Marker vaccine potential of a foot-and-mouth disease virus with a partial VP1 G-H loop deletion

Abstract

Previous work in cattle and pigs demonstrated that protection against foot-and-mouth disease (FMD) could be achieved following vaccination with chimeric foot-and-mouth disease virus (FMDV) vaccines, in which the VP1 G-H loop had been substituted with that from another serotype. This indicated that the VP1 G-H loop may not be essential for the protection of natural hosts against FMDV. If this could be substantiated there would be potential to develop FMD marker vaccines, characterised by the absence of this region. Here, we investigate the serological responses to vaccination with a virus with a partial VP1 G-H loop deletion in order to determine the likelihood of achieving protection and the potential of this virus as a marker vaccine. Inactivated, oil adjuvanted, vaccines, consisting of chemically inactivated virus with or without a partially deleted VP1 G-H loop, were used to immunise cattle. Serum was collected on days 0, 7, 14 and 21 and antibody titres calculated using the virus neutralisation test (VNT) to estimate the likelihood of protection.

We predict a good likelihood that cattle vaccinated with a vaccine characterised by a partial VP1 G-H loop would be protected against challenge with the same virus containing the VP1 G-H loop. We also present evidence on the potential of such a construct to act as a marker vaccine, when used in conjunction with a novel serological test.

Introduction

Current foot-and-mouth disease (FMD) vaccines consist of chemically inactivated whole virus antigen that are formulated with either aluminium hydroxide/saponin or mineral oil adjuvant, depending on the target species [1]. Although these vaccines are capable of protecting animals from clinical disease they do not confer sterile immunity. The possibility of undisclosed infection in vaccinated animals necessitates methods to identify this and these rely on serological tests that can differentiate the immune response elicited by vaccination from that due to infection. Currently, this is achieved by purifying the vaccine antigen to remove FMD virus (FMDV) non-structural proteins (NSP) and then using detection of NSP antibodies as an indicator of infection [2].

However, vaccine preparations, depending on their source, can contain traces of NSP, reducing the specificity of the NSP assays [2]. Additionally, some vaccinated animals exposed to infection can become asymptomatic carriers, without an associated NSP seroconversion [3]. Therefore, there is a need for an additional and more reliable means of discriminating vaccinated and infected animals.

Foot-and-mouth disease virus shares many structural features with that of other picornaviruses, for example the structural proteins VP1, VP2 and VP3 have their β-barrel cores at essentially the same radius and orientation. However, unlike rhino- and enteroviruses, which have a ‘canyon’ or pit to prevent antibodies binding to their receptor binding site, FMDV has a relatively smooth surface with a prominent loop structure protruding from the capsid protein VP1, referred to as the G-H loop. The loop possesses an RGD binding site for attachment of the virus to integrin receptor molecules on the surface of susceptible cells [4]. Although the VP1 G-H loop has been regarded as an immunodominant antigenic site (site 1) on the viral capsid surface, there is considerable evidence to suggest that other antigenic sites are important in eliciting antibodies and protection against FMDV, not least that: (i) G-H loop peptide vaccines perform poorly in protecting target species such as cattle [5], (ii) pigs vaccinated with a chimeric vaccine virus possessing a serotype A backbone and a serotype C VP1 G-H loop were protected from challenge with serotype A virus but only partially protected from challenge with serotype C virus [6], (iii) cattle vaccinated with a virus which differed at sites other than the VP1 G-H loop from the challenge virus were also not protected from challenge [7], (iv) the proportion of antibody directed towards the VP1 G-H loop varies substantially in convalescent or vaccinated sera [8], [9], (v) competition of sera from the three main target species with monoclonal antibodies (MAbs) demonstrated that no one antigenic site (1, 2 and 3) could be considered immunodominant [10], (vi) MAbs raised against serotype O virus are often to site 2 [11] and (vii) MAbs to conformational sites outside the VP1 G-H loop are more efficient at opsonising virus and protecting mice than those generated to the VP1 G-H loop [12].

Overall, the role and importance of the VP1 G-H loop in induction of protective immunity in target species is still not fully understood. A recent study which experimentally substituted the VP1 G-H loop with 10 glycine residues, Frimann et al. [13] showed that the removal of this dominant B cell epitope can dramatically enhance the immune response to less dominant B cell epitopes leading to broader cross-reactivity within and between serotypes. This could be advantageous in the development of negatively marked FMDV vaccines which are characterised by the partial or complete absence of the VP1 G-H loop.

This paper describes detailed comparisons of the antibody responses to two plaque purified virus variants discovered within a single vaccine strain, one containing an unmodified VP1 G-H loop and one containing a 13 amino acid deletion within the VP1 G-H loop. The antibodies generated were tested for reactivity against a selection of serotype A field isolates to determine whether a vaccine lacking a significant portion of the VP1 G-H loop is likely to (i) protect against challenge with a virus which includes the wild type VP1 G-H loop and (ii) be sufficiently discriminatory to suggest that an FMD vaccine lacking the VP1 G-H loop could act as an efficient marker vaccine.