mataaai

T1

Toxicology and Applied Pharmacology176,64–71 (2001)doi:10.1006/taap.2001.9277, available online at http://www.idealibrary.com on

Arsenite Is a Cocarcinogen with Solar Ultraviolet Radiation for MouseSkin: An Animal Model for Arsenic Carcinogenesis

Toby G. Rossman, Ahmed N. Uddin, Fredric J. Burns, and Maarten C. Bosland

Nelson Institute of Environmental Medicine and Kaplan Cancer Center, New York University School of Medicine,57 Old Forge Road, Tuxedo, New York 10987

Received June 1, 2001; accepted June 29, 2001

idendiicodinas

s

e es n ofc rgans( ver,n odelf d inI archC re,w ani-m olece fewe basiso can-c

is hasb , theU ed al

e noe ofusingshold(low

stud-lentinor-

ncil,nousvingthisoter.

sultsomo-

estedtiators

, are

Arsenite Is a Cocarcinogen with Solar Ultraviolet Radiation forMouse Skin: An Animal Model for Arsenic Carcinogenesis. Ross-man, T. G., Uddin, A. N., Burns, F. J., and Bosland, M. C. (2001).Toxicol. Appl. Pharmacol. 176, 64–71.

Although epidemiological evidence shows an association be-tween arsenic in drinking water and increased risk of skin, lung,and bladder cancers, arsenic compounds are not animal carcino-gens. The lack of animal models has hindered mechanistic studiesof arsenic carcinogenesis. Previously, this laboratory found thatlow concentrations of arsenite (the likely environmental carcino-gen) which are not mutagenic can enhance the mutagenicity ofother agents, including ultraviolet radiation (UVR). This enhanc-ing effect appears to result from inhibition of DNA repair byarsenite. Recently we found that low concentrations of arsenitedisrupted p53 function and upregulated cyclin D1. These resultssuggest that the failure to find an animal model for arsenic carci-nogenesis is because arsenite is not a carcinogen per se, but ratheracts as an enhancing agent (cocarcinogen) with a genotoxic part-ner. We tested this hypothesis with solar UVR as carcinogenicstimulus in hairless Skh1 mice. Mice given 10 mg/l sodium arsenitein drinking water for 26 weeks had a 2.4-fold increase in yield oftumors after 1.7 KJ/m2 UVR three times weekly compared with

ice given UVR alone. No tumors appeared in mice given arsenitelone. The tumors were mostly squamous cell carcinomas, andhose occurring in mice given UVR plus arsenite appeared earliernd were much larger and more invasive than in mice given UVRlone. These results are consistent with the hypothesis that arseniccts as a cocarcinogen with a second (genotoxic) agent by inhib-ting DNA repair and/or enhancing positive growth signaling.

© 2001 Academic Press

Key Words: arsenic; skin; tumor; mice; ultraviolet radiation.

Arsenic contamination of drinking water is a worldwproblem. It is especially severe in the Bengal region of Iand in parts of Bangladesh, Taiwan, Chile, Argentina, Mexand China, but is also of concern in other countries, incluthe Western states of the United States and in small areNew England and the Midwest (Guha-Mazumderet al.,1998;

ondelet al.,1999; Chiouet al.,2001; Hopenhayn-Richet al.,996; Garcia-Vargaset al., 1994; Lewiset al., 1999; Karaga

640041-008X/01 $35.00Copyright © 2001 by Academic PressAll rights of reproduction in any form reserved.

a,gof

t al., 2001; Welch et al., 1999). Epidemiologic evidenctrongly implicates exposure to arsenic in the causatioancers of the lung, skin, and bladder and possibly other oIARC, 1980; National Research Council, 2000). Howeumerous attempts for over 30 years to find an animal m

or arsenic carcinogenesis have mostly failed (revieweARC, 1980; Wang and Rossman, 1996; National Reseouncil, 2000). This is particularly true for oral exposuhere arsenic compounds alone fail to induce tumors inals (Kanisawa and Schroeder, 1967; Milner, 1969; Germt al., 1997). Thus, arsenic has remained one of thexamples of a well-established human carcinogen (on thef epidemiological evidence) which has failed to induceers in animals.Understanding the mechanism of arsenic carcinogenes

een hindered by this lack of an animal model. RecentlyS Environmental Protection Agency (USEPA) suggest

imit of 10 mg/l (reduced from 50mg/l) arsenic in drinkingwater (Environmental Protection Agency, 2000). Becausanimal model was available at the time to establish modaction, the USEPA based the decision for risk assessmentthe default assumption that carcinogens do not have a threand that excess risk is proportional to dose at low valuesdose linearity).

The increased cancer risk observed in epidemiologicalies is attributed mainly to the presence of inorganic trivaarsenic (arsenite) (IARC, 1980), although pentavalentganic arsenic is readily converted to the trivalent formin vivo(reviewed in Rossman, 1998; National Research Cou2000). Arsenite is not significantly mutagenic at endogeloci in bacterial or mammalian cells at concentrations gihigh levels of survival (reviewed in Rossman, 1998). Forreason, arsenite is sometimes considered a tumor promThere is little evidence for this view, because negative rehave been obtained in bioassays testing arsenite for prtional activity (Milner, 1969; Germolecet al., 1997). Arseniccompounds were also not carcinogenic to animals when tat reasonable doses as complete carcinogens or as ini(IARC, 1980; National Research Council, 2000).

Low concentrations of arsenite, which are not mutagenic

VResn o

UVernd

ffecifincemth

h y ba ichc othe atw reai sioa 01T wea eneB2 seo tiveg ultd feco

ush n,e ens end litys

etors wwaandmemitt

oupic

umce

. Tls., mous

90

00Rmeterury-of the

kept

ment. The

andanimalumorsVR by. Theth

r-ctionsssified

cellsocyticand

typi-ermisider-into; fibro-fers tollularCCwhenement

stedradesumoror

w

kingd toain-withuallymice

-halftxpo-R on

r inlackther

UVRmice

65ARSENITE IS A MOUSE COCARCINOGEN

able to enhance the mutagenicity of ultraviolet radiation (U(Lee et al., 1985; Li and Rossman, 1991). Evidence suggthat this comutagenic effect of arsenite is due to inhibitiorepair of UVR-induced DNA damage (Hartwiget al., 1997).Comutagenic effects of arsenite for mutagens other thanhave also been demonstrated by us and others, and thevidence implicating inhibition of DNA repair (both base anucleotide excision repair) as the mechanism for these eas well (reviewed in Rossman, 1998). However, no sperepair enzyme has been found to be sensitive to the cotrations of arsenite used in these experiments (Li and Ross1989a,b; Huet al., 1998). These observations have led to

ypothesis that the effects on DNA repair by arsenite maresult of faulty DNA damage-inducible signaling wh

ontrols DNA repair (Rossman, 1999). In a test of this hypsis, this laboratory has recently shown that in cells treith arsenite and ionizing radiation the p53-dependent inc

n p21 expression, normally a block to cell cycle progresfter DNA damage, is deficient (Vogt and Rossman, 20his is expected to lead to faulty DNA repair. In addition,nd others have found that low (nontoxic) exposure to arsnhances positive growth signaling (Germolecet al., 1997;archowskyet al., 1999; Troubaet al., 2000; Chenet al.,001; Vogt and Rossman, 2001). We suggest that the abf normal p53 functioning along with increased posirowth signaling in the presence of DNA damage may resefective DNA repair and account for the comutagenic eff arsenite.Based on arsenite’s comutagenicity, we have previo

ypothesized that arsenite might act as a cocarcinogexample in the case of skin cancer with UVR as carcinogtimulus (Li and Rossman, 1991). Here we show that arsoes indeed act as a cocarcinogen by potentiating the abiolar UVR to induce skin cancer in hairless mice.

METHODS

Animals. Female hairless mice (Crl: SK1-hrBR) aged 3 weeks at thbeginning of the experiment, were purchased from Charles River Labora(Portage, MI). They were housed 5 to a cage in solid-bottom plastic cagewire mesh covers and fed autoclaved Purina lab chow 5001 and distilledad libitum. Room illumination was on an automated cycle of 12 h light12 h dark, and room temperature was maintained at 22–25°C. All expericonducted were approved by the Institutional Animal Care and Use Comof New York University School of Medicine.

A total of 40 mice were divided into four different groups as follows: Gr1, controls (no arsenite, no UVR), 5 mice; Group 2, arsenite only, 5 mGroup 3, UVR only, 15 mice; Group 4, arsenite1 UVR, 15 mice.

Arsenite and UV exposure. Mice in Groups 2 and 4 received sodiarsenite (Sigma Biochemicals, St. Louis, MO) via drinking water at a contration of 10 mg/l beginning at 21 days after birth. This corresponds to;5770mg/l arsenic. Water consumption was determined at weekly intervalsweight of each mouse was determined at approximately 5-week interva

Three weeks after the start of the arsenite (or control water) exposurein Groups 3 and 4 were irradiated with a bank of four parallel Westinghsolar FS-20 lamps three times per week at a dose of 1.7 KJ/m2. According tomanufacturer’s specification, 85% of the lamp output was in the UVB (2

)tsf

Re is

ctscn-

an,ee

-edsen).

ite

nce

ints

lyforiciteof

iesithter

ntsee

e;

n-

he

icee

–

320 nm) range,,1% in the UVC (,290 nm) range, 4% in the UVA (320–4nm) range, and the remainder was in the visible (.400 nm) range. The UVdose rates were measured with a IL1400A digital radiometer/photoequipped with a SEL240 UVB-1 detector (International Light, Inc., Newbport, MA). The UVR dose chosen was reported to be approximately 50%minimal erythemic dose (MED) for these mice (Hallidayet al., 2000). Toavoid intensity loss, the UVR dose was checked every month, andconstant during the experiment by altering the exposure time.

Tumors. The mice were monitored at various times during the experifor the presence of skin tumor-like lesions visible to the naked eyenumber of tumor-like lesions of diameter equal to or larger than 1 mmpersisting at least 2 weeks after appearance was recorded for eachseparately. Tumor yield was calculated at each time as the number of tper mouse. The mice were euthanized 182 days after the start of the Ucarbon dioxide asphyxiation, and all tumors were biopsied for analysisvolumes of all tumors was calculated using the standard formula leng3width 3 height3 0.5236.

Pathological analysis. Collected tumor-like lesions were fixed with fomalin, processed, and embedded in paraffin. Five-micrometer-thick sewere stained using hematoxylin and eosin (H and E). Tumors were claas follows: squamous cell carcinomas (SCC) had nests of atypicalinvading into the dermis or deeper subcutaneous tissue; the term keratinintraepidermal neoplasia was used for small tumors with acanthotichyperplastic growth of the epidermis with cells of varying degrees of acality, but no invasive growth through the basement membrane into the d(Cockerell, 2000); papilloma was applied to papillomatous growths of epmal cells without cellular atypicality or invasion of epidermal tumor cellsdermis and either a sessile or exophytic appearance (Bogovslei, 1994)sarcoma refers to malignant tumors of connective tissue; hyperplasia renonmalignant excessive growth of tissue with varying degree of ceatypicality without invasion. Minimally invasive and highly invasive Swere distinguished. A tumor was considered to be minimally invasiveonly one or a few invading strands of cells had broken through the basmembrane.

Statistical analysis. Differences in the total number of tumors was teby the Wilcoxon rank–sum test and differences in their histological gwere analyzed by Fisher’s exact test. Differences in the distribution of tvolumes were analyzed by thex2 test. The distribution of time-to-first-tum

as analyzed by the Kolmogarov–Smirnov test.

RESULTS

The concentration of arsenite (10 mg/l) used in the drinwater had no effect on the growth of the mice, comparecontrol mice (Fig. 1). The mice drinking the arsenite-conting water consumed a similar amount of water comparedmice drinking water without arsenite, and appeared eqactive and healthy. There were no deaths in any group offor the duration of the experiment.

The dose of UVR was chosen to be approximately onethe MED for these mice (Hallidayet al., 2000). A slighreddening of the skin was observed after the first UVR esure, but not in subsequent exposures. No effect of the UVgrowth, physical activity, or general health was seen.

No skin tumors were seen in control mice (Group 1) omice exposed to arsenite alone (Group 2), confirming theof carcinogenicity by oral arsenite alone observed in ostudies. The first tumors in the combined arsenite1 UVRgroup (Group 4) occurred at 8 weeks after the start of theexposure (Figs. 2a and 3), whereas the first tumors in

Throuesrstwant

( VRa mi af1 rou( VRa n thU

ski( ioniG ton eacm rou4 , 65 pc es

.hadmmg 10orseeks

ned

mord

5 rt ofU ne.

ret ) thata 11w camel or laterc t week1 ph wast

allywer

66 ROSSMAN ET AL.

given UVR alone (Group 3) did not appear until week 12.earlier appearance of tumors in Group 4 compared with G3 is statistically significant. The Kolmogorov–Smirnov Twas used to test for equality of the distribution of time-to-fitumor in mice from Groups 3 and 4. The null hypothesisrejected (P 5 0.0018).Although at the end of the experime26 weeks of UVR) every mouse in both the combined or Ulone group had at least one tumor, the time until 100% tu

ncidence was also much reduced in the combined group;9 weeks of UVR exposure, 100% mice in the combined gGroup 4) had tumor compared to 33.3% mice given Ulone (Group 3). The percentage of tumor-bearing mice iVR alone group took 25 weeks to reach 100%.Tumors developed only in the UVR-exposed regions of

i.e., backs) of the mice in either group. Nine tumor-like lesn the UVR alone Group 3 and 21 in the UVR1 arseniteroup 4 regressed and were not included in the count. Theumber of tumors in Group 3 was 53 (the numbers forouse are 4, 4, 2, 4, 2, 2, 6, 3, 5, 2, 6, 5, 2, 3, 3) and in Gwas 127 (the numbers for each mouse are 9, 3, 3, 6, 19, 15, 12, 3, 9, 8, 12, 9). The hypothesis that these 2 saman be samples from the same random distribution was t

FIG. 1. Growth of female weanling (21 days) Skh1 mice continuexposed or not to 10 mg/l sodium arsenite in drinking water. Miceweighed at various times and the mean6 standard errors were calculated (N 520 for each group).

ep

t-s

orterp

e

ns

talhp

, 8,lested

by the Wilcoxon rank–sum test and rejected (P 5 0.0006)Eighteen of the 127 tumors in Group 4 (in 11 mice)extremely rapid growth rates, often reaching a 4- to 5-diameter within less than a week, and eventually exceedinmm in diameter. Two mice in this group developed tumwith ulcerating centers necessitating euthanasia after 18 wof UVR. Thus, the total number of tumors in the combitreatment group (127) is underestimated (Table 1).

The tumor yields over time are shown in Fig. 3. The tuyield in mice exposed to arsenite1 UVR was increase

.3-fold at week 18 and 2.4-fold at week 26 after the staVR, compared with the yield in mice exposed to UVR alo

FIG. 2. Early tumors in mice exposed to UVR1 arsenite. (a) These wehe first two tumors (later diagnosed as squamous cell carcinomasppeared after 8 weeks of UVR1 arsenite. Photograph was taken aftereeks of UVR exposure. (b) The same tumors (top and bottom mice) be

arge and ulcerating by 14 weeks. The center mouse shows a small tumlassified as keratinocytic intradermal neoplasia (arrow), that appeared a1 and remained almost the same size at week 14 when this photogra

aken.

e

, arners

if-

hiss 3ig

5 n ot moi sw eret ) ot diff inv inU erei e

1 es. 5cupsnot

mini-ouper-ma-

aug-lessted)celer-s in

tenti-d by

withse dataen fores of

that, noice.

me-

mayalsoandoterle rat

S

K

timd

a r

67ARSENITE IS A MOUSE COCARCINOGEN

The final volumes of all tumors, regardless of tumor typedepicted in Fig. 4. It is clear that the mice in the combitreatment (arsenite1 UVR) group had much larger tumothan those treated with UVR alone. Based on thex2 test, thetwo distributions of final tumor volumes are significantly dferent (P , 0.0001).

All tumors observed in this experiment were examinedtopathologically. The types of tumors found in Group(UVR) and 4 (UVR1 arsenite) are shown in Table 1 and F

. There were no significant differences in the distributioumor types except for a higher malignant phenotype of tun the combined arsenite1 UVR group. Most of the tumorere SCC with variable degrees of invasiveness and diff

iation. In the combined treatment group, 64 of 127 (50.4%he tumors were highly invasive SCC (moderate to wellerentiated) exhibiting typical keratin pearls (Fig. 5a) andasion into deeper structures (Figs. 5a and 5b), whereasVR alone group only 14 of 53 (26.4%) of the tumors w

nvasive SCC. A large number of small tumors (volum

FIG. 3. The tumor yield per mouse. Tumors were scored at eachperiod and number of tumors per mouse were calculated to give theshown. Error bars are standard deviations from the mean.N 5 15 for UVR

lone.N 5 15 for UVR 1 arsenite up to week 18, andN 5 13 thereaftebecause 2 mice with ulcerating tumors were euthanized.

ed

-

.frs

n-f--the

s

–100 mm3) from both the arsenite1 UVR group and thUVR alone group showed the characteristics of SCC (Figand 5d). In addition, a number of tumors from both grohave characteristics of SCC with invasive cancer cellsgoing through deeper tissues. These were classified asmally invasive SCC (Fig. 5e) (38 of 127 in the combined grand 25 of 53 in the UVR alone group). Keratinocytic intradmal neoplasia (Fig. 5f), papilloma, fibrosarcoma, and prelignant hyperplasia were also seen (Table 1).

DISCUSSION

Results of this study clearly demonstrate that arsenitements the carcinogenic effect of UVR in the skin of the hairmouse. It was recently reported that in Tg.AC (H-ras mutatransgenic mice, arsenite can act as a co-promoter by acating the onset and growth rate of benign skin tumorcombination with TPA as a promoter (Germolecet al.,1998).This report is the first demonstration that arsenite can poate the onset and growth of malignant skin tumors inducea genotoxic carcinogen in a strain of mice that is wild typerespect to oncogenes and tumor suppressor genes. Thesupport the hypothesis that arsenite acts as a cocarcinogthe mouse skin, when given along with carcinogenic doschronic UVR.

No tumors occurred on the skin of control mice or micereceived only arsenite in the drinking water. Upon autopsytumors were noted in any other organs in any of the mRecently, it was found that trivalent monomethyl arsenictabolites are more toxic than arsenite (Stybloet al., 2000;Petrick et al., 2000), suggesting that these metabolitesplay an important role in arsenic carcinogenesis. It hasbeen shown that high concentrations of the herbicidemetabolite dimethylarsinic acid can act as a tumor promand possibly as a complete bladder carcinogen in the fema

TABLE 1Histopathological Analysis of Tumors

Tumor type

Number of tumors (%)

SignificanceaUVR alone UVR1 arsenite

Squamous cell carcinoma(highly invasive)

14 (26.4) 64 (50.4) P 5 0.003

quamous cell carcinoma(minimally invasive)

26 (49.1) 38 (29.9) P 5 0.017

eratinocytic intraepidermalneoplasia

10 (18.9) 16 (12.6) NSb

Fibrosarcoma 1 (1.9) 1 (0.79) NSPapilloma 2 (3.8) 4 (3.1) NSHyperplasia 0 (0) 4 (3.1) NS

Total 53 127

a Fisher’s exact test.b Not significantly different (P . 0.05).

eata

cainaiatthup

acat tinken

ths30

o 340of

thnlywethastim

nsis-thatandairageseduldo in-rgerxper-ationndar-

enic-e lighte unex-p fromtw ingk d sur-f Them a thatc heret ctionf onlys -ChiC osedp ands entss aredi ha n is2T enw orea senge ratioo tantm ing,b coverm n inw VR.W , thel pen-e bep esulto hatt y.I evereD elld

ini con-

dsbm

68 ROSSMAN ET AL.

(reviewed in Kenyon and Hughes, 2001). Since micemethylate arsenite to dimethylarsinic acid as the major urmetabolite (Vahter, 1981), the obligate trivalent intermedmonomethylarsonous acid must also be formed. It ispossible that some or all of these arsenic metabolites maya role in the arsenite cocarcinogenesis seen here, but the ltumors at sites other than UVR-treated skin is evidence thamounts of metabolites formed from 10 mg/l arsenite in dring water were not sufficient to act as complete carcinog

The sodium arsenite concentration (10 mg/l) used indrinking water corresponds to;5770 mg/l arsenic, which ionly 4.4 times higher than the highest concentration (1mg/l) found in Nevada drinking water (Warneret al.,1994) and

nly 1.7 times higher than the highest concentrations (mg/l) found in drinking water in the West Bengal regionIndia (Guha Mazumderet al., 1998).

When the experiment ended at 182 days, tumors inarsenite1 UVR group outnumbered those in the UVR ogroup by 2.4:1. Tumors in the combined treatment groupmuch larger (Fig. 4) and more invasive (Table 1, Fig. 5)in the UVR alone group, suggesting an arsenite-induced

FIG. 4. Final volume (mm3) of skin tumors. Comparison of tumor yiely tumor volume in Group 3 (UVR alone) and Group 4 (arsenite1 UVR)ice. Tumor yields and volumes were assessed at time of euthanasia.

nryes

layk ofhe-

s.e

0

0

e

ren

-

ulation of the growth rate of the SCCs. These data are cotent with, but do not constitute proof of, our hypothesisarsenite acts as a cocarcinogen by inhibiting DNA repairup-regulating positive growth signaling. Inhibiting DNA repshould result in a higher rate of conversion from DNA dam(in this case by UVR) to mutations, leading to increatumorigenicity. Up-regulating positive growth signaling shoalso lead to increased mutagenesis and, in addition, tcreased tumor growth. Both greater tumor yield and latumor size are seen in mice treated with arsenite in the eiments reported here. Preliminary data show the up-regulof cyclin D1 in the skin of mice treated with arsenite, aexperiments to demonstrate inhibition of DNA repair bysenitein vivo are planned.

It is often stated that the skin cancers seen in arsxposed individuals differ from those associated with sunxposure in that arsenic-associated cancers develop in “osed” parts of the body. This statement is based on data

he southwest Taiwan arsenic endemic area (Yehet al., 1968)here it was claimed that 48.14% of “lesions” (includeratosis as well as cancers) appeared on “unexposeaces,” defined as “palms, soles, extremities, and trunk.”ost frequent site of lesions in males was the back, an are

annot be said to be unexposed in a subtropical region whe main occupations are farming, fishing, and salt produrom sea water, and where men usually work wearinghorts; children also dress in little more than shorts (Lunghen, personal communication). If we consider as unexparts of the body the axilla, groin, genitals, buttock, palm,ole (the latter two because the thick keratin layer prevunlight penetration), then only 5.2% of skin cancers appen unexposed parts of the body in men (calculated from Yeetl., 1968). The frequency of skin cancers in this populatio.9-fold higher in males than in females (Tsenget al., 1968).he explanation often given for this difference is that mork outside more than women, and therefore may drink mrsenic-contaminated water. However, as pointed out by Tt al. (1968), this raises the question of why there is a sexf 1:1 for hyperpigmentation and keratosis, two imporarkers of arsenic exposure. Some women work at farmut because fair skin is seen as desirable, women usuallyore of their bodies and wear hats. Body clothing worarm climates does not generally completely block Uhile the average men’s shirt transmits 20% of solar UVR

ight gauzy weaves preferred by women may allow 50%tration (World Health Organization, 1979). It should alsoointed out that in order for human skin cancers to be the rf combined arsenic1 UVR exposure, it is not necessary t

he cancers appear on themostUVR-exposed parts of the bodn fact, an argument can be made that in the case of sNA damage, inhibition of DNA repair might lead to ceath, and not to increased tumorigenesis.In addition, UVR may not be arsenic’s only partner

nducing skin cancer. There are a number of important

r

cinoma

gnant cells(300at were

69ARSENITE IS A MOUSE COCARCINOGEN

FIG. 5. Histopathology of the tumors (H and E-stained paraffin sections). (a) One of the squamous cell carcinomas shown in Fig. 2 (volume.1000 mm3)appearing in a mouse in the combined arsenite1 UVR group after 8 weeks of exposure. A large keratin pearl is shown by the arrow (303). (b) Higher powemagnification of (a) showing typical keratinizing pearls (arrows) and invasion of tumor cell nests into subcutaneous structures (arrowheads) (1503). (c) Onesmall tumor (volume5 15 mm3) from the arsenite1 UVR group showing two small keratin pearls (arrow), a typical characteristic of squamous cell car(1503). (d) Higher magnification of (c) showing small strands of atypical cells invading the dermis (arrowheads) (3003). (e) A minimally invasive SCC froma mouse in the UVR alone group. Unlike the large SCCs shown in (a) and (b), this minimally invasive SCC had only a few small strands of maliinvading the epidermis (arrowheads), while the remainder of the tumor did not clearly break through the basement membrane (not shown)3). (f)Keratinocytic intraepidermal neoplasia from a mouse in the UVR1 arsenite group. These, usually small, tumors consisted of atypical epidermal cells thentirely confined to the epidermis and did not show any evidence of dermal invasion through the basement membrane (603).

cepulain tmaceg

nis otri-2,

e ag( owt2 oti 86

ab gea thec lunc osw

iwenrsce

skecdet

wcitth

ssmges

ThentEnvancatiope

Pea

ly to

mitlat

ress

mmon

B fromocar-

B in5,

C antaiwan:.

C entsns of

C -L.,rsenicsis-

C s cell

E rurce

1).

G sky-is, F.roni-

G y of

G , T.,t, R.,can

rived

G mble,ster,ation

G hak-and

H picalsolar-

H R.,tide

H , G.king in

H E.,ated

H en,skin

70 ROSSMAN ET AL.

founding variables that must be considered. Genetic susbility to arsenic carcinogenesis may vary in different poptions (see discussion in Gebel, 2000). Other carcinogensenvironment, such as polycyclic aromatic hydrocarbons,also be involved in skin as well as lung and bladder can(Boffetta et al., 1997). Humic acid byproducts in drinkinwater in Taiwan may contribute to its carcinogenesis (Luet al.,1986). Skin cancer rates in Taiwan are higher for arseexposed individuals who work in salt fields, are carrierhepatitis B virus with liver dysfunction, or have poor nutional status (Hsuehet al.,1995). Deficiencies in vitamin B1folate, vitamin B6 (pyridoxine), niacin, ascorbate,a-tocoph-

rol, iron, or zinc are associated with increased DNA damAmes, 2001). Both folic acid and selenium have been sho protect against arsenic toxicity (Gebel, 2000; Ruanet al.,000) and selenium protects against arsenic-induced gen

city in human lymphocytes (Beckman and Nordenson, 19Since arsenic exposure is also associated with lung

ladder cancers in humans, it is possible that a cocarcinorsenic exposure protocol may lead to animal models forancers as well. Arsenic appears to act synergistically inarcinogenesis with tobacco use in occupationally exporkers (Pershagenet al., 1981; Hertz-Picciotoet al., 1992),

and cigarette smoking in the arsenic endemic area of Tawas 40% (Chenet al.,1985). Cigarette smoking was indepdantly associated with increased risk of urinary tract canceTaiwan (Chiouet al.,2001). Arsenic also enhances lung canrisk from radon gas in tin miners (Xuanet al., 1993).

The development of an animal model for arsenic-relatedcarcinogenesis now makes it possible to evaluate the mnisms underlying the dose/response relationship and tomine the proper model for risk assessment. Animal modelsalso allow molecular analysis of arsenic’s cocarcinogenidetermination of the role of specific environmental agentsmay act as arsenite’s partners in cocarcinogenesis, asseof the effects of dietary deficiencies in arsenite cocarcinoesis, and for development of chemopreventive strategiearsenic-exposed populations.

ACKNOWLEDGMENTS

This work was supported by NIEHS Grant ES09252 and is part ofNelson Institute of Environmental Medicine and the Kaplan Cancer Cprograms supported by Grant ES00260 from the National Institute ofronmental Health Sciences and Grant CA16087 from the National CInstitute. We thank Eleanor Cordisco for help with the manuscript preparthe Nelson Institute Animal Care and Histopathology staff for their excontributions, Arthur Nadas for help with statistical analysis, and M. J.for his encouragement.

REFERENCES

Ames, B. N. (2001). DNA damage from micronutrient deficiencies is likebe a major cause of cancer.Mutat. Res.475,7–20.

Barchowsky, A., Roussel, R. R., Klei, L. R., James, P. E., Ganju, N., SK. R., and Dudek, E. J. (1999). Low levels of arsenic trioxide stimu

ti--heyrs

c-f

en

ox-).ndnicseg

ed

an-inr

inha-er-illy,atent

n-for

eeri-ern,rtk

h,e

proliferative signals in primary vascular cells without activating steffector pathways.Toxicol. Appl. Pharmacol.159,65–75.

Beckman, L., and Nordenson, I. (1986). Interaction between some cogenotoxic agents.Hum. Hered.36, 397–401.

offetta, P., Jourenkova, N., and Gustavsson, P. (1997). Cancer riskoccupational and environmental exposure to polycyclic aromatic hydrbons.Cancer Causes & Control8, 444–472.

ogovslei, P. (1994). Tumours of the skin. InPathology of the TumoursLaboratory Animals(Turosov, V. S., and Mohr, U., Eds.), Vol II, pp. 1–4IARC Scientific Publication No. 111, IARC, Lyons, France.

hen, C.-J., Chuang, Y.-C., Lin, T.-M., and Wu, H.-Y. (1985). Malignneoplasms among residents of a blackfoot disease endemic area in THigh-arsenic artesian well water and cancers.Cancer Res.45, 5895–5899

hen, H., Liu, J., Merrick, B. A., and Waalkes, M. P. (2001). Genetic evassociated with arsenic-induced malignant transformation: ApplicatiocDNA microarray technology.Mol. Carcinogen.30, 79–87.

hiou, H.-Y., Chiou, S.-T., Hsu, Y.-H., Chou, Y.-L., Tseng, C.-H., Wei, M.and Chen, C.-J. (2001). Incidence of transitional cell carcinoma and ain drinking water: A follow-up study of 8,102 residents in an arseniaendemic area in Northeastern Taiwan.Am. J. Epidemiol.153,411–418.

ockerell, C. J. (2000). Histopathology of incipient intradermal squamoucarcinoma (“actinic Keratosis”).J. Am. Acad. Dermatol.42, 11–17.

nvironmental Protection Agency (2000).National Primary Drinking WateRegulations; Arsenic and Clarifications to Compliance and New SoContaminants Monitoring.Proposed rule (Federal Register 65, No. 12

arcia-Vargas, G. G., Del Razo, L. M., Cebrian, M. E., Albores, A., OstroWegman, P., Montero, R., Gonsbatt, M. E., Lim, C. K., and De Matte(1994). Altered urinary porphyrin excretion in a human population chcally exposed to arsenic in Mexico.Hum. Exp. Toxicol.13, 839–847.

ebel, T. (2000). Confounding variables in the environmental toxicologarsenic.Toxicology144,155–162.

ermolec, D. R., Spalding, J., Boorman, G. A., Wilmer, J. L., YoshidaSimeonova, P. P., Bruccoleri, A., Kayama, F., Gaido, K., TennanBurleson, F., Dong, W., Lang, R. W., and Luster, M. I. (1997). Arsenicmediate skin neoplasia by chronic stimulation of keratinocyte-degrowth factors.Mutat. Res.3386,209–218.

ermolec, D., Spalding, J., Yu, H.-S., Chen, G. S., Simeonova, P. O., HuM. C., Bruccoleri, A., Boorman, G. A., Foley, J. F., Yoshida, T., and LuM. I. (1998). Arsenic enhancement of skin neoplasia by chronic stimulof growth factors.Am. J. Pathol.153,1775–1785.

uha-Mazumder, D. N., Haque, R., Ghosh, N., De, B. K., Santra, A., Croborty, D., and Smith, A. H. (1998). Arsenic levels in drinking waterthe prevalence of skin lesions in West Bengal, India.Int. J. Epidemiol.27,871–877.

alliday, G. M., Robertson, B. O., and Barnetson, R. St C. (2000). Toretinoic acid enhances, and a dark tan protects, from subedemalsimulated photocarcinogenesis.J. Invest. Dermatol.114,923–927.

artwig, A., Groblinghoff, U. D., Beyersman, D., Natarajan, A. T., Filon,and Mullenders, L. H. F. (1997). Interaction of arsenic(III) with nucleoexcision repair in UV-irradiated human fibroblasts.Carcinogenesis18,399–405.

ertz-Picciotto, I., Smith, A. H., Holtzman, D., Lipsett, M., and Alexeeff(1992). Synergism between occupational arsenic exposure and smothe induction of lung cancer.Epidemiology3, 23–31.

openhayn-Rich, C., Biggs, M. L., Fuchs, A., Bergoglio, R., Tello, E.Nicolli, H., and Smith, A. H. (1996). Bladder cancer mortality associwith arsenic in drinking water in Argentina.Epidemiology7, 117–124.

sueh, Y.-M., Cheng, G.-S., Wu, M.-M., Yu, H.-S., Kuo, T.-L., and ChC.-J. (1995). Multiple risk factors associated with arsenic-inducedcancer: Effects of chronic liver disease and malnutritional status.Br. J.Cancer71, 109–114.

anpai

eml

ctsmi

. Etionstu

an

esf u

L Cality

L by

sis

L e w.

L shipte

M eth

N -

P actiship

P hianite

R c. I–

R ren

lth, pp.

R ., andasts

S L.,D. J.and

.

T . H.,king

T nite-

T h, S.nicism

V nioc

V 1 andm for

W ic. Iness,

W andls ofada.

W enicdL.,han-

W -

X ng,iners

Y kin:

71ARSENITE IS A MOUSE COCARCINOGEN

Hu, Y., Su, L., and Snow, E. T. (1998). Arsenic toxicity is enzyme specificits affects on ligation are not caused by the direct inhibition of DNA reenzymes.Mutat. Res.408,203–218.

IARC (1980).Monographs on the Evaluation of Carcinogenic Risk of Chicals to Man: Some Metals and Metallic Compounds,Vol. 23. InternationaAgency for Research on Cancer, Lyon, France.

Kanisawa, M., and Schroeder, H. A. (1967). Life term studies on the effearsenic, germanium, tin, and vanadium on spontaneous tumors inCancer Res.27, 1192–1195.

Karagas, M. R., Stukel, T. A., Morris, J. S., Tosteson, T. D., Weiss, JSpencer, S. K., and Greenberg, E. R. (2001). Skin cancer risk in relatoenail arsenic concentrations in a US population-based case-controlAm. J. Epidemiol.153,559–565.

Kenyon, E. M., and Hughes, M. F. (2001). A concise review of the toxicitycarcinogenicity of dimethylarsinic acid.Toxicology160,227–236.

Lee, T.-C., Huang, R. Y., and Jan, K. Y. (1985). Sodium arsenite enhanccytotoxicity clastogenicity and 6-thioguanine-resistant mutagenicity otraviolet light in Chinese hamster ovary cells.Mutat. Res.148,83–89.

ewis, D. R., Southwich, J. W., Ouellet-Hellstrom, R., Rench, J., andderon, R. L. (1999). Drinking water arsenic in Utah: A cohort mortastudy.Env. Health Perspect.107,359–365.

i, J.-H., and Rossman, T. G. (1989a). Inhibition of DNA ligase activityarsenite: A possible mechanism of its comutagenesis.Mol. Toxicol.2, 1–9.

Li, J.-H., and Rossman, T. G. (1989b). Mechanism of comutagenesodium arsenite withN-methyl-N-nitrosourea.Biol. Trace Element Res.21,373–381.

i, J., and Rossman, T. G. (1991). Comutagenesis of sodium arsenitultraviolet radiation in Chinese hamster v79 cells.Biol. Metals4, 197–200

u, F. J., Guo, H. R., Chiangs, H. S., and Hong, C. L. (1986). Relationbetween the fluorescent intensity of well water and the incidence rabladder cancer.J. Chinese Oncol. Soc.2, 14–23.

ilner, J. E. (1969). The effects of ingested arsenic exposure on mcholanthrene induced skin tumours in mice.Arch. Environ. Health18,7–11.

ational Research Council (2000).Arsenic in Drinking Water.National Academy Press, Washington, DC.

ershagen, G., Wall, S., Taube, A., and Linnman, L. (1981). On the interbetween occupational arsenic exposure and smoking and its relationlung cancer.Scand. J. Work Environ. Hlth7, 302–309.

etrick, J. S., Ayala-Fierro, F., Cullen, W. R., Carter, D. E., and AposH. V. (2000). Monomethylarsonous acid (MMIII) is more toxic than arsein Chang human hepatocytes.Toxicol. Appl. Pharmacol.163,203–207.

ossman, T. G. (1998). Molecular and genetic toxicology of arseniEnvironmental Toxicology: Current Developments(Rose, J., Ed.), pp. 171187, Gordon and Breach Science Publishers, Amsterdam.

ossman, T. G. (1999). Arsenite genotoxicity may be mediated by interfe

dr

-

ofce.

.,tody.

d

thel-

l-

of

ith

sof

yl

onto

n,

n

ce

with DNA damage-inducible signaling. InArsenic Exposure and HeaEffects(Chappell, W. R., Abernathy, C. O., and Calderon, R. L., Eds.)233–241, Elsevier Science Ltd.

uan, Y., Peterson, M. H., Wauson, E. M., Waes, G.-V. J., Finnell, R. HVorce, R. L. (2000). Folic acid protects SWV/Fnn embryo fibroblagainst arsenic toxicity.Toxicol. Lett.117,129–137.

tyblo, M., Del Razo, L. M., Vega, L., Germolec, D. R., LeCluyse, E.Hamilton, G. A., Reed, W., Wang, C., Cullen, W. R., and Thomas,(2000). Comparative toxicity of trivalent and pentavalent inorganicmethylated arsenicals in rat and human cells.Arch. Toxicol.74,1178–1183

ondel, M., Rahman, M., Magnuson, A., Chowdhury, I. A., Faruquee, Mand Ahmad, S. A. (1999). The relationship of arsenic levels in drinwater and the prevalence rate of skin lesions in Bangladesh.Environ. HealthPerspect.107,727–729.

rouba, K. J., Wauson, E. M., and Vorce, R. L. (2000). Sodium arseinduced dysregulation of proteins involved in proliferative signaling.Toxi-col. Appl. Pharmacol.164,161–170.

seng, W. P., Chu, H. M., How, S. W., Fong, J. M., Lin, C. S., and Ye(1968). Prevalence of skin cancer in an endemic area of chronic arsein Taiwan.J. Natl. Cancer Inst.40, 453–463.

ahter, M. (1981). Biotransformation of trivalent and pentavalent inorgaarsenic in mice and rats.Envirn. Res.25, 286–292.

ogt, B., and Rossman, T. G. (2001). Effects of arsenite on p53, p2cyclin D expression in normal human fibroblasts—a possible mechanisarsenite’s comutagenicity.Mutat. Res.478,159–168.

ang, Z., and Rossman, T. G. (1996). The carcinogenicity of arsenToxicology of Metals(Louis W. Chang, Ed.), pp. 219–227, CRC PrBoca Raton, FL.

arner, M. L., Moore, L. E., Smith, M. T., Kalman, D. A., Fanning, E.,Smith, A. H. (1994). Increased micronuclei in exfoliated bladder celindividuals who chronically ingest arsenic-contamined water in NevCancer Epidemiol. Biomarkers Prevent.3, 583–590.

elch, A. H., Helsel, D. R., Focazio, M. J., and Watkins, S. A. (1999). Arsin ground water supplies of the United States. InArsenic Exposure anHealth Effects(Chappell, W. R., Abernathy, C. O., and Calderon, R.Eds.), pp. 9–17, Elsevier, Amsterdam-Lausanne, New York-Oxford, Snon, Singapore-Tokyo.

orld Health Organization (1979).Environmental Health Criteria 14: Ultraviolet Radiation.World Health Organization, Geneva. p. 93.

uan, X. Z., Lubin, J. H., Li, J. Y., Yang, L. F., Luo, A. S., Lan, Y., WaJ. Z., and Blot, W. J. (1993). A cohort study in southern China of tin mexposed to radon and radon decay products.Health Physics64, 120–131.

eh, S., How, S. W., and Lin, C. S. (1968). Arsenical cancer of sHistological study with special reference to Bowen’s Disease.Cancer21,312–339.