Human Papillomavirus Type 16 DNA in Ocular and Cervical Swabs of Women With Genital Tract Condylomata

Human Papillomavirus Type 16 DNA in Ocular and Cervical Swabs of Women With Genital Tract Condylomata Jan M. McDonnell, M.D., David Wagner, M.D., Siu T. Ng, M.D., Gerald Bernstein, M.D., and Yan Yu Sun, M.D. Human papillomavirus type 16 is associated with dysplasias and carcinomas of the conjunctiva and of the uterine cervix. To explore the relationship between cervical and ocular human papillomavirus infection, we examined DNA from bilateral limbal swabs and cervical swabs from 17 women (age range, 17 to 46 years; median, 31.7 years) with biopsyproven human papillomavirus-related cervical dysplasia who had a normal ocular surface. Using polymerase chain reaction, we identified human papillomavirus 16 DNA in one or both eyes of 13 (76.5%) patients, six (46.2%) of whom had demonstrable human papillomavirus 16 DNA in cervical swabs as well. It thus appears that human papillomavirus 16 is present in the conjunctivae of some patients with human papillomavirus-related genital warts who have no ocular manifestations of infection. Although autoinoculation of conjunctiva may be the source of some ocular human papillomavirus, data suggest that other modes of transmission to the eye also exist. Additional study of the epidemiologic characteristics of ocular human papillomavirus, a widely prevalent virus known to be associated with dysplasias/atypias and cancer,.is warranted. HUMAN PAPILLOMAVIRUSES are responsible for an epidemic of sexually transmitted disease in the United States, with an estimated 12 million

Accepted for publication April 24, 1991. From the Doheny Eye Institute (Drs. McDonnell, Wagner, and Sun), and Departments of Ophthalmology (Drs. McDonnell and Wagner), Pathology (Dr. McDonnell), and Obstetrics and Gynecology (Drs. Ng and Bernstein), University of Southern California School of Medicine, Los Angeles, California. This study was supported by grant GA 89-077 from Fight for Sight-Prevent Blindness. Reprint requests to Jan M. McDonnell, M.D., Doheny Eye Institute, 1355 San Pablo St., Los Angeles, CA 90033.

©AMl!RICAN JOURNAL OF OPHTHALMOLOGY

112:61-66,

JULY,

people currently being treated for anogenital lesions and millions more with inapparent or quiescent infection.' Approximately 750,000 human papillomavirus-related genital tract lesions are diagnosed each year. Over 50 distinct strains of human papillomavirus have been identified,' some of which, particularly types 16 and 18, have been implicated in the development of several carcinomas in humans, including those in the uterine cervix, anogenital region, upper respiratory tract, and skin," Human papillomavirus DNA from types 6 and 11, commonly associated with respiratory papillomatosis and benign anogenital condylomata, are also present in conjunctival papillomas,3-6 which indicates an association between human papillomavirus and ocular lesions. Human papillomavirus type 16 DNA has recently been identified in squamous dysplasias and carcinomas of the conjunctiva and cornea? and of the eyelid," and other conjunctival dysplasias have been shown to contain DNA from both human papillomavirus 16 and 18. 9 Although transmission of human papillomavirus to the cervix and anogenital region is through intimate contact,' it is not clear how the infection reaches the conjunctiva. Some sexually transmitted diseases, such as chlamydia, can be transmitted to the eye through hand-eye contact." We attempted to determine whether or not such a mode of transmission might be possible for human papillomavirus and specifically whether individuals with human papillomavirus-related cervical infection and lesions might also have asymptomatic ocular human papillomavirus infection.

Patients and Methods We studied 17 patients who were examined at the Outpatient Gynecology Clinic at Los Angeles County/University of Southern California Medical Center for clinically evident cervical 1991

61

62

AMERICAN

lesions, presumed to be of human papillomavirus origin, or biopsy-proven condylomata. Patients were examined for the study when they came to the clinic for examination after treatment of the cervical lesions. After obtaining informed consent, a slit-lamp examination of the anterior segment was performed on each patient. Limbal swabs were then performed after a drop of topical anesthetic was applied. A sterile cotton swab was gently passed over the inferior corneoscleral limbus and inferior fornix of each eye and immediately placed in sterile, cell-free virus transport medium. Samples from each eye were obtained with separate cotton swabs and placed in separate vials. Cervical swabs of cervical lesions were performed. After placement of an unlubricated speculum, a sterile cotton swab was passed gently across the endocervix and exocervix and then immediately placed in sterile, cell-free virus transport medium. Vials were stored at 4 C until DNA analysis. Polymerase chain reaction was performed as described previously." Briefly, the contents of each vial were boiled to extract DNA. An aliquot was added to individual microcentrifuge tubes already containing 100 ....molyl of each deoxyribonucleotide, 10 mmol of Tris buffer (pH 8.3), 50 ....moIjI of potassium chloride, 1.5 TABLE 1 HUMAN PAPILLOMAVIRUS 16 DNA IN SWABS OF LIMBUS AND CERVIX CASENO.• AGE(YRS). ETHNICITY'

R.E.

L.E.

CERVIX

1.20. H 2.46, H 3.36. H 4.23. H 5.27. H 6.31. H 7.28. H 8.39. H 9.23. B 10. 17. H 11.29, H 12.31, H 13.39. H 14,43. H 15.40, H 16.35, H 17.29. H

Absent Absent Present Present Present Absent Present Absent Present Present Present Present Absent Present Absent Present Absent

Present Absent Present Present Absent Absent Absent Present Absent Present Absent Absent Present Absent Absent Present Absent

Absent Absent Absent Present Absent Absent Present Present Absent Absent Absent Present Absent Absent Absent Present Absent

*H equals Hispanic; B equals Black.

July, 1991

JOURNAL OF OPHTHALMOLOGY

....mol /I of magnesium chloride, and 5 ....g per tube of gelatin as stabilizer. Each reaction mixture contained a pair of oligonucleotide primers specific for human papillomavirus type 16 or 18,1l at a concentration of 1 ....mol /I, and two units of AmpliTaq DNA polymerase (Perkin Elmer Cetus, Emeryville, California). The reaction mixture was overlaid with an equal volume of mineral oil. Temperature cycling consisted of denaturation at 94 C, annealing of oligonucleotides at 50 C for two minutes, and primer extension at 72 C for one minute. Twenty-five cycles were carried out for human papillomavirus 16; 30 cycles were done for human papillomavirus 18. Aliquots of the aqueous phase were analyzed by dot blot hybridization. For hybridization of amplified DNA,S ....1 of aqueous phase product was mixed with 250 ....1 of a denaturation solution consisting of 0.4 N sodium hydroxide and 25 mmol of ethylenediaminetetraacetic acid. The mixture was applied to nylon-filter membranes (Oncor, Gaithersburg, Maryland) with a dot blot apparatus (Oncor). Samples were washed, and the DNA was fixed by baking at 80 C under 16 em of water vacuum for two hours. Hybridization with 32p-Iabeled probes specific for human papillomavirus type 16 or 18 was performed as described previously." Positive controls included commercially available whole-genome DNA from human papillomavirus types 16 and 18 (Oncor) and conjunctival tissue samples known to be positive for human papillomavirus DNA. 7 Negative controls included paraffin sections from a conjunctival melanoma and three samples of ligneous conjunctivitis, processed as described previously," Additional negative controls included two vials run through the polymerase chain reaction, dot blot, and hybridization that contained all elements except target DNA or Taq polymerase, respectively, and in which these elements were replaced with an equal volume of distilled water. Results Swabs were obtained from the right eye, left eye, and cervix from 17 women who ranged in age from 17 to 46 years, with a median age of 31.7 years. Sixteen (94.1 %) were Hispanic, and one (5.9%) was black. The duration of medically documented cervical lesions ranged from one month to ten years with a median duration

Vol. 112, No. 1

Papillomavirus 1YPe 16 and Conjunctiva

of 21.6 months. All patients had abnormal results of Papanicolaou smears suggestive of human papillomavirus infection or condyloma. None of these women had a history of ocular problems. Ocular examination disclosed no abnormalities in any of the 17 women. Culposcopic examination with biopsies of cervix or endocervix confirmed the presence of condyloma in 16 (94.1 %) patients. The single patient without condylomata had cervical intraepithelial neoplasia grade III (cervical intraepithelial neoplasia III)/carcinoma in situ (full-thickness epithelial atypia) on examination of tissue from the endocervical canal. DNA from human papillomavirus 16 was identified in ocular specimens from 13 (76.5%) of the 17 patients (Table 1). In six (35.3%) patients, human papillomavirus 16 DNA was detected in only the right eye, and three (17.6%) patients had DNA present in only the left eye. Four (23.5%) patients had human papillomavirus 16 DNA present bilaterally (Figures 1 and 2). Six (35.3%) patients had cervical swabs positive for human papillomavirus 16 DNA. Three of these had human papillomavirus DNA present also in both eyes, two had human papillomavirus DNA in only the right eye, and the sixth was strongly positive for human papillomavirus in the left eye only. Every patient in whom human papillomavirus was identified in the cervix had human papillomavirus-positive conjunctivae, but seven (53.8%) of the patients with human papillomavirus 16-positive conjunctivae did not have human papillomavirus 16 DNA in their cervical swabs. Human papillomavirus 18 DNA was present in the right eye of one (9.1 %) patient, who also had human papillomavirus 16 DNA bilaterally but whose cervix was negative for human papillomavirps 16. Human papillomavirus 18 was present in the cervix of another patient with human papillomavirus 16-positive conjunctivae and cervix (Table 2). All other samples tested were negative for human papillomavirus 18 DNA. Controls of human papillomavirus 16 or 18 DNA were positive when reacted with type 16 and 18 primers and probes, respectively. Specimens of conjunctival melanoma and ligneous conjunctivitis were negative for human papillomavirus types 16 and 18 DNA, as were reactions run without target DNA and reactions containing type 16 or 18 whole-genome DNA but no Taq.

63

Fig. 1 (McDonnell and associates). Dot blots of polymerase chain reaction product, using human papillomavirus 16 primers, hybridized to IlIp-Iabeled human papillomavirus 16 probe. Data from two patients are illustrated as follows: Case 8 (Fl to HI), right eye is negative (F1), left eye and cervix are positive (G1 and HI, respectively). Case 11 (D2 to F2), right eye is positive (D2), left eye and cervix are negative. Additional samples represent controls: AI, human papillomavirus 16 DNA 0.5 ng; 81 to E1, positive tissue controls; A2, no DNA; C2, human papillomavirus 16 DNA 0.5 ng; G2, human papillomavirus 16 DNA 0.001 ng; H2, human papillomavirus 16 DNA, no Taq. Exposure time was 22 hours.

Discussion We identified human papillomavirus type 16 or 18 DNA in ocular swabs from 13 (76.5%) of 17 women with clinically evident, biopsy-proven, human papillomavirus-related genital tract condylomata or dysplasia. In six (46.2%) of these 13 women, the same human papillomavi-

64

AMERICAN JOURNAL OF OPHTHALMOLOGY

Ae e 1

2

3

B

c D E F Fig. 2 (McDonnell and associates). Dot blots of polymerase chain reaction product; data from three additional patients. A composite was made from different parts of a larger membrane to illustrate the results of duplicate polymerase chain reactions on two of the patients. Case 5 (AI to C1, A3 to C3) has positive right eye on two different polymerase chain reactions (AI, A3), and left eye and cervix are negative (B1, Cl; B3, C3). Case 2 had no identifiable human papillomavirus 16 DNA in any of the three swabs on two separate polymerase chain reactions (D1 to F1, A2 to C2). Case 7 had positive swabs of the right eye and cervix (D2 and F2, respectively), but the left eye was negative (E2). D3, human papillomavirus 16 DNA; E3, no DNA.

rus type was present in cervical swabs. None of the patients had any evidence of ocular disease on slit-lamp examination of the anterior segment. These findings suggest several novel concepts regarding the possible role of human papillomavirus in ocular surface neoplasia and raise many more questions. Human papillomavirus can exist without gross evidence of disease in the anogenital tract. Using polymerase chain reaction, filter in situ hybridization, or Southern blotting, Melchers and associates" found a 5% to 12% inci-

July, 1991

dence of DNA from several types of human papillomavirus in cervical samples from women with no overt cervical abnormalities; the types most frequently identified were 6, II, 16, and 18, human papillomavirus types previously shown also to be associated with ocular lesions.t" We reported DNA from human papillomavirus 16 in uninvolved and involved eyes of a patient with unilateral conjunctival epithelial dysplasia or carcinoma/ (and unpublished data) which establishes that human papillomavirus can similarly exist in the conjunctiva in the absence of an overt lesion. We have now demonstrated human papillomavirus 16 or 18 DNA in the apparently uninvolved conjunctivae of women with genital tract lesions presumed to be of human papillomavirus origin. Human papillomavirus DNA appears to exist in only one eye in some of the women we studied. Additionally, slightly over one half of the patients with unilateral or bilateral conjunctival human papillomavirus 16 had no demonstrable human papillomavirus 16 DNA on cervical swabs, even though cervical lesions were present both clinically and on biopsy of the cervix (Table 1). A variety of human papillomavirus types are associated with cervical lesions," and it is not unusual for a patient to have simultaneous infections with more than one human papillomavirus type in the genital tract alone. IS Recurrent condyloma may be associated with a human papillomavirus type different than that of an earlier lesion, which indicates that the human papillomavirus complement detectable in a given patient may vary over time, perhaps representing new infections." In a recent report, human papillomavirus 16 and 18 were both identified in two of five conjunctival epithelial tumors," and in our study there were two patients who had both human papillomavirus 16 and 18 in the same conjunctival swab. It thus appears that several different human papillomavirus types may infect the same individual at one or more body sites, and although one type is present in the genital tract, another type may infect the conjunctiva. We did not test for human papillomavirus 6, 11,31, or 33, or for other types associated with clinically evident anogenital lesions." Cervical condylomata and dysplasia are almost uniformly associated with human papillomavirus, and our negative results for human papillomavirus 16 in some patients suggest that other human papillomavirus types are present in the cervix in these women. Based upon the absence of human papilloma-

Papillomavirus 'Iype 16 and Conjunctiva

Vol. 112, No. 1

65

There are several possible explanations for the discrepancy. It may be that there is a long interval between human papillomavirus infection of the conjunctiva and development of epithelial lesions. This would explain why conjunctival epithelial neoplasia is an uncommon process seen only in the elderly, but it would not explain its overwhelming propensity to occur in men." Alternatively, our findings may herald an increased incidence of conjunctival squamous diseases in younger people, including a number of women. On the basis of such evidence,3.7.9 we now know that certain human papillomavirus types thrive in the conjunctival epithelium. With other organisms capable of infecting both the genital tract and the eye, such as chlamydia and syphilis, epidemics of sexually transmitted disease involving the genital tract are rapidly followed by similar epidemics of ocular disease. 10•15 Studies indicate more and more women develop sexually transmitted human papillomavirus infections of the genital tract, and infections are developing at a younger age," The median age for development of invasive cervical carcinoma is also declining. 16 The changing epidemiologic characteristics of human papillomavirus infection should be expanded to include the conjunctiva, and our patients may be the earliest manifestation of a future epidemic of ocular human papillomavirus infection and human papillomavirusrelated lesions, particularly among young people whose lifestyles render them susceptible to contracting sexually transmitted infections. The significance of human papillomaviruspositive conjunctival swabs for individual patients is unclear, although some human papillomavirus types may playa role in the development of conjunctival epithelial neoplasia. Human papillomavirus does not appear to be eradicated from the cervix of some patients over time." If the same is true of conjunctival human papillomavirus infection, our patients potentially have many years ahead of them during which they may develop conjunctival

virus 16 and 18 DNA in the cervical swabs from seven (53.8%) of the 13 women with subclinical infection of the conjunctiva, it appears that in only six (46.2%) patients is it feasible that the women infected their own conjunctivae with cervical human papillomavirus 16 through hand-to-eye contact or a systemic infection. This may imply that ocular infection in the other women was not contracted through selfinoculation but by some other mode of transmission. In one patient whose Papanicolaou smear results had been abnormal for ten years before sampling of the cervix for polymerase chain reaction, it is possible that the human papillomavirus type now in the eye may have been transmitted from an initial cervical infection that was successfully treated and no longer amenable to sampling. In most of the patients, the time course between abnormal Papanicolaou smear results and biopsy precludes such an interpretation. Thus, although anogenital human papillomavirus is clearly a sexually transmitted infection, the way that human papillomavirus reaches the eye, and therefore the population at risk for such infection, is unknown. The patients we previously described who had human papillomavirus-positive conjunctival neoplasms or swabs were older (median age, 65 years), and most were male. 6•7 All had conjunctival or corneal lesions that were severe enough to lead them to seek medical attention and to necessitate surgery. The current study demonstrates that human papillomavirus may also be present in individuals who have no overt conjunctival disease, at a younger median age than that reported for dysplasias (31.5 years in the current study). If, as we believe, human papillomavirus is somehow related to the development of conjunctival dysplasias and carcinomas." such individuals may be at risk of developing conjunctival epithelial neoplasia. If this is the case, it would be a striking departure from current thinking about the patient group at risk of developing conjunctival epithelial neoplasia.

TABLE 2 HUMAN PAPILLOMAVIRUS 16 AND 18 DNA IN SWABS OF THE CORNEOSCLERAL LIMBUS AND CERVIX R.E.

L.E.

CERVIX

PATIENT

16

18

16

18

16

18

3

Present Present

Absent Absent

Present Present

Present Absent

Absent Present

Absent

4

Present

66

AMERICAN JOURNAL OF OPHTHALMOLOGY

human papillomavirus -related neoplasia. Whether or not such lesions develop may depend upon the relative roles played by human papillomavirus and other factors in the development of neoplasia. Only large epidemiologic studies conducted over many years will disclose the potential effects of human papillomavirus in the conjunctivae of these and other asymptomatic individuals and will identify factors that may be involved in contracting ocular human papillomavirus infection and in the development of conjunctival epithelial neoplasia.

References 1. National Institute of Allergy and Infectious Disease. Fact sheet on sexually transmitted disease. Bethesda, Maryland, N.I.A.I.D., 1989. 2. Arends, M. J., Wyllie, A. H., and Bird, C. C.: Papillomaviruses and human cancer. Hum. PathoI. 21:686,1990. 3. Lass, J. H., Grove, A. S., Papale, J. J., Albert, D. M., Jenson, A. B., and Lancaster, W. D.: Detection of human papillomavirus DNA sequences in conjunctival papilloma. Am. J. Ophthalmol. 96:670,1983. 4. Naghashfar, Z., McDonnell, P. J., McDonnell, J. M., Green, W. R., and Shah, K. V.: Genital tract papillomavirus type 6 in. recurrent conjunctival papilloma. Arch. Ophthalmol, 104:1814, 1986. 5. Pfister, H., Fuchs, P. G., and Volcker, H. E.: Human papillomavirus DNA in conjunctival papilloma. Graefes Arch. Clin. Exp. Ophthalmol. 223:164, 1985. 6. McDonnell, P. J., McDonnell, J. M., Kessis, T., Green, W. R., and Shah, K. V.: Detection of human papillomavirus type 6/11 DNA in conjunctival papillomas by in situ hybridization with radioactive probes. Hum. PathoI. 18:1115, 1987. 7. McDonnell, J. M., Mayr, A. J., and Martin, W. J.: DNA of human papillomavirus type 16 in dysplastic and malignant lesions of the conjunctiva and cornea. N. EngI. J. Med. 320:1442, 1989.

July, 1991

8. Lauer, S. A., Malter, J. S., and Meier, J. R.: Human papillomavirus type 18 in conjunctival intraepitheliaI neoplasia. Am. J. Ophthalmo!. 110:23, 1990. 9. McDonnell, J. M., McDonnell, P. J., Stout, W. c., and Martin, W. J.: Human papillomavirus DNA in a recurrent squamous carcinoma of the eyelid. Arch. OphthalmoI. 107:1631, 1989. 10. Mannis, J. M.: Chlamydial diseases. In Kaufman, H. E., Barron, B. A., McDonald, M. B., and Waltman, S. R. (eds.): The Cornea. New York, Churchill Livingstone, 1988, pp. 201-216. 11. Shibata, D. K., Arnheim, N., and Martin, W. J.: Detection of human papilloma virus in paraffinembedded tissue using the polymerase chain reaction. J. Exp. Med. 167:225, 1988. 12. Melchers, W., van den Brule, A., Walboomers, J., de Bruin, M., Burger, M., Herbrink, P., Meijer, c., Lindeman, J., and Quint, W.: Increased detection rate of human papillomavirus in cervical scrapes by the polymerase chain reaction as compared to modified FISH and Southern blot analysis. J. Med. Virol. 27:329, 1989. 13. Young, L. S., Bevan, I. S., Johnson, M. A., Blomfield, P. I., Bromidge, T., Maitland, N. J., and Woodman, C. B. J.: The polymerase chain reaction. A new epidemiological tool for investigating cervical human papillomavirus infection. Br. Med. J. 298:14, 1989. 14. Erie, J. C., Campbell, R. J., and Liesegang, T. J.: Conjunctival and corneal intraepithelial and invasive neoplasia. Ophthalmology 93:176, 1986. 15. Smith, R. E., and Nozik, R. A.: Syphilis, tuberculosis and miscellaneous infections. In Smith, R. E., and Nozik, R. A.(eds.): Uveitis. A Clinical Approach to Diagnosis and Management. Baltimore, Williams and Wilkins, 1989, pp. 213-221. 16. Galloway, D. A., and McDougall, J. K.: Human papillomaviruses and carcinomas. Adv. Virus Res. 37:125, 1989. 17. Ferenczy, A., Mitao, M., Nagai, N., Silverstein, S. J., and Crum, C. P.: Latent papillomavirus and recurring genital warts. N. Eng!. J. Med. 313:784, 1985.