Respiratory syncytial virus is a negative strand RNA virus in the family Paramyxoviridae, the subfamily Pneumovirinae and the genus Pneumovirus. RSV was first discovered in 1952 as the cause of a serious lower respiratory tract disease, most pronounced among children in their first year of life [1]. Other individuals who are susceptible to severe RSV disease include patients with cardiac and pulmonary disorders, patients with immunodeficiencies, and the elderly [2,3,4,5,6]. RSV infection generally presents as an upper respiratory tract (URT) infection which progresses for several days before virus traffics to the lung [1,7]. Most infants are able to clear virus without extreme adverse events, but in the United States approximately 2% require hospitalization [8,9]. Globally, RSV infections are estimated to result in up to 199,000 deaths annually in children younger than 5 years of age [10] and treatment options remain controversial. Re-infections can occur, but generally result in milder disease; risks of infection are inversely associated with RSV-specific serum neutralizing antibody titers [11,12]. For infants at risk for first infections, the passive transfer of monoclonal RSV F-specific antibodies is recommended, but this type of prophylaxis is expensive and unavailable for most individuals who need it [13]. A licensed vaccine, the single best health care solution to infectious disease, is unavailable in the RSV field.

The RSV genome consists of 11 coding sequences including NS1, NS2, N, P, M, SH, G, F, M2-1, M2-2 and L [1,14]. The predominant proteins on the outer membrane of the mature virus are G, F and SH. Initiation of infection is usually mediated by the interaction between RSV G, the major attachment glycoprotein, and the host cell membrane. The viral receptors for G (and F) proteins on mammalian cells are not fully defined, but G can bind highly sulfated heparin-like glycosaminoglycans [15] as well as the CX3CR1 fracktalkine chemokine receptor [16]. G attachment facilitates F-mediated membrane fusion, although F protein can bind cell membranes independently (e.g., via binding to cellular heparan sulfate and nucleolin [16,17]) and mediate virus infection in the absence of G [1,18]. Fusion occurs when the F1 subunit of F protein is inserted into the target membrane and the protein refolds into a hairpin structure bringing virus and cell membranes into close proximity [19]. Upon membrane fusion, virus material including a single negative strand RNA genome is released into the cell cytoplasm [19]. Viral RNA is then transcribed and replicated yielding viral mRNA and new virus genome. New virions are assembled at the cell membrane, bud using cell membranes as their outer coat, and then target new cells for additional rounds of infection. Non-structural proteins support virus production while down-regulating host cell growth and defense mechanisms [1].

A main hurdle for RSV vaccine development was encountered in the 1960s during the testing of a formalin-inactivated RSV vaccine (FI-RSV). Unfortunately, the vaccinated infants were not protected from RSV infection. Instead, an unusually large percentage of the vaccinated children, when subsequently exposed to RSV by natural causes, developed disease severe enough to require hospitalization and there were two vaccine-related deaths [20]. It is essential that such a scenario never be repeated, and because the reason for the vaccine-related adverse events has not been confirmed, scientists are unsure precisely how to proceed. There remains no licensed vaccine product after one-half century of additional research.

The cause of vaccine-related deaths remains a point of discussion. One potential explanation is as follows: The formalin treatment of RSV altered membrane proteins on the virus and in so doing, rendered a vaccine that induced antibody responses that were non-neutralizing [21,22]. The inactivated vaccine also failed to elicit robust CD8+ T cells, because these classical killers of virus-infected targets are best induced by endogenously expressed viral antigens. In the absence of robust neutralizing antibodies and CD8+ T cells, RSV persisted and induced an aggressive CD4+ T cell and cytokine response in the lower respiratory tract [23,24]. Uncontrolled RSV replication and persistent inflammation led to blockage of the small airways in infants, leading to substantial morbidity and mortality [1,4,14,25].

There is now debate as to which immune cell populations or effector molecules were responsible for disease [21,26,27,28,29,30,31,32,33]. In a murine model designed to recapitulate the clinical outcome with FI-RSV, it was demonstrated that a subset of RSV G-specific CD4+ Th2 cells and eosinophils associated with disease. [23,24,31,32,34,35]. Additional mouse experiments showed that inhibition of IL-4 and IL-10 Th2 cytokines abrogated pulmonary histopatholgy [24]. Some researchers argue that the induction of Th2 cells, or perhaps the induction of any RSV-specific T cells should be avoided [36], even though T cells may be key providers of help for B cells and cytotoxic T lymphocyte functions at the time of RSV exposure. Others argue that the large majority of granulocytes described in original clinical autopsy reports were neutrophils rather than eosinophils [37], questioning the absolute comparison between mouse and human responses to FI-RSV. The potential benefits provided under the appropriate conditions by Th2 cells, other CD4+ and CD8+ T cell subsets, and eosinophil populations are contemplated [37,38,39,40,41,42]. For example, it is noted that individuals who lack T cells can experience great difficulty in RSV clearance and suffer worse outcomes than their immunocompetent counterparts [4,14,43,44]. It is further noted that eosinophils can be beneficial in that their transfer to the lungs of RSV-infected mice can enhance RSV clearance and inhibit airway hyper-reactivity [45]. Until debates are resolved, researchers and regulatory boards struggle to define 'go' and 'no go' criteria for the advancement of candidate vaccines, particularly when clinical studies target the pediatric arena. It is possible that past RSV vaccine candidates may have proved safe and efficacious in children, but were never tested in seronegative, pediatric populations. Perhaps a solution resides in comprehensive data analyses, revealing the complexities of the immune system and that each lymphocyte subset need not be categorically designated as beneficial or injurious [46]. Rather, as is the case in most RSV-experienced adults, a variety of adaptive and innate immune effectors including B cells, CD4+ T helper cells and CD8+ cytotoxic T lymphocytes function synergistically to mediate safe elimination of virus and virus-infected cells.

Another issue pertinent to RSV vaccine development concerns the selection of an appropriate human test population. Vaccine studies in young infants are performed with hesitancy due to the devastating experience with the FI-RSV vaccine. One suggested possibility is that vaccines might be tested in seropositive adults including the elderly rather than in infants. A number of strategies are considered including: (i) standard vaccination of placebo and control groups with subsequent assessment of disease caused by natural RSV exposure, (ii) the use of an experimental RSV challenge virus [47], and (iii) the vaccination of pregnant females to measure protection afforded to the infant at birth. One or more of these strategies may prove fruitful provided that attention is paid to a number of confounding variables. First, due to varying degrees of seropositivity in older individuals, immune responsiveness toward vaccination may be difficult to interpret. Individuals with high pre-existing immunity might be expected to show a boost in antibody titers following vaccination, but may instead clear vectors and antigens so rapidly that there is little opportunity for immune cells to re-activate. Second, vaccine safety data may be difficult to interpret due to frequent, unrelated disease complications in the oldest adults. One additional variable to be considered (upon design of either pediatric or adult vaccine protocols) is that individuals with poor diets or poor metabolism (e.g., vitamin deficiency) may exhibit general defects in immune responses toward respiratory virus vaccines [48,49]. With these considerations in mind, it has been proposed that RSV vaccine studies in the elderly are feasible, but would require the recruitment of thousands of participants to ensure fair vaccine assessment [50]. Another recent proposal has been that RSV vaccine studies might best be conducted not in the youngest infants or in adults, but in seronegative children who are at least 6 months of age [51]. Perhaps efficacy studies in older, seronegative children could assist vaccine licensure for a restricted age group while prompting additional clinical studies in the youngest infants.

The first formal vaccine development effort occurred two centuries ago when Edward Jenner demonstrated that material from a cowpox lesion could serve as a vaccine for smallpox [52]. Successful vaccination, both then and now, relies on the safe introduction of a pathogen's antigenic determinants to the immune system. If the vaccine’s antigenic determinants are well matched to those of the native pathogen (as was the case for cowpox and smallpox), the antigens will activate pathogen-specific lymphocytes. In the case of virus-specific B cells, some effectors will mature to the plasma cell stage and constitutively secrete antibodies into blood, lymph and mucosal secretions, while others will maintain memory status, capable of immediate re-activation upon pathogen exposure. Activated CD8+ T cells are classically known for their killing of virus infected cells, while CD4+ T cells, (including Th1, Th2, Th17, TFh, Treg, and other subsets) are known for their support and regulation of CD8+ and B cell activities [53,54,55,56,57]. The heightened or “primed” state of immune surveillance may last months or years to inhibit pathogen entry and pathogen-mediated damage to the host [58]. Generally, vaccines are developed by: (i) inactivating the virus, (ii) identifying a related virus in another species that is safe in humans (the Jennerian approach), (iii) attenuating the virus, or (iv) using recombinant technology to present viral antigens, often in the context of a replication competent or replication-incompetent vector. In the RSV vaccine field, because of the outcome of the FI-RSV vaccine study, the first approach is generally discouraged, even if candidate vaccines retain RSV neutralizing determinants. The second strategy, the Jennerian approach, has not been advanced due to an insufficient antigenic match between bovine and human RSV [59]. The third and fourth strategies are the topics of most current vaccine research, described in greater detail below. Preclinical testing of vaccine candidates is generally accomplished first in small animals and then in non-human primates. Most small animal studies use hamsters, BALB/c mice, or cotton rats [60,61,62,63] while non-human primate studies generally use African green monkeys, rhesus macaques and chimpanzees [64,65,66]. Each animal is at least semi-permissive for RSV infection and is therefore advantageous in that immune responses and immunopathological events can be measured, but no animal model fully predicts the course of immune responses and disease in humans. In one instance, for example, an attenuated RSV vaccine appeared to be safe in non-human primates, but proved unsafe when tested in seronegative children [67]. Again, the debates described above concerning: (i) interpretation of data from animal experiments, and (ii) selection of clinical trial target populations, must be considered to determine when and how to advance vaccine candidates from pre-clinical to clinical trials.