Respiratory syncytial virus vaccine development

Abstract

Respiratory syncytial virus (RSV) is the leading cause of lower respiratory tract viral disease in infants and young children. Presently, there are no explicit recommendations for RSV treatment apart from supportive care. The virus is therefore responsible for an estimated 160,000 deaths per year worldwide. Despite half a century of dedicated research, there remains no licensed vaccine product. Herein are described past and current efforts to harness innate and adaptive immune potentials to combat RSV. A plethora of candidate vaccine products and strategies are reviewed. The development of a successful RSV vaccine may ultimately stem from attention to historical lessons, in concert with an integral partnering of immunology and virology research fields.

Respiratory syncytial virus: the disease

Respiratory syncytial virus (RSV) is the most common cause of lower respiratory tract infections in young children and is responsible for an estimated 160,000 deaths annually worldwide. In a single year, approximately 34 million episodes of RSV-associated lower respiratory tract infections may occur in children under the age of 5 years, and among these, 10% may require hospitalization [1]. Infant hospitalization costs may well exceed US$400 million annually in the USA alone [2–12]. For patients with bronchiolitis and pneumonia, RSV has been identified as the etiologic agent in as many as 90 and 50% of cases, respectively [9,10].

Newborn infants are partially protected from disease by maternal RSV-specific neutralizing antibodies, but this protection is not long lasting [13–16]. Infants are at particularly high risk for RSV-associated morbidity and mortality between 2 and 4 months of age when maternal antibodies have fallen and have not yet been replaced by an endogenous antibody response [14,17]. RSV infections usually occur in the first year after birth and by the second year, 90% of individuals will have been infected at least once [3]. Particularly prone to RSV infection are individuals born prematurely, of low birth weight and/or with congenital cardiopulmonary defects, but the majority of serious RSV infections occur among infants with no known risk factors. Individuals with underlying immunodeficiencies (e.g., bone marrow transplant recipients) and the elderly are also at high risk for RSV disease [18–21] and HIV-1-infected children shed RSV for remarkably long periods [6,22–24].

Respiratory syncytial virus is usually spread by droplets, direct contact or fomites. Several days (2–8) after viral infection, upper respiratory tract symptoms may occur (e.g., rhinorrhea, cough and sneezing). There may also be a loss of appetite and low-grade fever. Within 1–3 days following the onset of rhinorrhea, the virus and associated symptoms may spread to the lower respiratory tract. Severe bronchiolitis may occur, leading to airway blockage and respiratory distress. As an aftermath to RSV infection, airway hyper-responsiveness and obstruction (e.g., asthma and wheezing) can continue for many years. Debates are ongoing as to whether these long-term symptoms are caused by RSV [25] or define a predisposition that exacerbates RSV disease [3,17,26,301].

The first RSV infection is most severe and reinfections are common, particularly of the upper respiratory tract. In a clinical study when adults were deliberately challenged repeatedly with RSV, infections occurred in at least a quarter of the participants per challenge. Mild symptoms demonstrated that pre-existing immune responses did not afford complete protection [6,27,28].

There are no clear recommendations for treatment of RSV apart from supportive care. Ribavirin is sometimes prescribed, but this drug is expensive and is not certain to improve outcome [5,29]. A new strategy for RSV control involving siRNA is now in early stages of clinical development [28]. As prophylaxis for RSV infection, a monoclonal antibody against a highly conserved RSV F epitope has been used (Synagis^®/palivizumab [MedImmune, Gaithersburg, MD, USA]). This antibody has significantly reduced severe disease in high-risk infants [30]. In premature infants without chronic lung disease, the efficacy of Synagis/palivizumab has been as high as 80% [31]. A new product with a similar specificity was recently developed by MedImmune (motavizumab) to replace Synagis/palivizumab, but tests did not support superiority and development of the new product for licensure has been discontinued [302]. Prophylaxis is not available to many who need it due to the cost and logistics of injections. It is generally prescribed for individuals with predetermined risk factors, and as stated earlier, risk factors are identified in only a fraction of susceptible children. In developing countries, constraints in infrastructure and funding further impede access [32]. Clearly, further development of RSV vaccines and therapies is warranted, and novel successful products could have an enormous positive impact on human health.

RSV: the virus

Respiratory syncytial virus is a negative-strand nonsegmented RNA virus of the family Paramyxoviridae, the subfamily Pneumovirinae and the genus Pneumovirus [17]. There are two subtypes, A and B. Virus size is pleiomorphic, typically 100–350 nm yet filaments may be as long as 10 µm. There are ten viral genes that are transcribed sequentially into separate mRNAs by the viral polymerase. Genes are linearly ordered from the 3′ to 5′ end of the negative-sense RNA genome. On the 3′ end are genes NS1, NS2 (putative nonstructural proteins 1 and 2, anti-inflammatory proteins), N (nucleoprotein) and P (phosphoprotein, a cofactor in RNA synthesis). These are followed by M (matrix protein), SH (small hydrophobic protein, a nonessential protein that may function as a pentameric ion channel [3,33]), G (a type II integral membrane attachment glycoprotein), and F (a type I integral membrane fusion protein). Most 5′ are M2 (a gene that encompasses two open-reading frames encoding M2-1 and M2-2, supportive of mRNA transcription and the balance between RNA replication and transcription) and L (large protein, the major polymerase) [3,17].

G and F proteins comprise the major glycoprotein spikes on the viral membrane and are major targets of neutralizing antibodies. SH is a third integral membrane protein, but is not required for efficient viral growth in vitro or in vivo. The G protein is approximately 300 amino acids in length, but is heavily glycosylated with several N-linked glycosylation sites and perhaps as many as 25 O-linked glycosylation sites, yielding a glycoprotein of molecular weight approximately 90 kDa [3,17]. RSV G also exists as a secreted protein product owing to an alternative initiation codon in the coding region of the transmembrane domain [6,34]. Some researchers suggest that secreted product may act as a decoy to block the host’s G-specific neutralizing antibody response [6,35,36]. The F protein is approximately 500 amino acids in length and glycosylated much less heavily than G, yielding a 70-kDa glycoprotein product. F protein is first synthesized as an inactive F₀ precursor that assembles into a homotrimer [17]. The precursor is sensitive to cleavage by a furin-like intracellular protease in the trans-golgi of an infected cell (although full cleavage is not required for transport of F to the infected cell surface [37,38]). Three products of cleavage are F₂, p27 and F₁. F₁ contains the transmembrane domain at one end, two heptad repeats (HRA and HRB), and a hydrophobic fusion peptide at the opposite end. The F₁–F₂ disulfide linkage forms the active molecule essential for virus/cell fusion or cell/cell fusion (syncytia formation) [39]. RSV replicates almost exclusively in the mammalian respiratory tract [6] where the furin-like protease is synthesized. Apart from facilitating virus fusion, the RSV F protein is also capable of signaling through Toll-like receptor (TLR)-4 and CD14 [40,41]. This signaling may induce proinflammatory cytokines including interferons with antiviral activities. By contrast, the G, NS-1 and NS-2 proteins may in some cases downregulate inflammatory events [3,17,42].

The virus life cycle typically begins by attachment of the G glycoprotein to the mammalian host plasma cell membrane (F can also facilitate attachment), followed by F-mediated fusion. Paramyxoviruses typically fuse at the cell surface. Some siRNA studies have suggested that clathrin-mediated endocytosis and early endosomes may also play a role in RSV infectivity [3,43]. The cellular targets of G and F proteins are not fully defined. G protein can bind glycosaminoglycans such as heparin sulfate and chondroitin sulfate B as well as C-type lectins such as surfactant protein [6,44]. F protein also has heparin binding domains [45] and like G, can bind glycosaminoglycans [46,47]. Other possible receptors for RSV include the intercellular adhesion molecule-1 [48] and the CX3CR1 fracktalkine chemokine receptor [3,17,49].

Virus–cell fusion occurs when the fusion peptide is inserted into the target cell membrane and the protein refolds into a hairpin structure assisted by HRA and HRB interactions to bring viral and target cell membranes into close proximity [39]. Viruses that lack G (ΔG) are capable of fusion and infection in vitro, although ΔG viruses are highly attenuated in vivo [17]. Once virus and cell membranes are fused, internal viral components are released into the cytoplasm where RSV replication occurs. P and L proteins support RNA synthesis, both for the production of mRNAs and for the production of new negative-strand genomes (~15 kb in size). During transcription and replication, the nucleoprotein supports and protects the genome and antigenome RNA. Perhaps shielding of RNA by the nucleoprotein helps avoid additional onset of host-defense mechanisms by TLR and other forms of signaling. Following de novo RNA and protein production, new virions assemble at the cell surface where they acquire an envelope membrane by budding and release from the mammalian cell [3,17].

Vaccine development & the immune system

Harnessing the immune response by vaccination is the single most effective method for the control of human infectious disease. The strategies associated with most of today’s licensed vaccines were developed decades or centuries ago. In recent years, highly sophisticated molecular techniques have simplified vaccine production and improved safety [50]. New techniques have also tempted researchers to test an extraordinary number of antigen manipulations and peptide reassortments [51], not all of which have been successful [52]. The application of fundamental immune concepts to the development of new vaccines is encouraged to ensure that products mimic pathogen structures and induce positive immune responses in the majority of humans [52]. This section reviews basic immune concepts that may impact vaccine efficacy, and may assist evaluation of old and new RSV vaccine candidates.

The first vaccine success

Edward Jenner was the first to formally demonstrate the benefit of vaccination in the 1790s, by vaccinating a young boy with material from cowpox lesions, and then deliberately exposing the boy to smallpox virus. The vaccine was successful in that protection against smallpox was observed [303]. It is now known that the cowpox virus was successful, because it activated durable B- and T-cell (lymphocyte) responses with cross-reactivity for smallpox virus [53–55]. Lessons from this event cannot be overstated as Jenner’s work advanced the only vaccine to eradicate a human disease [56].

Amplification of antigen-specific immune effectors

Historical vaccine successes have shown that vaccines must mimic the proteins and peptides presented by a target pathogen so that the immune system can be pre-activated before pathogen exposure. The strength of the immune system lies in its vast array of B and T lymphocytes. Each lymphocyte carries a different receptor with a unique binding site determined by the cell’s previous history of complex immunoglobulin or T-cell receptor (TCR) gene rearrangements. For any given pathogen, a fraction of lymphocytes will have ‘specificity’, in that their preformed receptor (antibody for B cells or TCR for T cells) will bind intricately to the foreign material (antigen, either native or processed) with a lock-and-key-type interaction. The ‘specific’ cells are those that are targeted by vaccination. Each person carries a different repertoire of lymphocytes reflecting their unique array of gene rearrangements. Despite differences among humans, the sheer quantity of lymphocytes in each individual ensures that he/she will respond to essentially any complex foreign pathogen in nature [50,57].

Upon first pathogen exposure, specific lymphocyte receptors are triggered by their interactions with antigens, after which the respective parent cells divide to generate large numbers of identical daughter cells. For B cells, somatic mutation and selection can further improve antibody binding potentials [58]. A portion of B cells also mature to become plasma cells (also termed antibody-forming cells or antibody-secreting cells), which constitutively secrete antibodies. After priming, memory B cells and T cells can provide long-term surveillance of the blood, lymph, peripheral tissues and mucosal surfaces, in some cases for the lifetime of an animal [59–61].

Vaccines that successfully mimic their target pathogen in terms of processed and unprocessed antigens will prime the three classical subsets of immune effectors (B cells, T helper cells and cytotoxic T lymphocytes) and may harness the greatest benefit from the immune system. If the match of vaccine and pathogen is not precise, the receptors on primed lymphocytes will provide limited benefit and may possibly cause harm. When a pathogen’s antigens are well conserved in nature, a single-component vaccine can be successful. When a pathogen’s antigens are diverse, cocktail vaccines are most effective, as each antigen in the cocktail can trigger a different subpopulation of specific lymphocytes [62,63]. To date, researchers have successfully mimicked pathogen structures by:

• Attenuating the pathogen (e.g., cold-adapted influenza virus vaccine)

• Killing the pathogen (e.g., killed polio virus vaccine)

• Identifying a related virus from another species (the Jennerian approach)

• Isolating membrane molecules from a pathogen’s coat (e.g., hepatitis B virus vaccine)

Mechanisms of antibody-mediated protection

In a vaccinated individual, antibodies provide a powerful first line of adaptive immune defense. These preformed effectors naturally circulate through blood and lymph and bathe mucosal surfaces so that they may immediately bind pathogen without additional cell stimulation or antigen processing. Mechanisms of antibody-mediated protection [64–67] include, but are not limited to:

• Aggregation of pathogen;

• Direct inhibition of receptor–ligand interactions between pathogen and host target cell;

• Inhibition of membrane fusion;

• Inhibition of virions within the cytosol of an infected cell [64];

• Labeling of the pathogen for destruction by phagocytes (opsonization);

• Induction of complement-mediated kill;

• Induction of degranulation by mast cells or eosinophils;

• Labeling of virus-infected host cells for destruction by innate immune effectors such as natural killer cells (antibody-dependent cell-mediated cytotoxicity).

IgA is particularly well suited for the control of respiratory virus infections as this isotype frequently bathes the mucosal lining of the respiratory tract. With assistance from the poly-Ig receptor, IgA antibodies transcytose across epithelial cells to enter the lumen of the respiratory tract, and are tethered to the airway lining by secretory component [58]. IgG can also cross epithelial cells into the respiratory tract and, in the absence of IgA, can protect against RSV disease [68,69]. Neutralizing antibody titers correlate well with protection against RSV disease in some, but not all cases. Humans with low serum antibody neutralization titers are often more likely to be hospitalized, whereas a titer of >100 associates with a better outcome [70,71]. In one cotton rat passive transfer study, the serum antibody titer necessary for protection against RSV lung infection was 380 units. To protect the upper respiratory tract from infection, a ten-times higher serum antibody titer was required [71–73].

Mechanisms of CD4⁺ T-cell-mediated protection

CD4⁺ T cells secrete cytokines, assist B- and T-cell activation/function, and interact with the innate immune system to control pathogen growth [58,74,75]. Except in rare situations (T-cell-independent responses) CD4⁺ T cells may be essential for the induction and durability of protective immunity. CD4⁺ T cells can also reduce virus load in the absence of their CD8⁺ T- and B-cell partners [76,77]. CD4⁺ T cells do not recognize free antigens, but recognize the antigen’s peptides, processed by B cells or other antigen-presenting cells to associate with MHC class II antigens on the mammalian cell surface. Antigen processing for CD4⁺ T-cell activation usually initiates when an antigen–antibody complex or particle is engulfed by a B cell or antigen is engulfed by another APC, after which proteins are fragmented in endosomes, bound into the peptide binding grooves of MHC class II molecules, and returned to the cell surface. Because humans have widely disparate MHC antigens, a pathogen’s most immunogenic (immunodominant) peptides will differ from one individual to the next. Algorithms are in some cases useful to predict which peptide sequences will ‘fit’ into the peptide-binding groove of a particular MHC molecule. The context of a peptide within a viral protein will also influence its immunogenicity, because a peptide’s position will determine its accessibility to the proteases and other enzymes necessary for protein fragmentation and peptide assembly with MHC [78,79]. Certain peptides also compete with one another for presentation; when taken into a new context, an otherwise weak peptide may claim immunodominance [80]. An additional factor that will influence immunogenicity is the similarity of peptide to ‘self ’ proteins, because lymphocytes responsive to ‘self ’ are usually eliminated or inactivated by tolerance mechanisms [81–85]. However another influence on immunogenicity is the precise repertoire of lymphocytes within each host, as not every individual carries a TCR that will bind a given peptide. In summation, a peptide that elicits a strong response in the context of a pathogen may not do so when placed in an artifactual context, and a peptide that elicits a strong response in one individual may not do so in another. This explains in part why the frequency of vaccine responsiveness in the human population can be disappointing when antigen scope is reduced and when peptide contexts are altered [52].

CD4⁺ T-cell subpopulations have been the topic of much debate in the RSV vaccine community (see formalin-inactivated RSV vaccine discussion later). CD4⁺ T cells are subdivided into multiple subsets including Th1, Th2, Th17 and Tregs. Th1 responses typically associate with interleukins IL-2, IFN-γ and IL-12 (the latter produced by macrophages and dendritic cells), and IgG antibody isotypes of various subclasses (e.g., IgG2a in mice). Th2 populations more often associate with interleukins IL-4, IL-5, IL-6, IL-10 and IL-13, and antibody isotypes IgM, IgG, IgE and IgA [58,86–90]. It should be noted that Th1 and Th2 patterns and profiles are not absolute or mutually exclusive. The appearance of one cytokine from a particular CD4⁺ T-cell subset during an immune response does not ensure the appearance of all others [86]. Furthermore, there is no globally accepted standard measurement of Th1 and Th2 cells; one laboratory may measure IL-4 to quantify a Th2 response while another may measure IL-10 or IL-13. The quality and quantity of membrane markers used to define different T-cell subsets also vary among laboratories. Caution must therefore be taken to avoid the mistaken assignment of dissimilar cells to a single population.

Treg and Th17 cells, like other T-cell types, are defined in some cases by phenotype and in other cases by function. The Treg cell is typically CD4⁺CD25⁺ and FoxP3⁺. These cells can downregulate immune responses using antigen-specific and nonspecific mechanisms [91–95]. Unlike the Treg cell, the IL-17-producing Th17 cell is most often associated with inflammation [96,97]. Environmental factors influence the development and ratios of Tregs, Th17 cells and other T-cell subsets [98,99]. A fine balance of subsets is essential to ensure that the immune response is sufficiently robust to clear virus, but is not sufficiently robust to inflame tissues and block respiratory airways.

Mechanisms of CD8⁺ T-cell-mediated protection

CD8⁺ T cells can secrete cytokines and are best known for their CTL capacity. They act as a second line of defense when antibodies fail to clear viral particles before infection. Their kill of infected target cells inhibits the production of new virus particles. CD8⁺ T cells are classically induced after:

• Virus infection of a susceptible target

• Endogenous expression of viral proteins

• Fragmentation of proteins in proteasomes

• Transporter activated protein-mediated entry of peptides into the endoplasmic reticulum

• Association of peptides with MHC class I proteins

• Presentation of MHC–peptide complexes on the infected cell surface

CD8⁺ T cells can also be responsive to exogenous antigens after cross-presentation, but this mode of antigen processing is generally less efficient [100–102]. The preference of CD8⁺ T cells for peptides from endogenously expressed viral proteins ensures that uninfected cells will not often be killed.

The combined activities of antibodies, and CD4⁺ and CD8⁺ T cells present an impressive defense against infectious disease. The vaccine developer benefits in that he/she need not create an immune system, but must simply pre-activate the antigen-specific lymphocytes that naturally exist in the healthy human.

In vitro & in vivo models for RSV vaccine development

There is currently no single globally accepted model for the study of RSV vaccines and no model in which RSV disease matches that of humans. Instead, a variety of different systems are used. Small animal models include the BALB/c mouse, the hamster [103] and the cotton rat (Stenopus hispidus) [41,104,105]. The mouse model is attractive owing to the plethora of associated reagents and the availability of congenic, mutant, transgenic and knock-out strains. Disadvantages are that high RSV doses are required for infection. The cotton rat is more susceptible to infection, but lacks a comparable reagent pool. Primate models include rhesus macaques [106], African green monkeys [107] and chimpanzees, but none are highly susceptible to RSV disease. When a cold-adapted RSV vaccine was proven safe in chimpanzees, it nonetheless caused lower respiratory tract infection when translated to a clinical study in seronegative children [5,108]. Apart from animal models, in vitro models are also used to mimic the respiratory tract [109–113]. Cultures can be varied in terms of cell origin, cell type and environment. For example, a 3D air–liquid interface model of pediatric bronchial epithelium has been established to compare cells/functions between asthmatic and nonasthmatic children [111]. Cultures share some characteristics with the respiratory tract, but lack the complexity of the whole animal including the natural defenses of innate and adaptive immunity. Each model can provide important information concerning vaccine potential, but has limitations in terms of predicting clinical trial outcome.

RSV vaccine development

Respiratory syncytial virus vaccine development has progressed for a half century, yielding an extraordinary number of vaccine candidates. Products include killed virus, purified RSV proteins, attenuated RSV, nanoparticles, virus-like particles (VLPs; particles that mimic native virus conformation but lack genetic material), virosomes (vesicular membranes carrying virus-derived proteins without nucleocapsids or genetic material), replication-competent vectors carrying RSV genes, and replication-deficient vectors carrying RSV genes. Molecular tools include DNA-based, viral-based and bacterial-based vectors, which may be expressed in tissue culture, plants or animal hosts. The RSV proteins that have been studied most often are F and G, because they each appear on the native RSV particle and are important targets for neutralizing antibodies and effector T cells. Of the two molecules, F is better conserved and is perhaps better suited to induce lymphocytes that cross-react with antigens from RSV A and B isolates.

Later, a number of strategies will be described. An attempt is made to mention old and new strategies, at least briefly, to indicate the breadth of research activity. Despite many decades of research, most vaccine candidates have been tested only preclinically or in RSV seropositive humans. For many of these, features of safety and efficacy in a seronegative human target population are presumed, but unknown.

The formalin-treated RSV vaccine: an early disappointment

In the 1960s, a formalin-inactivated RSV vaccine (FI-RSV) was combined with alum for intramuscular injection of babies. Unfortunately, the vaccine was not efficacious and enhanced disease when participants were subsequently exposed to RSV. Hospitalizations were far more prevalent in the vaccinated group than among controls and there were two fatalities attributed to the vaccine [5,17,114–116].

Both small and large animal models have been used in attempts to recapitulate the immunopathology of the FI-RSV vaccine. Models have included mice, cotton rats and primates. These models are instructive, but are in some cases influenced by artifacts including impurities that may exist in a vaccine or challenge virus [304]. RSV-infected humans have also been studied extensively in attempts to dissect correlates of immunopathology. Again caution must be taken when interpreting results, as pathology is influenced by a complexity of factors including virus growth, target cell death, airway size, virus-associated signals to the immune system, and inflammation. Minor fluctuations in any of these parameters (e.g., a slightly higher virus load or a slightly smaller airway) may tip the balance between safety and morbidity.

One scenario that could explain the FI-RSV result of the 1960s is as follows: the formalin inactivation of RSV destroyed key antibody-neutralizing determinants on the virus. Therefore, the antibodies that were primed by FI-RSV were capable of antigen binding and precipitation, but they did not recognize neutralizing determinants on the native virus and did not act as a first line of defense to clear RSV [6,117–119]. CTL responses were also minimal owing to the inert nature of the vaccine. Given that these important lines of defense were lacking, there was a high antigen burden in the lung after RSV infection, and consequent recruitment of cells into the lower respiratory tract. Included among cells were RSV-specific T cells, eosinophils and neutrophils [41,120]. Inflammation ultimately obstructed small airways leading to morbidity in many vaccinees [121–123].

An additional suggestion to explain the 1960s result is that the formalin treatment of the vaccine altered surface antigens to block natural TLR-4 signaling by the RSV F protein, which further impeded the natural processes of virus-specific antibody development [124]. In support of the notion that TLR signaling impacts RSV infection and disease, it has been shown that TLR-4 knockout mice clear RSV less well than their wild-type counterparts [41,125]. In humans, TLR polymorphisms have been associated with disease severity in some, but not all studies [3,126–129].

Debates continue as to which cell type(s) was the most important instigator of disease. Contradictory results make absolute conclusions difficult and suggest that the outcome was influenced by a composite of events rather than a single cell type. In a BALB/c mouse model, high Th2/Th1 cell ratios (IL-4 and IL-10 production) have been associated with immunopathology [130–132]. T cells marked with Vβ14 TCR and with specificity for RSV G were deemed responsible for immunopathology following some, but not all strategies of vaccination [132,133]. In a monkey model, disease was associated with IL-5 and IL-13, as well as neutrophils and eosinophils. These types of results have lead some investigators to suggest that RSV vaccines should induce negligible neutrophil, eosinophil and/or Th2 responses (or negligible T-cell responses of all types) [3,5,6].

On the other hand, there is evidence that Th2 cells need not always promote immunopathology during RSV infection [134]. For example, in infants with severe RSV infections, disease is sometimes associated with a predominance of Th1-like responses or high Th1/Th2 ratios [41,135–138]. In cotton rats, vaccination with the original Lot 100 FI-RSV vaccine (used in the failed clinical trials of the 1960s) [120], did not upregulate Th2 cytokines exclusively. Rather, the FI-RSV vaccine additionally upregulated the expression of chemokine genes and Th1-associated cytokine genes such as IL-2, IFN-γ and IL-12p40 [41,139]. Furthermore, it has been shown that some RSV vaccines that induce Th2 responses do not show immune-enhanced disease after RSV infection [140], and that infants who are deficient in T-cell-mediated immune responses suffer worse outcomes following RSV infections than their normal counterparts [3,141]. Perhaps this is not surprising, as Th2 responses can associate with IgA antibody production, and IgA is a key isotype for protection of the respiratory tract’s mucosal surface [58].

The FI-RSV vaccine outcome suggested that eosinophils might be associated with risk, but eosinophils can be beneficial in other scenarios [142]. For example, transfer of eosinophils to the lungs of RSV-infected mice has, in some cases, been shown to enhance clearance of RSV and inhibit hyper-reactivity in the airways [143].

There are multiple means by which immunopathology can be modulated in small animal models. Prestimulation of CTL responses in a mouse model was demonstrated to be advantageous [5,144,145]. Tregs could also downregulate immunopathological effects [146]. In a mouse model, the administration of anti-RSV G monoclonal antibody 1 day prior to RSV infection reduced inflammation and disease [147], highlighting the advantage of partnering potent B- and T-cell responses for virus control. Adjuvants can modify immunopathological events, as demonstrated by researchers who formulated FI-RSV with CpG oligodeoxynucleotide, a host defense peptide, and polyphosphazene [148]. In a separate study, monophosphoryl lipid adjuvant was administered with FI-RSV to cotton rats, after which there was a decrease in the cytokine storm associated with vaccine-enhanced disease. In that model, expression of both Th1 and selected Th2 (IL-4, IL-5, IL-13, but not IL-10) cytokine genes was reduced [41,139]. Antisense RNAs have additionally been used as a means to reduce immune-mediated lung pathology [149]. In one model, mice were initially infected with RSV at 1 week of age and were reinfected at 6 weeks of age, after which they exhibited immunopathology (a situation that is not necessarily comparable to that in humans). The administration of IL-4Rα antisense oligonucleotides during the primary RSV infection enhanced the ratio of Th1/Th2 responses and inhibited pulmonary dysfunction.

Taken together, the results described earlier re-emphasize the complexities of vaccine-mediated protection versus immunopathology. There are clearly numerous subtleties that impact the overall outcome of vaccination. One result of the FI-RSV clinical trial is that the development of subunit vaccines or killed vaccines for the pediatric arena is often discouraged [3,5,6]. These vaccines are at some disadvantage because booster(s) may be required to elicit sustained immune activity whereas replication-competent vectors can confer rapid immunity after a single inoculation [59–61]. Rapid induction of protective immunity is clearly desired for infants who are at immediate risk of RSV infection. An additional disadvantage of subunit or killed vaccines is that they do not promote endogenous antigen expression and therefore induce relatively low CTL activity (dependent on antigen cross-presentation as described earlier). Despite these limitations, subunit or killed vaccines also provide advantages. They benefit in that they can be administered by intramuscular injection and do not require cell transfection, transduction or infection in the host. Also, they have succeeded historically in preventing human disease (e.g., hepatitis B virus vaccine). These features suggest that new candidate vaccines of multiple modalities, if capable of inducing robust neutralizing B- and T-cell responses, deserve consideration. An unbiased evaluation of each new candidate vaccine based on overall outcome may help avoid the possible rejection of vaccines that are beneficial, or approval of vaccines that are not.

RSV G-based protein vaccine, BBG2Na

BBG2Na is a fragment of the RSV G protein (residues 130–230) fused to the albumin-binding domain of streptococcal protein G, produced in bacteria and often formulated in an alum-based adjuvant [17,150,151]. In small and large animals, BBG2Na has been shown to elicit an immune response [106,150,152,153]. Initial studies of BBG2Na in RSV seropositive humans were also promising [154], but clinical trials did not progress owing to rare unexpected development of purpura in some vaccinees [155,156]. Efforts are underway to reformulate RSV G protein fragments to improve upon durable immune responses while avoiding adverse events [157].

Purified F protein

There have been numerous attempts to formulate F protein vaccines with adjuvants or as a chimeric protein with RSV G [158–163]. Purified fusion protein (PFP) has been administered in alum to humans, both alone and in combination with infectious RSV vaccines [164]. PFP defines a series of candidate vaccines including PFP-3, an ion-exchange-purified F protein from a cold-adapted temperature-sensitive isolate of RSV A2 with enhanced levels of F expression [165]. Generally, PFP vaccines were well tolerated in seropositive humans and were able to induce neutralizing antibody activities. Nonetheless, the impact of vaccines on lower respiratory tract infections with RSV was not high enough to encourage continued investigation in children, pregnant women or the elderly [160,166–168]. For example, when PFP-2 was administered to expectant mothers, the changes in neutralizing antibody activities in mother and infant were modest and RSV incidence in infants was not significantly altered [168]. Purified F proteins remain a topic of interest, with recent attention paid to the protein’s postfusion structure [169].

RSV F, G and/or M subunit vaccines

Combinations of RSV F, G and M proteins have been clinically tested. In one case, a combination product was tested in the elderly (>65 years age group) with or without alum. The best response was seen with a 100-µg dose in the absence of alum. In this case, more than 50% of individuals experienced a greater than four-fold increase in neutralizing antibody titers [6,170]. The testing of subunit vaccines in seropositive young or elderly populations may possibly underestimate full vaccine potentials and complicate efficacy evaluations but, as stated earlier, there is reluctance in the field to advance subunit or killed vaccine candidates to the seronegative pediatric arena.

Epitope scaffold vaccines

As structure–function technological methods improve, it is now possible to examine the fine detail of immune interactions. For example, crystal structures are currently available of monoclonal RSV-specific neutralizing antibodies complexed with linear RSV peptide targets (e.g., structures exist for monoclonals motavizumab and 101F bound to peptide [171,172]). Information gained from studies of monoclonal B and T cells drives endeavors to isolate a select viral epitope(s) and integrate this artifactually with another protein sequence (carrier or scaffold [6,173–175]) to create vaccines. There are several challenges associated with the epitope-scaffold approach. First, the vaccine should mimic native RSV protein structures as necessary for the induction of cross-reactive RSV-specific antibodies (3D structures can influence antibody response patterns even when the target epitope is ‘linear’ [171,176–178]). Second, the processing of vaccine should yield MHC–peptide complexes matching those of RSV (T-cell receptor recognition of an MHC–peptide complex depends on protein fragmentation that can be affected by peptide flanking sequences and 3D context [78,179–182]). Third, the selected epitope(s) should induce a high-frequency of immune responses in humans with disparate B-/T-cell receptors and MHC haplotypes [52,183].

Nanoparticle vaccines

Nanoparticle vaccines have been a current topic of considerable interest in the vaccine field. Biodegradable synthetic polymers (e.g., poly [d,l-lactic-co-glycolic acid]) can be used to create particles loaded with or chemically conjugated with antigens and adjuvants of choice [184,185]. Nanoparticle vaccines that incorporate antigenic components from RSV have been shown to stimulate mucosal and systemic immune responses in preclinical studies. These studies are in early stages of development [186].

Propiolactone-treated RSV

An inactivated RSV vaccine can be produced via treatment of virus with β-propiolactone, an alkylating agent that reacts with many nucleophilic reagents including nucleic acids. Combinations of inactivated RSV with CpG (a TLR-9 ligand) and MDP (a NOD2 receptor ligand), administered intranasally, induced mucosal IgA and serum IgG, and were capable of protection against RSV in BALB/c mice [187]. As for other killed vaccines, the induction of robust CD8⁺ T-cell responses may prove more difficult than with replication-competent vaccines, and boosters may be desirable. Owing to the FI-RSV vaccine experience, it remains to be seen if another killed vaccine approach will be advanced in the RSV field.

Attenuated RSV

Respiratory syncytial virus can be attenuated by multiple mechanisms including cold passage (cp), mutagenesis and reverse genetics. Previously, the serial cp of RSV has been partnered with chemical mutagenesis to select for attenuated temperature-sensitive (ts) mutants. Reverse genetics technology now lends to this process by facilitating rescue of viable RSV (or other viruses) bearing mutations of choice [188]. A first step in paramyxovirus reverse genetics methodology is to create a plasmid containing the entire viral genome, either with or without mutations. The plasmid is then transfected into a receptive host cell, usually in the presence of helper constructs that heighten viral protein expression. Infectious viral particles rescued from the host cell are expanded for further characterization. The reverse genetics strategy is a useful tool both for the study of mutation phenotypes and for the creation of desired vaccines.

A number of cold-passaged ts (cpts) viruses with mutations in N, F, L and M2 genes have been selected for use as attenuated vaccine candidates. Genome manipulations also include deletions of nonessential genes NS1, NS2, SH or M2-2, or manipulations of gene positions. Challenges are to identify an RSV derivative that is well tolerated in humans, is immunogenic, and replicates sufficiently to support scale-up of manufacturing [6].

One cpts vaccine, cpts248/404, was developed after CP and mutagenesis. Mutations were present in N, F and L genes and in the M gene start signal, but the vaccine caused unacceptable nasal congestion in human infants [5,189]. Reverse genetics was then employed to rescue rA2cp248/404/1030ΔSH with a mutated RSV genome devoid of the SH gene and including a 1030 mutation in the L gene. In early clinical trials, this vaccine was tested at two doses (10^4.3 and 10^5.3 PFU). Following vaccination of RSV seronegative children, peak vaccine titers were 10^2.5 PFU in nasal secretions. Only 44% of infants who were vaccinated twice with cA2cpts 248/404/1030ΔSH at the 10^5.3 dose had detectable antibody responses. Nonetheless, the first dose of vaccine may have impacted the second dose in that replication of the second vaccine dose was restricted [5,190]. This vaccine is now undergoing further testing in a Phase II clinical trial with seronegative infants [191]. Preclinical studies also continue to improve stability of the attenuated phenotype [192].

Another strategy has been to delete the NS2 gene in conjunction with the introduction of point mutations. Two vaccines, rA2cpts 248/404ΔNS2 and rA2cpts530/1009ΔNS2, were developed with this strategy, but in contrast to the cpts 248/404, these were not strongly immunogenic and have not been advanced [5]. Other live-attenuated RSVs include NS1 or M2-2-deletion mutants (RSVΔM2-2 [191]). These have already been tested in nonhuman primates and may advance to human testing [193].

The MEDI-559 clinical trial conducted by MedImmune is currently testing an intranasal, recombinant, live-attenuated, ts RSV in a randomized, double-blind placebo-controlled study to evaluate viral shedding in healthy children from 1 to <24 months of age [6,305].

Merck is developing a live-attenuated vaccine using an RSV287 strain, initially isolated in the 1970s. The virus was subjected to serial in vitro passage in VERO cells yielding mutations in the G, F and L genes and 5′ untranslated region. The vaccine induced an immune response in cotton rats. The response was weaker when cotton rats were administered anti-RSV immune sera to mimic maternal antibodies in infants [194].

One additional attenuated RSV vaccine candidate originated from the clinical RSV isolate 98-25147-X. A virus lacking the G gene was produced by reverse genetics and was found to be nonreplicative in cotton rat lungs. After an intranasal vaccine inoculation, neutralizing antibodies were generated and protective immunity lasted for at least 10 weeks [195,196].

In general, the attenuation of virus is considered an attractive approach and has been highly successful in other vaccine fields. The difficulty experienced over several decades has been to find a balanced RSV vaccine candidate that is sufficiently attenuated to avoid disease and is sufficiently robust to elicit a protective immune response in humans.

RSV envelope-based virosomes

A relatively new vaccine approach has been to present RSV viral envelopes in membranes without nucleocapsids or nucleic acids [197,198]. Tests with a RSV envelope-based virosome vaccine in vitro demonstrated NF-κB activation in macrophages, perhaps owing to TLR triggering by RSV F. When virosomes were reconstituted with RSV envelopes and incorporated the TLR-2 ligand P3CSK4, virus-neutralizing antibodies and IFN-γ-positive T cells were induced in mice. The RSV virosomes were protective against virus challenge in small animals and there were no symptoms of enhanced disease.

Influenza-based virosomes carrying RSV F protein

Another relatively new virosomal vaccine (PEV4) is influenza based and formulated to contain 10 or 50 µg of RSV F protein. Microsomes were injected into BALB/c mice by subcutaneous inoculation twice or three times successively with 3-week intervals. Vaccinated animals generated an anti-RSV response associated with neutralizing activity. The vaccine prevented RSV replication upon challenge. Curiously, empty vectors were also associated with some nonspecific protection, a phenomenon that will be the topic of further study [199].

Bovine RSV or bovine/human RSV chimeric viruses

Bovine RSV vaccines are currently sold for use in cows, but the vaccine has not been protective against human RSV in chimpanzees, even when chimeric viruses were produced with human RSV F and G proteins substituted for bovine RSV F and G. Results were possibly due to the limited growth of bovine RSV in an unnatural host species [5,200].

Newcastle disease virus recombinants or virus-like particles

A recombinant Newcastle disease virus (NDV) was created by reverse genetics to express the RSV F glycoprotein. BALB/c mice vaccinated with this recombinant were protected from RSV challenge, an outcome that correlated with RSV-specific CD8⁺ T-cell activity [5,201]. In a separate study, NDV virus-like particles (VLPs) containing both RSV F and G proteins were prepared by combining NDV nucleocapsid and membrane proteins with chimeric proteins containing the ectodomains of RSV F and G proteins fused respectively to NDV F and HN protein transmembrane and cytoplasmic domains. A single immunization with the nonreplicating VLPs induced a relatively high IgG2a/IgG1 response in mice, and conferred protection against lung infection following an RSV challenge without immunopathology [202]. An additional VLP was developed using NP and M proteins of NDV and a chimeric protein containing the ectodomain of RSV G fused to transmembrane and cytoplamic domains of the NDV HN protein [203]. In this case no F protein was expressed. In a recent experiment, vaccination of mice with 10 or 40 µg total VLP protein by intraperitoneal or intramuscular injection stimulated antibody responses to G protein. One or two doses resulted in modest levels of neutralizing antibodies and the complete protection of mice from RSV replication in the lungs with no enhanced pathology.

Advantages of the VLP vaccine approach are that there are no concerns of virus spread or reversion, and for NDV, pre-existing immunity in humans is rare. NDV vaccines are currently used in veterinary medicine and VLP vaccines have advanced to clinical testing.

Bovine/human PIV-3-recombinants

Chimeric bovine/human parainfluenza virus type 3 (b/hPIV-3)-based vaccines have been constructed by reverse genetics to express RSV F [3,103,107,204]. One of these products, MEDI-534, is currently advancing through clinical trials. In African green monkeys, the vaccine was infectious. A dose of 2–3 × 10⁵ PFU given intranasally and intratracheally reached peak titers in upper respiratory tract samples of >10⁵ PFU/ml and in lower respiratory tract samples of approximately 10⁷ PFU/ml. Significant protection against subsequent RSV challenge was observed [107]. This vaccine has progressed through Phase I clinical trials in the 6–24-month age group. Three doses were tested (10⁴, 10⁵ and 10⁶ tissue culture infectious doses₅₀). There was a statistically significant increase in rhinitis/rhinorrhea in the group that received the 10⁴ vaccine dose compared with controls (p = 0.017), but not at higher vaccine doses, suggesting that the phenomenon may have been coincidental rather than indicative of vaccine-related disease. There was no evidence of enhanced disease following natural exposure to RSV in this study. At the highest vaccine dose, seroconversion to RSV was at the 67% level and seroconversion to PIV-3 was 100%. A large international clinical trial in infants is now underway [205].

Sendai virus-based recombinants

An unmanipulated Sendai virus (SeV; a parainfluenza virus type I of mice) has been studied as a vaccine for human parainfluenza virus type I (hPIV-1), and a SeV recombinant virus carrying the RSV F gene (SeVRSV) has been studied as a vaccine for RSV and hPIV-1. Despite ample opportunity for exposure between humans and mice, there has never been a confirmed clinical case of SeV-related disease in humans. This may be due in part to the unique sensitivity of SeV to human interferon-associated innate immune responses [206]. SeV has been shown to induce B- and T-cell responses in the upper respiratory tract (the point of entry for the target pathogen) and lower respiratory tract within 7–10 days of an intranasal inoculation in small animals. After vaccination, antibody-forming cells and T cells were maintained in diffuse nasal-associated lymphoid tissue for the animal’s lifetime. Coincident with the immune response, SeV and recombinant SeV vaccines conferred rapid and durable protection against their target pathogens [59,60,207–211]. Furthermore, SeVRSV elicited neutralizing antibodies and protection against a variety of RSV A and B isolates. These results have encouraged the conduct of an ongoing Phase I clinical trial of SeV in adults and children, in which the vaccine has been well tolerated to date [212]. In African green monkey studies, immune activities against SeV or SeVRSV were generated without adverse events [213,214]. SeVRSV conferred complete protection against lower respiratory tract infection after RSV challenge. No clinically relevant adverse events were demonstrated after either vaccination or challenge in the large animal model. SeV and SeVRSV were more attenuated in the upper and lower respiratory tracts of primates than the b/hPIV-3-based RSV vaccine (described earlier), which is currently progressing through clinical trials. These results support plans to advance the SeVRSV vaccine product to the clinic [214].

Modified Vaccinia Ankara recombinant vaccine

Modified Vaccinia Ankara (MVA) has been tested as a vector for RSV gene delivery [215]. The strategy has demonstrated immunogenicity in chimpanzees, but the vaccine has not advanced to clinical trials. MVA gained considerable popularity as a smallpox vaccine when smallpox bioterrorism threats arose and when the safety profile of the standard smallpox vaccine (based on unmodified vaccinia virus) came into question [216–218]. Clinical trials with MVA-recombinant vaccines have been conducted in other vaccine fields [219,220], a situation that may prompt future studies with MVA-RSV recombinants. MVA is somewhat difficult to manufacture owing to its attenuated growth, but efforts have been made to perfect the process [221].

DNA plasmid vaccines & DNA in human papillomavirus pseudovirions

DNA plasmid vaccines carrying a number of different RSV genes have been characterized in small and large animal models [222–226]. In addition, DNA immunizations have been administered following encapsidation of RSV nucleic acid sequences by the capsid proteins of human papillomavirus [227]. In one example, self-assembled pseudovirions carrying RSV M and M2 genes were produced. These were delivered intravaginally. Researchers found that immunization of mice with the pseudovirions induced M- and M2-specific CD8⁺ T-cell and antibody responses similar to those induced by a 10,000-fold higher dose of naked DNA. In general, experiences in other vaccine fields have found that despite various formulations and modes of injection [228,229], DNA vaccines do not elicit robust and sustained immune responses when used in isolation. They nonetheless serve as useful priming tools in the context of heterologous vaccination regimens [230,231], and are most often advanced in this capacity.

Adenovirus recombinants

Adenovirus has been considered an attractive candidate vaccine backbone, in part because manufacturing is facilitated by robust virus growth in tissue culture [6]. Challenges to this approach include the natural exposure of humans to adenovirus 5 and consequent B- and T-cell immune responses, which can dampen vaccine efficacy [232,233]. Other adenovirus serotypes (e.g., 14, 26, 28 and 35), chimeric viruses, manipulated viruses or nonhuman primate adenoviruses provide additional options for vaccine design [234,235]. One recent study used a replication-deficient adenovirus (rAd)-based codon-optimized vaccine expressing the soluble F1 fragment of F protein (amino acids 155–524) [236]. In mice, mucosal IgA responses were induced by one intranasal immunization with the vaccine. There was potent protection against subsequent RSV challenge, but neither serum antibodies nor T cells could be detected. When cholera toxin B subunit was used as an adjuvant with the vaccine, IgA responses were further enhanced.

Semliki Forest virus recombinant

Semliki Forest virus is an alphavirus, a small enveloped positive-strand RNA virus. Recombinant technology has been used to create both replication-competent vaccines and vaccines capable of only one round of infection [237,238]. In the latter case, recombinant RNA is packaged into virus particles by cotransfection with packaging-deficient helper RNAs. Resultant virus particles will infect cells and express high levels of recombinant gene products without expression of structural genes. Protein expression is supported by self-amplification of viral replicons and downregulation of host protein synthesis. Antiviral immunity has been elicited in BALB/c mice by a recombinant Semliki Forest virus particle-based vaccine encoding the RSV F and G proteins [239]. Intranasal immunization reduced lung RSV titers upon challenge without disease enhancement.

Venezuelan equine encephalitis virus recombinant

Another alphavirus-based vaccine is a Venezuelan equine encephalitis virus-based replicon particle that instructs expression of RSV F and/or G glycoproteins. Recombinant Venezuelan equine encephalitis vaccines elicit protective humoral and cellular responses in cotton rats and mice [240,241]. Most recently, a nonreplicating virus-based replicon particle carrying a RSV F gene induced antibodies (including mucosal IgA) and CD8⁺ T cells in small animals, even in the context of passively transferred RSV antibodies. The vaccine controlled virus in the lungs and also reduced virus in the upper respiratory tract after RSV challenge. Future goals for the alphavirus-based vaccines are to demonstrate protection in a primate model and to establish methods for scale-up manufacturing [6,240].

Recombinant bacteria

Recombinant Bacillus Calmette-Guérin has been produced to express RSV genes. In mice, recombinant Bacillus Calmette-Guérin induced IFN-γ-producing Th1 cells [242]. In addition, adoptive transfer of both CD4⁺ and CD8⁺ T cells from vaccinated mice contributed to the clearance of RSV RNA, as measured by PCR, after RSV infection of host animals. Other bacteria used to deliver or express RSV genes include Salmonella [243] and Streptococcus [244]. One challenge associated with the expression of RSV genes by bacteria is that the antigens do not exhibit post-translational modifications typical of the mammalian host, and may fail to generate robust CD8⁺ T-cell activities. Vaccines based on bacteria may also pose safety issues and problems with pre-existing immunity.

Plant-based vaccines

Recombinant technology also provides capacity to develop plant-based vaccines. For example, transgenic apples have recently been described that express the human RSV F protein [245]. Specifically, a gene driven by the CaMV 35S promoter was introduced into apple leaf tissues by Agrobacterium-mediated transformation. Two transgenic lines were developed in which the RSV F gene was identified by PCR. Stable integration of the RSV F transgene was confirmed and levels of F protein expression were as high as 20 µg per gram of tissue. Plant-based vaccines are attractive in concept, but also face challenges concerning post-translational modifications atypical of mammalian cells, weak induction of CD8⁺ T cells and oral tolerance.

Heterologous prime–boost vaccine regimens

Prime-boost strategies are now being considered in which two different modes of vaccination are used in succession. Because there are many vaccine candidates and strategies, the number of potential prime–boost regimens is vast. In one study, rhesus monkeys were immunized with RSV F using a DNA recombinant vaccine to prime and a recombinant adenovirus to boost. DNA was administered by tattooing, intramuscular injection, intramuscular electroporation or intradermal electroporation [246]. RSV-neutralizing antibodies and IFN-γ ELISpot responses were identified, particularly when DNA was administered by intramuscular electroporation.

Another complex prime–boost strategy was described by researchers at Emory University using a BALB/c mouse model [247]. First, B cells were activated ex vivo and pulsed with an M2 (82–90) peptide. These cells were infused by intravenous inoculation. Mice were boosted three times with M2 (82–90) peptide, PolyI:C and costimulatory anti-CD40 antibody (TRIVAX). In first studies, researchers found an increase in CD8⁺ T cells responsive to the peptide and some protection in mice from RSV in the lungs postchallenge. The strategy in which cells are removed from an individual and reinfused after in vitro manipulations has been previously used in clinical research to treat cancer and HIV. Difficulties relate to cost, logistics, risks associated with in vitro manipulations and proof of efficacy [248–251]. There are also risks associated with antigen-nonspecific activation or killing of immune populations when antibodies specific for lymphocyte membrane proteins are used [252–254,306].

An additional prime–boost protocol has utilized an attenuated Salmonella typhimurium recombinant to deliver DNA followed by a recombinant rAd vector (somewhat similar to the DNA/adenovirus constructs described earlier) [255]. Specifically, researchers used an oral SL7207/pcDNA3.1/F prime and intranasal FGAd/F boost to generate immune responses to RSV in mice. The regimen generated stronger RSV-specific humoral and mucosal immune responses than the oral SL7207/pcDNA3.1/F regimen alone, and stronger cellular immune responses than the FGAd/F regimen alone. Upon RSV challenge, the prime–boost regimen showed increased efficacy based on histopathological analyses.

The strategy of prime–boost with heterologous carriers is attractive in theory in that any immune response elicited toward a first delivery vehicle during the primary immunization need not inhibit the ‘take’ of the booster immunization [231]. Nonetheless, the increased complexity of the vaccine may cause logistical difficulties at the manufacturing stage, a situation that could ultimately discourage advancement of products.

Expert commentary & five-year view

The RSV field has gained a great deal of experience and knowledge upon which successful vaccine design may be based. The immune system comprises complex B- and T-cell populations that together combat invading pathogens. Vaccines that activate all three populations of RSV-specific lymphocytes (B cells, CD4⁺ T cells and CD8⁺ T cells) might best harness the full strength of the immune system. When evaluating new vaccines, researchers must acknowledge the vast complexity of the immune system. The precise environment in which cells are found may influence final outcome, explaining the many discrepancies in RSV literature. As of yet, there is not sufficient scientific evidence to base acceptance or rejection of a vaccine candidate on a single cytokine or cellular measurement. Rather, comprehensive evaluations of outcome following vaccination may serve to select beneficial products and increase our understanding of protective immune correlates.

One of the difficulties of advancing any RSV vaccine candidate is that infants comprise an important target population, and the dire consequences of testing an ineffective RSV vaccine in infants has already been realized. When vaccines are tested in older persons who are already seropositive for RSV, data may be difficult to interpret and results may underestimate full vaccine potential. Vaccine opponents argue that infants usually recover from RSV infection (particularly in the USA) and that the risk of vaccine development is too high. There is also the logistical difficulty of coordinating RSV vaccination with the already-complex schedule of routine pediatric vaccinations.

Another concern relates to maternal antibodies which may impede vaccination in the newborn [256]. At birth, maternal antibody titers can be equal to, or more than titers in adults. However, as described previously, maternal antibodies fall substantially by 2 months after birth [14]. At this time, infants with the lowest antibody levels and greatest vulnerability to RSV infection are most likely to respond positively to an RSV vaccine.

Yet another concern among researchers is that the developing immune system may not respond to vaccination. Fortunately, the human immune system is more advanced at birth than that of some other mammalian species (e.g., mouse) and human infants are already known to respond positively to licensed pediatric vaccines [257,258].

RSV vaccines can be tested in the elderly, but this strategy also poses difficulties [259]. As with any adult RSV vaccine study, preexisting immune responses can hinder efficacy and complicate data interpretation [6]. Additionally, it may be difficult to assess the positive impact of a vaccine when unrelated abnormalities or infections mask the vaccine’s protective effects [19,167,260]. For these reasons it is proposed that clinical vaccine trials in elderly populations must be designed with very large group sizes to distinguish between vaccine-related and vaccine-unrelated outcomes [261].

Despite the difficulties described earlier, a number of pharmaceutical companies (e.g., Novartis, Sanofi Pasteur, Novavax, GenVec and Crucell) and governments have recently asserted that the development of an RSV vaccine is of high priority. These and other initiatives toward preclinical and clinical development of novel and classical vaccine approaches lend optimism to the field. There are probably a number of vaccines that will ultimately prove safe and efficacious in infants and adults. One may gain much information from preclinical research using tissue culture and animal models. Cautious, steady advancement toward clinical research may then expedite vaccine discovery.

A successful RSV vaccine may well be developed within the next 5–10 years if preclinical and clinical trials progress efficiently and if candidate vaccines are tested in parallel. Clinical research may expedite vaccine discovery, illustrate true correlates of protection, and prevent the devastating disease caused by RSV worldwide.

Key issues.

• Respiratory syncytial virus (RSV) is a catastrophic disease of children.

• The B cells and T cells of the immune system have an extraordinary capacity to ward off serious respiratory pathogens. To date, vaccines that activate only one of these populations (e.g., by preserving T-cell epitopes, but omitting or distorting neutralizing B-cell epitopes) have not been successful. Experiences in other vaccine fields encourage the development of vaccines that harness antibody activity as a first line of defense to destroy free pathogens, and harness T cells to recognize processed antigens and kill infected cells.

• Owing to the intricacies of antigen processing/presentation and immune repertoires, each individual may respond differently to a given pathogen. Furthermore, a peptide that is immunodominant in the pathogen’s context need not be immunodominant when placed elsewhere. While advanced molecular techniques tempt antigen scrambling, past vaccine successes encourage preservation of antigenic structures and contexts to ensure the induction of pathogen-specific immune activities in the majority of humans.

• There are a multitude of strategies available to the field, several of which may yield a successful RSV vaccine. Most strategies have not yet been advanced to testing in the seronegative pediatric arena. Current enthusiasm for RSV vaccine development in academia and industry may prompt the implementation of new pediatric vaccine clinical trials and giant steps toward the prevention of RSV disease.