase in PKC? phosphorylation also occurred in the primary keratinocytes

system after H-RasV12G suppression (Figure 6) indicating that it is a cell autonomous property.

More work will be needed to determine if these pathways are active in apoptotic cells.

Interestingly, in primary keratinocytes cell culture (Figure 6), the significant apoptotic event did

not occur in the reversible manner upon H-Ras suppression (data not shown), suggesting the

possibility that the reversibility following doxycycline addition may reflect a continued

requirement for H-Ras in promoting and sustaining cell-autonomous hyperplasia–host

interactions that are essential for the skin phenotype and maintenance that are not faithfully

mimicked in tissue culture.

In summary, we have identified protein and phosphoprotein changes that accompany H-

RasV12G induction and de-induction in mouse skin. These molecular changes may mediate the

poorly understood phenomenon of oncogenic addiction.

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 18

Acknowledgements : We thank Stuart Yuspa, Ronald Wolf, Christophe Cataisson, and Lyuba

Varticovski for insights into Ras signaling.

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 19

Figure legends

Figure 1. A) Change in H-RasV12G protein in K5:H-RasV12G mice when de-induced for 0, 3, 7, or

30 days following expression for 3 weeks (0 days). B) Dorsal view of mouse following 3 weeks

of H-RasV12G expression. Reversal of skin hyperplasia (arrows) following suppression of H-

RasV12G expression. C) H-RasV12G expression for 3 weeks. H-RasV12G expression for 3 weeks

and subsequently suppression for D) 10 days and E) 30 days. F) Dissected skin expressing H-

RasV12G for 3 weeks showing extensive angiogenesis.

Figure 2. Histological differences in K5:H-RasV12G mice skin during H-RasV12G induction and

de-induction. Hematoxylin and eosin-stained sections from A) H-RasV12G expressing skin for 3

weeks (left), H-RasV12G suppression for 3 days after 3 week induction (right panel), and B) H-

RasV12G suppression for 3 weeks. Time-dependent changes in skin histology of K5:H-RasV12G

mice. Eight time points are presented. H-RasV12G expression was induced and samples were

examined at 0, 3, 7, and 21 days. H-RasV12G expression was suppressed and samples were

examined at 3, 7, 14, and 30 days. C) H&E. D) BrdU immunohistochemistry to monitor cell

proliferation. E) TUNEL assay to monitor apoptosis. Number of positive cells/mm at the eight

time points presented in Figure 2D-E and supplementary figure for F) BrdU positive cells. G)

TUNEL positive cells.

Figure 3. Serial sections of skin showing TUNEL assay and immunohistochemistry using CD31

or pan-keratin antibody in samples expressing H-RasV12G for 3 weeks and subsequent de-

induction for 3 days. A) TUNEL assay to identify apoptotic cells (arrows). B) CD31

immunohistochemistry to identify endothelial cells (arrows). E, epidermis; S, stroma. Dotted line,

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 20

epidermis-stroma borderline. C) Pan-Keratin immunohistochemistry to identify epidermis shown

by arrows. D) Western analysis of VEGF following H-RasV12G suppression.

Figure 4. KinetworksTM KAPS1.0 analysis of 25 apoptosis related proteins for three skin

samples. A) wild-type skin. B) Skin expressing H-RasV12G for 3 weeks, and C) Skin following

H-RasV12G expression for 3 weeks and then suppression for 1 week. D) The percent change in

concentration for 17 detected proteins relative to wild-type skin with H-RasV12G expression

(black bars) and H-RasV12G de-induction (white bars) are shown. The average intensity of the

ECL signal (counts per minute) from wild-type skin is in the right column. E) Bax protein

concentration at the same time points used in B and C. F) Protein concentrations of four

epidermal differentiation markers at 0, 3, 7, and 30 days following H-RasV12G suppression.

Figure 5. KinetworksTM KPSS1.3 analysis of 37 phospho-sites in cell signaling proteins for

three skin samples (A-C) described in Figure 4. The positions of detected target

phosphoproteins are indicated with arrows. D) The percent changes in expression for 17

detected target proteins with H-RasV12G expression (black bars) and H-RasV12G de-induction

(white bars) are shown. Error bars indicate the standard error of the means from four separate

experiments performed on different samples. E) The percent change in concentration for 48

detected proteins in KinetworksTM KPKS 1.2 analysis. Error bars indicate the standard error of

the means from two separate experiments performed on different samples.

Figure 6. Effect of H-RasV12G expression and suppression on primary keratinocytes morphology

and expression profile of PKC? . A) Primary keratinocytes in which H-RasV12G protein was

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 21

expressed for 2 days. After 2 days, H-RasV12G was suppressed or expressed for another 7 day

and cell morphologies were monitored by phase-contrast microscopy. B) Time-dependent

expression levels of H-RasV12G , phospho PKC? , and total PKC? protein.

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 22

References

1. Malumbres, M. and Barbacid, M. (2003) RAS oncogenes: the first 30 years. Nat Rev

Cancer, 3, 459-65. 2. Nikiforova, M.N., Lynch, R.A., Biddinger, P.W., Alexander, E.K., Dorn, G.W., 2nd,

Tallini, G., Kroll, T.G. and Nikiforov, Y.E. (2003) RAS point mutations and PAX8-PPAR gamma rearrangement in thyroid tumors: evidence for distinct molecular pathways in thyroid follicular carcinoma. J Clin Endocrinol Metab, 88, 2318-26.

3. Sekiya, T., Fushimi, M., Hori, H., Hirohashi, S., Nishimura, S. and Sugimura, T. (1984) Molecular cloning and the total nucleotide sequence of the human c-Ha-ras-1 gene activated in a melanoma from a Japanese patient. Proc Natl Acad Sci U S A, 81, 4771-5.

4. Aoki, Y., Niihori, T., Kawame, H., Kurosawa, K., Ohashi, H., Tanaka, Y., Filocamo, M., Kato, K., Suzuki, Y., Kure, S. and Matsubara, Y. (2005) Germline mutations in HRAS proto-oncogene cause Costello syndrome. Nat Genet, 37, 1038-40.

5. Vitale-Cross, L., Amornphimoltham, P., Fisher, G., Molinolo, A.A. and Gutkind, J.S. (2004) Conditional expression of K-ras in an epithelial compartment that includes the stem cells is sufficient to promote squamous cell carcinogenesis. Cancer Res, 64, 8804-7.

6. Bailleul, B., Surani, M.A., White, S., Barton, S.C., Brown, K., Blessing, M., Jorcano, J. and Balmain, A. (1990) Skin hyperkeratosis and papilloma formation in transgenic mice expressing a ras oncogene from a suprabasal keratin promoter. Cell, 62, 697-708.

7. Bos, J.L. (1989) ras oncogenes in human cancer: a review. Cancer Res, 49, 4682-9. 8. Yuspa, S.H. (1994) The pathogenesis of squamous cell cancer: lessons learned from

studies of skin carcinogenesis--thirty-third G. H. A. Clowes Memorial Award Lecture. Cancer Res, 54, 1178-89.

9. Marshall, M.S. (1995) Ras target proteins in eukaryotic cells. Faseb J, 9, 1311-8. 10. Shuman, J.D., Sebastian, T., Kaldis, P., Copeland, T.D., Zhu, S., Smart, R.C. and

Johnson, P.F. (2004) Cell cycle-dependent phosphorylation of C/EBPbeta mediates oncogenic cooperativity between C/EBPbeta and H-RasV12. Mol Cell Biol, 24, 7380-91.

11. McFall, A., Ulku, A., Lambert, Q.T., Kusa, A., Rogers-Graham, K. and Der, C.J. (2001) Oncogenic Ras blocks anoikis by activation of a novel effector pathway independent of phosphatidylinositol 3-kinase. Mol Cell Biol, 21, 5488-99.

12. Aplin, A.E., Short, S.M. and Juliano, R.L. (1999) Anchorage-dependent regulation of the mitogen-activated protein kinase cascade by growth factors is supported by a variety of integrin alpha chains. J Biol Chem, 274, 31223-8.

13. Eckert, L.B., Repasky, G.A., Ulku, A.S., McFall, A., Zhou, H., Sartor, C.I. and Der, C.J. (2004) Involvement of Ras activation in human breast cancer cell signaling, invasion, and anoikis. Cancer Res, 64, 4585-92.

14. Shields, J.M., Pruitt, K., McFall, A., Shaub, A. and Der, C.J. (2000) Understanding Ras: 'it ain't over 'til it's over'. Trends Cell Biol, 10, 147-54.

15. Yuspa, S.H., Dlugosz, A.A., Cheng, C.K., Denning, M.F., Tennenbaum, T., Glick, A.B. and Weinberg, W.C. (1994) Role of oncogenes and tumor suppressor genes in multistage carcinogenesis. J Invest Dermatol, 103, 90S-95S.

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 23

16. Roper, E., Weinberg, W., Watt, F.M. and Land, H. (2001) p19ARF-independent induction of p53 and cell cycle arrest by Raf in murine keratinocytes. EMBO Rep, 2, 145-50.

17. Chin, L., Tam, A., Pomerantz, J., Wong, M., Holash, J., Bardeesy, N., Shen, Q., O'Hagan, R., Pantginis, J., Zhou, H., Horner, J.W., 2nd, Cordon-Cardo, C., Yancopoulos, G.D. and DePinho, R.A. (1999) Essential role for oncogenic Ras in tumour maintenance. Nature, 400, 468-72.

18. Felsher, D.W. and Bishop, J.M. (1999) Reversible tumorigenesis by MYC in hematopoietic lineages. Mol Cell, 4, 199-207.

19. Politi, K., Zakowski, M.F., Fan, P.D., Schonfeld, E.A., Pao, W. and Varmus, H.E. (2006) Lung adenocarcinomas induced in mice by mutant EGF receptors found in human lung cancers respond to a tyrosine kinase inhibitor or to down-regulation of the receptors. Genes Dev, 20, 1496-510.

20. Weinstein, I.B. (2002) Cancer. Addiction to oncogenes--the Achilles heal of cancer. Science, 297, 63-4.

21. Varmus, H. (2006) The new era in cancer research. Science, 312, 1162-5. 22. Diamond, I., Owolabi, T., Marco, M., Lam, C. and Glick, A. (2000) Conditional gene

expression in the epidermis of transgenic mice using the tetracycline-regulated transactivators tTA and rTA linked to the keratin 5 promoter. J Invest Dermatol, 115, 788-94.

23. Dlugosz, A.A., Glick, A.B., Tennenbaum, T., Weinberg, W.C. and Yuspa, S.H. (1995) Isolation and utilization of epidermal keratinocytes for oncogene research. Methods Enzymol, 254, 3-20.

24. Nunez, N.P., Oh, W.J., Rozenberg, J., Perella, C., Anver, M., Barrett, J.C., Perkins, S.N., Berrigan, D., Moitra, J., Varticovski, L., Hursting, S.D. and Vinson, C. (2006) Accelerated tumor formation in a fatless mouse with type 2 diabetes and inflammation. Cancer Res, 66, 5469-76.

25. Pelech, S., Sutter, C. and Zhang, H. (2003) Kinetworks protein kinase multiblot analysis. Methods Mol Biol, 218, 99-111.

26. Raimondi, A.R., Vitale-Cross, L., Amornphimoltham, P., Gutkind, J.S. and Molinolo, A. (2006) Rapid development of salivary gland carcinomas upon conditional expression of K-ras driven by the cytokeratin 5 promoter. American Journal of Pathology, 168, 1654-65.

27. Okada, F., Rak, J.W., Croix, B.S., Lieubeau, B., Kaya, M., Roncari, L., Shirasawa, S., Sasazuki, T. and Kerbel, R.S. (1998) Impact of oncogenes in tumor angiogenesis: mutant K-ras up-regulation of vascular endothelial growth factor/vascular permeability factor is necessary, but not sufficient for tumorigenicity of human colorectal carcinoma cells. Proc Natl Acad Sci U S A, 95, 3609-14.

28. Felsher, D.W. (2004) Reversibility of oncogene-induced cancer. Curr Opin Genet Dev, 14, 37-42.

29. Young, T., Mei, F., Liu, J., Bast, R.C., Jr., Kurosky, A. and Cheng, X. (2005) Proteomics analysis of H-RAS-mediated oncogenic transformation in a genetically defined human ovarian cancer model. Oncogene, 24, 6174-84.

30. Cregan, S.P., Fortin, A., MacLaurin, J.G., Callaghan, S.M., Cecconi, F., Yu, S.W., Dawson, T.M., Dawson, V.L., Park, D.S., Kroemer, G. and Slack, R.S. (2002) Apoptosis-

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 24

inducing factor is involved in the regulation of caspase-independent neuronal cell death. J Cell Biol, 158, 507-17.

31. Maruoka, Y., Harada, H., Mitsuyasu, T., Seta, Y., Kurokawa, H., Kajiyama, M. and Toyoshima, K. (1997) Keratinocytes become terminally differentiated in a process involving programmed cell death. Biochem Biophys Res Commun, 238, 886-90.

32. Downward, J. (1998) Ras signalling and apoptosis. Curr Opin Genet Dev, 8, 49-54. 33. Song, K.S., Seong, J.K., Chung, K.C., Lee, W.J., Kim, C.H., Cho, K.N., Kang, C.D., Koo,

J.S. and Yoon, J.H. (2003) Induction of MUC8 gene expression by interleukin-1 beta is mediated by a sequential ERK MAPK/RSK1/CREB cascade pathway in human airway epithelial cells. J Biol Chem, 278, 34890-6.

34. Fisher, G.H., Wellen, S.L., Klimstra, D., Lenczowski, J.M., Tichelaar, J.W., Lizak, M.J., Whitsett, J.A., Koretsky, A. and Varmus, H.E. (2001) Induction and apoptotic regression of lung adenocarcinomas by regulation of a K-Ras transgene in the presence and absence of tumor suppressor genes. Genes Dev, 15, 3249-62.

35. Raj, D., Brash, D.E. and Grossman, D. (2006) Keratinocyte apoptosis in epidermal development and disease. J Invest Dermatol, 126, 243-57.

36. Shachaf, C.M., Kopelman, A.M., Arvanitis, C., Karlsson, A., Beer, S., Mandl, S., Bachmann, M.H., Borowsky, A.D., Ruebner, B., Cardiff, R.D., Yang, Q., Bishop, J.M., Contag, C.H. and Felsher, D.W. (2004) MYC inactivation uncovers pluripotent differentiation and tumour dormancy in hepatocellular cancer. Nature, 431, 1112-7.

37. Mao, J.H., To, M.D., Perez-Losada, J., Wu, D., Del Rosario, R. and Balmain, A. (2004) Mutually exclusive mutations of the Pten and ras pathways in skin tumor progression. Genes & Development, 18, 1800-5.

38. Thannickal, V.J., Lee, D.Y., White, E.S., Cui, Z., Larios, J.M., Chacon, R., Horowitz, J.C., Day, R.M. and Thomas, P.E. (2003) Myofibroblast differentiation by transforming growth factor-beta1 is dependent on cell adhesion and integrin signaling via focal adhesion kinase. J Biol Chem, 278, 12384-9.

39. Fujio, Y. and Walsh, K. (1999) Akt mediates cytoprotection of endothelial cells by vascular endothelial growth factor in an anchorage-dependent manner. J Biol Chem, 274, 16349-54.

40. Shimizu, H., Banno, Y., Sumi, N., Naganawa, T., Kitajima, Y. and Nozawa, Y. (1999) Activation of p38 mitogen-activated protein kinase and caspases in UVB-induced apoptosis of human keratinocyte HaCaT cells. J Invest Dermatol, 112, 769-74.

41. Kawai, T., Akira, S. and Reed, J.C. (2003) ZIP kinase triggers apoptosis from nuclear PML oncogenic domains. Mol Cell Biol, 23, 6174-86.

42. Okuda, H., Adachi, M., Miyazawa, M., Hinoda, Y. and Imai, K. (1999) Protein kinase Calpha promotes apoptotic cell death in gastric cancer cells depending upon loss of anchorage. Oncogene, 18, 5604-9.

43. Cataisson, C., Joseloff, E., Murillas, R., Wang, A., Atwell, C., Torgerson, S., Gerdes, M., Subleski, J., Gao, J.L., Murphy, P.M., Wiltrout, R.H., Vinson, C. and Yuspa, S.H. (2003) Activation of cutaneous protein kinase C alpha induces keratinocyte apoptosis and intraepidermal inflammation by independent signaling pathways. J Immunol, 171, 2703-13.

44. Guillem, J.G., Hsieh, L.L., O'Toole, K.M., Forde, K.A., LoGerfo, P. and Weinstein, I.B. (1988) Changes in expression of oncogenes and endogenous retroviral-like sequences during colon carcinogenesis. Cancer Res, 48, 3964-71.

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

Oh et al 6/4/07 7:29 PM 25

45. Colapietro, A.M., Goodell, A.L. and Smart, R.C. (1993) Characterization of benzo[a]pyrene-initiated mouse skin papillomas for Ha-ras mutations and protein kinase C levels. Carcinogenesis, 14, 2289-95.

46. Weyman, C.M., Taparowsky, E.J., Wolfson, M. and Ashendel, C.L. (1988) Partial down-regulation of protein kinase C in C3H 10T 1/2 mouse fibroblasts transfected with the human Ha-ras oncogene. Cancer Res, 48, 6535-41.

47. Haliotis, T., Trimble, W., Chow, S., Bull, S., Mills, G., Girard, P., Kuo, J.F. and Hozumi, N. (1990) Expression of ras oncogene leads to down-regulation of protein kinase C. Int J Cancer, 45, 1177-83.

48. Wolfman, A. and Macara, I.G. (1987) Elevated levels of diacylglycerol and decreased phorbol ester sensitivity in ras-transformed fibroblasts. Nature, 325, 359-61.

49. Imamdi, R., de Graauw, M. and van de Water, B. (2004) Protein kinase C mediates cisplatin-induced loss of adherens junctions followed by apoptosis of renal proximal tubular epithelial cells. J Pharmacol Exp Ther, 311, 892-903.

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from

by guest on June 7, 2013http://carcin.oxfordjournals.org/

Dow

nloaded from