Respiratory syncytial virus (RSV) is a highly prevalent childhood pathogen and represents a common cause of infant hospitalization. Attempts to develop an RSV vaccine have so far floundered owing to issues with stability, purity, reproducibility and potency. Now, writing in Science, McLellan and colleagues report the structure-based design of a novel RSV vaccine, which shows a high level of RSV-specific neutralizing activity in mice and rhesus macaques.

Credit: MedicImage/Alamy

Severe disease caused by RSV in at-risk infants may be prevented by using the marketed prophylactic palivizumab, which is a monoclonal antibody that inhibits an epitope at the antigenic site II present on both the pre-fusion and post-fusion conformation of the RSV fusion (F) glycoprotein (a trimeric glycoprotein used by the virus to enter host cells via membrane fusion). However, the passive immunization provided by this antibody does not last from season to season and it is very costly.

More recently, RSV-neutralizing antibodies elicited by natural infection (such as D25) have been found to be substantially more potent than palivizumab. The target of these antibodies is the antigenic site zero (Ø), which is a metastable site located at the membrane-distal apex of the pre-fusion RSV F glycoprotein.

With this in mind, McLellan and colleagues set out to develop a vaccine that induced neutralizing antibodies against pre-fusion RSV F. Using information gleaned from a previous analysis of the structure of the antigenic site Ø in complex with the D25 antibody, they engineered antigens based on variants of RSV F exhibiting a stably exposed antigenic site Ø. This involved appending a carboxy-terminal trimerization domain combined with other means of stabilization including the introduction of cysteine pairs and cavity-filling hydrophobic substitutions.

In all, more than 100 RSV F variants were constructed, 3 of which — DS, Cav1 and TriC — were selected for further crystallographic and antigenic characterization owing to their stability when exposed to extremes of pH, osmolality and temperature. Crystallographic analysis of individual — or a combination of — variants provided atomic-level details regarding the specific structural features responsible for their stability.

Next, the authors assessed the immunogenicity of their RSV F variants. Mice and rhesus macaques were injected with post-fusion forms of RSV F or RSV F variants, combined with the poly-ICLC adjuvant, and the authors measured the ability of sera to prevent RSV infection in the human tumour cell line HEp-2.

In mice, DS, Cav1 and TriC each elicited high titres of neutralizing activity, which was ∼4 times greater than that elicited by post-fusion RSV F and ∼20 times greater than the protective threshold, whereas the DS–Cav1 combination elicited neutralizing activity roughly 8 times higher than that of post-fusion F and 40 times higher than the protective threshold. Similarly, in rhesus macaques, DS–Cav1 injection resulted in neutralizing activity that was 70 times and 80 times greater than that of the post-fusion F protein against RSV subtype A and subtype B, respectively, by 8 weeks.

Together, these studies demonstrate the concept of structure-based vaccine design and highlight the potential of pre-fusion RSV F as a promising target for the development of an effective RSV vaccine.