Prophylactic human papillomavirus vaccines: will they do their job?

Authors


Margaret Stanley, Department of Pathology, Tennis Court Road, Cambridge CB2 1QP, UK.
(fax: +44 1223 333346; e-mail: mas1001@cam.ac.uk).

Abstract

Abstract.  Stanley M (University of Cambridge, Cambridge, UK). Prophylactic HPV vaccines: will they do their job? (Foresight). J Intern Med 2010; 267: 251–259

The prophylactic HPV L1 virus like particle vaccines are highly efficacious preventing >98% of HPV 16/18 caused high grade cervical intra-epithelial neoplasia (CIN2/3) according to protocol group in phase III randomized clinical trials over approximately 4 years. Data from the trials together with postvaccine surveillance indicates that they have a good safety profile. As the current HPV vaccines include only two of the 15 oncogenic HPVs, they will not eliminate cervical cancer or the other HPV associated ano-genital diseases but could reduce, for example, the incidence of cervix cancer by up to 70–80%, if effectively delivered to adolescent females.

Introduction

Human papillomaviruses (HPVs) were long assumed to be unimportant infections, causing warts which, although cosmetically unsightly excrescences, were of trivial clinical significance. However, this large family of small DNA viruses, it turns out, include individual types that are major global carcinogens causing an estimated 3.7% of cancers mainly in the ano-genital tract Fig. 1. Even the so-called trivial warts, particularly if they occur on the ano-genital skin [1] or in the larynx [2], have significant morbidity and are a substantial health economic burden [3] for their management and treatment.

Figure 1.

 The global burden of HPV associated malignancies [38, 39]. HPV is a carcinogen of squamous epithelia with HPV 16 the major player at all sites.

The global burden of HPV associated ano-genital disease is huge and the development of prophylactic vaccines to prevent or reduce HPV infection and HPV associated disease has been an important objective. This has not been an easy task as the inability to efficiently propagate these viruses in tissue culture was a serious impediment. Despite this, over the past two decades huge progress has been made in the development of prophylactic HPV vaccines culminating in the widespread licensing of two commercial products. Both vaccines significantly reduce infection and intra-epithelial disease caused by the major oncogenic HPV types 16 and 18 and one of the vaccines, in addition, targets the commonest causes of genital warts HPV types 6 and 11.

Prophylactic HPV vaccines

Immune response in natural infections

Human papillomaviruses are exclusively intra-epithelial infections, and in natural infections HPV only infects productively primitive basal cells in squamous epithelia or cells with the potential for squamous maturation such as the reserve cell at the cervical squamo-columnar junction. HPV infection is determined by the detection of HPV DNA, not by virus isolation, and after infection there is a latent period between infection and HPV DNA detection. This interval is of variable duration extending for weeks months or even years and is probably related to the dose of infectious virus. Eventually, the virus is kick started into action, virus replication is initiated, viral DNA can be detected in cervical swabs or genital skin scrapes and clinical lesions become evident. Cycles of virus replication, assembly and shedding, apparently in the absence of an immune response, then continue for a variable period that may extend at least for 12–18 months for the high risk HPV types, but eventually a cell mediated immune (CMI) response is initiated with cytotoxic effectors accompanied or followed by seroconversion and antibody to the major coat protein of the virus L1. The vast majority of infected individuals – 80–90% of women – it is speculated make a good CMI, they become HPV DNA negative and in general remain refractory to re-infection or reactivation by the HPV type against which they have mounted a successful immune response [4]. A minority of women, 10–20%, makes an inadequate CMI, they remain DNA positive with persistent virus infection. These groups of women are at risk for progression to high grade cervical, vulval and vaginal intra-epithelial neoplasia (CIN2/3, VIN3, VaIN3) and therefore invasive cancers at these sites.

The humoral or antibody response to HPV infection is modest. Serum neutralizing antibody responses are made to the major coat protein of the virus L1 and are type specific. The average time to seroconversion for HPV 16 infections is 8–9 months after the first detection of HPV DNA and apparently only 50–70% of women with incident infection seroconvert [5]. Peak serum antibody concentrations are low but at least 25% of women continue to have detectable antibody over a 10 year period postseroconversion [6]. Evidence from the classical experiments in the rabbit by Shope showed that serum neutralizing antibody is protective [7]. Passive immunization experiments in dogs with purified immunoglobulin G from the serum of animals post wart regression demonstrated that the immunized animals were protected against viral challenge and disease [8].

HPV vaccines

These preclinical studies in animals strongly suggested that the neutralizing antibody would be effective in preventing HPV associated disease and were a significant contribution to the scientific case for the development of prophylactic HPV vaccines [9]. HPVs cannot be grown bulk in tissue culture and therefore live or killed viral vaccines are not possible. The current HPV vaccines are sub-unit protein vaccines consist of the L1 coat protein expressed via recombinant yeast or baculovirus vectors. This method of expression results in the assembly of L1 into macromolecular structures – virus like particles – that mimic the wild type virus capsid morphologically and antigenically, but that contain no DNA and are therefore not infectious (Fig. 2). Passive immunization experiments using the dog [10] and rabbit [11] showed that the antibodies generated after immunization with the relevant virus like particles (VLPs) protected against viral challenge and wart formation.

Figure 2.

 HPV vaccines are sub-unit vaccines consisting only of the major capsid protein L1. The L1 protein is expressed via recombinant eukaryotic expression vectors and self assembles into macromolecular structures or virus like particles morphologically and antigenically similar to the wild type virus [21].

Two HPV L1 VLP prophylactic vaccines have been developed. These are Cervarix®, a bivalent HPV 16/18 L1 VLP vaccine from GlaxoSmithKline Biologicals, Rixensart, Belgium and Gardasil® also known as Silgard, a quadrivalent HPV 16/18/6/11 L1 VLP vaccine from Merck & Co. Inc., Whitehouse Station, NJ, USA (Fig. 3). Both vaccines have undergone randomized, placebo controlled, double blind clinical trials (RCTs) in women in North America, Latin America, Europe and Asia Pacific. Both have been licensed in many countries including the nation states of the European Union (EU). Gardasil® has been licensed by the Federal Drugs Agency (FDA) of the USA, since June 2006.

Figure 3.

 Characteristics of the commercially available HPV VLP vaccines.

Vaccine efficacy

Outcomes from the phase III RCTs over an approximately 4 year period are now available for both vaccines (Fig. 4). The design of the RCTs differs significantly for both vaccines in terms of population demographics, baseline inclusion criteria for case assignment, serological and DNA detection methods. However, both vaccines have shown very high efficacy against HPV 16/18, which caused high grade CIN2/3 the end-point accepted as the ethically acceptable proxy for vaccine efficacy against cervical cancer [12]. In the phase III RCT according to the protocol group (TVC-E group) of the bivalent vaccine in women who were 15–25 years of age with six or less lifetime sex partners, sero negative and DNA negative for the relevant oncogenic HPV type in the vaccine at trial entry and at month 6 postimmunization, a vaccine efficacy of 92.9% against HPV 16/18 caused CIN2 and 80% against 16/18 CIN3 was achieved [13]. Multiple HPV infections are a feature of genital tract infection making the assignment of causality to a specific HPV type problematic. Post hoc analysis using the HPV type assignment algorithm (TAA) of HPV 16 or 18 DNA in the lesion and in the preceding cytology samples resulted in an efficacy of 97.7% for CIN 2 and 100% for CIN 3.

Figure 4.

 Outcomes of the phase III randomized control trials of the bivalent and quadrivalent vaccines – efficacy in the per protocol groups [13–15].

In the phase III RCTs of the quadrivalent vaccine the per protocol group (PPG) were women aged 16–24 years with <4 lifetime sex partners and naïve for one or more HPV vaccine genotypes at enrolment through 1 month post the 3rd immunization. In the FUTURE I RCT designed to evaluate efficacy of the quadrivalent vaccine in preventing HPV 6/11/16/18 caused ano-genital disease, vaccine efficacy in the PPG was 100%, i.e. no cases of HPV 6/11/16/18 caused VIN, VaIN or external genital warts detected in the vaccine group over 3–4 years [14]. A combined analysis of RCTs (phase IIb and phase III trials) of the efficacy of the quadrivalent vaccine against HPV 16/18 caused CIN has been published [15]. In these trials involving 20 583 women efficacy against HPV 16/18 caused CIN 2/3 was 99% (one case in the vaccine group) and 100% for HPV 16/18 related adenocarcinoma in situ (AIS). If the TAA used for the bivalent vaccine post hoc analysis is applied to these data, vaccine efficacy against HPV16/18 caused CIN2/3 is 100%.

Vaccine immunogenicity

The vaccines are highly immunogenic inducing high levels of serum antibodies in virtually 100% of individuals. The current assumption is that the major basis for protection is neutralizing antibody and this assumption is supported by animal models that demonstrate protection against viral challenge in animals immunized by passive transfer of antibody [10, 11] and animals immunized with formalin treated VLPs [16]. To date, there is no immune correlate of protection, no antibody threshold or other immune measurement has been defined that correlates with protection. In the phase IIb immunogenicity trial 001/007 of the bivalent vaccine anti HPV 16 and 18 antibody persisted at levels about 10× that observed in natural infections up to 4.5 years postimmunization [17] and evidence presented at medical conferences shows that these levels continue up to 6.5 years at least. Antibody levels induced by the quadrivalent vaccine persist in vaccinees up to 5–6 years postvaccination [18] and mathematical modelling of the kinetics of antibody decay suggests that the detectable antibody could persist for 30 years [19]. In about, 40% of subjects immunized with the quadrivalent vaccine HPV 18 antibody concentrations fall to background levels but efficacy against HPV 18 associated CIN2/3, AIS and VIN/VaIN3 remains at 100% over a 4 year period irrespective of antibody level [20]. However, the competition Luminex bead assay used in the quadrivalent vaccine trials measures only one monoclonal neutralizing antibody species, whereas the conventional ELISA used in the bivalent vaccine trials measures total antibody (both neutralizing and non-neutralizing) [21]. Evidence presented at medical meetings shows that if total antibody concentrations induced by the quadrivalent vaccine are measured all subjects in the vaccine group are seropositive for HPV 18 antibodies 4 years postimmunization.

Importantly, there is good evidence that robust immune memory is generated by these vaccines. The quadrivalent vaccine has shown a good anamnestic/recall response to antigen challenge, the functional read out for memory, 5 years postimmunization [22] and circulating B memory cells can be detected 1 month after the third and final immunization with the bivalent vaccine [23]. Furthermore, the persistence of antibody levels in excess of that found in natural infection for protracted periods strongly suggests robust induction of B and T cell memory. Immune memory is fundamental to successful immunization and the observations of persistence of antibody and robust recall from the VLPs in the trials leads to optimism that the duration of protection might be measured in decades as has been shown for hepatitis B sub-unit vaccines [24].

Cross-protection

In natural HPV infections, the humoral immunity induced is type specific and type specific neutralizing antibodies appear to be the predominant species generated by the VLPs. However, the VLP vaccines induce much higher concentrations of antibody than natural infection and there is considerable amino acid sequence homology in L1 between closely related HPV types [25] raising the possibility of cross-neutralizing antibody species. Cross-protection against nonvaccine types has been now demonstrated for both vaccines. Subjects immunized with the quadrivalent vaccine showed protection against CIN2+ disease caused by several HPV types including HPV 31, 33, 35, 52 and 58 [26, 27]. Very high levels of cross-protection against HPV 31 and HPV 45 caused CIN2+ has been reported for the bivalent vaccine together with lower levels for some other oncogenic types [13] but the duration of such cross-protection is unknown and the mechanism by which such cross-protection is effected is not clear at the present.

Will these vaccines be effective – will they do their job?

The emerging scientific evidence together with the results of the large phase RCTs are clear, the VLP vaccines, if delivered to girls and women (and men) with no evidence of infection with the HPV types in the vaccine at the time of immunization will prevent disease caused by those HPV types. In the best case scenario assuming maximum coverage and long term duration of cross-protection this would result in 20–50 years time in a reduction in incidence of cervical carcinoma of 70–80% leaving 20–30% of cases caused by other oncogenic HPV types. Genital warts are caused predominantly by HPV 6 and 11 and it would be expected that the incidence of these conditions would be reduced within 5–10 years by >90% in women and over time, as a result of herd immunity, in men. However, demonstrating efficacy in trials is one thing, achieving population effectiveness is another issue. These vaccines will only live up to their promise if:

  • 1 they are implemented as part of a successful immunization programme a challenge that requires political will and commitment with the full engagement of governments and public health authorities,
  • 2 the general public is convinced that they work and are safe.

Implementation in industrialized countries

The criteria for HPV vaccine introduction in industrialized countries have been reviewed recently [28]. A key element in any vaccination programme is vaccine coverage and the maintenance of this coverage in successive cohorts. HPV vaccines are prophylactic, not therapeutic, preventing, not treating infection and they are not effective in individuals with already established infections. Genital infection is usually sexually transmitted and the most important risk period for acquisition of a genital HPV is soon after the onset of sexual activity. The average age of sexual debut varies widely between societies [28] but to be assured that the vaccine recipients receive protection, young adolescents in the 9–15 year age group should be targeted, an age that is also immunologically optimal [29]. The recommendations for HPV vaccination in most countries recognize this and are remarkably uniform in targeting 9–15 year females as the primary group for immunization (Fig. 5). Catch up programmes are recommended in some countries, but there is variability in the age of the catch up populations. There has been uncertainty about vaccinating sexually active women but the phase III trials clearly show that women previously infected with a vaccine HPV type, seropositive but DNA negative at time of immunization do benefit from vaccination http://www.cdc.gov/vaccines/recs/acip/downloads/mtg-slides-feb08/14-3-hpv.pdf.

Figure 5.

 Recommendations for HPV immunization in 16 countries.

Furthermore, the quadrivalent vaccine has been evaluated in older women (25–45 years) and efficacy equivalent to that observed in the younger age groups has been shown [30].

The experience with Hepatitis B vaccination is that rapid introduction and high coverage rates for adolescents are best achieved with school based immunization programmes [28]. Australia and the UK are amongst those countries that have introduced HPV vaccination via school programmes and coverage rates >80% have been achieved in 11–13 year age group for the complete three dose schedule http://www.immunisation.nhs.uk/publications/HPV_Uptake_June09.pdf, [31]. This is in contrast to the USA, where there are few school based programmes; services are usually delivered via private providers and coverage rates of 25–30% are reported http://www.cdc.gov/vaccines/stats-surv/imz-coverage.htm#nisteen.

Both the UK http://www.immunisation.nhs.uk/Professional_Information/Key_vaccine_information/HPV and Australia have catch up programmes (14–18 years in the UK and 13–26 years in Australia) with delivery to girls not in school and women relying upon general practitioners. Considerable success in targeting the older catch up group 18–26 years has been achieved in Australia with estimates of 50–70% coverage. This is considered to be a consequence of the extraordinarily intense media coverage of the HPV vaccine and intensive advertising campaigns funded by both Government and the manufacturer. The quadrivalent vaccine was publicly funded for the Australian programme that commenced in April 2007 and there is already evidence of population effectiveness with reports of a decline of up to 48% in the number of women aged 28 and under presenting with new genital warts at a large sexual health clinic in Melbourne (Fairley personal communication 2009).

However, it must be remembered that the current vaccines include only two of the 15 oncogenic genital HPV types and even if delivered optimally with 100% coverage of the target age group, would prevent only 70–80% of cancers in the long term (and this takes into account some contribution from cross-protection). Countries with good cervical cancer screening programmes will have to maintain them and screening therefore will have to continue for the foreseeable future. There will be the large unvaccinated population outwith the vaccinated cohort that remain at risk for the development of HPV caused premalignant and malignant disease. Both immunized and nonimmunized birth cohorts will have to continue in the screening programme, as the immunized group will continue to be at risk for the nonvaccine oncogenic types [32]. Nonetheless the combination of screening and vaccination in countries with high population coverage for the female cohort for both these interventions could result in the elimination of cervical carcinoma in the medium term. Vaccination of men is a controversial issue, but if the coverage for women falls below 50% then immunization of men, if the objective is the reduction in incidence of cervix carcinoma, becomes cost effective [33].

Implementation in nonindustrialized countries

However, 80% of cervical cancers occur in the developing world, where access to secondary interventions such as screening is limited or nonexistent. These are predominantly countries with a low gross domestic product (GDP) and access to vaccines in general and expensive new vaccines, in particular, is difficult [34]. A major impediment to implementation in such settings is cost and mechanisms to reduce the costs of HPV vaccines to affordable levels for such countries will be essential, if the women who desperately need these vaccines will benefit [35].

Even if the vaccines are made affordable to the poorest countries there still remain formidable practical and cultural obstacles for the delivery of a three dose vaccine to female adolescents. The infrastructure for immunization exists in most developing countries and the WHO Extended Programme for Immunization (EPI) has reached a very high coverage for infants with about 70% of the world’s children receiving basic immunization services [34]. The challenge is to exploit the established immunization networks that deliver vaccines to babies, children and pregnant women to include adolescents and young adult women before the sexual debut and marriage. This will require new strategies involving health services and agencies not traditionally involved in immunization such as sexual and reproductive health and cancer control [36].

Vaccine safety

Public confidence in the safety of a new vaccine is crucial. The evidence from the randomized control trials and the postvaccine surveillance to date is clear, the VLP vaccines have a very good safety profile. Injection site reactions: pain, swelling, etc., as would be expected for a protein vaccine, are the most commonly reported but serious adverse events are no more frequent in vaccinees than in the unvaccinated population in that age range [37] http://www.whale.to/vaccines/Cervarix_-_safety_analysis_04_Dec_08%5B1%5D.pdf.

Clear statements regarding HPV vaccine safety have been made by both the US FDA http://www.fda.gov/BiologicsBloodVaccines/SafetyAvailability/VaccineSafety/ucm179549.htm and European EMEA, regulatory agencies http://www.emea.europa.eu/humandocs/PDFs/EPAR/gardasil/Gardasil_press_release.pdf, http://www.emea.europa.eu/humandocs/PDFs/EPAR/cervarix/H-721-en6.pdf.

The public perception of risk is not necessarily a reflection of a rational intellectual analysis of the evidence but quite understandably reflects fear of harm to themselves or their child. Public confidence will be strengthened by statements from medical professionals based on a rational assessment of the evidence and it is critical that such statements are based on a scientifically robust and in depth analysis of the data. The harm to public health by superficially based opinion particularly if it is then championed by media outlets can be substantial as the experience with the measles mumps rubella (MMR) vaccine has shown. http://news.bbc.co.uk/1/hi/talking_point/1802192.stm.

Conclusion

The prophylactic HPV vaccines are a major breakthrough in modern medicine, they are highly efficacious and the available evidence indicates that they are ‘safe’ vaccines. The current HPV vaccines will not eliminate cervical cancer or the other HPV associated ano-genital diseases but could dramatically reduce the incidence, if effectively delivered to the populations that need it both in the industrialized and nonindustrialized world.

Conflict of interest statement

Dr. Stanley is consultant for the following: SPMSD Lyoft France, GSK Biologicals, Rixensart, Belgium and Merck research laboratories, USA.

Ancillary