Human Papillomavirus 16 E6 Antibodies in Individuals without Diagnosed Cancer: A Pooled Analysis

A rapid increase in the incidence of oropharyngeal cancer has been reported in many parts of the world with a high development index (1–8), especially among men younger than 60 years of age (9). This upsurge has been attributed to an increase in human papillomavirus (HPV)–driven oropharyngeal cancers (7). In the United States, incidence has increased by more than 200% over the past several decades (10). HPV16 infection alone accounts for approximately 90% of HPV-positive oropharyngeal cancers (11, 12) and is estimated to be responsible for at least 50% of oropharyngeal cancer cases in parts of the world with a high development index (10, 13, 14).

Unlike cervical cancer, a precursor lesion for oropharyngeal cancer has yet to be identified, making early detection of oropharyngeal cancers difficult (15). However, numerous case–control studies have shown that the presence of circulating HPV antibodies is strongly associated with cancer of the oropharynx (12, 16–24). Recently, HPV16 E6 antibody positivity has been identified as a potentially promising marker for oropharyngeal cancer (25). A prospective study conducted with prediagnostic sera found that 35% of patients with oropharyngeal cancer were seropositive for HPV16 E6 compared with only 0.6% of controls; these antibodies were present up to 10 years before diagnosis and were not associated with cancers at other head and neck cancer sites (25). The specificity of HPV16 E6 marker for detection of oropharyngeal cancer makes biologic sense considering that the oropharynx (unlike other anatomic sites of the head and neck) is rich in lymphoid tissue and therefore is more likely to induce an antibody response to HPV infection.

Because of the high specificity of HPV16 E6 seropositivity for oropharyngeal cancer, this marker has the potential to be further developed into a screening tool for identifying high-risk individuals. Therefore, characterization of HPV16 E6 seropositivity within healthy individuals without diagnosed cancer is merited. However, HPV16 E6 seropositivity is extremely rare among healthy individuals without cancer (<1%), making it difficult to study (23–25).

To overcome this issue, we conducted a descriptive epidemiologic analysis of pooled controls from several studies of cancer and HPV seropositivity whose samples were all tested within the same laboratory with a bridging panel that allowed for interpretation across studies (23–25). The goals of this analysis were to investigate demographic and serologic factors associated with HPV16 E6 seropositivity among individuals without diagnosed cancer.

Our analytic population consisted of 4,666 controls pooled from four large studies of HPV seropositivity; three studies of head and neck cancer and one study of anogenital cancers (23–26). Controls were pooled from (i) two nested case–control investigations within the European Prospective Investigation into Cancer and Nutrition (EPIC); one focused on head and neck cancer (n = 1,599 controls), and one focused on HPV-driven anogenital cancers (n = 718 controls; refs. 25, 26); (ii) one nested case–control investigation within the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO; n = 924 controls; unpublished data); and (iii) one case–control study, the Alcohol-Related Cancers and Genetic Susceptibility in Europe (ARCAGE; n = 1,425 controls; ref. 23). Informed consent was obtained from all participants in each study, and each study was approved by their respective institutional ethics review boards.

In brief, 521,330 individuals were recruited to the cohort between 1992 and 2000 from 10 European countries, of whom 385,747 contributed a blood sample (27). Participants completed self-administered questionnaires on lifestyle factors and diet. Two control participants (one in Denmark) were randomly assigned for each patient with cancer (including head and neck and anogenital cancers) from appropriate risk sets consisting of all cohort participants alive and free of cancer (except non-melanoma skin cancer) at the time of diagnosis (and hence, age) of the index case. Additional study specific matching criteria are discussed below.

Matching criteria were: country, sex, date of blood collection (±1 month, relaxed to ±5 months for sets without available controls), and date of birth (±1 year, relaxed to ±5 years for sets without available participants). The final study included 1,599 controls (25).

Matching criteria included: study center, sex, date of blood collection (±3 months, relaxed to ±6 months for sets without available controls), and age at blood collection (±3 months, relaxed to ±2 years for sets without available controls), fasting status, and where relevant, menopausal status, and postmenopausal hormone replacement therapy use, and menstrual cycle. The final study included 718 controls (26).

PLCO recruited approximately 155,000 55- to 74-year olds from the general population during 1993–2001 who had not been diagnosed previously with prostate, lung, colorectal, or ovarian cancer. Blood (screening arm only), demographic, and behavioral data were collected (28). Blood samples were obtained at baseline and five subsequent annual visits; the earliest available sample was used for this study. For each case, four controls were randomly chosen from appropriate risk sets consisting of all cohort members alive and free of cancer (except non-melanoma skin cancer) at the time (and hence age) of diagnosis of the index head and neck cancer case. Matching criteria were: year of entry into the study, year the material was collected, study year of cancer diagnosis (for cases; the same year was used for the matched control), birth year, and smoking status (never, former, or current; unpublished data).

Briefly, 2,227 control subjects were recruited from 10 European countries during the period from 2002 to 2005 using a standardized protocol in all centers (except France; ref. 29). All subjects underwent personal interviews to record lifestyle exposures; details are described elsewhere (29). Controls were recruited in each center and frequency matched for age, sex, and area of residence to cases with head and neck cancer. ARCAGE centers mainly used hospital controls to facilitate collection of blood samples, with the exception of the UK centers that used population-based controls randomly chosen from the same family medical practice list as the corresponding cases. Hospital controls were selected from the following diagnoses: endocrine and metabolic, skin, subcutaneous tissue and musculoskeletal, circulatory, nervous system diseases, genitourinary, gastrointestinal, ear, eye and mastoid, plastic surgery cases, and trauma patients (23).

Given the variation between studies in how the covariates were ascertained, we were unable to create single unified definitions of smoking and alcohol consumption. As a result, individuals were categorized according to study-specific definitions of smoking and alcohol consumption (Supplementary Table S1).

Plasma (EPIC and ARCAGE) and serum (PLCO) samples were sent on dry ice to the German Cancer Research Center (DKFZ, Heidelberg, Germany) and testing was performed using multiplex assays (24, 30–32). Samples were analyzed for HPV16 antibodies to the major capsid protein (L1), the early oncoproteins (E6 and E7), and other early proteins (E1, E2, and E4) with the exception of PLCO where seroreactivity against HPV16 E4 was not assessed. In addition, seroreactivity against the E6 protein from the following HPV types was also evaluated; HPV6, HPV11, HPV18, HPV31, HPV33, HPV45, and HPV52 with the exception of the EPIC anogenital study where seroreactivity against HPV52 was not assessed. A bridging panel was included in all studies so that mean fluorescent intensity (MFI) values could be normalized to account for variation between assays. Briefly, this bridging panel consisted of a subset of 188 samples with known seroprevalence for each HPV antigen as defined by standard MFI cutoffs. For each individual analysis, study-specific cutoff values were calculated for each antigen such that the seroprevalence of each HPV antigen within the bridging panel was the same across all studies. If the study-specific cutoff differed from the standard MFI cutoff by more than 10% for a particular antigen, the MFI values for that antigen were normalized by multiplying each value by the quotient of the standard MFI cutoff divided by the study-specific MFI cutoff. This normalization allowed for the same standard MFI cutoff values for seropositivity to be applied across all three studies.

For HPV16 E6, two mutually exclusive categories of HPV16 E6 seroreactivity were created: (i) moderate: MFI ≥ 484 and < 1,000; and high: MFI ≥ 1,000. Previous work from our group showed that increasing the seropositivity cutoff of HPV16 E6 from 484 to 1,000 resulted in an increased specificity for oropharyngeal cancer without an associated decrease in sensitivity (25). For the other HPV16 proteins, the MFI cutoffs used to define seropositivity were: L1, 422; E1, 200; E2, 679; E4, 876; E6, 484; E7, 548. For the E6 proteins of non-HPV16 types, the MFI cutoffs for seropositivity were: HPV6 E6, 500; HPV11, 260; HPV18, 243; HPV31, 890; HPV33, 253; HPV35, 260; HPV45, 249; HPV52, 271.

Characteristics of the control participants were evaluated overall and by the study. The proportion of HPV16 E6 seropositive controls by demographic categories was computed for the pooled studies. Demographic and serologic determinants of moderate and high HPV16 E6 seroreactivity were evaluated through odds ratios (OR) and 95% confidence intervals (CI) calculated in univariate analyses by logistic regression. Demographic variables evaluated included: gender, age, world region, smoking status, and alcohol consumption. Serologic variables evaluated included seroreactivity against: (i) other HPV16 proteins (L1, E1, E2, E4, and E7) and (ii) E6 proteins from non-HPV16 types (HPV6, HPV11, HPV18, HPV31, HPV33, HPV45, and HPV52).

A total of 4,666 controls without diagnosed cancer were included in this analysis (Table 1); 1,425 (30.5%) individuals from ARCAGE; 2,317 (49.7%) from EPIC; and 924 (19.8%) from PLCO. The majority of the controls were male (63.7%), 60 years of age or younger (59.6%) and ever alcohol drinkers (80.3%); smoking status appeared evenly distributed among the categories (i.e., never, former, and current). Small differences between studies were noted for gender, age, and smoking status (Table 1).

HPV16 E6 seropositivity was rare. Of the 4,666 pooled controls, a total of 32 individuals (0.7%) were seropositive for HPV16 E6. Prevalence of HPV16 E6 seropositivity was similar by study; ARCAGE (0.8%), EPIC (0.6%), and PLCO (0.9%). Of the 32 HPV16 E6 seropositive controls, 17 (0.4%) were classified as having moderate HPV16 E6 seropositivity (MFI ≥ 484 and <1,000) and 15 (0.3%) were classified as having high HPV16 E6 seroreactivity (MFI ≥ 1,000; Table 2).

Age, gender, smoking status, and alcohol consumption did not elevate the odds of moderate or high HPV16 E6 seroreactivity. Only former smoking was significantly associated with high HPV16 E6 seroreactivity (OR, 5.5; 95% CI, 1.2–51.8). No other significant associations for either moderate or high HPV16 E6 seroreactivity were observed (Table 2).

Seroreactivity against HPV16 proteins, including L1, E1, E2, E4 or E7, was common. Similar seroprevalence for any of these proteins was observed among the HPV16 E6 seronegative and moderately HPV16 E6 seroreactive controls; 27.9% and 29.4%, respectively (Table 3). No significant associations between moderate HPV16 E6 seroreactivity and seroreactivity against any of the other HPV16 proteins either individually or combined was observed.

In contrast, prevalence of seroreactivity against 1 or more HPV16 proteins (L1, E1, E2, E4, or E7) in addition to E6 was greatest among the controls with high HPV16 E6 seroreactivity; 46.7%. Of the 5 HPV16 proteins evaluated, seroreactivity against HPV16 L1 was most common (n = 5 out of 15, 33.3%). Controls with high HPV16 E6 seroreactivity were also more likely than HPV16 E6 seronegative controls to be seroreactive against all HPV16 proteins with the exception of E7; however, only HPV16 L1 (OR, 4.3; 95% CI, 1.1–13.8) and E2 (OR, 7.7; 95% CI, 1.4–29.1) reached statistical significance. High HPV16 E6 seroreactivity was also significantly associated with seroreactivity against multiple HPV16 proteins (OR, 25.3; 95% CI, 2.6–119.6) for seroreactivity against 3 HPV16 proteins in addition to E6, although in absolute terms, seroreactivity against multiple HPV16 proteins was rare (4 of 4,666; 0.08%).

Of note, 7.9% of HPV16 E6 seronegative controls were seroreactive against a non-HPV16 E6 protein compared with approximately 20% for both the moderate and high HPV16 E6 seroreactive controls. Only seroreactivity for HPV33 E6 was significantly associated with high HPV16 E6 seroreactivity (OR, 17.7; 95% CI, 1.9–81.8; Table 4).

In this large descriptive epidemiologic analysis of more than 4,000 pooled controls from several studies of HPV seroreactivity and HPV-associated cancer (23–25), HPV16 E6 seropositive controls were rare (<1%), particularly those with high HPV16 E6 seroreactivity (0.3%). Furthermore, of the determinants of HPV16 seropositivity evaluated, significant associations were observed only among controls with high HPV16 E6 seroreactivity. Of the demographic risk factors assessed, none were predictors of HPV16 E6 seropositivity except for former smoking. Of the serologic determinants assessed, seroreactivity against other HPV16 proteins was common among all controls and was greatest among controls with high HPV16 E6 seroreactivity (47%). A marker of cumulative lifetime HPV16 exposure, HPV16 L1 seropositivity (5 out of 15; 33.3%) was most commonly detected among high HPV16 E6 seroreactive controls compared with the other four HPV16 proteins tested. High HPV16 E6 seroreactivity was significantly associated with seroreactivity against HPV16 L1, E2 and multiple HPV16 proteins. Seropositivity for any non-HPV16 E6 proteins was less common; 8% among HPV16 E6 seronegative controls and approximately 20% among controls with both moderate and high HPV16 E6 seroreactivity. Only HPV33 E6 seroreactivity was significantly associated with high HPV16 E6 seroreactivity; however, due to its high sequence homology with the E6 protein of HPV16, this finding may be the result of antibody cross-reactivity (33).

On the basis of studies of cervical cancer (34), individuals without an underlying HPV-driven cancer would not be expected to have antibodies against the HPV16 oncoproteins. One potential explanation for the small percentage of controls with seroreactivity against HPV16 E6 may be due to potential laboratory error or sample misclassification. Therefore, the 15 strongly HPV16 E6 seroreactive controls identified may reflect the error rates within these large epidemiologic studies. Alternatively, the HPV16 E6 seroreactive controls identified in our study may be harboring a yet to be diagnosed cancer or precancer. Recent findings have suggested that induction of HPV-specific antibodies, most notably HPV16 E6 antibodies, may be a response to an underlying HPV-driven neoplastic process that may take years to manifest into a diagnosable cancer (25). Studies conducted at the time of diagnosis have shown that the presence of detectable HPV16-specific antibodies in diagnostic serum is highly sensitive for HPV-driven head and neck squamous cell carcinomas (35). A large proportion (7 out of 15) of controls with high HPV16 E6 seroreactivity was also seroreactive against at least one other HPV16 protein in addition to HPV16 E6. However, of all HPV16 proteins, a recent prospective study showed that seroreactivity against HPV16 E6 is the most strongly associated with development of oropharyngeal cancer (OR, 274; 95% CI, 110–681). Of note, all the HPV16 E6 seropositive oropharyngeal cancer cases identified in the previous study had HPV16 E6 MFI values greater than 1,000 and therefore would have been classified as having high seroreactivity in this current analysis (25). Taken together, these findings raise the possibility that the HPV16 E6 seropositive controls with high E6 seroreactivity described in this study, especially those with antibodies against multiple HPV16 proteins, may be on the path to developing an HPV-driven cancer or precancer that, as follow-up accrues may eventually be diagnosed. Updated record linkage of the EPIC head and neck cancer study (25) revealed that one HPV16 E6 seropositive control was subsequently diagnosed with invasive anal cancer, however, additional follow-up time is needed to fully investigate this possibility.

Therefore, an important limitation of this study is that we are unable to extend follow-up to further ascertain the health status of the HPV16 E6 seropositive controls. For the cohort studies (EPIC and PLCO), we will continue to monitor record linkage updates and investigate this question accordingly; for the ARCAGE case–control studies, no additional follow-up will become available. In addition, controls in this analysis were initially matched to cases, therefore skewing their distribution of certain variables, such as age and gender, toward that of cases. For the case–control studies, controls were recruited only when they were eligible based on a list of nonchronic diseases unrelated to smoking and alcohol. Although this does not jeopardize the validity of our findings, it limits their generalizability. There may have also been some minor misclassification in terms of smoking status and alcohol consumption due to the differences between studies in how these variables were ascertained on the questionnaires. In addition, we do not have information regarding other covariates such as sexual behavior, and host immunogenetics, which may have been potentially helpful in explaining why some individuals develop these antibodies. Finally, even with over 4,000 pooled controls, we were still limited by power given that our outcome was so rare.

In conclusion, given the rarity of HPV16 E6 seropositivity among individuals without diagnosed cancer, our data suggests that HPV16 E6 serology results in an extremely low rate of misclassification among controls potentially implying that serology may have a higher specificity than other biomarkers in oropharyngeal cancer cases as well. Because of the high specificity of the HPV16 E6 marker (25), HPV16 E6 antibody testing may have the potential to identify individuals at high-risk for developing HPV-driven oropharyngeal cancer within a general population of older adults.

X. Castellsagué reports receiving a commercial research grant from Merck & Co. and Genticel; has received speakers bureau honoraria from Vianex and SPMSD; and is a consultant/advisory board member for SPMSD. No potential conflicts of interest were disclosed by the other authors.

Conception and design: K.A. Lang Kuhs, M. Johansson, P. Brennan, W. Ahrens, L. Simonato, F. Merletti, K. Kjaerheim, T.V. Macfarlane, X. Castellsagué, W.-Y. Huang, V. Krogh, H.B. Bueno-de-Mesquita, E. Weiderpass, A. Tjønneland, M. Pawlita, A.R. Kreimer

Development of methodology: K.A. Lang Kuhs, D. Anantharaman, T. Waterboer, H.B. Bueno-de-Mesquita, E. Weiderpass, M. Pawlita, A.R. Kreimer

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): T. Waterboer, M. Johansson, P. Brennan, A. Michel, M. Willhauck-Fleckenstein, M.P. Purdue, I. Holcátová, W. Ahrens, P. Lagiou, J. Polesel, L. Simonato, F. Merletti, C.M. Healy, K. Kjaerheim, D.I. Conway, T.V. Macfarlane, P. Thomson, X. Castellsagué, A. Znaor, A. Black, W.-Y. Huang, V. Krogh, A. Trichopoulou, H.B. Bueno-de-Mesquita, F. Clavel-Chapelon, E. Weiderpass, J. Ekström, A. Tjønneland, M.-J. Sánchez, R.C. Travis, M. Pawlita, A.R. Kreimer

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K.A. Lang Kuhs, T. Waterboer, M. Johansson, P. Brennan, M. Willhauck-Fleckenstein, X. Castellsagué, W.-Y. Huang, E. Weiderpass, A. Hildesheim, M. Pawlita, A.R. Kreimer

Writing, review, and/or revision of the manuscript: K.A. Lang Kuhs, D. Anantharaman, T. Waterboer, M. Johansson, P. Brennan, A. Michel, M. Willhauck-Fleckenstein, M.P. Purdue, W. Ahrens, P. Lagiou, J. Polesel, L. Simonato, F. Merletti, C.M. Healy, K. Kjaerheim, D.I. Conway, T.V. Macfarlane, P. Thomson, X. Castellsagué, A. Black, W.-Y. Huang, V. Krogh, A. Trichopoulou, H.B. Bueno-de-Mesquita, E. Weiderpass, J. Ekström, E. Riboli, A. Tjønneland, M.-J. Sánchez, R.C. Travis, A. Hildesheim, M. Pawlita, A.R. Kreimer

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K.A. Lang Kuhs, D. Anantharaman, M. Johansson, A. Michel, F. Merletti, X. Castellsagué, E. Weiderpass, J. Ekström, M.-J. Sánchez, A.R. Kreimer

Study supervision: K.A. Lang Kuhs, M. Johansson, W. Ahrens, X. Castellsagué, H.B. Bueno-de-Mesquita, E. Weiderpass, A. Hildesheim, A.R. Kreimer

The authors thank the following people for their contributions to the article: David Castenson (Information Management Services, Calverton MD); Dana Mates (Institute of Public Health, Bucharest, Romania); Vladimir Bencko (Charles University in Prague, Prague, Czech Republic); Victor Wünsch-Filho (School of Public Health, University of Sao Paulo, Sao Paulo, Brazil); Elena Matos (Institute of Oncology Angel H. Roffo, University of Buenos Aires, Buenos Aires, Argentina); and Jose Eluf-Neto (Universidade de São Paulo, Sao Paulo, Brazil). The authors also acknowledge contribution of Prof. Gary J. Macfarlane, Dr. Ann-Marie Biggs, Prof. Martin Tickle, Prof. Phil Sloan, and Prof. Nalin Thakker with study conduct in UK centers.

This work was supported by the Intramural Research Program of the National Cancer Institute (to A.R. Kreimer).

The EPIC study has been supported by the Europe Against Cancer Program of the European Commission (SANCO); Deutsche Krebshilfe, Deutsches Krebsforschungszentrum, German Federal Ministry of Education and Research; Danish Cancer Society; Health Research Fund (FIS) of the Spanish Ministry of Health, Spanish Regional Governments of Andalucia, Asturias, Basque Country, Murcia, and Navarra; Catalan Institute of Oncology, Spain; the ISCIII of the Spanish Ministry of Health (RETICC DR06/0020); Cancer Research UK; Medical Research Council, United Kingdom; Greek Ministry of Health; Stavros Niarchos Foundation; Hellenic Health Foundation; Italian Association for Research on Cancer (AIRC); Italian National Research Council, Fondazione-Istituto Banco Napoli, Italy; Associazione Italiana per la Ricerca sul Cancro-AIRC-Milan; Compagnia di San Paolo; Dutch Ministry of Public Health, Welfare, and Sports; World Cancer Research Fund; Swedish Cancer Society; Swedish Scientific Council; Regional Government of Västerbotten, Sweden; NordForsk (Centre of excellence programme HELGA), Norway; French League against Cancer (LNCC), France; National Institute for Health and Medical Research (INSERM), France; Mutuelle Générale de l'Education Nationale (MGEN), France; 3M Co, France; Gustave Roussy Institute (IGR), France; and General Councils of France (to P. Brennan).

The ARCAGE study was supported by the grant from European Commission's 5th Framework Program (contract QLK1-2001-00182). This project was partly funded by the Health General Directorate of the French Social Affairs and Health Ministry. The serology testing was supported in part by a grant from the European Commission's 7th Framework Program (contract FP7-HEALTH-2011–282562; to P. Brennan).

PLCO was supported by the Intramural Research Program of the Division of Cancer Epidemiology and Genetics and by contracts from the Division of Cancer Prevention, National Cancer Institute, NIH, DHHS.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.