Development of respiratory syncytial virus (RSV) vaccines for infants

Summary

2017 will mark the 60^th anniversary since the first isolation of RSV in children. In spite of concerted efforts over all these years, the goal of developing an effective vaccine against paediatric RSV disease has remained elusive. One of the main hurdles standing in the way of an effective vaccine is the fact that the age incidence of severe disease peaks within the first 3 months of life, providing limited opportunity for intervention. In addition to this complexity, the spectre of failed historical vaccines, which increased the risk of illness and death upon subsequent natural infection, has substantially increased the safety criteria against which modern vaccines will be assessed. This review traces the history of RSV vaccine development for young infants and analyses the potential reasons for the failure of historic vaccines. It also discusses recent breakthroughs in vaccine antigen design and the progressive evolution of platforms for the delivery of these antigens to seronegative infants.

Background

RSV is the most important viral cause of acute lower respiratory tract infection (ALRI) among infants globally. In 2005 at least 33.8 million cases of RSV associated ALRI occurred in the under 5 age group and of these, 3.4 million cases were associated with severe illness that necessitated hospital admission.¹ Between 66,000 and 199,000 deaths were estimated to have occurred in the same year as a result of RSV associated ALRI, with a vast majority of the deaths occurring in developing countries.¹ RSV mortality is associated with certain risk factors. Certain co-morbidities are associated with an increased risk of death following RSV infection. For example, the mortality rate in children with congenital heart disease or chronic lung disease is significantly higher than RSV associated mortality rate in otherwise healthy children.² Apart from its effect on the health of the patient, RSV imposes a heavy economic cost. In the United States for example, the average total cost of hospitalisation for term and pre-term infants infected with RSV exceeds the hospitalisation costs of infants admitted with other conditions by an average of $9,151 and $17,465 respectively.³ In view of the fact that the burden of severe RSV disease is highest among infants, these data suggest that the monetary cost associated with severe RSV infection among infants is substantial.

In addition to infants, RSV exerts a substantial morbidity burden in adults. Whilst RSV-attributable mortality in immunocompetent adults is rare, it is a common cause of recurrent respiratory illness whose cost can be measured in terms of days of work or school missed. Previous studies have shown that RSV infection in previously healthy adults is associated with periods of absence from work of up to one week.^4,5 As a result of the health and economic costs linked with infection, development of an RSV vaccine is considered to be a research priority.

Surprisingly little is known about the pathogenic processes of RSV in the infant lung. In small rodent models, severity of symptoms following infection appears to correlate not with viral load, but with excessive host inflammation⁶ suggesting that severe disease is an immunopathological phenomenon. A few histological studies of post-mortem samples from infants who died of suspected RSV pneumonia suggest that severe disease in infants is largely an outcome of airway obstruction. This obstruction is caused by the occlusion of small airway lumens by necrotic epithelial and inflammatory cell debris, coupled with the accumulation of fibrin, oedema and excessive mucus production.^7,8

Historical paediatric RSV vaccines

In the mid 1960s a formalin inactivated RSV vaccine (FI-RSV) was developed by culturing the virus at 36°C, filtering it on a membrane and inactivating it using 0.025% formalin at 36°C for 72 hours.⁹ This vaccine was subsequently tested in a randomised control trial in which children received either a full dose of the FI-RSV vaccine or a parainfluenza control vaccine.⁹ Although the FI-RSV was reasonably well tolerated with no immediate adverse reactions,¹⁰ the incidence of medically attended RSV infections in FI-RSV group was about 3 times greater than the control group in the RSV epidemic that followed. In addition, the severity of clinical symptoms in the FI-RSV group was much greater than the control group: up to 44% of the hospitalised children in the FI-RSV group had severe or very severe pneumonia compared with 5.6% among the controls.¹¹ Similar observations were made in a separate FI-RSV vaccine trial in which 80% of FI-RSV vaccinees required hospitalisation following natural infection compared to only 5% in the active control group.⁹ As in the previous trial, the severity of illness was greater among the FI-RSV vaccinees relative to the control group. Tragically, two toddlers aged 14 and 16 months who were in the FI-RSV vaccine group died upon natural exposure to RSV, with post-mortem evidence of extensive bronchopneumonia, pneumothorax and eosinophilia.⁹

Designing a safe and effective RSV vaccine: Structural properties of protective immunity targets

Investigations of the immune responses that were mounted by children following FI-RSV vaccination showed that young infants developed high levels of antibody to the virus fusion (F) protein but had poor responses to the attachment (G) protein.¹² None of the FI-RSV vaccinated children developed neutralising antibody titres comparable to that of age-matched individuals who had undergone natural infection.¹² Among other reasons, it was concluded that formalin inactivation had somehow altered epitopes on the F and G proteins resulting in the development of non-functional (non-neutralising) antibodies and that these non-neutralising antibodies could have potentiated disease through the formation of immune complexes in the lung.¹² Subsequent studies found that in addition to their poorly neutralising capacity, F protein specific antibodies to the FI-RSV vaccine did not display fusion inhibiting activity which resulted in the increased spread of the virus in the respiratory tract upon natural infection.¹³

Remarkably, new studies of the structural properties of the F protein are now finally providing powerful evidence of the effect of formalin inactivation on the antigenic properties of this critical target of protective immunity. Recent advances in the structural biology of the F protein have unravelled critical changes in its antigenic characteristics that occur during the fusion process between the virus envelope and host cell membrane. The F protein is initially expressed as a high energy, metastable pre-fusion isoform (pre-F) that upon insertion into the host cell membrane, irreversibly transitions into a stable low energy post-fusion structure (post-F). This process results in the fusion of the viral envelope and the host cell membrane^14,15 – a prerequisite step in the delivery of viral RNA into the host cell. Although structurally dissimilar, pre- and post-F share two antigenic regions (sites II and IV) that are the targets of moderately neutralising antibodies. However three other antigenic sites that are highly neutralisation sensitive (sites Ø, III and V) are only present in pre-F. Importantly, it has now been definitively shown that the process of formalin and heat inactivation mediates a rapid and irreversible shift from the pre-F to the post-F conformation¹⁶ and as a consequence of this change, there is a near complete loss of neutralisation sensitive antibody epitopes.¹⁷. These results provide a compelling explanation for the tragic failure of the FI-RSV vaccines, and enhance the prospect of developing effective and safer vaccines for paediatric RSV.

Vaccine platforms for seronegative infants

The recent breakthroughs in the structural biology of the F protein have opened up the possibility of achieving licensure of an RSV vaccine within a decade. However, development of appropriate platforms for delivery of these antigens in the main risk group for severe disease – seronegative infants – has not kept pace with the structural biology advances. Based on the clinical experiences with the non-replicating FI-RSV vaccine, there are still lingering safety concerns regarding the administration of non-replicating vaccines to seronegative infants. As a result of these misgivings, the most practical route for active infant immunisation is likely to be through the use of replicating antigen delivery platforms. The most advanced replicating platforms for RSV antigen delivery are live-attenuated vaccines and virus-vectored vaccines.

Live-attenuated vaccines

Following the failure of the FI-RSV vaccines of the 1960s, attention shifted to the development of live-attenuated vaccines. Live-attenuated RSV vaccines were first developed through extensive serial passaging of the A2 strain of RSV at progressively lower temperatures i.e. cold passaging (cp).¹⁸ Adaptation of a live, intranasally delivered vaccine for replication at lower temperatures is an important safety feature since the temperature of the upper respiratory tract is lower than the core body temperature.¹⁹ To achieve further attenuation, chemical mutagenesis has been used to induce a temperature sensitive (ts) phenotype in the cp vaccine strains.²⁰ Temperature sensitive mutants have a shutoff temperature above which they are unable to replicate – thus mutants with shutoff temperatures that are close to the core body temperature are unlikely to replicate in the lower respiratory tract and cause severe disease.²¹ Initial clinical trials with cp-ts live-attenuated RSV vaccines showed that they were still insufficiently attenuated and were genetically unstable.²⁰ Subsequent studies reported the generation of cp-ts strains with remarkably low shutoff temperatures and greater genetic stability.²² Further attenuation has been achieved by reverse genetics, through the incorporation of attenuating mutations in some virus genes or the deletion of entire genes from the virus backbone.²³

A number of features make live-attenuated vaccines an especially attractive platform for immunising young infants. The intranasal route of delivery of live-attenuated vaccines provides the opportunity to avoid the potent immune suppression mediated by maternally acquired antibodies.²⁴ In addition, it provides direct immune stimulation of the respiratory tract, thereby inducing local functional immunity that is broader and more effective than that induced by subunit vaccines.^25–27 Perhaps the main advantage offered by live-attenuated vaccines, is the large safety database accumulated from decades of conducting trials of RSV live-attenuated vaccines in seronegative infants. These studies have consistently shown that RSV live-attenuated vaccines do not prime infants for enhanced respiratory disease upon subsequent natural exposure,²³ a critical safety checkpoint for any modern RSV vaccine. Furthermore, through reverse genetics, the immunogenicity of these viruses could potentially be enhanced through the expression of potent neutralising antibody targets such as pre-F. Despite their potential, live-attenuated vaccines have struggled to strike the right balance between achieving sufficient attenuation while retaining immunogenicity. In spite of these challenges, the use of innovative reverse genetics manipulations to the virus backbone have yielded viruses with excellent attenuation and immunogenicity profiles in animal models,²⁸ and provide some grounds for optimism regarding the future use of these vaccines in seronegative infants.

Virus vectored vaccines

The concept of viral-vectored vaccines is predicated on the fact that viruses possess an inherent capability of delivering foreign genes to host cells while at the same time providing the potent adjuvanting signals that are intrinsic to this mode of antigen delivery. Viral vectors been used widely over the years and have a demonstrable record of success in the induction of both cellular and humoral immune responses against a variety of antigens derived from pathogens such as influenza,^29,30 TB,³¹ malaria³² and RSV³³ and are the technology being applied to develop Ebola vaccines.³⁴ They elicit strong immune responses in humans and have acceptable safety profiles, including in 10 week-old infants.³⁵ A further advantage of this vaccine platform is that their safety profiles can be further increased by incorporating mutations that render the viruses replication-defective;³⁶ thus reduce the likelihood of vaccine virus shedding and transmission. In addition to the ease with which transgenes of stabilised pre-F can be inserted into the vector backbone, their capacity for limited replication makes these vectors an acceptable platform for the delivery of viral antigens in the seronegative infant population. Although no trials of virus-vectored RSV vaccines have yet been conducted in this population, the results of phase I safety and immunogenicity trials in young adults are promising. In this population these vaccines have been shown to have a remarkable safety profile and have been shown to induce strong RSV-specific T and B cell responses, including serum neutralising antibody.³³ These results open the pathway for further age-de-escalation trials down to seropositive children and finally to seronegative infants.