Tobacco Documents Online

GA~TR OF,.%'T~RO LOG Y 1992:103;X33~-I34S SPECIAL REPORTS AND REVIEWS Infectious Agents in the Etiology of Esophageal Cancer FU~U CHANG, STINA SYRIANEN, LIJUAN WANG, and KARl SYRIANEN Department.of Pa~holosy and Kuopio Cancer ]~esesz'ch Centre. University of Kuopio. Kuopio, Finland: and Departmen~ of Precancerous Studies. Henan Medic~l Unive~ity. Zhen~zhou, Henan. China Extensive epidemioiogical and experimental stud- ies have su~ested t}iat some chemical agents, nu- tritional defcienci~, and physical factors are asso- ciated with the devdlopment of esophageal cancer (EC). Recent evidenc~ also su~ests an etiologic role of certain mic~oorga.nlsms in esophageal carcino- genesis either by prd~ducing carcinogens or promo- tots or by acting directly on the host cells. The mu- tagenic and carcinogenic effects of several fungi and bacteria isolate~, from the grains and foodstuffs in high-risk areas h~ve been shown by in vitro and in vivo studies, Cert..in v/ruses, e.g,, human papil- lomavirus, herpes simplex virus, cytomegMovirus, and Epstein-Barr virus, have been implicated in the pathogenesis of a ~ariety of human cancers, and all of them are know~ to produce tumors in animals and cell transformatibn in vitro. These viruses also have been shown to iinfect the esophageal epithe- lium. Therefore, although many of the key issues of their mechanisms of bction are unclear as yet, they should be consider~ potential etiologic agents of EC. The present revi .e~ summarizes the data avail- able on the eilology o[EC, emphasizing the current evidence implicatin~ian etiologic role of microor- sanisms in the pathoo~enesis of this mal~suancy. ~"~ne of the most int~guing features of esophageal %.Jcancer (EC) is its qonsiderable geographic va.ria- tion. Data worldwide present a mosaic of changing inc/dence rates and sex rat/os.~'~ In most countries. incidence rates of EC per 100,000 are approximately 2.5-5.0 for males and 1.5-2.5 for females. However, in certain areas the incidence rates are remarkably high, varying up to 500-foId from one area to an- other, from one country to another, and between dif- ferent ethnic groups within the same country.~-~ Epi- demiological studies have identified the high-risk countries for EC: the People's Republic of China.*'s Singapore,~ [ran,~z-~ the former Soviet Union,~" Puerto Rico.m.u Chile,~°-~ Brazil.m'~ Switzer- land.*'°'" France.m-'' and South Africa?:'~° Among these high-risk countries, the highest rates of EC have been reported in the northern parts of China2-' the Caspian littoral of IranY-~s and the Tmnskei area of South A~'ica.~-~° In the People's Republic of China, the annual deaths due to EC account for 27% of all cancer deaths among males and 20% among females, ranking EC among the two leading causes of cancer deaths, sec- ond only to stomach cancer.~'~ In Linxian, a county in the Henan Province of North China with a popula- tion of 800,000, the ase-adiusted average incidence rates were 162.33/10~ in males and 102.88/10s in females ~rom 1971 to 1974.s'~ The deaths due to EC in this area account for 16% of all deaths and 65% oE cancer deaths,s'~ Similar fi~res have been reported in Iran~-~e and South Africa,~ where up to 180/10~ and 246/10s new cases of EC have been recorded per year, respectively. The reasons for these major reg/onal variations in the incidence of EC are l~orly understood, However, it seems clear that no single etiolos/c factor could account for such a dramat/c variation in the ~re- quency of EC in distinct geographic areas. The wide range of the incidence rates between the high- and low-risk areas, as well as the dramatic differences within d/st/oct geograph/c areas between sexes and ethn/c groups, suggest a predominant role for envi- ronmental factors in EC. The esophagus is one of the most frequent mucosal sites to come in contact with environmental factors, and as such it is also a signifi- cant route o~ entry for foreign, often harm~l and noxious agents into the human body. Such poten- tially harmful agents include pathogenicmicroorgan- isms. chemical irritants, environmental pollutants, or food additives. Therefore. it seems likely that the pathogenesis of EC is in some way associated w/th these factors. This is also strongly suggested by the fact that in domestic chickens, which usually share 1992 by the Arner|c~n Gastroenter~lo~ical ..%ssoc'~tion 0016-508S/9~/$3.00 0 0 0"~ Oo BATCo document for Mayo Clinic 28 March 2002

Oct~erl~2 ET~OLOGY OF ESOPHACEAL CANCEK 1337 the food sources and natural environment of their hosts, the incidence Of pharyngeal and EC is lO-fold #eater in Linxian chickens compared with that in chickens in low-inc/~ence areas,s,7.~-~ Large epidemiolo~/cal and experimental studies have suggested that excess/ve consumption of. alco- hol and tobacco, ~xposure to special chemical agents, and presence~ of nutritional deftc/encies are important risk factors for EC.~'~ Substantial evidence has been prov/ded during the past several decades implicating an etiolog/c role ofinfectiousmicroorgan- isms in the pathogen~sis of aerodigestive cancer. No- table examples include the close assoc/ation of herpes simplex virus. (HSV) and human papilloma- virus ~IPV) infections with oral cancer,~'~= HPV in- fection with laryngeal carc/noma,~-°= Epstein-Barr virus {EBVJ infectiop with nasopharyngeaI carc/- noma,=7"~= Helicobacter p~'lori infection with stom- ach carc/noma,~z ~cytomegalovirus (CIvlV] and Schistosomo ]oponi~m infections with colorectal c~ncer,~'=~ snd he~atit.is B virus [HBV) infection and mold Asperg/~lu~s j~ovus contamination with liver carcinomas.~i Similarly, evidence is also ac- cumulating to su~es{ an etiologic role of certain mi- croorgan/~r~ in esop~agealcarcinogenesis. The pres- ent article reviews the current data on the etiology of EC, paying particular attention to the role of microor- gamsms. Current Conce~ts of the Etiology of EC Extens/ve comparisons of the dietary and cul- rural habits of peoplb living in geographically dis- tinct, high-incidencel areas have revealed little/n common to sug~est'a similar pathogenetic reecho- nism of EC. Thus, it ~eeras more likely that each of the high-risk areas h~s its reg/onsl peculiarities that are probablyofetiolo~csignificance. Cigarette smok- . ing and excessive alcohol intake may be the major risk factors in some ~reas, e.g:, Western coun~es and South A~rica, especially when these two ~actors are combined.~-z~ However, these agents do not appear to be major~ risk factors in China and Iron.~v'~z'~ On the other hand, more attention has been focused in China on specific nutritional deft- cienc/es, including those of vitamins A, B. and C and certain minerals as well as nitresamines formed in moldy foodstuffs,~-~.~x In Iran, similar nutritional deficiencies were also noted, and opium tar was also blamed as a contributing factor.~z-" Among the chemical carcinogens related to the de- velopment of EC, nitrosamines and nitmsamides or ~'-nitroso compounds are remarkable for their abil- ity ~o induce tumors in many organs.~=-=-:x"-*s-~= On the basis of the results of animal experiments, sev- ersl nitrosamines are known to produce malignant tumors in the esophagus. Human expose:re ro nitro- samines and their precursors may come [tom in- gested lroods, drinking water, the volatile fract/on of tobacco smoke, and industrial emissions.~ N-Nitroso compounds can also be formed endogenously by the reaction of nitrite with secondary or tertiary amines.~='*~= Nitrite is commonly forr~ed by bacte- rial or plant enzymatic reduction of exogenous ni- trate, the most common natural ~orm of n/trogen. Pickled and stored vegetables, cured meats and fish, a variety of vegetables, and alcoholic beverages are the major sources of ingested nitrates.~.~'~z The ecologic factors that can enhance the formation of N.nitroso compounds include microorganism con- tamination in food and water, low molybdenum con- tent in soil and food, and z/no deftc/ency. Increased r/sk of exposure to nitmsamines in the development of" EC has been reported in several high-inci- dence areas, including China,s~.~'~ Iron,~z-" South Africa,~'~ and the Kashmir region of India,~ etc. Opium abuse has been reported to be closely re- lated with the local variations o~EC in Iran. Tests for urinary morphine metabolites indicated that the ma- jorlty of the adult population are opium addicts in the high-incidence areas oft hat country.~-~-~ Bacte- rial mutagenic/ty assays showed mutagenic activity of opium residues,xs Furthermore, the tarry residues from the opium pipes are frequently eaten. ~imilar practices of ingesting opium tar have also been found in the Transkei area of South Aft'ira. A poor diet, comprising mainly carbohydrates with a low intake of an/real protein, fruits., and green vegetables, has repeatedly been found to be asso- c/areal with EC in both high- and low-inc/dence areas.~'xx~-z°-~ Similarly, an imbalanced d/et has also been shown to be a risk factor for EC.~s~ Low molybdenum content in soil and food has been sug- gested as a contributing factor to EC in several high- incidence areas.~ Biochemically. molybdenum is a cofactor of the enzyme nitrate reductase, which af"- fects nitrite and nitrate content in plants.~ Animal stud/es show that the addition of ammonium molyb- date could decrease gastric adenocarc/nomas asso- ciated with exposure to N-nltroso compounds,s More- over, z/no deficiency has been shown to increase nitrosamine-induced EC in rats.~= P3boflavin deft- ciency may be another important nutritional factor in the development of EC, because it can produce epithelial changes in various experimental animals and in humans. It has been shown to increase the mitotic rate of" the skin as well as the buccal and esophageal mucosa of baboons.~" l~iboflav/n de~- ciency is known to affect the rate of carcinogen me- tabolism. Flavocoenzymes are essential constituents of the mixed-function oxidase system, which is re- sponsible for the detoxification and activation o~car- cinogens.~'~° Ano-:her vitamin that has long been in- BATCo document for Mayo Clinic 28 March 2002

~a C~NG~7, AL G.-LSTROENTEROL(3GY Vo[. 103..so. ~ criminated is vitamin~ A. The deficiency of this vita- min has been known (~ be assoc/ated with hyperker- atin~ation, metaplbsia, dysplasia, and the appearance of grossI epithelial .tumors.=t'~ Such changes have been s~ggested to be involved in the development ofEC.=" .:~'~ Vitamin C may have a role in the metabolism of nitrates as well and therefore, when in insufficient s.:upply in the body, may be im- pliceted in the pathogenesis of EC.~ It seems un- likely that nutrition~l deficiencies could directly cause EC. However. ~or nutrition of the epithelial cells may alter the in.:~egrity of the esophageal mu- cosa and may thus in~/rectly enhance the suscepti- bility of the epitheliurh to a variety of infectious mi- croorganisms and chemical carcinogens. Epidenziolog/cal stu~l/es showed that people in the high-risk areas of China, Iron, and Brazil customarily eat food or drink bevdrages that are hot.~-~ In one survey in Linxian, approximately 77% of the inhabit- ants regularly enjoyed, their food at temperatures of 60-70°C or higher [8.0--~8°C}.~ In animal experi- ments, thermal inju~ has been shown to increase the inc/dence of esopl~ageal tumors and shorten the induction time of chemical carcinogens/~ Although the intake of hot food does not seem to have a carcin- ogenic effect per se. it ~ possible that thermal injury may lead to nonspec~c chronic esophagitis that not associated with g~." tric reflux or eros/on. Endo- scopic studies have shbwn that such a chronic eso- pha#t.is is the most frequent finding in the h/gh-risk populations of China i80%}~'~ and Iran {70%}.n'~ The consumption of b~verages at burning tempera- tures has been reporte:d as the strongest risk factor for the development of.bhronic esophagitis.~'~s Thus, there are good reasons !o presume that this condition is a precursor lesion o~ EC.':-'~ Fungal Contamination and Infection in the Etiology of EC ~ Fungul Contamination of Foodstu~s in areas where EC is prevalent, grains and foodstuffs are frequently contaminated with fungi.~.~'~'-= Studie~ in the high-risk areas for EC in China showed that some common species of gus belonging to Fusorfum, AJternar/a, Geotrichum, Aspergiilus, Cladospodum, Penicillium. and other genera not only could reduce nitrates to nitrites but also could decompose proteins and increase the amount of amines in food, consequently promoting the formation of nitrosamines.*'~'~'~'~'~ Fusar/um and Alternorio are the two most common fungi that grow on maize, millet. ~nd other grains cultivated in Linxian. A higher lev*l of nitrosamines has been identiAed by ~as and mass chromatography in grains or food sample.~, e.g.. wheat, corn. millet, m. tiler bran. dried sweet potato, and pickled vegetables, from the high-incidence areas in China.=~'~-~ Similarly. a 17- fold increase in secondary amines was noted in sarium moniliforme-incubated cornbread. When a small amount of NaNOz was added to moldy corn- bread, which was then incubated for 8 days. several N-nitroso compounds, e.g., dimethylnitrosamine, diethylnitrosamine, methylbenzylnitrosamine, and dimethylthiotetranitrosod/iron, could be de- tectedY~-~'s~'~ Laboratory tests carried out both in vitro and in vivo have shown that F. monili/orme and Alternoria alternata have mutagenic and carcino- gen/c effects as shown by the sister chromatid ex- change test, V 79 cell mutation, and Ames test with extracts of the fungi. Furthermore, forestomach and esophageal papillomas and carcinomas were Induced in rats fed fungus-contaminated cornbread.~'s~ Another dietary habit of the high-risk popuIatious in China and some other areas of the world is the heavy consumption of a special variety of pickled vegetables.~'r These pickled vegetables are prepared by placing chopped, blanched leaves of Chinese cab- bage, turnip, soybean, sweet potato, sesame, and other vegetables in a water container to be fer- mented for several months. The final products are regularly covered with white mold. Geotrichum can- didum/s the predominant species of fungus occur- ring in the pickled vegetables, and additional con- tamination by A. flavus, F. moniliforme, Aspergillus fumigatus, and Aspergillus nidulons is a regular oc- currence as wells-r'4~-s~ Epidemiological studies in China have shown that the absolute amounts and the number of months per year of consumption of the pickled vegetables were directly correlated w/th the mortality rate o[ECY~'~ In vitro experimental studies have shown that the extracts of Linxian pickled vegetables were highly mutagenic as showzi by the sister chromatid ex- change test, V 79 cell mutation..Ames test, and counting of transformed loci in culture of Syrian hamster embryonic cells.~'~'~'~z'~'s-~'m Experimen- tal studies showed that the daily gastric incubat/on ofG. candidum culture fluid in mice or rats induced various hyperplastic and dysplastic lesions of the forestomach and esophagus, and. when adminis- tered simultaneously, the fungus culture fluid en- hanced the carcinogenic action of methylbenzylni- trosamine in the forestomach2#'*~'~-~'~z Firm evidence from the high-risk areas of South A~ca and France indicate fungal contamination of grains and foodstuffs. In South Africa. maize grown in poor soil with zinc and molybdenum deAciencies has been identified in the high-risk areas by agricul- tural surveys. Molybdenum deficiency increases the susceptibility of maize to a variety of fungai species. especially Fusarium andAspergillu~, which are capa- BATCo document for Mayo Clinic 28 March 2002

O~ober I9~2 ETIOLOGY OF ESOPHAGEAL C.~/CER 1339 hie of producing carcinogenic metabolites?7~'=j Fur- thermore, the maize "is also fermented to make tradi- tional Xhosa beer. in which high concentrations of nitrosamines have bben identified.1" In the high-risk area of France. epidemiological studies have sug- gested that the quantity of ethanol consumed as ap- ple c/dot ls related to::the prevalence of EC. However, it has been suggast~ that the origin of the apples used in making the cider could also be a risk factor, it is usually made from discarded apples, many of which are more or Id.'ss spoiled2 Funga| Infectibns in the Esophagus Of the various infections in the esophagus, fungal infections, pa~'ticularly those of Candida spa. cias, are by far the most common.~'~ In the study by Kodisi et at., Condid.a.induced esophagitis was de- tected in 7.3% of cbnsecutive patients by endos- copy.~ This is not su.'rprising because Candida a/hi- cans, the organism U'sually implicated, is a normal commensal of the mbuth, the oropharynx, and the upper gastrointestin~ tract in humans. Although in- fections can occur i~ otherwise normal individuals, those with predisposing conditions in which host de- fenses are comprom.'.'lsed are more frequently letted.~n'~ The lesio.ns due to Candida are usually confined.to the Iow~r third of the esophagus, but they may be multifodal and diffuse as well, with the gross appearance of l~yperemia and white plaques or membranes on esophageal mucusa.~-m Mucosal biopsy specimens show an acute and chronic inflam- matory reaction in tl~e epithelium and lamina pro- prin, with occasional ..hlceratious and granulation tis- sue. The typical fungal spores and nonseptate pseudohyphae are be~t seen with special stains such as the periodic acid~Schiff reaction aRer diastase treatment or the met'.henamine silver stain. The or- ganism may also be detected by examination of di- rect cytological smea~ taken from the plaques with an endoscopic brush.F In the hlgh-incideni:! e areas in China, fungal inva- sion of the esophagu~ is common. The presence of fungus as determine~i by cytology correlated with the degree of dysplas!a.~'~ For example, esophageal fungus was found in 31% (64/207} of normal people and those with mild dysplasia, in 72% {108/248} of those with severe dysplasia, and in 90% {90/100} of those with EC.~ Fungal invasion of the esophagus was also studied in biopsy specimens and resected specimens. In a series of 185 samples, fungalinvasion was found in 30% of the hyperplastic epithelium in noncancerous patients, in 50% of hyperplastic epi- thelium in early EC p~tients, in lS% of cancer tissue in early EC patients, and in 3.1.% of apparently nor- real epithelium,rr Electron microscopy showed that the fungi invaded the area between esophageal epi- thelial cells as well as the cytoplasm of the cells.: The epithelial cells adjacent to the invading fungi fre- quently showed various degrees of change, ranging from hyperplasia to mild and severe dysplasia to early malignant changes,z The area of infection is normally in the middle third of the esophagus, a cation similar to that of EC. Patients with fungal in- fections are on the average 7 years younger than EC patients.7 As in the general population. Candida spe- cies are the most common invaders of the esophagus in the high-incidence areas as well, and a pure cul- ture of C. albicans can frequently be isolated from -the hypezplastic epithelium and carcinoma in situ of the esophagus. Candida trice|is, Candida kruseL Candida parepsilosis, Torulopsis glabrota, and Toru- Iopsis tomata have also been isolated from oral and esophageal samples,r'~'r~-z~ However, these direct as- sociations of fungal infection and EC remain to be confirmed by other investigators. Being normal com- mensals in the upper aerodigestive tract, fungi may easily contaminate the research materials; therefore, great care should be taken to interpret the observa- tions of fungi in esophageal biopsy samples. Bacterial Contamination and Infection in the Etiology of EC Bacterial Contamination of Drinking Water and Foodstu~s The production of carcinogens, mutagens and promoters by bacteria has not been studied widely, and the data available at present are limited. It is known that N-nitroso compounds are formed by the action of nitrite on a suitable nitrogen compound at acid pH, but the reaction may be catalyzed by bacte- ria at neutral pH.~-~ Consequently, they are formed under the conditions in which nitrite, bacteria, and nitrostable amino compounds coexist. Although there is little nitrite in the diet, it is readily formed by bacterial reduction of nitrate. The high-risk areas in China andSouth Africa gen- erally are situated in arid or semiarid re~ons where water supply has historically been a serious prob- lem.~'~"~'z° In China, although wells are available in certain areas, most people in Linxian traditionally rely on water from "dry wells" or artificial ponds. Dry wells or ponds are used to collect rain water, which is stored for use throughout the year. The water is infested with microorganisms and fre- quently contaminated with refuse from humans and domestic animals.*'~ The nitrate and nitrite contents in dry-well water are often much higher than those in well water.~-~-~ Bacterial Infections in the Esophagus Bacteria are normally present in the upper all- • mentary tract, ir~cb.'ding the esophagus. Thus. they BATCo document for Mayo Clinic 28 March 2002

G.4~-TROENTEROLOGY VoL ~u~. ~o. ~ are well positioned tO interact with dietary compo- nents and promote the format/on of rdtrosamines. Nitrate-balance studi~ have shown that dietary ni- trate is rapidly absorbed from the upper small intes- tine and secreted prln~ipally in the saliva and gastr/c secretions, reaching ~aximum levels at approxi- mately I hour after in~estion. It is finally excreted in the urine, reaching ~ maximum concentration ap- proximately 3-5 bourn aRer ingestion.~ In the hu- man oral cavity, nit're is partly converted {20%} into nitrite by the oral microbial flora.~ Extremely high concentrations el nitrates and nitrRes were de- tected in gastric juice!and saliva in Linxian inhabit- ants.~'1°° The salivary nitrites in patients with marked epithelial dysi~lasia or EC were significantly higher than those in fiormal controls.1°° The poten- tial hazard of the high hitrite concentrations in saliva is the possible reactiob with nitrostable amino com- pounds to yield carcii~ogenic N-nitrgso compounds such as N.mtmsamm ,es. As mentioned, this reaction is catalyzed wheneve/.:., bacteria are present. When host immunblogic defenses are compro- mised or when norma~ flora is suppressed in patients who are on long-terni antibiotic therapy, the nor- really saprophytic bacteria may become pathogenic and invade the esophageal mucosa.~ Bacterial eso- phagitis may be freque~afly underdiagnosed, because the oropharyngeal flor~ normally bathes the esopha- gus at~d makes the cli~.ical diagnosis difficult. It was recently reported that~ bacterial esophagitis can be fonnd in up to 11% Of endoscopic biopsies and in 16% of autopsies among immunocomprom/sed pa- tients.~: Bacterial infd.ctJons are associated with a severe inflammatory p~ocess characterized by a ~use neutrophilic/nfd~ration w/th necrosis and de- generation. It is genera~lly more diffuse and spreads deeper than other ulcerative lesions2z Along with this !nflammatory process, esophageal epithelial cells show a sustained r~generative proliferation that may render them mor~ susceptible to carcinogenic agents. It has been reporte~ that the presence of dental caries is more serious i~ EC patients than in normal subjects. Oral hyg/ene Of high-risk populations/s gen- erally poor.~-~ which may contribute to carcinogene- sis of the esophagus (a} t.hrough the production of carcinogens or precursors such as nitroso com- pounds and n|trosamines by oral microorganisms and (b} by prodding the source of microorganisms for in~ection of the esophagus. Viral Infections and EC HPV Infection Papillomaviruses are a group of small, double- stranded ONA v/ruses, some o~'.:'~ich have remark- able host and target cell speciAcity. They induce hy- perplastic, papillomatous, and verrucous squamous cell lesions in the skin and at various mucosal sites in a wide range of hosts including humans. HPV infec- tions I/ave been reported in a number of body sites, includ/ng the anogenital tract, urethra, skin. la~'nx, tracheobronchial mucosa, nasal cavity-paranasal sinuses, oral cavity, con]unctiva, and esopha- gus}~'~-~ Recent advances in molecular biology have significantly increased our knowledge of the molecular biology and biochemistry o~ HPVs. So far, more than 60 different types of HPVs have been identified, man~ of which are shown to have a rela- tively spe'~c site of infection in the human body2°~ Besides HPV-induced benign epithelial lesions, strong evidence has accumulated/n the past few years to implicate an etiologic role ~or speciOc HPV types in the development of squamous epithelial precancerous lesions and carcinomas. Such HPV-as- sociated malignancies include anogenital carci- nomas, squamous cell carcinomas developed ~rom epidermodysplasia verrucifomis lesions and warts in |mmunocomprom/sed patients and ~rom carcinomas arising in the upper aerodigestive tract, verrucous carcinomas, and Bowen's disease.~-~ Up to 80% of cervical intraepithelial neoplasias and 88% of ana- lyzed squamous cell carcinomas of the uterine cer- vix have been shown to contain HPV DNA, mainly that of HPV 16 and 18, and, less frequently, HPV 31. 33. and 352~'~'~°~ In addition to the primary tumors, HPV DNA has also been found/n lymph node metas- tases. Cell lines derived from cervical carcinomas such as HeLa, CaSk/. and .S/Ha cells have been shown to conta/n HPV DNA and express HPV RNA.~°~ Ex- perimental studies have shown that some I-IPVs are capable of transforming rodent cells and immor- tal/z/ng human foreskin and cervical keratino- cytes.~e~='~°~ The early genes, i.e., E6 and ET. of HPVs seem ~o be the main transforming genes2~'~ New insight into the possible mechanisms o~ HPV- associated transformatior~ has been derived from the recent findings that these transforming gene prod- ucts can interact with some cellular tumor suppres- sor genes, e.g.. Rb and p53. presumably leading to inactivation of the tumor suppressor genes and con- sequently leading to the uncontrolled proliferation of the infected cells2°~'~°~ A substantial amount of evidence is available to suggest that specific HPV types are necessary but insu~cient in oncogenesis: synergistic action with other initiating dvents seems to be requi.red.=~s'~o~'~°~ Substantial evidence [or a papiilomavirus etiology of EC has beeh found in cattle. Studies in the Scottish Highlands show it to be an area with a remarkably high incidence of squamous cell carcinoma of the upper aF.mentary tract, especially the esophagus. BATCo document for Mayo Clinic 28 March 2002 C)

O~aber lgg" ETIOLOCY OF ES.OPHAGEAL CANCER among animals,t~-I°* In cattle, squamous cell papil- Iomas of the upper a~mentary tract, including the tongue, palate, pharTz~.', esophagus, and tureen, are known to be caused b~ bovine papi]lomavirus (BPV) 4 infection. Persistentiand widespread papillomato- sis and even carcinom.~s could be experimentally re- produced with BPV:; 4 infection in these ani- mals.~°~ When fe~al bovine palatinal tissue ~ra~'nents were infected with BPV 4 in vit_m and in- serted into renal capsules of nude mice, squamous call papillomas could ..be subsequently identified in 90% (19/21] of the mice. One of these papillomas was accompanied bya~ invesive squamous cell carci- noma with spleen me~astas/s-,u°'m Field stud/as in the high-incidence ar.~'~s show that 96% of cancer. bearing an/mals concurrently had benign papillomas in their alimentary tra~t, and 40% of the animats had more than lS papilIom~s. In some instances, the pro- gression from benign p =apiilomas to careinomes could be dearly idantified.l=.:"t~ Ingestion of bracken fern has been shown to be ia cri~/cal factor in the malig- nant conversion of th#se papillomas. Bracken fern contains carcinogenic ::agents and immunosuppres- sants, e.g.. azathioprin.."e?°~ BPV-4 DNA sequences could be de.tected/n .~ high copy number in both naturally occurring o~ experimentally/nduced pa- pillomas. However, nq viral DNA or viral antigens were detected in eithe~ of the naturally occurring or experimental canc~rs, iindicatin8 that the viral ge- homes are not require~I for the maintenance of the mal/gnant stateJ=-m T:.hese data suggest that {a) BPV 4"'may execute one of ~.he early events in cell trans: ~rmation, and its gun,tic informat/on may not be required for maligr/antipmgression; (b) immunosup- press/on caused by theiingestion of bracken fern al- lows the spread and th.'e persistenc~ of the BPV-in- duced papilIomes; anc~ {cJ the bracken fern would supply cocarc/nogens a~dmrcinogens, thus lead/rig to full cell transformation and progress/on. I-IPV/nvolvement iR human esophageal lesions was first suggested in ~9.6z by Syrjanen,~= who found that 40% {24/60} of pat!ents with EC presented with histological changes iddntical to those of genital con- dylomatous lesions. He an~i h/s associates subse- quently found the existence of HPV structural pro- reins in one case of esophageal squamous cell papilloma?~= These observations have been subse- quently confirmed by a number of other invest/go- tom who found HPV-sug~estive changes, HPV ant/- gens. and HPV DNA sequences in esophageal squamous cell lesious,t~'=== althoughsome ~others found no evidence of HPV involvement in esopha= geal lesions?"~'~= These data. being in alignment with the currently available evidence on the etio- logic role of HPV in the pathogenesis o~ squamous ceil carcinomas at other mucosul sites, have impli- cated HPV as a potential etiologic agent in human EC as wellJz'~t== To assess the role of HPV in the deveIopment of EC, a series of esophageal precancerous lesions and cancers derived from the high-risk area in Linxian were systematically analyzed usfiug 1/ght micros- copy, electron microscopy, in situ hybridization. filter in situ hybridization, Southern blot hybridiza- tion, and polymerase chain reaction [PCR) tech- niques in our laboratory. HPV infections hi the esophagus could be demonstrated by these tech- niques, although the diagnostic rates of HPV infec- t/ons vary in different series of esophageal lesions as well as in different methods employed. Using light microscopy, HPV-suggestive lesions were found in 49.0% (25 of 51] of EC spec/mens,m-== Using the • DNA in situ hybridization technique. 43.1% [22 of 5~) of the lesions were shown to contain HPV DNA?= HPV-like parr.icles located in the nuclei o£ koilocytotic cells were found in 2 of the ~ specimens previously shown to be HPV positive by in situ hy- bridization?= Using PCR, 49.0% o~ the lesions were shown to contain HPV DNA?~ Similarly, HPV DNA sequences were also found in more than 40% of the DNA samples extracted from fresh ECs and their surrounding Lissues by Southern blot hybrid/zaLion. As detected by Southern blot hybrid/zation~ HPV DNA in these carcinomas appears to be present mainly as an integrated form.t== Using the filter in situ hybridization technique, more than 66.3% of esophageal cytological samples were demonstrated to be HPV DNA positive. HPV DNA was detected in 22.2% o~ pat/ents (2/9} without cytological atypia, in 50% (3/6] with mild dysplasia, in 80.6% (25/3:t) with moderate dysp]asia, in 67.9% (~9/28] with severe dysp]asia, and in 66.7% (4/6) with invasive squa- mous cell carcinoma?=~ These results indicate that HPV in[ecLion in hu- man esophagus does occur, and it seems- to be re- markably prevalent in people living in the high-risk area of the Henan Province of China. The histologi- cal features and viral types in esophageal HPV infec- Lions are comparable with those described at other mucesal sites.~''~-~=-~ These results suggest that HPV infections may play an etiologic role in the pathogenesis dEC, and. like BPV 4, HPVs might also act syner~isLically with other risk factors that have previously been related to this malignancy. Similar observations were recently reported by Williamson et al?= and Miller et al.~== in the high-incidence pop- ulations o£ South Africa and Alaska Natives. Many sirrdlarities apparently exist between the pathogenesis of. alimentary BPV-carcinoma se- quence and human ECs. Both lesions show remark- able geographic distribution. In many instances, a high inddence ofany~articular malignancy in well- BATCo document for Mayo Clinic 28 March 2002 Oo

G.s~rROENTEROLOG¥ VoL I03. No. 4 defined geographic ar~as appears to be associated to some extent with ini~ectious agents. ThLs has been shown with HBV, EBV..', and human T-cell leukemia/ lymphoma viruses.~'~ The same is true with upper alimentary tract cancer in cattle; the high incidence is ascribed to synergistic actions between' BPV 4 and carcinogenic factors in bracken fern.~°s'm Simio lady, a high preval.ence of HPV infections has been recognized in the high-risk areas of South Africa,"~'n'n~=T Al~ka,~" and China.~.1=-~ Fur- thermore, in alignment with the hypothesis of the synergistic actions between BPV 4 and ingested bracken fern, I-IPV infect/on in human esophagus may aIso act synergistically with other high-risk ~ac- tots, including the chem/cal carcinogens, physicaI trauma, and nutritional deficienc/es discussed prev/- ously. Therefore, the ~tabllshed mechanisms for the pathogenesLs of bovin.e alimentary carcinomas may also be applicable to that of EC in humans. Herpesvirus and EC Herpesv/ruses a~e a group of enveloped, icosa- hedral v/ruses that cdntain linear, double.stranded DNA with a virion ~meter of approximately ~00 nm?~-They are wide,read in human beings, mates, and other animals, and they are the causative agents of diseases ran~/ng from relatively mild skin erupt/ons to severe and frequently fatal encephal/- tis.~'~ Some member~ of the Herpesviridae are con- vincingly assoc/ated .with the pathogenesis of roots.~'~ Three herp~sv/mses are known to cause esophageal infections fin humans: HSV, CMV, and EBV. Each of these v/..i'uses has been implicated in " certain malignant d/se~es as well. HSVin~.~ction. Among the visceral organs, the esophagus IS the mostlfrequent site of involvement by HSV.~'~'~-~'= Unt/il recently, however, herpetic esophagitis was rarely s~ uspected or diagnosed during life. elthough postmortem exam/nat.ions have shown that it is quite frequent ~s a cause ofesophageaI ulcer- ation-m.'~'m The disease is characterized by the presence of multiple d!screte, yellowish mucosal ul- cerations, particularly fin the distal esophagus,ua-~ It seems likely that HS~V infections of the esophagus have been previously overlooked because many agents induce esophagitis, and herpes is o.ften ne- glected unless the pat/ent belongs to a high-risk group for the disease Or unless one recognizes the typical viral inclusions. In the study conducted by Nash and Roes.~s herpetic esophagitis accounted for 25,% of all cases of esophageal uIceration. Esophageal herpes infections may be complicated by the pres- ence of other infectious agents such as fungi. CMV, and bacteria.~-s" On the other hand. herpetic infec- t/on may also precede secondary bacterial or fungal in~ections. The most notable characteristics of HSV are latent infections and subsequent recurrences. When HSV infects sensory nerve endings, the virus is trans- ported by way of the axons to the neurons of sensory ganglia, where it remains latent. As to the pathogene- sis of herpes infections, the superior cervical and vagus ganglions of humans are considered to be the source of recurrent herpes esophagitisY~ HSV al- most invariably undergoes a lytic infect/on, leading to cell death. During either an in/tial or a recurrent infection. HSV may rarely undergo an abortive in- fection with the expression of only a limited number of viral gene products. In th/s way. it may be in- volved in oncogenesis.~'~'~.~s HSV-t frequently causes herpetic lesions in the head and neck region in two identifiable antigeRic types o~ HSV. whereas HSV-2 is most frequently as- soc/ated with genital infections.~ Epidemiological studies have suggested a relationship between the presence of HSV antibodies and cancer of the uterine cervix and oral cavity. HSV-2 infection IS mainly as- sociated with cervical cancer,~-~* whereas HSV-1 infection is more closely associated with oral cancer.~T'~z-"~ Almost all investigat/ons have found that women with cervical cancer have in. creased levels of HSV-2 antibody compared with carefully matched control subjects?~.~-~-~e~s° In many cases, patients with carcinoma in situ or dys- plasia had HSV.2 antibody "t/tars between those of the cancer and control groups.~ Similarly, the titers of HSV ant/bodies were also significantly higher in patients w/th head and neck cancer,.partico ularly oral and laryngeal cancers.~r-~* Shillitoe et al?~ report that pat/ents with oral c~ncer have higher levels ofimmunoglobulin {Ig} M antibodies to HSV compared with those of control patients. Smith et al.''~ found significantly higher titers of anti-HSV IgA in such patients. As with HSV infection in oral mucosa, viral culture studies have shown that esoph- ageal HSV infect/on is mostly caused by HSV- Laboratory studies have shown that HSV-1 and HSV-2 are able to transform the morphological phe- notype of rodent cells.~'~-~-m'~u The failure to detect viral DNA in transformed cells led to the"hit- and-run" hypothesis of HSV transformation.~s~.~sz So far, the exact mechanism by which HSV induces transformation is not understood. Various lines of in- vestigation have shown that HSV is able to cause mutations, either point mutations or gene rearrange- ments. HSV can also induce gene amplification, par- ticularly of the sequences harboring an origin of replication such as SV40 or PVs.~°-~z Other exper/- ments have shown that HSV can activate the expres- sion of endogenous type C retroviruses.~s~'~z More broadly'. HSV has been shpwn to activate cellular BATCo document for Mayo Clinic 28 March 2002

October 1.99~' ETIOLOGY OF ESOPHAGEAL CANCER 1343 transcription or to switch on the s~'n thesis of host cell proteins not normall~r expressed in nontransformed cells.1=t'z=z These results stress the init/atorlike func- tions of HSV infectj.'ons in oncogenesis. It seems likely that pr/roary and recurrent herpetic infections lead to an increased ~umber of in/tiating events, re- suiting in a higher risk for cancer developroent,t~'ls~ The association of HSV infection with EC has not been extensively studied. A possible etiologic rela- tionship between HSV infect/on and EC has been suggested by the re~ent deroonstration of herpes- virus particles in biopsy specimens froro EC.~m "As discussed above, eso~hagitis caused by HSV infec- tion has been recosni~ ed as a distinct entity for soroe t/me. From the h/stoSene~ic point of v/ew, cancer of the esophagus resemijles that of the mouth and uter- ~ne cerv/x, in which ~quamous cell carc/nomas are predominant tumor Wpes, and they frequently de- velop through prem~ignant lesions. All these sites are targets of pr/roar~ HSV infections, and they can also have recurrent l~erpetic infections. Like genital and oral infectinns, I-~V infect/on can persist in the esophagus. Thereforel further stud/es are apparently justified to elucidate ~e role of HSV infection in the developroent of EC. : CMV in~ectioni Infection with C~V is very common in huroans, ~vith apprex/roately 80% of the populat/on older tha~ 35 years showin8 evidence of prior CtVlV infection i as measured by coropIeroent fixation.~ The vir~ can be transmitted by blood transfusion, transpla~entally, and via milk, urine, and respiratory secre.fions.~ It has been associated with a wide range of .diseases from birth defects and interstitial pneumo~i~ to f~quently mild subclin/cal infections, and it has .~lso been iroplicated in a num- ber of huroan turoorsi e.g.. prestatic carc/noma, ce- Ion carcinoma, and K~posi's sarcoroa.~z~-~'~z~z~. Ex- perimentally. C/vfV is ~ble to transforro cells in vitro either as a UV light-i.nactivated virus or as cloned fragments of viral DNA.~"'~° As with HSV, a de- fined fragment of viral DNA has been used to induce transformat/on of rodent cells, but to date no C~V DNA has been detected in these transformed ceils.~.~s.~,~o After HSV infections, C~vIV/s the second roost common virus to involve the esophagus.=~.~ss-~s~ However, unlike HSV. C,'v~/tends to involve the stomach and intestines more frequently than the esophagus.~'~z The association o~ C~[V with hu- man EC has not been stud/ed. However, ~l~v" is sociated with herpetic esophag~tis and can persist in the esophagus. This virus has been implicated in the pathogenesis of a variety of human cancers, and it has the ability to pmd.'uce tumors in animals and transform ceils in vit,'o. Therefore. C~iV should not be neglected while considering the viruses as etio- logic agents of EC. E~V i~[ection. EBV is an ubiquitous human herpesvirus found as widespread and Iargely asyrop- toroatic infections in huroan beings. The virus has been d~scribed as B lymphotropic because o fits c.lose association with three [ymphoproliferat/ve diseases of B-cell or/gin, i.e., Burkitt's lymphoroa, B-cell lyro- phoroas in immunocoropmmised individuals, and infectious mononucleosis.~'~.l~'~s However, re- cent data suggest that the priroe ta~et for EBV infec- t/on might be the epithelial cells and that infection B lymphocytes is of secondary importance with re- gard to the viral life cycle.~°'~" Beside the etiologic association of EBV infection with lyrophat/c tissue malignancies. EBV/s also assoc/ated with nasophar- yngeal carcinoma. More than 90% of Burkitt's lyro- phoma and nasopharyngeal carcinnroa show evi- dence ofEBV infection.~°~.~ In add/tion, EBV has also been identified in some undifferentiated carci- nomas with intense lyrophoid infiltration [lyropho- epithelioroalike carcinomas) at other sites, including the thymus,z~.~'~ stomach,z~-Is lung,z~.~° and salivary gland.~'z~ EBV infection has also been found in association with oral haixy leukoplakia, a condition associated with acquired, irorounodefi- ciency syndrome (AIDS}.~ Recently, Kitchen et a]. described a gzoup of ulcerative esophageal lesions occurring in patients with AIDS in which EBV DNA sequences were shown by in situ hybridization. These findings suggest the involveroent of EBV in esophageal epithelial lesions and support the hy- pothesis that EBV is an etiologic agent in soroe cases of AIDS-associated esophageal ulceration. The close association of EBV with nasopharyngeal carcinoroa clearly indicates that the virus can have oncogenic potential within an epithelial ce11. Al- though EBV DNA has also been detected in a number of other epithelial malignancies of the head and neck, the role o~this virus in these tumors is less well understood. The fact that esophageal epitheliuro can be infected with EBV in vivo, and the close proxim- ity of the esophagus to the nasopharynx, raise the possibility that this virus may be involved in esopha- geal carcinogenesis as well, although there is no di- rect evidence as yet implicating EBV as an etiologic agent of Summary Generally. cancer davelops secondary to roul- tiple genetic events. Any environmental agent or disease processes that increase the possibility of ge- netic damage or produce sustained increased cell proliferation can enhance the likelihood of cancer development. Many chronic inflammatory processes caused by chemical irritants, physical trauma.s, and BATCo document for Mayo Clinic 28 March 2002

GASTROE~. ~'TEROLOGY vo{. ~0~. No. -1 infectious microorganisms result in sustained regen- erat/ve proliferat/oq of cells and increase the risk of cancer developmenL No|able examples are the asso, ciarion of chronic at~ phic gastr/t/s with gastr/c carci- noma. chronic ulce/ztive colitis with colonic carci. noma. galIstones wiih cancer of the gallbladder and biliary ducts, tropic.~l phagedenic ulcers with squa- mous cell carcinom~ of the skin, and chronic eso- phagitis secondary l to gastric reflux leading to Earrett's esophagus W/th adenocarc/noma,l~ In this connection, chronic ~sophagitis (non-reflux related) has been described ~ the most frequent ~nding in the high-risk popul~t/ons of Iron and Ch/na, and it has been presumed ~s a precursor lesion of EC, It/s well known that hifectious with microorganisms cause acute and/or ~hronic inflammatory processes; therefore, in~ect/ou~, agents may be involved in esophageal carcinogenesis s/reply by stimulating the chronic inflammato~ process. In addition, microor. ganisms may also cohtribute to the development of EC by producing car .c~nogeus or promoters or by act- ing directly on esopl~a, ge~l epithelial cells. Fungal contaminaiion and infection may be in- volved in the pathog~nesis of EC by producing nitro- samines and their pr~.cursors or hy producing myco- toxins. Studies in th~ high-incidence areas of EC in China have shown t~.'t the ingestion of moldy food- stu~ and pick.Ied vegetables are important etiologic factors for the development of EC. Laboratory inves- t/gations have demo~trated that some common spa. cies of ~unzi belon~ng to Fusariurn, C, eotrichum, Asperg//|us, and other, genera not only could reduce nitrates to nitrites butialso could decompose proteins and increase the amo.unt of amines in food, conse- quently promoting th~ formation of carcinogenic ni- trosamines. The mut~erdc and carcinogenic effects of the extracts of so#oral fungi isolated from the ~rains and foodstuffs! in the high-risk ar_eas have been sho~ by both in!vitro and in vivo studies. Simi- lar to fungi, bacteria n~ay also be associated with EC by producing carcinogenic chemicals and increasing cell proliferation by ~t/mulating the inflammatory process. A substantial amount of co/deuce has suggested that certain tumor viruses capable of cell transfor- mation in vitro play ~ role in growth regulation in vivo, and under certain circumstances such cells con- tribute to the oncogenic process. Of these viruses, HPV. HSV, CMV, and EBV have been implicated in the pathogeuesis of a variety of human carcinomas. and they have the ability to produce tum6rs in ani- mals and transform cells in vitro. These viruses have also been shown to infect esophageal epithelium. thus making them potential candidates as etiologic agents of EC. Unlike the fungi and bacteria that are ind,.'rectly ,.'.-.vo!ved in ~n¢vgenesis b.v producing car- cinogenic and cocarcinogenic compounds or bystim- ulating the inflammatory process, the viruses usually cause spec/fic cancers by acting directly on the host cells at a complex molecular level. Despite the substantJaI amount of data on EC ob- tained during the recent years, the causat./ve factors of this disease still remain to be established. Except for the chemical agents (e.g., nitrosamines, mycotox- ins, opium abuse, excessive tobacco use, and alcohol drinking}, nutr/rional deficiencies (including vita- mins A, B, and C and certain trace elements), and physical factors (e~arse and hot food intake), the data summarized in this review clearly indicate that in- fectious agents may play an important ale in the etinlogy of EC either by contaminating the food chain or by directly irJecting the esophageal mu- nose. Esophageal carcinogenesis is a complex, multi- step process, and R certainly has a multifactor/al eti- ology. Some factors may be important in the initiation o~ the neoplastic state, whereas others may act in the promot/on and progression of the lesions. Most certainly, the development ofinvas/ve EC will result from the synergistic actions between .some or many of these etiologic factors. Further studies cused on the etiology o~ EC and synergistic actions between these risk factors would greatly increase our understanding of esophageal carcinogenesis. References .1. Waterhouse ~. Muir CS. Shanmugarat'm~n K, Fowell .Cancer incidence in f~ve continents. Volume ]~ Lyon. Fr~lc~ [ARC SclentJfic'P~:~lic, ar~on 42,1976. 2. I}unham LJ, Bailzr IC. World maps'o[ c~ncer morality rates and frequency ra~ios. J Natl C~ncet" Inst 196~:41:15.q-2o:L 3. P~kin DM0 L~r~ Eo Mui~ CS. Esthnates o~ the wor[dw'Ide frequency of sixteen major cancers in 1980. Int J C~ncer 1988;41:184-197. 4. Coordinating Group for Research on Etiology Cancer |n North China. The epidemiology and etiology o~ emph~Seat cancer in Chin~. Chin Meal | 197S:I:157-I83. $. M M. L/P. Li B. Recent progress in research on esophageal cancer in Chiru~. Adv Cancer Res 19~0:33:I73-2~9. 6. Wu HH. Esophageal cancer resez,'ch in the Feo[:tle's Republic o~ China. Am10to11981~9Q:359-353. 7. YanS CS. Resezrr.h on ~sophzgezI cancer in China: a review. Cancel" Res 1980;40:2as3-2644. 8. Lu lB. YanS WX. Lin JM, Li YS. Qin YM.'rr~nds in morbidity and rnor,~lRy for oesophzgeal cancer in Linxian country. J Cancer 19as:36:643-645. 9. Li JY. Ershow AG, Chen ZJ. Wacholder $, Li GY. _Guo W. Li B. 8|or WJ. A case-cvutm| study of cancer of the e_~pha~us and gastric cardia |n Linzian. int J Cancer 19~9;43:75S-761. 10. Day NF_.. SORe aspects Of the epiCleminlosy of ~ophageal cancer. Cancer Res 1975~3S:3304-3307. 11. Sons HU. Etloiogic and epidemio|o~c factors ofcarcinoma o~" the esophagus. Surg Gynecol Obstet IgST:165:ls3-190. 12. Joint lran/IARC Study Group. Esophageal cancer studies in the Caspian Littoral or" Iz'zn: result of population studies: a p~c~dome. J Nazl Cancer InsZ |977~11ZT-113e. 13. Mabbcubi EO..=.ramesh B. E..":.demZolozy of esophagea.l BATCo document for Mayo Clinic 28 March 2002

~-'~OLOGY OF ESOPHAGEAL CANCER 1345 cancer in Inn, with ~l~-~:ial reference to nulritional and cul- tural aspects. Preven~ t MUd 1~0:9:613~21. 14. ~et I. ~h~ubi ~ ~phage~ ~ncer in ~ ~ ~tt~ ~. Mah~ubi EO. Kme~L ~k PJ. Day ~ G~ P. ~teb • Ompha~ ~ncer stu~ in the ~ litto~i o~ln~ the ~pian ~cer m~t~. BrJ ~r 1~:1~- 214. 16. Dowla~ ~ Miller RL Role of opi~ in ~hage~ ~ncec a h~th~ ~n~r Res IL Sflb~ W. ~noma 9f the o~pha~ ~ ~demfol- la. Wa~ GP. ~me ~ of the epide~ol~ and I~ of ~p~eal ~ncer ~ ~i~lar emp~ on Tm~k~ ~u~ A~ Adv ~n~r R~ 1~ J~ ~ ~ ~O, Van der W~t ~ and oth~ m~ ~n~r ~tte~ in f~ ~ 20. R~ E ~p~ ~ncer ~ the Triter 1~1~ J Nail 2L A~ • ~e e~ide~ol~ of ~n~r ~ ~, R~ubllc of Ch~ I~t J Epldeminl ~ ~ ~ T~ for ~e P~venUon ~ T~t of ~ha~i ~ncer. P~pha~ ~ ~hageal ~. D~nt of Path~o~ of ~n~r l~t~ ~ C~ ~ P~Ze~ and ~pha~ ~ce~ ~ dam~c ~ 24. Detriment ~ Pa~I~ of ~cer ~te of the ~r H~tal ~f Hu~ ~. ~~ in ~o~S ~u~ty, Hu~i Pm~ ~ ~1 Sin 197~31N16. 25. Rubio ~ Uu F~ S~n~nm~ ~o~ ~a of~e 2& Shl~ ~, Sflve~ ~ O~ mn~r and ~ ~plm 2~. ~UF ~ Om~clal he~ ~mplex ~ ~o~ ~ ~n~p~ In ~e ep~de~ioIo~, path~n~ a~ and ~ in whi~ the ~ ~y be impH~ O~ 26. Ve~ ~F:~-137. ~ ~ DNA s~u~nc~ in o~ ~n~ I~ and ~uamo~ ~11 ~in~a demo~tmt~ byin~ tio~ J O~ Pathol ~:17:273-278. 30. ~HF ~ ~x MF, P~e~.Maiziand N. Fa~ll~ ~e cu~nt ~t~ in relation to o~l d~e~. O~ Su~O~l M~ 3L Chang F, Sy~ln~ & ~u~nen J, ~ J. S~nen ~ ~on of h~an ~o~ms ~ DNA in oral ce~ ~om~ by in situ hyb~n a~ ~i~e~ chain ~eactio~ A~ Oe~atol R~ Chang F. Sy~anen 5. gellokm~ I, Sy~anen ~ H~n ~ma~ms ~P~ inf~zio~ and their ~o~ wi~ ~e. J Oral Pathol ~ ~I:~O~3~7. M~n~ P. Shah ~. Respi~to~ ~pillomat~ rela~on to genital tract ~api~omzvi~s. P~ M~ ~ 37.. 33. 34. zur Hansen H. Papillomaviruses as carcinomaviruse~ Adv Viral Onc~l 1989:8:1-26. 35. Gissmann L Papillomaviruses and their ~mdation with cancer in animal~ and in man. Cancer Sur~ 1984;3:151-181. 36. Chang F. Role of papillomaviruses. J Clin Pathol 199~,43:289-.27& 3)'. El~tein MA. AchonS BG, ads. The Epsteln-Barr v~rus: recent advances. London: Heinemann, 19~6. 38. Young L& Sixby JW. Epstein-Ba~r vit~s and epithelial cell,: A possible role for the virus in the development of cervical carcinoma. Cancer Sur¢ 39, Grafey ML Weiss LM. Viral oncegenesis: Ep~tein-Barr virus. Am J Otolary~So11990:11:375-381. 40, Nomura A, Stemmermann GN. Chyou P-H. gato I. Purer. Pere~ G[. l~laser M/. Hdlcebacter py/orf infection and ~stric carcinoma among Japanese Americans in Hawaii. N Engl J Mud 1~9I;325:1132-1136. 41. Pzrsonnet J, Friedman GD. Vandemteen DP, Chang Yuan. Vogelman JH. Orentreigh N. Sthley Rg. HeJicobectex pylorl fnfecilon and the risk of gastric cat'c~noma. N Engl J MUd 1991;325:1127-1131. 42. Correa P. Is S~strJc caroinoma an Infectious disease.:' N Engl J MUd 1991;325:1170--1171. 43. Spector DH. Spector SA. The oncogenic potential of human cytomegalovirus. Pros Mud Viral 1984",2~.48-89. 44. Rapp F. Cytomegalovlrus and carcinogenesis. J NaZI Cancer ln.q I984;)'2:783-787. 45. Lung ML. Chang ~. Huans ML, Guo HY, Choy D. Sham J. Tsao SY. Chang P. NS MH. Epstein-Barr virus ganotypes soc:~atud with nasopharyngeal carcinoma in southern China. Virology 1-q90;,177:44-53. 48. Yah F-S. Yu MC, Mo CC, Luo S, Tons MJ, Heuderson BE. Hepatitis B viru~ aflatoxin.s, and hepatocellular carcinoma fn southern Guan~xi. China. Cancer Res 1~89;49:2~6..2509. 4).. Sun T-T. Chu YY. Carcinogenesis and prevention strategy of liver cancer in "areas of prevalence. J Cell Fhy~inl 1984~$uppl}:39-44. 48. Wynder EL, Br~s 1~. A study of etiological factors in cancer of the esophagu~ Cancer 1961:14:389-413. 49. Lu SH. Rug@m M. Zhang MS, Fens L, Luo FJ, Chui SX. Diana U, Saffh~ R, Rajewsky MF, Relevance of N.niu~sa. mines to esophageal cancer in China. [ Cell Physiol 1986;4{Suppl}:51-5~. ~0. Lu SH. Chul SX, YangWX. Hu XN. Guo LF. Li FM. Retevan~e o! N.nitl~amines to oesophageal cancer in China. IARC Sci 5~. LI M~. Chang b"J. Carcinogenesis of oe~ophaseal cancer in Lin:dan. China. Chin MUd ~ 1984:9"/'.311-316. ~2. Singer GM, Chuan J, Roman J. MinoHsln L. Lljinskl W. Nitro- samines and nitrosamine precursors in fo4xis kom Linxian. China. a high incidence area for esophageal cancer. Caroino- genesis 1986:~.'733-736. 53. Schottenfeld D. Epidemiology of cancer of the esophagus. gamin Oncol 1~4:11:92-100. ~4. Siddiqi M, Preussmann R. F.sophaseal cancer in Kashmir-- an assessment. J Cancer Res Clin Oncel 1989;115:111-117. 5& Pottern L,X,L Morr~ LE. Blot W]. Zlesler RG, Fraumeni Emphagea[ cancer among black men in Washington. D.C. I. Alcohol. tohacc~, and other risk hcto~ | Nail Cancer lnst 1981;67:777-783, 2]egler RG. Morris LE. Blot WJ. Pottern LM. Hoover R. Frau- meni JF. F~opha~-al cancer among black men in Washing- ton. D.C. |L Role of nutrition. | Nail Cancer |not I981"-67:1199-t2{]6. Tuyns AJ. Ep]demlolo~" o[ e~ophageal cancer in France. |m Pfeif~er CJ. ed. Cancer of the esophagos. Volume I. Boca ton. FL: CRC. t~82:3-18. 56. 57. 0 0'~ BATCo document for Mayo Clinic 28 March 2002