Impact of Serum Antibodies to HPV Serotypes 6, 11, 16, and 18 to Risks of Subsequent Genital HPV Infections in Men: The HIM Study

Genital human papillomavirus (HPV) prevalence among men exceeds 70% in some regions of the world (1), with HPV DNA detected in 29% to 82% of penile cancers (2, 3) and 80% to 100% of genital warts (4, 5). Furthermore, nearly 10,000 new cases of HPV-related oropharyngeal cancers among men are diagnosed in the United States each year (6). Although the antibodies produced following HPV vaccination among men provide protection against future anogenital HPV infections and related diseases (7), it is unclear whether the antibodies produced after natural HPV infection are sufficient to protect against subsequent infection in men.

Among women, antibodies produced in response to natural HPV infection are markers of past infections and have been shown to provide partial immunity against subsequent infections and precancerous lesions (8–10); however, not all studies observed these protective effects (11–13). Differences in study findings may be due to the use of different antibody assays, serum antibody levels, and time since first exposure to HPV (14). Furthermore, the virus-like particle (VLP)–based assay and reagents used in two prior studies (11, 12) were in early stages of investigation to assess the role of naturally acquired antibodies for immunity against subsequent HPV infections. A prospective study of HPV infection among men in Arizona did not show protective effects of circulating HPV antibodies (15). However, this study was limited by a short follow-up time, small sample size, and lack of a quantitative serum antibody assessment. An initial study of 2,187 participants in the multinational HPV Infection in Men (HIM) Study also did not show an association between serum antibodies and reduction in subsequent HPV16 infections (14). However, this study was limited to only one HPV type with a median duration of 2 years follow-up. A recent study among HIV-negative and HIV-positive men also did not show protective effects against subsequent HPV infection for multiple HPV types, but the study was restricted to men who have sex with men (MSM; ref. 16).

In the current study, we provide the first comprehensive evaluation of incident genital HPV 6, 11, 16, 18 (any duration infection and 6-month persistent infections) by baseline antibody status among the entire HIM Study cohort (n = 4,123) followed for a median 4.1 years.

The HIM Study is an ongoing multinational study of the natural history of HPV among men in Tampa, Florida, São Paulo (Brazil), and Cuernavaca (Mexico). Details of this study have been described previously (17). Briefly, healthy men were enrolled at each study site and followed for a median follow-up of 4.1 years, with an average interval of 6.9 months between visits. Men were eligible for the study if they (i) were 18 to 70 years of age; (ii) were residents of one of the study sites; (iii) had no previous diagnosis of penile or anal cancers; (iv) had never been diagnosed with genital or anal warts; (v) had no symptoms of a sexually transmitted infection (STI) and were not receiving treatment for an STI; (vi) were not participating in an HPV vaccine study; (vii) had no history of HIV or AIDS; (viii) had no history of imprisonment, homelessness, or drug treatment during the past 6 months; and (ix) were willing to comply with 10 scheduled visits every 6 months for 4 years with no plans to relocate during that time. Extensive sexual history and health questionnaires were administered using computer-assisted self-interviewing at baseline and at each follow-up visit. All eligible participants signed an informed consent, and approval was obtained from the human subjects committees of the University of South Florida (Tampa, FL), Ludwig Institute for Cancer Research (São Paulo, Brazil), Centro de Referencia e Treinamento em Doencas Sexualmente Transmissíveis e AIDS (São Paulo, Brazil), and Instituto Nacional de Salud Publica de Mexico (Cuernavaca, Mexico). After excluding subjects who reported never being sexually active (n = 235) and subjects who received at least one dose of an HPV vaccine (n = 27), a total of 4,123 HIM Study participants were eligible for inclusion in this study.

A 10-mL venous blood sample was collected at baseline to measure serum antibodies against four HPV genotypes (HPV 6, 11, 16, and 18). Serum anti-HPV antibody assessments for HPV genotypes (6, 11, 16, and 18) were carried out using a VLP-based ELISA (18). Details of this assay have been described previously (19). Briefly, insect cells from recombinant baculoviruses expressing HPV L1 capsid proteins were used to produce HPV VLPs. Absorbance values in optical density were measured to assess seroreactivity. Seroreactivity from children (1–10 years old) was used as a negative control and to estimate the mean and SD of the absorbance value for an antibody-positive specimen. The cut-off point for seropositivity for each HPV genotype was selected as 5 SD above the mean absorbance value among the children. Positive and negative laboratory serum controls were used in each assay run for quality control purposes. Laboratory staffs were blinded to participant's HPV DNA status.

Three prewetted Dacron swabs were used to collect genital cell specimens from the coronal sulcus/glans penis, penile shaft, and scrotum and were later combined to form a single sample (17, 20). All specimens were stored at −80°C until genotyping was conducted. The QIAamp DNA Blood Mini Kit (QIAGEN) was used to extract DNA from genital cell specimens. Roche Linear Array Kits were used for PCR and HPV DNA genotyping to detect 37 different types of HPV (21). The presence of human β-globin was used to assess specimen adequacy with an overall β-globin positivity of >98%.

Demographic characteristics were compared between seronegative and seropositive men at baseline using χ² tests with Monte Carlo estimation of exact P values. Only men who were DNA negative for the respective HPV type at the baseline visit were included in analyses. Separate analyses were carried out for each of the four HPV genotypes (HPV 6, 11, 16, and 18) by duration of infection (any duration for incident infections and ≥6-month duration for persistent infection).

Incidence proportions (all incident HPV 6, 11, 16, and 18 infections regardless of duration) and 6-month persistent incidence proportions were calculated at each 6-month interval among subjects who had one or more follow-up visits after baseline. Follow-up times were calculated on the basis of visit dates, and participants who were late for their visit were included in the appropriate time interval for the analysis. To estimate incidence proportions, the first detection of HPV DNA at a follow-up visit was defined as an incident HPV infection (numerator). The number of subjects who tested negative for HPV DNA at the beginning of each study interval was defined as the at-risk population (denominator). For 6-month persistent infections, a persistent infection was defined as HPV DNA detection at two or more consecutive visits. Persistent infections also included subjects with an intervening negative result (n = 20 for HPV 6, n = 5 for HPV 11, n = 22 for HPV 16, and n = 11 for HPV 18). Participants who were DNA negative for the given HPV type (at risk of acquiring a new infection) at the beginning of each study interval were defined as at-risk subjects for 6-month persistent infection.

Kaplan–Meier curves were constructed for incident and 6-month persistent infections for each of the four HPV genotypes (6, 11, 16, and 18). Cumulative incidence for each genotype was compared between seropositive and seronegative subjects, with the log-rank test used to determine significant differences by serum antibody status. Cox proportional hazard models were used to calculate crude and adjusted HRs (aHR) and 95% confidence intervals (CI). A list of candidate variables was created on the basis of descriptive analyses, previous literature, and assessment of confounding by each variable. Variables were evaluated in both backward and forward stepwise models. Final models included variables based on best-fit approach based on the lowest Akaike information criterion values. As sexual behavior prior to acquisition of incident infections strongly influences HPV acquisition, we included the following sexual behavior variables as time-varying covariates in final adjusted Cox models: lifetime number of female sexual partners, number of new female sexual partners in the past 6 to 12 months, frequency of sexual intercourse with female partners in the past 6 to 12 months, lifetime number of male sexual partners, and number of new male sexual partners in the past 6 to 12 months. Because of the differences in HPV seroprevalence by sexual orientation, stratified analyses were also conducted. To evaluate whether antibody levels were associated with subsequent detection of incident infections, we categorized serum antibody levels into the highest tertile and lowest two tertiles and compared these two groups with seronegative men (9, 16).

Serum antibody and HPV genotyping data were available for 4,103 subjects at baseline. We excluded 250 (HPV6), 55 (HPV11), 311 (HPV16), and 90 (HPV18) subjects who were DNA positive at the baseline visit for the respective HPV type. Therefore, baseline seroprevalence analyses included 3,851 (HPV6), 4,046 (HPV11), 3,790 (HPV16), and 4,011 (HPV18) subjects who were DNA negative for the specific HPV genotype. Among the excluded subjects with prevalent HPV 6, 11, 16, and 18 infections, baseline seroprevalence was 13.6%, 25.5%, 18.6%, and 12.2%, respectively. Among these men positive for a genital HPV infection at baseline, those who cleared their infections did not differ from those who reacquired infection with the same type with respect to baseline serum antibody status for HPV 6 (χ²P = 0.626), HPV 11 (P = 0.468), HPV 16 (P = 0.457), and HPV 18 (P = 0.448).

Overall, baseline HPV6, 11, 16, and 18 seroprevalence among eligible participants was 8.1%, 13.9%, 12.7%, and 10.8%, respectively. Significant differences were observed between seropositive and seronegative men for various covariates (Table 1). Seroprevalence increased with increasing age, with the highest seroprevalence for HPV6 (9.4%), 11 (16.2%), and 16 (16.4%) observed among men ages 31 to 44 and the highest HPV18 seroprevalence (17.9%) among men ages 45 to 70. When we compared seroprevalence across the three age categories, we found no differences between the two older age categories (31–44 and 45–70) for HPV 6, 11, and 16; however, HPV18 seroprevalence was significantly higher among the oldest age group. Brazil had the highest seroprevalence for HPV6 (10.8%), 16 (16.9%), and 18 (13.0%), while Mexico had the highest seroprevalence for HPV11 (18.8%). MSM had the highest seroprevalence for all four genotypes (HPV6, 23.2%; HPV11, 27.6%; HPV16, 28.6%; and HPV18, 29.3%). Seroprevalence was highest among men reporting ≥10 lifetime male sexual partners (30.0%, 37.8%, 39.0%, and 40.4% for HPV 6, 11, 16, and 18, respectively).

For HPV 6, 11, 16, and 18, the total numbers of incident infections were 319 (10.3%), 111 (3.5%), 362 (12.4%), and 200 (6.2%) among seronegative men and 30 (10.6%), 17 (3.3%), 70 (16.7%), and 22 (5.6%) among seropositive individuals, respectively. Cumulative incidence among seropositive individuals was higher after 18 months for HPV16 (P = 0.007, Fig. 1). For HPV11 and 18, incidence proportions were generally lower among seropositive individuals until the 13- to 18-month interval, followed by similar proportions (HPV11 and 18) after 24 months. Kaplan–Meier curves indicated a similar pattern with overall nonsignificant P values (P = 0.76 for HPV11 and P = 0.69 for HPV18) for the association between serostatus and HPV acquisition. For HPV6, cumulative incidence among seropositive and seronegative subjects was similar (P = 0.728).

The total numbers of incident 6-month persistent infections were 100 (3.5%), 37 (1.3%), 120 (4.5%), and 74 (2.5%) among seronegative men and 7 (2.8%), 2 (0.4%), 21 (5.5%), and 2 (0.6%) among seropositive individuals for HPV 6, 11, 16, and 18, respectively. Six-month persistent HPV16 infections did not differ significantly between seronegative and seropositive individuals (Fig. 2). The overall number of 6-month persistent HPV11 and 18 infections observed was low (two infections for each type over 4.1-year follow-up period). Six-month persistent infections were not observed prior to month 31 among HPV11-seropositive individuals and prior to month 19 for HPV18-seropositive individuals. Significantly lower incidence of 6-month persistent HPV18 infections was observed among seropositive individuals (P = 0.02). Median duration of persistence was 28.8 months, 37.1 months, 34.8 months, and 46.2 months for HPV 6, 11, 16, and 18, respectively. Duration of persistent infection did not differ by seropositivity for HPV6 (t test P = 0.586), HPV11 (P = 0.551), HPV16 (P = 0.769), and HPV18 (P = 0.132).

Risk of an incident HPV16 infection was significantly higher among seropositive compared with seronegative men (aHR, 1.37; 95% CI, 1.01–1.86; Table 2). A similar nonsignificant pattern was observed for risk of 6-month persistent HPV16 infections (aHR, 1.26; 95% CI, 0.79–2.01). Serostatus was not associated with risk of HPV 6, 11, and 18 incident infections. However, risk of persistent HPV18 infection was significantly lower among seropositive men in unadjusted analyses (crude HR, 0.22; 95% CI, 0.06–0.91) but failed to reach significance in the adjusted model (aHR, 0.19; 95% CI, 0.03–1.37). In separate analyses adjusting for demographic and sexual behavior variables one at a time, similar results to those shown in Table 2 were observed.

Table 3 shows the risk of incident and 6-month persistent HPV16 and 18 infections by serostatus and sexual orientation. No significant associations were observed, likely due to the relatively small sample sizes in each stratum. No evidence of effect modification was observed for HPV18. However, serostatus was nonsignificantly associated with a decreased risk among MSM and increased risk among heterosexual men. Similarly, no significant associations were observed for HPV6 and 11 (data not shown). In analyses by antibody levels, high levels of antibody were not significantly associated with reduced risk of subsequent genital HPV infection, with risk estimates similar to those presented for the whole cohort in Table 2 (data not shown). To address the issue of small sample size and to increase power, we conducted analyses by combining seropositive individuals for any of the four HPV types and compared them with seronegative individuals for all four HPV types. These analyses did not indicate significantly higher HR for subsequent HPV16 incident infection, but for persistent HPV18 infection, a significant protective effect was observed in adjusted models (aHR, 0.42; 95% CI, 0.19–0.93; Table 4). Sensitivity analyses were conducted by (i) restricting the analysis to men who were negative to all 4 vaccine types at baseline and (ii) restricting the analysis to men who were ages 18 to 50 years. The magnitude of the association for the various HPV types did not differ from what is presented in Table 2. However, as the sample size was reduced in these analyses, these associations were no longer statistically significant for risk of persistent HPV18 infection. Furthermore, in a separate analysis adjusting for the same set of adjustment variables for each of the four HPV types and forcing the country variable into the model, results did not change from what is presented in Table 2. 

This is the first study to report the effect of baseline anti-HPV antibodies for subsequent HPV6 and 11 infections among men, along with data for HPV16 and 18. In this study, baseline seropositive status was not associated with a reduction in subsequent incident or 6-month persistent HPV 6, 11, and 16 infections. In fact, HPV16-seropositive men had a higher HPV16 incident infection rate compared with seronegative men. A possible protective effect was observed for HPV18 persistent infections among seropositive individuals, but the small number of infections requires a cautious interpretation of these results. In general, natural immunity after HPV infection was not associated with a reduced risk for subsequent HPV infections among men.

This is the first study to report statistically significant higher risk of incident genital HPV16 infections and possible lower risk of persistent HPV18 infections among seropositive men. In the few studies that have assessed the association of antibodies generated following natural HPV infection and protection against subsequent genital HPV infections in men, results have either been inconclusive or have not shown an association (14–16). The current study is an extended analysis of previous work from our group (14), in which we showed a nonsignificant increased risk for incident (aHR, 1.22; 95% CI, 0.83–1.79) and persistent HPV16 genital infections (aHR, 1.05; 95% CI, 0.51–2.16).

A higher overall incidence of HPV16 infections was observed in the current study. However, when examined over time, a similar or lower genital HPV16 incidence was observed among seropositive and seronegative men up to month 18 of follow-up. Similar initial protective effects against HPV16 and other HPV genotypes have been documented among women, followed by gradual waning of naturally induced IgG antibody levels over time (22–28). High-risk sexual behavior among seropositive men, compared with seronegative men, may have led to higher rates of incident infections among them. However, if this was the case, higher incidence among seropositive men should have been observed for other HPV types as well, and not just for HPV16. Although we adjusted for sexual behavior variables as time-varying covariates, there is still a possibility that residual confounding occurred. There is also a possibility of not capturing precise sexual behavior–related risk due to error in reporting sexual behavior in self-reported questionnaire. Higher risk of incident and persistent infections among HPV16-seropositive individuals may also indicate possible reactivation of prior latent infection (29).

In this study, incident and persistent HPV11 and 18 infections were generally lower among seropositive subjects. Lower incidence (especially for persistent HPV18) may be suggestive of a transient protective effect, which gradually wanes after 18 months, as the Kaplan–Meier analyses indicated. However, due to the small number of HPV18 events, this association failed to reach statistical significance. Additional studies may be needed to understand the role of anti-HPV18 antibodies among men.

Factors including differential immune response by gender, differences in assay techniques, and different durations of antibody levels have been suggested for the difference in antibody response observed among men and women (7, 14, 23, 30). Numerous studies have indicated a higher seroprevalence of both oncogenic and nononcogenic HPV types among women compared with men (31–37). A lower level of antibody response following infection of keratinized epithelium among heterosexual men compared with mucosal epithelial infection in heterosexual women may be a possible explanation for this gender difference (14). Inclusion of nonneutralizing antibodies in our study may have led to overestimation of overall serum antibody levels and may have biased the results toward the null. Serum antibody cut-off levels may also affect the observed association. We therefore carried out analyses using the highest versus the lowest two tertiles of serum antibodies, and the results were not significantly different.

In our analysis, difference in seropositivity seen by anatomic genital location was not a marker for circumcision status. Because of overall large sample size and low seroprevalence among men, significant differences by seropositivity were observed for many covariates, including circumcision status, but circumcision status was not associated with HPV serostatus and risk of subsequent infection in further analyses. The highest seroprevalence was observed among men reporting ≥10 lifetime male sexual partners and having two or more new male sexual partners. These results are consistent with our previous report, which indicated that same sex anal intercourse among men is an independent risk factor for seroprevalence for HPV 6, 11, 16, and 18 (38).

To the best of our knowledge, our study is the first to assess the effect of serum antibody status for all four major vaccine type HPVs in a large, multicenter cohort of men with a long-term follow-up. Despite these strengths, some limitations may have influenced the results. Serum antibody levels were determined at a single baseline visit in our study. Men with prior exposure may have been misclassified as seronegative due to lower seroconversion rates, waning antibody response, and time lag between exposure and antibody response. This misclassification is likely to be nondifferential and may have attenuated findings toward the null. Furthermore, we did not have antibody information for participants at subsequent visits to assess changes in antibody status during the follow-up. Misclassification of HPV infection is also a possibility, although we utilized robust methods for DNA detection that are well established (17). Furthermore, single baseline visit DNA-negative subjects were assumed to be the at risk population and were considered eligible for the study. We ran a separate analysis using more stringent criteria of requiring two negative visits (baseline and first 6-month visit) to qualify for entry into the analyses. Use of this stringent criterion also did not alter study results. Information reported on the eligibility survey prior to study enrollment was self-reported and may have led to inclusion of older men with a history of genital warts; as such, this is a potential limitation for our study.

We did not have information about infections that may have occurred prior to the study start. These infections may not have cleared but instead remained latent at low copy numbers (29). Latency for HPV infection has been described in animal models, and recently, reactivation of latent infection has been demonstrated among immunocompromised individuals (29). Assessment of previous infections is challenging, especially among men who have been sexually active for a long period of time. Furthermore, current HPV DNA detection methods may not be sensitive enough to detect low viral load of latent infection or reactivations that last for a short time period (14). Future studies are needed to explore the effect of latent HPV infections and their reactivation among healthy individuals.

This is the first study to provide results about the effect of anti-HPV antibodies on subsequent risk of HPV6 and 11 infections, which are responsible for 90% of the genital warts (5). Genital warts associated with HPV6 and 11 account for the majority of HPV-related diseases among men. This study indicates that men with naturally acquired HPV6 and 11 antibodies remain susceptible to subsequent infections from these genotypes. Furthermore, our results confirm that seropositive men remain susceptible to subsequent HPV16 and 18 infections as well. Only two previous studies have examined HPV16 and 18 among men, which include a small study (n = 285) of men in Tucson Arizona (15) and a study in the Netherlands among MSM (16). This is the first multinational study examining HPV16 and 18 among a general population. Our results have important public health implications, supporting the need for HPV vaccination of both males and females. HPV infection is readily transmitted between sexual partners, and male HPV infection significantly contributes to infection and subsequent disease in women (39). As natural antibodies do not protect men from subsequent HPV infection and vaccination coverage is low among men (13.9% for all three doses) and women (37.6% for all three doses; ref. 40), it is important to promote vaccination and other preventive methods to target HPV-associated diseases in both the sexes.

In conclusion, among men participating in the HIM study, baseline seropositivity after natural infection with HPV 6, 11, and 16 was not associated with protection against subsequent type-specific genital infection, with a possible protective effect against persistent HPV18 infections. For HPV16, a significantly higher risk of incident infection was observed among seropositive men. The effect of anti-HPV antibodies may be influenced by duration of the antibody response, time period between initial and subsequent genital infection, and overall antibody levels. These results highlight the importance of HPV prevention measures, such as vaccination, which are proven to protect against subsequent HPV infections and related genital diseases.

S.L. Sudenga reports receiving a commercial research grant from Merck & Co. L.L. Villa has received speakers bureau honoraria from and is a consultant/advisory board member for Merck, Sharp & Dohme. A.R. Giuliano is a consultant/advisory board member for Merck. No potential conflicts of interest were disclosed by the other authors.

Conception and design: S.J. Pamnani, S.L. Sudenga, R. Viscidi, L.L. Villa, A.R. Giuliano

Development of methodology: S.J. Pamnani, R. Viscidi, D.J. Ingles, E. Lazcano-Ponce, J. Salmeron, Y. Huang, A.R. Giuliano

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Viscidi, D.J. Ingles, M. Abrahamsen, L.L. Villa, J. Salmeron, M. Quiterio, A.R. Giuliano

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.J. Pamnani, S.L. Sudenga, D.E. Rollison, B.N. Torres, J. Salmeron, Y. Huang

Writing, review, and/or revision of the manuscript: S.J. Pamnani, S.L. Sudenga, R. Viscidi, D.E. Rollison, D.J. Ingles, M. Abrahamsen, L.L. Villa, E. Lazcano-Ponce, J. Salmeron, Y. Huang, A. Borenstein

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.J. Pamnani, J. Salmeron

Study supervision: S.L. Sudenga, M. Abrahamsen, L.L. Villa, E. Lazcano-Ponce, J. Salmeron, Y. Huang, A.R. Giuliano

The authors thank the HIM Study teams in the United States (Hui-Yi Lin, Jane Messina, Christine Pierce Campbell, Bradley Sirak, Christine Gage, Kim Isaacs, Kayoko Kennedy, Andrea Bobanic, Shams Rahman, Matthew Schabath, Alan Nyitray, and Julie Rathwell), Brazil (Maria Luiza Baggio, Roberto Carvalho da Silva, Lenice Galan, Ricardo Cintra, Filomena Cernicchiaro, Graça Ribeiro, Rosária Otero, Roberta Bocalon, Juliana Antunes, Fernanda Silva, Rossana Terreri, and the CRT-DST/AIDS nursing team), and Mexico (Aurelio Cruz Valdez, René de Jesús Alvear Vásquez, Oscar Rojas Juárez, Rossana del Carmen González Sosa, Rosangel Ríos Vences, Martha Huerta Segura, Alicia Rodríguez Galván, Paula Román Rodríguez, Ana Laura Landa Vélez, Griselda Díaz García, Verónica Chávez Abarca, Gisela Flores Quevedo, María del Pilar Hernández Nevárez, Guillermina Sánchez Martínez, Adriana Ortiz Rojas, and Carlos Omar Barrera Flores).

A.R. Giuliano, B.N Torres, D.J. Ingles, M. Abrahamsen, L.L. Villa, E. Lazcano-Ponce, J. Salmeron, and M. Quiterio received financial support from the NCI, NIH CA R01CA098803 (principal investigator: A.R. Giuliano).

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.