Immunogenicity of Gardasil-9 HPV Vaccine in People Living With HIV

BACKGROUND AND SIGNIFICANCE

HPV is the etiological agent for most if not all cervical and anal cancer and for over 50% of head and neck squamous cell carcinomas. HPV infection, pre-cancerous and cancerous lesions are more prevalent in people living with HIV (PLWHIV), including those with HIV infection that is well-controlled with anti-retroviral therapy. HPV vaccines have been developed based on expression of the L1 major capsid protein to make virus-like particles and demonstrate impressive efficacy, with type-specific protection against persistent infection and incident pre-cancer (cervical/anal intraepithelial neoplasia, CIN/AIN) in HIV uninfected populations. These vaccines include quadrivalent Gardasil (4vHPV) and the more recent nonavalent Gardasil (Gardasil-9).

The 4vHPV has been utilized to immunize PLWHIV. In general, excellent seroconversion rates (over 90%) were seen with slightly lower rates seen for HPV-18, and in those with low CD4 cell count, low CD4 cell count nadir or uncontrolled HIV replication. The seroconversion rates appear to be slightly lower than that seen in HIV-negative populations (>95%). Additionally, titers for HPV-6 and 18 were noted to be lower compared to HIV-negative historical controls in children ages 7-12. This is consistent with observations that vaccination of PLWHIV for influenza, hepatitis A or hepatitis B has resulted in lower than desirable seroconversion rates (50-70%) with higher rates in those on cART and with higher CD4 cell count. To date, only one study has described immunogenicity of Gardasil-9 in adult PLWHIV. Investigators have reported seroconversion to all 9 HPV types in all 100 immune-intact/virally suppressed PLWHIV who received Gardasil-9, and higher antibody GMTs in those with pre-existing seroreactivity. This cohort study conducted in Belgium included predominantly Caucasian (68%) males (85%) with a mean age of 38 years at enrollment. There is a paucity of literature exploring the immunogenicity of Gardasil-9 in diverse demographic populations of adults living with HIV.

Efficacy studies using 4vHPV in HIV-positive adults have provided mixed results at best. One study in HIV-positive men and women who had significant prior HPV exposure showed no efficacy of HPV vaccination in preventing future AIN2/3 development. This may be related to the inclusion of men and women with AIN2/3 at baseline (30% and 50%, respectively) and the older age (average, 47) of this cohort. In another study of females aged 9-65 (average, 39) with 2 years of follow up, 4vHPV showed modest efficacy with rates of 1.0/100 patient-years for a composite endpoint of the development of warts or persistent infection with the vaccine types. This compares to rates of 0.1/100 patient-years for HIV-uninfected women and 1.5 for HIV-negative unvaccinated women. The reason for this modest efficacy is not clear but may be due to deficiencies in cellular immunity and/or the age of the cohorts. In addition, many of these HIV infected individuals have been previously exposed to some of the vaccine types and they often have concurrent infections with other oncogenic HPV types. These excessive HPV exposures may promote CIN/AIN and confound efficacy endpoints. Importantly, HIV positive women with CIN3+ have a nearly 4-fold higher prevalence of non-traditional oncogenic HPV types compared to HIV negative women with CIN3+. This finding suggests that the modest efficacy of 4vHPV in HIV positive women may be improved substantially by use of Gardasil-9. To date, no studies have reported efficacy of Gardasil-9 for prevention of HPV infections, CIN or AIN in adult PLWHIV.

There is clearly an unmet need to evaluate the immunogenicity and immune protection generated by Gardasil-9 vaccination in diverse populations of PLWHIV, who remain at increased risk of HPV infection and related pathology despite adequate management of their HIV. We hypothesize that people with well-controlled HIV will seroconvert following Gardasil-9 vaccination and will develop high titer antibodies to all 9 HPV types. In addition, we postulate that the well-controlled HIV-positive individual will also show short-term efficacy with fewer incident HPV infections and reduced incidence of CIN/AIN attributed to HPV types in the vaccine. This efficacy will primarily be seen in those who are seronegative and HPV DNA negative for a given type at baseline. It is predicted that those who are seropositive but HPV DNA negative for a given vaccine type will demonstrate higher titers of response after vaccination. Efficacy in this subset (seropositive, DNA negative) as well as the other subsets has not been studied in detail. The goal of these proposed studies is to show immunogenicity and short-term efficacy of Gardasil-9 in a well-controlled HIV-positive population.

STUDY DESIGN AND PROCEDURES

This study is a single-center, prospective, intervention trial evaluating the immunogenicity of Gardasil-9 in adult men and women living with adequately managed HIV infection. The study will include both men and women and will reflect the gender, race, ethnicity, sexual orientation and socioeconomic status of the patient population seeking care at University Medical Center.

After informed consent is secured, enrolled participants will receive the 3-dose Gardasil-9 vaccine per product label at 0, 1 and 6 months. Subjects will be observed for at least 30 minutes after the vaccination is given. Participants consenting to the study will complete a sociodemographic survey at the enrollment visit. At 0, 7, 12, and 18-month visits, participants will complete a social history survey and whole venous blood, saliva, and anal swabs will be collected. A cervical/vaginal swab will also be collected from female participants.

A second arm of the study includes men and women with prior receipt of all three doses of Gardasil HPV vaccine (quadrivalent or nonavalent). Individuals are eligible for this single time-point, cross-sectional study if their HIV is adequately managed and date of Gardasil vaccine series completion can be documented. Participants will complete a social history survey and whole venous blood, saliva, and anal swabs will be collected. A cervical/vaginal swab will also be collected from female participants.

Whole blood will be processed to obtain serum, which will be aliquoted and stored at -80C. Serum samples will be tested for HPV antibodies by the competitive Luminex immunoassay (cLIA, Merck) to determine seroconversion rates and antibody titers. Genomic DNA isolates from saliva and swabs will be tested for HPV genotype-specific infection by high-throughput sequencing (HPV MY-Seq). Results of clinical laboratory tests pertinent to the study will be captured from the participant's electronic medical record. For patients who undergo cervical colposcopy or anoscopy with biopsy for clinical care, residual biopsy specimens will be obtained from UMC Pathology for HPV DNA genotype testing.

The immunological response of study participants will be compared to previously published historical (HIV negative adult) controls. Additionally, clinical outcomes of prevalent and incident HPV genotype-specific infection and anogenital dysplasia will be evaluated, comparing those participants who seroconvert/generate high titer antibodies to those who do not seroconvert/generate low titer antibodies, and comparing high-risk HPV genotypes included in the vaccine to high-risk HPV genotypes that are not included in the vaccine.

PLANNED ANALYSIS

The primary outcome of the study is seroconversion to the 9 HPV genotypes in the Gardasil-9 vaccine. This outcome will be analyzed as a binary variable (seropositive/seronegative) and as a continuous numerical value (antibody titers). Secondary outcomes of the study will evaluate the impact of vaccination on HPV infection and disease. Variables of interest include prevalent HPV infection, incident HPV infection, prevalent HPV-associated dysplasia, and incident dysplasia. Clinical outcome variables will also be interpolated from the data, including clearance and/or persistence of HPV infection and clearance, persistence or progression of HPV-associated disease. All disease will be evaluated on a per-site (mucosal tissue) basis and definitions for HPV related variables will depend on genotype-specificity. All variables will be collected at each 6-month timepoint. Time-to-event (e.g. viral clearance, progression of lesion) will also be extracted from the prospective data to evaluate clinical impacts of vaccination. Data will be stratified by demographic and social history variables to identify predictors of seroconversion and clinical outcomes and control for confounding factors.

Statistical Methods: Categorical covariates will be described by reporting counts and percentages, while continuous covariates will be reported using means and standard deviations. Categorical covariate distributions will be compared between seropositive/negative groups using a Fisher exact test, and continuous covariates will be compared using Wilcoxon rank sum tests. To test the primary hypothesis - that seropositivity rates do not differ from published rates of 95% in HIV-negative populations - a one-sided one-sample test of proportions will be used.

In order to determine what demographic factors influence variables in this population after potential confounder adjustment, several regression analyses will be used. For binary variables, logistic regression will be used, while for continuous variables typical linear regression is used. For time to event variables, Cox regression will be used, and for variables that are proportions, Beta regression will be used. When using repeated measurements, independence will be maintained by controlling for subject ID as a random or fixed effect. If necessary, Bayesian methods will be used.

Power/Sample Size: Assuming a seroconversion rate of 95% in HIV-negative patients, 150 enrolled patients would give power of about 99% to detect a seroconversion rate of 80%. The design also has a power of 80% if the true seroconversion rate in Gardasil patients is 89%. The design will not be well powered to detect decreases \<6% in seroconversion rates from the published literature.